i

Regional Economic Development

and History

Regional Studies is inextricably intertwined with history. Cultural and institu-

tional legacies inform choices between different policy options, meaning that
the past plays a crucial role in how we think about regional economic develop-
ment, planning and policy.
Through a selection of accessible theoretical, methodological and empirical
chapters, this book explores the connections between regional development
and history. Drawing on the expertise of scholars in several disciplines, it links
history to topics such as behavioural geography, interdependence, divergence
and regional and urban policy.
This innovative book will be of interest to researchers across regional studies,
planning, economic geography and economic history.

Marijn Molema is a historian and working as programme leader at the Frisian

Institute of Social Research. He is also a guest researcher at the Fryske Akademy,
a research institute in the city of Leeuwarden of the Royal Netherlands Academy
of Arts and Sciences.

Sara Svensson is a political scientist specializing in regional collaboration across

national borders. She is Senior Lecturer at Halmstad University in Sweden
and Research Fellow at the Center for Policy Studies at Central European
University in Budapest (Hungary).
ii

Regions and Cities
Series Editor in Chief
Joan Fitzgerald, Northeastern University, USA
Editors
Ron Martin, University of Cambridge, UK
Maryann Feldman, University of North Carolina, USA
Gernot Grabher, HafenCity University Hamburg, Germany
Kieran P. Donaghy, Cornell University, USA
In today’s globalised, knowledge-​ driven and networked world, regions and cities have
assumed heightened significance as the interconnected nodes of economic, social and cultural
production, and as sites of new modes of economic and territorial governance and policy
[Link] book series brings together incisive and critically engaged international
and interdisciplinary research on this resurgence of regions and cities, and should be of interest
to geographers, economists, sociologists, political scientists and cultural scholars, as well as to
policy-​makers involved in regional and urban development.
For more information on the Regional Studies Association visit [Link]
There is a 30% discount available to RSA members on books in the Regions and Cities
series, and other subject related Taylor and Francis books and e-​books including Routledge
titles. To order just e-​mail Emilia Falcone, [Link]@[Link], or phone on +44 (0)
20 3377 3369 and declare your RSA [Link] can also visit the series page at www.
[Link]/​Regions-​and-​Cities/​book-​series/​RSA and use the discount code: RSA0901

138. Metropolitan Economic Development

The Political Economy of Urbanisation in Mexico
Alejandra Trejo  Nieto

139. Regional Economic Development and History

Edited by Marijn Molema and Sara Svensson

For more information about this series, please visit:

[Link]/​Regions-​and-​Cities/​book-​series/​RSA
 iii

Regional Economic
Development and History

Edited by Marijn Molema

and Sara Svensson
iv

First published 2020
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
52 Vanderbilt Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2020 selection and editorial matter, Marijn Molema and Sara Svensson;
individual chapters, the contributors
The right of Marijn Molema and Sara Svensson to be identified as the authors of the
editorial material, and of the authors for their individual chapters, has been asserted
in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to infringe.
British Library Cataloguing-​in-​Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-​in-​Publication Data
A catalog record has been requested for this book
ISBN: 978-​1-​138-​33413-​7  (hbk)
ISBN: 978-​0-​429-​44554-​5  (ebk)
Typeset in Bembo
by Newgen Publishing UK
 v

Contents

List of contributors  vii

Acknowledgements  ix

Prolegomenon  1

Bringing the past back in: Taking history seriously in the

study of regional development  3
RO N M ART I N, PE TE R SUNLE Y AND E MI L E VE NH U IS

Introduction  9

1 The importance of history for regional economic development  11

S A RA S V E N S S ON AND MARI JN   MO LE MA

PART I
Disciplinary and theoretical explorations  19

2 Regional development, history and the institutional lens  21

M A RI JN M O L E MA AND JO HN TO MANE Y

3 Behavioural economic geography and regional history:

Explaining uneven development from a human perspective  36
RO B E RT H U G G I NS AND PI E RS THO MPSO N
vi

vi Contents
PART II
Innovations in research design and methodology  61

4 An interdisciplinary approach to the persistent effects of

Polish partitions on educational achievements  63
JU STY N A KO Ś CI Ń SK A AND MI KO ŁAJ  HE RB ST

5 Regional GDP before GDP: A methodological survey

of historical regional accounts  82
K E R STI N E NFLO AND ANNA MI SSI AI A

6 Comparative research designs: Interdependence as challenge

and opportunity in regional studies  98
M ART I N ÅB ERG AND THO MAS  D E NK

PART III
Empirical case studies  117

7 Catching the ladder: The formation and growth of the

São Paulo automotive industry cluster  119
TO M ÀS F E R NÁND E Z- ​D E -​S E VI LLA AND ARMA N DO J. DA L L A  COSTA

8 Urban and regional development policy: Its history and its

differences  145
K E V I N C OX

9 Spatial-​economic development: The effect of urbanization

on education in China, 1890–​present  164
M E I M E I WA NG AND BAS VAN LE E UWE N

Conclusions  183

10 Setting an agenda for a “New Regional History”  185

M ARI JN  M O LE MA

Index  197
 vii

Contributors

Martin Åberg is Professor of History at Karlstad University (Sweden), and

specializes in 19th and 20th century political organization and mobilization
in a comparative perspective.
Kevin Cox is Distinguished Emeritus Professor of Geography at the Ohio
State University (US).
Armando J. dalla Costa is an economist at the Federal University of Paraná
(Brazil) and Research Productivity Scholar at the Brazilian National Council
for Scientific and Technological Development (CNPq).
Thomas Denk is Professor of Political Science at Örebro University (Sweden).
His research focus is on comparative interdependence, state formation, dem-
ocratization and political culture.
Kerstin Enflo is Associate Professor at the Department of Economic History
at Lund University (Sweden) and is a leading expert on long-​term regional
development.
Emil Evenhuis is a researcher in the Department of Urbanisation and Transport
at the Netherlands Environmental Assessment Agency (Planbureau voor de
Leefomgeving, Netherlands).
Tomàs Fernández-​de-​Sevilla was the Kurgan van-​Hentenryk Postdoctoral
Fellow in Business History at the Université Libre de Bruxelles (Belgium)
from 2015 to 2018.
Mikołaj Herbst is Assistant Professor at the University of Warsaw, the
EUROREG Centre for European Regional and Local Studies (Poland),
specializing in regional studies and the economics of education.
Robert Huggins is Professor of Economic Geography at the School of
Geography and Planning at Cardiff University (UK).
Justyna Kościńska is a PhD candidate at the Institute of Sociology, University
of Warsaw (Poland), working on the spatial accessibility of educational ser-
vices and urban sociology.
viii

viii Contributors
Ron Martin is Professor of Economic Geography at University of Cambridge
(UK) and President of the Regional Studies Association.
Anna Missiaia is a post-​doctoral researcher at the Department of Economic
History of Lund University (Sweden) and her main research interest is on
regional industrialization and regional development in the long-​run.
Marijn Molema is a historian and working as programme leader at the Frisian
Institute of Social Research, Leeuwarden (Netherlands).
Peter Sunley is Professor of Economic Geography at Southampton
University (UK).
Sara Svensson is a political scientist specializing in regional collaboration
across national borders. She is Senior Lecturer at Halmstad University in
Sweden and Research Fellow at the Center for Policy Studies at Central
European University in Budapest (Hungary).
John Tomaney is Professor of Urban and Regional Planning in the Barlett
School of Planning, University College London (UK).
Piers Thompson is Associate Professor of local and small business economics
at Nottingham Business School, Nottingham Trent University (UK).
Bas van Leeuwen is an economic historian working on global history at the
International Institute of Social History, a research institute of the Royal
Netherlands Academy of Arts and Sciences.
Meimei Wang is a PhD student working on Chinese economic and social
­history at Utrecht University (Netherlands).
 ix
newgenprepdf

Acknowledgements

This volume could not have been produced without the Research Network
Regional Economic and Policy History (ReHi-​Network), supported by the
Regional Studies Association. We want to thank the RSA and their dedicated
staff members Klara Sobekova and Daniela Carl, among others, for their
involvement and interest. Conversations and cooperation with the co-​founders
of the ReHi-​Network, Arno van der Zwet (University of the West of Scotland),
Martin Åberg and Silke Reeploeg (Karlstad University), helped to shape this
volume as well. John Tomaney (University College London) was a valuable
advisor in the early stage of the network, and hosted its inaugural meeting in
April 2017. Lilla Jakobs at the Center for Policy Studies at Central European
University gave valuable advice on the practicalities of producing a book.
The Center as well as the Fryske Akademy also provided intellectual stimu-
lating environments for the duration of the work on this book. Lisa Lavelle
and Natalie Tomlinson at Taylor & Francis have been helpful throughout.
Ruth-Anne Hurst did a good job as production [Link] book was edited by
a historian and a social scientist (political science).We are grateful to each other
for the additional insights brought together by this cooperation, all of which we
will take with us in our further research work.
The editors, 17 September 2019
x
 1

Prolegomenon
2
 3


Bringing the past back in
Taking history seriously in the study
of regional development
Ron Martin, Peter Sunley and Emil Evenhuis

We want to start in the past by recalling Sidney Pollard’s (1981) classic regional
economic history Peaceful Conquest that explained how industrialization spread
across European regions. As it did so, industrialization was influenced by timing
and by the pre-​existing conditions and policies in each region and state, but it was
also reinforced and invigorated by regional processes. Just as regional economies
were formed by industrialization, so they in turn shaped the course of indus-
trialization itself. The current phase of capitalist development, characterized by
a deepening crisis of carbon-​based industrial production, the disruptive effects
of new economic and technological paths, and the faltering of globalization,
involves a different set of creative and destructive forces, and associated policy
problems. But these transformations and challenges are also profoundly regional
in nature and impact, involving new relationships and interactions between
regions within and between nations. As many of the contributors to this timely
collection of essays argue, a regional-​historical approach is no less essential to
understanding contemporary economic and societal change as it is to expli-
cating previous phases of capitalist development.
How then might history fruitfully contribute to a better understanding of
regional socio-​ economic development? And indeed, how might the study of
regional socio-​economic development inform not just the practice of historical
enquiry but also our understanding of the process of social economic evolution and
progress? Not merely by knowing more about particular places at particular times
under particular circumstances, and then adding it to the discussion –​though that
would certainly be useful. More fundamentally, it can be argued that by adopting
a historical perspective –​or perhaps better termed a historical epistemology –​in
which the aim is to reveal how at any particular time an observed pattern of regional
socio-​economic differentiation has come to be what it is, we not only improve
our understanding of regional development as an inherently historical process, but
also derive a clearer view of how far and in what ways policy interventions might
help to steer future regional development along some more favourable path. In
this respect, policy-​makers would then need to be more realistic about the way in
which policies mix into the flow of an economy’s spatial history and not simply
imagine policies will achieve the “experimental” results they wish for them.
4

4  Ron Martin, Peter Sunley and Emil Evenhuis

No doubt several alternative views might be proffered as to how we should
go about taking history seriously in regional development studies. But a fruitful
starting point is to turn to Charles Tilly’s (1984) brilliant exegesis on the uses
of history, on the critical role of time in affecting development, and on the
value of time and place comparisons in arriving at plausible explanations.
He distinguishes four main research strategies, or “ways of seeing”. We might
for example adopt an “individualizing” approach, and focus on the historical
processes, events, path dependencies, reactive sequences and key decisions that
explain a given region’s development as a means of grasping the peculiarities of
that case. Such an approach might, for example, help to explain the divergent
development of, say, two regions or cities in terms of, for example, their different
and distinctive structural, institutional and governance set-​ups. Secondly, we
might adopt a “universalizing” approach in which comparison is made across
several regions in an effort to reveal how regional socio-​economic develop-
ment over time is driven by some key common factors, forces and processes,
regardless of the specificities of local context. One such common force fre-
quently referred to in the literature is the tendency for regions to specialize to
some degree in one or other particular economic activities in which they have
some natural, contingent or constructed comparative advantage. Essentially, the
aim here is to search for what is “universal” in the “particular”, with a view to
formulating or confirming some generalizations about the historical processes
and forces of regional socio-​economic development.
A third strategy is that of “variation-​finding” comparison. Here the aim is to
establish some key principles of variation in the character or pace of regional
development by examining systematic differences among the histories of
different regions. It might be, for example, that there are systematic differences
across regions in their social-​political class histories, including their propensity
to trade unionism, which in turn have impacted on their economic evolution.
Fourth, and most ambitiously, we might embark on an “encompassing” com-
parison approach. This seeks to place the development experiences of different
regions within a larger system on the way to explaining those development
histories in terms of the varying relationship of the regions to the larger system
as a whole. Can we explain the diversity of regional development experiences
over time as deriving from their (changing) roles and interdependencies within,
say, the overall national economic system to which they belong? Or even more
broadly, in relation to the international system? To what extent have the different
regional paths of socio-​economic development been shaped by big historical
processes of capitalist progress and accumulation? Encompassing comparisons
typically begin with a large structure or process and then explain similarities or
differences among cases –​in our instance, regions –​as consequences of their
relationships to the whole. Such an approach can yield invaluable insights. But it
also carries two potential dangers: first, the presumption of a pre-​given “whole”
that “governs” its parts (regions), and second, a functionalist trap, wherein a
region’s economic development is explained in terms of its functional role in
 5

Bringing the past back in 5

that pre-​given whole. In certain respects, these were two of the problems that
characterized the idea of “combined and uneven geographical development”
that was a key Marxian research paradigm in economic geography and regional
studies in the late 1970s to the late [Link] paradigm then fell out of favour
from the late 1980s onwards, in part because of the general intellectual shift
away from top-​down grand theory and meta-​narratives that swept through the
social sciences.
To avoid these potential dangers, McMichael (1990) advocates a revised
encompassing approach which he calls “incorporated comparison”, a mode of
historical enquiry in which the “whole” –​such as a big historical economic
process, say the industrialization of a national economy –​does not exist inde-
pendently of or pre-​given to its parts, say the regional economies that make up
the national system, but itself emerges via comparative analysis of those parts,
as moments in a self-​forming “whole”. Large-​scale historical meta-​processes
of capitalist development are not, therefore, “out-​there”, independent of indi-
vidual, pre-​given regions which they (differentially) impact, but are themselves
historically constituted by the processes and patterns of “regionalization” of
economic activity and relations at any given historical moment. However, at the
same time, while such (time-​specific) self-​forming “wholes” (meta-​processes
and metasystems), emerge from their (regional) parts, they are not simply redu-
cible to the latter, but themselves develop macro-​and meta-​level emergent
effects that can then exert “downward causation” on those parts (see Martin
and Sunley, 2012). This could open up possibilities for a more multi-​scalar and
recursive relational approach to the notion of combined and uneven geograph-
ical development.
All the above strategies will work for some [Link] will be instances
when what we are most interested in is a clear understanding of the singu-
larities of a particular region’s historical experience. Getting those singular-
ities right can provide considerable theoretical insight, either by confirming
existing theoretical or conceptual schemas or by serving as a firm basis for the
construction of new grounded theory. A universalizing approach to revealing
the importance of history can be especially illuminating. To show that regions
in widely separated settings undergo the same sequence, pattern or con-
junction of development forces and processes over time not only has highly
positive implications for theory-​building, but potentially also helps to iden-
tify suitable forms of policy intervention that have wide applicability. As Tilly
argues, however, variation-​finding research strategies can help make sense of
socio-​economic structures and processes that never recur in the same local
form across regions, but which appear to vary in some systematic or regular
way, which systematic variation then calls for an explanation. Such research
also has immediate value for designing policies intended to reduce regional
economic disparities. Exploring the role of history in regional development by
means of an encompassing or incorporated research strategy, though certainly
the most challenging, is arguably ultimately the most rewarding. It not only
6

6  Ron Martin, Peter Sunley and Emil Evenhuis

has the advantage of assessing the role of the interconnectedness (and patterns
of economic dominance) between different regions in explaining their his-
torical experiences, but also in helping to understand the historical evolution
of larger socio-​economic systems in terms of the differing but interrelated
histories of their constituent regions. Of course, mixed research strategies are
also possible, indeed perhaps desirable, depending on the specific purpose of
the study in hand.
Elements of these different approaches to taking history seriously in regional
development studies can be found in this welcome and timely book. As the
editors note, the chapters provide some defining points for interdisciplinary
conversations about the principles, aims and methods of a “new regional his-
tory”. In doing so, they discuss aspects of comparative research design and
exemplify the value of some of the regional-​historical strategies of the sort
suggested by Tilly’s work. They also discuss and reflect on a further set of key
questions. As several of the contributors emphasize, a key challenge is to avoid
both ahistorical theory and atheoretical, purely descriptive history. But what
exactly does historical or historically-​appreciative theory mean, and how best
should theory incorporate the significance of time and history at different paces
and rates of change?
These are also questions we have grappled with in our own work, as part
of our endeavours to broaden and develop the conceptual, methodological
and empirical basis of Evolutionary Economic Geography (Martin and Sunley,
2006; Martin and Sunley, 2007; Martin and Sunley, 2012; Martin and Sunley,
2015; Evenhuis and Dawley, 2017). In the past two decades or so, Evolutionary
Economic Geography has come to the fore as an influential “paradigm” in
regional studies and economic geography. It seeks to examine “the processes by
which the economic landscape –​the spatial organization of economic produc-
tion, circulation, exchange, distribution and consumption –​is transformed from
within over time” (Boschma and Martin, 2010, pp. 6–​7). Much of the existing
work within this paradigm has however been particularly concerned with the
micro-​level dynamics explaining the development of patterns of “innovation”
across regions and cities over time (and the spatial configurations of firms, indus-
tries, clusters, technologies, capacities and networks involved). Together with
others, we have attempted to develop a more holistic, socialized and multi-​scalar
approach that places these micro-​level dynamics within a larger perspective,
taking account of macro-​level structures and processes, such as developments
in national economic policies and governance, structural changes in economies
(like deindustrialization and financialization), and changes in global competitive
pressures (also see MacKinnon et al., 2009; Hassink et al., 2014; Pike et al., 2016).
In extending this approach to Evolutionary Economic Geography, not only
have we sought to look at developments over a much longer timeframe, tracing
current patterns and issues back to important critical junctures in the past (also
Henning, 2019), but have also stressed an important role for comparisons, as
these help to uncover the various “mechanisms” by which economic evolution
takes place in regions, how these are tied up with structures and processes at
 7

Bringing the past back in 7

different scales, and the role that particular contextual factors, unique events and
decisions may have played.
In advancing Evolutionary Economic Geography in the directions outlined
above there is a key role for the sort of theoretical explorations, methodological
innovations, and empirical case studies, contained in this volume. Indeed, in
our view, the volume helps to lay the groundwork for bringing the approach
to Evolutionary Economic Geography that we champion, and regional his-
tory –​perhaps reformatted as “new regional history” –​much closer together.
We should clearly not underestimate the challenges involved in developing
a historically-​oriented, interdisciplinary, comparative approach, combining an
enhanced geographical sensitivity with a profound historical epistemology.
Breaking down old institutionalized walls between social sciences and history
will be difficult, but this book provides us with a very valuable set of steps
towards that multidisciplinary endeavour.

References
Boschma, R. and Martin, R. L., 2010. The aims and scope of Evolutionary Economic
Geography. In:  R. Boschma and R. L. Martin, eds. The Handbook of Evolutionary
Economic Geography. Cheltenham: Edward Elgar, pp. 3–​39. [Link]
9781849806497.00007
Evenhuis, E. and Dawley, S., 2017. Evolutionary perspectives on economic resilience
in regional development. In:  N. Williams and T. Vorley, eds. Creating resilient econ-
omies: Entrepreneurship, growth and development in uncertain times. Cheltenham: Edward
Elgar, pp. 192–​205. [Link]
Hassink, R., Klaerding, C. and Marques, P., 2014. Advancing Evolutionary Economic
Geography by engaged pluralism. Regional Studies, 48, pp. 1295–​1307. [Link]
org/​10.1080/​00343404.2014.889815
Henning, M., 2019. Time should tell (more): Evolutionary Economic Geography and
the challenge of history. Regional Studies, 53, pp. 602–​613. [Link]
00343404.2018.1515481
MacKinnon, D., Cumbers, A., Pike, A., Birch, K. and McMaster, R., 2009. Evolution
in Economic Geography: Institutions, political economy, and adaptation. Economic
Geography, 85, pp. 129–​150. [Link]
Martin, R. L. and Sunley, P. J., 2006. Path dependence and regional economic evolution.
Journal of Economic Geography, 6, pp. 395–​437.
Martin, R. L. and Sunley, P. J., 2007. Complexity thinking in Evolutionary Economic
Geography. Journal of Economic Geography, 7, pp. 573–​601. [Link]
jeg/​lbl012
Martin, R. L. and Sunley, P. J., 2012. Forms of emergence and the evolution of economic
landscapes. Journal of Economic Behaviour and Organisation, 82, pp. 338–​351. [Link]
[Link]/​10.1016/​[Link].2011.08.005
Martin, R. L and Sunley, P. J., 2015. Towards a developmental turn in Evolutionary
Economic Geography? Regional Studies, 49, pp. 712–​732. [Link]
00343404.2014.899431
McMichael, P., 1990. Incorporating comparison within a world-​historical perspec-
tive: An alternative comparative method. American Sociological Review, 55, pp. 385–​
397. [Link]
8

8  Ron Martin, Peter Sunley and Emil Evenhuis

Pike, A., MacKinnon, D., Cumbers, A., Dawley, S. and McMaster, R., 2016. Doing evo-
lution in Economic Geography. Economic Geography, 92, pp. 123–​144. [Link]
org/​10.1080/​00130095.2015.1108830
Pollard, S., 1981. Peaceful conquest:  The industrialization of Europe 1760–​ 1970.
Oxford: Oxford University Press.
Tilly, C., 1984. Big structures, large processes, huge comparisons. New  York:  Russell Sage
Foundation.
 9

Introduction
10
 11

1 
The importance of history for
regional economic development
Sara Svensson and Marijn Molema

It is perhaps deeply human to feel that the time in which we happen to live
is unprecedented in multiple ways. As this volume was being prepared, we still
witnessed aftershocks of the 2008 financial crisis and the further rise of popu-
list parties, and, in some places, governments. In Europe, the decision by the
United Kingdom to leave the EU and the political and policy uncertainty
related to that, caused many to worry, as did the regional conflict between (parts
of) Catalonia and Spain. In the US, the election of Donald Trump as President
increased uncertainty. These current, complex and to some extent unforeseen
processes have regional dimensions in common, in the sense that the processes
have much to do with geographical “cold spots”. These are regions that find
themselves below average and falling further behind in social, economic and
sometimes cultural, terms. This was demonstrated by voting patterns in British
and US elections, and have caused what some refer to as “touring journalism”
in which reporters seek out small places to try to “understand these people”.
What we argue though is that understanding regional problems and how they
have political, social and economic ramifications at the national level asks for
“deep explanations” in which the historical dimension cannot be absent. This
is a view that is gaining currency within Regional Studies, and to which we
adhere as well.

Aim and key questions of the book

The past plays a crucial role in our understanding as well strategically supporting
regional economic development. History signifies the role played by insti-
tutional and cultural legacies that inform choices between different policy
options. In that respect history becomes crucial for both Regional Studies and
the practice of regional policy. However, while regional economic evolution
and development policies are the object of research in both the social sciences
and the humanities, interdisciplinary connections are scarce. To develop this
a bit further, this book is based on two empirical premises reinforced by our
observations as researchers and teachers on regions. First, academic teaching
and research on regions is often bound up in either nationalist methodologies,
or very localist/​local government issues. In our own teaching and professional
12

12  Sara Svensson and Marijn Molema

lives, we have observed the need for works dealing with regions that can be
used in courses directed towards undergraduate and graduate students, but also
for researchers in the field. Second, while disciplinary walls sometimes cause
problems within the social sciences (geography, sociology, political science etc.)
those walls are thicker and more persistent between social sciences and the
humanities. In order to develop our understanding of regional development
and regional policy, multidisciplinary perspectives are needed.
Through this volume we wanted to explore the connections between both
domains, extracting benefits of interdisciplinary cooperation for the devel-
opment of Regional Studies as a field. As editors this meant we also had to
question our own assumptions about ways to think about and approach know-
ledge, since one of us is a trained historian, the other a political scientist. We
experienced ourselves how scholars from the social sciences can learn from
historians that a longer time perspective is crucial to understand how regions
develop, and what the impact of regional policies are. Likewise, we saw how
historians can benefit from the theoretical and conceptual strengths of social
science. It was our hope that this collection of theoretical and empirical essays
should stimulate the study of regions across disciplinary boundaries, and we set
out the work with three aims in sight:

• to stimulate the study of regional development and its accompanying pol-

icies across disciplinary boundaries.
• to connect the study of history with the social sciences through examples
that both illustrate its usefulness and give directions for the future.
• to show how understandings of the past (history) is vital for shaping pol-
icies for the present and the future, thus fulfilling the promise of regional
studies not only to be multidisciplinary but also to span the gap between
research and policy practice.

This volume critically investigates theoretical approaches to regional studies in

the intersections of social science and history, explores possibilities that can open
up through methodological innovation and cooperation between the disciplines
and showcases the values of improved theory and methodology through a series
of case studies. Thus, the first part of this book (Disciplinary and theoretical
explorations) critically reviews analytical approaches that incorporate elements
of the past, such as evolutionary economic geography, historical institutionalism
and behavioural psychology. The second part (Innovations in research design
and methodology) engages with different analytical frameworks, including
attention to methods and data like time series. Finally, the third part marks the
interdisciplinary field with the help of case studies in several countries, regions
and sectors.
This book would not have happened without the stimulus provided by the
Regional Studies Association (RSA). As part of the 50th anniversary celebra-
tion of the Regional Studies Association, the Research Network on Regional
Economic and Policy History (ReHi) was launched as one of several supported
 13

The importance of history 13

research networks. The aim of the network was similar to the broader aim of
this book; to connect social scientists and historians within the field of regional
research, which would in turn facilitate a deeper understanding of the roles
played by history in regional development in general, and in regional policy-​
making specifically. Many of the contributions to this book have as their origin
presentations that took place at the inaugural workshop of the network, held
in London on April 25–​26, 2017. The workshop focused on “stock-​taking” of
the connections that already exist between history and regional studies, and
explored existing points of contact between historians and social scientists
involved in regional studies, including geographers, political scientists and
sociologists. Many geographers, political scientists and economists include his-
torical perspectives in their work. This can include an occasional sketch of the
historical context of economic and/​or political processes. Additionally, concepts
such as “learning region” or “regional resilience” are closely related to histor-
ical studies and analyze long-​or medium-​term processes which have a his-
torical dimension. Simultaneously, the historical sciences have a long tradition
of studying regional economic development and a vast body of work exists.
Agglomerated economies are an important subject in the historiography about
economic growth. Besides, historians make use of concepts borrowed from
economic geography. For example, business historians have operationalized
concepts such as “industrial districts” and “clusters” in order to explain the
dynamics of regional economies. Furthermore, planning and economic policies
also include themes for historical analysis. These and other dimensions of the
social sciences provide a point of contact between historians and scholars with
a regional studies background. The two days were divided into the guiding
themes of ideas, methodologies and case studies, with four panels exploring
theory, conceptual developments, variations in measurements and regional
policies. That and subsequent workshops organized by the network have also
inspired the editors when looking for complementary contributions to the
volume.

Content of the book

Disciplinary and theoretical explorations

The benefits of interdisciplinary collaboration for theoretical development of
Regional Studies and History, is explored in the first part of the [Link] scene
of the book is set in Chapter 2, in which Marijn Molema and John Tomaney
demonstrate how a particular lens –​a focus on analysis of institutions –​can
serve as a binding agent between regional and historical studies. Together, this
duo, consisting of a historian and an urban planner, show that research objects
of both regional and historical studies are situated in a (regional) society, and
that the societal basis of both sciences makes it possible to find a middle way
between atheoretical histories on the one hand and ahistorical social science on
the other. The chapter sketches the relationship between the two domains of
14

14  Sara Svensson and Marijn Molema

the sciences, arguing that a period of divergence (1890–​1945) was followed by
a period of convergence (1945–​1980). In this period historians were inspired
by social scientific approaches and introduced explanatory models, as well as
classifications and typologies in their methodologies. After the advances of
regional and historical studies, institutional thinking became influential during
the final decades of the 20th century in disciplines like sociology, economy
and political sciences. The authors argue that this opens up new possibilities
to strengthen the connection between regional and historical studies. Through
institutional analysis, scholars from social sciences and historical studies can
cooperate in order to better grasp the mechanisms of regional development
that are driven by institutional factors and circumstances.
In Chapter  3 Robert Huggins and Piers Thompson push the theoret-
ical inquiries a step further by scrutinizing the role of human behaviour in
shaping regional economic trajectories. Emerging theories of economic devel-
opment and growth differences across regions and cities are moving toward
a (re)turn to addressing the role of individual and collective behaviour in
determining regional development outcomes. The chapter seeks to propose a
behavioural theory of how such places evolve and develop. To achieve this, the
key constructs and principles relating to this behavioural theory of urban and
regional development are defined. It is argued that the roots of behavioural
differences across cities and regions are co-​determined by two key factors,
namely:  culture and personality psychology. In essence, it is the interaction
of these two factors in the form of a psychocultural profile that generates the
human agency of individuals. Along with developing a new perspective on
the factors underlying human agency and its role in explaining the growth
of cities and regions, the chapter further explores this for a number of case
study cities. Evidence is provided illustrating the role played by key individ-
uals in ensuring urban and regional development through innovation. This
provides insights into the policies that can be embraced to encourage the
psychocultural conditions that promote the types of human agency boosting
development and avoiding lock-​in.

Innovations in research design and methodology

In this volume’s Part II, the demonstration of how collaboration can lead
to methodological innovation and therefore new insights is done through
chapters about three distinctly different uses of innovation in research design
and methodology: interdisciplinary composition of the research team, use of
new data sets and advancement of comparative methods applied to studies of
regions.
Chapter  4 is written by Polish scholars Justyna Kościńska and Mikołaj
Herbst, both at the University of Warsaw. They describe how they applied an
interdisciplinary approach in examining the impact of historical institutions on
present-​day educational achievements in Poland. They make use of a distinct
feature with relation to the history of education in Poland; during the time at
 15

The importance of history 15

which the Polish education system was formed in the 19th century, Poland was
not an independent country but was partitioned between the three empires
of Russia, the Kingdom of Prussia, and the Habsburg-​ruled dual monarchy
of Austria-​Hungary. The project team included representatives from different
fields, such as economics, sociology, geography, mathematics and history, and
combined different methods including quantitative techniques, qualitative tools
and elements of spatial analysis resulting in the use of new variables. Using
this varied methodology, they found that the institutions shaped during the
period of the territorial division of Poland had a significant effect on average
achievements in contemporary schools.
In Chapter 5, economic historians Kerstin Enflo and Anna Missiaia show
the importance of coming up with new and innovative ways to enable long-​
time analysis of phenomena that are core to current regional development
debates such as the source of inequality. Both national and regional GDP are
relatively recent economic indicators: they have been systematically produced
by statistical offices only well after the end of the Second World War. For earlier
periods, GDP at any geographical level can only be reconstructed using data
that were not gathered to this end. The estimation of historical regional GDP
relies on even thinner evidence compared to the national one. This chapter
provides an overview of the existing literature devoted to estimating historical
regional GDP and discusses the methodological challenges faced by scholars.
The potential yields of long-​term regional time-​series is shown through an
illustrative case study of the Swedish NUTS-​3 regions for which regional
GDP estimates in selected benchmark years are available from 1571 to today.
These estimates provide the longest series of regional GDP to date for any
given country, giving a unique perspective of regional inequality from pre-​
industrial to present times.
The historian Martin Åberg and political scientist Thomas Denk pose the
question in Chapter 6 of how social science and historical approaches can be
combined in order to enrich the regional studies theme. Their chapter focuses
on designs for small-​N comparative research, specifically the problem of how
to apply systematically structured synchronic and diachronic comparisons as
a means to analyse interdependence. They hold that comparative sequence
design (CSD) poses an important methodological option in that perspec-
tive. In contrast to standard research designs, such as cross-​sectional design
and development design (“historical case studies”), CSD allows for both sys-
tematic (1)  cross-​unit and cross-​case comparisons as well as (2)  within-​unit
analyses. It (3) opens up for the use of process-​oriented models with a focus
on explaining developmental outcomes among sub-​national level units, and
thereby (4)  facilitates analysis of both spatial, temporal and temporal-​spatial
interdependence among sub-​national level units (regions). CSD provides a tool
for examining, for instance, policy-​making and policy diffusion as dynamic,
historically contingent processes; and, from the perspective of the historical
disciplines, allows analysts to move beyond the limits of traditional case studies
of regions and localities.
16

16  Sara Svensson and Marijn Molema

Empirical case studies

Part III invites the reader to look at long-​term empirical studies based on
regional development from Latin America (Brazil), North America and Europe,
as well as Asia (China). Chapter 7, written by Tomàs Fernández-​de-​Sevilla and
Armando J. dalla Costa, uses a common narrative approach in historical analyses.
The chapter seeks to discover the role played by factors such as the presence
of external (or Marshallian) economies, local institutions, industrial policy
and big companies in the formation and expansion of the São Paulo automo-
tive industry cluster. The evidence presented shows how the clustering of the
automotive industry around São Paulo in the 1940s to the detriment of other
Brazilian regions is explained by the presence of external economies generated
in the 1920s. However, the expansion of the cluster was only possible with the
application of strategic policies to protect the nascent industry in the 1950s. It
was precisely the associations of local manufacturers which campaigned for the
application of this type of policy, that had a bearing on the establishment of
strict requirements for national content. The closing of the market forced the
establishment of the big brands in the sector which, from the 1960s, acted as
hub companies and hierarchized the cluster. In this phase, the role of the asso-
ciations of local manufacturers consisted of providing advice and training for
the autoparts industry.
In Chapter 8, Kevin Cox focuses explicitly on traditions of regional eco-
nomic policies in modern, industrialized nations. In his comparison between
the United States and Western Europe, he uncovers essential differences. Urban
and regional development policy emerged as a clear government focus after
the Second World War. Legitimated by the earlier emergence of city and
regional planning, it based itself on a new mobility of firms and people in
the space economy, albeit mediated by a newly-​emergent property capital. It
would assume very different forms in the countries of Western Europe and the
USA. In the former, central states would play an important role, facilitating the
movement of employment to backward areas and those of high unemploy-
ment. New towns were constructed to take the pressure off major metropol-
itan centres. In the US, the emphasis would be bottom-​up. Local governments
competed for inward investment in a market for locations. Immediate reasons
for this difference include contrasting state structures, particularly in their ter-
ritorial aspects. These structures, though, are one aspect of more fundamental
variations in social formations having to do with the fact that in Western
Europe capitalism emerged from pre-​capitalist social relations, accompanied by
the absolutist state. In the US this was not the case. These archetypes of policy
traditions help to characterize regional economic policies in a general way.
Meimei Wang and Bas van Leeuwen provide the last empirical chapter (9),
in which they examine economic dynamics by looking at urban centres in
China as cores of educational development and modernization. As economic
historians, Wang and Van Leeuwen combine various disciplines like political
science and sociology and history, and discuss systematically the relation between
 17

The importance of history 17

urbanization and education in multiple political [Link] findings are, first,
that academic fragmentation on this topic has persisted over time, with the
exception of the 1949–​1978 period when research was dominated by political
scientists. Second, the chapter demonstrates how different statistical definitions
of urbanity and urbanites have been translated into policies of essential rele-
vance for the schooling opportunities of children, but also how the effects of
this (urban versus rural schooling) must be seen in relation to other long-​term
factors.
Finally, Marijn Molema draws together the picture as based on the different
contributions in the volume, and sketches the way ahead for the field. It
delineates a field of research in which regional historians exchange knowledge
with colleagues from the social sciences, such as geographers, economists, pol-
itical scientists and sociologists. The field, programmatically described as “New
Regional History”, is demarcated by four points of departure or “corner flags”.
The first corner flag argues that every multidisciplinary exchange should start
from a societal problem or challenge. The second corner flag stands for the idea
that history and regional studies should together increase our understanding
of how regions were constructed, by whom and on the basis of what kind
of arguments. The third corner flag introduces multidisciplinary exchanges
between regional and historical studies as a rule of conduct, and the fourth
commands a comparative approach and a preference for analyzing the transfer
of people, ideas and commodities between regions.

Time to look forward, and time to look back

No social science consists solely of events that are static in [Link] the relation
to time in social science has often been an uneasy one, since both quantitative
and qualitative methods struggle with well-​known and often-​discussed diffi-
culties. Examples of these constraints are the difficulty to assess the direction of
causality, or the (often faulty) memories of interviewees. Short time-​frames and
attention spans of policy-​makers, which increasingly ask for quick and directly
applicable research results, does not favour a more evolutionary approach either.
However, we hope that the edited volume will help set a research agenda in
which regional development will also be studied from a historical perspective.
18
 19

Part I

Disciplinary and theoretical

explorations
20
 21

2 
Regional development, history
and the institutional lens
Marijn Molema and John Tomaney

Introduction
What is the role of history and historians in understanding patterns of regional
economic development? Most politicians, policy-​makers, entrepreneurs, and
indeed, scholars are forward-​looking. But there is also a growing awareness
that history matters in understanding patterns of regional change. Scholars of
regional development frequently sketch historical contexts of economic and/​or
political processes in their work. In this chapter we explore this increased his-
torical awareness, especially with regard to the study of economic development
in regions. Understanding the role of history is part of a broader trend in the
social sciences, to which the interdisciplinary field of regional studies belongs.
Viewing the development of cities and regions as embedded in frameworks of
institutions favours a longer-​time horizon for the study of regions. As such, the
recent past offers a laboratory in which all kinds of concepts and theories can
be tested and refined.
Conversely, concepts such as “resilience”, that originate in the social sciences,
are closely related to historical studies, because they analyze long-​or medium
term processes which have dimensions of continuity and change (Martin and
Sunley, 2015). Simultaneously, historians make use of concepts borrowed from
the social sciences. For example, historians have operationalized concepts such
as “institutions”, “industrial districts” and “clusters” (Molema, Segers and Karel,
2016). But research which aims to be policy relevant often lacks signs of multi-
disciplinary exchange between historical and regional studies. The gap has its
origins in methodological differences between historians and social scientists,
in which, broadly, the former apply a narrative approach while the latter search
for general theories.
Seen from the perspective of the research object, however, historical and
regional studies may overlap. Both scientific domains study phenomena that
are situated in a (regional) society. The societal basis of both sciences makes it
possible to find a middle way between atheoretical histories on the one hand
and ahistorical social science on the other. A search for this middle way needs
some understanding of how the traditions within historical and regional studies
have evolved. Disciplinary traditions include customary practices about the
22

22  Marijn Molema and John Tomaney

incorporation of research themes, theories and [Link] have trav-
elled from the social sciences in to history, and vice versa. This chapter will
begin with an overview of the dynamic relationship between history and the
social sciences between 1890 and 1980. Next, we will examine the institu-
tional turn in the social sciences during the last quarter of the 20th century,
and its repercussion for historical studies. We propose that institutional analysis
can function as the binding agent for multidisciplinary connections between
regional and historical studies. Finally, we explore the relevance of these themes
for the study of regional economic development. The chapter shows how the
institutional turn impacts on regional perspectives in both the social sciences
and the study of history. We conclude by discussing why it is important to
engage historians in current debates on regional development.

History and social sciences: A dynamic relationship

The connections between history and the social sciences are of crucial import-
ance for those who are interested in historiography, theory and methodology.
During the 19th century, history and the social sciences became professionalized
and established themselves within universities as new academic disciplines. As
such, they sought to define their research objects and demarcate themselves in
the wider academic landscape, resulting in a series of intellectual conflicts.

Divergence (1890–​1945)
At the end of the 19th century, history and the social sciences met at a cross-
roads. The methodological and theoretical differentiation was most visible in
the German academic dispute about the correct form of historiography and
the difference between history and the social sciences. The narrative approach
was prescribed by the influential historian Leopold von Ranke (1795–​1886).
It was the task of historians to gain insight into the growth and development
of the state. This insight was necessary for an understanding of the present and
the achievement of good governance. As the master of a ship has to know the
difference “between a warship and a cargo ship”, the politician has to grasp “not
only the nature of the sea in which he is to sail, but, above all, the complete rec-
ognition and grasping of the nature of his state” (Ranke, 1925, p. 43). History
served politicians and policymakers, and good history had to show, on the basis
of primary sources, “how it actually has been”. The link between history, criti-
cism of the sources and politics meant that historiography was strongly focused
on individuals and their role in political processes, such as the contribution of
important statesmen and their role in the great events of the national past.
By the end of the 19th century, the Rankean view on history was being
criticized (Iggers, 1984). One of the challengers was Karl Lamprecht (1856–​
1915), who pursued two goals in his writing of history. First, Lamprecht
contended that the field of historical science had to move its centre of attention
to the social and cultural aspects of history. Second, he stressed the value of
 23

The institutional lens 23
analytical frameworks and comparative approaches to ensure that history was
scientifically based. Lamprecht criticized the lack of clear theories and meth-
odologies of the (Neo)-​Rankeans, and pleaded for the exploration of causal
laws. Lamprecht’s attempt to modernize the historical sciences itself attracted
opposition. The debate in German historiography between 1891 and 1899 is
known as the “Methodenstreit” and set the tone for academic debates around
the world. Lamprecht was accused of neglecting peculiarity and variation in
history. The search for causality did not belong in history.
In the debate about methods, two different approaches emerged which can be
called “nomothetic” and “ideographic” (Windelband, 1915). The ideographical
approach centres on individual and collective experiences and particularities.
The nomothetic method, on the other hand, aims to generalize, categorize
and find general laws. Scientists use the nomothetic method to explain, and
the ideographical method to understand. The controversy inspired a distinction
between history and the social sciences in the first half of the 20th century.
The “ideographic” method was accepted by the historians and theoretically
deepened by hermeneutic philosophers such as Gadamer (1960). Although the
“nomothetic” procedure also found adherents among historians, the search for
generalizations and regularities was assigned to social sciences such as soci-
ology, economics and political science. Between historiography and the social
sciences, two different (prototypical) methods emerged. The historians worked
narratively, focusing on context peculiarities and tried to present a synthetic
account of temporal change. Social scientists focused on the general, investigated
structures and processes with a theoretical reference and tried to analyze social
structures and regularities.
There were important exceptions to the broad categorization we have made
so far. Combinations of nomothetic and ideographic approaches can be found
during the whole first half of the 20th century as well. One of these strands in the
historiography consists of works inspired by the Marxist materialist conception
of [Link] conception rested on a clear theory about the development of the
society, driven by the accumulation of capital and class-​based forms of exploit-
ation and attendant [Link] Industrial Revolution and the concomitant for-
mation of class consciousness became a central theme for historians such as John
and Barbara Hammond (1920).The French Annales School of History eschewed
teleological views on the historical process, focusing on the long-​run develop-
ment of social structures to explain historical change (Burke, 1990). Continuing
on the path laid by Lamprecht, historians in the Annales tradition emphasized the
importance of social and economic processes in history, opening the possibility of
multidisciplinary cooperation with cognate disciplines such as sociology, demog-
raphy and anthropology, among others (Febvre and Bloch, 1929).

Convergence (1945–​1980)
Marxist history (Thompson, 1963) as well as the Annales School (Braudel, 2009)
remained influential after the Second World War. At the same time however,
24

24  Marijn Molema and John Tomaney

historians searched for connections with the social sciences on a much broader
scale. Social science, as Iggers (2006, p. 238) reminds us, “gained momentum in
the three decades immediately following World War II”, which had implications
for the discipline of history. Historians sought to place the discipline on more
scientific foundations, paying special attention to questions of method (Stone,
1977). In this view, every historical project should start with a clearly-​defined
problem, derived from rigorously determined concepts that direct the research
process. Awareness that history is not just a collection of facts derived from the
past, and that historical narrations are mediated by assumptions of the historian,
required explication of their intentions and expectations. Such new basic rules
were inspired and influenced by the social [Link] impact of neighbouring
disciplines such as sociology and economics was undeniable in another way as
well. In their attempt to make reflections on the past more scientific, historians
embraced comparative and quantitative techniques. The adaptation of these
methodological aspects was closely connected with a reassessment of theories.
Historical reasoning moved toward the use of hypotheses that could be tested
and refined theories about societal development as a consequence. As such,
historians learned to work with explanatory models, as well as classifications
and typologies, so that general processes and structures from the past could be
revealed (Tosh, 1999, p. 134).
The methodological and theoretical transformations led to new histor-
ical movements in various countries. Among Anglo-​American historians for
example, “social history” emerged which dedicated itself to research on the
development on the capitalist, industrial society and the rise of national states
(Tilly, 2002). Anglo-​American social history was inspired by an international
comparative study about long-​ term political development, with social and
economic structures as important variables that explained differences in polit-
ical regimes of 20th century America, Asia, Europe and Russia (Moore, 1966).
Strong influence of the social sciences, together with multi-​case comparisons,
became paradigmatic for an approach that has been labelled as “social science
history” (Monkkonen, 1994). The Nordic and the Low Countries, among
others, also contributed to this movement. Scholars in West Germany modified
the quest for innovation in history into a “historical social science” or Historische
Sozialwissenschaft (Nolte, 2002). This approach was programmatic in the sense
that it sought a new narrative of Germany’s past in which societal processes
and structures, instead of political events, were revealed and analyzed as the
drivers of history. Compared to the social science history that was produced in
other countries, the German historical social science was less characterized by
international comparisons. The national focus was in line with the historians’
engagement:  reconstructions of the past could contribute to democratic
developments in Germany. We underscore this feature, because it links to our
own aspirations regarding a stronger role of historians in current debates about
regional development.
As a result of the new, interdisciplinary orientation among historians and
their references to socio-​economic structures and processes, history and social
 25

The institutional lens 25
science converged in the third quarter of the 20th century. While historians
discovered the value of scientific methods, comparative research and the
development of hypotheses, social scientists, in turn, stretched their research
objects to incorporate a concern with the past leading to the emergence of
disciplines such as historical sociology and historical geography. Sociologists
dealing with “big issues in social theory” acknowledged the value of looking
into the past for understanding present-​day life. The rapid transition from an
agrarian to an industrial society called for an interpretation which included
historical perspectives and led to the establishment of “historical sociology”
as a sub-​discipline (Calhoun, 2003). Historical geographers developed a focus
on the changing relationship between people and the environment over time
(Baker, 2003).

The institutional turn
The multidisciplinary connections between the social sciences and history
were strengthened, during the third quarter of the 20th century, by an overlap
in research objects as well as methodological and theoretical approaches.
During the fourth quarter of the 20th century, we suggest, institutional ana-
lysis acted as another binding agent between both scientific domains. The con-
cept of institutions originated and survives in the study of sociology (Schelsky,
1970). From a sociological point of view, institutions can be broadly defined
(Acham, 1992). First, the concept refers to a normative behavioural structure
or a system of social rules. This behavioural structure can be general or spe-
cific. For example, a religion influences the general behavioural structure and
Catholicism influences specific behaviour. Second, organized associations can
be seen as institutions. Organizations have formal characteristics and are typified
by explicitly formulated goals, rules and practices. This regulates the behaviour
of their members. Power and hierarchical structures determine the distribution
of areas of competence and the responsibility for making decisions. An example
of an organization is the convent. The members of the convent are nuns who
have specific tasks, such as praying and working. Third, specific forms of know-
ledge can be analyzed as an institution. These forms of knowledge ensure a
conscious standardization of action through an explicit formulation of values
and norms, in relation to beliefs, such as religion, ideology, philosophical or sci-
entific convictions. In addition to the canonization of behavioural structures,
specific forms of knowledge justify and legitimize human action. An example
is theology.

Institutions and the economy

“Institutions” refers to a wide range of phenomena, but in Douglass North’s
classic definition consist of, “the humanly devised constraints that struc-
ture political, economic and social interaction. They consist of both informal
constraints (sanctions, taboos, customs, traditions, and codes of conduct), and
26

26  Marijn Molema and John Tomaney

formal rules (constitutions, laws, property rights)” (North, 1991, p. 97). Within
contemporary economics, the definitions presented by the sub-​discipline of
the “new institutional economics” include: “(…) a rule or system of rules, a
contract or system of contracts (including their enforcement mechanisms)
through which the behaviour of individuals is channelled” (Erlei, Leschke and
Sauerland, 1999, p. 23). The neoclassical approach to economics that emerged
at the end of the 19th century incorporated a strong methodological individu-
alism and largely eschewed a concern with the institutional foundations of
economic life. Although periodically modified, neoclassical approaches remain
mainstream within the discipline, which has taken on an increasingly math-
ematical character, focused on the building of increasingly complex models.
The neoclassical approach eclipsed alternative theorizations, which existed in
the 19th century and afterwards. For instance, in Germany from 1830 to about
1880, the “Historical School” focused less on building models and more on
historical data; it did not only rely on mathematics, but attended also to social-​
cultural interactions between agents as a development factor (Drukker, 2006).
The institutional approach was used to explain how markets were ordered,
and how ownership relationships developed as well as the evolution of eco-
nomic standards such as measurement and monetary systems. Other alterna-
tive theorizations of the economy such as those proposed by Henry George
were also marginalized in this period (Ryan-​Collins, Lloyd and MacFarlane,
2017). Although marginalized, these alternative ways of viewing the economy
did not disappear. Later, economists such as Polanyi (1944) stressed the role of
institutions as providing the framework within which markets and economic
development processes operated.
“New institutional economics”, as a term, was introduced by Williamson
(1975), simultaneously invoking and distinguishing the old institutional
approach represented by the Historical School and others. The new institu-
tional economics remained embedded in a neoclassical framework, accepting
the underlying rules that theory determined the economic process, regarding
institutional theories as complementary to (or a development of) the ortho-
doxy. The new institutional economics paid particular attention to the role of
phenomena such as property rights, transaction costs and principal-​agent the-
ories (Hazeu, 2007, p. 86). Although highly influential, this approach was not
revolutionary and was seen by its protagonists as amounting to a loosening of
some of the strictures of neoclassical orthodoxy rather than their displacement.
These developments impacted upon economic history. North identified
a science of “cliometrics”, which used the “building blocks” of neoclassical
economics (North, 1984). Such approaches developed the use of quantitative
approaches to examine the evolution of economies, such as the reconstruction
of historical national accounts (Bolt and Van Zanden, 2014) or the role of infra-
structural development on economic growth (Fogel, 1964). New institutional
economics encouraged economic historians to go a step further and draw the
social context in which economic relationships originate into the centre of
academic attention. Institutions originate in the society; insight and knowledge
 27

The institutional lens 27
of societal development offered the key for understanding how, why and when
institutions that favoured (or hampered) economic growth came into being in
the first place. Instead of excluding institutions from explanatory models by
taking them as given, economic historians tried to take institutional develop-
ment as the starting point of grasping economic development. An important
purpose of economic history became, in the words of North (1991, p. 98) “to
account for the evolution of political and economic institutions that create an
economic environment that induces increasing productivity”.

Institutions and the political

The institutional turn was not limited to the economic realm. Political scientists
focused their attention on the formation of political institutions from the 1980s.
This led to new approaches, summarized under the term “new institutionalism”
(March and Olsen, 2008). Among the different kinds of institutionalism (Hall
and Taylor, 1996) the approach of the new Historical Institutionalism is most
relevant in our context, which focuses on the influence of informal procedures,
routines, norms and conventions of an organizational structure or certain polit-
ical policy (Skocpol and Pierson, 2002). The focus of historical institutionalists
was on longer periods, with particular reference to context and political vari-
ation. Institutions provided a normative framework for interpretation.
The concern for institutional development in political science, economics
and economic history also impacted on historical studies. The rise of the new
social and economic history during the second half of the 20th century aroused
criticism from some American scholars in the 1980s. Skocpol’s (1985) call to
“bring the state back in” signalled a concern to reemphasize the importance of
the political in explaining longer-​term patterns of change. On the one hand, the
state was regarded as an actor: state organizations impacted on society in general
and events of public policy in particular. On the other hand, it was emphasized
that the state with its arrangement and patterns of action influenced political
cultures. So, the formation of groups, agendas and the way in which polit-
ical groups drew attention to their claims became important research topics.
This novel approach stimulated comparative studies which, while remaining
concerned with the impact of socio-​economic structures, focused on the evo-
lution of the political structure of society.
The rediscovery of the political, and renewed concern with the structures of
the state, was the backdrop to the emergence of “territorial politics” (Keating,
2008). Although sub-​national politics is always present, the degree to which it
attracts the attention of scholars varies over time. An underlying teleological
assumption of the post-​war period was that territorial differences would be
eroded by processes of national integration and, more broadly, the march of
modernity. The “political sciences” contributed to the neglect of the local and
regional in its pursuit of universal laws of [Link] persistence of local
political claims was conceived theoretically as “revolts against modernity” (Lipset,
1985). The focus on territorial politics emerged in response to the neglect of
28

28  Marijn Molema and John Tomaney

sub-​national geography in political [Link] rise of “small nationalisms”, the
claims of stateless nations and the emergence of regionalist movements attracted
the attention of political scientists from the 1970s. Decentralization of political
authority to local and regional government was often the response to these
developments. New theories sought to explain these developments through the
prism of centre/​periphery relationships, by problematizing assumptions about
the formation of nation states and by rediscovering the importance of regional
political economy. Bulpitt (2008, p. 59) defined territorial politics as:

(…) that arena of political activity concerned with the relations between
the central political institutions in the capital city and those interests, com-
munities, political organisations and governmental bodies outside the cen-
tral institutional complex, but within the accepted boundaries of the state,
which possess, or are commonly perceived to possess, a significant geo-
graphical or local/​regional character.

Political scientists typically made great use of historical evidence in order to

make sense of the new terrain.

Institutional analysis and development in space

The preceding review of the evolution of the relationship between history
and the social sciences is a necessary prelude to our main task which is to
apply these insights to the relationship between history and regional studies.
In particular, we have identified the “institutional turn” as fertile ground for an
explanation of the development of more historically informed understanding
of regional development.

Regional development in the social sciences

Region is a key concept in human geography. A concern with the region is
bound up with the beginnings of the discipline in the modern era. Regional
geographers placed the study of areal delineation at the heart of the discipline.
Regional geography in different forms became influential in many countries
by the turn of the 20th century but, in essence, concerned the interrelationship
of landscape and life. The “chorographical tradition” provides rich historically
informed accounts of the development of the region but was left marginalized
by the social scientific revolution. The growth of regional science after the
Second World War drew on and reflected the growing importance of eco-
nomics as a discipline in itself and in the policy world, as well as the growth of
computational forms of analysis and, more broadly, the pervasive influence of
logical positivism. In addition, it spoke to the needs of rapidly-​growing indus-
trial economies in the post-​war period providing data to aid planning. A con-
cern with the region did not vanish as a result of the quantitative revolution
and the rise of regional science, but the problem was reformulated. Defining
 29

The institutional lens 29
the region became a special type of classification problem. While classification
of populations into mutually exclusive categories is a vital element of quantita-
tive techniques in social science, the need to define contiguous areas meant that
they were not necessarily as homogeneous as other classified groups. Thus, the
development of complex mathematical models and theories became central to
the definition of regions (Tomaney, 2009).
Political and economic crises in the 1970s and 1980s brought into question
regional science approaches that looked for statistical regularities, when what
was required was more attention to the qualitative factors that shape different
patterns of development. In an era of increasing global integration geographers
argued that regions were being reshaped by accelerating material and discur-
sive flows that, simultaneously, were undermining the political authority of the
nation-​state. New research showed that the region remained an important cul-
tural, political, and ecological category in countries across the world, albeit with
great diversity in their formation and transformation. There is no unfolding
logic of regionalization.
On the other hand, according to some geographers, the intensification of
global flows is undermining the notion of territorial political identities. In these
accounts, regions are understood as open and discontinuous spaces, constituted
by social relationships and material, discursive, and symbolic forms stretched
across and through regions, founded on a topological rather than topographical
approach to space that ultimately rejects the notion that regions are territorially
fixed, bounded or closed.“Regions” are defined by their place in networks and by
their relationship within and beyond predefined territorial boundaries. Regions
are “unbound”. Ideas of stable and clearly demarcated hierarchies of scale are
eschewed in this perspective. Indeed, the notion of the region is replaced by the
idea of a “regional assemblage” or, a multi-​actor, multi-​scalar topological geom-
etry. This relational view draws attention to the way in which (post-​)modern
life is embedded in socially-​constituted and spatially stretched networks. But, in
this perspective, there is a danger in overlooking the persistence of territoriality
and the reproduction of boundary setting and the enduring presence of scales as
expressions of social practice, discourse and power (Tomaney, 2009).
Attention to the fate of regions has been central to the emergence of “evolu-
tionary economic geography”. This sub-​discipline emerged in the 1980s under
the influence of institutional and evolutionary economics. The Schumpeterian
view on economic renewal, so influential among evolutionary economists, was
now applied to specific regions and territories. The basic concern of this sub-​
discipline is the “processes by which the economic landscape –​the spatial organ-
ization of economic production, distribution and consumption –​is transformed
over time” (Boschma and Martin, 2007, p. 539). Key concepts deployed in evo-
lutionary economic geography include regional “path dependency”, “lock-​in”
and “resilience”, each of which are historical formations (Martin and Sunley,
2011). Moreover, the institutional conditions in which regional economies are
embedded are critical to the shaping of patterns of development and to the
extent of resilience in the face of economic shocks (Pike et al., 2010). While
30

30  Marijn Molema and John Tomaney

influenced by the social-​cultural approach of the institutionalists, the emerging
evolutionary perspective focused on Schumpeterian processes of economic
change and renewal. With his emphasis on creative destruction and new com-
binations of economic resources, Schumpeter offered a theoretical direction to
study the process of innovation and provided the foundations for another sub-​
discipline: evolutionary economic geography. Evolutionary economists focused
on the dynamics within an economic process and studied “the generation
and impact of novelty as the ultimate source of self-​transformation” (Boschma
and Martin, 2007, p. 537). However, the evolutionary economic geographers
combined the evolutionary angle with a spatial view on the economic process.
They studied the relationship between regional concentration and economic
renewal. As such, evolutionary economic geographers introduced research
themes that attracted the attention of economic and business historians, who
embraced concepts like path dependency (Valdaliso et al., 2014) and clusters
(Popp and Wilson, 2007; Molema, Segers and Karel, 2016).

Regional development in history

Regional economic development has been an important theme in (economic)
history, albeit at the periphery of the discipline. Mainstream history is more
inclined to study national economic development  –​and more recently global
differences between, for example, Europe and Asia. Recently, some historians
have sought to explain differences in national patterns of development with
reference to the historical evolution of institutions. Notably, Acemoglu and
Robinson (2012) stress the inclusiveness of economic and political institutions
as the determining factor in the prosperity of nation states. Since the late 1970s,
historians have paid attention to the uneven geography of economic growth
between and within states. Work by Fremdling and Tilly (1979, Pollard (1981)
and Kiesewetter (2004), among others, showed how economic processes are
inscribed geographically. Against the dominant national frame of reference, they
stressed the role of regional business concentrations within economic and social
history. Their programmatic goal was to complement the historical view on
the Industrial Revolution with its important regional dimensions. These kinds
of studies of regional development, with a combination of quantitative and
qualitative approaches, largely fell out of fashion. However, the last couple of
decades have seen progress in the quantification of regional growth. Like the
reconstruction of national accounts, economic historians have worked on the
quantification of historical regional development throughout Europe (Wolf and
Rosés, 2018; Enflo and Missiaia, this volume).
Another, more continuous line of inquiry is the field of urban history. Cities
are obviously the critical unit of analysis within this historical sub-​field, which
took root from the beginning of the 20th century. There are two reasons why
urban history is of relevance in the search for the overlap between regional and
historical studies. First, urban historians cultivated a deep concern with the cat-
egorization and conceptualization of space and place on the sub-​national level.
 31

The institutional lens 31
They developed a rich typological framework (Ewen, 2016) which is helpful
for regional research as well. For instance, cities are topologized by virtue of
their scale (e.g. megacities, metropoles); political-​economic function (e.g. port
city, capital city); and/​or thematic categories (e.g. “manufacturing” cities). Such
typologies serve the comparative approach and help urban historians to grasp
the development of specific places in their wider, global contexts.
This brings us to the second reason why urban history is instructive in our
multidisciplinary search for connections between historical and regional studies.
Urban history has been a multidisciplinary field of study from its outset, and
was highly influenced by the social sciences. Especially in the last quarter of the
20th century, inspired by sociological studies of urbanization and its accom-
panying societal challenges, urban histories employed empirical studies inspired
by theoretical models. Another characteristic of urban history which brings the
sub-​field close to regional studies is its preoccupation with planning. Modern
cities and city-​regions are, for a great part, the result of human intervention in
urban structures and networks. The intellectual history of planning measures is
an important theme for urban historians and overlaps with the field of planning
history, to which urban historians contribute as well (Hein, 2018).

Conclusion: A new engagement for regional history

Let us return to the initial position of this essay, in which we questioned the
absence of historians in debates about the development of European regions
(see Applegate, 1999 for an exception to this statement). A possible explanation
might be that contributors to this debate (academics, politicians, policymakers,
etc.) are not interested in the past at all. But this argument is not valid: impactful
social scientists include historical arguments in their analysis as well. A  more
plausible reason seems to be the national-​and global-​centred view of main-
stream history, and the subsequent marginalization of regional history. Regional
history has typically been associated with provincial themes, with Heimat, and
traditionally seen as the field of the amateur or antiquarian. The program-
matic goals of the historians who worked on a regionalized view on industrial
development did not inspire a new generation of historians, and, with notable
exceptions, historians who engage with regional perspectives do not connect to
current debates about how to deal with key policy questions such as the fates of
hi-​tech clusters, the emergence of economically vulnerable regions and the evo-
lution of spatial inequalities more generally –​they leave this to regional studies.
Fundamental to this problem are the methodological and theoretical differences
between historical and regional studies. Against the background of a disciplinary
overview, we have argued that this gap has been narrowed thanks to the institu-
tional turn in political science, economics and geography; indeed, that the insti-
tutional approaches bind history and regional studies together. Acham’s (1992)
threefold, sociological definition of institutions (1) as a system of social rules, (2) as
organized associations and, (3)  as forms of knowledge, connects to the holistic
view of historians. Within the social sciences, the conviction that institutions are
32

32  Marijn Molema and John Tomaney

the result of an evolutionary process has taken [Link] evolutionary convictions
suggest links with historians’ engagement with processes of continuity and change.
When historians are willing to embrace the concept of institutions and give it a
central position in their research designs, a practical bridge between the historical
and the social sciences is constructed. For sure, this bridge should be strengthened
by greater theoretical, methodological and conceptual clarification.
What historians can learn from social scientists is, first of all, to become
engaged with current problems such as regional economic development.
Regional inequality is part of history just as other phenomena such as nation-
alism or global connections. Regional inequality is an important theme for
policymakers on the European level, as well as in individual countries and
many regions suffer from economic vulnerability. Indeed, the salience of
regional questions has risen in the age of Brexit, Trump and European popu-
lism (Rodrìguez-​Pose, 2018). Social scientific research tries to understand how
regional inequality evolves, and what kind of human action helps vulnerable
regions. For example, political scientists study power structures, and evolu-
tionary economic geographers explore the capacity of regions to [Link]
use historical perspectives in their research approaches, but do not interact with
the historical sciences systematically. Such interaction has the potential to be
highly profitable. Social scientists run the risk of oversimplifying the past. Since
most social scientists tend towards the nomothetic approach, the warning as
formulated in the Methodenstreit comes into play once again: do social scientists
properly respect peculiarities, contexts and variations in history?
The added value of historians lies in their empirical strength and holistic
approach. Most historians eschew highly specified models and schemes, because
it hinders the exploration of the contingent factors that shape processes or
big events. Contingency refers to the more or less coincidental circumstances
which drive (regional) developments. The narrative skills of the historian help
not only to understand how, but also why things happened: by analyzing the
interplay between different circumstances and factors they help to grasp the
phenomenon of regional development (Popp and Wilson, 2007). As such, they
are well equipped to signify and understand processes of continuity and change.
However, an extremist stance on the ideographical approach has its own limits.
A  historian’s claim that every process or event is singular and incomparable
gives ammunition to the idea that “history is just one damned thing after
another” (Toynbee, 1954, p. 195). At the same time, there is a need to go fur-
ther than understanding these mechanisms in singular cases. A multidisciplinary
research agenda should, inspired by earlier traditions such as historical soci-
ology, be comparative in nature. The discrepancy between atheoretical history
and ahistorical theory might be bridged by a shared interest in institutions. As
such, the research approaches of historians and social scientists can cross discip-
linary divides and become complementary to each other. There is potential for
these domains to cooperate in order to better grasp the mechanisms of regional
development that are driven by institutional factors and circumstances.
 33

The institutional lens 33
References
Acemoglu, D. and Robinson, J., 2012. Why nations fail: The origins of power, prosperity, and
poverty. London: Profile Books.
Acham, K., 1992. Struktur, funktion und genese von institutionen aus
sozialwissenschaftlicher sicht. In: G. Melville, ed. Institutionen und [Link]
aspekte und mittelalterliche befunde. Cologne: Böhlau. Pp. 25–​71.
Applegate, C., 1999. A Europe of regions: Reflections on the historiography of sub-​
national places in modern times. American Historical Review, 104(4), pp. 1157–​1182.
[Link]
Baker, A. R. H., 2003. Geography and history: Bridging the divide. Cambridge: Cambridge
University Press.
Bolt, J. and Van Zanden, J. L., 2014. The Maddison Project: Collaborative research on
historical national accounts. The Economic History Review, 67(3), pp. 627–​651. [Link]
[Link]/​10.1111/​1468-​0289.12032
Boschma, R. and Martin, R., 2007. Editorial: Constructing an Evolutionary Economic
Geography. Journal of Economic Geography, 7(5), pp. 537–​ 548. [Link]
10.1093/​jeg/​lbm021
Braudel, F., 2009. History and the social sciences: The longue durée. Review (Fernand
Braudel Center), 32(2), pp. 171–​203.
Bulpitt, J., 2008. Territory and power in the United Kingdom: An interpretation. Manchester:
Manchester University Press.
Burke, P., 1990. The French historical revolution: The Annales School, 1929–​89. Cambridge:
Polity Press.
Calhoun, C., 2003. Why historical sociology? In: G. Delanty and E. Isin, eds. Handbook
of historical sociology. London, Thousand Oaks and New Delhi: Sage. Pp. 383–​393.
Drukker, J. W., 2006. The revolution that bit its own tail: How economic history changed our
ideas on economic growth. Amsterdam: Aksant.
Enflo, K. and Missiaia, A., 2019. Regional GDP before GDP: A methodological survey
of historical regional accounts. In: M. Molema and S. Svensson, eds. Regional economic
development and history. London: Routledge. Pp. 82–97.
Erlei, M., Leschke, M. and Sauerland, D., 1999. Neue institutionenökonomie. Stuttgart:
Schäffer-​Poeschel Verlag.
Ewen, S., 2016. What is urban history? Cambridge: Polity.
Febvre, L. and Bloch, M., 1929. À nos lecteurs. Annales d’histoire économique et sociale,
1(1), pp. 1–​2.
Fogel, R. W., 1964. Railroads and American economic growth: Essays in econometric history.
Baltimore: John Hopkins Press.
Fremdling, R. and Tilly, R. H., 1979. Industrialisierung und raum: Studien zur regionalen
differenzierung im Deutschland des 19. Jahrhunderts. Stuttgart: Klett-​Cotta.
Gadamer, H.-​G., 1960. Wahrheit und methode: Grundzüge einer philosophischen hermeneutik.
Tübingen: J. C. B. Mohr.
Hall, P. A. and Taylor, R. A., 1996. Political science and the new institutionalism. Political
Studies, 44, pp. 936–​957. [Link]
Hammond, J. L. and Hammond, B., 1920. The skilled labourer 1760–​1832. London:
Longmans, Green and Co.
Hazeu, C.A., 2007. Institutionele economie: Een optiek op organisatie –​en sturingsvraagstukken.
Bussum: Coutinho.
34

34  Marijn Molema and John Tomaney

Hein, C., 2018. The what, why, and how of planning history. In:  C. Hein, ed. The
Routledge handbook of planning history. London: Routledge. Pp. 1–​10. [Link]
10.4324/​9781315718996
Iggers, G. G., 1984. The “Methodenstreit” in international perspective. The reorienta-
tion of historical studies at the turn from the nineteenth to the twentieth century.
Storia Della Storiografia, 6, pp. 21–​32.
Iggers, G. G., 2006. The professionalization of historical studies and the guiding
assumptions of modern historical thought. In: L. Kramer and S. Maza, eds. A com-
panion to western historical thought. Malden, Massachusetts:  Blackwell. Pp. 225–​242.
[Link]
Keating, M., 2008. Thirty years of territorial politics. West European Politics, 31(1–​2),
60–​81. [Link]
Kiesewetter, H., 2004. Industrielle revolution in Deutschland. Regionen als Wachstumsmotoren.
Stuttgart: Franz Steiner Verlag.
Lipset, S. M., 1985. The revolt against modernity. In: S. M. Lipset, ed. Consensus and con-
flict. Essays in political sociology. New Brunswick, NJ: Transaction. Pp. 253–​294.
March, J. G. and Olsen, J. P., 2008. Elaborating the “New Institutionalism”. In: R. A. W.
Rhodes, S. Binder, and A. B. Rockman, eds. The Oxford handbook of political institutions.
Oxford:  Oxford University Press. Pp. 3–​ 20. [Link] oxfordhb/​
9780199548460.001.0001
Martin, R. and Sunley, P., 2011. Conceptualizing cluster evolution:  Beyond the life
cycle model? Regional Studies, 45(10), pp. 1299–​ 1318. [Link] [Link]/​10.1080/​
00343404.2011.622263
Martin, R. and Sunley, P., 2015. On the notion of regional economic resilience:
Conceptualization and explanation. Journal of Economic Geography, 15(1), pp. 1–​42.
[Link]
Molema, M., Segers, Y. and Karel, E., 2016. Introduction:  Agribusiness clusters in
Europe, 19th and 20th centuries. The Low Countries Journal of Social and Economic
History, 13(4), pp. 1–​16. [Link]
Monkkonen, E. H., 1994. Engaging the past:  The uses of history across the social sciences.
Durham: Duke University Press.
Moore Jr., B., 1966. Social origins of dictatorship and democracy:  Lord and peasant in the
making of the modern world. Boston: Beacon Press.
Nolte, P., 2002. Historische sozialwissenschaft. In: J. Eibach and G. Lottes, eds. Kompass der
geschichtswissenschaft: Ein handbuch. Göttingen: Vandenhoeck & Ruprecht. Pp. 53–​68.
North, D. C., [Link] costs, institutions and economic history. Zeitschrift für die
Gesamte Staatwissenschaft, 140(1), pp. 7–​17.
North, D. C., 1991. Institutions. Journal of Economic Perspectives, 5(1), pp. 97–​112. [Link]
[Link]/​10.1257/​jep.5.1.97
Pike, A., Dawley, S. and Tomaney J., 2010. Resilience, adaptation and adaptability.
Cambridge Journal of Regions, Economy and Society, 3(1), pp. 59–​70. [Link]
10.1093/​cjres/​rsq001
Polanyi, K., 1944. The great transformation. The political and economic origins of our time.
Boston, MA: Beacon Press.
Pollard, S., 1981. Peaceful conquest. The industrialization of Europe 1760–​1970. Oxford:
Oxford University Press.
Popp, A. and Wilson, J. F., 2007. Life cycles, contingency, and agency: Growth, develop-
ment, and change in English industrial districts and clusters. Environment and Planning
A, 39, pp. 2975–​2992. [Link]
 35

The institutional lens 35
Rodrìguez-​Pose, A., 2018. The revenge of the places that don’t matter (and what to
do about it). Cambridge Journal of Regions, Economy and Society, 11(1), pp. 189–​209.
[Link]
Ryan-​Collins, J., Lloyd, T. and MacFarlane, L., 2017. Rethinking the economics of land and
housing. London: Zed Books.
Schelsky, H., 1970. Zur soziologischen Theorie der Institution. In: H. Schelsky ed. Zur
Theorie der Institution. Dusseldorf: Bertelsmann Universitätsverlag. Pp. 9–​26.
Skocpol, T., 1985. Bringing the state back in: Strategies of analysis in current research.
In:  P. B. Evans, D. Rueschmeyer and T. Skocpol, eds. Bringing the state back in.
Cambridge: Cambridge University Press. Pp. 3–​35.
Skocpol, T. and Pierson, P., 2002. Historical institutionalism in contemporary political
science. In: I. Katznelson and H. V. Milne, eds. Political science: State of the discipline.
New York: W. W. Norton. Pp. 693–​721.
Stone, L., 1977. History and the social sciences in the twentieth century. In: C. F. Delzel, ed.
The future of history: Essays in the Vanderbilt University centennial. Nashville: Vanderbilt
University Press. Pp. 3–​43.
Thompson, E. P., 1963. The making of the English working class. London: Victor Gollancz.
Tilly, C., 2002. Neue angloamerikanische Sozialgeschichte. In: J. Eibach and G. Lottes,
eds., Kompass der Geschichtswissenschaft:  Ein Handbuch. Göttingen:  Vandenhoeck &
Ruprecht.
Tomaney, J., 2009. Region. In: R. Kitchin and N. Thrift, eds. International encyclopedia of
human geography, vol. 9. Oxford: Elsevier. Pp. 136–​150.
Tosh, J., 1999. The pursuit of history. Harlow: Longman.
Toynbee, A., 1954. A study of history, vol. 9. London: Oxford University Press.
Valdaliso, J. M., Magro, E., Navarro, M., Aranguren, M. J. and Wilson, J. R., 2014. Path
dependence in policies supporting smart specialisation strategies: Insights from the
Basque case. European Journal of Innovation Management, 17(4), pp. 390–​408. [Link]
[Link]/​10.1108/​EJIM-​12-​2013-​0136
von Ranke, L., 1925. Das politische Gespräch und andere Schriftchen zur Wissenschaftslehre,
Philosophie und Geisteswissenschaften; Bd. 2; Philosophie und Geisteswissenschaften. Halle:
M. Niemeyer.
Williamson, O. E., 1975. Markets and hierarchies: Analysis and antitrust implications (a study
in the economics of internal organization). New York: Free Press.
Windelband, W., 1915. Präludien: Aufsätze und Reden zur Einleitung in die Philosophie und
ihrer Geschichte. Tübingen: Mohr.
Wolf, N. and Rosés, J. R. eds., 2018. The economic development of Europe’s regions: A quan-
titative history since 1900. New York: Routledge.
36

3 
Behavioural economic geography
and regional history
Explaining uneven development from a
human perspective
Robert Huggins and Piers Thompson

Introduction
Explanations of economic development and growth differences across regions
are generally rooted in factors based on the structure, dynamics and organiza-
tion of firms, industries and capital. Emerging theories, however, are moving
toward a (re)turn to addressing the role of individual and collective behav-
iour in determining regional development outcomes (Obschonka et al., 2013).
Contemporaneously, a greater focus has been given to the behaviour of those
associated with the political economy of cities and regions as a form of agency
that determines the future of these places (Beer and Clower, 2014). Indeed,
leading scholars such as Chang (2013) and Piketty (2014), for example, high-
light the role of political leadership in determining economic outcomes. More
generally, a number of concepts relating to the behaviour of individuals and
groups of individuals have taken an increasingly central role in shaping an
understanding of why some places are better able to generate higher rates of
development and growth, and avoid the low-​road development trajectories and
associated higher rates of inequality found in weaker regions (Huggins and
Thompson, 2019; Huggins et al., 2018; Tabellini, 2010; Tubadji, 2013).
A behavioural approach is not strictly “new” in either comparative eco-
nomics or economic geography. Myrdal (1968), for instance, takes a behavioural
and cultural approach to understanding economic development across Asian
economies, in particular the role of religious and social (caste) systems. From
the 1960s there was also an emerging school of behavioural geography largely
concerned with identifying the cognitive processes that led to individuals and
communities codifying, reacting to, and recreating their environments. Pred
(1967), in particular, argued that economic geographic and locational distribu-
tion patterns are a consequence of the aggregate manifestation of decisional acts
made at the individual, group and/​or firm level.
Given a range of theoretical developments in recent years, it would appear
appropriate to examine behavioural explanations of regional growth and devel-
opment. Based on thinking from behavioural economics, some scholars suggest
that individual decision-​ making results from local influences experienced
through situations. Such “situations” equate to the dominant cultural traits
 37

Behavioural economic geography 37

embedded within the local communities where these “influences” are formed
(Storper, 2013). Behavioural economics generally concerns the integration
of psychological theories of behaviour as a means of explaining economic
action (Cartwright, 2014). Such theories have increasingly shown the limits
of rational-​choice theories in explaining economic as well as social action and
the underlying decision-​making processes of individuals in determining such
action (Hodgson, 2013).
The aim of this chapter is to propose a behavioural theory of how such
cities and regions evolve and develop. The chapter outlines the key constructs
and principles relating to this behavioural theory of urban and regional devel-
opment, and argues that the roots of behavioural differences across cities and
regions are co-​determined by two key factors, namely:  culture and person-
ality psychology. In essence, it is the interaction of these two factors that form
the human agency of individuals. Based on a historic perspective and building
on institutional theory, the chapter proposes that research needs to pay more
attention to the “pyschocultural life” of cities and regions when seeking to
understand differing types of behaviour and their relationship with economic
development outcomes. In summary, the fundamental objective is to argue that
in order to explain why economic development is more advanced in some
places than others, it is necessary to move beyond the known factors and
mechanisms of such development, and to analyze deeper and more fundamental
and historically entrenched causes of uneven development.

Human behaviour and regional development

Drawing on Simon’s (1982) notion of “bounded rationality”, behavioural eco-
nomics suggests that the minds of individuals are required to be understood in
terms of the environmental context in which they have evolved, resulting in
restrictions to human information processing, due to limits in knowledge and
computational capacity (Kahneman, 2003). Similarly, the rise in importance
given to cultural values in regional development theory has led to the emer-
gence of a “new sociology of development” that entwines the role of geog-
raphy with factors relating to individual and collective behaviour (Sachs, 2000;
Hadjimichalis, 2006; Spigel, 2013). As Clark (2015) argues, human behaviour
is fundamental to the social sciences in terms of understanding what people
do, where and why they do it, and the costs and benefits of this behaviour. In
order to understand the “aggregate” differences in socio-​economic activities
and performance there is a need to explore how these differences stem from
the experiences and actions of individual actors (Ariely, 2008; Storper, 2013).
The issue of how cultural factors impact on urban and regional develop-
ment, in innovation and entrepreneurial-​led development (Spigel, 2013), has
been increasingly debated in recent years (Tubadji and Nijkamp, 2015). In par-
ticular, recent work on socio-​spatial culture and the spatial nature of personality
psychology has sought to address knowledge gaps relating to the role of context
and environment in shaping behaviour (Beugelsdijk, 2009; Obschonka et al.,
38

38  Robert Huggins and Piers Thompson

2013). The fact that concepts such as “tolerance” are found to be associated
with the form and trajectory of urban and regional development is an indica-
tion of the apparent importance of psychological and behavioural dimensions
in determining these trajectories (Florida, 2002; Storper, 2013). From this psy-
chological perspective, Obschonka et  al. (2015), for example, draw from the
“five-​factor theory of personality –​the big five traits”, which is the predom-
inant personality model in contemporary psychological science  –​to explain
such differences in behaviour across places. Rentfrow et al. (2013) have identi-
fied three psychological profiles of regions –​friendly and conventional, relaxed
and creative, temperamental and uninhibited  –​covering the US states. They
find that in terms of economic prosperity a positive link exists with openness
and extraversion, while conscientiousness displays a negative association. Taken
together, it can be suggested that culture and personality psychology form the
psycho-​cultural behaviour of a city.
These psychocultural behaviours have the potential to be persistent and deeply
rooted in previously dominant economic activities, so that their influence is felt
many decades later. For example, a historically high level of mining is found
to be associated with lower entrepreneurial activity (Glaeser et al., 2015), posi-
tive attitudes to collective behaviour in the form of unionism (Holmes, 2006),
and preferences against redistribution (Couttenier and Sangier, 2015). Therefore,
there are strong reasons to suggest that the concentration of large-​scale coal-​based
industries in mature regions, for example, leave a lasting psychological imprint on
local culture of the old industrial regions, with selective outmigration resulting
in more optimistic and resilient individuals with positive and agentic mindsets
seeking new environments that offer new economic opportunities, resulting
in an indigenous population in the home region lacking in “entrepreneurial
spirit” (Stuetzer et al., 2016; Obschonka et al., 2018). Others suggest that these
regions now suffer from a case of “social haunting”, whereby there is a kind of
“ghosted” affective atmosphere that has endured long after the traditional indus-
tries associated with these regions have disappeared (Gordon, 1997; Bright, 2016).
In general, an acknowledgement of these cultural and psychological factors
has given rise to calls to examine the role of behavioural traits in promoting or
hindering regional renewal, resilience and transformation (Martin and Sunley,
2015). In recent years, scholars have also increasingly highlighted the role of
agency and associated institutions in fostering urban and regional development,
particularly through the welfare effects it generates within and through com-
munities (Bristow and Healy, 2014). A growing literature also suggests that the
development of cities and regions is linked to the types of key agents within
these places (Storper, 2013; Savini, 2016). Others note the role of agency in
hindering urban and regional development, with rent-​seeking behaviour being
an example of the potentially negative development outcomes resulting from,
for example, political agency (Ayres, 2014).
As indicated above, studies have frequently found a role for personality traits,
culture and institutions in determining economic growth, but equally it should
be noted that there are important differences between each of these factors.
 39

Behavioural economic geography 39

In  fundamental terms, these factors work at different levels of aggregation.
Whilst personality traits are individually held, community culture concerns the
shared values, beliefs and expectations of a group (Van Maanen and Schein,
1979). Hofstede (1980) refers to culture as systems of meaning within and
across ascribed and acquired social groups, and the collective programming of
the mind. Institutions, on the other hand, have been described as the rules of
the game (Hwang and Powell, 2005). In the literature stemming from economic
and political science, in particular, institutions act through rules, procedures and
agreements (North, 1990; 2005). The following sections analyze the role of
institutions, culture, personality psychology and agency in further detail.

Institutions and culture

Effective institutions have a key role in shaping the success of cities and regions
(Storper, 2010). In this sense, formal institutions are not necessarily restricted
to being the rules of the game, but also include regulatory aspects such as the
monitoring and enforcement of these rules (Scott, 2008). Institutions, there-
fore, can promote, hold back or even shape particular economic activities that
facilitate urban and regional competitiveness and development, such as entre-
preneurship and innovation (Peng et al., 2010; Stenholm et al., 2013). Although
enforcement and monitoring may vary across the cities and regions of nations,
informal institutions in the form of socially accepted norms within professional
and organizational interaction, as well as the cognitive dimensions of institutions
are also likely to vary across cities and regions (Scott, 2008). This means that
alongside the role of more formal institutions, it is becoming increasingly
recognized that the individual actions leading to places being better positioned
to take higher economic development roads are encouraged or limited by local
and regional influences, whereby these influences are formed by the dominant
cultural traits embedded in local communities (Storper, 2013). Furthermore,
the role of institutions in development has been acknowledged, and it is, there-
fore, a fairly reasonable assumption to extend this to cultural influences. As well
as the incentives and constraints that institutions and culture provide, they are
themselves also reflective of human agency (Bristow and Healy, 2014). As such,
a growing number of studies have considered the link between community
culture and economic activities, and resultant rates of economic development
(Huggins and Thompson, 2014; Tubadji and Nijkamp, 2015).
One of the most commonly analyzed aspects of culture is social cohesion,
which reflects the complexity of the cultural-​ economic growth relation-
ship. Whilst studies such as Easterly et  al. (2006) have found social cohesion,
as captured by the lack of ethnic fractionalization, to be positively associated
with economic growth, greater social cohesion is thought to reduce transac-
tion costs and improve cooperation and information flows (Beugelsdijk and
van Schaik, 2005; Kwon and Adler, 2014; Putnam et al., 1993). This is achieved
through the generation of greater trust from the development of social capital
(Dasgupta, 2011).
40

40  Robert Huggins and Piers Thompson

Institutions associated with publically-​funded education may have a key
role in developing the common social norms that benefit society (Gradstein
and Justman, 2000). Such cooperation and collaboration are considered to be
key components of the innovative activities required to achieve lasting eco-
nomic growth (Rutten and Boekema, 2007). It is no surprise, therefore, that
where deep divisions exist within communities these are often associated with
poorer economic performance (Aghion et al., 2004). However, social cohesion
may also have a downside when it leads to rent-​seeking behaviour by dom-
inant groups and produces insider–​outsider problems (Rodríguez-​Pose and
Storper, 2006).
As well as the role of social cohesion, another group of studies inspired by
Florida (2002) have found that open tolerant societies grow faster, reflecting the
attraction of both conventional human capital and a greater presence of the cre-
ative class (Florida et al., 2008).This may allow access to more ideas, but can also
help exploit the knowledge held and developed within an area as more diverse
sets of skills become available. Studies also suggest that migrants may be better
placed to see the opportunities available by providing a fresh pair of eyes and
drawing on international and extra-​local networks (Levie, 2007). For example,
Rodríguez-​Pose and Hardy (2015) examine the link between diversity and
entrepreneurial activity and find that place of birth diversity, rather than ethnic
background, has the strongest relationship with entrepreneurship.
Given these counteracting forces, it is no surprise that empirical studies
have provided mixed evidence in terms of the association between economic
growth and social cohesion with both significant positive (Guiso et al., 2004;
Knack and Keefer, 1997; Zak and Knack, 2001), and insignificant results found
(Rodríguez-​Pose, 2001). A  further group of studies find evidence of a rela-
tionship between social capital and improved performance at the level of indi-
vidual firms, but there is less evidence of this when considering a region or
locality as a whole (Cooke et al., 2005). Conflicts may also arise from the need
to distinguish between bonding and bridging social capital (Putnam, 2000).
Whereas bridging social capital may reduce rent-​seeking behaviour and pro-
vide access to knowledge, bonding social capital may promote the achievement
of non-​materialistic objectives, sometimes at the expense of economic growth
(Beugelsdijk and Smulders, 2003). In a similar vein, Bjørnskov (2006) finds that
only social trust is related to outcomes such as improved governance and life
satisfaction. On the other hand, some studies find that associational activity, and
the weak ties this generates, is of particular importance for economic activity
related to innovation (Hauser et  al., 2007). Furthermore, the form of social
capital and associated policy interventions may also vary depending on the
type of innovation sought, for example traditional, hidden or social (Murphy
et al., 2016).The complexities of the association between cultural measures and
economic performance measures are further compounded by studies that find
links between economic growth and individualism and a lack of collectivism
and more “masculine” cultures (Gorodnichenko and Roland, 2016; Huggins
and Thompson,  2016).
 41

Behavioural economic geography 41

A further direction for investigation returned to in more detail below is
that it is not always appropriate to study cultural dimensions purely in terms of
one aspect, but rather through specific combinations. For example, social cohe-
sion may have positive effects when combined with an openness to ideas, but
equally could form a further barrier when combined with a limited acceptance
of ideas from outside a community (Adler and Kwon, 2000). Similarly, Stephan
and Uhlaner (2010) use seven of House et  al.’s (2004) nine dimensions of
descriptive norms to create measures reflecting performance-​ based culture
(rewarding individual accomplishment rather than collective membership,
family relationships), and socially supportive culture (promoting support and
cooperation). They find that the norms combining to generate a performance-​
based culture are associated with demand for entrepreneurial activities, whilst
the norms generating socially supportive culture with close links to social cap-
ital increase the supply of entrepreneurs.

Psychology and personality science

In parallel with theoretical and empirical developments concerning the influ-
ence of culture on economic growth, another stream of literature has considered
how individual behaviour may have an impact at the aggregate and spatial level
(Obschonka et al., 2013; Stuetzer et al., 2016). There have long been studies
within psychology and personality science with regard to the different person-
ality traits possessed by individuals (Cattell, 1943). As indicated above, one of the
most commonly utilized approaches is that associated with the Big Five frame-
work, which consists of the identification and measurement of the following
concepts:  extraversion, agreeableness, conscientiousness, neuroticism and
openness (Costa and McCrae, 1992). Studies have found that more extravert
individuals tend to exhibit higher levels of sociability and energy, whilst pro-
social behaviour as captured by the notion of agreeableness is found to be linked
to factors such as social capital and reduced crime (Rentfrow, 2010; Rentfrow
et al., 2008). Conscientiousness is associated with individual levels of organiza-
tion and self-​discipline, with neuroticism reflecting differences in anxiety and
depression. The concept of openness is associated with individual differences in
curiosity and liberal values. Whilst such measures have traditionally been used
to examine how particular personalities may lead to particular behaviour and
outcomes at the individual level (Judge et al., 1999), the use of large surveys has
allowed much bigger databases that compare personality traits across nations
to be established (Schmitt et al., 2007). The size of these surveys has allowed
an examination of the distribution of personality traits across different areas of
countries such as the United States (Rentfrow, 2010; Rentfrow et al., 2009) and
the United Kingdom (Rentfrow et al., 2015).
Unlike cultural norms, which are formed at the group level, these personality
traits are based on the individual, but where a place has a relatively larger pro-
portion of individuals with particular types of personality, this may affect local or
regional factors such as economic or other quality of life outcomes (Obschonka
42

42  Robert Huggins and Piers Thompson

et al., 2013). There is a recognition that activities such as innovation and entre-
preneurship are likely to be promoted by certain cultural characteristics or the
presence of particular personality traits (Wyrwich, 2015). Rentfrow et al. (2015)
find a positive link between economic prosperity and openness and extraver-
sion, whilst conscientiousness displays a negative association. This is interesting
as Lee (2017) finds that conscientiousness in England and Wales is positively
associated with innovation as captured by patenting activity.
As with community culture, the majority of psychological research has
examined the impact of particular individual personality traits on a variety of
outcomes in isolation. However, in order to move beyond these single variable
perspectives at the local and regional level, there is a need for a more holistic
conceptualization of these factors (Rentfrow et al., 2013), especially as certain
configurations of traits have been found to be good predictors of developmental
outcomes such as: achievement at school (Hart et al., 2003); the development
of social support networks (Caspi, 2000); older age health issues such as the
prevalence of strokes and heart disease (Chapman and Goldberg, 2011); and
the likelihood of having spells in unemployment (Caspi, 2000). Understandably,
where such configurations are more prevalent in a locality or region it would
be expected that community outcomes will differ. Rentfrow et al.’s (2013) study
introduced a spatially-​oriented perspective on personality psychology by finding
and examining three spatial clusters across the United States described as: friendly
and conventional; relaxed and creative; and temperamental and [Link]
study is one of the first to develop a holistic spatially-​oriented psychological
perspective, and found numerous associations between the geographic clustering
of personality types and economic outcomes. However, it does not account for
the role of local cultural aspects when examining these relationships.

The psychocultural profile of cities and regions

When examining the culture and personality traits present within a city or
region, studies have frequently noted that the two are likely to be closely linked,
without explicitly examining this link. For example, in their study of voting
patterns Rentfrow et al. (2009) suggest a bi-​directional relationship between cul-
ture and the presence of particular personality [Link] is understandable given
research suggesting that in the long-​term the genetic and cultural evolution of
humans is interactive, that is cultural-​genetic co-​evolution (Van den Bergh and
Stagl, 2003). This co-​evolution can be related to theories of “generation” and
“collective memory”, or as “generational units” of meaningful collectives that
move through time with high degrees of self-​awareness (Lippmann and Aldrich,
2016). Given the above, it can be proposed that it is this interactive and co-​
evolving psychocultural behaviour, rather than an individual trait or aspect of
community culture that is most likely to be important for economic growth. In
order to understand this co-​evolution, it is necessary to examine the mechanisms
that have been suggested by previous studies linking the development of one to
the other. First, those links stemming from culture and influencing personality
are examined, followed by those running in the opposite direction.
 43

Behavioural economic geography 43

Initially, it should be recognized that personality traits are usually found
to be stable or slowly evolving at the individual level (Cobb-​ Clark and
Schurer, 2012). Rentfrow et al. (2015) highlight three routes that may result in
differences in personality developing within countries or regions. These three
mechanisms act through:  traditions and social norms; physical environment;
and selective migration. With regard to the first of these, community culture
provides the social norms that may influence an individual’s attitude and behav-
iour (Hofstede and McCrae, 2004). This can include a pressure to conform and
fit with the prevailing culture; for example exposure to a more diverse and tol-
erant population is found to be positively associated with greater acceptance
and openness (Pettigrew and Tropp, 2006). Similarly, a prevailing culture that
adheres to social rules is strongly linked to an individual’s habits and perceptions
of others (Bourgeois and Bowen, 2001).
The second mechanism, physical environment, is less likely to be directly
linked to community culture but can influence personality traits (van de Vliert,
2009) and the underpinning culture present (Huggins and Thompson, 2016).
For example, agreeableness and conscientiousness may develop as a coping
mechanism in challenging environments (Steel et al., 2008; Jokela et al., 2015).
The third mechanism, selective migration, may also be linked to community
culture, whereby migrant individuals base their choice of location on commu-
nity cultures that provide a good psychological fit with their own personality
traits. Indeed, Jokela et al.’s (2015) finding that those with high openness seek
out communities with similar traits is consistent with this proposition.
As well as community culture influencing the personalities of those residing
in these communities, it is just as plausible that personality at an individual
level will affect the development of community culture through its influence
on social norms and attitudes. Although a particular community culture may
attract or dissuade the inward migration of certain personalities, once within
a locality such personality traits may influence community culture evolution.
This may be a slow process, but where, for example, a less socially cohesive
community culture attracts individuals of a more extravert and less agreeable
nature, such individuals are likely to reinforce the reproduction of existing
social norms associated with such a local community culture. The potential for
a reinforcing pattern to behavioural development is captured by studies such
as Florida (2002), which suggests that the presence of bohemians attracts other
high-​skilled individuals. This presumably operates through those pursuing a
bohemian lifestyle helping to generate a tolerant community culture that does
not exclude outsiders, particularly more extravert individuals who are willing
to explore new ideas. At the other end of the spectrum, where agreeableness is
higher it is suggested that outward migration is reduced (Boneva et al., 1998;
Jokela et  al., 2008). This helps to generate a more socially cohesive society,
potentially to such an extent that outsiders are excluded (Rodríguez-​Pose and
Storper, 2006). Societies with more bonding social capital have been found
to place greater weight on non-​materialistic outcomes and a higher value on
family life (Beugelsdijk and Smulders, 2003), which may become engrained in
the social norms of the community culture present.
44

44  Robert Huggins and Piers Thompson

In an attempt to empirically measure differences in psychocultural behav-
iour across cities, regions and localities, Huggins et al. (2018) undertook a study
of the UK that identified measures for three different forms of psychocultural
behaviour at the local level, namely: Inclusive Amenability, Individual Commitment
and Diverse Extraversion. These three forms of psychocultural behaviour inte-
grate both cultural and personality traits, with inclusive amenability refer-
ring to places that are high in agreeableness, social cohesion, feminine and
caring activities and adherence to social rules, but low in openness. Individual
Commitment consists of places high in conscientiousness, engagement with
education and adherence to social rules but with low levels of collective activity.
Diverse Extraversion concerns places where there are high levels of extraversion
and openness, as well as low levels of [Link] regard to psychocultural
behaviour that can be regarded as inclusively amenable, localities in more geo-
graphically peripheral parts of Britain tend to display higher levels of such
behaviour, as shown by Figure 3.1.
At the regional level, Wales, Scotland and North East England have the
highest rates of inclusive amenability, with London having by far the lowest rate.
At the local level, there is found to be a significant negative relationship between
rates of inclusive amenability and economic competitiveness. This suggests that
places portraying behaviour that tends to be agreeable and cohesive do not
generally generate the highest rates of competitiveness and economic perform-
ance. In other words, whilst such culture and psychology may have significant
positive attributes with regard to social development, such attributes do not
always appear to be the “right” ingredients for stimulating economic growth
and development (Huggins and Thompson, 2017).
Behaviour based on individual commitment shows strong geographical
differences across Britain. Localities situated in South East England, the East of
England, South West England and the East Midlands have the highest concen-
tration of this behavioural form. Although parts of London appear to have above
average concentrations of these individuals, the region as a whole has, perhaps
a lower density of these types of individuals to that which might be expected.
The regions where this behaviour is least prevalent are Wales, Scotland, North
East England and North West England.
As shown by Figure  3.1, diverse extravert behaviour is most commonly
clustered in London and parts of South East England. Whilst London, South
East England and South West England have the highest rates of diverse extra-
version, the least diverse extravert behaviour is found in the regions of West
Midlands, East Midlands and Yorkshire and Humber. A  positive relationship
is found between diverse extraversion and economic competitiveness at the
local level (Huggins and Thompson, 2017). This suggests that cosmopolitanism
and outwardly-​facing behaviour tends to foster greater economic strength and
competitiveness, and hint at the possibility that some regions and localities
possess the “wrong” type of behaviour when it comes to catalyzing economic
development.
 newgenrtpdf
 45
Inclusive amenability Individual commitment Diverse extraversion

2.4 1.6 3.2

1.6 2.4
0.8
0.8 1.6
0.0 0.8
0.0
–0.8 –0.8 0.0
–1.6 –0.8
–1.6
–1.6
–2.4 –2.4 –2.4
–3.2 –3.2 –3.2
–4.0 –4.0
–4.0

Figure 3.1 Psychocultural behavioural profiles

46

46  Robert Huggins and Piers Thompson

Human agency and regional economic history

The above sections suggest that the pyschocultural profile of a city or region
has a strong impact on its economic development trajectory. However, it says
little about this process and how the associated development mechanisms are
likely to operate in practice. In this section it is argued that the psychocultural
profile of a city or region is a key determinant for shaping the extent to which
a sufficient number of the types of human agents in a city or region are cap-
able of catalyzing the entrepreneurship and innovation required for economic
transformation and long-​term development. To achieve this, some examples
from the economic history of a number of cities and regions during periods
of transformation are utilized. Recent research has increasingly indicated that
human agency is based on a rationality that is spatially bounded (Huggins and
Thompson, 2019; Pike et  al., 2016). In particular, cities and regions them-
selves produce a spatially bounded rationality based on their underlying
psychocultural behavioural profiles that determines the forms and types of
human agency apparent in a given city or region, and subsequently the nature
of knowledge, innovation, and development. Innovation, entrepreneurship and
creativity are social processes that involve groups of people who build off one
another historically, and are the products of cities and regions that act as the
key organizing unit for these activities, bringing together the firms, talent and
other regional institutions necessary for them to thrive (Florida et al., 2017).
Similarly, the symbiotic relationship between key agents and their location is
found in research relating to the role of a limited number of “star” scientists in
promoting the innovation performance of certain cities and regions (Zucker
et al., 1998; Moretti, 2012).
Fundamentally, agency refers to acts carried out intentionally, and during
periods of economic transformation in particular cities and regions there are
likely to be a relatively small number of key human agents that are the core
drivers of such transformation. Through such agency, regions can become
“incubators of new ideas” and provide opportunities for entrepreneurship to
take place, as well as for discovering valuable new knowledge (Huggins and
Williams, 2011; Hülsbeck and Pickavé, 2014). Successful urban and regional
economic development and transformation may emerge from forms of agency
that promote institutional and cultural change, especially through the intro-
duction of economically efficient institutions and cultural change and diversity
across time. Regions that are unable to effectively transform economically and
industrially may be marked by agents that promote institutional and cultural
persistence, in particular through rent-​seeking institutions, and cultural repro-
duction and homophily across time. In terms of the spatiality of these changes,
whilst the agency promoting economic transformation is based on localized
networks of agents, broader transformations relating to “ideas” that often pre-
cede industrial and economic innovation are based on agents that are connected
through networks across cities and regions. Therefore, the notion of human
agency is a useful concept with regard to integrating industrial, structural, insti-
tutional and evolutionary explanations of urban and regional development and
 47

Behavioural economic geography 47

transformation, as it represents a fundamental building block of the mechanisms
and processes underpinning economic change.
A compelling analysis of the role of human agency in propelling economic
transformation is provided by Mokyr (2017) who argues that from 1500 to
1700, parts of European society –​largely the educated elite –​developed a set
of cultural traits and accompanying institutions that were highly attuned to
fostering the forms of intellectual innovation and knowledge that ultimately
propelled the Industrial Revolution. Specifically, Mokyr (2017) suggests that
key cultural changes relating to the increased value placed on innovation and
ideas occurred during this period, and through the formation of a market for
ideas, a relatively small number of cultural entrepreneurs across Europe were
responsible for driving this cultural change. Mokyr (2017) describes how these
entrepreneurs stimulated (directly and indirectly) economic evolution on an
unprecedented scale, with the interaction of cultural evolution and evolutionary
biology resulting in the emergence of adaptive agents who chose whether to
adopt new ideas from a series of cultural menus that led to intellectual innovation.
Following in the footsteps of Mokyr (2017), it is of interest to examine
whether his key agent thesis of development at a pan-​European level holds
up when regional and city territories are analyzed. Although a full analysis
would require extensive new data collection, one useful and bounded source
of information is Peter Hall’s (1998) Cities in Civilization. In this book, which
consists of more than 1,000 pages, Hall describes the evolution of 18 cities
and regions at the height of their innovative and transformative prowess on a
chapter-​by-​chapter basis. Although Hall’s own analysis may be partial in terms
of its coverage, it does give a good indication of the protagonists at the heart
of the regional transformation process. Therefore, it is of value to undertake a
content analysis of Hall’s text in order to identify the extent of key agents and
their role in catalyzing transformation.
As part of the results of the content analysis, Table  3.1 presents a sum-
mary of the key agents and the time of their most important agentic activity
highlighted by Hall (1998) in nine of the leading cities and regions he describes,
ranging from the emergence of Silicon Valley and the Bay Area of San Francisco
between 1950 and 1990 to as far back the Roman Empire from 50 BC to AD
100. Many of the agents will be familiar to those with a historic knowledge of
these cities, and what is marked is that in the majority of cases approximately
20–​30 agents are considered to have been the major catalysts that fuelled the
urban and regional innovation and development that took place. This begins
to suggest that much like Mokyr’s (2017) arguments regarding the role of an
elite group of entrepreneurial agents in triggering the role of the Industrial
Revolution, as well as research on the role of star scientists in underpinning
regional innovation success (Zucker et al., 1998; Moretti, 2012), a more historic
analysis of urban and regional transformation is likely to pinpoint a relatively
small number of agents as being central catalyzing forces propelling the evolu-
tion of their respective urban and regional ecosystems.
In order to consider the process of transformative regional evolution a little
more closely,Tables 3.2 and 3.3 present more detailed findings from the content
48

Table 3.1 Agents of innovation, transformation and development for a cross section of

cities and regions at key points in their evolution

City (time Key agents

period)

Silicon Valley/​ Alexander M. Poniatoff (1944–​1956); Bill Gates (1981); Charles Litton (1932);
Bay Area Cy Elwell (1909); Ed Roberts, Leslie Solomon, Roger Melen (1974);
(1950–1​ 990) Frederick Terman (1951); Gary Kildall (1973); Harold Lindsay (1947–​
1956); Lee De Forest (1906); Robert Noyce (1956); Stephen Wozniak and
Steven Jobs (1976);William Weber, Sigurd Varian and Russell Varian (1939).
Los Angeles Adolph Zukor (1903–​06); Alexander Black (1893–​1994); Thomas Armat
(1910–1​ 945) and Charles Francis Jenkins (1895–​1896); Carl Laemmle (1912); David
W. Griffith (1915); Harry Cohn (1920); Horace Henderson Wilcox
(1883); John P. Harris and Harry Davis (1905); Louis B. Mayer (1924);
Thomas Alva Edison (1887–​1889); Harry, Albert, Sam and Jack Warner
(1918); William Fox (1903).
New York Alexander G. Bell (1876); Clifford M. Holland (1919–​1927); John
(1880–1​ 940) B. Dunlop (1888); Thomas Alva Edison (1879); Frank Julian (1887);
John Augustus (1844); John Francis Hylan (Red Mike) (1920s);
Fiorello La Guardia (1934–​1940); Ottmar Mergenthaler (1885);
Othmar Hermann Ammann (1931); Robert Fulton (1807); George
Westinghouse (1868); William J. Wilgus (1903–​1913).
Berlin Emil Rathenau (1889–​1890); Frank J. Sprague (late 1880s); Friedrich
(1840–1​ 930) Wilhelm Anton von Heynitz (1778–​1799); Sir Humphry Davy (1808);
John Gibbs and Lucien Gaulard (1880); Nikola Tesla (1887); Samuel
Morse (1837); Werner von Siemens (1879).
Detroit Charles E. Duryea and J. Frank Duryea (1893); Gottlieb Daimler and Karl
(1890–1​ 915) Benz (1885); Etienne Lenoir (1860); Henry Ford, Ransom E. Olds and
Charles B. King (1870s–​1890s); Oliver Evans (1875); Nicolaus A. Otto
(1876); Ransom E. Olds (1899); Alfred P. Sloan (1918 onwards);
Wilhelm Maybach (1893).
London Jeremy Bentham (1784); Sir Edwin Chadwick (1839–​1942); Dr. John Snow
(1825–1​ 900) (1849); Henry Fielding (1750); John Howard, Sir William Blackstone and
William Eden (1778); John Nash (1820s); Lord John S. Eldon (1819);
Lord Charles Grey (1830); Lord John Russell (1839); Messrs Haden of
Trowbridge (1842); Prince of Wales (1865); Colonel Sir Charles Rowan
and Sir Richard Mayne; William Allen (1817); William Farr (1841).
Glasgow James Beaumont (1828); Henry Bell (1812); William Denny (1818); John
(1770–1​ 890) Elder and Alexander C. Kirk (1854); John Golborne (1775); Samuel Hall
(1834); Patrick Miller and William Symington (1788); David Mushet
(1801); David Napier (1819 and 1822); James Napier (1830); J. C. Perier
(1775); John Robertson (1812); John Roebuck (1760); Sir John Biles;
William Symington (1801); Tod and MacGregor (1836 and 1850); James
Watt (1769); John Wilkinson (1787); Thomas Wingate (1838).
Manchester John Aikin (1795); Richard Arkwright (1771); Edmund Cartwright
(1760–1​ 830) (1786); Peter Clare (1778); Samuel Crompton (1779); Abraham Darby
(1709); Edward III (1400); William Galloway (1790); Grimshaw of
Gorton (1790); James Hargreaves (1765); John Kay (1733); William Lee
(1589); Thomas Lombe (1721); Robert Peel; Richard Roberts (1825);
Andrew Ure (1835); John Wyatt and Lewis Paul (1741).
Rome Augustus Caesar (22BC); Julius Caesar (1BC); Claudius (52AD); Domitian;
(50BC–​ Sextus Julius Frontinus (96AD); Marcus Vipsanius Agrippa (20BC–12BC);
AD100) Marcus Vitruvius Pollio (30BC–15BC).

Source: Based on the authors’ analysis of Hall (1998)

 49

Table 3.2 Agents of innovation, transformation and development in Manchester

1760–​1830

Agent Occupation Year Key contribution to economic development

(Time)

Edward III Monarch 1400 Brought Flemish weavers to settle in

various places in the north of England
including Manchester, Rossendale and
Pendle in 14th century (latterly the
Weaver’s Act 1558 freed the weaving
industry from medieval regulations,
and thus Lancashire enjoyed a rare
degree of economic freedom)
William Lee Clergyman 1589 Invented stocking frame, creating
complex domestic production
Abraham Darby Ironmaster and 1709 Smelted iron with coal
Quaker
Thomas Lombe Silk-​thrower 1721 Built first recognizable factory
John Kay Innovator and 1733 Innovated flying shuttle which increased
Iindustrialist the efficiency of weaving two-​fold.
This eventually created pressure on
the supply of weft needed for weave
John Wyatt and Carpenter and 1741 Innovated and applied the system of
Lewis Paul innovator spinning cotton by rollers
James Hargreaves Weaver, 1765 Invented the spinning jenny which
carpenter reduced the labour required to
produce yarn
Peter Clare Clockmaker 1778 Proposed to establish philosophical
school emphasizing mechanics and
similar subjects
Samuel Biographer, 1779 Invented the mule machine, which was
Crompton industrialist cheap, compact, light and could be
hand-​operated in an ordinary house
Edmund Professor of 1786 Improved weaving machine model
Cartwright Poetry which stopped if the thread was
accidentally broken
Robert Peel Industrialist 1787 Improved the factory system. In 1787 he
built an integrated spinning, weaving
and printing factory
William Mining engineer, 1790 Established firms to make water wheels
Galloway professor, and
industrialist
Richard Roberts Millwright 1825 Developed self-​acting mule in 1825
at the request of the manufacturers
who were afflicted by the strikes of
spinners.
Richard Economist, 1771 Arkwright did the most to make
Arkwright industrialist the spinning machine useful for
production
Grimshaw of Industrialist 1790 Developed power loom
Gorton

Source: Based on the authors’ analysis of Hall (1998)

50

Table 3.3 Agents of Innovation, Transformation and Development in Glasgow

1770–​1890

Agent Occupation Year (Time) Key Contribution to Economic Development

James Watt Instrument 1769 Invented Watt Steam Engine

maker
Sir John Biles Professor Devised new means of turbine efficiency
John Roebuck Inventor and 1760 Established first Scottish blast furnaces
Industrialist
John Golborne Contactor, 1775 Deepened the Clyde Channel
Navigation
engineer
Robert Wilson Founder of blast 1780 Set up a foundry with first furnace
furnaces
John Wilkinson Ironmaster 1787 Iron was first used in part-​construction
of a barrage called ‘Trial’
Patrick Miller Banker and 1788 Sailed steamboat on Dalswinton Lake.
and William Engineer The speed of the boat was about 5
Symington miles per hour
David Mushet Engineer 1801 Found that materials mined in the
district abandoned as “wild coal”
contained at least 30–​50% iron when
raw and up to 70% when calcined
William Engineer and 1801 Fitted a steamboat with a Watt engine
Symington Inventor
Henry Bell Steamboat 1812 Started first commercial steamship
developer services
John Robertson Engineer 1812 The original ‘Comet’ engine was
developed. He developed it mainly for
land travel
William Denny Ship builder 1818 Built 30 horsepower engines
David Napier Marine 1819 and Built the boiler for ‘Comet’ and engines
engineer 1822 for the Talbot. Also established the first
commercial steamship line between
Liverpool, Greenock and Glasgow
James Manager 1828 Discovered how to use hot instead of
Beaumont cold air in furnace blast
James Napier Engineer and 1830 Invented the horizontal tubular boiler,
Inventor giving 25–​30 percent fuel saving
Samuel Hall Engineer 1834 Patented a condenser
David Tod Engineers and 1836 and Opened the first Clyde iron shipyard.
and John Shipbuilders 1850 Regular transatlantic passenger traffic
MacGregor by iron steamship
Thomas Engineer 1838 Made the first Glasgow based
Wingate transatlantic steam voyage in 1838, the
‘British Queen’
John Elder and Engineer 1854 Elder’s compound engine of 1854 and
Alexander Kirk’s triple-​expansion engine of 1886
C. Kirk offered price and speed advantages

Source: Based on the Authors’ Analysis of Hall (1998)

 51

Behavioural economic geography 51

analysis of Hall (1998) for the cases of the growth of the textile industry in
Manchester between 1760 and 1830 and the shipping industry in Glasgow
between 1770 and 1890. In both cases it is clear that a series of entrepreneurs,
industrialists and intellectuals built on the success of their predecessors in each
city through a process of knowledge accumulation and both radical and incre-
mental innovation that typifies the regional evolutionary framework discussed
above. Similarly, the establishment of economic scale and interdependence across
a range of associated economic activities within each city echoes not only the
Perroux-​Hirschman theses of development, but are also markedly in line with
analyses of the contemporary ecosystem of Silicon Valley (Perry Piscione, 2013).
As with Silicon Valley, the initial triggers of the regional transformation can be
much discussed and debated, and in the case of Manchester it is suggested by
Hall (1998) that it can be partly traced to the migration of Flemish weavers to
the north of England at the beginning of the 15th century, which indicates some
of the temporal issues that need to be considered when looking at notions of
economic evolution. Indeed, the ecosystems in both Glasgow and Manchester
were ultimately to enter a period of demise. However, they are instructive in
terms of highlighting the role of agency in fostering transformation within
regional ecosystems.
Clearly, the source of long-​term regional economic success and renewal is to
create a behavioural environment that fosters positive lock-​in allowing innov-
ation to become psychoculturally embedded. As illustrated by Table  3.4, this
would seem to require agents within regional economic systems that promote
institutional and cultural change. In particular, such agents will be responsible
for creating local economic institutions that are efficient in the sense that they
incentivize innovation and help remove barriers to change, which can be seen
in the case of both Glasgow and Manchester. Also, successful regions are those
that are more likely to experience social and economic change and diver-
sity across time, which again helps avoid an environment of stasis and instead
stimulates one of vibrancy. The problem for mature regional economies is that
they are likely to have suffered from negative lock-​in, whereby the available
agency has led to institutional and psychocultural persistence both economic-
ally and socially, resulting in an institutional environment within which rent-​
seeking behaviour becomes commonplace (Huggins and Pugh, 2015).

Table 3.4  Behavioural sources of positive and negative regional lock-​in

Positive lock-​in regions Negative lock-​in regions

Agents Agents promoting institutional and Agents promoting institutional and

psychocultural change psychocultural persistence
Culture Psychocultural change and diversity Psychocultural reproduction and
across time homophily across time
Institutions Economically efficient institutions Rent-​seeking institutions
52

52  Robert Huggins and Piers Thompson

Conclusions
This chapter has principally argued that the economic development histories
and future trajectories of cities and regions are at least partly determined by
their behavioural life in terms of the underlying and dominant cultural and psy-
chological traits of individuals within these places. These traits determine the
forms of human agency to be found in particular regions, and in terms of
the mechanisms and processes through which the psychocultural profile of a
place may impact upon development outcomes, it can be proposed that some
psychocultural profiles may better facilitate the type of entrepreneurial and
innovative agency that leads to economic development and growth (Hauser
et al., 2007; Stuetzer et al., 2016; Wyrwich, 2015). Indeed, the type and nature
of human agency existing within cities and regions at particular points in their
development is a significant factor in explaining the capacity of these places to
achieve economic transformation and continued upward development trajec-
tories. In other words, human agency is likely to be one of the key rooted drivers
associated with more traditional explanatory causes of economic development
and transformation, and should be considered seriously when addressing the
routes available to cities and regions in their bid to foster transformation.
It has been shown that complementary community cultures and personality
traits reinforce one another to create quite distinct psychocultural behavioural
profiles (Boneva et  al., 1998; Hofstede and McCrae, 2004; Rentfrow et  al.,
2013).Three forms of psychocultural behavioural profile appear [Link]
one, Diverse Extraversion, displays lower levels of social cohesion and neuroti-
cism and higher levels of extraversion and openness, the other psychocultural
profiles display higher levels of agreeableness, social cohesion and collective
traits  –​defined as Inclusive Amenability  –​or independent and self-​sufficient
characteristics, defined as Individual [Link], although individual
aspects of community culture and personality psychology traits have been
linked to urban and regional economic development and growth (Huggins
and Thompson, 2015; Obschonka et al., 2015; Stuetzer et al., 2016), it can be
concluded that they may be even more strongly influenced by combinations
that generate specific forms of a holistic psychocultural behavioural profile
(Rentfrow et al., 2013).This is supported by emerging empirical evidence indi-
cating that the psychocultural behaviour of cities, localities and regions helps
to shape their long-​term development trajectories. Regions that have relatively
atomized behavioural environments with high levels of individual commitment
tend to enjoy competitiveness and development benefits. Similarly, places with
high rates of cultural diversity and extravert individuals also have relatively high
levels of competitiveness.
On the other hand, more socially inclusive regions, with a significant number
of people with amenable and agreeable personality traits, experience relatively
low rates of competitiveness. To a large extent, the findings make intuitive
sense with, for example, the individual commitment found in competitive
localities and regions being a manifestation of a “personal competitiveness”
 53

Behavioural economic geography 53

that subsequently becomes visible at an aggregated spatial level. Clearly, how-
ever, the relationship between psychocultural human behaviour and urban and
regional development and competitiveness is unlikely to be a direct one. It is
more likely that behaviour initially impacts upon on other sources of com-
petitiveness such as the form and efficiency of local agents and institutions.
Positive urban and regional growth and development, therefore also requires
high-​quality institutions, in the form of growth-​enabling rules and incentives,
alongside the types of capital suggested by regional competitiveness theory.
Whilst institutions can be considered to be the rules of the game governing
growth processes, cultural and psychological traits encompass the extent to
which such rules are adhered to, as well as the way in which they foster future
institutional change.
In conclusion, it is suggested that within the field of urban and regional
economic development research there is a need for further theoretical integra-
tion, particularly through the deployment of a behavioural conceptual lens that
gives more consideration to historical context. Behavioural economic geog-
raphy, encompassing culture, psychology and the agency of individuals, poten-
tially provides new insights into the persistence of the long-​term unevenness of
development across regions. In particular, psychocultural behavioural patterns,
and their evolution, provide a basis for understanding the type and nature of
human agency that exists within cities and regions, and the institutions such
agency generates. Furthermore, behavioural-​based frameworks incorporating
cultural and psychological aspects help us to understand why particular agents
within a city or region may possess a proclivity towards fostering the forms of
innovation that propel development and transformation. Finally, although the
focus of urban and regional development narratives concerns explanations of
economic outcomes, there is scope to consider further theoretical connections
with wider development goals beyond economic growth, such as those related
to social development, well-​being and the sustainable development of cities and
regions.

References
Adler, P. S. and Kwon, S. W., 2000. Social capital:  The good, the bad, and the ugly.
In: E. L. Lesser, ed. Knowledge and social capital: Foundations and applications. Boston,
MA: Butterworth Heinemann. Pp. 89–​115.
Aghion, P., Alesina, A. and Trebbi, F., 2004. Endogenous political institutions.
Quarterly Journal of Economics, 119(2), pp. 565–​ 612. [Link] [Link]/​ 10.1162/​
0033553041382148
Ariely, D., 2008. Predictably irrational: The hidden forces that shape our decisions. New York,
NY: Harper.
Ayres, S., 2014. Place-​ based leadership:  Reflections on scale, agency and theory.
Regional Studies, Regional Science, 1(1), pp. 21–​ 24. [Link]
21681376.2013.869424
Beer,A. and Clower,T., 2014. Mobilizing leadership in cities and regions. Regional Studies,
Regional Science, 1(1), pp. 5–​20. [Link]
54

54  Robert Huggins and Piers Thompson

Beugelsdijk, S., 2009. Entrepreneurial culture, regional innovativeness and economic
growth. In: A. Freytag and [Link], eds. Entrepreneurship and culture. Heidelberg and
New York, NY: Springer. Pp. 129–​154.
Beugelsdijk, S. and Smulders, S., 2003. Bonding and bridging social capital: Which type
is good for economic growth? In: W. Arts, L. Halman and J. Hagenaars, eds. The cul-
tural diversity of European unity: Findings, explanations and reflections from the European
Values Study. Boston, MA: Brill Leiden. Pp. 147–​184.
Beugelsdijk, S. and van Schaik,T., 2005. Differences in social capital between 54 Western
European regions. Regional Studies, 39(8), pp. 1053–​1064. [Link]
00343400500328040
Bjørnskov, C., 2006. The multiple facets of social capital. European Journal of Political
Economy, 22(1), pp. 22–​40. [Link]
Boneva, B. S., Frieze, I. H., Ferligoj, A., Jaršová, E., Pauknerová, D. and Orgocka, A.,
1998. Achievement, power, and affiliation motives as clues to emigration desires: A
four-​countries comparison. European Psychologist, 3(4), pp. 247–​254. [Link]
10.1027/​1016-​9040.3.4.247
Bourgeois, M. J. and Bowen, A., 2001. Self-​organization of alcohol-​related attitudes
and beliefs in a campus housing complex: An initial investigation. Health Psychology,
20(6), pp. 434–​437. [Link]
Bright, N. G., 2016. ‘The lady is not returning!’:  Educational precarity and a social
haunting in the UK coalfields. Ethnography and Education, 11(2), pp. 142–​157. [Link]
[Link]/​10.1080/​17457823.2015.1101381
Bristow, G. and Healy, A., 2014. Regional resilience: An agency perspective. Regional
Studies, 48(5), pp. 923–​935. [Link]
Cartwright, E., 2014. Behavioral economics. 2nd ed. Abingdon and New  York,
NY: Routledge.
Caspi, A., 2000. The child is father of the man: Personality continuities from childhood
to adulthood. Journal of Personality and Social Psychology, 78(1), pp. 158–​172. [Link]
[Link]/​10.1037//​0022-​3514.78.1.158
Cattell, R. B., 1943. The description of personality: Basic traits resolved into clusters.
Journal of Abnormal and Social Psychology, 38(4), pp. 476–​ 506. [Link]
10.1037/​h0054116
Chang, H. J., 2013. Hamlet without the Prince of Denmark:  How development has
disappeared from today’s ‘development’ discourse. In:  D. Held and C. Roger, eds.
Global governance at risk. Cambridge: Polity Press. Pp. 129–​148.
Chapman, B. P. and Goldberg, L. R., 2011. Replicability and 40 year predictive power of
childhood ARC types. Journal of Personality and Social Psychology, 101(3), pp. 593–​606.
[Link]
Clark, G. L., 2015. Behavior, cognition and context, mimeo. Oxford:  Smith School of
Enterprise and the Environment, Oxford University.
Cobb-​Clark, D. A. and Schurer, S., 2012. The stability of big-​five personality traits.
Economic Letters, 115(1), pp. 11–​15. [Link]
Cooke, P., Clifton, N. and Oleaga, M., 2005. Social capital, firm embeddedness and
regional development. Regional Studies, 39(8), pp. 1065–​ 1077. [Link]
10.1080/​00343400500328065
Costa Jr., P. T. and McCrae, R. R., 1992. Revised NEO personality inventory (NEO-​PI-​R)
and NEO five-​factor inventory (NEO-​FFI) professional manual. Odessa, FL: Psychological
Assessment Resources.
 55

Behavioural economic geography 55

Couttenier, M. and Sangier, M., 2015. Living in the Garden of Eden: Mineral resources
and preferences for redistribution. Journal of Comparative Economics, 43(2), pp. 243–​
256. [Link]
Dasgupta, P., 2011. A matter of trust: Social capital and economic development. In: J.
Y. Lin and B. Pleskovic, eds. Annual Bank Conference on Development Economics
(ABCDE)  –​Global 2010:  Lessons from East Asia and the global financial crisis.
Washington, D.C: World Bank. Pp. 119–​155.
Easterly, W., Ritzen, J. and Woolcock, M., 2006. Social cohesion, institutions, and
growth. Economics and Politics, 18(2), pp. 103–​ 120. [Link]
j.1468-​0343.2006.00165.x
Florida, R., 2002. The rise of the creative class and how it’s transforming work, life, community
and everyday life. New York, NY: Basic Books.
Florida, R., Adler, P. and Mellander, C., 2017. The city as innovation machine. Regional
Studies, 51(1), pp. 86–96. [Link]
Florida, R., Mellander, C. and Stolarick, K., 2008. Inside the black box of regional
development:  Human capital, the creative class and tolerance. Journal of Economic
Geography, 8(5), pp. 615–​659. [Link]
Glaeser, E. L., Kerr, S. P. and Kerr, W. R., 2015. Entrepreneurship and urban growth: An
empirical assessment with historical mines. Review of Economics and Statistics, 97(2), pp.
498–​520. [Link]
Gordon, A., 1997. Ghostly matters: Haunting and the sociological imagination. Minneapolis,
MN: University of Minnesota Press.
Gorodnichenko,Y. and Roland, G., 2016. Culture, institutions and the wealth of nations.
Review of Economics and Statistics, 99(3), pp. 402–​ 416. [Link] 10.1162/​
REST_​a_​00599
Gradstein, M. and Justman, M., 2000. Human capital, social capital, and public schooling.
European Economic Review, 44(4/​6), pp. 879–​889. [Link]
S0014-​2921(99)00044-​6
Guiso, L., Sapienza, P. and Zingales, L., [Link] role of social capital in financial devel-
opment. American Economic Review, 94(3), pp. 526–​556. [Link]
0002828041464498
Hadjimichalis, C., 2006. Non-​economic factors in Economic Geography and in ‘new
regionalism’:  A sympathetic critique. International Journal of Urban and Regional
Research, 30(3), pp. 690–​704. [Link]
Hall, P., 1998. Cities in civilization. New York, NY: Pantheon Books.
Hart, D., Atkins, R. and Fegley, S., 2003. Personality and development in childhood: A
person centred approach. Monographs of the Society for Research in Child Development,
68(1), pp. 1–​122. [Link]
Hauser, C., Tappeiner, G. and Walde, J., 2007. The learning region: The impact of social
capital and weak ties on innovation. Regional Studies, 41(1), pp. 75–​88. [Link]
org/​10.1080/​00343400600928368
Hodgson, G. M., 2013. From pleasure machines to moral communities: An evolutionary eco-
nomics without homo economicus. Chicago, IL: University of Chicago Press.
Hofstede, G., 1980. Culture’s consequences: Internal differences in work related values. Beverly
Hills, CA: Sage.
Hofstede, G. and McCrae, R. R., 2004. Personality and culture revisited: Linking traits
and dimensions of culture. Cross-​Cultural Research, 38(1), pp. 52–​88. [Link]
10.1177/​1069397103259443
56

56  Robert Huggins and Piers Thompson

Holmes, T. J., 2006. Geographic spillover of unionism. National Bureau of Economic
Research working paper no. 12025. [Link]
House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W. and Gupta, V., 2004. Culture,
­leadership and organizations:  The GLOBE study of 62 societies. Thousand Oaks, CA:
Sage.
Huggins, R. and Pugh, R., 2015. Regional competitiveness and Schumpeterian devel-
opment: Policy evolution in Wales. In: J. [Link] and J. R. Wilson, eds. Strategies
for shaping territorial competitiveness. Abingdon: Routledge. Ch.8.
Huggins, R. and Thompson, P., 2014. Culture, entrepreneurship and uneven develop-
ment: A spatial analysis. Entrepreneurship and Regional Development, 26(9/​10), pp. 726–​
752. [Link]
Huggins, R. and Thompson, P., 2015. Culture and place-​based development: A socio-​
economic analysis. Regional Studies, 49(1), pp. 130–​159. [Link]
00343404.2014.889817
Huggins, R. and Thompson, P., 2016. Socio-​spatial culture and entrepreneurship: Some
theoretical and empirical observations. Economic Geography, 92(3), pp. 269–​ 300.
[Link]
Huggins, R. and Thompson, P., 2017. Human behaviour and economic development: Culture,
psychology and the competitiveness of Britain’s regions and localities. Cardiff:  School of
Geography and Planning, Cardiff University.
Huggins, R. and Thompson, P., [Link] behavioural foundations of urban and regional
development: Culture, psychology and agency. Journal of Economic Geography 19(1),
pp. 121–146. [Link]
Huggins, R. and Williams, N., 2011. Entrepreneurship and regional competitive-
ness: The role and progression of policy. Entrepreneurship and Regional Development,
23(9/​10), pp. 907–​932. [Link]
Huggins, R., Thompson, P. and Obschonka, M., 2018. Human behaviour and eco-
nomic growth:  A psychocultural perspective on local and regional development.
Environment and Planning A: Economy and Space, 50(6), pp. 1269–​1289. [Link]
org/​10.1177/​0308518X18778035
Hülsbeck, M. and Pickavé, E. N., 2014. Regional knowledge production as deter-
minant of high-​ technology entrepreneurship:  Empirical evidence for Germany.
International Entrepreneurship and Management Journal, 10(1), pp. 121–​138. [Link]
org/​10.1007/​s11365-​011-​0217-​9
Hwang, H. and Powell, W. W., 2005. Institutions and entrepreneurship. In: S. A. Alvarez,
R. Agarwal and O. Sorenson, eds. Handbook of entrepreneurship research: Interdisciplinary
perspectives, New York, NY: Springer, pp. 201–​232.
Jokela, M., Elovainio, M., Kivimäki, M. and Keltikangas-​Järvinen, L., [Link]
and migration patterns in Finland. Psychological Science, 19(9), pp. 831–​837. [Link]
[Link]/​10.1111/​j.1467-​9280.2008.02164.x
Jokela, M., Bleidorn, W., Lamb, M. E., Gosling, S. D. and Rentfrow, P. J., 2015.
Geographically varying associations between personality and life satisfaction in the
London metropolitan area. Proceedings of the National Academy of Sciences of United
States of America, 112(3), pp. 725–​730. [Link]
Judge, T. A., Higgins, C. A., Thoresen, C. J. and Barrick, M. R., 1999. The big five per-
sonality traits, general mental ability, and career success across the life span. Personnel
Psychology, 52(3), pp. 621–​652. [Link]
Kahneman, D., 2003. Maps of bounded rationality:  Psychology for behavioral eco-
nomics. The American Economic Review, 93(5), pp. 1449–​ 1475. [Link] [Link]/​
10.1257/​000282803322655392
 57

Behavioural economic geography 57

Knack, S. and Keefer, P., 1997. Does social capital have an economic impact? A cross-​
country payoff. Quarterly Journal of Economics, 112(4), pp. 1251–​1288. [Link]
org/​10.1162/​003355300555475
Kwon, S.W. and Adler, P. S., 2014. Social capital: Maturation of a field of research. Academy
of Management Review, 39(4), pp. 412–​422. [Link]
Lee, N. D., 2017. Psychology and the geography of innovation. Economic Geography,
93(2), pp. 106–​130. [Link]
Levie, J., 2007. Immigration, in-​migration, ethnicity and entrepreneurship in the United
Kingdom. Small Business Economics, 28(2/​3), pp. 143–​170. [Link]
s11187-​006-​9013-​2
Lippmann, S. and Aldrich, H., 2016. A rolling stone gathers momentum: Generational
units, collective memory, and entrepreneurship. Academy of Management Review, 41(4),
pp. 658–​675. [Link]
Martin, R. and Sunley, P., 2015. On the notion of regional economic resilience:
Conceptualization and explanation. Journal of Economic Geography, 15(1), pp. 1–​42.
[Link]
Mokyr, J., 2017. A culture of growth:  The origins of the modern economy. Princeton,
NJ: Princeton University Press.
Moretti, E., 2012. The new geography of jobs. New  York, NY:  Houghton Mifflin and
Harcourt.
Murphy, L., Huggins, R. and Thompson, P., 2016. Social capital and innovation: A com-
parative analysis of regional policies. Environment and Planning C:  Government and
Policy, 34(6), pp. 1025–​1057. [Link]
Myrdal, G., 1968. Asian drama: An inquiry into the poverty of nations. London: Allen Lane.
North, D. C., 1990. Institutions, institutional change and economic performance. Cambridge:
Cambridge University Press.
North, D. C., 2005. Understanding the process of economic change. Princeton, NJ: Princeton
University Press.
Obschonka, M., Schmitt-​Rodermund, E., Gosling, S. D. and Silbereisen, R. K., 2013.
The regional distribution and correlates of an entrepreneurship-​prone personality
profile in the United States, Germany, and the United Kingdom: A socioecological
perspective. Journal of Personality and Social Psychology, 105(1), pp. 104–​122. [Link]
[Link]/​10.1037/​a0032275
Obschonka, M., Stuetzer, M., Gosling, S. D., Rentfrow, P. J., Lamb, M. E., Potter, J. and
Audretsch, D. B., 2015. Entrepreneurial regions:  Do macro-​psychological cultural
characteristics of regions help solve the “knowledge paradox” of economics? PloS
One, 10(6), e0129332. [Link]
Obschonka, M., Stuetzer, M., Rentfrow, P. J., Shaw-​Taylor, L., Satchell, M., Silbereisen,
R. K., Potter, J. and Gosling, S. D., 2018. In the shadow of coal:  How large-​scale
industries contributed to present-​day regional differences in personality and well-​
being. Journal of Personality and Social Psychology, 115(5), pp. 903–927. [Link]
org/​10.1037/​pspp0000175
Peng, M. W., Yamakawa, Y. and Lee, S-​H., 2010. Bankruptcy laws and entrepreneur-​
friendliness. Entrepreneurship Theory and Practice, 34(3), pp. 517–​530. [Link]
10.1111/​j.1540-​6520.2009.00350.x
Perry Piscione, D., 2013. Secrets of Silicon Valley: What everyone else can learn from the innov-
ation capital of the world. New York, NY: Palgrave Macmillan.
Pettigrew, T. F. and Tropp, L. R., 2006. A meta-​analytical test of intergroup contact
theory. Journal of Personality and Social Psychology, 90(5), pp. 751–​783. [Link]
10.1037/​0022-​3514.90.5.751
58

58  Robert Huggins and Piers Thompson

Pike, A., MacKinnon, D., Cumbers A., Dawley, S. and McMaster, R., 2016. Doing evo-
lution in Economic Geography? Economic Geography, 92(2), pp. 123–​144. [Link]
org/​10.1080/​00130095.2015.1108830
Piketty, T., 2014. Capital in the twenty-​first century. Cambridge:  The Belknap Press of
Harvard University Press.
Pred, A., 1967. Behavior and location. Foundations for a geographic and dynamic location theory.
Lund: C. W. K Gleerup.
Putnam, R., 2000. Bowling alone: The collapse and revival of American community. New York,
NY: Simon and Schuster.
Putnam, R. D., Leonardi, R. and Nanetti, R. Y., 1993. Making democracy work:  Civic
traditions in modern Italy. Princeton, NJ: Princeton University Press.
Rentfrow, P. J., 2010. Statewide differences in personality: Toward a psychological geog-
raphy of the United States. American Psychologist, 65(6), pp. 548–​558. [Link]
org/​10.1037/​a0018194
Rentfrow, P. J., Gosling, S. D. and Potter, J., 2008. A theory of the emergence, per-
sistence, and expression of geographical variation in psychological characteristics.
Perspectives on Psychological Science, 3(5), pp. 339–​ 369. [Link]
[Link]/​10.1111/​
j.1745-​6924.2008.00084.x
Rentfrow, P. J., Jokela, M. and Lamb, M. E., 2015. Regional personality differences
in Great Britain. PloS One, 10(3), e0122245. [Link]
pone.0122245
Rentfrow, P. J., Jost, J. T., Gosling, S. D. and Potter, J., 2009. Statewide differences in per-
sonality predict voting patterns in 1996–​2004 U.S. presidential elections. In J. T. Jost,
A. C. Kay and H. Thorisdottir, eds. Social and psychological bases of ideology and system
justification. Oxford: Oxford University Press. Pp. 314–​350.
Rentfrow, P. J., Gosling, S. D., Jokela, M., Stillwell, D. J., Kosinski, M. and Potter, J., 2013.
Divided we stand: Three psychological regions of the United States and their polit-
ical, economic, social and health correlates. Journal of Personality and Social Psychology,
105(6), pp. 996–​1012. [Link]
Rodríguez-​Pose, A., 2001. Local production systems and economic performance in
Britain, France, Germany, and Italy. In:  C. Crouch, P. Le Gales, C. Trigilia and H.
Voelzkow, eds. Local production systems in Europe. Rise or demise? Oxford:  Oxford
University Press, pp. 25–​45.
Rodríguez-​Pose, A. and Hardy, D., 2015. Cultural diversity and entrepreneurship in
England and Wales. Environment and Planning A, 47(2), pp. 392–​411. [Link]
10.1068/​a130146p
Rodríguez-​Pose, A. and Storper, M., 2006. Better rules or stronger communities? On
the social foundations of institutional change and its economic effects. Economic
Geography, 82(1), pp. 1–​25. [Link]
Rutten, R. and Boekema, F., 2007. Regional social capital: Embeddedness, innovation
networks and regional economic development. Technological Forecasting and Social
Change, 74(9), pp. 1834–​1846. [Link]
Sachs, J., 2000. Notes on a new sociology of economic development. In: L. E. Harrison
and S. P. Huntington, eds. Culture matters: How values shape human progress. New York,
NY: Basic Books. Pp. 29–​43.
Savini, F., 2016. Self-​organization and urban development. International Journal of Urban and
Regional Research, 40(6), pp. 1152–​1169. [Link]
Schmitt, D. P., Allik, J., McCrae, R. R. and Benet-​Martinez,V. [Link] geographic dis-
tribution of big five personality traits: Patterns and profiles of human self-​description
across 56 nations. Journal of Cross-​Cultural Psychology, 38(2), pp. 173–​212.
 59

Behavioural economic geography 59

Scott,W. R., 2008. Institutions and organizations: Ideas and interests. 3rd [Link] Oaks,
CA: Sage.
Simon, H. A., 1982. Models of bounded rationality. Cambridge, MA: MIT Press.
Spigel, B., 2013. Bourdieuian approaches to the geography of entrepreneurial cultures.
Entrepreneurship and Regional Development, 25(9/​10), pp. 804–​818. [Link]
10.1080/​08985626.2013.862974
Steel, P., Schmidt, J. and Shultz, J., 2008. Refining the relationship between personality
and subjective well-​being. Psychological Bulletin, 134(1), pp. 138–​161. [Link]
10.1037/​0033-​2909.134.1.138
Stenholm, P., Acs, Z. J. and Wuebker, R., 2013. Exploring country-​level institutional
arrangements on the rate and type of entrepreneurial activity. Journal of Business
Venturing, 28(1), pp. 176–​193. [Link]
Stephan, U. and Uhlaner, L. M., 2010. Performance-​based vs socially supportive cul-
ture:  A cross-​national study of descriptive norms and entrepreneurship. Journal
of International Business Studies, 41(8), pp. 1347–​ 1364. [Link] [Link]/​10.1057/​
jibs.2010.14
Storper, M., 2010. Why does a city grow? Specialisation, human capital or institutions?
Urban Studies, 47(10), pp. 2027–​2050. [Link]
Storper, M., 2013. Keys to the City: How economics, institutions, social interaction and politics
shape development. Oxford and Princeton, NJ: Princeton University Press.
Stuetzer, M., Obschonka, M., Audretsch, D. B., Wyrwich, M., Rentfrow, P. J., Coombes,
M., Shaw-​Taylor, L. and Satchell, M., 2016. Industry structure, entrepreneurship, and
culture: An empirical analysis using historical coalfields. European Economic Review,
86(1), pp. 52–​72. [Link]
Tabellini, G., 2010. Culture and institutions: Economic development in the regions of
Europe. Journal of the European Economic Association, 8(4), pp. 677–​716. [Link]
org/​10.1111/​j.1542–​4774.2010.tb00537.x
Tubadji, A., 2013. Culture-​based development  –​culture and institutions:  Economic
development in the regions of Europe. Society Systems Science, 5(4), pp. 355–​391.
[Link]
Tubadji, A. and Nijkamp, P., 2015. Cultural gravity effects among migrants: A compara-
tive analysis of the EU15. Economic Geography, 91(3), pp. 343–​380. [Link]
10.1111/​ecge.12088
Van den Bergh, J. C. and Stagl, S., 2003. Coevolution of economic behaviour and
institutions: Towards a theory of institutional change. Journal of Evolutionary Economics,
13(3), pp. 289‒317. [Link]
Van de Vliert, E., 2009. Climate, affluence, and culture. Cambridge:  Cambridge
University Press.
Van Maanen, J. and Schein, E. H., 1979. Toward a theory of organizational socialization.
Research in Organizational Behavior, 1(1), pp. 209–​264.
Wyrwich, M., 2015. Entrepreneurship and the intergenerational transmission of
values. Small Business Economics, 45(1), pp. 191–​ 213. [Link] [Link]/​10.1007/​
s11187-​015-​9649-​x
Zak, P. and Knack, S., 2001. Trust and growth. The Economic Journal, 111(470), pp. 295–​
321. [Link]
Zucker, L. G., Darby, M. R. and Armstrong, J., 1998. Geographically localized know-
ledge:  Spillovers or markets? Economic Inquiry, 36(1), pp. 65–​86. [Link]
10.1111/​j.1465–​7295.1998.tb01696.x
60
 61

Part II

Innovations in research
design and methodology
62
 63

4 An interdisciplinary approach to the

persistent effects of Polish partitions
on educational achievements1
Justyna Kościńska and Mikołaj Herbst

Introduction
An interdisciplinary approach is considered to be the key to innovatory and
influential academic research. In some fields of study, cooperation between
scholars from different disciplines seems to be essential. This is equally true in
the case of research on the connections between history and regional develop-
ment. For instance, despite many attempts, it seems to be impossible to explain
observed differences in educational achievements between regions by using the
concepts applied within one discipline. Working over traditional boundaries is
also recommended by scholars studying the connections between the histor-
ical evolution of institutions and economic (North, 1990), social (Moore, 1966)
and political development (historical institutionalists, e.g. Pierson and Skocpol,
2002; Hall and Soskice, 2001). Effective collaboration allows scholars to share not
only their practical knowledge, but also different ideas and experiences. Social
scientists can learn from historians that a longer-​time perspective is crucial to
understanding regional divisions. Meanwhile, historians can benefit from the the-
oretical background of social sciences. However, interdisciplinary work is often
a challenge, as every discipline has its own perspective and distinctive methods.
In this chapter, we describe how we applied an interdisciplinary approach
in examining the impact of historical institutions on present-​day educational
achievements in Poland. This work is based on a project conducted by seven
researchers representing different academic fields, such as economics, sociology,
geography, mathematics and history. The main objective of the research was to
gain a better understanding of whether, and through what channels, historically
rooted differences in institutional arrangements may exert a persistent influence
on the outcomes of the education system. We expected that the institutions
shaped during the period of the territorial division of Poland (1795–​1918) will
have had a significant effect on the average present-​day school achievements.
Poland is a valuable case for research on the persistence of historical
institutions. In the second half of the 18th century, as the result of a three-​stage
partitioning, the Polish-​Lithuanian Commonwealth was carved up between its
three neighbours: the Russian Empire, the Kingdom of Prussia and Habsburg
in Austria.2 As a result, Poland did not exist as a sovereign country until the end
64

64  Justyna Kościńska and Mikołaj Herbst

of the First World War in 1918. During that time, each partitioner3 imposed
different policies concerning various areas of life, including educational prac-
tice. Access to primary and secondary education varied significantly across the
partitions (Herbst and Kaliszewska, 2017). After the Second World War, the
boundaries of Poland were redefined once again. In our research we did not
include the interwar period, the Polish People’s Republic (1945–​1989), or the
democratic transition beginning in 1989. Since the reintegration of the country
in 1918, political decisions concerning different sectors, including education,
were aimed at the reintegration of the state (Trzebiatowski, 1970). However,
recent studies show that the divisions between three former partitions are still
visible (e.g. Bartkowski, 2003; Herbst, 2012; Wysokińska, 2011). Therefore,
regional differences in Poland remain a very interesting subject for researchers.
Due to the numerous border changes, the Polish case is well-​suited for the
theory of path dependency and longue durée. It also offers methodological
lessons on how to design and implement such research.
This chapter is structured as follows: in the first section, we show examples of
previously studied connections between history and educational achievements.
In the second part, we introduce the methodology that we used. In the third
section, we explain the results of the research. In this section we also present
existing studies on the impact of Polish partitions on educational achievements.
The last part is devoted to conclusions.

Why history matters

Institutional persistence and cumulative causation have been widely
documented within various contexts. Many empirical studies show that insti-
tutional arrangements can have an impact on the functioning of societies long
after the disappearance of the original processes. While studying connections
between history and contemporary societies, we can distinguish two different
effects: on the one hand, of formally established rules and arrangements, and
on the other, of social norms and conventions. North (1990) proposes studying
institutions as “any form of constraint that human beings devise to shape
human interaction” (p. 4). Thus, present-​day choices may be influenced by his-
torical processes, including both informal norms and values shared in society
and formal regulations imposed by the authorities. An analysis of the evolution
of these institutions may be the key to understanding present-​day differences
between, as well as within, contemporary states.
History also matters for education. Scholars agree that the differences in
educational outcomes between various regions cannot be fully explained by
current trends, such as educational policy, the labour market, or present demo-
graphic conditions. Like many other aspects of social life, educational outcomes
have been studied in the context of economic development. For example,
in the United States, Parcel and Dufur (2009) highlighted the inequalities
in ­educational achievements between the northern and southern states, and
tried  to explain them by regional variations in human and financial capital.
 65

An interdisciplinary approach 65
The sociologists stated that students in richer regions tend to have higher edu-
cational achievements, with family variables controlled (Parcel and Dufur, 2009).
The authors referred mostly to demographic and socioeconomic differences,
rather than to historical arguments.
The historical perspective was included by Lynn (2010) in his analysis of the
projection of educational achievements in Italy, a country that was unified in
the mid-​19th century. Differences in living standards between the rich north
and poor south of Italy persist to the present day (Putnam et al., 1993), with
northern regions performing better than southern [Link] psychologist Lynn
(2010) calculated the regional IQs from the Program for International Student
Assessment (PISA) studies and compared them with average per capita income
in euro. The study showed that academic achievement is much higher in the
northern region of the country, and declines steadily into the South, in line
with per capita income (Lynn, 2010). According to Lynn (2010) this situation
results from the genetic differences between people living in the southern and
northern regions of the [Link] author refers to the processes of migration
from North Africa and the Near East to Sicily and the southern Italian main-
land that took place in the 8th century BC.
However, positive correlation between economic development and educa-
tional outcomes is not a universal rule. In Germany, students in the economic-
ally lagging eastern regions tend to have higher education achievements than
students living in regions with a higher per capita income (Welt-​Online, 2008).
Differences between the historical eastern and western part of the country are
tangible. Thus, the economic factor is not the only one explaining regional
differences in academic achievements. Student efforts at school can also be
affected by the social norms and attitudes towards education in a given region.
It is most likely that regions with a prevalence of norms and attitudes such as
motivation to study and the importance of education have higher educational
achievements. Sociologist Chauvel (1999) distinguished three possible models
of social structure in Europe with reference to the historical [Link] first
is the German model, with its weak hierarchies and limited stratification. It is
characterized by high economic development and a low market value of formal
education. Meanwhile, in less developed Catholic countries such as Spain or
Portugal, education is strongly connected with socio-​economic status. The
Latin model presents an intermediate structure. Consequently, the inhabitants
of countries with lower economic development have stronger motivation to
gain an education. In richer states, the schooling system is more closely linked
to the labour market.
The relationship between history and present-​day educational achievements
can also be considered in the context of past political and administrative
decisions. In Europe, the public schooling system was founded in the 19th
century. It should be underlined that the development of public schooling
depended on the economic position of the country. Using the example of the
multiethnic Habsburg Empire, economists Cvrcek and Zajicek (2013) claimed
that the politics of demand for schools was not motivated by economic return
66

66  Justyna Kościńska and Mikołaj Herbst

in terms of wage or employment prospects, but rather by the conflict between
nationalities. Public schooling was supported by the local political elites, if taught
in the “right” language of instruction. At the same time, the Austrian system
was known for its tolerance and inclusiveness, along with its well-​functioning
administration. This may have translated into positive social attitudes towards
education and, in consequence, higher academic achievements. Becker et  al.
(2016) conducted an analysis of attitudes among border area inhabitants of the
former Habsburg Empire. The economists used a border specification and a
two-​dimensional geographic regression discontinuity based on a micro dataset
of the 2006 Life in Transition Survey (LiTS). The results showed that historical
Habsburg affiliation increases current trust toward local public services of any
kind, including education.
While some studies analyze connections between educational achievements
and different socio-​economic factors (e.g. demography, per capita income), others
include the historical perspective and institutional characteristics. In researching
connections between history and present-​day educational achievements, scholars
use econometric tools such as regression discontinuity design, and conduct data
analysis based on external sources (as PISA or LiTS). In studying the influence
of history on educational outcomes there is still little research which combines
both qualitative and quantitative techniques from different fields. Meanwhile,
according to leading researchers from the domain of education, as well as his-
torical institutionalists (e.g. Pierson and Skocpol, 2002; Hall and Soskice, 2001),
studying path dependency requires crossing traditional boundaries between dis-
ciplines, and combining different approaches and research techniques. In this
chapter we present a study which corresponds with these proposals.

Methodological explorations
In this chapter we attempt to explain the influence of historical processes on
present-​day educational outcomes by using extensive data from diverse sources.
To reiterate, our project applied an interdisciplinary methodological approach,
including historical analysis, quantitative techniques, qualitative tools and elem-
ents of spatial analysis. In this section, we show how different tools applied
within various disciplines can be used in the analysis of path dependency. We
begin with a description of our historical research. In the second part, we
explain so-​called quasi-​experimental techniques and their application in eco-
nomic studies. Next, we describe how we measured the effect of historical
partitions on educational achievements in Poland by combining geographic
and economic approaches. Finally, we show how sociological tools helped us
in tracking history.

Historical insight
The existing research devoted to the persistent effect of former borders on the
present geography of educational achievements in Poland is very much focused
 67

An interdisciplinary approach 67
on measuring the gap in performance between different regions and on veri-
fying its robustness in different diagnostic tests. When it comes to explaining
channels and mechanisms, the literature is somewhat speculative, drawing
different hypotheses, but not discussing them in depth. We realized that careful
analysis of the foundations of educational systems in each of the three his-
torical partitions, going back to the late 18th century, is necessary in order to
proceed to the next step, and to provide a qualitative insight in the institutional
differences behind the regional education systems, which would be compli-
mentary to the economic and geographical approaches.
The research was based on a careful review of historical literature on the
beginnings of public education in the territory of today’s Poland. The analyzed
material included monographs, articles in academic journals and daily press,
legal acts and memorials. As the number of sources and the scope of the poten-
tial analysis exceeded the capability of our small research team, we decided
to narrow the review to the institutional arrangements already mentioned in
the existing literature. This led to us defining three major areas of interest: the
introduction of compulsory schooling and school accessibility, formation of
the teaching profession and common attitudes (norms) towards education. The
results of our review were published in a standalone research article by Herbst
and Kaliszewska (2017). However, the literature review was also used as input to
design the quantitative part of our research including the parental questionnaire,
and the scenario of the interview with educational leaders.

Quasi-​experimental methods in economic research

Many economic studies tracking the impact of history on present differences in
socio-​economic development rely on so-​called quasi-​experimental techniques.
This approach mimics the methods applied in evaluation studies, in which the
effectiveness of implemented programs, policies or interventions is [Link]
quasi-​experiment (or natural experiment) allows the researcher to compare, in
terms of a certain measure of effectiveness, the performance of units covered by
the program or intervention, with the performance of a properly selected con-
trol group. The terms “quasi” or “natural” indicate that the experiment was not
designed by any research team, but happened unintentionally as a consequence
of some policy implementation, historical event or by natural force.
To be effective as an evaluation tool, the experimental design requires the
participation (or lack of participation) in the evaluated program to be the
only factor that differentiates both groups. However, in quasi-​experiments, in
contrast to standard experiments, the division of the observed units into the
treatment group and the control group is neither random, nor designed by
the investigator. Therefore, some additional steps need to be taken in order
to ensure the comparability of observations from both groups. One way to
achieve this goal is to match observations from the two groups by analyzing
the similarity of selected characteristics. The eventual comparisons are made
only within small groups of observations sharing the same characteristics, but
68

68  Justyna Kościńska and Mikołaj Herbst

differing with respect to treatment status (where being “treated” means being
subject to a certain policy or being affected by some event, as opposed to non-​
treated observations which belong to the control group). This method is called
propensity score matching (Rosenbaum and Rubin, 1983). Another approach is
to use instrumental variables in order to distinguish the effect of treatment from
the impact of other factors that may differentiate the observations.
Public policy evaluators frequently apply the method called difference-​in-​
difference which relies on assessing the performance of both the treated group
and control group before and after the intervention. However, this approach
is less useful in research on the consequences of historical events, policies or
institutions, as it requires detailed data on the period preceding the interven-
tion, which is rarely available. Instead, historically-​oriented economic research
sometimes uses a method called regression discontinuity design (Thistlethwaite
and Campbell, 1960). This approach elicits the causal effects of interventions
by assigning a cut-​off or threshold, above or below which an intervention is
assigned. By comparing observations lying closely on either side of the threshold,
it is possible to estimate the average treatment effect in environments in which
randomization is infeasible.

Combining geographic and economic approaches: A

Geographical Regression Discontinuity Design
In order to estimate the effect of different historical experience on present
achievements in education we used the historical administrative border as a factor
determining regression discontinuity. In this kind of quasi-​experimental approach,
it is assumed that location (living, attending school) on one or the other side of this
no-​longer existing border is random. The “treatment” relies on the exposure of
individuals to certain cultural and institutional arrangements due to their location
on one side of the geographical boundary, while location on the other side of the
border implies lack of treatment. The driving factor is the distance between the
unit and the border. This approach is referred to in the literature as Geographic
Regression Discontinuity (GRD) (e.g. Keele and Titiunik, 2015; 2016) or spatial
regression discontinuity (e.g. Egger and Lassmann, 2015; Hidano et al., 2015).
Our assumption, based on the studies cited in the previous section, is that
local communities living along the two sides of the border may have different
norms and attitudes towards education because, back in the 19th century, they
were exposed to different cultural, administrative and economic arrangements.
However, we must take into account that the border of interest (between the
Austro-​Hungarian Empire and Russia) was abolished 100  years ago, when
Poland regained its independence in 1918. It is obvious that populations from
both sides of the border have been mixing since then, and thus so have their
social norms. Instead of assuming the sharp cut-​off effect of the border, we
should associate living on one side of the historical border with a certain (high)
probability of originating from the same region, and thus sharing the norms
that are typical for that region.
 69

An interdisciplinary approach 69
We decided to define treatment as a continuous, rather than a discrete vari-
able. We assigned each school with a certain probability of belonging to the
treatment group. For obvious reasons, this probability depended on the side of
the historical border where the school was located. However, it was also related
to the distance between the school and the border. We assumed that in close
proximity to the border, the percentage of crossovers (that is people originating
from one partition, but living in another) would be higher, so the treatment
variable would take a value between 0 and 1. Moving away from the border,
we expected to observe a rapid increase in the probability of treatment for the
schools on the “Austrian” side, and a decrease in the probability for those in the
former Russian partition. In other words, we assumed that the further from the
border a family lives today, the less likely it is that their ancestors migrated across
the border in the past, and the higher their probability of having “consistent”
norms towards education.
Because of the association with fuzzy experimental design4 we have called
our approach a quasi-​ fuzzy geographical Regression Discontinuity Design
(RDD). The equation estimated within the RDD is:

p
Yi = β0 + β1 Di + β 2 disti + β3 Di disti + ∑β j X ji + ε i
j=4

Where: i indexes observations, Di is a treatment variable (discrete or continuous

depending on specification), disti is a distance from historical border and Xji are
additional control variables.

Sociological research
After conducting historical and econometric research, we decided to deepen
our analysis by including sociological methods. This part of the research aimed
at discovering social attitudes and the perception of educational systems, and
consisted of field research and a questionnaire survey. In June 2017, we carried
out individual in-​depth interviews in three cities in regions of the former
Russian, Austrian and Prussian partitions. The chosen cities had populations of
between 25,000 and 50,000 people, and were located at least 60km from the
nearest voivodeship capital city. We chose medium-​sized towns, because we
expected that agglomerations may no longer reflect the former partitions, as
they have experienced shifts of migrants from different parts of the [Link]
tried to ensure comparability by choosing cities characterized by a similar level
of: income per habitant, unemployment rate, percentage of the population aged
less than 18 years, and availability of educational services.
Qualitative methods are aimed at gaining a broader understanding of the
subject studied, and the reasons and motivations that underlie people’s decisions
and behaviours. By using the method of individual in-​depth interviews, we
wanted to discover people’s attitudes towards history and the partitions, and
70

70  Justyna Kościńska and Mikołaj Herbst

identify local traditions in the field of education. For this reason, we decided
to select respondents according to their occupation and profession. In each
city studied we conducted about 12 interviews with local experts, each lasting
from one to two hours. We conducted interviews with local journalists, school
principals, history teachers, political leaders and members of non-​governmental
organizations from the field of history and education. During the field research
we also carried out observations by visiting local schools and regional museums.
Since qualitative research helps to gain a better understanding of deeper
motives and explain decisions made by people, quantitative research can show
general trends in populations. In October and November 2017, a question-
naire survey was conducted using computer-​assisted personal interviews. The
study was conducted on a sample of 2,500 parents of students from the last
two grades of secondary schools. The respondents were selected by quota sam-
pling in two steps. First, schools in pre-​war Germany that became part of the
People’s Republic of Poland after the Second World War and schools with fewer
than ten students were excluded from the survey. Then, the sample was strati-
fied according to the territory (formerly belonging to the Russian, Austrian
or Prussian partition). Additionally, each layer was divided into ­ categories
according to town size. A total of 250 schools were selected for the survey with
a probability proportional to the number of students attending the given school.
In each school ten respondents filled out the survey. Thus, the chosen sample
can be said to be representative of the population.

Results of the research

In this section we summarize the results of our interdisciplinary project. Our
intention was to design the research in such a way that the different disciplinary
approaches would be complimentary, and provide outcomes which could be
presented as stand-​alone results, but could also be used as input to further stages
of the project. With such a design, the order of research steps is very important.
We started by reviewing the existing empirical literature on the impact of
partitions on educational achievements. We proceeded with a historical inves-
tigation focused on the education-​related institutions in partitioned Poland in
the early 19th century. We then conducted a careful cartographic analysis of
recent educational achievements in Poland, followed by econometric model-
ling. Finally, we performed a survey among parents and in-​depth interviews
with local leaders in three towns, belonging historically to different partitions.
We will preserve this order in presenting the major results of our research in
the following sections.

Historical partitions and educational achievements in

Poland –​literature review
Since the first national examination in Polish primary schools in 2002, many
studies have been conducted. They show that the territorial distribution of
 71

An interdisciplinary approach 71
educational achievements clearly reflects historical partitions, with eastern and
south-​eastern regions performing better in school tests than the western part
of the country (Herczyński and Herbst, 2002; Herbst, 2004). Eastern Poland
outperforming the West is not a common finding in regional studies investi-
gating the persistent effects of partitions on present-​day Polish society. In other
respects, the regions formerly under Prussian rule mostly outperform the former
Austrian and Russian partitions. For example, authors show that the western
parts of Poland governed by Prussia are better developed in terms of economy
than the rest of the country (Herbst, 2012).This includes indicators such as: per
capita income, urbanization level and infrastructure density (CPM, 2009).
Moreover, former Prussian-​controlled regions are characterized by a higher
general trust indicator than former Russian territory (Wysokińska, 2011).
Scholars have tried to explain the fact that educational indicators do not
follow patterns common in Polish society by proposing various hypotheses.
Geographers Czapiewski and Śleszyński (2007) linked the differences in educa-
tional outcomes with the agricultural history of the territories formerly under
Russian and Austrian rule. By referring to data derived from external nationwide
surveys, they claimed that the inhabitants of eastern parts of the country are more
devoted to their land and, accordingly, they feel more responsible for their fam-
ilies and the education of their children. The differences may also be connected
with the fact that the inhabitants of territories which were formerly under
Russian rule are conservative in their choices (Bartkowski, 2003). Furthermore,
the authors highlight a second dimension of differentiation in educational
outcomes, which is a division between urban and rural areas (Czapiewski and
Śleszyński, 2007). Another possible explanation is the dependence of educa-
tional outcomes on the level of social capital in the region. Economists Herbst
and Rivkin (2013) used data from the Polish Central Examination Board,
Polish Central Statistical Office, and national census conducted in 2002 to elab-
orate on the possible role of social capital and uneven returns to education. In
low ­economic performance areas higher academic achievements may be seen,
according to the authors (Herbst and Rivkin, 2013), as generally more profit-
able. One possible reason is that economically lagging regions have low employ-
ment in the industrial sector, in which career opportunity is less dependent on
academic credentials. They also tend to have a large share of jobs in the public
sector, where education level matters. Finally, high academic achievements
facilitate migration to more endowed regions (for further work-​related studies),
which increases the motivation to work hard at school.
More recently, Bukowski (2018) applied the regression discontinuity
approach in a municipality-​level analysis, and found a statistically significant
difference between test results in post-​Austrian and post-​Russian territories.
As mentioned above, the former Austrian regions perform better in school
tests than the former Russian territories. The economist tried to explain the
differences in educational achievements by referring to the general policy of
former partitioners (Bukowski, 2018). For instance, in the Prussia-​controlled
territory, the public education system was oppressive. The Polish language
72

72  Justyna Kościńska and Mikołaj Herbst

was forbidden in schools and education was seen as a tool for spreading the
state language and culture (Bukowski, 2018). Processes of Germanization and
Russification may have resulted, according to Bukowski (2018), in nega-
tive social attitudes towards education. The author claims that higher educa-
tional achievements in territories which were formerly under Austrian rule
result from its liberal and tolerant politics towards national minorities in the
19th century. The Austrian partitioner allowed Polish people to teach and
learn in their mother tongue, which may have created a positive attitude to
learning. They also had the opportunity to associate. This is also expressed in
a higher level of satisfaction with received education in the former Austrian
Partition and higher expectations regarding the schooling effort (Bukowski,
2018, p. 21).
Many authors have tried to explain the low educational achievements in the
territories acquired from Germany after the Second World War (e.g. Bański,
Kowalski and Śleszyński, 2004; Herczyński and Herbst, 2002; Herbst, 2004;
Herbst and Rivkin, 2013). In 1945, Poland lost its eastern territories to the
Soviet Union, and acquired parts of pre-​war East and North East Germany.
Borders shifts caused mass relocations of Polish, Russian and German citi-
zens. The post-​war Polish authorities used these newly-​acquired territories to
develop state-​owned farms (PGR or State Agricultural Enterprises) and agri-
cultural cooperatives. Most of them were closed after the beginning of eco-
nomic transition in 1989. Herbst and Rivkin (2013) claim that mass migrations
throughout the 20th century resulted in the absence of long-​lasting social
institutions, which may be the source of the educational failure. To this day,
western and northern territories of Poland are struggling with high structural
unemployment, slow economic growth and lack of infrastructure (Herczyński
and Herbst, 2002; Bański, Kowalski and Śleszyński, 2004). Overall, there has
been a lot of research on path dependence and differences in educational
achievements between different partitions in analysis of regional development
in Poland. However, most existing empirical studies which attempt to explain
aforementioned differences use aggregated data for municipal level research. In
contrast to these earlier studies, we base our work on school-​level data obtained
from questionnaire [Link] also include qualitative data: individual in-​depth
interviews and observations.

Results of cartographic and econometric analysis5

The conclusions from the literature review are consistent with the pattern in
Figure  4.1 showing the municipal averages of 9th-​grade student test scores
in mathematics (2016). The achievements in south-​ eastern Poland, which
belonged to the Austro-​Hungarian Empire until 1918, are high compared to
other regions of Poland. The intervals shown on the map refer to the per-
centage maximum achievable score. For example, municipalities shaded in
black (highest interval) had an average test score between 50.8 per cent and
74.6 per cent of the maximum score that could be achieved in this test. As one
 73

An interdisciplinary approach 73

Territories
Prussian partition
acquired from
Germany after
WWII

Russian
partition

Territories Partition
acquired from municipalities
Germany after 0.274–0.411
0.411–0.444
WWII 0.444–0.475
0.475–0.508
0.508–0.746
Austrian partition

Figure 4.1 Municipal average scores on 9th-​grade test in mathematics (2015) as a per-

cent of maximum achievable score

can see, most of the municipalities in the former Austrian partition are ranked
within the top two intervals.
More precise calculations were made by Herbst and Wójcik (2018), who used
school level data weighted by the number of students. First, the actual scores
were standardized so that the country average was 0 and standard deviation
equalled 1. The average school score in the Austrian partition is higher than
the country average by 0.18 standard deviations. The gap between the Austrian
partition and the Prussian partition is even larger (0.25 standard deviations).
The ex-​Russian territory performs better than average, but still 0.13 standard
deviations below the achievements in the ex-​Austrian region. The worst score
is observed in the territories acquired by Poland after the Second World War
(from Germany). Students in this area perform 0.16 standard deviations below
the country average, and 0.35 standard deviations below the mean score in the
Austrian partition.
In the next step we applied a geographic regression discontinuity design
(GRDD) to measure the effect of the partition in close proximity to the former
border. We found that location in the former Austrian partition had a signifi-
cant and positive effect on student performance in mathematics, as measured
by the test score achieved by students graduating from middle schools. Even if
74

74  Justyna Kościńska and Mikołaj Herbst

2.5
Exam results

0.0

–2.5

–50 –25 0 25 50
Distance from historical border
Former partition
Austrian Russian

Figure 4.2 RDD plot with discontinuity at former border between the Austrian and
Russian partitions

we controlled for school and municipal characteristics (including measures of

poverty and educational attainment in the population) the magnitude of the
effect was 0.4 standard deviations. Moreover, the effect remained significant in
the value-​added specification, when student achievements at an earlier stage
of education were controlled, with the magnitude reduced to 0.15 standard
deviations.

How does history matter? Sociological research based on historical insight

Our insight into the genesis of the education system in partitioned Poland,
reaching back to the early 19th century, led to the conclusion that one fun-
damental difference between the early education system in Poland’s terri-
tory under Austrian, Russian and Prussian rule may relate to the degree of
social acceptance of school, teacher status and the role of education as a lever
of social advancement. The former Austrian partition was superior to other
regions with respect to all these factors (for details see Herbst and Kaliszewska,
2017). Another important difference between the partitions was how the state
perceived the mission of public education: in terms of acquisition of practical
skills, or rather in terms of personal development. In this respect, there was a
 75

An interdisciplinary approach 75
clear dissimilarity between the former Austrian partition (which corresponds
to the historical region of Galicia), where the curriculum was focused on the
humanities, and other partitions (particularly Prussia), where schools were more
closely linked to the economy.
In order to assess to what extent these historical differences may still affect
the social norms towards education in Poland we conducted a research survey
among parents. Using the methods of individual in-​depth interviews, obser-
vation and a questionnaire survey (CAPI), we were searching for mechanisms
which could have enabled the reproduction of historical institutions. In this
section, we discuss two factors: the level of social capital in the region, and the
norms and values prevailing in given territories.

Social capital and trust in the education system

Many authors have shown that there exists a significant relationship between
social capital and educational achievements (Rogosic and Baranovic, 2016).
“Social capital” refers to “connections among individuals –​social networks and
the norms of reciprocity and trustworthiness that arise from them” (Putnam,
2000, p. 19). Differences in educational outcomes can be explained by different
levels of existing social capital, as it can support schooling in the form of appro-
priate values and motivation. While studying this phenomenon, it is important
to analyze the level of trust in public institutions, including educational ones. It
can be assumed that in regions with higher levels of social capital, respondents
will declare a higher level of trust in the education system.
When asked about various public institutions (e.g. government, police, muni-
cipality) in the questionnaire survey, inhabitants of Galicia declared that they
undoubtedly trust in them. In the questionnaire, 18.52 per cent of respondents
from the former Galicia declared that they definitely trust the education system.
In the former Prussian and Russian partitions, in turn, this percentage is 9.55
per cent and 9.30 per cent, respectively. Inhabitants of the former Austrian par-
tition are also more satisfied with their jobs and family relations; they see good
prospects for their [Link] underline that they are satisfied with their child’s
school (25.93 per cent are very satisfied, in the Prussian partition 13.58 per
cent; in the Russian 11.94 per cent). Of the respondents in the former Galicia,
43.15 per cent believe that their child’s school is better than the country’s
average (in the Prussian partition 38.51 per cent, and in the Russian 26.12 per
cent). To illustrate, during the individual in-​depth interview, the principal of a
primary school and local politician in the former Austrian partition said “when
it comes to the value of education, it is certainly perceived as very, very high,
and it must be admitted that it is important to people here in our region” (A1F,
interviewed on 2017/​06/​06)6. Trust in education in Galicia can be summed up
with the words of a local activist, who said “education is the issue number one”
(A2M, interviewed on 2017/​06/​05). In contrast, the leader of a local group of
scouts in the former Russian partition suspected that “if parents let children do
what they want, they wouldn’t even go to high school” (R3M, interviewed on
76

76  Justyna Kościńska and Mikołaj Herbst

2017/​06/​08).These results confirm the thesis proposed by Becker and his team
(2016) and by Bukowski (2018). In the 19th century, Austrian rule was lib-
eral and tolerant. In contrast, in the Russian and Prussian partitions inhabitants
experienced many forms of [Link] can assume that the more oppressive
Russian and Prussian systems spawned reluctance towards formal education.
By contrast, the liberal Austrian rule may have translated into positive attitudes
towards education, and, as a result, higher educational achievements.
Social trust, as well as satisfaction from life and education may also result
from the level of attachment to the region. Of the respondents from the former
Austrian partition, 71.48 per cent were born in the town where they live right
now. In contrast, in Prussian and Russian partitions this percentage is 60.75
per cent and 62.48 per cent, respectively. Accordingly, when asked about their
relationship with the region, 26.30 per cent of inhabitants from the former
region of Galicia declared a very strong attachment (in Prussian and Russian
partition this was 19.25 per cent and 20.93 per cent, respectively). During our
observations, we also discovered that in former Galicia local history is more
often cultivated. Historical memory in the former Austrian partition is vivid.
For example, when asked about the partitions, a history teacher answered: “from
my perspective, and the perspective of this school, it was a very favourable
time” (A4F, interviewed on 2017/​06/​06). In contrast, in a town studied in the
Russian partition, neither schools nor streets have local patrons, and no writers
or educational activists are commonly known. In former Galicia, inhabitants
also express greater trust in the Church (both in the questionnaire survey and in
individual in-​depth interviews). It can be assumed that religion can strengthen
social bonds, as it encourages people to meet each week and participate in
parish life. In this dimension, religion can produce social capital, which can
be later transferred to the other spheres of social life. To conclude, it can be
stated that the high educational achievements in the former Galician region
may result from the higher social capital and trust in educational institutions in
these territories.

Formative versus applicative role of education

Prevailing social norms and educational motivation are other important
channels for the persistence of institutions. Education can be recognized as
having a formative or applicative role. Parents can believe that the role of the
school is to shape a child’s personality, or to teach directly applicable, practical
skills. In the questionnaire, respondents were asked to prioritize the goals of
school education by choosing three priorities from various options. Inhabitants
of the former Austrian partition pointed more often to formative values such
as patriotism or respect for the authorities. In contrast, residents of the former
Prussian partition tended to select “utilitarian” goals such as self-​reliance and
knowledge about reality. Respondents were also asked to choose three factors
which can help to “achieve success in life”. In general, in first place most of
the respondents chose “education”. However, in further responses, differences
 77

An interdisciplinary approach 77
between regions could be observed. In the former region of Galicia, emphasis
was placed on resourcefulness, in the Prussian partition, on hard work, and in
the Russian partition, on contacts.
Consequently, the interviewees in former Galicia agreed that school should
provide general education and knowledge which will help students adapt to
all possible conditions. To illustrate, a historian from the local high school in
Galicia in the personal interview said that one of the main benefits of education
is “understanding the world, first and foremost. The more you are closed in a
specific niche and do not want to gain an education, expand your knowledge,
the more difficult it will be for you to find your place in this contemporary
world” (A4F, interviewed on 2017/​06/​06). Parents in Galicia share an opinion
that schools should be teaching values such as justice, honour and patriotism,
and should not be correlated with economic conditions. In contrast, inhabitants
living in the former Prussian partition tend to declare that it is important to
learn practical skills and have a concrete profession. For example, the leader of
a local group of scouts in the former Prussian territories claimed that school
should teach things that are needed in everyday life, “such as dealing with office
work or writing letters” (P5M, interviewed on 2017/​06/​07). According to the
principal of a primary school, students want “to have a concrete profession,
which will bring them money, what does not necessarily have a lot in common
with the knowledge” (P6F, interviewed on 2017/​06/​07).
In the questionnaire survey parents were also asked to choose three subjects
they care about the most while monitoring their children’s school achievements.
In the Austrian partition parents tended to choose history, or artistic subjects
(such as music or painting), while in the Prussian partition, they chose entrepre-
neurship and STEM subjects (i.e. science, technology, engineering and math-
ematics). Respondents in the former Russian partition also shared the opinion
that it is more important to learn STEM subjects than the humanities. To illus-
trate, the principal of a local primary school said:

R‌ecently you can clearly see that there are some directions of education
that are more likely to be valuable and useful later in life […] if you finish
a good technical university and good department there, well, life is com-
pletely different than after the humanities, which, not always […] give you
the possibility to choose a job.
(R7M, interviewed on 2017/​06/​08)

Studying STEM subjects is seen as a key factor in finding a good job in the
future. This may result from the more general norms shared in society about
the market value of different disciplines as well as from present-​day conditions
in the local labour market.
In the former region of Galicia, belief that school should play a formative
role may result from the historical events that took place in this territory in
the 19th century. The Austrian partition was not well-​developed in an eco-
nomic sense. It was very difficult to pursue a career in technical professions. At
78

78  Justyna Kościńska and Mikołaj Herbst

the same time, the liberal Habsburg rule allowed minorities to associate and
to have administrative and political representation. The education system was
inclusive and open. As a result, careers were often pursued in official positions
(Herbst and Kaliszewska, 2017). In contrast, in the Prussian partition, industries
and agricultural technologies were well-​developed, as these territories were
meant to be an agricultural base for the Kingdom of Prussia. This generated a
demand for workers with technical skills: in agriculture and in industry as well.
Accordingly, available schooling in the Prussian partition offered vocational
education in different [Link] Russian partition was the least advanced eco-
nomically, although several industries existed in these territories. These institu-
tional frameworks may have persisted in society as routines and customs.
The greater emphasis placed on the formative role of education may be an
important factor in explaining the high educational achievements in former
Galicia. More trust in personal development through education implies that
the level of student effort is less dependent on the perceived material benefits
from education, and on whether the achievement is measured using high stake
examinations or just diagnostic tests. The observed regional differences in par-
ental attitudes towards education may also suggest that the achievement gap
between the Austrian partition and the other regions of Poland should be
larger with respect to humanities (history, literature) than in mathematics or
science. However, this last hypothesis was not confirmed in the quantitative
analysis.

Conclusions
The aim of this chapter was to describe how an interdisciplinary approach can be
applied in conducting research on the connections between history and regional
development. We began our work with cartographic and econometric analysis.
This showed that there exists a clear difference in educational achievements
between regions which were governed by different partitioners in the 19th
and early 20th centuries. However, by using econometric tools we could only
measure the effect; we could not explain the mechanisms that underlie it. For
this reason, we conducted a historical analysis which provided a qualitative
insight into the regional differences of the education system. It helped us to
designate main areas of interest, which later became the basis for sociological
research. We used this to prepare scenarios for individual in-​depth interviews
and a questionnaire survey. During the sociological research we discovered that
regional differences in educational outcomes can be explained with reference to
two dimensions: the level of social capital, and the norms and values prevailing
in the [Link] complementarity of the perspectives offered by different fields
was one of the greatest advantages of this interdisciplinary research.
Additionally, our research shows that the intergenerational transfer of long-​
inculcated norms may have a powerful impact on territorial patterns of socio-​
economic development, even if the administrative and political divisions that
were at the foundation of these norms vanished a long time ago. This supports
 79

An interdisciplinary approach 79
the earlier findings of Bisin and Verdier (2001), and Patacchini and Zenou
(2011) on how parental effort combined with local social norms contributes
to the petrification of regional differences in educational performance. With
respect to educational policy, our results show that traditional measures of
coping with inequalities that rely on redistribution of resources, providing extra
teaching hours, or intensified evaluation of achievements may be ineffective if
deeply rooted local norms towards education are not taken into account, and
if policies fail to address different generations, and not just students at schools.
Due to the interdisciplinary approach, we were able to analyze the mechanisms
and channels of persistence. This kind of research could be applicable in other
culturally diverse countries.

Notes
1 This work was supported by the National Science Centre (NCN), under the project
“Persistent effects of historical partitions on educational achievements. An interdis-
ciplinary approach” (2014/​14/​E/​HS4/​00089).
2 For a more detailed historical description of the partitions of Poland see Davies
(1981).
3 The term “partitions” is commonly used in the international literature to describe
the conquering of Poland’s territory by the three superpowers between 1772 and
1795 (e.g. Davies, 1981; Encyclopaedia Britannica, 2018). We will therefore use this
expression, as well the term “partitioners”, to refer to the conquering countries.
4 Fuzzy Regression Discontinuity Design (Hahn et al., 2001) helps to assess the effect
of treatment in the presence of crossovers. However, this method requires that we
actually know the treatment status of each observation in the data. In our case how-
ever, we only know where students attend school now, but we can’t be sure whether
their families originate from the same region in which they live now, or from another
partition. For this reason, on our analysis we replaced a dichotomous treatment vari-
able (taking value 0 or 1) with the probability of treatment.
5 A more comprehensive discussion of results can be found in Herbst and Wójcik
(2018).
6 All citations from interviewees are in the authors’ possession. Chosen quotes are
translated by the authors of the article. Quotes are described in the following way: A,
R, or P (former Austrian, Russian, or Prussian partition), number of respondent, F or
M (female or male), date of interview.

References
Bański, J., Kowalski, M. and Śleszyński, P., 2004. Zarys problemów związanych z
uwarunkowaniami zróżnicowań przestrzennych wyników sprawdzianu dla uczniów szkół
podstawowych w 2002 r. Warsaw: Polish Central Examination Board.
Bartkowski, J., 2003. Tradycja i polityka. Wpływ tradycji kulturowych polskich regionów na
współczesne zachowania społeczne i polityczne. Warsaw: Żak.
Becker, S., Boeckh, K., Hainz, C. and Woessmann, L., [Link] Empire is dead, long live
the empire! Long-​run persistence of trust and corruption in the bureaucracy. The
Economic Journal, 126(590), pp. 40–​74. [Link]
80

80  Justyna Kościńska and Mikołaj Herbst

Bisin, A. and Verdier, T., 2001. The economics of cultural transmission and the dynamics
of preferences. Journal of Economic Theory, 97(2), pp. 298–​ 319. [Link]
10.1006/​jeth.2000.2678
Bukowski, P., (2018) How history matters for student performance: Lessons from the
partitions of Poland. Journal of Comparative Economics. (Accepted for publication
October 2018.) [Link]
Chancellery of the Prime Minister (CPM), 2009. Polska 2030. Wyzwania rozwojowe.
Warsaw: Chancellery of the Prime Minister.
Chauvel, L., 1999. Existe-​ il un modèle européen de structure sociale? Revue de
t-​
l’OFCE, 71(1), pp. 281–​298. [Link]
Cvrcek, T. and Zajicek, M., 2013. School, what is it good for? Useful human capital and
the history of public education in central Europe. Cambridge, MA: National Bureau of
Economic Research. [Link]
Czapiewski, K. and Śleszyński, P., 2007. Geografia zróżnicowania wyników egzaminów
zewnętrznych. In:  Polish Central Examination Board, 2007. Egzamin. Biuletyn
Badawczy Centralnej Komisji Egzaminacyjnej. Warsaw:  Polish Central Examination
Board. Pp. 52–​79.
Davies, N., 1981. God’s playground. A  history of Poland:  In two volumes. New  York,
NY: Columbia University Press.
Encyclopedia Britannica, 2018. Partitions of Poland. Available online at: [Link].
com/​event/​Partitions-​of-​Poland [Accessed 19 September 2018].
Egger, P. H. and Lassmann, A., 2015. The causal impact of common native language
on international trade: Evidence from a spatial regression discontinuity design. The
Economic Journal, 125(584), pp. 699–​745. [Link]
Hahn, J., Todd, P. and Klaauw, W. (2001). Identification and estimation of treatment
effects with a regression-​ discontinuity design. Econometrica, 69(1), pp. 201–​ 209.
[Link]
Hall, P. A. and Soskice, D., eds. 2001. Varieties of capitalism: The institutional foundations of
comparative advantage. Oxford: Oxford University Press.
Herbst, M., 2004. Human capital formation in Poland: Where does educational quality
come from? Studia Regionalne i Lokalne, 17(3), pp. 89–​104.
Herbst, M., 2012. Edukacja jako czynnik i wynik rozwoju regionalnego: Doświadczenia Polski
w perspektywie międzynarodowej. Warsaw: Scholar.
Herbst, M. and Kaliszewska, A., 2017. The partitions and education: The beginnings of
education establishments on the Polish territory in the context of today’s differenti-
ation of academic achievement. Studia Regionalne i Lokalne, 2(68), pp. 5–​29.
Herbst, M. and Rivkin, S., 2013. Divergent historical experiences and inequality in
academic achievement: The case of Poland. Journal of Socio-​Economics (42), pp. 1–​12.
[Link]
Herbst, M. and Wójcik, P., 2018. The persistent legacy of the fallen empires:  Assessing the
effects of Poland’s historical partitions on student achievement. Warsaw:  University of
Warsaw EUROREG Working Paper.
Herczyński, J. and Herbst, M., 2002. Pierwsza odsłona: Społeczne i terytorialne zróżnicowanie
wyników sprawdzianu szóstoklasistów i egzaminu gimnazjalnego przeprowadzonych wiosną
2002 roku. Warsaw: Friedrich Naumann Stiftung.
Hidano, N., Hoshino,T. and Sugiura, A., [Link] effect of seismic hazard risk information
on property prices: Evidence from a spatial regression discontinuity design. Regional
Science and Urban Economics, 53, pp. 113–​122. [Link]
2015.05.005
 81

An interdisciplinary approach 81
Keele, L. J. and Titiunik, R., 2015. Geographic boundaries as regression discontinuities.
Political Analysis, 23(1), pp. 127–​155. [Link]
Keele, L. J. and Titiunik, R., 2016. Natural experiments based on geography: ERRATUM.
Political Science Research and Methods, 4(2), p.  449. [Link] 10.1017/​
psrm.2015.37
Lynn, R., 2010. In Italy, North–​South differences in IQ predict differences in income,
education, infant mortality, stature, and literacy. Intelligence, 38(1), pp. 93–​100. [Link]
[Link]/​10.1016/​[Link].2009.07.004
Moore Jr., B., 1966. Social origins of dictatorship and democracy:  Lord and peasant in the
making of the modern world. Boston: Beacon Press.
North, D. C., 1990. Institutions, institutional change and economic performance. Cambridge,
MA: Cambridge University Press.
Parcel, T. and Dufur, M., 2009. Family and school capital explaining regional variation
in math and reading achievement. Research in Social Stratification and Mobility, 27(3),
pp. 157–​176. [Link]
Patacchini, E. and Zenou,Y., 2011. Neighbourhood effects and parental involvement in
the intergenerational transmission of education. Journal of Regional Science, 51(5), pp.
987–​1013. [Link]
Pierson P. and Skocpol, T., 2002. Historical institutionalism in contemporary political
science. In: I. Katznelson and H. V. Milner, eds. 2002. Political science: State of the dis-
cipline. New York, NY: W. W. Norton. Pp. 693–​721.
Putnam, R. D., 2000. Bowling alone:  The collapse and revival of American community.
New York, NY: Simon & Schuster.
Putnam, R. D., Leonardi, R. and Nanetti, R. Y., 1993. Making democracy work:  Civic
traditions in modern Italy. Princeton, NJ: Princeton University Press.
Rogosic, S. and Baranovic, B., 2016. Social capital and educational achievements: Coleman
vs. Bourdieu. CEPS Journal, 6(2), pp. 81–​100.
Rosenbaum, P. R. and Rubin, D. B., 1983. The central role of the propensity score in
observational studies for causal effects. Biometrica, 1(70), pp. 41–​55.
Thistlethwaite, D. L. and Campbell, D. T., 1960. Regression-​discontinuity analysis: An
alternative to the ex post facto experiment. Journal of Educational Psychology, 51(6), pp.
309–​317. [Link]
Trzebiatowski, K., 1970. Szkolnictwo powszechne w Polsce w latach 1918–​1932. Wroclaw:
Ossolineum.
Welt-​Online, 2008. Pisa study results: East German education outperforms West. Die
Welt, 19 Nov.
Wysokińska, A., 2011. Invisible wall:  Role of culture in long term development. Florence:
European University Institute.
82

5 
Regional GDP before GDP
A methodological survey of historical
regional accounts
Kerstin Enflo and Anna Missiaia*

Introduction
In a broad sense, economic inequality and ways of tackling it have for a long time
been a focus in both academic research and policy-​making. But different types
of inequality are regarded as more or less relevant in different historical periods.
In the aftermath of the financial crisis of 2008, the debate largely focused on
the surge of interpersonal inequality (Piketty, 2014; Milanović, 2016). Growing
interpersonal inequality has been proposed as a driver of political movements
since the crisis. However, not all scholars agree that focusing on this particular
type of inequality should be the priority. Rodríguez-​Pose (2018) noted how the
recent political turbulence originated not from people “that don’t matter” but
from places “that don’t matter”.The latter are the most depressed regions in both
the US and Europe. As regards interpersonal inequality, the study of regional
inequality prefers a long-​term perspective to better understand its determinants.
Rosés and Wolf (2018b) recently coordinated a collective effort by several eco-
nomic historians to evaluate the regional GDP from 1900 to the present of sev-
eral NUTS-​2 European regions. The results are surprisingly similar to Piketty
(2014) in terms of long-​run aggregate tendency: regional inequality declined in
Europe from an overall Gini of 0.55 in 1900 until the 1980s, when it reached
0.47; it then increased up to today.A similar U-​shaped curve is observed for inter-
personal inequality. Why then is regional inequality important? Rodríguez-​Pose
(2018) suggests that regional inequality cannot be neglected because its effects
on voting behaviour are far stronger than those of interpersonal inequality. If
political choices are driven by living in a poor region rather than living in a poor
household, government policies should address regional inequality.
Historical research can provide a long-​term view of the drivers of both
types of inequality and their long-​run evolution. Piketty (2014), in his study
of historical interpersonal inequality suggested that the reduced inequality in
the period following the Second World War was the exception rather than the
rule, and warned that without the appropriate policies, inequality would by and
large continue to increase. Similarly, Rosés and Wolf (2018a) suggest that the
post-​Second World War period, when regions experienced convergence, was an
exception that should not be taken as a policy benchmark for our own times.
 83

Regional GDP before GDP 83

Reliable empirical evidence on both the outcome variable and the explana-
tory variables is essential for evaluating the extent of regional inequality and its
causes and for eventually designing appropriate policies. For current regional
inequality, there are extensive data from a range of geographical units that can
be used for this [Link] main measure used is the Gross Domestic Product
(GDP), which is also the main measure for comparing countries. However,
the level of precision and coverage that can be reached at the national level is
not achievable at the regional one. For this reason, regional GDP is typically
reconstructed using evidence at the industry rather than the household level
and often neglects intra-​regional imports and exports which would be too
costly to record (Eurostat, 2013).
The evaluation of long-​term regional inequality is even more complicated.
Thus, the purpose of this chapter is to provide a methodological overview
of existing approaches to measuring historical regional GDP. Even at national
level, GDP is a late bloomer. Kuznets (1934) was a frontrunner of national
accounts: he introduced the concept of national income in the 1930s, but with
some differences from today’s practice in what the measure included. The goal
was primarily to measure economic growth. Standard GDP as we know it
today was developed well after the end of the Second World War and is mostly
based on a mix of direct information about production such as tax records and
estimates for otherwise unrecorded production.
For estimates of national GDP before the mid-​20th century, economic
historians rely on both direct and indirect measures of production or income
from sectors of the economy that are later aggregated into national series. One
of the first to propose estimates of national accounts for various countries from
a historical perspective was Gerschenkron (1962). Later, a massive effort to
harmonize all the existing estimates worldwide over the very long run was
made by the late Angus Maddison (Maddison, 2007). The Maddison dataset
was constructed on the basis of two elements: the 1990 related income levels
for all countries and the growth rates for each country, taken from the existing
literature and with sometimes ad hoc adjustments decided by Maddison him-
self. Using these two, comparable estimates of GDP at a constant of 1990’s
international dollar rate were compiled. The efforts of Maddison are today
continued by the members of the Maddison Project (Bolt et al., 2018).
For historical regional accounts, the challenges are even greater. Direct
estimates of income or production similar to national ones are rarely achiev-
able because of the lack of quantitative evidence at the regional level. In spite
of the challenges, in the last 20 years scholars using various methods have been
working towards the reconstruction of regional GDP series. Early attempts
were based on the use of proxy variables. For instance, Good (1994) applied
a method similar to that used at national level for the UK by Crafts (1983) to
estimate the GDP per capita of the different regions of the Habsburg Empire.
The method is based on various regional proxies that are plugged into a regres-
sion. The coefficients were then used to estimate the GDP for each of the
regions. A similar attempt has been carried out by Esposto (1997) to estimate
84

84  Kerstin Enflo and Anna Missiaia

the regional GDP of the Italian regions after 1861. Another method, which is
currently the most widely adopted by economic historians, is that proposed
by Geary and Stark (2002). The Geary and Stark (G-​S) method allocates the
national series of the various sectors according to employment at regional and
at sectoral level. Employment figures in benchmark years are often provided
by population censuses. To correct for productivity differentials, regional wages
are used. Quite frequently, scholars use a mix of direct estimates for sectors
which are under a state monopoly (for instance, mining or tobacco), or find
better direct evidence elsewhere. For the remaining sectors they employ the
G-​S method. Examples of these mixed estimates are Schulze (2007) on Austria-​
Hungary, Felice (2011) on Italy and Enflo and Missiaia (2018) on Sweden.
The period that can be covered varies from country to country and depends
on the availability, quality and consistency of sources over time. Historical
regional GDP can typically be estimated for the entire 20th century world-
wide. Most European countries can rely on population censuses (or other
sources) from the mid-​19th century onwards. Before the mid-​19th century, the
regional GDP time series can be regarded as an [Link] recent examples
are the series from 1819 for Belgium (Buyst, 2018) and those from 1750 for
Sweden (Enflo and Missiaia, 2018), along with the early benchmark of 1571.
The Swedish case will be illustrated in detail below as a case study, where a
mix of direct estimates for the mining sector and of G-​S method estimates for
the other sectors is used for the period 1750–​1850. The 1571 benchmark was
estimated using regional principles other than employment, providing a novel
approach that may be adopted in other historical [Link] new Swedish series,
jointly with those by Enflo, Henning and Schön (2014), represent a unique
opportunity to consistently trace regional inequality in a European country
from the pre-​industrial period to the present day.
The structure of the chapter is as follows. Section 2 surveys the existing
methodologies for estimating historical regional GDP and describes the types
of source that have been employed for the task; Section 3 showcases a particular
methodology through a case study of the NUTS-​3 Swedish regions from 1571
to the present; Section 4 concludes with some reflections on future research.

Historical regional GDP: Methodological challenges

The estimation of regional GDP is for various reasons intimately connected
to the effort to estimate national series. First, whenever possible, the regional
series are calculated using the same methodology as the national ones. This is
the case for bottom-​up series seeking to estimate production directly, which
represent the standard for most national GDP estimates but only a fraction of
the corresponding regional ones.1 National and regional series also share meth-
odology in the case of the proxy-​based approach. For instance, Crafts (1983)
estimated the GDP of a group of European countries where the data were too
poor to pursue a direct estimate of the production series (Spain, Switzerland,
Austria-​Hungary and the Netherlands). The estimates run from 1850 to 1910
 85

Regional GDP before GDP 85

and are based on a regression equation in which GDP per capita is a function
of a number of proxy variables such as letters posted per capita, share of non-​
agricultural active population, coal consumed and infant mortality rates. As we
will see in the next section, this methodology was also the one used for some
early regional series. It should also be noted that regional GDP figures are often
the result of a breakdown of the national ones according to some regional
principle; therefore, the quality of the regional series depends on that of the
national ones.
We have already observed that compiling regional GDP series for the pre-
sent day that are up to the standards of the national is a challenge. This is due
to conceptual issues that one encounters in trying to estimate GDP for sub-​
national units that are not sovereign and independent. According to Eurostat
(2013), examples of conceptual problems are “the regional allocation of finan-
cial transactions of enterprises with establishments in several different regions
or the regional allocation of the collective consumption of general govern-
ment.” The allocation of these transactions to one region rather than another is
very problematic. Moreover, an estimation of regional GDP following the same
conceptual guidelines as the national ones would require the quantification of
imports and exports for every [Link] administrative burden of this quanti-
fication for both firms and authorities would not be sustainable (Eurostat, 2013).
Historical GDP estimations carry with them the same conceptual problems as
the current estimations do, but face several additional data constraints. Under
these constraints, suitable methodologies can be used to estimate regional GDP
before the mid-​20th century. This section illustrates the most widely used one.

Proxy-​based approach for GDP

The estimation of regional GDP based on proxies was first introduced in the
1990s following the work by Crafts at the national level (1983). Good (1994)
applied a similar strategy to that illustrated above to estimate the regional GDP
for the Habsburg successor-​state territories, consisting of 22 regions of the
Habsburg Empire, with a ten-​year benchmark from 1870 to 1910. The nov-
elty of this approach at the time lay in using a regression equation to obtain
coefficients that were used in the next step to weight the different proxies.
This approach “let the data speak” in terms of the different weights for the
proxies, rather than having to decide the weight of each one arbitrarily. This
approach was, however, heavily criticized by Pammer (1997), who referred to
the estimates by Good (1994) as “statistical artifacts.” The criticism was mostly
directed at the specification of the estimating equation and the application of
the resulting coefficients to the most backward regions of the Empire. The first
claim was demonstrated, for instance, by replacing the share of non-​agricultural
labour force with the agricultural [Link] variable should have complemented
the one used by Good (1994), but it led to dramatically different results. The
reason suggested for this discrepancy was that GDP and the shares of workers in
the labour force show a non-​linear relationship because the latter (unlike GDP)
86

86  Kerstin Enflo and Anna Missiaia

are capped at 100 per cent. From a historical point of view, the use of the shares
of employment as a proxy of GDP was challenged because of the different levels
of development in the regions: the proxy was particularly poor for more back-
ward regions, because the benchmark for the relationship over time was a sample
from more advanced European countries. Similar points were brought forward
regarding the other proxy variables used. The critique by Pammer (1997) led
to a comprehensive revision of the series, published by Good (1999). Both the
estimating procedure and the underlying data were revised: the share of agri-
culture in the active population in the latest series followed a logistic curve over
time to accommodate the different rates of decline at different stages of indus-
trialization. Finally, maximum likely estimation was used rather than ordinary
least squares (OLS) and the other proxy variables relied on more disaggregated
district-​level data that helped to overcome measurement issues. The revised
series showed a higher GDP level with lower growth rates and less regional
inequality, which confirms that the previous estimates penalized the backward
regions of the Empire too heavily. All in all, Austria-​Hungary appears not to
have lagged far behind the frontrunners of Europe. Another country for which
the proxy variable approach to regional GDP was attempted in the 1990s was
Italy. The first scholar to approach the estimation of regional GDP for post-​
unification Italy was Esposto (1997). His estimates largely relied on the method
by Crafts (1983) that was also used by Good (1994; 1999). This work opened
the door to the estimation of regional GDP for a country that presents one of
the highest levels of regional inequality. The results showed a very wide GDP
gap between north and south as early as 1861, which we will later see reduced
by the estimates of Emanuele Felice using the G-​S method (Felice, 2011; 2018).

Proxy-​based approach for regional business cycles

Another approach to regional GDP should be mentioned; it is based on proxy
variables, aiming at the more limited scope of studying regional business cycles.
Two studies using such an approach focused on the historical cases of Italy
and the Habsburg Empire. The first was by Ciccarelli, Fenoaltea and Proietti
(2010), who used annual data on constructions to study the convergence of
regional short-​run business cycles in post-​unification Italy. The underlying
assumption here is that new constructions are a very sensitive measure of short-​
term fluctuations of the economy, and therefore of GDP.2 The authors there-
fore employed some standard time-​series techniques, such as the Hodrick and
Prescott filter and the measurement of the inter-​quantile range to assess the
regional convergence or divergence of the cycles. Another study to mention
among those using proxy variables is Ciccarelli and Missiaia (2018).The authors
here made use of a composite annual series based on indirect taxes in order to
assess the short-​term movements of regional GDP in Imperial Austria from
1867 to 1913. The taxes were collected from official statistical yearbooks and
cover all the 14 Austrian [Link] inclusion of specific taxes in the series was
tied to the concept of aggregate demand (following which GDP is composed
 87

Regional GDP before GDP 87

of consumption plus investment plus net exports). Each tax reflected one of
these components: for consumption, they used consumption taxes, lotto, salt,
tobacco; for investment, they used stamp duties and registration fees paid on
business transactions; for net exports, they used import duties. The authors
were of course aware that such a measure can provide only a sense of the co-​
movements of the regional series and that they reveal little on levels (either
absolute or relative). For this reason, they took some measures to correct the
proxy-​based series with the information available on the ten-​year benchmarks
from Schulze (2007). In particular, they used only the tax-​based series to fill the
gaps between benchmark years, letting the superior series from Schulze (2007)
set the level in the first and last year of each ten-​year period. For this reason,
the tax-​based series are used only to indicate movements between two known
levels. Since the series of taxes can be quite volatile from year to year, Ciccarelli
and Missiaia (2018) also used a moving average to smooth out possible outlier
years. Once the series were obtained, they were treated using standard filtering
techniques to separate trend and cycles and a measure of dynamic correlation
showed how the synchronization of the series evolved over time.

Regional employment-​based method
The methodology introduced by Geary and Stark (2002) takes a step for-
ward from the proxy variable approach. The “G-​S method” is a short-​cut
method through which national value-​added series are allocated to the regions
according to their share of total employment. In doing so, the authors assumed
that regional production depends on the number of workers employed in each
region. The other assumption was that differences in regional wages in each of
the main sectors of the economy reflect differences in productivity. Wages are
therefore used to correct the allocation based on employment. Regional GDP
Y in county i is defined in the following way:

j
 wij 
Yi = ∑ y j � β j   × Lij
 wj 

where yij is the average value added per worker in county i and sector j, and
Lij the number of workers in county i and industry j. � β j is a scalar reflecting
regional relative differences, ensuring that regional totals add up to the national
figure for each [Link] Geary-​Stark method in essence allocates value added
at the nation/​sectoral level to the regions on the basis of their shares of the
labour force corrected by wage differentials to take into account differences in
labour productivity.
The G-​S method has been widely used by economic historians for cases
in which direct estimates of production were not viable but census data pro-
vide good enough information on employment and at least some evidence on
wages was available. Geary and Stark (2002) used their method to estimate the
88

88  Kerstin Enflo and Anna Missiaia

GDP of the four countries of the UK for the census years between 1861 and
1911. As a proof of concept, the authors also compared the estimates reached
with this method with the direct ones for the year 1971, for which both were
[Link] results were close enough to consider the short-​cut method valid.
Since 2002, the G-​S method has been widely applied to other historical cases
and other periods in history. Examples of works using this method are Felice
(2005a; 2005b) on Italy; Rosés, Martínez-​Galarraga and Tirado (2010) on Spain;
Buyst (2010) on Belgium; Enflo, Henning and Schön (2014) on Sweden; Enflo
(2014) on Finland; and, recently Geary and Stark (2016) revising the series for
the UK.3
The G-​S method has several advantages. The procedure is fairly simple to
implement, provided the necessary data are available. It is also a flexible pro-
cedure, in that it allows employment and wage series from different sources
(and possibly of a different quality) to be used for different sectors. It is also flex-
ible in the number of sectors that have to be used in order to estimate regional
GDP. Conceptually, the minimum requirement is to be able to estimate separ-
ately the three main sectors of the economy (agriculture, industry and services);
however, a finer breakdown of the economy is welcome and leads to more
precise estimates because of the attribution of different average value added
per worker in different sectors and the ad hoc correction of the productivity
of each sector. This for instance is the case in the first generation series for Italy
produced by Felice (2005a; 2005b), which made use of the detailed regional
breakdown of employment from the population censuses already reclassified by
Fenoaltea (2003). An example of flexibility in the various sources used comes
from Enflo and Missiaia (2018), for which the series from 1750 to 1850 have
been estimated using the G-​S method (except for the mining sector). In their
case, the evidence for agricultural wages was far richer than that for industrial
wages:  the G-​S method allowed both sets of wages to be used in the same
estimation.

Direct estimates, mixed and alternative method

The flexibility of the G-​S method also allows the merging of sectors that
have been reconstructed with different methods. Two cases can be conceptu-
ally distinguished here. The first is the one where direct evidence on produc-
tion is available in some sectors but not all. In this case the rationale for not
using a G-​S approach is empirical in nature. Probably the most interesting of
these cases occurs in the work by Emanuele Felice, who produced a first set
of estimates in 2005 (Felice 2005a; 2005b) that was planned to use the G-​S
method for industry and later progressively replaced all the industrial series
with those produced by Ciccarelli and Fenoaltea (2009; 2014) (see Felice, 2011
and the subsequent revisions).4 The estimates of industrial value-​added for
the Italian regions represent an impressive and unprecedented effort because
they contain bottom-​up measures of production at the regional level. Their
inclusion in the regional GDP estimates of Italy places these estimates among
 89

Regional GDP before GDP 89

the most precise ever produced for any given country. Another example of
mixed methods is the work of Schulze (2007) on Austria-​Hungary. Here some
series were reconstructed directly, some were a breakdown of the national fig-
ures according to employment and wages, and others were inferred using a
proxy-​based approach. In particular, the estimates for the Hungarian part of
the Empire mirror the work by Good (1999), while direct estimates for the
Austrian part were made possible by following the sources used also at the
national level. Finally, manufacturing and construction were allocated according
to employment and wages.
The second modification to the G-​S method to be found in the litera-
ture is conceptual rather than being indicative of data availability. Crafts (2005)
proposed a modified version of Geary and Stark (2002) that takes non-​wage
income into account, in order to calculate the regional GDP of the UK regions.
This used the tax assessments for Great Britain in the period 1871–​1911.
According to Crafts, the G-​S method based on employment and wages uses
information on only half of the GDP (income from wages is estimated to be
50 per cent of the total). Geary and Stark (2016) in their reply to Crafts (2005)
maintained that the joint use of wages, employment and tax data creates bias
in the estimates and should be avoided. They justified their criticism on the
grounds that wage and employment data provide an output perspective, while
tax data provide an income perspective. The two should not be mixed to pro-
vide regional relative shares. Moreover, from an empirical point of view, they
argued that their method was demonstrably robust by estimating the bench-
mark of 1971 and comparing it to direct estimates, an exercise that cannot
be replicated for the estimates using the modification by Crafts (2005). The
response by Geary and Stark (2016) is still relatively recent and the debate
among scholars is still open. It is reassuring, however, that the overall long-​
term trend of regional inequality is similar in both methodologies, at least for
the British regions. Finally, the set of regional estimates of personal income for
the US between 1880 and 1910 by Klein (2009) is an important reference for
readers interested in income calculated from the consumption side rather than
that of production.5
It is beyond the scope of this chapter to survey all the sources and meth-
odological issues encountered by the numerous scholars who have produced
regional GDP estimates from the mid-​19th century to the present. Nevertheless,
in the next section we propose a more detailed discussion of one recent case
study that gives a flavour of working with historical regional GDP estimates.
This is the case of the Swedish regions in the period 1571–​2010.

Swedish historical national accounts: Sources and methods

Swedish historical sources are well known for their exceptional quality. Swedish
historical national accounts are no exception:  the national series compiled
by Schön and Krantz (2012; 2015) starting from the 16th century is just one
example. The reason for such richness in the sources lies in the relative stability
90

90  Kerstin Enflo and Anna Missiaia

of Sweden as a unified country (in spite of the several wars it waged with its
neighbours over the centuries). Moreover, after the Reformation, the overlap
between the State and the Church allowed the King to rely on the religious
institutions to collect taxes and record information on the population. In 1749,
the first national population census based on parish-​level data was released by
the Royal Census Committee (Kungliga Kommissionen över Tabellverket),
which can be considered one of the first statistical offices in the world. Censuses
have since been undertaken every ten years without interruption to the present
day. These population statistics constituted the backbone of our GDP estimates
from 1750 onwards and allowed us to create the longest series of regional GDP
for any given country (Enflo, Henning and Schön, 2014; Enflo and Missiaia,
2018) following the G-​S method. The 1571 benchmark presented in Enflo
and Missiaia (2018) is also a product of the extensive network of parishes con-
trolled by the King. The main source for this benchmark was a special wealth
tax called “Älvsborgs ransom” intended to regain control of the strategic castle
of Älvsborg after a seven-​year war against Denmark (1563–​1570).
The availability of these exceptional sources allowed us to start a regular
series of benchmark estimates about one hundred years earlier than gener-
ally possible for European countries. This brought a completely new contribu-
tion, in that it allowed regional inequality patterns to be observed well before
modern industrial growth. It also posed more methodological challenges than
19th-​century estimates did. A detailed account of all the sources and the meth-
odological choices made in estimating all benchmarks is well beyond the scope
of this [Link] therefore focus in the sub-​paragraph below, on the challenges
that the 1750–​1850 series posed, compared to those of 1860–​2010. After that,
in the sub-​paragraph that follows, we give an account of the estimation of the
1571 benchmark as an interesting case study for scholars, in our view, which
could promote access to similar sources for other countries.

Swedish regional accounts (1750–​2010): The longest regional series

The ten-​year benchmark series from 1750 to 2010 were produced by two
different studies: Enflo and Missiaia (2018) for the period 1750–​1850 and Enflo,
Henning and Schön (2014) for 1860–​2010. The two studies can be connected
without further input; they both sum up to common national GDP estimates.
In the case of Sweden, the longest and most detailed national series in terms
of sectorial breakdown are the Swedish Historical National Accounts (SHNA)
from 1560 onwards (Schön and Krantz 2012; 2015).The SHNA provide annual
estimates for agriculture, manufacturing, building and construction, transpor-
tation and communication, private services, public services and services of
dwelling.6 The entire series of regional GDP relies on the regional labour force
for each county and sector. For 1750–​1850, they come from the Tabellverket
provided by the CEDAR Umeå Demographic database. Here the number of
individuals in each social group and the corresponding job titles are provided
for each parish from its church registers. For the 1860–​2010 benchmarks,
 91

Regional GDP before GDP 91

various publications released by Statistics Sweden are used (for a detailed review,
see Table 2 of Enflo, Henning and Schön, 2014).
Some methodological issues had to be solved with regard to the labour force
figures. For instance, for the period 1750–​1850, women were not included
because of the difficult task of clearly defining their occupation. This is a fre-
quent challenge when reconstructing historical labour force figures and it is
often necessary to restrict the labour force to [Link] second challenge related
to the estimation of the labour force in services before 1860. The labour force
figures for services as a whole in 1850 appear overestimated by some 10 per
cent compared to those by Enflo et al. (2014) for 1860. Given that the share
of GDP at national level from the SHNA is fairly stable over time, the share
of workers in services was kept constant at the 1860 level in all years to avoid
possible bias from misreporting by the Tabellverket. The excess workers were
redistributed to the other sectors.
The G-​S method also required using the wages for each sector to correct
for regional productivity differentials. For the 1750–​1850 series, agricultural
wages were taken from the standard source, Jörberg (1972), which gives good
coverage of all the 24 Swedish counties and was also one of the main sources
for the 1860–​2010 estimates. The industrial wages in the earlier series proved
to be much harder to ascertain than in the 1860–​2010 series, for which Lundh
et al. (2004) is the main source. For the 1750–​1850 period a series of builders
for Stockholm from Söderberg (2010) and one for Malmö from Gary (2017)
were used to compile a “Stockholm versus the rest” index. To reconstruct the
wages for all counties, the wages of manufacturing workers in 1860 for all the
24 counties from Collin (2016) were extrapolated backwards in time using the
Stockholm-​Malmö ratio as a proxy of movements in time.7

A historical accident: The 1571 benchmark

The peace treaty of Stettin between Sweden and Denmark in 1570 stated that
Sweden had to pay 150,000 silver coins after the war to regain the Älvsborgs
castle, which was strategic for western access to the sea. This historical event
represents a unique opportunity to gather parish-​level data for 1571: in order to
pay the ransom, a wealth tax comprising a tenth of all cattle, agricultural surplus
and metal goods was imposed on the Swedish population. The tax collection
was carried out and recorded by local priests under the supervision of regional
bailiffs. These records were re-​organized and aggregated at county level by
Forsell (1869–​1875; 1872–​1883). Forsell remarks that the ransom can be seen
as an enumeration of all that was considered of value in Swedish homes, seeing
that chattels and smaller presses are often included. In this way, the source also
gives an indication of the value of consumer goods in a pre-​modern economy.
In order to move from a measure of wealth to a measure of the added
value of production, however, a few adjustments had to be made. The material
had to be made compatible with the industrial classification system in modern
statistics, and the values of the input were estimated and deducted from the
92

92  Kerstin Enflo and Anna Missiaia

values of the output. For a pre-​modern economy, this was in some ways less
complicated since most goods were produced relatively low down the value-​
added [Link] pioneering efforts to construct national GDP from the wealth
tax were made by Krantz (2004). He adapted the modern industrial classifica-
tion system to the source material and derived value-​added measures for all
the sectors. Unsurprisingly, the results show that agriculture made up the lar-
gest share of the pre-​modern GDP, although the value of production was not
completely negligible in “industries” such as mining and metal goods, wood
products and food and textiles.
Since employment figures and wages by county or industry do not exist for
Sweden in 1571, Enflo and Missiaia (2018) could not opt for the usual Geary
and Stark approach to regional GDP. Instead, the geographical scope of the
wealth tax data was used for the regional disaggregation of industrial value-​
added. As an example, regional breakdowns of agricultural yields and various
forms of cattle could be readily used to regionalize agricultural value-​added
while the existence of mines and metal works recorded in historical sources
helped to regionally distribute the value of metal production. A full account of
the guiding principles for the regionalization of industry value-​added can be
found in Enflo and Missiaia (2018). As far as possible, in the absence of regional
wage data, the estimates were adjusted to take into account different levels of
sectoral productivity between regions.
Although it is beyond the scope of this chapter to interpret the results of these
estimations, a brief overview of the Swedish evolution of regional inequality
can be provided by connecting the series by Enflo and Missiaia (2018) to those
of Enflo, Henning and Schön (2014). Figure 5.1 shows the coefficients of vari-
ation for all the available benchmarks.

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0
1571

1770
1780
1790
1800
1810
1820

1850
1860
1870
1880
1890

1920
1930
1940
1950
1960
1970

1990
2000
2010
1750
1760

1830
1840

1900
1910

1980

Figure 5.1 Coefficients of variation of Swedish regional GDP per capita.

Source: Enflo and Missiaia (2018).
 93

Regional GDP before GDP 93

For 1571, as expected for a highly agricultural and pre-​industrial economy,
regional inequality appears to be fairly low. What is surprising is the dramatic
increase in the coefficient of variation in 1750, which pre-​dates industrializa-
tion by some 100 years. Similarly, the decrease of regional inequality after 1850
corresponds to the industrialization of the country. The determinants of both
pre-​industrial inequality and of industrialization without regional inequality are
discussed at length by Enflo and Missiaia (2017) and Enflo and Rosés (2015).

Conclusions and future research

Economic inequality can be considered in many dimensions. In the past few
years, wealth and income inequality have been at the core of economic research
because of the recent surge of interest in these themes. However, interper-
sonal inequality has not been the only type that has engaged scholars. Regional
inequality within countries has also been increasing, with possible consequences
for growth, welfare and political outcomes (Rodríguez-​Pose, 2018). The study
of regional inequality today is also intimately related to inequality in the past.
This is because regional inequality is often rooted in imbalances that date back
at least to the onset of modern industrialization in countries, if not earlier.
Moreover, although regional inequality during the first industrialization has
been studied in the past few years for a number of countries (Rosés and Wolf,
2018b), the assessment of pre-​industrial inequality has proved to be exception-
ally difficult to assess. This is because of data limitations that will require more
effort to overcome. Until then, it will be hard to establish whether regional
divergence is a phenomenon that predates modern economic growth or is
observed only in connection with it.
This chapter is intended to provide a guide to existing empirical methods for
estimating regional GDP in different historical [Link] have here introduced
the most widely used methods, going from the proxy method approach for
cases in which there is little evidence about regional employment and wages,
to the popular Geary and Stark method of allocating national value-​added
according to employment and wages, finishing with the case of Sweden, where
a mix of different methods, including direct estimation, can be employed. The
challenges in estimating regional GDP before the mid-​20th century are mostly
empirical, since suitable series are hard to come by or are of poorer quality than
current ones. Some of the challenges are also conceptual, in that the further
back in time, the harder it is to divide economic activity into separate sectors
without overlap. For the case of Sweden, we see that the correct assessment
of women’s employment presents researchers with a hard task. Moreover, the
overlap between agriculture and private service employment makes it hard to
correctly assess the size of these two sectors from original sources: the use of
assumptions and proxies cannot be avoided. The Swedish series represent an
exceptional effort to estimate pre-​industrial regional inequality whether using
the conventional Geary-​Stark method for 1750–​1850 or using a less conven-
tional fiscal source for 1571.
94

94  Kerstin Enflo and Anna Missiaia

Future research might do well to go in three directions. First, even for periods
already covered by some estimates there is room for improvement through
the collection of better data and further analysis of the existing material. For
instance, a more targeted effort should be made to include women in the evi-
dence from Sweden. Second, it is probable that the time-​span in more countries
could be expanded by using other sources, as is the case for Sweden in 1571.
Some of these sources may be known but not yet employed for quantitative
reconstructions. Finally, the amplification of the series from benchmark years
to annual through short-​cut methods could provide interesting insights into
regional business fluctuations.

Notes
* The chapter was written thanks to the financial support of the Knut and Alice
Wallenberg Foundation for the project ‘The evolution of regional economies in
the Nordic region – a long run approach’ (ID: KAW2014.0138), Länsförsäkringars
forskningsfond: and the Crafoord and Handelsbanken Foundations for the project
‘Trade, market and regional development in pre-industrial Sweden (1750–1850)’ (ID:
20160665 and ID: P2016-0247:1 respectively).
1 Most of the series included in the Maddison dataset, at least for the 20th century,
are based on direct estimates of production following the framework of the histor-
ical national accounts; some regional GDP follow the same methodology for at least
some sectors. One example comes from the second generation series used for the
regional industrial value added by Ciccarelli and Fenoaltea (2009; 2014) that are
incorporated in the most recent series of regional GDP for Italy by Felice (2018).
2 The authors cite the classic French expression “quand le bâtiment va, tout va” to
evocate the economic relevance of the construction sector and its elasticity to short-​
term movements in the economy as a whole.
3 In this chapter we are going to discuss the sources used and the methodological
challenges faced for the Swedish GDP series. For a more detailed and organic review
of the existing series we refer the reader to Rosés and Wolf (2018b).
4 Daniele and Malanima (2007) also produce regional GDP estimates that have
prompted much debate among Italian economic historians because of their sur-
prising result that the North-​South gap was not visible in 1861. The estimates have
been criticized for the methodology adopted to allocate national production to
regions, which is not based on the G-​S methodology but makes use of the interpol-
ation of regional quotas over time.
5 The alternative approach is common between national and regional estimates in the
US and originates from different choices by the official statistical offices producing
income estimates.
6 The exception is the period 1750–​1850, for which mining is separated from the rest
of the industrial sector and information at parish level from Olsson-​Spjut (2007) is
used to directly estimate the level of production.
7 Wages for services are not available for 1750–​1850. Private services are proxied using
the agricultural wages since it is assumed that a high level of substitution prevailed
between agricultural workers and private services workers. Public services do not
have a wage correction, given the assumption of homogeneity of the type of employ-
ment across counties.
 95

Regional GDP before GDP 95

References
Bolt, J., Inklaar, R., de Jong, H. and van Zanden, J. L., 2018. Rebasing ‘Maddison’: New
income comparisons and the shape of long-​run economic development. Maddison
Project Working Paper 10.
Buyst, E., 2010. Reversal of fortune in a small, open economy:  Regional GDP in
Belgium, 1896–​2000. Rivista di Storia Economica 26, pp. 75–​92. [Link]
10.2139/​ssrn.1586762
Buyst, E., 2018. Growth and regional imbalances during Belgium’s classical Industrial
Revolution. Journal of Interdisciplinary History, 49(1), pp. 71–​ 92. [Link]
10.1162/​jinh_​a_​01232
Centre for Demographic and Ageing Research (CEDAR). The Demographic Data Base.
Umeå: Umeå University.
Ciccarelli, C. and Fenoaltea, S., 2009. La produzione industriale delle regioni d’Italia, 1861–​
1913:  Una ricostruzione quantitativa, 1.  Le industrie non manifatturiere. Roma:  Banca
d’Italia.
Ciccarelli, C. and Fenoaltea, S., 2014. La produzione industriale delle regioni d’Italia, 1861–​
1913: Una ricostruzione quantitativa, 2. Le industrie manifatturiere. Roma: Banca d’Italia.
Ciccarelli, C., and Missiaia, A., [Link] fall and rise of business cycle co-​movements in
imperial Austria’s regions. Annals of Regional Science, 60(1), pp. 171–​193. [Link]
org/​10.1007/​s00168-​017-​0850-​5
Ciccarelli, C., Fenoaltea, S., Proietti, T., 2010. The effects of unification: Markets, policy,
and cyclical convergence in Italy, 1861–​1913. Cliometrica, 4(3), pp. 269–​292. [Link]
[Link]/​10.1007/​s11698-​009-​0046-​z
Collin, K., 2016. Regional wages and labor market integration in Sweden, 1732–​ 2009.
Gothenburg: Gothenburg University Dissertation.
Crafts, N. F.  R., 1983. Gross national product in Europe:  Some new estimates.
Explorations in Economic History, 20, pp. 387–​ 401. [Link]
0014-​4983(83)90026–​8
Crafts, N. F.  R., 2005. Regional GDP in Britain, 1871–​ 1911:  Some estimates.
Scottish Journal of Political Economy, 52(1), pp. 54–​ 64. [Link]
j.0036-​9292.2005.00334.x
Daniele, V. and Malanima, P., 2007. Il prodotto delle regioni e il divario Nord-​Sud in
Italia (1861–​2004). Rivista di Politica Economica, 97, pp. 1–​49.
Enflo, K., 2014. Finland’s regional GDPs 1880–​2010: Estimates, sources and interpret-
ations. Lund Papers in Economic History, 135.
Enflo, K. and Missiaia, A., 2019. Between Malthus and the industrial take-​off: Regional
inequality in Sweden, 1571–​1850. Economic History Review, forthcoming.
Enflo, K. and Missiaia, A., 2018. Regional GDP estimates for Sweden, 1571–​1850.
Historical Methods, 51(2), pp. 115–​137.
Enflo, K. and Rosés, J. R., 2015. Coping with regional inequality in Sweden: Structural
change, migrations and policy, 1860–​ 2000. Economic History Review, 68(1), pp.
191–​217.
Enflo, K., Henning, M., and Schön, L., 2014. Swedish regional GDP 1855–​
2000:  Estimations and general trends in the Swedish regional system. Research in
Economic History, 30, pp. 47–​89.
Esposto, A. G. 1997. Estimating regional per capita income: Italy, 1861–​1914. Journal of
European Economic History, 26(3), pp. 585–​604.
Eurostat, 2013. Manual on regional account methods. Luxemburg: Eurostat.
96

96  Kerstin Enflo and Anna Missiaia

Felice, E., 2005a. Il reddito delle regioni italiane nel 1938 e nel 1951. Una stima basata
sul costo del lavoro. Rivista di Storia Economica, 21, pp. 3–​30.
Felice, E., 2005b. Il valore aggiunto regionale. Una stima per il 1891 e per il 1911 e
alcune elaborazioni di lungo periodo (1891–​1971). Rivista di Storia Economica, 2, pp.
273–​314.
Felice, E., 2011. Regional value added in Italy, 1891–​2001, and the foundation of a
long-​term picture. Economic History Review, 64(3), pp. 929–​950.
Felice, E., 2018. The roots of a dual equilibrium:  GDP, productivity, and structural
change in the Italian regions in the long run (1871–​2011). European Review of
Economic History, forthcoming.
Fenoaltea, S., 2003. Peeking backward: Regional aspects of industrial growth in post-​
unification Italy. Journal of Economic History, 63, pp. 1059–​102.
Forsell, H., 1869–​1875. Sveriges inre historia från Gustaf I, 1–​2. Stockholm: P. A. Norstedet
& Söners Förlag.
Forsell, H., 1872–​1883. Sverige 1571. Försök till en administrativ-​statisktisk beskrifning öfver
det egentliga Sverige, utan Finland och Estland. Stockholm: P. A. Norstedet & Söners
Förlag.
Gary, K., 2017. Constructing equality? Women’s wages for physical labor, 1550–​1759.
Lund Papers in Economic History, 158.
Geary, F. and Stark, T., 2002. Examining Ireland’s post-​famine economic growth per-
formance. Economic Journal, 112(482), pp. 919–​935.
Geary, F. and Stark, T. 2016. Regional GDP in the UK, 1861–​1911:  New estimates.
Economic History Review, 68(1), pp. 123–​144.
Gerschenkron, A. 1962. Economic backwardness in historical perspective, a book of essays.
Cambridge, Massachusetts: Belknap Press of Harvard University Press.
Good, D. 1994. The economic lag of central and eastern Europe:  Income estimates
for the Habsburg successor states, 1870–​1910. Journal of Economic History, 54(4), pp.
869–​891.
Good, D. 1999. The economic growth of Central and Eastern Europe in comparative
perspective, 1870–​1989. European Review of Economic History, 3(2), pp. 103–​137.
Jörberg, L. 1972. A history of prices in Sweden 1732–​1914. Part II. Lund: CWK Gleerup.
Klein, A., 2009. Personal income of U.S. states: Estimates for the period 1880–​1910.
Warwick Economic Research Papers, 916.
Krantz, O., 2004. An estimate of Swedish GDP in 1571. In S. Heikkinen, and J. L. van
Zanden, ed. 2004. Exploring economic growth: Essays in measurement and analysis; a fest-
schrift for Riita Hjerppe on her 60th birthday. Amsterdam: Aksant. Pp. 105–​130.
Kuznets, S. 1934. National income, 1929–​1932. National Bureau of Economic Research.
Lundh, C., Olofsson, J., Schön, L., and Svensson, L., [Link] formation, labor market
institutions and economic transformation in Sweden 1860–​2000. Lund Studies in
Economic History, 32.
Maddison, A., 2007. Contours of the world economy 1-​2030 AD: Essays in macro-​economic
history. Oxford: Oxford University Press.
Milanović, B., 2016. Global inequality: A new approach for the age of globalization. Cambridge,
Massachusetts: Harvard University Press.
Olsson-​Spjut, F., 2007. Järnhanteringens dynamik: Produktion, lokalisering och agglomerationer
i Bergslagen. Umeå: Department of Economic History Dissertation.
Pammer, M. 1997. Proxy data and income estimates: The economic lag of central and
eastern Europe. Journal of Economic History, 57(2), pp. 448–​455.
 97

Regional GDP before GDP 97

Piketty, T., 2014. Capital in the twenty-​first century. Cambridge, Massachusetts: Harvard
University Press.
Rodríguez-​Pose, A., 2018. The revenge of the places that don’t matter (and what to do
about it). Cambridge Journal of Regions, Economy and Society, 11(1).
Rosés, J. R. and Wolf, N., 2018a. Regional economic development in Europe, 1900–​
2010: A description of the patterns. CEPR Discussion Paper 12749.
Rosés, J. R. and Wolf, N., 2018b. The economic development of Europe’s regions: A quantita-
tive history since 1900. London: Routledge.
Rosés, J., Martínez-​Galarraga, J. and Tirado, D., 2010. The upswing of regional income
inequality in Spain (1860–​1930). Explorations in Economic History, 47(2), pp. 244–​257.
Schön, L. and Krantz, O., 2012. The Swedish economy in the early modern
period:  Constructing historical national accounts. European Review of Economic
History, 16, pp. 529–​549.
Schön, L., and Krantz, O., 2015. New Swedish historical national accounts since the
16th century in constant and current prices. Lund Papers in Economic History, 140.
Schulze, M. S., 2007. Regional income dispersion and market potential in the late nine-
teenth century Hapsburg Empire. LSE Economic History Working Papers 106/​07.
Söderberg, J., 2010. Long-​ term trends in real wages of labourers. In:  R. Edvinsson, T.
Jacobsson, and D. Waldenström, ed. Exchange rates, prices, and wages, 1277–​ 2008.
Stockholm: Ekerlids Förlag & Sveriges RB. Pp. 453–​478.
98

6 
Comparative research designs
Interdependence as challenge and
opportunity in regional studies
Martin Åberg and Thomas Denk

Introduction: Purpose, terminology, and layout of the chapter

Any combination of historical and social science approaches to regional devel-
opment problems involves methodological challenges (cf. Capocci and Ziblatt,
2010; Møller, 2013). The application of comparative research designs is a case
in point. On the one hand, the classic approach taken by historians, as well as
by many historical sociologists, is intrinsically narrative, as illustrated by, among
others, Bloch (1952 [1928]), Anderson (1974), Bendix (1964), and Sewell (1967
see also Kocka, 1999). On the other hand, social scientists typically tend towards
systematically structured research designs for making comparisons, although
often exclude the temporal dimension of societal and political change (Almond
and Verba, 1963; Esping-​Andersen, 1990; Lijphart, 1999). We argue that his-
torical comparative research can be of great value for exploration of the sub-​
national, regional level, but in order to enhance its explanatory potential this
requires systematic application of structured research designs. In turn, this draws
attention to a number of challenges involved with making comparisons.
Typically, regional-​level studies involve only a limited number of units of
analysis. In comparisons that involves the study of only a few regions, however,
it is not possible to test several, alternative explanations, since the effects of these
explanations cannot be isolated from each other. For example, we might for-
mulate two competing hypotheses, each one claiming a specific condition as an
explanation as to why two different regions are wealthy; yet, if the regions share
the same two conditions, it is impossible to test which one explains regional
wealth, or whether it is the combination of the two conditions that is important.
In addition, comparisons of few units have limitations with respect to how far
we may assume that the empirical observations represent general patterns. From
both challenges it follows that the selection of units, but also the selection of
cases and data, becomes a crucial task.
Finally, and importantly, a key premise of structured comparative research is
the assumption that the units of comparison are independent of each other. This
assumption, though, often proves incorrect. Rather than being independent
of each other, the units are empirically interdependent and this, in turn, may
lead the analyst to erroneous conclusions about causality. Arguably, therefore,
 99

Comparative research designs 99

interdependence represents the single most important challenge involved with
making synchronic and diachronic comparisons of few regions.
Among historians the problem is often conceptualized in terms of how the
units of analysis, such as regions or countries, share “entangled” histories, and
how processes take colour from cultural and other types of “transfers” across pol-
itical and administrative borders (Werner and Zimmermann, 2006; Ther, 2009;
Christian, 2016, pp. 3–​9). Among social scientists, the problem is, as indicated,
analyzed in terms of how the units of comparisons are interdependent by force
of spill-​over effects, or by diffusion of traits between them. More to the point,
interdependence are mutual connections between cases, such as regional-​level
industrial enterprises, or regional policy-​makers. Interdependence –​if not con-
trolled for –​creates uncertainty about the extent to which conditions (inde-
pendent variables) relate causally to the observed outcome. To fail to identify
correlations between cases, and not account for potential sources of inter-
dependence consequently means that any explanation of outcomes will suffer
from decreased reliability, whether the topic of analysis is industrial development
or policy changes (Braun and Gilardi, 2006; Franzese and Hays, 2008; Naroll,
1965; Przeworski and Teune, 1970; Simmons, Dobbins and Garrett, 2006).
Whereas there have been considerable advancements in the study of inter-
dependence within the framework of large-​N (“large number”) studies in
the last decades, a similar development has not occurred considering small-​
N (“small number”) comparative research designs. Regardless of whether the
latter type of research designs draw on qualitative or quantitative methods,
however, or on any combination of such methods, the main challenges involved
with the study of interdependence remains the same. This holds for historical
comparative research, too. Importantly, and on condition that other challenges,
e.g. those mentioned in the above, are addressed by the research design, we also
propose that interdependence is not just a methodological problem to small-​N
comparisons. Rather, the study of interdependence provides small-​N compara-
tive research with an opportunity as well.
Among others, Rokkan (1966) has questioned the assumption that nation-​
states are homogeneous units of analysis. More recently, Snyder (2001, p. 93) has
suggested that “subnational comparisons can expand and strengthen the meth-
odological repertoire available to social science researchers”. Whereas Snyder
does not pay attention to historical perspectives, we add to his observation that
similar gains may be won by historians making comparisons. Adding systematic,
comparative research designs within their methodological repertoire, provides
historians with a powerful tool for moving beyond the limits of traditional case
studies of regions and localities. Specifically, the study of interdependence and
its temporal dimension has the potential to provide students of regional themes
with more accurate explanations, for example with respect to problems such as
regional resilience, and policy analysis (cf. e.g. Martin, Sunley and Tyler, 2015).
The topic of this chapter, then, are designs for comparative research
involving small numbers of regions, and approaches for inclusion of synchronic
as well as diachronic comparisons with these designs as a means to analyze
100

100  Martin Åberg and Thomas Denk

interdependence. Since policy analysis and policy diffusion, among other
regional themes, involves problems pertaining to values, culture, and “history”,
there is, indeed, a strong case for comparative, historical studies in the [Link], a
clarification with respect to terminology is required before we proceed. As pre-
viously indicated with the help of the notion of “transfers”, there are parallels
between historians and social scientists when dealing with interdependence.
However, historical transfer approaches lack of clarity with respect to key
concepts and definitions (recently e.g. Ihalainen, 2017, pp. 23–​30). For example,
Werner and Zimmermann claim that “comparison favor[s]‌the implementa-
tion of a synchronic reasoning [in explanations], and [that] transfer studies tend
toward an analysis of diachronic [sic] processes” (2006, p. 50). At the same time,
e.g. Ther (2009, pp. 217–​218), stresses that the definition of “cultural transfer”
remains an open question.
At the same time, the study of interdependence and causality is a well-​
established topic among social scientists, notably among students of policy
diffusion (Simmons and Elkins, 2004; Weyland, 2006; Shipan and Volden, 2008;
Volden, Ting and Carpenter, 2008). In the following, we therefore use the ter-
minology involved with this strand of research, but include, when deemed
necessary, clarifications. Notions such as, e.g. “unit” and “case” have slightly
different meaning and usage among different disciplines. In order to avoid any
confusion in the following, we use a terminology that includes unit of ana-
lysis (e.g. regions), case (in order to describe units of observation, e.g. regional
policies), factor, and variable in order to refer to condition (in terms of inde-
pendent variable, or factor), and outcome (dependent variable or factor).
In the following sections, we first expatiate on the definition of interdepend-
ence. We identify three critically important dimensions of interdependence
(spatial, temporal and temporal-​ spatial interdependence). In connection to
this, we elaborate on our reasoning on interdependence, causality, and explan-
ation. Second, we apply a simple typology in order to structure similarities and
differences between three basic types of comparative research design. We dem-
onstrate that they offer different options for dealing with the three dimensions
of interdependence that we identify. Notably, approaches that combine syn-
chronic and diachronic comparisons are particularly challenging  –​but also
promising –​with respect to the analysis of interdependence. In line with this,
finally, we suggest Comparative Sequence Design (CSD) as a systematic means
of facilitating controlled synchronic and diachronic comparisons. Application
of CSD e.g. within policy analysis, a theme which traditionally tends towards
synchronic and hence static comparisons, allows for the study of policy-​making,
policy diffusion and policy implementation as dynamic, historically contingent
processes.

Interdependence: Spatial and temporal dimensions

Almost any number of approaches are possible with respect to the problem of
explaining historical change, including the matter of whether there are societal
 101

Comparative research designs 101

and other regularities across time. Likewise, e.g. to Skocpol (1984, p.  385),
though, we agree that history does involve certain patterns and regularities.
Hence, too, the exploration and causality of such patterns and regularities is a
valid, methodological concern, as demonstrated by systematically structured,
comparative research designs. However, it is precisely in connection to this, that
interdependence emerges as a challenge –​and an opportunity.
As indicated, one chief source of interdependence is diffusion-​processes.
Conditions and outcomes spread among the spatial units of analysis (spatial
interdependence). Let us by way of example consider a sample of ten administra-
tive units (regions) from one country –​chosen, let’s say, to analyse how regional
innovation policies are formulated and implemented. Let us for the sake of
argument also assume that the administrative make-​up of the ten regions is
similar. Finally, let us also assume that, albeit each policy differs with respect to
the specific challenges that they address, there are nevertheless substantial simi-
larities between them, in terms of the overall strategies that the policies propose
in order to promote innovation. Administrative similarities between the regions
would lead the analyst to hypothesize that this condition explains the outcome,
i.e. observed similarities in policy-​making. This hypothesis cannot be verified,
though, if we do not at the same time examine and rule out the possibility
that the similarities are a result of diffusion between the ten regions, by way
of learning or, perhaps, emulation. If the latter is indeed the case, diffusion of
traits between regions means that ten supposedly independent cases of policy-​
making are, in fact, just one, single case. Interdependence caused by diffusion,
then, provides a poignant illustration of the challenges involved with analysing
causality with the help of controlled comparisons:  it is necessary to separate
correlations between factors on the unit-​level from effects caused by factors in
other units.1
Cross-​border regions, i.e. regions that connect with other regions across
nation-​state borders, provide further illustrations to the problem. For example,
cross-​border regions are suitable for formulating and testing hypotheses
about the malleability of administrative borders to various types of processes
(e.g. Wolff, 2003; Laven and Baycroft, 2008). Among other Askarov and
Doucouliagos (2015), demonstrate that foreign aid spreads between countries
not only through official channels but, importantly, also by informal channels,
specifically through migration of guest workers. At the same time, this illustra-
tion also draws attention to an important distinction –​that between direct and
indirect interdependence.
Direct interdependence by diffusion (or “outcomes diffusion”) means that
outcome Y in region A also causes outcome Y in region B. In our first example
in the above, a regional policy may emerge in region A, and then spread to
region B, which explains why both regions come to share the same policy.
Indirect interdependence, though, involves diffusion that causes effects on the
independent variable(s). That is, condition X in region A, changes condition X
in region B, thereby leading to outcome Y. In line with our second example, on
foreign aid, diffusion of aid between regions in two countries may explain why
102

102  Martin Åberg and Thomas Denk

the level of economic wealth increases in both countries, if we consider foreign
aid in terms of condition (X) that spreads, and leads to a similar outcome (Y) in
two different cases. Analyzing interdependence in this form too therefore opens
up opportunities within the regional studies theme for the exploration of new
problems involving causality and explanation.
In addition to the distinction between direct and indirect interdependence,
there is also the difference between spatial interdependence and spatial auto-
correlation to consider. Occasionally, these are confused with each other in
research. However, the crucial difference is that spatial autocorrelation occurs
when attributes (e.g. conditions or outcomes) cluster together spatially in
countries or regions. This creates a strong correlation between location and
attributes. Possible examples involve situations in which democratic regimes are
concentrated in certain macro-​regions, whereas non-​democratic regimes tend
to be more frequent in other macro-​regions (O’Loughlin et al., 1998). Another
example are situations, for instance, when an observed degree of institutional
efficacy is larger in some sub-​national regions, compared to other regions in
the same country (Putnam, 1993).2 By contrast, spatial interdependence refers
to dependence between attributes not solely in terms of spatial co-​location
(note that the difference between autocorrelation and interdependence can also
occur temporally).3
This brings us to what is at the heart of the historian’s craft –​temporality.
First, change, such as by diffusion of traits, results from  –​in a wide sense  –​
“historical” processes (Rogers, 1995; Strang, 1991; Elkins and Simmons, 2005).
Second, causal explanations of change therefore necessarily involve an element
of time and, consequently, must consider temporal [Link] inter-
dependence means that factors within the same spatial unit of analysis exercise
effects on outcomes; but, importantly, these factors appear in the unit prior
to outcomes and, hence, the effects extend across a shorter or longer period
of time. Processes that illustrate this type of dynamics can include anything
from regional diffusion and adoption of new crops in early modern society
(Cipolla, 1972; Overton, 1985), to changes brought about by any type of tech-
nical innovation, business behaviour or policy in modern-​day regions (Shipan
and Volden, 2008; Balland, Boschma and Frenken, 2015; Boschma, Martin and
Minondo, 2017). Some of these processes might occur with great strength and
at relatively high speed, perhaps even more so in the era of globalization, such
as indicated by studies of policy diffusion (Jahn, 2006; Franzese and Hays, 2008).
To include temporal interdependence in the analysis offers regional studies
with opportunities for the systematic incorporation of process-​oriented models
and approaches with the research design; but, importantly, with consideration
for the limitations imposed by available, historical sources.
The above illustrations demonstrate that interdependence can involve
both region-​to-​region and nation-​to-​region connections as well as global
processes. Interdependence can be direct or indirect, synchronically as well
as diachronically. Interdependence is also, in both these dimensions, dis-
tinct from autocorrelation. More important for the purpose of this chapter,
 103

Comparative research designs 103

however, is that sources of spatial and temporal interdependence can also
combine into temporal-​ spatial interdependence, and thereby provide new
challenges to the possibilities of formulating causal explanations (see below,
Table 6.1; cf. Wellhofer, 1989).
In connection to this, it is important to note that the logic underpinning
causal explanations, tested by the help of controlled comparisons, are the same,
regardless of the type of interdependence involved, and regardless of discip-
linary [Link] logic of explanation by means of comparison harks back
to John Stuart Mill (1875), who introduced the “method of agreement” and
the “method of difference” and which, later, was developed by Przeworski and
Teune (1970) into “most different systems design” (MDSD) and “most similar
systems design” (MSSD).4 Controlled comparisons aim systematically to elim-
inate all conditions that (by way of hypothesis) lack any effect on the observed
outcome. For example, the method of agreement compares units in which the
outcomes are similar to each other. The analysis excludes all conditions that
make the units different from each other; potentially, these conditions cannot
explain the outcome. Explanations of this type are therefore essentially similar
to what McCullagh (2004, pp. 174–​176) has termed “genetic explanations” in
the historical disciplines. That is, when social scientists in comparative research
define relevant independent variables and examine these, this is equivalent to
when historians eliminate step-​by-​step less likely conditions from the analysis,
and focus on those key conditions which, in all probability, caused a specific out-
come (McCullagh, 2004, p. 176).5
Table 6.1 summarizes the three dimensions of interdependence in terms of
the types of effects that they exercise with respect to causality.6 The first type
of effect occurs when factors –​either a condition (X), or an outcome (Y) –​in
one unit (s2) affect the outcome in another unit (s1) simultaneously in time (t1)
(Yt1,s2 => Yt1,s1; Xt1,s2 => Yt1,s1). Spatial interdependence has, by definition, no
temporal dimension. However, since causality always involves an element of
temporality (i.e. factors lead to an outcome), and since interdependence can
be not only direct, but also indirect, it should again be stressed that Table 6.1
only indicates how spatial and temporal dimensions are normally dealt with in
different research designs.

Table 6.1 Spatial and temporal dimensions of interdependence

Types of effects Spatial Temporal Form of causality

interdependence interdependence

Effects from spatial + -​ Yt1,s2 => Yt1,s1

interdependence on Yt1,s1 Xt1,s2 => Yt1,s1
Effects from temporal -​ + Yt-​n,s1 => Yt1,s1
interdependence on Yt1,s1 Xt-​n,s1 => Yt1,s1
Effects from temporal-​spatial + + Yt-​n,s2 => Yt1,s1
diffusion on Yt1,s1 Xt-​n,s2 => Yt1,s1
104

104  Martin Åberg and Thomas Denk

The second type of effect are effects from factors on the outcome in the
same unit, but factors that relate to an earlier stage in the development of the
unit (Yt-​n,s1 => Yt1,s1; Xt-​n,s1 => Yt1,s1). By definition, this type of effect has a tem-
poral dimension but no spatial dimension, since the factors are endogenous to
one unit. That is, the notion of temporal interdependence, first, underlines how
the temporality of cause and outcome is a neglected problem in analyses that
only account for spatial interdependence. Second though, and similar to spatial
interdependence, temporality is at the same time indicated, since effects causing
temporal interdependence, too, can be direct or indirect. For example, condi-
tion X in region A changes X in region B, which eventually leads to outcome
Y in region B.
Temporal-​spatial interdependence, finally, refers to situations when factors
pertaining to preceding periods affects the outcome in another unit (Yt-​n,s2 =>
Yt1,s1; Xt-​n,s2 => Yt1,s1). In research designs made to include this type of caus-
ality, the effects have temporal as well as spatial dimensions. As in our pre-
vious examples, interdependence in these cases can be direct but also indirect.
Drawing on the logic outlined in Table 6.1, we will now review different com-
parative research designs with respect to how they deal with interdependence.

Comparative research designs and interdependence: A

typology
Social scientists and historians use several different approaches to classification
of comparative research designs. For example, Peters (2014) but also Landman
and Carvalho (2016), define research designs in terms of case studies (of single
units), comparative studies (involving few units), and statistical studies (involving
many units). Similarly, historians have discussed how case study design relates
to micro-​history approaches in cases when the latter is applied to single units
of analysis (Joyner, 1999; Alapuro, 2012). For our purpose, classifications such
as these are misleading. Considering the combinations of spatial and temporal
dimensions that interdependence may involve, it is the distinction between syn-
chronic and diachronic comparisons, rather than the number of units, that is
important to our classification, i.e. the potential for each research design to
allow analysis of spatial, temporal and temporal-​spatial interdependence.
Consequently, we suggest an alternative typology, one that focuses on how
the synchronic and diachronic dimensions are dealt with in different research
designs (Table  6.2). Typically, a comparative study will tend towards stressing
either a synchronic (cross-​sectional design) or a diachronic analysis (devel-
opment design). Studies based on the principles of comparative sequence
designs are an exception since they systematically incorporate both types of
comparisons. CSD therefore has the potential to control for effects causing
both spatial, temporal and spatial-​temporal interdependence in small-​N studies
(note that we ignore case study design, i.e. static studies of a single region,
since case studies neither involve controlled comparisons, nor include options
for analyzing interdependence).7 In addition, Table 6.2 includes references to a
 105

Comparative research designs 105

Table 6.2 Synchronic and diachronic dimensions in comparative research designs

No synchronic Synchronic
dimension dimension

No diachronic Case study design Cross-​sectional design (e.g. Kubicek,

dimension 2000; Doloreux, 2004; Reiser,
Rademacher and Jaeck, 2008)
Diachronic Development design (e.g. Comparative sequence design
dimension Loughlin, 2007; Møller, 2017) (e.g. Evenhuis, 2016, Basta, 2018)

few examples of studies that, albeit in different ways, involve regions as units of
comparison, and which illustrate the logic of each research design.
In Table  6.2, the logic of cross-​sectional design is illustrated by Kubicek’s
(2000) analysis of political behaviour in the L’viv and Donets’k oblasts (adminis-
trative regions) in Western and Eastern Ukraine; by Doloreux’s (2004) study of
entrepreneurship in the two, Canadian regions of Ottawa and Beauce; and by
Reiser, Rademacher and Jaeck’s (2008) study of independent local lists (ILLs) in
different German Länder.8 Development design, i.e. studies of a single unit across
time, is demonstrated with the help of Møller’s (2017) analysis of the forma-
tion of representative political institutions in medieval Aragon, and Loughlin’s
(2007) study on ethno-​nationalism and policy-​making in Northern Ireland,
c.1880–​1920. As illustrations to the logic of CSD, we draw on Basta (2018)
and Evenhuis (2016). The first of these two studies focuses on the problem
of territorial reform and political stability in Canada and Spain, whereas the
latter is a multi-​scalar analysis of economic change, policy and government in
South Saarland and Teesside, c. 1960–​1980. We will now proceed by focusing
on cross-​sectional design and in particular the challenge of spatial interdepend-
ence. After that, we devote one section to development design and temporal
interdependence and, finally, comment on comparative sequence design and
temporal-​spatial interdependence.

Cross-​sectional design and spatial interdependence

Cross-​sectional design is a standard approach not only for comparisons of macro-​
level regions and nation-​states, but also sub-​national level units. Comparisons in
the latter case, though, typically include only a small number of units (Kubicek,
2000; Doloreux, 2004; Table  6.2). Cross-​sectional design is also the simplest
illustration of the logic of explanation by comparison, although its stress on
synchronic comparisons indicates some of the key problems involved with ana-
lysing interdependence.
Put simply, cross-​sectional design is a means of structuring comparisons in
order to avoid comparing “apples with pears”. By way of illustration we may
return to our two hypothetical cross-​border regions  –​region A  and region
B –​but we now include the assumption that these regions do not only share
106

106  Martin Åberg and Thomas Denk

an observed similarity in terms of e.g. political behaviour, or regional develop-
ment policy. We also assume that regions A and B are fundamentally different
from each other with respect to a number of key conditions: region A might
be highly industrialized, whereas region B is agricultural. Provided that political
behaviour nevertheless tends to depend on one certain and similar key condi-
tion, consequently, the aim of the comparison of region A and region B is to
determine, according to the logic of the “method of agreement”, whether there
is any such, so far undetected similarity, typical to both regions. The “method
of difference” simply follows the opposite logic:  a difference between units
explain a difference in terms of outcomes (see previous section).
Importantly, though, comparisons made on the basis of cross-​sectional design
assume unit-​homogeneity, i.e. that the units of analysis do not somehow relate
to each other empirically. Consequently, diffusion of conditions or outcomes
between regions A and B can, hypothetically speaking, exercise effects that lead
to spatial interdependence (Table 6.1; Przeworski and Teune, 1970, p. 52). Still,
there are also situations in which it is more or less unlikely that there are any
such unit-​to-​unit effects on the observed outcome. For example, geograph-
ical distance between units, but also regional and nation-​level institutional
differences (in studies involving regions in different countries), suggests that
unit-​to-​unit diffusion of policies is less likely to intervene as a factor in the ana-
lysis (Balland, Boschma and Frenken, 2015; Boschma, Maroccu and Paci, 2016).
Presumably the first of these two conditions is relevant in Doloreux’ (2004;
Table 6.2) study of entrepreneurship in Ottawa and Beauce, at least to some
extent. First, Doloreux describes the two Canadian regions as differentiated and
highly specialized, i.e. features that perhaps make unit-​to-​unit effects leading to
spatial interdependence less likely, even within the same nation-​state context
(Doloreux, 2004, pp.  489–​490). Second, though, Doloreux at the same time
notes that entrepreneurship in both regions actually draws a lot on “extra-​
regional networks” (2004, pp. 489–​490).
By comparison, a situation that is more likely to involve spatial interdepend-
ence is illustrated by Kubicek’s (2000; Table  6.2), analysis of West and East
Ukrainian regions (L’viv and Donets’k). Whereas the two regions are mostly
different from each other, notably in terms of political preferences among the
voters (Kubicek, 2000), we could still be dealing with a case of spatial inter-
dependence, caused by direct or indirect diffusion. Outcomes in terms of voting
behaviour among citizens in L’viv may cause effects on the voting behaviour in
Donets’k (or vice versa), since elections in the two regions are part of the same,
nation-​level institutional arrangements.
As already indicated, though, cross-​sectional design does include options for
controlling spatial interdependence. Within the framework of small-​N cross-​
sectional studies, either the selection of units, or spatial-​lagged variables,9 can
be used as means to control for spatial interdependence (Caramani, 2008).With
respect to the first strategy, Doloreux’s (2004) above-​mentioned study is some-
what ambiguous, but the guiding principle is nevertheless to select regions
that, beyond reasonable doubt, lack any connections whatsoever between them
 107

Comparative research designs 107

(real-​world illustrations can include regional development policies in Australia
and Sweden). According to the second strategy, the regional development policy
e.g. in region A is included to control for diffusion effects on the development
policy in region B.
But cross-​ sectional research design also suffers from methodological
weaknesses. It can involve analysis of certain elements of temporality, as reflected
in the use of temporal-​lagged variables,10 or by inclusion of the “historical
backgrounds” of the units of analysis. Yet, cross-​sectional comparisons of units
from different, historical contexts are, by definition, problematic in this type of
research design, since it is “unlikely that the same bases of [causal] inference are
equally valid for all [social, political et  al. types of] systems” (Przeworski and
Tune, 1970, p.  113). Although cross-​sectional design has inspired models for
process-​tracing, i.e. one of the methods often used in studies made according
to development design (see next section), in its original form lacks options
for specifying and analysing temporal and temporal-​spatial interdependence
(Table  6.1). The opportunities that analysts miss out because of this can be
illustrated with the help of the last example listed in Table 6.2, that is, Reiser,
Rademacher and Jaeck’s (2008; Table  6.2) study of German independent
local lists.
In a de-​centralized, political system such as the Bundesrepublik, local polit-
ical parties of this type have a potentially important role to play from a local
and regional development perspective. However, their performance, including
voter support, differs between regions and localities, and the purpose of Reiser,
Rademacher and Jaeck’s study is to explain these differences. Their conclusion
is that the performance of ILLs is most likely tied to historically-​transferred
differences in “political culture” (2008, pp.  142–​143). These variations, how-
ever, are analysed by studying cross-​sectional differences in institutional design
among the units (regions). Apart from the fact that behavioural and/​or attitu-
dinal data is more suitable as an indicator of political culture, the study, because
of its cross-​ sectional design, overlooks the opportunity to investigate how
historically-​transferred differences in this relate causally to each other in the
temporal dimension (Table 6.1).

Development design and temporal interdependence

Methodologically, the opposite of cross-​sectional design is development design,
since the latter does not include synchronic comparisons (and therefore, typ-
ically, no options for specification and analysis of spatial interdependence,
although there are certain exceptions, see below). Rather, development design
serves the purpose of making diachronic comparisons in analyses of one single
unit. Studies made according to development design therefore face other types
of challenges compared to cross-​sectional design, but also open up opportun-
ities for which the former type of study does not allow.
First, development design indicates the problem of causal lags in connection
to diachronic comparisons. As previously pointed out, processes and, therefore,
108

108  Martin Åberg and Thomas Denk

change always includes a temporal dimension. It takes time for the independent
factor to cause effects in the dependent factor: how long does it take for a con-
dition X in region A to cause an observed outcome with respect to outcome Y?
Examples of outcomes can include features of regional policy or regional gov-
ernance. That is, contrary to synchronic comparisons in which focus is on the
causal relation between X and Y at a given, static moment in time, diachronic
comparisons focus on changes between X and Y across time (Table 6.1).
Second, temporal interdependence also involves the contingency that the
dependent factor might exercise effects on the independent factor, rather than
the other way around. Causality is not necessary linear. Contrary to theoretical
predictions in a study of region A, outcomes relating e.g. to regional policy, can
involve a causality characterized by interaction, or feedback effects between
conditions in the historical context of region A.  The strategy to include
temporal-​lagged variables, or process-​tracing models for the purpose of policy
analysis (Beach and Pedersen, 2012), therefore helps to clarify the causal order
between variables in studies made according to development design. Similar
challenges, but also opportunities, face students who apply historical approaches
to regional themes, but we will start by clarifying the logic of controlled dia-
chronic comparisons with the help of Møller (2017; Table 6.2).
Møller’s (2017) concern is the causality involved with formation of rep-
resentative, political institutions (notably the cortes) from the 12th century
onwards in Aragon, one of Europe’s many principalities which were grad-
ually emerging as sovereign states. The study stresses the role of discontinuity
and specific events to change; it is, indeed, an analysis of a period in his-
tory during which distinguishable differences between institutional levels in
society started to appear between local and (however primordial) state-​levels
of administration. Consequently, Møller starts by specifying possible sources
of spatial interdependence, pertaining to the cases within the unit (Aragon),
before considering temporal interdependence (labelled “historical diffusion”
in the study). In so doing, he also devises a solution to the problem posed
by the lack of extensive historical data. In the diachronic comparisons of
Aragonese representative institutions, (scarce) historical narratives are sys-
tematically converted into within-​case binary observations (2017, p.  2339).
Binary categories are, according to Møller, easier to detect in the sources,
compared to detailed information about institutions. Note therefore, that his
approach is different to the manner in which historians apply source criticism
as a tool to sift a limited number of sources in order to extract as much valid
information as possible.
Yet, despite such disciplinary differences with respect to managing data, the
explanatory logic on which development design studies draw, and the logic
underpinning conventional historical case studies, remains the same. Provided
that the data collected from the sources can confirm that similar outcomes
in Aragon repeatedly relate to similar conditions in Aragon across time, these
conditions and the observed outcomes are in all probability causally related. In
analogy to Møller’s (2017) study, historians can therefore exploit the advantages
 109

Comparative research designs 109

provided by detailed, in-​depth analysis to examine generic processes (cf. Joyner,
1999). Consequently, the above-​mentioned types of research design can poten-
tially help explain developmental outcomes within single units.
In addition to shedding some light on the “root causes” of certain, generic
phenomena, such as representative, political institutions, development design
studies can however also highlight the historical specificity of particular cases.
Loughlin’s (2007; Table 6.2) study of Northern Ireland is a case in point. Since
it is the relation between ethno-​nationalism and policy-​making, c. 1880–​1920,
that is the focus of the study, Loughlin involves processes that are typical to
cross-​border regions such as Northern Ireland, i.e. how identities are forced
upon regions by their ‘host states’ (Wolff, 2003). Hence, the study, among
other things, draws attention to travelogues as a means to construct a sense of
Britishness in Northern Ireland, and, thereby, the creation of what Loughlin
labels a “social and geographical fact” (Loughlin, 2007, p. 162).
At the same time, though, the above examples also point to the weaknesses
of diachronic comparisons within the same unit. Although of theoretical and
empirical value, studies made according to development design are less suitable
for testing explanations of regularities and causality. Rather, for such purposes,
proponents of quantitative research favour spatial-​temporal lag models designed
to capture both spatial as well as temporal interdependence in synchronic and
diachronic comparisons (Darmofal, 2015). Spatial-​temporal lag models require
extensive access to historical data, and suffer from limited options for analysing
causality between variables, but still allow considering interdependence in terms
of an analytical opportunity. Comparative sequence design is a case in point.

Comparative sequence design and temporal-​spatial interdependence

Falleti and Mahoney (2015) provide the best illustration to the logic underpin-
ning CSD.11 Being process-​oriented, the purpose of this type of research design
is comparisons of “events” ordered into historical “sequences”. Events, then,
are the basic unit of analysis in CSD. They are contextual in temporal and spa-
tial terms, i.e. events occur in specific historical settings that infuse them with
meaning and –​depending on the sequence in which they occur, events relative
to each other –​causally decide their outcomes. Nevertheless, the analyst can
still decide whether different events share enough commonalities to allow for
comparisons. Following Falleti and Mahoney (2015, pp. 212–​213), events are
theoretically distinct from “occurrences”, since the latter are singular and, thus,
defy the logic of comparative analysis.
By the same logic, a sequence is a string of events that relate to a single,
coherent mode of activity across time, such as democratization, decentraliza-
tion, or mobilization. Also, sequences can be part of bigger, and more complex
processes. Not surprisingly, Falleti and Mahoney (2015, p. 211) refer back to
Barrington Moore’s seminal (1966) study of democratization and dictatorship
as a precursor, and in which study Moore compared sequences (of events) and
societal processes in order to formulate conclusions about the determinants for
110

110  Martin Åberg and Thomas Denk

political regimes (cf. also the analogy to Anderson, 1974, on the lineages of the
Absolutist state).
First, according to the logic of CSD, e.g. process-​tracing models are applied
to order (inductively or deductively) the events temporally, and into at least two
separate sequences for the purpose of diachronic and synchronic comparisons.
The aim of the diachronic comparisons in CSD is to analyze causal links
between different events, i.e. whether a temporal and/​or causal order between
the events can be determined. An important feature of this step is also to deter-
mine the extent to which the “speed” or duration of a specific event proves to
be a causal condition for the observed outcome. This opens up the possibility
for testing hypotheses about the causality of processes, and for the formulation
of theoretical statements about case-​specific modes of causality.
Second, CSD at the same time includes synchronic comparisons of sequences,
according to the logic of cross-​sectional comparisons previously described here
(i.e. the methods of agreement and difference, but also qualitative comparative
analysis). In this step of the analysis, the research design allows researchers to
identify which types of sequences are sufficient or even necessary, to cause the
observed outcome. That is, the inclusion of controlled, synchronic comparisons
with the research design means that CSD meets the demands for observing the
degree of regularity involved with causality among several spatial units, such as
regions.
CSD therefore invites analysts to explore interdependence between the (spa-
tial) units included in the analysis. An illustration to this is the opportunity to
determine whether events in region A, such as the launching of a new devel-
opment policy, connect to similar events in region B, or if events are part of
independent processes in region A and region B. Furthermore, the diachronic
dimension of CSD also allows inclusion of dynamic models focusing on
explaining developmental outcomes and analysis of temporal interdependence.
In technically more advanced models applicable to the research design, it is, as
indicated, possible to identify the temporal order of events, and to analyse the
temporal order of causality with respect to the dependent variable (outcome).
This includes control for reverse effects, e.g. in terms of feedback-​effects and
time lags. Among others, the idea that critical junctures create path dependencies
over time –​and, hence, that historical “legacies” exercise effects on decision-​
making with respect to policy and governance (cf. Capocci and Ziblatt, 2010;
Møller, 2013) –​can be systematically tested in studies made according to CSD.
So far, however, actual examples of small-​N regional-​level studies, in which
the analysis of both spatial, and temporal, as well as temporal-​spatial inter-
dependence is systematically included, are scarce. Yet, Basta’s (2018; Table 6.2)
study of territorial reform in Canada and Spain during two time periods is
one possible illustration. Basta’s focus lies on the reinforcing effects exercised
by sequences of accommodation and reactiveness respectively on political sta-
bility (in terms of demands for regional independence). Basta’s aim is to test the
hypothesis that territorial reform, through policy substance, produces effects
on political stability. Canadian and Spanish sequences of accommodation and
 111

Comparative research designs 111

reaction are first identified (two sequences in each country, but under different
periods). Basta then compares the sequences in two pair-​[Link] con-
clusion is that a combination of decentralization and symbolic recognition of
minority nationhood, rather than political and economic divisions between
governmental levels, better explains political stability (Basta, 2018, pp. 66–​67).
Finally, Evenhuis’ (2016; Table  6.2) multi-​scalar analysis of economic change,
policy and government in two industrial regions, South Saarland and Teesside,
c. 1960–​1980, should be mentioned as an illustration to the logic of compara-
tive sequence design. Although described by the author as a “comparative case
study”, the methodology, in fact, draws on a combination of cross-​sectional
design and path-​dependency analysis for the study of resilience and adaptation
across time in the two regions (Evenhuis, 2016, pp. 65, 80–​85).
Unlike cross-​sectional design and development design, CSD provides guiding
principles for how to analyse combinations of spatial and temporal interdepend-
ence (temporal-​spatial interdependence, Table 6.1). For instance, provided that
a particular policy initiative (X)  was implemented in region A, in 2000, and
region B chose to introduce the same policy 15 years later –​does this indicate
that policy X has, indeed, travelled across time, from region A to region B? CSD
opens up for an exploration of precisely how events of this kind in one unit
affect later events in yet other units, for example, by force of diffusion.

Conclusions
We have demonstrated the methodological challenges, but also the opportun-
ities involved with the analysis of interdependence in comparative research.
Different types of comparative research designs, however, offer different options
for specifying and analyzing spatial, temporal and temporal-​spatial interdepend-
ence (summarized in Table  6.3). We conclude that above all, comparative
sequence design poses an important methodological option, ripe for further
exploration within the framework of small-​N comparative research on regions.
CSD invites the researcher to combine controlled synchronic and dia-
chronic comparisons. In relation to other types of research design, it offers the

Table 6.3 Spatial and temporal dimensions of interdependence

Type of research design Spatial Temporal Dimension of interdependence

interdependence interdependence

Cross-​sectional + -​ Spatial interdependence

design
Development design -​ + Temporal interdependence
Comparative + + Spatial independence
sequence design Temporal interdependence
Temporal-​spatial
interdependence
112

112  Martin Åberg and Thomas Denk

most extensive toolbox for specifying and analysing interdependence and its
role in causal explanations: CSD allows for both (i) cross-​unit and cross-​case
comparisons as well as (ii) within-​unit analyses. It (iii) allows the researcher to
use dynamic (process-​oriented) models with a focus on explaining develop-
mental outcomes among sub-​national level units. Thus, allowing analysis of the
temporal and spatial order of causality, including a control for reverse effects,
and time and spatial lags, CSD (iv) enhances the opportunities to analyse spatial,
temporal and temporal-​spatial interdependence among sub-​national level units.
Regional themes that can profit from this historically-​oriented and methodo-
logically versatile research design include policy-​making, policy diffusion, and
policy implementation as dynamic, historically contingent processes.

Notes
1 Originally, the challenges involved with interdependence were addressed by Francis
Galton (1889) in a comment on a study by Edward Taylor (1889) on social com-
plexity, marriage and descent (hence the reason why interdependence is also known
as “Galton’s problem”). Galton questioned Taylor’s conclusion about a strong correl-
ation between degree of social complexity and the degree of modernity of marriage
institutions, simply by pointing out that these institutions often diffuse between soci-
eties. It should also be noted that diffusion-​processes, in addition to learning and
emulation, can also involve coercion, competition, migration and also commonality
of cultural norms and symbolic imitation, only to mention some of the most com-
monly researched mechani[Link] and Gilardi (2006); Franzese and Hays (2008);
Simmons, Dobbins and Garrett (2006).
2 Sample size is an important parameter for estimating the strength of spatial autocor-
relation, but this is not necessarily always a problem in small-​N comparative studies,
since small-​N studies can involve large samples of empirical observations.
3 This happens when attributes cluster closely together diachronically. Let us consider
democratization of political regimes, and a situation in which the degree of democ-
racy in one year correlates strongly with the degree of democracy in the previous
year. In such cases, the duration of conditions favouring democratization must also be
included in the analysis in order to decide whether we are dealing with temporal
autocorrelation, or temporal interdependence, see Coppedge (2015).
4 For the purpose of this chapter, we find that commenting any further e.g. on the
“method of concomitant variation” is superfluous.
5 We have added probability to McCullagh’s definition, since quantitative and quali-
tative approaches in social sciences and historical research alike always, to some
extent, involves interpretation and, because of this, a certain amount of (theor-
etically informed) probability-​reasoning. Note also that we refer to societal et al.
regularities, similarly to McCullagh’s reasoning about explanations of structural
phenomena.
6 Note that Table 6.1 does not include effects appearing as contemporaneous effects
of factors within the unit on the observed outcome, since this effect includes neither
spatial nor temporal dimensions. This type of effect involves the relation between
factors in the same time-​period and unit (Xt1,s1=> Yt1,s1), and it causes no inter-
dependence between units. Spatial autocorrelation, as mentioned above in the text, is
one illustration to this type of effect.
 113

Comparative research designs 113

7 It should however be noted that studies made according to case study design nor-
mally include references e.g. to the nation-​state “context” of a specific region, or offer
“parallels”, or “analogies” or anything such to yet other regions; “context” indicates
a spatial dimension but, importantly, not controlled (synchronic or diachronic)
comparisons. Similarly, if we consider studies made according to development design,
we can note that these are fundamentally similar to historical case studies. However,
whereas development design includes the potential for specifying and analysing tem-
poral interdependence, it lacks options for controlling spatial interdependence.
8 ILLs are local political parties that organize and compete for votes locally, and are
independent from the national party system.
9 A spatial-​lagged variable refers to a condition or outcome in another unit than the
unit of the dependent variable.
10 Temporal-​lagged variables are included in order to define the temporal location of
the compared cases. This strategy challenges the notion of cross-​sectional design,
since this design focuses exclusively on synchronic comparisons. Note also that
comparative analyses use the time dimension as a response to the classic problem of
“too-​many-​variables-​too-​few-​cases”; see Lijphart (1971). That is, in order to avoid
overloading the analyses with too many variables in relation to the number of cases,
researchers measure the same variables and cases at different points in time, which
provides the analysis with more observations.
11 Falleti and Mahoney use the term “comparative sequential method” to describe this
type of research design. We prefer “comparative sequence design”, since the former
term signals that there is, indeed, a single method for analysing temporal-​spatial
interdependence.

References
Alapuro, R., 2012. Revisiting microhistory from the perspective of comparisons. In: S.
Fellman and M. Rahikainen, ed. 2012. Historical knowledge. In quest of theory, method
and evidence. Newcastle upon Tyne: Cambridge Scholars Publishing. Pp. 133–​154.
Almond, G. and Verba, S., 1963. The civic culture: Political attitudes and democracy in five
nations. Princeton: Princeton University Press.
Anderson, P., 1974. Lineages of the absolutist state. London: NLB.
Askarov, Z. and Doucouliagos, H., 2015. Spatial aid spillovers during transition.
European Journal of Political Economy, 40(1), pp. 79–​ 95. [Link]
[Link].2015.10.005
Balland, P-​A., Boschma, R. and Frenken, K., 2015. Proximity and innovation:  From
statistics to dynamics. Regional Studies, 49(6), pp. 907–​920. [Link]
0034340052000320887
Basta, K., [Link] state between minority and majority nationalism: Decentralization,
symbolic recognition, and secessionist crisis in Spain and Canada. Publius: The Journal
of Federalism, 48(1), pp. 51–​75.
Beach, D. and Pedersen, R. B., 2012. Process-​tracing methods: Foundations and guidelines.
Ann Arbor: University of Michigan Press.
Bendix, R., 1964. Nation-​ building and citizenship. Studies of our changing social order.
New York: Wiley.
Bloch, M., 1952 [1928]. Toward a comparative history of European societies. In: F. C.
Lane and J. C. Riemersma, eds. Enterprise and secular change. London: Allen & Unwin.
Pp. 494–​521.
114

114  Martin Åberg and Thomas Denk

Boschma, R., Maroccu, E. and Paci, R., 2016. Symmetric and asymmetric effects of
proximities. The case of M&A deals in Italy. Journal of Economic Geography, 16(2), pp.
505–​535. [Link]
Boschma, R., Martin,V. and Minondo, A., 2017. Neighbour regions as the source of new
industries. Papers in Regional Science, 96(2), pp. 227–​245. [Link]
pirs.12215
Braun, D. and Gilardi, F., 2006. Taking “Galton’s problem” seriously: Towards a theory
of policy diffusion. Journal of Theoretical Politics, 18(3), pp. 298–​322. [Link]
10.1177/​0951629806064351
Caramani, D., (2008). Introduction to the comparative method with Boolean algebra. Quantitative
applications in the social sciences, no. 158. London: Sage.
Capocci, G. and Ziblatt, D., 2010. The historical turn in democratization studies:  A
new research agenda for Europe and beyond. Comparative Political Studies, 43(8/​9),
pp. 931–​968.
Christian, S., 2016. What is global history? Princeton: Princeton University Press.
Cipolla, C. M., 1972. The diffusion of innovations in early modern Europe. Comparative
Studies in Society and History, 14(1), pp. 46–​52.
Coppedge, M., 2015. Democratization and Research Methods. Cambridge:  Cambridge
University Press.
Darmofal, D., 2015. Spatial analysis for the social sciences. Cambridge:  Cambridge
University Press.
Doloreux, D., 2004. Regional innovation systems in Canada: A comparative study. Regional
Studies, 38(5), pp. 479–​492. [Link]
Elkins, Z. and Simmons, B. A., 2005. On waves, clusters, and diffusion: A conceptual
framework. ANNALS of the American Academy of Political and Social Science, (598),
pp.  33–​51.
Esping-​Andersen, G., 1990. The three worlds of welfare capitalism. Princeton:  Princeton
University Press.
Evenhuis, E., 2016. The political economy of adaptation and resilience in old industrial
regions: A comparative study of South Saarland and Teesside. Newcastle University: PhD
thesis, Centre for Urban and Regional Development Studies, School of Geography,
Politics and Sociology. Faculty of Humanities and Social Sciences.
Falleti, T. G. and Mahoney, J., 2015. The comparative sequential method. In: J. Mahoney
and K. Thelen, ed. 2015. Advances in comparative-​historical analysis (strategies for social
inquiry). Cambridge: Cambridge University Press. Pp. 211–​239.
Franzese, R. J. and Hays, J. C., 2008. Interdependence in comparative politics: Substance,
theory, empirics, substance. Comparative Political Studies, 41(4/​5), pp. 742–​780.
Galton, F., 1889. Discussion of Dr. Tylor’s memoir. Journal of the Royal Anthropological
Institute of Great Britain and Ireland, (18), pp. 270–​272.
Ihalainen, P., 2017. The springs of democracy:  National and transnational debates on con-
stitutional reform in the British, German, Swedish and Finnish parliaments, 1917–​1919.
Helsinki: Studia Fennica Historica.
Jahn, D., 2006. Globalization as ‘Galton’s problem’:  The missing link in the analysis
of diffusion pattern in welfare state development. International Organization, 60(2),
pp. 401–​431.
Joyner, C. W., 1999. Shared traditions: Southern history and folk culture. Urbana: University
of Illinois Press.
Kocka, J., [Link] historical comparison: The case of the German Sonderweg.
History and Theory, 1(38), pp. 40–​50. [Link]
 115

Comparative research designs 115

Kubicek, P., 2000. Regional polarisation in Ukraine: Public opinion, voting, and legis-
lative behaviour. Europe-​Asia Studies, 52(2), pp. 273–​294.
Landman, T. and Carvalho, E., 2016. Issues and methods in comparative politics. An introduc-
tion. New York: Routledge.
Laven, D. and Baycroft, T., 2008. Border regions and identity. European Review of History,
15(3), pp. 255–​275. [Link]
Lijphart, A., 1971. Comparative politics and the comparative method. American Political
Science Review, 65(3), pp. 682–​693.
Lijphart, A., 1999. Patterns of democracy: Government forms & performance in thirty-​six coun-
tries. New Haven: Yale University Press.
Loughlin, J., 2007. Creating a “social and geographical fact”. Regional identity and the
Ulster question 1880–​1920s. Past & Present, 195(1), pp. 159–​196. [Link]
10.1093/​pastj/​gtm002
Martin, R., Sunley, P. and Tyler, P., 2015. Local growth evolutions: Recession, resilience
and recovery. Cambridge Journal of Regions, Economy and Society, (8), pp. 141–​148.
McCullagh, C. B., 2004. The logic of history. Putting postmodernism in perspective.
London: Routledge.
Mill, J. S., 1875. A system of logic: Ratiocinative and inductive: Being a connected view of the
principles of evidence and the methods of scientific investigation, Vol 1 (9th ed). London:
Longmans, Green, Reader & Dyer.
Møller, J., 2013. When one might not see the wood for the trees:  The “historical
turn” in democratization studies, critical junctures, and cross-​ case comparisons.
Democratization, (20)4, pp. 693–​715.
Møller, J., 2017. A framework for congruence analysis in comparative historical analysis
of political change. Quality and Quantity, 51(5), pp. 2337–​2355.
Moore, Jr, B., 1966. Social origins of dictatorship and democracy. Lord and peasant in the making
of the modern world. Boston: Beacon.
Naroll, R., 1965. Galton’s problem: The logic of cross-​cultural analysis. Social Research,
32(4), pp. 428–​451.
O’Loughlin, J., Ward, M. D., Lafdahl, C. L., Cohen, J. S., Brown, D. S., Gleditsch, K. S.
and Shin, M., 1998. The diffusion of democracy, 1946–​1994. Annals of the Association
of American Geographers, 88(4), pp. 545–​574.
Overton, M., 1985. The diffusion of agricultural innovations in early modern
England: Turnips and clover in Norfolk and Suffolk, 1580–​1740. Transactions of the
Institute of British Geographers, 10(2), pp. 205–​221.
Peters, G., 2014. Strategies for comparative research in political science. Theory and methods.
Basingstoke: Palgrave.
Przeworski, A. and Teune, H., 1970. The logic of comparative social inquiry. Malabar: Krieger.
Putnam, R. D., 1993. Making democracy work. Civic traditions in modern Italy. Princeton,
NJ: Princeton University Press.
Reiser, M., Rademacher, C. and Jaeck, T., 2008. Präsenz und erfolg kommunaler
wählergemeinschaften im bundesländervergleich. In:  A. Vetter, ed. 2008.
Erfolgsbedingungen lokaler bürgerbeteiligung. Wiesbaden: VS Verlag. Pp. 123–​147.
Rogers, E. M., 1995. Diffusion of innovations. New York: Free Press.
Rokkan, S., 1966. Comparative cross-​national research: The context of current efforts.
In: R. L. Merrit and S. Rokkan, ed. 1966. Comparing nations: The use of quantitative
data in cross-​national research. New Haven: Yale University Press. Pp. 3–​25.
Sewell, W. H. Jr., 1967. Marc Bloch and the logic of comparative history. History and
Theory, 6(2), pp. 208–​218.
116

116  Martin Åberg and Thomas Denk

Shipan, C. R. and Volden, C., 2008. The mechanisms of policy diffusion. American
Journal of Political Science, (52)4, pp. 840–​ 857. [Link]
j.1540-​5907.2008.00346.x
Simmons, B. A. and Elkins, Z., 2004. The globalization of liberalization: Policy diffusion
in the international political economy. American Political Science Review, 98(1),
pp. 171–​189.
Simmons, B., Dobbins, F. and Garrett, G., 2006. Introduction: The international diffusion
of liberalism. International Organization, 60(4), pp. 781–​810.
Skocpol, T., 1984. Emerging agendas and recurrent strategies in historical sociology.
In:  T. Skocpol, ed. Vision and method in historical sociology. Cambridge:  Cambridge
University Press, pp. 356–​291.
Snyder, R., 2001. Scaling down:  The subnational comparative method. Studies in
Comparative International Development, 36(1), pp. 93–​110.
Strang, D., 1991. Adding social structure to diffusion models: An event history frame-
work. Sociological Methods and Research, 19(3), pp. 324–​353.
Taylor, E. B., 1889. On a method of investigating the development of institutions: Applied
to laws of marriage and descent. Journal of the Royal Anthropological Institute of Great
Britain and Ireland, (18), pp. 235–​270.
Ther, P., 2009. Comparisons, cultural transfers, and the study of [Link] a trans-
national history of Europe. In: H.-​G. Haupt and J. Kocka, eds. Comparative and trans-
national history. Central European approaches and new perspectives. New York: Berghahn.
Pp. 204–​225.
Volden, C., Ting, M. M. and Carpenter, D. P., 2008. A formal model of learning and
policy diffusion. American Political Science Review, 102(3), pp. 319–​332.
Wellhofer, S. E., [Link] comparative method and the study of development, diffusion,
and social change. Comparative Political Studies, 22(3), pp. 315–​342.
Werner, M. and Zimmermann, B., 2006. Beyond comparison. “Histoire croisée” and the
challenge of reflexivity. History and Theory, 45(1), pp. 30–​50.
Weyland, K., 2006. Bounded rationality and policy diffusion. Social sector reform in Latin
America. Princeton: Princeton University Press.
Wolff, S., 2003. Disputed territories: The transnational dynamics of ethnic conflict settlement.
New York: Berghahn.
 117

Part III

Empirical case studies

118
 119

7 
Catching the ladder
The formation and growth of the São
Paulo automotive industry cluster1
Tomàs Fernández-​de-​Sevilla and
Armando J. dalla Costa

Introduction
A direct connection between industrialization and economic growth has largely
been established (Gerschenkron, 1962; Landes, 1969; Pollard, 1981; Allen, 2010).
The automobile industry became a major driver of economic growth in late
industrialized countries from the mid-​20th century (Amsden, 1989; Jenkins,
1995). Likewise, regional concentration of industry has also been indicated as
a source of competitive advantage in both developed and developing countries
(Porter, 1990: Schmitz, 1999). However, in order to be able to overcome the
difficulties to develop industrial clusters in an endogenous manner, the per-
ipheral economies tend to have to resort to direct foreign investment, often in
the form of the installation of subsidiaries of multinational companies (MNCs)
(Amin and Thrift, 1994;Young, Hood and Peters, 1994). Nevertheless, it is not
surprising that the result is an independent evolutionary process through which
the development of the cluster in turn determines the strategies of the MNCs
(Enright, 2000).
The aim of this research is to analyze the emergence and growth of an auto-
mobile industry cluster in São Paulo (Brazil) from its origins at the beginning
of the 20th century until its expansion phase was completed in the mid-​1970s,
when production stabilized in the region at 900,000 vehicles per year for the next
25 years. The São Paulo automobile industry cluster considered in this research
encompasses companies and institutions located in São Paulo city and in the areas
organized around the Anchieta and Dutra motorways. The former connects São
Paulo to the Port of Santos, crossing the so-​called São Paulo ABC (the muni-
cipalities of Santo André, São Bernardo do Campo, and São Caetano do Sul),
whereas the latter connects São Paulo with Rio de Janeiro. The cluster stretches
as far as the municipality of São José dos Campos (85km from São Paulo).
The formation and expansion of the São Paulo cluster was an evolutionary
process which, far from being linear or automatic, arose from the combination
of premeditated decisions, historical situations and trial and error. Historical
analysis therefore has great explanatory potential, making it essential to study
this subject from a long-​term dynamic perspective. Understanding the process
of the emergence and development of the São Paulo industrial automotive
120

120  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

district in a comprehensive manner will lead to an increase in the empirical
evidence on this type of phenomena, which are currently taking place both
in other regions of Brazil and in other countries which are seeking to become
industrialized. This will facilitate the practice of regional policy and planning,
in Brazil and elsewhere.
Greater São Paulo is one of the few tropical or sub-​tropical areas in the
southern hemisphere where an advanced industrial system has developed. The
origins of São Paulo’s industrialization can be traced to the late 19th century,
when the investment of capital accumulated through coffee exportation caused
the first factories to emerge (Dean, 1969; Silva, 1976; Cano, 1981; Suzigan,
1986). Although the take-​off of the automotive industry took place in the mid-​
20th century, its roots go back to the beginning of the century. By using a
historical approach, this study attempts to assess the key determinants in the
formation and growth of the São Paulo auto-​industry cluster.
This research will follow the interpretative models and empirical strategies
defined by Markusen (1996), Catalan (2017), and Carli and Morrison (2018).
Markusen’s (1996) concept of hierarchical cluster, in which few big firms acted
as coordinating centres or hubs of the regional economy, is used. The concep-
tual framework distinguishes between internal and external explanatory factors
and allows for acknowledging that the different phases in the evolution of the
cluster do not occur in a rigid manner but rather follow an adaptive path (Carli
and Morrison, 2018). Finally, the empirical strategy used to analyze the evo-
lution of the São Paulo automotive cluster is defined in Catalan (2017), who
studied the life-​cycle of the Barcelona automotive district exploring the role
played by four possible factors in the explanation of the relative success attained
by that cluster: the presence of external economies; the capacities provided by
hub companies; the adoption of national government strategic industrial pol-
icies; and the emergence of adequate local institutions.

History matters: Cluster institutions, strategic policies and

large enterprises in the formation of the São Paulo auto-​
industry cluster
Clusters emerge and develop due to a process of positive mutual reinforcement
arising from the geographic concentration of industrial activity, their life-​cycles
differing from those of industries (Swann, 2009, c­ hapter 13; Menzel and Fornahl,
2010). Bergman (2008) identifies three main phases in the cluster life-​cycle: the
existence, which includes what must happen before the cluster achieves crit-
ical mass; the expansion, which involves the cluster growth; finally, the exhaus-
tion, which arises when maturity and saturation hinder its viability, breaking its
ability to innovate. In this last phase, which is beyond the scope of this study
and in which congestion costs play a key role (Maskell and Kebir, 2006), the
cluster may follow a divergent path, since it either enters a lock-​in phase or may
experience a renaissance (Enright, 2003; Tödtling and Trippl, 2004).
 121

Catching the ladder 121

However, recent studies maintain that the majority of the work on the life-​
cycle of clusters has focused almost exclusively on their internal dynamics,
neglecting the role of external factors, socio-​ economic contingencies and
historical situations (Martin and Sunley, 2011; Trippl et  al., 2015; Carli and
Morrison, 2018). The conclusion of these studies is that clusters tend to follow
an “adaptive” path rather than evolving in accordance with a predefined pattern,
thus revealing the importance of history. Through a historical analysis which
uses a narrative structure, this study seeks to understand the role played by
various economic factors, such as Marshallian externalities, local institutions,
industrial policy and big companies, in the formation and expansion of the São
Paulo automotive cluster.
The developmental state has been pointed out as a crucial agent of industrial-
ization in developing countries (Johnson, 1982; Amsden, 1989 and 2003; Evans,
1995; Woo-​Cumings, 1999), including Brazil (Schneider, 1991; Shapiro, 1994;
Catalan and Fernández-de-Sevilla, 2018).2 Chang (2002) recovered the protec-
tionist argument for the defence of emerging industries in developing coun-
tries described by Friedrich List –​who was in turn influenced by Alexander
Hamilton –​to argue that all economies in the process of modernization have
employed industrial, trade and technological policies to boost their economic
development. According to him, historical experience shows that the transi-
tion towards higher added-​value activities does not happen naturally, but rather
requires the use of a wide array of strategic policies, including protection of
targeted industries. Their insistence on the need for developing countries to
adopt active industrial policies is shared extensively (Cimoli, Stiglitz and Dosi,
2009; Stiglitz and Lin, 2013; Norman and Stiglitz, 2016).
Nevertheless, industrial policy alone does not explain why the automobile
industry clustered in São Paulo, a region that produced practically all of the
near one million vehicles manufactured in Brazil in 1974. The competitive
advantages associated with the geographic concentration of industry were first
characterized by Alfred Marshall, who argues that Britain was the workshop
of the world in the 19th century due to certain external economies of scale
present in its industrial districts. Together with a specific set of services and
infrastructures –​in Lancashire, for example, the regional railway system –​these
include a pool of skilled workers and specialized suppliers, forming a vigorous
intra-​district market and permitting the (almost) free spread of knowledge and
information (Marshall, 1890, book 4, ­chapter 10).
The Marshallian district was revisited by Italian economic geographers, who
pointed out that the competitive advantage of industrial districts of the so-​
called “Third Italy” came from an institutional environment characterized by
an abundance of small and medium-​sized firms with a high level of flexi-
bility, enabling cooperation between firms and local governments in order to
share risks and stabilize markets (Bagnasco, 1977; Brusco, 1982; Becattini, 1990).
However, it has been argued that these works understated the role played by big
firms in shaping the Italian districts (Harrison, 1994, ­chapter 4).
122

122  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

The Italian perspective has been challenged in Markusen (1996) by presenting
three typologies of industrial districts articulated around large companies (the
hub-​and-​spoke district, the satellite platform and the state-​anchored district),
which would better explain the flourishing and consolidation of regional econ-
omies in the USA, Japan, Korea and Brazil. Hub-​and-​spoke districts, the most
predominant typology, are formed by a few large companies that act as anchors
or hubs to the regional economy. In economic and business history, the greater
efficiency of large companies has been also defended by emphasizing that their
success lies in their accumulated technological and organizational capabilities
(Chandler, 1977; Lazonick, 1990).
By offering access to public or semi-​public goods such as infrastructures,
vocational schools, universities and research centres, the cluster institutional
environment, usually articulated by industrial and trade associations, could also
provide part of the regional competitive advantage (Porter, 2008). According
to Brusco (1982), the interaction between the productive structure, the labour
market, and the local institutions to link salaries, productivity and investment in
Emilia-​Romagna (Italy) explained the regional competitive advantage. From a
macroeconomic perspective, Marglin (1990) and Eichengreen (2007) maintain
that the remarkable growth of Western European economies during the Golden
Age was due to a series of neo-​corporate institutional agreements between
unions, employers’ organizations and governments that explain why the former
accepted the restriction of wage increases in return for the latter channelling the
higher profit into new investment. Conversely, Olson (1982) and Acemoglu and
Robinson (2012) explain that diverse groupings of industrialists or workers can
be motivated to form lobbies to defend their personal interests, often pushing
for the application of protectionist policies that damage economic growth.
Detailed studies have been undertaken in economic and business history on
the structural transformations of central districts of the Industrial Revolution in
the United Kingdom and the United States (Scranton, 1997; Wilson and Popp,
2003; Popp and Wilson, 2007). In the case of a typical industry from the second
Industrial Revolution, such as the automotive sector, studies such as Klepper
(2007, 2010), Donnelly, Begely and Collis (2017), and Pardi (2017) illustrate
the potential of the historical analysis to understand the regional concentration
of this industry, showing the rise and decline of clusters such as that of Detroit
in the United States and the West Midlands in England. Likewise, Catalan,
Miranda and Ramon-​Muñoz (2011a) have presented an analysis that integrates
the evidence presented in almost 20 case studies which offer a historical vision
of the formation and growth of districts and clusters in Southern Europe. They
conclude that, in this region, although cases have occurred of the emergence
of competitive advantage in the framework of the neo-​Marshallian district, the
cluster model has been more common, with leading companies which have
complemented the Marshallian externalities with the organizational capabilities
characteristic of big companies (Catalan, Miranda and Ramon-​Muñoz, 2011b).
Outside the scope of industrial clusters, Molema, Segers and Karel (2016) indi-
cate that, in the case of various agribusiness clusters developed in Europe in
 123

Catching the ladder 123

the 19th and 20th centuries, interactions between governments, entrepreneurs
and associations of producers in the region played a pre-​eminent role in the
expansion through regulatory frameworks along with educational and scientific
policies. On the other hand, disagreement and competition could also promote
the clustering of activity. All of these studies reveal the potential of the histor-
ical analysis to understand an economic phenomenon such as the tendency of
industrial activity to concentrate on a regional level.
The following three sections chronologically study the formation and
expansion stages of the cluster, beginning with its origins, continuing with
the formation of the critical mass essential for the cluster to be recognized as
such, and ending with its growth until it reaches maturity. Finally, a last section
integrates the three phases and the role played by the determinant factors in
each of the stages.

The cluster’s origins, 1900s–​1929

At the same time as the wealthy coffee barons from inland were building their
mansions in the first three decades of the 20th century, 1.57 million immigrants
(82 per cent from abroad, mainly Europe) arrived in the city of São Paulo in
search of employment in trade, construction and in its budding industry. The
number of industrial establishments in the overall state went from 327 in 1907
to 4,112 in 1919 and 9,603 in 1928, with the number of employees in industry
increasing from 25,000 to over 150,000, raising the weight of São Paulo’s industry
in the Brazilian total from 16 per cent to 37 per cent (Negri, 1996, pp. 36–​37).
If in 1901 there were five cars registered in São Paulo, five years later there
were 84 (Forest, 2002, p.  42). The chauffeurs that usually accompanied the
imported vehicle introduced the first knowledge of mechanics to the region.
They were responsible for repairing the vehicles and they taught the profession
to their first Brazilian colleagues (Wolfe, 2010, pp. 21–​22). The origins of the
São Paulo cluster, however, lie in the small repairs and car parts manufacturing
workshops that began to emerge during the 1920s to serve the automotive
aftermarket. During those years, the number of cars registered in São Paulo state
experienced constant growth, reaching 70,000 vehicles in 1929 (Forest, 2002,
p. 31). In 1928, there were 317 establishments dedicated to the manufacture
of transport material, which together employed almost 5,000 people (Negri,
1996, p. 37).
It was under this expansion that Ford, GM and International Harvester
opened their first assembly lines for completely knock-​down (CKD) kits.3
After having opened a small office in the city of São Paulo in May 1919, Ford
rented a large skating rink in Praça da República, right in the city centre, which
became its first assembly line in January 1920. A few months later, the company
decided to build a three-​storey plant (which accounted for 100 employees) in
the Bom Retiro neighbourhood, where it remained for several years (Wilkins
and Hill, 2011, pp.  93–​94). In 1925, Ford’s production reached 14,861 units
(Wilkins and Hill, 2011, p. 146), while sales reached 24,500 cars (Gonçalves,
124

124  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

1989, p. 34). According to Wilkins (1974, p. 171), Ford also tried to assemble
in Recife (1925), Porto Alegre (1926) and Rio de Janeiro (1927), but without
success, maintaining its assembly operations in São Paulo.
For its part, GM inaugurated a factory in the Ipiranga neighbourhood, the
industrial heart of the city, in 1925. According to General Motors (1995, p. 11),
the decision to set up the factory in São Paulo “resulted from an in-​depth feasi-
bility study conducted by the Corporation’s management”. The GM managers
placed special value on the availability of electric power, petrol, raw materials,
accessories and spare parts and components. A year after its inauguration, the
rate of production was 40 cars per day. In 1927, having already assembled 25,000
Chevrolets, GM began the construction of a new factory in São Caetano do
Sul, in the ABC region. The new factory was inaugurated in 1929, with an
installed capacity of 100 units per day. The GM workforce (600 people) tripled
Ford’s workforce. Together with International Harvester, which also installed a
small truck assembly line in 1926, many other brands established networks of
representatives to import complete vehicles.
A set of related associations accompanied the first phases of São Paulo’s
motorization. The first car race in Latin America took place in São Paulo on
26 July 1908, organized by the Automobile Club of São Paulo, which had
been founded 15 days earlier. Approximately 10,000 people gathered in Parque
Antártica to see the start and finish of the race (Gonçalves, 1989, p. 24). After
that, a wide range of car races took place involving São Paulo. Ten years later,
in 1917, the 1st Congress of São Paulo Roadways was held, promoted by the
state government. Some 70 cars circulated with relative ease along the road
that connected São Paulo to the port of Santos, whose infrastructure devel-
opment works, which began in 1913, were at a very advanced stage (Forest,
2002, pp. 40–​42). In fact, in 1921, the four main cities in the state of São Paulo
(Campinas, Jundiaí, Santos and São Paulo) were connected by road, while the
connection with Rio de Janeiro did not appear until 1928. Furthermore, on
13 October 1923, the First Automobile Exhibition was inaugurated, with five
shows being held during the decade (Wolfe, 2010, p. 47). In terms of worker
training, it is worth highlighting that the Ford School of Mechanics and the
GM School of Advanced Mechanics both geared towards training workers for
the assembly lines and for the aftersales service offices (Nascimento, 1976, p. 14).
Beyond the automotive industry, the creation of the Centre of Industries of the
State of São Paulo (CIESP) in March 1928 and the Federation of Industries of
the State of São Paulo (FIESP) in 1931, both founded by Jorge Street, Francisco
Matarazzo and Roberto Simonsen, was also of special importance, as they were
the first associations to call for a containment of the deindustrialization of Brazil
and to defend industry protection.
During this period, neither the government nor the firms called for or tried
to implement industrial policies to foster the industry’s development. Ford and
GM’s main interest was the liberalization of trade, especially with regard to
the importation of CKD kits. If both were assembling in São Paulo, this was
due to their conviction that this was more profitable than importing complete
 125

Catching the ladder 125

vehicles. In fact, all the relations that Ford established with the Brazilian gov-
ernment involved rubber plantations in the Amazon basin, which was its main
business in Brazil until 1945 when it was sold at a substantial loss; none involved
assembly plants (Wilkins, 1974, pp. 102–​122).The manufacturers from the aux-
iliary industry did not push for the application of industrial policies either,
since they were not yet organized and did not have a unified discourse, which
arrived with the deindustrialization that accompanied the impact of the Great
Depression. Indeed, the creation of the CIESP coincided with the begin-
ning of the slowdown in industrial production just before the outbreak of the
Great Depression, while the formation of the FIESP took place at the height
of the same. Moreover, the sectors most affected by the crisis were the chem-
ical industry and metal products, within which the automotive industry falls.
The manufacturing levels of 1929 did not recover until 1933, reaching a min-
imum of 72 per cent in 1931 (Baer, 2014, p. 29). Thus, in this phase, no type of
industrial policy geared toward promoting the automotive industry was applied,
either on a regional or a federal level.
At the dawn of the Great Depression, subsidiaries of the two biggest
multinationals in the sector were operating in São Paulo and had great poten-
tial to become hub firms for the region. Ford’s unsuccessful attempts to establish
more assembly lines in other Brazilian regions could indicate the presence of
external economies, as described by Marshall, in São [Link] hypothesis could
be reinforced by the fact that GM also established its subsidiary in São Paulo,
defined by GM as the main industrial hub in Latin America (General Motors,
1995, p.  11). Together with the Michigan giants, dozens of small workshops
geared towards the spare parts market were starting to shape the cluster. This
involved the formation of potential suppliers and skilled workers. They had
no problem in finding as many workers as they required, not only from the
roughly 150,000 industrial workers in the region, but also from immigration.
Moreover, some associations were emerging and accompanying the district’s
first steps. However, the cluster’s potential remained low, mainly because the
two large firms restricted their operation to the simple assembly of CKD kits
from the United States. At this stage, no industrial policy was implemented, and
the degree of development achieved was not enough to consider the cluster as
existent.

Looking for the critical mass: A trade association behind the

cluster’s first boost, 1930–​1955
The Great Depression placed the Brazilian economy in dire straits. São Paulo’s
recovery consolidated in 1933, boosted by industrial expansion driven mainly
by national capital that erected small and medium-​sized plants with low capital
[Link] state’s share of industrial added-​value for the whole of Brazil went
from 32.2 per cent to 48.9 per cent between 1919 and 1949, a percentage that
reached 72.4 per cent in the case of durable goods (Negri, 1996, p. 80). This
was no longer an industrialization process subordinated to the coffee industry’s
126

126  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

dynamics, but an expansion whose main driver was the city of São Paulo itself.
In this period, Brazil was mainly ruled by Getúlio Vargas, who divided his sym-
pathies between Roosevelt’s New Deal, Mussolini’s Charter of Labour and
German-​inspired dirigisme (Fausto, 2006). Vargas was appointed provisional
president in 1930 by the Military Junta that led the 1930 Revolution, remaining
in power first as constitutional president (1934–​1937), and then as a dictator
(1937–​1945) by imposing a constitution that eradicated political parties, the
right to vote and the right to [Link] recovered the power in 1951 through
a free and secret ballot, ending his presidency by committing suicide in 1954
(Fausto, 2006).
The autoparts industry boosted its development by taking advantage of
the impossibility of importing autoparts during the Second World War. When
the war ended, around a hundred workshops were scattered throughout São
Paulo with the capacity to produce parts such as electrical accumulators, brake
drums, wheel hubs, crown gears, pinions and axle shafts (Dean, 1969, p. 244).
Although the liberalization of the market at the end of the war damaged the
industry, most of Brazil’s autoparts workshops were clustered in São Paulo in
1949 (Table  7.1), hosting 3,219 workers who accounted for 80 per cent of
Brazil’s total autoworkers (Nascimento, 1976, p. 27).
The externalities present in São Paulo can explain why the Veículos e
Máquinas Agrícolas (Vemag) factory was established in the Ipiranga neigh-
bourhood, although neither of its owners came from the region  –​the
Domingos Fernandes family, established in Rio since the 19th century,
where they had interests in casinos and hotels; the Swedish investor Svend
H. Nielsen; and Melvin Brooks, a Studebaker representative (Sandler, 2005,
p. 134). Its factory, which had one of the most important fleets of heavy-​duty
equipment in Latin America, assembled its first truck in 1948. Two years later,
the company began negotiations with Scania on importing CKD truck kits,
with the first units arriving in 1952. One year earlier, the Cia. Americana
Industrial de Omnibus (CAIO), established in 1945, had started the pro-
duction of chassis for its own buses, importing only engines and drive shafts
(Nascimento, 1976, p. 25).

Table 7.1 Autoparts workshops in Brazil: 1949, 1953 and 1957

São Paulo Rio de Janeiro Rest of the States TOTAL

N º % N º % N º % N º %

1949 65 65.0 3 3.0 32 32.0 100 100.0

1953 103 71.0 24 16.6 18 12.4 145 100.0
1957 118 88.1 5 3.7 11 8.2 134 100.0

Source: Authors’ elaboration. For 1949 from Nascimento (1976, p. 27). For 1953 and 1957 the data cor-
responds to firms with displays at the Mostras da Indústria Nacional de Autopeças celebrated in Rio de
Janeiro on 20 January 1953 and 22 July 1957, obtained from Gattas (1981, p. 107 and p. 249).
 127

Catching the ladder 127

However, neither Vemag nor CAIO had the potential to act as anchor for
the regional economy, as defined in Markusen (1996). It was GM and Ford
who should have played this role, but they did not. In 1955, Ford assembled
ten trucks per day and GM only half that, even adding buses and trucks, having
an installed daily capacity of 125 and 200 vehicles, respectively (Nascimento,
1976, p. 53). The Brazilian economy was not attractive enough to persuade the
American giants to transform their facilities into fully-​fledged factories. In fact,
until the mid-​1950s, the main activity of GM’s plant in São Caetano was fridge
production (General Motors, 1995, pp. 40–​45). Under these circumstances, the
autoparts industry was limited to working only for the aftermarket.
The lifting of trade restrictions once the war ended (coinciding with the
ousting of Vargas in 1945) once again filled the domestic market with imported
parts, leaving many of the district’s small companies in a critical situation.
Therefore, the autoparts manufacturers took advantage of Brazil’s serious foreign
payment crisis of 1947, which depleted practically all of the currency reserves
accumulated during the war, to promote the protection of the sector. After sev-
eral meetings in the headquarters of the FIESP, the Professional Association of
the Autoparts Industry of São Paulo (Associaçao, transformed into Sindipeças
since September 19534) was founded in October 1951 and involved the partici-
pation of 122 companies (Gattas, 1981, pp. 57–​61). This association would turn
out to be a key for the future of the cluster.
The Associaçao intensively demanded a protectionist programme inspired by
the agenda formulated by Roberto Simonsen,5 the direction of which recalls
the nascent industry protection argument formulated by Friedrich List in 1841
and the design of which presents numerous points in common with the pol-
icies which, at the time, were beginning to be applied in developing coun-
tries, such as South Korea (Chang, 1993; Catalan, 2010) and Spain (Catalan
and Fernández-​de-​Sevilla, 2013; Fernández-​de-​Sevilla, 2014). Its demands were
considered favourable by Vargas’s new government, which had been established
after a democratic election on 31 January 1951. The government created the
Sub-​Commission on Jeeps, Tractors, Trucks and Cars, which reported to the
Commission on Industrial Development on 20 March 1952 that it would study
the viability of manufacturing cars in Brazil. One month later, the Associaçao
delivered a report in which it maintained the existence of 250 companies cap-
able of manufacturing 162 groups of parts and accessories. The rewards were
reaped in August 1952 when “Aviso” No. 288 was published, vetoing the
importation of the items on a list of 104 groups of parts and accessories already
manufactured in the country. Because of this measure, many companies from
the metal-​mechanical branch began to produce parts that appeared on the list
(Gattas, 1981, pp. 77–​88).
In order to strengthen its position, the Associaçao promoted the holding
in Rio de Janeiro, then the capital of the Federation, of the First Show of the
Brazilian Autoparts Industry, which took place in Santos Dumont Airport on
20 January 1953. The geographical distribution of the almost 145 companies
involved reveals that the industry was already concentrated in São Paulo (see
128

128  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

Table 7.1). A little later, in April 1953, “Aviso” No 311 prohibited the import-
ation of pre-​assembled motor vehicles, authorizing only the entry of CKD kits
if they came without the parts specified in Decree 288. As a result,Volkswagen
(VW) and Willys-​Overland began to assemble CKD kits in workshops located
around São Paulo, while Mercedes-​Benz assembled its first units in Rio de
Janeiro. However, as with Ford and GM, these initiatives were far from being
fully-​fledged manufacturing facilities.
The industrial development achieved by São Paulo in this period likely
indicates the presence of external economies in the region, mainly the abun-
dance of skilled workers as well as the presence of related [Link] intensi-
fication of the car parts industry clustering around São Paulo (see Table 7.1) and
the establishment of Vemag also point in this direction. At the end of the period,
well over 100 firms were located in the cluster, mostly in the Ipiranga neigh-
bourhood and in the ABC. A pool of skilled workers and specialized suppliers
was present in the district, where knowledge of the art moved easily from
workshop to workshop. The lack of large carmakers with great demand-​pull
capacity was the main constraint to the cluster’s expansion. Neither GM nor
Ford had any desire to increase the range of their operations. To overcome
this constraint, the autoparts companies created their own association to boost
the sector through protectionist policies. Their demands found prompt echo in
the new Vargas administration. Before the end of the year, Willys and VW had
established assembly lines in São Paulo, while Mercedes-​Benz installed its line
in Rio de Janeiro. Soon afterwards, in June 1954, the Vargas government created
the Executive Commission of the Autoparts Industry to boost the industry’s
development (Latini, 2007, pp. 84–​97). However, after Vargas committed sui-
cide in August 1954, the new liberal government of João Café Filho placed it
on standby (Nascimento, 1976, pp. 45–​49). The definitive boost for the cluster
came from Juscelino Kubitschek (JK), appointed president of Brazil in 1956.

Growth: Industrial policy and hub firms, 1956–​1974

Kubitschek did not hesitate to use the force of the state to boost the industrial-
ization and modernization of Brazil (Sikkink, 1991; Shapiro, 1994).Automobiles
were included in JK’s flagship programme, known as the “Plano de Metas” or
“Goals plan”, which turned their manufacture into a priority.6 In June 1956,
the Executive Group of Automobile Industry (GEIA in its Portuguese initials)
was created to accomplish the requirements of the Goals plan regarding vehicle
manufacturing (Gattas, 1981; Shapiro, 1994). Its functions included setting the
production targets (80,000 trucks and buses, 50,000 jeeps and light commer-
cial vehicles and 40,000 passenger cars) and the percentages of local content
(90 per cent for trucks and commercial vehicles and 95 per cent for cars and
jeeps in 1960), as well as handling the acceptance and monitoring of individual
investment projects. Protection was the essential tool to force the installation
of big brands in the area, brands that would also gain some advantages in the
 129

Catching the ladder 129

exchange rate and some access to public credit. Beyond the fact that it was prac-
tically impossible to import automobiles, the central element which explains
the growth of the cluster was the requirement to produce them with nationally
manufactured parts and components.
The GEIA guidelines encouraged the development of the subsidiaries
established in 1953 by VW and Willys, as well as causing the move of Mercedes
to São Paulo, these being the companies that led the cluster’s growth. The first
project that VW presented to the GEIA was the manufacture of the Kombi van.
Although, in the end, pressure from the GEIA led the Germans to also include
the Beetle, their people’s car –​popularized in Brazil as the Fusca (Shapiro, 1991,
pp. 911–​918). Because of this pressure,VW built a new factory in São Bernardo
do Campo, strongly inspired by its flagship plant in Wolfsburg (Fleury and
Salerno, 1998, p. 281). At the same time,Willys erected a modern industrial com-
plex in São José dos Campos with assembly lines, bodywork stamping and engine
manufacturing (the engine block was cast in the neighbouring municipality of
Taubaté), adding a research and development centre in 1962. It is important to
note that, in order to accomplish its new and more ambitious goals, Mercedes-​
Benz moved its factory from Rua Bela in the city of Rio de Janeiro to São
Bernardo do Campo. In September 1956, Mercedes inaugurated a new factory
located at kilometre 15/​16 of the Anchieta motorway, in the heart of the ABC
region, thus further reinforcing the presence of external economies in the region.
The GEIA’s requirements also forced Ford and GM, which were much
more reluctant about manufacturing in Brazil (Wilkins and Hill, 2011, p. 414;
Shapiro, 1991, pp. 918–​933), to develop their facilities and become mass produ-
cers. GM modified its project to build only a truck assembly line in São José dos
Campos by establishing a new comprehensive Chevrolet engine factory with
iron casting, forging and engine block manufacturing, which was inaugurated
in March 1959 (Gattas, 1981, pp. 223–​224). Ford, which had no intention of
manufacturing using entirely local content in its new facilities inaugurated in
the Ipiranga neighbourhood in 1953, had to commit to erecting a fully-​fledged
engine factory next to the factory they already had (Wilkins and Hill, 2011,
p. 416). Ford had only agreed to increase the domestic content of their trucks
to 30 per cent, since they considered the Brazilian market too small to do so in
greater depth. However, to obtain the GEIA rubberstamp, Ford built an engine
and stamping factory in São Paulo and a foundry in Osasco, a municipality
located a few kilometres from the factory (Shapiro, 1994, p.  76). Because of
the GEIA’s requirements, Ford and GM transformed their long-​time Brazilian
assembly operations into a real manufacturing activity.
JK’s administration considered foreign direct investment to be the only way
to promote the transfer of technology and capital that the industry needed
(Nascimento, 1976, pp. 94–​97; Shapiro, 1991, pp. 876–​947). This explains why
the GEIA did not promote the creation of one or two public companies to
act as a ‘ “national champion”,7 a strategy which was applied in countries such
as South Korea, Spain and Argentina (Catalan, 2010). The GEIA approved 18
car manufacturing projects, of which 11 were implemented, and 154 autoparts
130

Table 7.2 Projects approved by the GEIA which achieved implementation (1957)

Company Established Factories Firm ownership 1962 Types of vehicles

Fábrica 1942 Rio de Janeiro Brazilian (95%) Heavy trucks

Nacional Acquired by Alfa Passenger cars
de Motores Romeo in 1967
Ford Motor 1920 São Paulo city Ford (100%) Light
do Brasil commercial
vehicles
Medium trucks
General 1925 São Caetano do GM (100%) Light
Motors do Sul (ABC) commercial
Brasil São José dos vehicles
Campos (SP) Medium trucks
International 1926 Santo André Int. Medium trucks
Harvester (ABC) Harvester (100%)
Acquired by Chrysler
in 1966
Mercedes-​ 1953 São Bernardo do Brazilian (50%) and Medium trucks
Benz do Campo (ABC) Mercedes (50%) Heavy trucks
Brasil Buses
Scania-​Vabis 1957 São Paulo city Scania (major) and Heavy trucks
do Brasil Vemag (minor) Buses
Simca do 1958 São Bernardo do Brazilian (80%) and Passenger cars
Brasil Campo (ABC) Simca (20%)
Acquired by Chrysler
in 1966
Toyota do 1958 São Paulo city Toyota (100%) Passenger cars
Brasil Light
commercial
vehicles
Vemag 1945 São Paulo city Brazilian (82%) and Passenger cars
DKW (18%) Light
Acquired by commercial
Volkswagen in vehicles
1965 and closed in
1967
Volkswagen 1953 São Bernardo do Brazilian (20%) Passenger cars
do Brasil Campo (ABC) and Volkswagen Light
(80%) commercial
vehicles
Willys 1952 São Bernardo do Brazilian (55%) and Passenger cars
Overland Campo (ABC) Kaiser (45%) Light
do Brasil Taubaté (SP) Acquired by Ford in commercial
1967 vehicles

Source: Authors’ elaboration from ANFAVEA (1961) and Gattas (1981, p. 380).

 131

Catching the ladder 131

manufacturing projects (ANFAVEA, 1962a, p. 27). As Table 7.2 shows, all of
the projects regarding auto manufacture located their facilities in São Paulo,
except for the Fábrica Nacional de Motores (FNM), which played a residual
role in the development of the [Link] FNM was a state-​owned company
established in 1942 to make aeroplane engines for the allied aviation that started
to assemble trucks using Issota licences in 1948 and Alfa Romeo licences from
1951, with a low level of production. In 1955, the FNM assembled 2,426 of the
13,950 trucks assembled in São Paulo (Nascimento, 1976, p. 50).The industrial
base of the region, with hundreds of workshops and the presence of potential
supplier industries, was decisive in the choice of location. Only Simca tried to
establish its factory outside the cluster.
During a European trip just after his electoral victory, JK visited Simca’s
factory near Paris and showed enthusiasm towards the establishment of a sub-
sidiary in Brazil, preferably in his native state of Minas Gerais. Convinced by
the direct commitment of the president of Brazil, the French became involved
in the project amidst a serious lack of definition and knowledge, to the point
that their plan implied importing CKD kits, which disobeyed the GEIA’s rules.
Simca do Brasil was founded in May 1958 in Belo Horizonte, capital of Minas
Gerais, where it received land assigned and prepared by the state governor. The
shareholders were the National Steel Company, a group of banks in Minas
Gerais, the Banco Francés e Brasileiro, and Simca itself, in a minority pos-
ition. Almost a year later, in March 1959, the company began to operate pro-
visionally in a rented workshop in São Bernardo (ABC), 800km from Belo
Horizonte. The decision was taken by the second technical authority of Simca,
who had been sent to Brazil to monitor the operation. The domestic content
requirements of the GEIA forced the temporary transfer to São Paulo. One year
later, it was admitted that the move to Minas Gerais was unviable. Practically all
of the nearly 1,000 suppliers were located in São Paulo (there were three in Rio
Grande do Sul and another one in Rio de Janeiro), including Ford do Brasil,
which supplied the engines (Sandler, 2005).
The case of Vemag illustrates the vibrant dynamism that the cluster
experienced in the mid-​1950s. Taking advantage of the GEIA’s guidelines,
Vemag established a joint venture with the German DKW to manufacture its
vans and jeeps, producing the first vehicle in 1958. Vemag also participated in
the production of trucks, this time through a joint venture with the Swedish
Scania Vabis, established in 1957. Scania do Brasil installed an engine and
chassis factory for buses and trucks in Ipiranga, very close to the Vemag factory
where the trucks were to be assembled. The engine blocks were cast by the
Companhia Fabricadora de Peças in its factory in Santo André (ABC), with
the help of Swedish technicians. The factory began activity in 1959, producing
1,000 engines with 40 per cent Swedish pieces. One year later, Scania decided
to take on solo production of the truck by dissolving the joint venture, after
accusing Vemag of lacking the capacity to assemble for both DKW and Scania.
However, it continued to use the Vemag line until December 1962, when it
inaugurated its new factory. The factory was located in São Bernardo, on land
132

132  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

assigned by Vemag in exchange for shares in the new company. With the move,
Scania located itself in the heart of the ABC region, where most of its 680
suppliers were also located (Scania, 2007, pp. 50–​59), although only some can
be defined as autoparts makers.
The results of the policies adopted by the GEIA are clear. In 1962, produc-
tion reached nearly 200,000 vehicles, with over 30,000 people employed by
carmakers and another 30,000 employed in the autoparts industry. Furthermore,
the percentages of domestic content exceeded 90 per cent (ANFAVEA, 1962b,
p. 13). Despite that, JK’s government had to face the rejection of international
institutions such as the IMF, an organization the government broke with in
1959. According to the Fund, the Brazilian government could not achieve
comprehensive development since it broke with its monetary and credit pol-
icies (Gattas, 1981, pp. 297–​298).
The expansion of the automobile industry was interrupted by the economic
and social turmoil which began in 1962 and led to a military dictatorship
in 1964. Production reached 190,000 vehicles in 1962, a figure that was not
exceeded until 1966. Behind this evolution, we find the political instability in
which Brazil had become immersed. Janio da Silva Quadros, appointed presi-
dent on 31 January 1961, proved to be a severe critic of policies to promote
industrialization, particularly due to the registered inflation rates. However,
amidst a serious political crisis, Quadros could not even finish the first year of his
term. In August, he was replaced by Joao Goulart, his vice president, who failed
to implement any consistent economic policy (Baer, 2014, p.  73). Incapable
of stabilizing the situation, Goulart, who was accused of having communist
tendencies after nationalizing a few small-​scale refineries and announcing his
willingness to carry out agrarian reform, was ousted by a right-​wing military
coup on 31 March 1964. Although the GEIA was taken over by the Executive
Group of Machine Industries, the military maintained basic aspects of the GEIA
policies, including market protection and the domestic content requirements.
However, with the military governments pretty much closed to foreign capital
(Skidmore, 1988, ­chapters 2–​4; Moraes, 2016), a process of denationalization of
the autoparts industry took place (Addis, 1999, pp. 108–​121).
The large carmakers gradually organized the cluster hierarchically as the
1960s progressed, transferring parts of the organizational capabilities that they
had been developing for years and that were at the heart of their competi-
tive advantage (Chandler, 1977; Lazonick, 1990). In order to achieve the level
of local content to which they had committed, the end manufacturers were
forced to teach the car parts industry concepts of industrial organization, to give
them equipment and tools, and to facilitate contact with their foreign suppliers,
helping them to obtain the transfer of licences and technical assistance (Shapiro,
1994, ­chapter  5; Addis, 1999, ­chapter  3). The signing of long-​term contracts
(mostly supported by the GEIA and under exclusivity agreements) also allowed
the long-​term development of the autoparts industry. As this process expanded,
the São Paulo cluster gradually adopted a similar form to what Markusen
(1996) characterized as hub-​and-​spoke districts, although the participation in
 133

Catching the ladder 133

international trade was almost token –​in the 1960s, 987 buses were exported,
26 trucks and 3 cars (ANFAVEA, 2016, p. 70) –​and, above all, the business asso-
ciations maintained their central position in the cluster.
In the 1960s, trade associations such as Sindipeças, the Federation of
Industries of the State of São Paulo, and the Industries Centre of the State of
São Paulo, as well as educational and vocational organizations, continued to
play a central role in the development of the autoparts industry, although their
function changed. In the 1950s, the main task of industry unions was to serve as
a lobby to protect the industry and to establish strict local content requirements.
However, in the 1960s, their efforts were focused on obtaining quality standards
and on training qualified and semi-​qualified workers. Technical assistance was
given to the manufacturers through agreements with the institutes of tech-
nology, especially the Instituto Tecnológico de Aeronáutica (Technological
Institute of Aeronautics), located in São José dos Campos, and the Instituto
de Pesquisas Tecnológicas (Institute of Technological Research) located in
São Paulo. Through these collaborations, they sought to maintain permanent
contact with the small and medium-​sized companies in order to guide them
towards the obtainment of higher-​quality standards. The industry union also
coordinated companies with technical teaching institutions to develop courses
in metalwork, mechanics and electricity, as well as to establish university degrees
in engineering, economics and business administration. The semi-​ qualified
operators were normally trained in the workplace through short-​term method-​
based courses. Some noteworthy training centres for medium and higher-​level
managers include the University of São Paulo, the Technological Institute of
Aeronautics of São José dos Campos and the Getúlio Vargas Foundation.
According to Teixeira, Correa and Fausto (1967, p.  160), 27 per cent of
the small and medium-​sized firms that manufactured autoparts in São Paulo
in 1965 had requested direct help from one of the technological institutes.
This research was funded by the Delft Programme, which is in itself a sign of
the cluster’s active institutional environment in the search for competitiveness,
since it was the result of an agreement between the Faculty of Economic and
Administrative Sciences at the University of São Paulo, the Industries Centre of
the State of São Paulo, the Industrial Social Service of São Paulo and the Research
Institute for Management Science at the Delft University of Technology in the
Netherlands, its goal being the training by consultants of the SME autoparts
makers of São Paulo. The land registry data analyzed in the study show that, in
1966, there were 489 companies registered as autoparts manufacturers, of which
62 were described as artisanal (less than eight workers), 309 as small (between
eight and 110 employees), 104 as medium (between 111 and 550 workers), and
14 were considered large companies (over 550 employees). Nevertheless, the
study detected that, of the 413 companies identified as small and medium-​sized
(SMEs), only 368 were effectively operating, as the rest manufactured autoparts
as a secondary activity, had moved outside the cluster or were inactive (Teixeira,
Correa and Fausto, 1967, pp. 8–​10).This leads one to conclude that the number
of SMEs manufacturing autoparts was around 400, since presumably part of the
134

134  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

250

200

150

100

50

0
1911–20 1921–30 1931–40 1941–50 1951–60 1961–65
Small Medium Total

Figure 7.1 Period of establishment of the SMEs operating in the cluster in 1965

Source: Authors’ elaboration from Teixeira et al. (1967, pp. 18–​19).The percentages obtained
for the 112 companies effectively studied have been applied to the group of 368 SMEs.

artisanal companies should also have been excluded from the sample. However,
since the study does not consider companies located in São José dos Campos
and Taubaté, the total number of companies present in the district would easily
be over 400. In terms of activity, a little over three-​quarters of the SMEs were
dedicated to manufacturing mechanical parts, one-​tenth to the production of
electrical material, 7 per cent to manufacturing rubber parts, and a little over 2
per cent both to upholstery and to plastic parts.
Figure  7.1 shows that, of the SMEs active in 1966, only 1 per cent were
established before 1920, 2 per cent during the 1920s and 8 per cent in the
1930s. The boost experienced due to isolationism resulting from the Second
World War is reflected in the fact that 14 per cent of the companies were
established in the 1940s. Nevertheless, the figure clearly illustrates the impact of
industrial policy: over 60 per cent of the companies were founded in the 1950s,
the decade in which the industrial policies that explain the cluster’s growth
were implemented, both by Kubitschek’s and Vargas’s governments.8 The ser-
ious crisis experienced by the Brazilian economy between 1962 and 1966 is
reflected in the fact that fewer companies were founded in the first five years of
the 1960s than in the 1940s, and represented only a quarter of those founded
in the 1950s. Moreover, as already indicated, at the heart of most of the SMEs
manufacturing autoparts, we find specialized workers from other companies
in the sector. At a time of considerable increase in the demand for inputs from
the final assemblers, it seems logical that workers who had enough knowledge
about the industry would venture to found their own companies. Therefore,
while most of the 122 companies that founded Sindipeças at the beginning
 135

Catching the ladder 135

of the 1950s did not remain active, they were the origin of the companies
established in the second half of the 1950s, which constituted the bulk of the
companies operating in the mid-​1960s.
The fact that a number of these SMEs were set up by specialized workers
from other companies in the sector reinforces the hypothesis of the presence of
Marshallian externalities, at the same time as pointing to inheritance as a key
factor in the cluster’s reproduction, which is in line with what Klepper (2010)
indicated regarding the formation of the auto cluster in Detroit. However, the
hierarchization of the cluster around the large-​end assemblers adopting a hub-​
and-​spoke form, as described in Markusen (1996), limited this model since
large companies with higher salaries attracted workers trained in the medium-​
sized companies of the ancillary industry, which in turn did the same thing
with specialized workers from the sector’s small companies (Teixeira, Correa
and Fausto, 1967, p. 145).
The vast majority of the companies were members of one of the business
organizations in the district, which proves the importance that the institutions
had in the creation of competitive advantage. Thus, only 10 per cent of the
district’s SMEs were not members of any association, in most cases small com-
panies. Sindepeças was by far the largest association of the autoparts industry,
with a membership percentage that reached 44 per cent of the SMEs in the
district. The Federation of Industries of the State of São Paulo and the CESP
had, together, a level of membership that represented 33.3 per cent of the small
companies and 28.6 per cent of the medium-​sized companies (Teixeira, Correa
and Fausto, 1967, p. 35). Also, a third of the SMEs surveyed confirmed having
received assistance from one of the trade associations in the district, a percentage
which increased to half in the case of medium-​sized companies. Likewise, only
14 per cent of the SMEs surveyed did not know that institutions in the cluster
could offer them assistance, whether technical, administrative or financial
(Teixeira, Correa and Fausto, 1967, p. 161). For their part, the end assemblers
were organized in the National Association of Motor Vehicle Manufacturers, or
ANFAVEA, created in May 1956, just at the beginning of Kubitschek’s devel-
opmental policies.
The automotive industry led the economic recovery of the period 1968–​
1974. During those years, the production of vehicles grew at an average annual
rate of 22 per cent (double the overall growth of the economy), exceeding
800,000 passenger cars and light commercial vehicles manufactured and close
to 100,000 trucks and buses in 1974. At the end of the period, the six carmakers
that were still active in São Paulo (see Table 7.2) directly generated 100,000 jobs,
while the ancillary industry generated almost 200,000 jobs (ANFAVEA, 2016,
pp. 44–​45). The driver of growth was domestic consumption, which expanded
under expansionary macroeconomic policies.
When the economic recovery of 1967–​1968 began, the structure of the
cluster was changing. The weight of the small and medium-​sized companies
was reducing, the end manufacturers were integrating more parts from the
136

136  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

production process, and the number of companies controlled by foreign capital
was increasing (Addis, 1999, pp. 93–​95). With the surge in inflation, the end
manufacturers pressed the ancillary industry (mainly by terminating the exclu-
sivity contracts) to achieve price reductions. When they could, they opted to
integrate production vertically, and when this was not profitable, they promoted
the installation of new suppliers, especially foreign ones that tended to enter
through joint ventures with local companies already in operation (Shapiro,
1994, pp.  191–​215). Despite the growth in production, the structure of the
district did not change much and its hierarchical organizational structure was
consolidated around the large final assembly plants.
The take-​off of the cluster began at the start of the 1950s and culminated
at the start of the 1970s. Although, when the expansion began, the structure
of the cluster could resemble that of a district with a Marshallian base, with
abundant labour, specialized suppliers and the relatively simple circulation of
knowledge and information through the movement of specialized workers
who gradually created their own companies. When the expansion ended, its
structure fitted in reasonably well with what Markusen (1996) described as the
hub-​and-​spoke district. During this transformation, local capital lost its pre-​
eminence compared with international capital, and the associations of small and
medium-​sized companies, the true promoters of the district at the beginning of
the 1950s due to their insistence on the application of policies protecting the
industry, shifted their activity toward the provision of training and specialized
consulting. From the mid-​1970s, and for the following 20  years, both pro-
duction and the number of employees remained relatively stable, indicating
that the cluster was passing through its maturity stage which, in view of its
characteristics, deserves a separate study.

Concluding remarks: A long-​run view

The research presented reveals how the regional concentration of industry
does not follow a predetermined path, but rather is affected by endogenous
and exogenous factors which interact with different results, strongly influenced
both by the different historical situations and by the regional area in which they
are inserted. Therefore, a full understanding of the emergence and expansion
of industrial districts and clusters also requires an analysis with a historical per-
spective which, in turn, increases in explanatory potential if it integrates the-
oretical and conceptual frameworks developed from multidisciplinary spheres
such as regional studies, economic geography and evolutionary economics.
During the first half of the 20th century, São Paulo was the most industrialized
region in Brazil, a backward and agriculture-​based country. However, by the
middle of the century, the production of cars in São Paulo was almost token.
Twenty-​five years later, São Paulo had placed Brazil in the world top ten of
car-​producing countries. Different leading brands in the industry agreed to
transfer their knowledge and technologies, forced by a closed market situation
and by strict local content [Link] also had to do this if they did not
 137

Catching the ladder 137

Table 7.3 Determinants in the development of the São Paulo auto industry cluster
by stage

Cluster life-​cycle Main sources of development Other sources of development

stage

Until Formation External Economies Institutional Environment

1929 1. Pool of labour 1. Infrastructures
2. Pool of industrial activity
1930–​55 Achieving critical Institutional Environment External Economies
mass 1. Trade associations 1. Pool of labour
2. Infrastructures 2. Pool of suppliers
3. Knowledge spillovers
1956–​74 Growth Active Industrial Policies External Economies
1. Protectionism 1. Pool of labour
2. Domestic content 2. Specialized suppliers
Large Companies 3. Knowledge spillovers
1. Hub firms Institutional Environment
2. Internal scale economies 1. Trade associations
3. Organizational capabilities 2. Infrastructures
3. Education and
vocational institutions

Source: Authors’ elaboration.

want to abandon an empty and rapidly expanding market. However, the out-
come was that the cluster evolved towards a hierarchical organizational struc-
ture headed by the large-​end manufacturers under foreign control, while the
weight of Brazilian capital, even in the autoparts industry, gradually declined.
The key factors in the performance of the São Paulo auto industry cluster
in each stage of development are indicated in Table 7.3. The origins of the São
Paulo auto industry cluster can be found in the dozens of small workshops
serving the spare parts market that progressively emerged during the 1920s
to facilitate the region’s motorization, while Ford, GM and International
Harvester established facilities to assemble CKD kits imported entirely from
the US. Their activity was facilitated by the good connection  –​by railroad
(1870s) and by highway (1910s) –​between the Port of Santos and the city of
São Paulo. In addition, from 1910, there was a group of associations that aimed
to promote the automotive sector. Despite the fact that early knowledge of the
art was widespread throughout the region in the form of semi-​skilled workers
and some specialized suppliers, the critical mass necessary for the cluster’s exist-
ence was still far off.
The opportunity to achieve a critical mass came with the trade restrictions
brought about by the Second World War. Restrictions on car part imports from
industrialized nations favoured the flourishing of hundreds of workshops that
were able to provide a wide range of parts and components, even though they
were undercapitalized and undersized. According to Table 7.1, 65 per cent of the
workshops and 80 per cent of the workers in the Brazilian automotive industry
138

138  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

were located in São Paulo in 1949. As local producers became aware of their
common interests, they joined together to push to maintain barriers to foreign
competitors once the war was [Link] association of autoparts makers created
in São Paulo in October 1952 to obtain market protection had its first success
a few months later when 108 groups of parts were protected. The importing of
complete vehicles was banned shortly after, with only the entry of CKD kits
and none of the aforementioned groups of parts being permitted. Forced by the
new legislation,Volkswagen and Willys-​Overland also established assembly lines
in São Paulo. External economies in the form of specialized suppliers and semi-​
skilled workers, together with a powerful industry association, worked well for
this location. Although the critical mass had been reached, the cluster’s growth
had not yet occurred.
The cluster’s expansion took place once JK’s government, established in
January 1956, applied a broad set of active industrial policies that can be defined
as infant industry protection a la List (see Chang, 2002). The high level of
domestic content required forced all project facilities to be located in São Paulo,
reinforcing the sector’s concentration in the region, as shown in Table 7.1. If
protection and, above all, domestic content requirements were key for the
cluster’s growth, they found fertile ground in the form of external economies
made up of hundreds of suppliers and thousands of workers. Proof of this prior
base can be found in the existence of an institutional framework that accom-
panied the industry’s progress, and which worked towards the comprehensive
development of the industry by facilitating the circulation of knowledge and
by providing assistance in worker training and business organization. The data
presented in Figure 7.1 on the year of establishment of the SMEs active in the
cluster in the mid-​1960s fit well with the interpretation proposed. However,
since then, large subsidiaries of global carmakers have led the cluster’s devel-
opment, causing the district to take on a hub-​and-​spoke form, as described by
Markusen (1996).
By targeting the auto industry for investment, placing barriers to foreign pro-
duction and requiring high levels of domestic content, strategic industrial pol-
icies fostered the development of the auto industry clustered around São Paulo,
as stated by Chang (1993, 2002).While the active industrial policy applied since
the 1950s can explain the automobile industry’s take-​off, it cannot explain its
concentration in São Paulo. The industry clustered in the region due to the
presence of external economies and the actions of its own institutions, mainly
trade associations, which led the cluster’s development until the mid-​1960s.
Nevertheless, the cluster’s path of development also presents some limitations,
since the development of local carmakers did not occur. Both the Japanese and
the Korean experience reveal that the goal is not only to force the introduction
of an industry, but also to create domestic firms with capabilities to compete in
the international arena.
This research shows how a certain region in a backward country was able
to develop one of the flagship industries of the Second Industrial Revolution,
proving the potential of infant industrial policy. However, the research also
 139

Catching the ladder 139

highlights how active policies only impacted the most industrialized region
of the country. A  pool of specialized suppliers and skilled workers, together
with vibrant industry associations and an appropriate provision of public goods,
increased the region’s competitive [Link] study has shown what factors
are behind the formation and growth of the São Paulo auto-​industry cluster.
However, only more research on the maturity stage will tell us how the cluster
has been able to sustain its activity over time.

Notes
1 The authors wish to express their gratitude to the Spanish Ministry of Economy and
to the European Regional Development Fund (ERDF) for their financial support
for this research through project HAR2015-​64769-​P, as well as to the Kurgan-​van
Hentenryk Chair in Business History of the Solvay Brussels School of Economics
and Management. Preliminary versions of this paper have been presented at several
venues including the 1st World Congress in Business History/​20th Congress of the
European Business History Association, the 2016 Economic History Society Annual
Conference, and the XII Congresso Brasileiro de História Econômica. We especially
thank organizers and audience members at the Inaugural Workshop RSA ReHi net-
work, as well as the editors of this volume for their insightful comments. Usual
disclaimers applied.
2 A nuanced point of view regarding the Brazil case is offered in Evans (1995); Amann
and Baer (2005); and Baer (2014).
3 A knock-​down kit is a kit containing the parts needed to assemble a product. A CKD
is a kit of the completely non-​assembled parts of a product and a method of supplying
parts to a market, particularly on shipping to foreign nations, being a common prac-
tice in the automotive industry.
4 In September 1953, the Associaçao was authorized to register itself as a trade asso-
ciation, which granted it a much greater capacity for action, turning it into the
powerful Sindicato da Indústria de Peças para Automóveis e Similares no Estado de
São Paulo, commonly known as Sindipeças.
5 The industrialist postulates formulated by Simonsen can be found in Simonsen and
Gudin (1977) and are analyzed in, for example, Bielschowsky (1988), Fonseca (2003),
Cepêda (2003), and Bruzzi and Mendes (2015).
6 The Plano de Metas sought the industrialization and modernization of Brazil and was
launched as a result of the work undertaken by the Conselho de Desenvolvimento,
established by Decree No. 38.744 of 1 February 1956, the first working day of the
Kubitschek administration. With the motto “50 years in five”, the plan presented 31
goals organized into five main groups –​energy (1 to 5), transport (6 to 12), food (13
to 18), basic industry (19 to 29), and education (30) –​plus one goal called a syn-
thesis goal: the construction of a new capital, Brasilia, which was to be located in the
centre of the country. As goal number 27, automobiles were the only consumer good
included in the plan.
7 Under a national champion policy, by setting policies which favour a particular
national firm (whether private or state-​owned), governments expect this domestic
corporation to seek profit and to advance the interests of the nation. This strategy
140

140  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

was largely used in the Gaullist dirigisme in France, also inspiring some developing
countries to adopt it.
8 The differences between the two administrations regarding the conditions to attract
FDI are explained in Campos (2017).

References
Acemoglu, D. and Robinson, J., 2012. Why nations fail: The origins of power, prosperity, and
poverty. New York: Crown.
Addis, C., 1999. Taking the wheel: Auto parts firms and the political economy of industrializa-
tion in Brazil. University Park PA: Pennsylvania State University Press.
Allen, R., 2010. The British industrial revolution in global perspective. Cambridge: Cambridge
University Press.
Amann, E. and Baer, W., 2005. From the developmental to the regulatory state:  The
transformation of the government’s impact on the Brazilian economy. Quarterly
Review of Economics and Finance, 45(2–​3), pp. 421–​431. [Link]
[Link].2004.12.012
Amin, A. and Thrift, N., eds., 1994. Globalization, institutions, and regional development in
Europe. Oxford: Oxford University Press.
Amsden, A., 1989. Asia’s next giant: South Korea and late industrialization. Oxford: Oxford
University Press.
Amsden, A., 2003. The rise of “the rest”: Challenges to the west from late-​industrializing econ-
omies. Oxford: Oxford University Press.
ANFAVEA, 1961. Industria automobilística Brasileira: Novembro 1961. São Paulo: ANFAVEA.
ANFAVEA, 1962a. Industria automobilística brasileira: Sua contribução para o desenvolvimento
nacional. São Paulo: ANFAVEA.
ANFAVEA, 1962b. Industria automobilística brasileira: Novembro 1962. São Paulo: ANFAVEA.
ANFAVEA, 2016. Anuário da indústria automobilística Brasileira 2015. São Paulo: ANFAVEA.
Baer, W., 2014. The Brazilian economy:  Growth & development. Boulder CO:  Lynne
Rienner.
Bagnasco, A., 1977. Tre Italie: La problematica territoriale dello sviluppo. Bologna: Il Mulino.
Becattini, G., [Link] Marshallian industrial district as a socio-​economic notion. In: F.
Pyke, G. Becattini and W. Sengenberger, eds. Industrial districts and inter-​firm cooperation
in Italy. Geneva: International Institute for Labour Studies. Pp. 37–​51.
Bergman, E., 2008. Cluster life cycles:  An emerging synthesis. In:  C. Karlsson, ed.
Handbook of research in cluster theory. Cheltenham: Edward Elgar. Pp. 114–​132
Bielschowsky, R., 1988. Pensamento econômico Brasileiro:  O ciclo ideológico do
desenvolvimentismo. Rio de Janeiro: IPEA/​INPES.
Brusco, S., 1982. The Emilian model:  Productive decentralisation and social integra-
tion. Cambridge Journal of Economics, 6(2), pp. 167–​ 84. [Link]
[Link]/​10.1093/​
[Link].a035506
Bruzzi, L. F. and Mendes, A., 2015. Redimensionando a contribuição de Roberto
Simonsen à controversia do planejamento (1944–​1945). America Latina en la Historia
Económica, 22(3), pp. 76–​107. [Link]
Campos, F. A., 2017. Internacionalização brasileira e Instrução 113 da SUMOC. América
Latina en la Historia Económica, 24(2), pp. 93–​124.
Cano, W., 1981. Raizes da concentraçao industrial em Sao Paulo. São Paulo: T. A. Queiroz.
 141

Catching the ladder 141

Carli, G. and Morrison, A., 2018. On the evolution of the Castel Goffredo hosiery
cluster:  A life cycle perspective. European Planning Studies, 26(5), pp. 915–​ 932.
[Link]
Catalan, J., 2010. Strategic policy revisited: The origins of mass production in the motor
industry of Argentina, Korea and Spain, 1945–​87. Business History 52(2), pp. 207–​230.
[Link]
Catalan, J., 2017. The life-​cycle of the Barcelona automobile-​industry cluster, 1889–​
2015. Revista de Historia Industrial, 26(66), pp. 77–​125.
Catalan, J. and Fernández-​de-​Sevilla, T., 2013. Die staatliche industriepolitik und die
entwicklung der automobilindustrie in Spanien 1948–​1985. In: [Link] and [Link],
eds. Automobilindustrie 1945–​2000. Munich: Oldenbourg. Pp. 254–​284.
Catalan, J. and Fernández-de-Sevilla, T., 2018. Emergence and maturity of the devel-
opmental state in Argentina, Brazil and Spain, 1930-1990: An economic history
approach. In: A. Ferraro and M. Centeno, eds. State and nation making in Latin America
and Spain:The rise and fall of the developmental state. Cambridge: Cambridge University
Press. Pp. 207–237.
Catalan, J., Miranda, J. A. and Ramon-​Muñoz, R., eds., 2011a. Distritos y clusters en la
Europa del sur. Madrid: LID.
Catalan, J., Miranda, J. A. and Ramon-​ Muñoz, R., 2011b. Ventaja competitiva y
capacidades empresariales de los distritos de la Europa meridional:  Evidencia
histórica para tres revoluciones tecnológicas. In:  J. Catalan, J. A. Miranda and R.
Ramon-​Muñoz, ed. Distritos y clusters en la Europa del sur. Madrid: LID. Pp. 367–​412.
Cepêda, V. A., 2003. Roberto Simonsen e a formaçao da ideologia industrial no
Brasil: Limites e impasses. PhD dissertation: Universidade de São Paulo.
Chandler, A., 1977. The visible hand:  The managerial revolution in American business.
Cambridge MA: Harvard University Press.
Chang, H.-​J., 1993. The political economy of industrial policy in Korea. Cambridge
Journal of Economics, 17(2), pp. 131–​157. [Link]
a035227
Chang, H.-​J., 2002. Kicking away the ladder:  Development strategy in historical perspective.
London: Anthem Press.
Cimoli, M., Stiglitz, J. and Dosi G., ed., 2009. Industrial policy and development.
Oxford: Oxford University Press.
Dean,W., 1969. The industrialization of Sao Paulo, 1880–​1945. Austin TX: The University
of Texas  Press.
Donnelly,T., Begely, J. and Collis, C., [Link] West Midlands automotive industry: The
road downhill. Business History, 59(1), pp. 56–​ 74. [Link]
00076791.2016.1235559
Eichengreen, B., 2007. The European economy since 1945:  Coordinated capitalism and
beyond. Princeton NJ: Princeton University Press.
Enright, M. J., 2000. Regional clusters and multinational enterprises:  Independence,
dependence, or interdependence? International Studies of Management and Organization,
30(2), pp. 114–​138. [Link]
Enright, M. J., 2003. Regional clusters: What we know and what we should know. In: J.
Bröcker, D. Dohse and R. Soltwedel, eds. Innovation clusters and interregional competi-
tion. Berlin: Springer Verlag. Pp. 99–​129.
Evans, P., 1995. Embedded autonomy: State and industrial transformation. Princeton: Princeton
University Press.
142

142  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

Fausto, B., 2006. Getúlio Vargas: O poder e o sorriso. São Paulo: Companhia das Letras.
Fernández-​ de-​
Sevilla, T., 2014. Inside the dynamics of industrial capitalism:  The
mass-​production of cars in Spain. Revista de Historia Económica–​Journal of Iberian
and Latin American Economic History, 32(2), pp. 287–​315. [Link]
S0212610914000093
Fleury, A. and Salerno, M., 1998. The transfer and hybridization of new models of pro-
duction in the Brazilian automobile industry. In: R. Boyer, E. Charron, U. Jürgens
and S. Tolliday, ed. Between imitation and innovation:  The transfer and hybridization of
productive models in the international automobile industry. Oxford:  Oxford University
Press. Pp. 278–​295.
Fonseca, P. C. D., 2003. Sobre a intencionalidade da politica industrializante do Brasil
na década de 1930. Brazilian Journal of Political Economy–​Revista de Economia Política,
23(1), pp. 133–​148.
Forest, M., 2002. Automóveis de São Paulo. São Paulo: Arquivo do Estado.
Gattas, R., 1981. A indústria automobilística e a 2ª revoluçao industrial no Brasil. São
Paulo: Prelo.
General Motors, 1995. General Motors do Brasil: 70 years of history. São Paulo: General Motors.
Gerschenkron, A., 1962. Economic backwardness in historical perspective. Cambridge,
MA: Harvard University Press.
Gonçalves,V., 1989. O século do automóvel no Brasil. São Paulo: Iconographia.
Harrison, B., 1994. Lean and mean: The changing landscape of corporate power in the age of
flexibility. New York: The Guilford Press.
Jenkins, R., 1995. The political economy of industrial policy: Automobile manufacture
in the newly industrialising countries. Cambridge Journal of Economics, 19(5), pp. 625–​
645. [Link]
Johnson, C., 1982. MITI and the Japanese miracle:  The growth of industrial policy 1925–​
1975. Redwood City CA: Stanford University Press.
Klepper, S., 2007. Disagreements, spinoffs, and the evolution of Detroit as the capital of
the U.S.  automobile industry. Management Science, 53(4), pp. 616–​631. [Link]
org/​10.1287/​mnsc.1060.0683
Klepper, S., 2010. The origin and growth of industry clusters: The making of Silicon
Valley and Detroit. Journal of Urban Economics, 67(1), pp. 15–​32. [Link]
10.1016/​[Link].2009.09.004
Landes, D., 1969, The unbound Prometheus: Technological change and industrial development in
Western Europe from 1750 to the present. Cambridge, MA: Cambridge University Press.
Latini, S., 2007. A implantaçao da indústria automobilística no Brasil:  Da substituçao da
importaçoes ativa à globalizaçao passive. São Paulo: Alaude.
Lazonick, W., 1990. Competitive advantage on the shop floor. Cambridge MA:  Harvard
University Press.
Marglin, S., 1990. Lessons of the golden age: An overview. In: S. Marglin and J. B. Schor,
ed. The golden age of capitalism: Reinterpreting the postwar experience. Oxford: Oxford
University Press. Pp. 1–​38.
Markusen, A., 1996. Sticky places in slippery space: A typology of industrial districts.
Economic Geography, 72(3), pp. 293–​313.
Marshall, A., 1890. Principles of economics. London: Macmillan. 8th ed. 1920.
Martin, R. and Sunley, P., 2011. Conceptualizing cluster evolution:  Beyond the life
cycle model? Regional Studies, 45(10), pp. 1299–​ 1318. [Link] [Link]/​
10.1080/​
00343404.2011.622263
 143

Catching the ladder 143

Maskell, P. and Kebir, L., 2006. What qualifies as a cluster theory? In:  B. Asheim, P.
Cooke and R. Martin, eds. Clusters and regional development:  Critical reflections and
explorations. London: Routledge, pp. 30–​49.
Menzel, M.-​P. and Fornahl, D., 2010. Cluster life cycles –​Dimensions and rationales of
cluster evolution. Industrial and Corporate Change, 19(1), pp. 205–​38. [Link]
10.1093/​icc/​dtp036
Molema, M., Segers, Y. and Karel, E., 2016. Introduction:  Agribusiness clusters in
Europe, 19th and 20th centuries. The Low Countries Journal of Social and Economic
History, 13(4), pp. 1–​16. [Link]
Moraes, R., 2016. O Governo Castello Branco e a Federação das Indústrias do estado de
São Paulo: As bases do “milagre” (1964–​1967). América Latina en la Historia Económica,
23(2), pp. 64–​90. [Link]
Nascimento, B., 1976. Formaçao da indústria automobilística brasileira:  Política de
desenvolvimento industrial em uma economia dependente. São Paulo: USP.
Negri, B., 1996. Concentraçao e desconcentraçao industrial em São Paulo (1880–​ 1990).
Campinas: Editora de Unicamp.
Norman, A. and Stiglitz, J., eds., 2016. Efficiency, finance and varieties of industrial Policy.
New York: Columbia University Press.
Olson, M., 1982. The rise and decline of nations:  Economic growth, stagflation, and social
rigidities. New Haven CT: Yale University Press.
Pardi, T., 2017. Industrial policy and the British automotive industry under
Margaret Thatcher. Business History, 59(1), pp. 75–​100. [Link]
00076791.2016.1223049
Pollard, S., 1981. Peaceful conquest:  The industrialization of Europe 1760–​1970. Oxford:
Oxford University Press.
Popp, A. and Wilson, J., 2007. Life cycles, contingency, and agency: Growth, develop-
ment, and change in English industrial districts and clusters. Environment and Planning
A, 39(12), pp. 2975–​2992. [Link]
Porter, M., 1990. The competitive advantage of nations. New York: Free Press.
Porter, M., 2008. Clusters and competition: New agendas for companies, governments,
and institutions. In: Michael Porter, ed. On competition: Updated and expanded edition.
Cambridge MA: Harvard Business Review Book. Chapter 7.
Sandler, P., 2005. DKW. A grande historia da pequenha maravilla. São Paulo: Alaude.
Scania, 2007. Scania no Brasil:  Passado, presente e futuro, 1957–​2007. São Bernardo do
Campo: Scania.
Schmitz, H., 1999. Collective efficiency and increasing returns. Cambridge Journal of
Economics, 23(4), pp. 465–​83. [Link]
Schneider, S., 1991. Politics within the state: Elite bureaucrats and industrial policy in authori-
tarian Brazil. Pittsburgh PA: Pittsburgh University Press.
Scranton, P., 1997. Endless novelty:  Specialty production and American industrialization,
1865–​1925. Princeton NJ: Princeton University Press.
Shapiro, H., 1991. Determinants of firm entry into the Brazilian automobile manufac-
turing industry, 1956–​1968. Business History Review, 65(4), pp. 876–​947. [Link]
org/​10.2307/​3117267
Shapiro, H., 1994. Engines of growth: The state and transnational auto companies in Brazil.
Cambridge: Cambridge University Press.
Sikkink, K., 1991. Ideas and institutions: Developmentalism in Brazil and Argentina. Ithaca
NY: Cornell University Press.
144

144  Tomàs Fernández-de-Sevilla and Armando J. dalla Costa

Silva, S., 1976. Expansao cafeeira e origens da industria no Brasil. São Paulo: Alfa Omega.
Simonsen, R. and Gudin, E., 1977. A controvérsia do planejamento na economia brasileira.
Rio de Janeiro: Ipea/​Inpes.
Skidmore, T., 1988. The politics of military rule in Brazil, 1964–​85. New York: Oxford
University Press.
Stiglitz, J. and Lin, J., ed., 2013. The industrial policy revolution I: The role of government
beyond ideology. New York: Palgrave.
Suzigan, W., 1986. A industrializaçao Brasilera:  Origen e desenvolvimento. São Paulo:
Brailiense.
Swann, P., 2009. The economics of innovation: An introduction. Cheltenham: Edward Elgar.
Teixeira, D., Correa L. and Fausto, F., 1967. Pequenas e medias indústrias de autopeças. São
Paulo: Programa Delft.
Tödtling, F. and Trippl, M., 2004. Like Phoenix from the ashes? The renewal of clusters
in old industrial areas. Urban Studies, 41(5–​ 6), pp. 1175–​ 1195. [Link]
10.1080%2F00420980410001675788
Trippl, M., Grillitsch, M., Isaksen, A., and Sinozic, T., 2015. Perspectives on cluster evo-
lution: Critical review and future research issues. European Planning Studies, 23(10),
pp. 2028–​2044. [Link]
Wilkins, M., 1974. The maturing of multinational enterprise: American business abroad from
1914 to 1970. Cambridge MA: Harvard University Press.
Wilkins, M. and Hill, F., 2011. American business abroad: Ford on six continents. Cambridge
MA: Cambridge University Press.
Wilson, J. and Popp, A., eds., 2003. Industrial clusters and regional business networks in
England, 1750–​1970. Aldershot: Ashgate.
Wolfe, J., 2010. Autos and progress:  The Brazilian search for modernity. Oxford:  Oxford
University Press.
Woo-​Cumings, M., ed., 1999. The developmental state. Ithaca: Cornell University Press.
Young, S., Hood, N. and Peters, E., 1994. Multinational enterprise and regional eco-
nomic development. Regional Studies, 28(7), pp. 657–​777. [Link]
00343409412331348566
 145

8 
Urban and regional
development policy
Its history and its differences
Kevin Cox

Context1
The emergence of a state interest in urban and regional development in the US
and in Western Europe has to be seen first in the context of the expansion of
the state in the 20th century: an expansion which combined both regulatory
aspects and a very, very substantial increase in state spending. There were many
conditions for its appearance, but two in particular command attention. The
first was an increasing socialization of production: notably the deepening of the
division of labour and the growth of a shared physical and social infrastructure.
These were both problematic in terms of the familiar challenges surrounding
the mitigation of transaction costs, externalities, monopoly power and sec-
toral imbalance. The second condition was the rise of the labour movement,
comprising the labour unions and the political parties that they founded or
sponsored, and which would eventually result in the growth of the welfare state.
Urban and regional development would be one particular area of state inter-
[Link] was little evidence for this as such in the 19th century but in the
course of the 20th century, and as a label, it would assume increasing coherence
of meaning attaching to a distinct set of state practices. It would burst forth in
very clear form subsequent to the Second World War. It was a response to the
same issues generated by the socialization of production and social inequalities
but now projected onto a spatial plane and mobilizing space relations as a means
of mitigation. There was, for instance, an urban question focusing on the exter-
nalities subsequent to agglomeration (Preteceille, 1976); and on questions of
social segregation and inequality of life conditions. A regional question would
emerge alongside in the form of anxieties about geographically uneven devel-
opment, regional imbalance and inequalities in employment opportunities.
Regional disparities would then be reflected in differences in the challenges
faced by different urban areas, while resolving the urban question could be
turned to resurrecting regional fortunes.
In general, policy would assume the form of interventions into the geography
of things: defining where they should go and moving them around. Therefore,
the approach to urban and regional crisis would be one of displacement,
though with varying degrees of a mixture between public and [Link] state
146

146  Kevin Cox
would always be a significant actor; sometimes more passive and sometimes
more active. The emphasis would be technical and the ultimately social basis
of urban and regional questions that gave them a refractory character tended
to be ignored. One result would be that, to a large extent, problems would be
simply moved around. Capitalism inevitably produces a scarcity of jobs and
unemployment somewhere, as in Marx’s industrial reserve army: rising wages
result in an attempt to economize on labour which then entails an increase in
the numbers of the unemployed. As wages rise again, the process repeats itself.
If the economy is in an expansionary phase, moving employment to areas of
high unemployment may not immediately show up in unemployment in the
areas from which the jobs are moving. But in the contractionary phase, as prof-
itability is reduced, it will be: which is what happened in Western Europe in the
1970s and subsequently led to calls for an end to the policy.
The concrete shape eventually assumed by policy in response to urban
and regional problematics was conditional on a number of developments of a
social and technical character (Cox, 2016, c­ hapter 2). These would then come
together to form an emergent structure of social relations. I would include here
at least four: the growth of city and regional planning as a state function; the
rise of property capital in the form of the real estate development industry; a
new mobility in the space economy; and new sorts of real estate product, as in
industrial parks, planned factory districts, shopping centres, and garden suburbs.
City and regional planning were already taking shape at the beginning of the
20th century as a response to urban growth, to health problems, housing and
transportation. It would provide legitimation for the sorts of state interventions
that would come later under the heading of urban and regional development.
Property capital as a distinct branch of the capitalist division of labour focusing
on the provision of premises, on the other hand, does not emerge till the post-​
Second World War period, though there are earlier intimations. Under its
stimulus new forms of real estate product would appear: some of them large
enough to affect uneven development in dramatic fashion and imbricating with
what I have called a new mobility (Cox, 2016, pp. 93–​97). The latter had its
own origins in the second Industrial Revolution:  electricity and the manu-
facture of an increasing array of consumer goods. It was more footloose than
the industries of the first Industrial Revolution and so auxiliary to the physical
displacement that would be the state response to urban and regional questions.
In the immediate post-​war period, urban and regional development policy
had clear outlines which would be quite different in the countries of Western
Europe than in the US. In the former, the partitioning of space and displacement
of industry, housing and the like would be subject to strong central state orches-
tration: planning systems organized from the centre, along with state financial
incentives. Policy would be legitimated as a search for the good geography: the
right things in the right places. In the US it would be very different: almost
a mirror image as policy would be subordinated to strong bottom-​up forces.
Local governments, aided and abetted by the famous growth coalitions, would
take the lead while more central branches of the state, particularly the federal
 147

Urban and regional development policy 147

level, would be more passive: a source of regulatory relief and monies prompted
by demands from below. The motif of planning would be displaced by one of
adapting to the dictates of the market; in that case, a market in locations in which
local governments acted on the supply side, facilitating the provision of prem-
ises in a struggle to alter the contours of the space economy to local advantage.2
In turn, each of these approaches has had distinctive socio-​historical conditions,
which have come together to create particular structures of social relations,
empowering and limiting in their different ways. The aim of this chapter is to
enlarge on these themes: how urban and regional development emerged as a
policy focus; and how and why this policy focus would be expressed quite dif-
ferently in the countries of Western Europe on the one hand, and in the US on
the other, even while they shared most of the underlying conditions.

The coming into being of urban and regional

development policy
The idea of a specifically urban and regional development policy has its roots
in a complex combination of urban and regional questions. The conditions
for this can be traced back a long way. It is useful at the outset, though, to
point to a shift in the competitive nature of the capitalist economy. As both
David Harvey (1982, pp. 144–​145) and John Weeks (1981, pp. 167–​168) have
indicated, the cut-​throat nature of the capitalist economy did not spring into
being, ready-​ made. The common notion of a pure competition (of small
firms) preceding a more contemporary dominance of monopolies is deceiving.
Rather it is the other way around. Under 19th-​century conditions, and more
so the further back one goes, firms tended to be protected from competition by
the cost of transportation and by obstacles faced by would-​be competitors in
organizing the invasion of a particularly lucrative market. The space economy
approximated more to a set of local and regional monopolies; space economies,
in other words. Railroads, and later the iron steam ship, would begin to make
a dent in this, along with the separation of capital into distinct industrial and
financial [Link] would allow firms to raise the money to enter a market
where previously there had been little competition. But in short, competitive
pressures tended to intensify in the course of the 19th century, both within
countries and internationally, forcing the seeking-​out of competitive advantage
in agglomeration, regional specialization, new products that would undermine
existing geographic divisions of labour and so on.
The subsequent tensions would then assume sharp relief in urban and
regional crises. An important marker here is the debate about tariff reform
in the United Kingdom at the beginning of the 20th century. The impetus
came from Birmingham and the surrounding Black Country in the English
Midlands. The rise of German industry was severely affecting the prospects
of the dominant metal-​working firms of the area. This prompted demand for
some modification of the free trade that had dominated British policy hitherto.
But it lacked serious support elsewhere in the country and there was opposition
148

148  Kevin Cox
from the City of [Link] Port of London did a healthy trade in the import
of goods from the rest of Europe prior to export elsewhere and City banks
prospered in part on financing these movements. In the US the shift of the tex-
tile industry to the Piedmont states of the South would promote similar angst
in New England, but, and significantly, was met from the 1920s on by a regional
response promoting transition in the regional economy (Koistinen, 2013).
The urban question was of longer standing:  issues of health, housing and
urban degeneracy long blamed on the moral failures of the poor and addressed
through an appropriate moralization. Temperance and social visits by the wives
of the well-​to-​do, often based on settlement houses, were part of the common
practice –​a sort of missionary approach to what were seen as those in need
of salvation. At the same time, particularly in Western Europe, there was a val-
orization of the rural, which was heavily steeped in class anxieties. The rural
was regarded as healthy, a repository of traditional values and of feelings of
belonging, while in urban areas alienation ran wild, church attendance lagged
and market values ruled. This was the founding motif of the new academic
discipline of sociology. Weber talked of tradition and modernity, Durkheim of
organic and mechanical solidarity, Tönnies of Gemeinschaft and Gesellschaft and
Henry Maine of status and contract, but the overlaps were striking.
Everywhere in Western Europe during the 1920s and 1930s the rural was
defined in positive terms, while the urban was seen as detrimental to “national
values”. Founded in 1902, the German Garden City Association was active in
proposals for new urban forms that would, among other things, bring people
back closer to the soil in the form of settlements that would mix industry and
agriculture. It also emphasized the creation of satellite cities in anticipation of
some aspects of post-​war planning policy in some of the Western European
countries.
In France, the contributions of Jean-​François Gravier (1947), the deeply
conservative writer who had enjoyed close associations with the Vichy gov-
ernment, came later but his book, Paris et le désert Français, written during the
war, would have a major impact on postwar French planning, even though the
details of his reasoning would be rejected. As per the title of his book, Gravier
railed against the concentration of people and industry in Paris and advocated
a radical decentralization. This was inspired by a view of the big city as deeply
injurious to the nation’s moral fibre. It destroyed families, uprooted people from
the communities that stabilized the social life of the country and lowered the
birth rate to the disadvantage of France as a great power.3
Great Britain shared some similarities with these cases, particularly that
of Germany (Dietz, 2008). It would also be different. From round about the
turn of the century, something called distributism emerges as a set of policy
recommendations (Cox, 2016, pp.  70–​72). This targeted the ills of economic
and political centralization and advocated a return to life in local, relatively self-​
sufficient communities, based on a greater democratization of property own-
ership. Free trade capitalism and the destruction of small-​town life that it had
caused were as anathema as socialism. Of these different strands of anti-​urbanism,
 149

Urban and regional development policy 149

all expressing a strong decentralizing sentiment, it was this that was most apparent
in the US, even when it did not adopt that label (Shapiro, 1972).4
Finally, in considering the issues and climate of opinion within which
we must situate urban and regional policy, the fundamental assumption that
inequality of outcomes was a bad thing, including those of a geographic nature,
owes at least something to the advent of the welfare state and the thinking which
preceded [Link] in turn was giving way to a recognition that state intervention
of a mitigating sort was an entirely legitimate activity. Provision for unemploy-
ment insurance was well underway in the European countries by 1930, even if
highly limited in what it provided. The idea that housing for the poor should
be subsidized in various ways, including through government provision of the
housing itself, had been realized in programs across much of Western Europe,
including in France, Germany, Great Britain and the Netherlands.
As far as the urban question was concerned, an early response would be
the development of city and regional planning as a state practice, but also as a
practice of those developers working on a large scale.5 The large metropolitan
area was a particular challenge, showing all the dysfunctionalities of excessive
agglomeration: pressures on housing, on the ability of firms to expand, on the
environment and, as the city expanded at its edges, moving people around.
It was within this context that notions of reordering space through zoning
controls and planning for expansion via the layout of new highway and rail
lines and the development of new nodes around them would take shape. It
assumed a strongly technical character –​a matter of moving things, including
people, around and arranging them in some rational form –​even while gen-
erating an almost missionary zeal among its practitioners. The city would be
reorganized so as to function more efficiently and, not least, as a competitive
mechanism (Cox, 2016, p. 82). At the same time, and as part and parcel of this,
class antagonisms would be mitigated. The idea of the new town or garden
city was an integral aspect, but so too was aiding the expansion of the city
so as to facilitate some resolution of the housing question. How these ideas
and practices would be later refined and incorporated into what would be
defined as urban and regional development was in the future, but not the least
of their achievements was to legitimize the idea of state intervention in the
space economy.
A second crucial part of the mix would be the rise of property capital: the
separation out from capitalist activity as a whole sector providing premises of
all sorts –​acquiring the land, constructing housing, shopping facilities, offices,
specialized industrial areas and then holding for rent or selling so as to provide
the funds for new rounds of development. During the 19th century and con-
tinuing well into the 20th, the separation was far from clear. A good deal of
housing for industrial workers and miners was owned by the employer. There
was also a lot of self-​build by workers (Harris, 1991; Stovall, 1989): purchasing
land from a landowner, speculative or otherwise, and then building a house
on it, perhaps over a fairly lengthy period of time. Retail outlets and office
buildings tended to be custom-​built and owned outright.
150

150  Kevin Cox
This clearly changed, particularly after the Second World War.6 The cre-
ation of premises, often for rent, as in the case of shopping centres, office
developments or industrial parks, has become a very competitive, speculative
industry and for developers, a clearly identifiable social interest. Developments
are highly leveraged and returns are put back into the business with a view
to further expansion. They are built ahead of any buyer commitment which
requires a constant attention to innovation to attract the customer. Some of this
innovation is also with a view to enhancing land rent through the internaliza-
tion of externalities; this helps to explain the sheer magnitude of many new
developments, including ever larger shopping centres, logistics parks, ambitious
residential developments incorporating lakes, green space and golf courses and
so-​called “mixed-​use” developments. Other changes relate to the specifics of the
buildings. Housing is exemplary in the transformations and changes of fashion
that it has experienced: from one-​car garages to two-​car and upwards; differen-
tiation into housing targeted at retirees versus that for young families and so on.
Likewise with office buildings which are open-​plan and now equipped with
underground garages and their own recreational spaces, as in the office campus
ideal. Meanwhile, the construction of modular units for industry and ware-
housing was at one time a novelty. Without a highly competitive, speculative
impulse, it is hard to imagine these changes.
The emergence of property capital has affected urban and regional devel-
opment in some obvious ways, with land extensive developments giving it a
strong suburbanizing impulse, for a start, but also in some less obvious ones; not
least in providing models that would then be incorporated into state practice. In
the 1930s, in an attempt to lure firms into areas suffering from high unemploy-
ment, the British government created what were called “trading estates”: these
included ready-​built factory units for rent at subsidized prices. This, however,
had already been anticipated by the earlier, private ones, notably the industrial
parks at Slough, west of London and Wythenshawe close to Manchester. In
the US, though an extreme case, local governments would be established by
developers as a tax-​avoidance ruse that consisted of nothing but an industrial
park. The sheer size of new developments has also been factored into develop-
ment policy, sometimes something to be fought, as in the aversion of down-
town retailers in the cities of Western Europe to suburban shopping centers; and
also as something, as in the American case, to be welcomed because of the large
and sudden impetus they can give to local tax bases.
The final condition for development policy has been what I have called “the
new mobility” (Cox, 2016, pp. 93–​97): a necessary one if things were indeed
to be moved around into areas of high unemployment, into new towns on
the edge of metropolitan areas or simply to satisfy the desires of local growth
coalitions. The contrast is with conditions under the first Industrial Revolution
where technologies and the nature of products tended to fix industry in highly
particular places. Most notably, and given reliance on steam power, these were
the coalfields and estuarine locations enjoying access to coal by relatively cheap
steam navigation:7 much cheaper than by rail. Textile industries clustered in
 151

Urban and regional development policy 151

this way, as did iron, steel and ship-​building: all leading sectors of the period.
The second Industrial Revolution would sow the seeds of change, though not
entirely on its own.
First, the rise of electricity as motive power liberated industry from the
coalfields. At the same time, the second Industrial Revolution was accompanied
by the creation of a whole new set of consumption goods industries, durable
and non-​durable. Some of these were downstream from existing ones, like the
increasing differentiation of the clothing industry. Others, like many household
appliances, depended on the extension of the electricity network if they were
to be bought. Meanwhile, there was the succession of new products revolution-
izing personal mobility: the bicycle, the automobile and the street car and tram.
The point about these new industries is they were far more indifferent in their
locations than their predecessors of the first Industrial Revolution. Partly this
was a matter of progressively deskilled labour needs; partly it was that they were
assembly industries that brought together components from many different
locations; in some cases it was a matter of market access, but in that instance,
depending on economies of scale in the industry in question, that could mean
a multiplication of plants in different locations. But for this new mobility to be
a condition for urban and regional policy, one of which was moving them into
what were judged to be desirable directions, either by a national government
or a local growth coalition, there had to be changes in industrial organization.
What makes a big difference here is the emergence of the multi-​locational
firm, either horizontally integrated, with several plants producing the same
thing; or vertically integrated, where plants were organized in supply chains, or
some mix of the two. No single firm with one point of production was going
to move lock, stock and barrel. That would have been highly risky, unless over
a very short distance, as did indeed happen in particular urban areas. But what
was now possible was the creation of a second, a third and more plants else-
where, perhaps in accord with calculations of shifting market advantages. This
would then give way to a phalanx of plants and shifts in capacity among them,
even to the closure of some, along with the opening of new ones. This would
in turn lay the basis for new spatial divisions of labour among places. Larger
cities might lose their production functions and even their headquarter, organ-
izing functions, to the extent that advantage in that direction lay in the cap-
ital city, or major regional centres like New York, Chicago and San Francisco.
So if through the branch plant, urban and regional policy had the weapon it
needed  –​moving them around in accord with policy goals  –​that self-​same
mobility could also challenge the economic bases of cities, re-​ posing the
regional question in new ways.
This is to focus on one aspect of the new mobility. It also extended to
workers and households. Successive revolutions in personal mobility –​the tram,
the street car and the omnibus, the automobile –​revolutionized the possibil-
ities of residential choice and had major impacts on the emerging development
industry. It allowed greater expansion at the periphery, creating some resolution
of the diseconomies of agglomeration, notably congestion and the concomitant
152

152  Kevin Cox
concentration of activities that results in increasing rents. This applied both to
employees seeking cheaper housing and to employers keen to expand and take
advantage of the larger layouts that electric power facilitated. The progressive
formation of metropolitan housing and labour markets then provided the scope
for the newer, larger real estate developments: the Levittowns, the industrial and
office parks and the regional shopping centres would follow.
But viewing the new mobility as a whole, the locational discretion of firms
as well as of households was also to some degree problematic. Not all firms or
types of activity were mobile to the same degree. Those depending on urban
rents, whether landlords or developers, could be particularly vulnerable to shifts
in their market as major employers closed plants and opened them elsewhere.
This could be a dilemma shared by the local working class to the extent that
people were, for whatever reason, locked into local labour markets under-
going a contraction as firms shifted operations elsewhere or simply went out
of business.
It is this tension between mobility and a relative fixity which lies behind
anxieties about local and regional economies and which is the sine qua non for
any notion of local and regional development policy. Harvey (1985, p.  203)
outlined one solution:

The more corporations used their powers of dispersal (…) the less urban
regions competed with each other on the basis of their industrial mix and
the more they were forced to compete in terms of the attractions they had
to offer to corporate investment as labor and commodity markets and as
bundles of physical and social assets that corporations could exploit to their
own advantage.

But in fact, this turns out to be a very country-​specific approach: that of the

US. In Western Europe it would be different. The problem of mitigating local
economic disadvantage is taken in hand by central government: a much more
top-​down approach, therefore. But in both instances, and as Harvey implied,
the new mobility would turn out to be not just a curse but also an opportunity.

National difference and national conjunctures

Accordingly, the way in which it would be exploited in development policy
would vary to a very considerable degree. In the US, the bits and pieces –​parts
of the divisions of labour of firms, and quickly defined as up for grabs –​would
become objects for local growth coalitions to attract in, or ignore, depending
on how they fitted into some local vision. In Western Europe, on the other
hand, the rise of the branch plant turned out to be something that could be
used by central governments as they sought to alter the distribution of industry
for their own purposes, which they did.8
Indeed, in the second half of the century, the opportunities for mobilizing
the footloose for purposes of local economic development would expand
 153

Urban and regional development policy 153

significantly beyond the industrial branch plant. Large corporations with mul-
tiple branches would come to be a major feature of developed economies in
both manufacturing and in the so-​called service industries:  bank branches,
hotel chains, back offices, retail chains and the shopping centres that serve as
their characteristic sites. Retail chains then entailed the creation of warehouses
and distribution centres, adding to the bits and pieces that could be the object
of location policy in the Western European case or of policies of attracting
inward investment in the American one. The expansion of the state increased
the possibilities through its multiplication of military, air and naval bases and the
hiving-​off of routine office work from capital cities for relocation elsewhere.
Eventually new residential developments, particularly the larger ones, would
become objects of strategic planning for local and regional development.
But it bears repeating that how such mobility would be exploited for the
purposes of advancing development agendas would vary considerably between
the US and the Western European countries. In the US they would be the
object of local, quite narrowly defined programs aimed at the enhancement
of rents from property, which would then be subject to forces of a market-​led
nature and a competition between localities:  for the population growth and
housing demand promised by investments in office or industrial employment;
and then a competition between the different local governments in a metro-
politan area to land the housing developments that would result –​a compe-
tition typically spearheaded by the landowners, the property companies and
local government itself. Planners would play a role but they would be in the
service of local government and local government in turn would find itself
subordinated to the pressures of local growth interests.
In Western Europe, it would be different. The various mobilities would be
drawn on by central governments as part of creating what can only be described
as “the good geography”: a set of roughly interlocking goals, reflecting national
debates about growth and distribution. There would be attention to questions
of equity over space as we have already seen in discussion of policies aimed
at redistributing employment. The national geography would also emerge
as a consciously conceived and planned-​for productive force as in specific-
ally national port/​airport/​transport policies, ensuring that housing was located
where it was most needed, and in attempts to create the compact city. Here, the
order of priorities would be reversed: central government rather than local, and
planners, particularly those at the centre, in a dominant position.
In Western Europe, there was considerable convergence in policy (Romus,
1958), most notably in directing employment towards areas of relatively
high unemployment, but also those deemed to be backward: as in Southern
Italy, Western France, the Scottish Highlands and Islands, and the far norths
of Norway and Sweden. Areas were designated and the state used financial
incentives to persuade firms to locate there and sometimes, just to remain. In
some instances, it was a matter of the investment decisions of firms in public
ownership. There again areas believed to be under severe housing and labour
market pressure were subject to controls over firm expansion. This applied to
154

154  Kevin Cox
much of the Midlands and Southern England, and as far as office employment
was concerned, to the Paris basin.
Other aspects of policy, including the dispersal of government offices from
capital cities to provincial towns in areas of high unemployment were more
variable from country to country:  France, Ireland, the Netherlands and the
United Kingdom all made use of this (Cox, 2016, p. 148). New towns were
common in France and the United Kingdom with other cases in Ireland, the
Netherlands and [Link] were in part designed to mitigate diseconomies
of agglomeration by moving people and employment out of major cities. The
new towns surrounding London, Birmingham, Manchester and Paris provide
examples. But in the old mining areas of South Wales, Northeast England and
the Pas de Calais/​Nord region of France, the idea was to concentrate labour
reserves in areas where population was notoriously dispersed over many small
colliery villages (Hudson, 1982). Again, in the Netherlands, new towns had a
different origin in the draining of the Zuyder Zee but, concomitantly, pro-
viding some relief for nearby Amsterdam.
Urban and regional development policy in the US has been quite different;
in fact, the “regional” part of it barely exists at all. Rather it is cities that take the
lead in promoting development, largely by attracting inward investment from
outside. Accordingly, they compete with each other by offering various finan-
cial incentives and infrastructural [Link] individual states are auxil-
iary to this in virtue of their responsibilities for not only their own taxation but
also large swathes of labour [Link] include the power to ban the closed shop
and so make the organization of workplaces more difficult, as well as control
of compensation for injury on the job: an important consideration that weighs
on firms in the form of insurance premia. State tax policy and labour law are
important contributions to what is called a state’s business climate and some-
thing over which local governments bring strong pressures to bear.
The undergirding structure governing urban development and, derivatively,
regional development, is what might be called a market in locations: locations
are “sold” to firms who play one local government off against another, seeking
the best deal. The competition in 2018 for the second Amazon headquarters,
promising 50,000 jobs to the “winning” city is indicative of how this works.
What is at stake is not so much development, even though that is the way in
which these practices will be sold to the public –​as growth. Significantly one
talks about “growth machines” and they are a crucial actor in urban develop-
ment policy:  something for which equivalents in Western Europe are much
harder to find.9
The growth in question is first and foremost in the demand for the products
of local developers; homebuyers for housing, tenants for shopping centres and
office parks, and conventioneers and tourists for hotels. Increasing demand via,
most obviously, an increase in population, is absolutely crucial for developers
in the American context. This is because of their heightened dependence on
demand in a particular locality. Developers tend to be locked into a particular
place in virtue of connections with local banks, builders and those “in the
 155

Urban and regional development policy 155

know” about what is about to happen: most importantly, the local politicians
who make the decisions about new freeway interchanges, re-​zonings of land
and major capital projects which promise to alter accessibility relations and so
the prospect of extracting rent. This is peculiarly American. The developers in
one city will be different from those in another: testimony to the unique qual-
ities of particular local markets; something connected, in turn, to the remark-
able institutional differences among US cities (Cox, 2018).
Historically they were joined by other agents, some of which remain
important. As a result of the anti-​trust legislation at the beginning of the last
century, banking was a notoriously decentralized industry. Branching out across
county and state lines was forbidden until the rules were relaxed in the 1970s.
This meant that banks were extraordinarily dependent on local markets: growth
in deposits and loan demand in the county where they were based was funda-
mental to their fortunes. Involvement in local development policy and politics
is now largely over for them, but not so the gas and electric utilities where there
were similar restrictions on inter-​state mergers. Rather they had unique service
areas, so that growth in demand there was fundamental to revenues. They are
still significant players: gatekeepers for inward investment in industry since any
owner of an industrial site will lodge the information with them. They control
the process with developers participating, and local government last of all (Cox
and Wood,  1997).
The stake of local government is largely fiscal. In the US local governments
are highly dependent for their revenues on local taxation: largely property and
sales taxes –​hence the interest in attracting large regional shopping centres –​
and increasingly, local income taxes. A lot rides on this: not just revenue but
also the ability to raise money for capital projects. This occurs through the sale
of what are called municipal bonds. But these carry ratings that determine the
rate of interest, and the rating depends significantly on the growth of the local
tax base (Cox, 2010, pp. 221 and 223).
Interestingly, until the 1970s, this worked quite well from the standpoint
of redistributing employment to relatively backward areas, most notably the
Southern states (Cox, 2016, pp.  154–​155), promoting a degree of regional
convergence, as in Western Europe. There were no new towns in the Western
European sense. Rather the answer to diseconomies of agglomeration in the
US has been the private one of massive urban sprawl and latterly the edge
city. There are “new towns” but they are simply coagulations of new develop-
ment around small towns slowly incorporated into a wider metropolitan area,
and perhaps the basis for a new nucleus of development around some massive
mixed-​use project for which it is hard to find parallels in Western Europe.
Since the 1970s, there have clearly been changes in both the US and Western
Europe. In the US policy largely continues much as it was: bottom-​up, a market
in locations. But now its effectiveness in countering uneven development, even
while it was never one of its goals, has markedly [Link] branch plants
that were the targets of many smaller towns are now going to Mexico or China
or have closed down in the face of heightened international competition. On
156

156  Kevin Cox
the other hand, the possibilities of attracting some high-​tech development with
demands for elevated labour skills are quite poor: the skills demands of mass-​
assembly branch plants were modest to say the least (Veltz, 2017).
In Western Europe, there has been a change in policy, though there too,
regional convergence has tended to go into reverse. The areas once eligible
for financial assistance to attract new employment have greatly contracted,
partly under pressure from an EU that wants to promote development in the
more backward areas of Eastern and Southern Europe, and partly, at least in
the United Kingdom, on ideological grounds and the view that market forces
should prevail. Meanwhile, there have been no new towns in Western Europe
since the late 1960s, though there is now talk in both France and the United
Kingdom of a new generation of them. On the other hand, there is also the
suggestion that the development of the single market has nudged policy more
in the direction of the American model: a more “bottom-​up” approach where
the member governments compete for inward investment. Firms now locate
with respect to the EU as a whole rather than with respect to individual coun-
tries. At the same time, the inclusion of the Eastern European countries has had
a dramatic effect on the range of wage rates within the EU. Both of these factors
have induced some attention to the “business climate”, akin to the practice of
the US states: notably national labour law and corporate taxation, but also the
willingness to provide the multinationals with various sweeteners. But while in
the US this is accepted as normal practice there has been greater criticism in
the EU, particularly in Western Europe where anxieties about the relocation of
firms to Eastern Europe has led to charges of what is called “social dumping”.
Even so, there has been continuity, and in particular in the central coordin-
ation of policy. There is a view that powers over new development have been
decentralized: a tendency in the American direction, and not just to the member
states of the EU as discussed above but also to local governments. This has been
exaggerated (Le Galès and Harding, 1998; Cox, 2009; Cox, 2016, pp. 171–​187).
Rather what is observable is a tendency on the part of central governments to
discriminate between those regions seen as able to compete more effectively
in an increasingly global market, and those less able. State money is targeted at
infrastructural improvements in the former and most notably the bigger metro-
politan areas:  the emergent post-​industrial cities. In some instances, this has
meant directing money to capital cities (Crouch and Le Galès, 2012). If there
is to be regional convergence in this policy-​making environment, then it will
be largely, though not exclusively, via trickle down effects: “not exclusively”,
since there have been other central government initiatives designed to pro-
mote development in more equalizing ways, as in the Urban Development
Corporations and enterprise zones of the Thatcher era (Anderson, 1990) and
France’s poles of competitivity (Ancien, 2005). This does not mean that local
and indeed regional governments have not become more involved in local eco-
nomic development and often in ways that depart from the simple competition
for inward investment: where there is a strong local IT presence, for example,
local governments seeking out ways of facilitating their expansion.10
 157

Urban and regional development policy 157

In short, while there has been some shift in the geographic division of state
labour in Western Europe as it pertains to urban and regional development, the
contrast with American practice remains a sharp one, which raises the question
of exactly why that might be. A first and obvious place to look is the territorial
structure of the state (Cox, 2004; Cox, 2016, pp. 295–​318). The contrasts are as
striking as the differences in approach to urban and regional policy that they
inform. The American state is a highly decentralized one. It is a radical feder-
ation in which the states have very considerable powers, including, as we have
seen, some aspects of labour law. The states have then seen fit to delegate many
of their powers to local governments, including land use policy. From a devel-
oper standpoint the functions that local governments can and do perform are
enviable. They include the power to:

• grant tax abatements for new developments.

• regulate land use without any superordinate supervision. Appeals to a
superior body from those adversely affected by a development are not
possible.
• raise money for public works from the sale of municipal bonds: for airports
and new highways with the potential to alter accessibility advantages and
so enhance the possibilities of new sources of land rent.
• annex land that is unincorporated. This is particularly important given the
interest of developers in land-​extensive projects like office parks, logistics
centres and regional shopping centres.

The highly fragmented nature of the state creates other possibilities for local
interests, whether organized as coalitions or even as individuals looking for
a favour. For the bottom-​up nature of development policy also extends to
relations with state and federal governments. They are a primary source
of favours, regulatory and monetary, and, in virtue of the structure of the
American state, highly vulnerable to entreaties from below (Arnold, 1981). In
part, this rides on the committee system. No piece of legislation can make
it to consideration in houses or senates unless it has been discussed and typ-
ically amended in the appropriate committee. By and large representatives
get to sit on committees relevant to their constituencies:  so congresspersons
and senators representing New  York and North Carolina are prominent on
finance and banking committees, given the importance of financial services
to respective major metropolitan centres: New York and Charlotte. This puts
them in a prime position to block legislation or seek amendments favourable
to the industry in their home bases. Should the legislation move forward for a
vote, then the weakness of party discipline means that, through horse trading
and the creation of coalitions that cross party lines, it stands a good chance of
[Link] is made possible by the radical separation of legislative from execu-
tive branches. Failure to pass a bill can never put a government in jeopardy as
it does under parliamentary systems, where it can create a crisis of confidence
and a government’s resignation. In fact, legislation can only be introduced by
158

158  Kevin Cox
a legislator, not by the government, and then has to be subject to committee
meetings. The same applies to the operations of the individual states, and with
similar implications for the politics of local and regional development (Weir,
1996). Local, and to a lesser degree, state government is utterly complicit in this
because of their own fiscal stakes. There are clearly government programs that
channel money from federal and state branches respectively. But with the pos-
sible exception of education, they have little equalizing effect. And the fact that
so much social policy is delegated or at least shared with the states, means that
inequality gets built in to the geography of America’s welfare state.
This is a lengthy explanation but important since misunderstanding of how the
American state functions is widespread. It is very different from the centralized
states of Western Europe, and even from the federal states of Belgium, Germany,
Spain and Switzerland which, compared with the American federation, are
quite weak, and where the separation of legislative from executive branches
does not exist. In the unitary states, local governments are subordinated to cen-
tral government in so many different ways: land use planning is coordinated
from the top, and funding of the welfare state has a strong equalizing effect
across local governments. Local governments have limited incentive to pro-
mote local development and when they do, central control of land use planning
limits their scope. The conjunction of legislative and executive branches then
translates into strong party discipline with scant possibility for the expression
of territorial interests in the way so characteristic of the American case. The
federal states are somewhat different and there is some scope for the individual
provinces to develop policies of their own. On the other hand, their powers
are much more limited than that of the American states and the vulnerability
of governments to dismissal through a vote of no confidence puts pressure on
representatives to toe the party line rather than that of the home district. And
in the German and Spanish cases, there is fiscal redistribution from the more
affluent provinces to the poorer ones.
Yet, and clearly, state structures do not exist independently of social
formations. The United States and the countries of Western Europe are cap-
italist states, but, in virtue of very different histories, the values, the dominant
discourses, the inherited structures, can vary and help condition the formation
and ongoing reproduction of, accordingly, different state forms. For a start, the
relation between state and economy is quite different. In the US the subor-
dination of the state to the forces of a capitalist economy is striking. This is
most dramatically on show in the way in which the state, through its structures
of relations, mimics the market. The market in locations through which local
governments compete for inward investment is one aspect of this; likewise, the
way in which state policy, via the committee system and weak parties, is seen
as an adjunct to competitive advantage. This in turn is underlain by a popular
embrace of the virtues of the market and a suspicion of the state as a mon-
opoly provider; accordingly, state powers are to be limited, except when they
involve grants to local governments for local development purposes or some
regulatory indulgence. Meanwhile, the mosaic of welfare states that feeds into
 159

Urban and regional development policy 159

business climate competition is celebrated on the grounds of expanding choice.
In the countries of Western Europe, it can seem to be the other way around.
Competition among local governments is more likely to be frowned on. Limits
are placed on the ability of local governments to grant tax concessions to new
developments. The competition of smaller airports for low cost airlines was
placed under severe limits by the EU (Cox, 2016, pp. 258–​260). Central gov-
ernment is free of the opprobrium of “big government”.
In part, it is a matter of a very different balance of class relations. In terms
of the subsumption of labour to capital, a process in which the working class
sees its interests as coincidental with those of employers –​i.e. profit and expan-
sion –​the US is sui generis. The labour movement is organizationally extremely
weak: there is no socialist representation in party politics and union member-
ship is relatively low. Labour unions are seemingly pragmatic bodies pursuing
an interest group approach to politics, which fits nicely into the dominant view
of what competitive politics should look like. They lack the ideological bite of
their Western European counterparts, even while that bite is no longer what it
was. This is particularly clear in the relative weakness of the welfare state and,
as an aspect of it, any serious spatial policy aimed at mitigating geographically
uneven development.
Crucial to understanding these differences are different histories. American
capitalism developed in and across a space uncluttered by pre-​ capitalist
residues: no peasantries, landed gentries or ancien regimes with limited interest
in buying and selling. In Western Europe it was different. A nascent capitalism
encountered resistances born of framings of social life that did not reduce it to
market relations and the unbridled conversion of everything into commodities.
Rather there were very different views of the respective claims of individual
and community, and of civil society and the state. Competition and its effects
on the community would be seen more critically and the state as a means of
limiting as well as facilitating capital. The two different aspects of the com-
modity form, so central to capitalism, would be given different emphases: in the
US exchange value could come to dominate use value, but not nearly so much
in Western Europe. There was, rather, a moral economy11 (Thompson, 1971) of
the form absent in the US, which fortified resistance (Calhoun, 2012) A cap-
italist economy is seen as having dysfunctionalities that need to be countered,
not least through the land use planning mechanism and a state-​orchestrated
reordering of the space economy.
There is also the question of inherited state forms. Certainly in the US,
the creation of the radical federal structure in 1789 had very little to do with
the class relations of capitalism because, for the most part, it pre-​dated them.
Despite this contingency of origins, it would become a condition for capitalist
hegemony and express the claims of the superiority of markets and of compe-
tition, albeit to the disadvantage of the labour movement. The formation of a
national welfare state would be frustrated by the competitive anxieties of the
different states; something with which the capitalist class as a whole could easily
ally itself (Farhang and Katznelson, 2005; Weir, 2005).
160

160  Kevin Cox
One can make analogous claims for the centralized states of Western Europe.
The absolutist state is antecedent to capitalist social relations, even while it can
be regarded as an important condition for them. Its centralizing institutions
would accordingly be the stage for class struggle: that was where power resided
and continues to reside. That the (central) state had a monopoly of power was
regarded by the labour movement as an advantage. This was because it allowed
a uniformity of labour law that would work against capitalist competition in
terms of wages and conditions of work. Even for those not inclined to vote for
parties of the left, the central state has been seen as an important counterweight
to the damaging effects of market relations: testimony to a different view of the
meaning of capitalism.

Concluding comments
The emergence of urban and regional development as a state preoccupation
after the Second World War has first to be situated with respect to the contra-
dictory development of capitalism and how that was projected onto a spatial
plane: regional questions of uneven development and urban questions of dis-
economies of agglomeration and its social implications. It would take some
time before it was defined and practiced by the state but the preconditions for
particular forms of intervention were emerging some time before:  city and
regional planning, property capital and its innovative forms of development
and a new mobility of production. How these bits and pieces would be put
together in policy form would vary. In Western Europe there have been strong
tendencies to top-​down planning:  notably attempts to displace employment
in the form, largely, of branch plants, to areas of relatively high unemployment
or sheer lagging development; and the relief of large cities by the construction
of new towns. In the US, instead, there was reliance on a market form: local
governments competed for the self-​same branch plants and local property
development assumed a dominant role instead of its more subordinate one in
Western Europe.
These differences are owing in the first place to very different state structures.
The decentralized state forms of the US are an important precondition for the
market in locations that is the form assumed by urban and regional develop-
ment there. The much more centralized states of Western Europe on the other
hand, have been a necessary precondition for top-​down policies. Beyond that,
state forms have to be seen as aspects of wider social formations which lend
discursive underpinnings to highly diverse policy approaches.

Notes
1 For a more detailed elaboration on these points, see Cox (2016, ­chapter 2).
2 Compare Goodman (1979).
3 Marchand and Cavin (2007) have discussed Gravier’s arguments, as well as outlining
the similar views expressed by the influential Swiss planner Armin Meili.
 161

Urban and regional development policy 161

4 See Woods (1939) for an example of this genre.
5 Whole cities, or segments of cities, have been planned by private developers: Irvine
CA in the first instance, and Edgbaston, the Birmingham suburb on the other; aka
the Calthorpe Estate (Cannadine, 1977). In some instances the private planning
came before public and provided templates of good practice.
6 Writing in 1975 about Great Britain, Ambrose and Colenutt (1975) recognized the
relative recency of this, arguing that “[b]‌efore the Second World War there were very
few large property companies. Instead, redevelopment was undertaken by individual
entrepreneurs, or by firms developing buildings for their own occupation. There
were few speculative office blocks and the tendency was for firms to own freeholds
to their premises rather than rent. Certainly there was no public awareness of
‘developers’ as such. Few property companies were quoted on the Stock Exchange,
there was virtually no property press dealing with commercial development, and
there was no coherent organization lobbying for the developers.” (p. 37).
7 For the influence of coal mining geography on urbanization in Europe as a whole,
in large part through the steam engine and its implications for the location of the
iron and steel industry, see Fernihough and O’Rourke (2014).
8 On the significance of the branch plant to urban and regional policy, see Heim
(1983).
9 See Cox (2017) for a critical discussion of some aspects of the growth machine idea.
10 For one example of this see While, Gibbs and Jonas (2004).
11 A social form in which economic life is not separate from moral judgment but
where economic practice inevitably has strong associations of what is fair and just,
and what isn’t.

References
Ambrose, P. and Colenutt. B., 1975. The property machine. Harmondsworth Middlesex:
Penguin.
Ancien, D., 2005. Local and regional development policy in France:  Of changing
conditions and forms, and enduring state centrality. Space and Polity, 9(3), pp. 217–​
236. [Link]
Anderson, J., 1990. The ‘new right’. Enterprise zones and urban development
corporations. International Journal of Urban and Regional Research, 14(3), pp. 468–​489.
[Link]
Arnold, R. D., [Link] local roots of domestic policy. In T. E. Mann and N. J. Ornstein,
eds. The new Congress. Washington D.C: The American Enterprise Institute. Ch. 8.
Calhoun, C., ed. 2012. The roots of radicalism: Tradition, the public sphere and early 19th-​
century social movements. Chicago: University of Chicago Press.
Cannadine, D., [Link] cities: How different? Social History, 2(4), pp. 457–​480.
Cox, K. R., [Link] politics of local and regional development, the difference the state
makes and the US/​British Contrast. In: A. Wood and D. Valler, eds. 2004. Governing
local and regional economies. Aldershot: Ashgate. Ch. 10.
Cox, K. R., 2009. ‘Rescaling the state’ in question. Cambridge Journal of Regions, Economy
and Society, 2(1), pp. 107–​121.
Cox, K. R., 2010. The problem of metropolitan governance and the politics of scale.
Regional Studies, 44(2), pp. 215–​227. [Link]
Cox, K. R., 2016. The politics of urban and regional development and the American exception.
Syracuse NY: Syracuse University Press.
162

162 Kevin Cox

Cox, K. R., 2017. Revisiting the city as a growth machine. Cambridge Journal of Regions,
Economy and Society, 10(3), pp. 391–405. [Link]
​ ​ ​ ​
Cox, K. R., 2018. American cities and their institutions. Unfashionable geographies.
Available at [Link]
​ ​ ​ ​ ​ ​ ​ ​
their-institutions.
​ ​
Cox, K. R. and Wood, A. M., 1997. Competition and cooperation in mediating the
global:  The case of local economic development. Competition and Change, 2(1),
pp.  65–94. ​
Crouch, C. and Le Galès, P., 2012. Cities as national champions. Journal of European Public
Policy, 19(3), pp. 405–419.
Dietz, B., 2008. Countryside-versus-city in European thought:  German and British
anti-urbanism between the wars. The European Legacy: Toward New Paradigms, 13(7),
pp. 801–814. ​
Farhang, S. and Katznelson, I., 2005. The Southern imposition: Congress and labor in
the New Deal and Fair Deal. Studies in American Political Development 19(1), pp. 1–30.
Fernihough,A. and O’Rourke, K. H., 2014. Coal and the European Industrial Revolution.
Discussion Papers in Economic and Social History, Oxford University, No. 124.
Goodman, R., 1979. The last entrepreneurs. New York: Simon and Schuster.
Gravier, J.-F., 1947. Paris et le désert Français. Paris: Portulan.
Harris, R., 1991. Self-building in the urban housing market. Economic Geography, 67(1),
pp. 1–21. ​ https:\\[Link]\10.2307/143633
​
Harvey, D., 1982. The limits to capital. Oxford: Blackwell.
Harvey, D., 1985. The urbanization of capital. Oxford: Blackwell.
Heim, C. E., 1983. Industrial organization and regional development in interwar Britain.
Journal of Economic History, 43(4), pp. 931–952.
Hudson, R., 1982. Accumulation, spatial policies, and the production of regional labor
reserves: A study of Washington New Town. Environment and Planning A, 14(5), pp.
665–680.
​
Koistinen, D., 2013. Confronting decline: The political economy of deindustrialization in 20th-
century New England. Gainesville, FL: University Press of Florida.
Le Galès, P. and Harding, A., 1998. Cities and states in Europe. West European Politics,
21(3), pp. 120–144. https:\\[Link]\10.1080/01402389808425260
​
Marchand, B. and Cavin, J. S., 2007. Anti-urban ideologies and planning in France
and Switzerland:  Jean-François
​ Gravier and Armin Meili. Planning Perspectives, 22,
pp.  29–53. ​
Preteceille, E., 1976. Urban planning:  The contradictions of capitalist urbanization.
Antipode, 8(1), pp. 69–76.
Romus, P., 1958. Expansion économique régionale et communauté Européenne. Leyde: Sythoff.
Shapiro, E. S., 1972. Decentralist intellectuals and the New Deal. The Journal of American
History, 58(4), pp. 938–957. [Link]
​ ​ ​
Stovall, T., 1989. French communism and suburban development: The rise of the Paris
red belt. Journal of Contemporary History, 24(3), pp. 437–460.
Thompson, E. P., 1971. The moral economy of the English crowd in the eighteenth
century. Past and Present, No. 50. Pp. 76–136.
Veltz, P., 2017. Introduction: Regions and territories: Evolutions and changes. Economie
et Statistique. 494-495-496.
​ ​ Pp.  5–17.
​
Weeks, J., 1981. Capital and exploitation. London: Edward Arnold.
Weir, M., 1996. Central cities’ loss of power in state politics. Cityscape, 2(2), pp. 23–40.
 163

Urban and regional development policy 163

Weir, M., 2005. States, race, and the decline of New Deal liberalism. Studies in American
Political Development, 19(2), pp. 157–​172.
While, A., Gibbs, D. C. and Jonas, A. E. G., 2004. Unblocking the city? Growth pressures,
collective provision, and the search for new spaces of governance in Greater
Cambridge. Environment and Planning A, 36, pp. 279–​304. [Link]
a3615
Woods, R., 1939. America reborn: A plan for decentralization of industry. New York: Longman,
Green and Co.
164

9 
Spatial-​economic development
The effect of urbanization on education
in China, 1890–​present1
Meimei Wang and Bas van Leeuwen

Introduction
Human capital is widely recognized as an important driver of regional economic
development; therefore, it is instructive to investigate the spatial differences in
schooling and learning (Koscinska and Herbst, this volume). Urbanisation is
a major spatial component affecting education. Because cities are generally
centres of learning and educational modernisation, urban education is prom-
inent in a wide range of academic debates, including those concerning urban
economic growth and the job market, the supply of essential social services in
cities, urban–​rural relations and migration.
In this chapter, we discuss the relationship between urban–​ rural spatial
differences and education in China. The Chinese case is particularly interesting
for several reasons. First, urbanisation has historically drawn significant govern-
ment attention in China because these urban areas are considered to have a
mainly administrative role (in contrast to Europe, where urban areas have many
additional functions), and also to act as drivers of economic modernisation and
development. Second, China’s education system was dominated for centuries
by a civil examination scheme having a curriculum based exclusively on the
humanities. Only since the late 19th century have more practical topics such
as mathematics, technology, languages, economics and physics been included.
Cities were the frontrunners in this educational reform process. Third, China
experienced a political period (c. 1949–​1979) during which the government-​
regulated relationship between urbanisation and education was reversed, and
educated people moved to the countryside. Finally, the remarkable economic
and urban growth experienced by China in recent decades has led to a return
to the traditional, pre-​1949 link between urbanisation and education.
In this chapter, the abovementioned massive changes that occurred over
time and across urban and rural regions are related directly to the theme of
this volume—​the multidisciplinary connection between historical and regional
studies. This chapter further elaborates on this topic by exploring urbanisation
(as a driver of knowledge-​based development of cities and regions) from the
angles of history and of the social sciences relevant to regional studies (e.g.
sociology and economics). However, the roles of these multiple fields vary
 165

Spatial-economic development in China 165

considerably by time [Link] analyse these issues, in the following section we
discuss the construct of urbanisation. Sections 3–​5 discuss Republican China
(pre-1949), the first phase of the New China period (1949–​1979) and the
Reform Period (1979–​present). We end with a brief conclusion.

Quantifying urbanisation

The historical dimension

Numerous studies have attempted to quantify urbanisation, most commonly
by defining cities as settlements with a certain number of persons (e.g. 5,000,
10,000 or 40,000) living together. Quantification has been related partly to a
desire to properly define the concept of urbanisation, and partly to the goal of
creating a historically and/​or geographically comparable indicator. However,
this type of definition goes beyond mere academic interest because (as we
will show below) it affects individuals directly in terms of allowing migration,
offering job opportunities and supplying social care, among others.
Examples of key scholars involved in quantifying urbanisation include
Chandler (1987), a historian; Modelski (2003), a political scientist and Bairoch
(1988), an economic historian. Chandler used censuses, travel reports, and other
direct and indirect information, including cities larger than approximately
20,000 before 1850 and cities larger than 40,000 after that year. Modelski
(2003) updated Chandler’s work by including more data and applying theories
to predict missing population counts for cities. One such theory was Zipf ’s law
which, when applied to cities, states that city sizes in a country will follow a cer-
tain “Zipf distribution”, thus allowing scholars to predict city sizes.2 Modelski
focused on cities with more than 10,000 (ancient times) or more than 1,000,000
inhabitants (modern times). Similar to Chandler and Modelski, Bairoch (1988)
used various sources, but took a more eclectic approach as, “when the data did
not exist, he found a way to collect or construct it” (Brezis, 2008). Bairoch
constructed data for a series of cities of 5,000 or more inhabitants for the
period before 1850. Presently, urbanisation data are published by the United
Nations as World Urbanization Prospects. In these publications, the definition of a
city is highly detailed and varies significantly across countries. In general, how-
ever, population clusters above 2,000 to 100,000 persons are considered cities,
depending on the country.
The above discussion on the population sizes of cities shows clearly that
urbanisation is viewed differently according to the region, field and time
period. Arguably, Paul Bairoch has been the most influential scholar in this
field. His work has been modified numerous times for application in various
countries (see e.g. Malanima, 2009), including China. Although numerous such
studies have been performed with regard to China (Skinner, 1977; Cao, 2000a,
2000b), most have been conducted by economic historians for the pre-​1949
period. Identifying the cities was a major problem faced by these studies. Yet
most scholars are agreed on the definition of a city as 2,000 or more inhabitants
166

166  Meimei Wang and Bas van Leeuwen

for the pre-​1900 period. In a recent study, Xu et al. (2018) provided trends by
region from AD 1100 to AD 1900. Based on the above definition, the urban-
isation ratio of Inner China remained stable from 1776 to 1893, at around 7
per cent. During the early stages of the 20th century, the total urbanisation in
China increased rapidly, approaching 13 per cent in 1918. Growth was also
rapid after 1918:  for cities with 2.500 or more inhabitants, the urbanisation
ratio was 34 per cent at the end of the 1920s (Xu, 1930), 32 per cent in 1933
(Ouyang, 1936) and 28 per cent in approximately 1936 according to the China
Yearbook (Sun, 1946).

Urbanisation in New China
Sociologists, in particular, objected to the approach of using city size as a measure
of urbanisation because many people who were counted in the urban popula-
tion were actually farmers. According to an official investigation of cities and
towns along main railway lines in the 1930s (with the exception of Chongqing
city and Guiyang city), farmers accounted for more than 70 per cent of the
urban population in all cities in the provinces of Guizhou and Sichuan. Even
in Guiyang city, the capital of Guizhou Province, more than 60 per cent of the
population consisted of peasants (Yin and Li, 2009).
This criticism of Republican urbanization rates led, after the establishment
of New China in 1949, to a change in the statistical definition of urbanisa-
tion to exclude the farming population. After 1958, the Hukou (household
registration) system divided the entire population into agricultural and non-​
agricultural [Link] who held farmland were designated as agricultural
and excluded from certain social benefits; e.g. they could not migrate to urban
areas unless they had an urban spouse. Hence, prior to 1978–​1982 when this
system was relaxed, the strict Hukou system and the existence of communes
almost completely separated the agricultural and urban populations, causing
the officially reported urbanisation ratio to be much lower than in previous
decades. Nevertheless, from 1949 to 1960, due to the demand for labour during
industrialization, the share of the non-​agricultural population increased quickly
and drove the urbanization ratio (defined at that time as the share of persons
with an urban Hukou) from 10.6 per cent (Bai, 2003) to 20 per cent (Zhao,
1988). However, this upward trend was interrupted by the Great Famine that
struck around 1960. The government was forced to reduce the urban popula-
tion because of the lack of food, and many labourers who had come to the cities
from rural areas were sent back. Hence, in 1966, the urbanization rate was only
17.9 per cent. The share of the urban population further stagnated when, from
the second half of the 1960s to the early 1970s, the government implemented
a strategy for the development of the hinterlands, for national security reasons.
Many urban students and workers were sent to rural areas to perform labour
(Wang, 2001). These two factors, combined with lower fertility rates in the
cities caused by the one-​child policy, led to the urbanization rate declining
slightly to 17.6 per cent in 1977.
 167

Spatial-economic development in China 167

After 1978, urbanisation resumed its upward trend, which has continued to
the present. This occurred because, first, the institutional barriers such as the
Hukou system became more flexible, reopening the possibility of rural–​urban
migration. Stimulated by more and better job opportunities created by the fast-​
growing urban economy, there was even an absolute decline in the rural popu-
lation. A second reason for this resumption of urban growth was a change in the
official definition of the urban population. Since the third population census
in 1982, and as was the case before 1949, all people living in urban areas—​
including farmers—​have again been defined as part of the urban population.
The definition of urban areas remains based on administrative divisions, but the
urban divisions have continuously expanded (National Bureau of Statistics of
China, 2008; see also Table 9.1).
The return to the practice of including the agricultural population living in
cities in the urbanisation statistics has met with the same criticism as it did in
the Republican Period, with multiple scholars arguing that rural labourers who
migrate to urban areas should be counted as part of the agricultural popula-
tion.3 These scholars considered the officially reported urbanisation of nearly
60 per cent in 2017 to be nominal urbanisation, with real urbanisation being
no more than 45 per cent.4

Table 9.1 Urbanisation ratios in New China

Note Urbanisation ratio

1776 Population >2,000; including farmers living in city 7%

1851 Idem 7%
1918 Idem 13%
1933 Population >2,500; including farmers living in city 32%
1958 Population > 2,500. Only those who had non-​ 16.2%
agricultural Hukou included
1966 Idem 17.9%
1977 Idem 17.6%
1986 All residents living in political centres of governments at 24.5%
all levels or other independent areas such as industrial
and mining areas, science parks, research institutions,
higher education facilities, the headquarters of farms
and forest farms with more than 3,000 residents;
including farmers living in city
1995 Idem 29.0%
2001 Idem, with as additional criterium a population density of 37.7%
1,500 or above
2009 Idem as in 2001 with the addition that now also towns 46.6%
and suburban regions with links to cities are included
2017 Idem 58.5%

Source: Xu et al. (2018; Ouyang (1936); Department of Comprehensive Statistics of National Bureau
of Statistics (2010); National Bureau of Statistics of China (2008); National Bureau of Statistics of
China (2018)
168

168  Meimei Wang and Bas van Leeuwen

The pre-​1949 effect of urbanisation on education

Traditional China was famous for its civil examination system, which was based
on Confucian canons. However, the military defeats in the Opium Wars in
the mid-​19th century created a Westernisation movement that strove for the
modernisation of industry and education. This resulted in accelerated indus-
trial and commercial growth. Economic historians (Skinner, 1977) have argued
that the resulting urban labour demand attracted a large-​scale inflow of people
to urban areas. Other scholars (Chi, 2001) added that after the 1930s, when
the government became increasingly weaker, refugees who wanted to escape
from famines, disasters or wars accounted for a large proportion of rural–​urban
migration.
This increasing urbanisation affected educational development in various
ways. First, urbanisation affected education since the latter was initially
funded by (and dependent on) donations from businessmen and missionary
organisations, as well as by self-​payment by students (Lu, 2011), who were more
abundant in cities. The opening of treaty ports and the concessions granted by
the Chinese government to foreign countries resulted in the further funding
and establishment of numerous missionary schools in urban areas. Some of the
educationalists (Zhu, 2009; Wang, 2013) who researched the development pro-
cess of modern education in China viewed these missionary schools favourably.
These schools’ external funding meant free educational opportunities for the
children of poor families, especially in elementary and girls’ schools. In addition,
these schools introduced flexible teaching methods and practical courses suited
to the modern sciences and technology. These schools were therefore deemed
to be promoters of China’s modern education reform and pioneers in modern
female education.5
The second factor driving the relation between urbanisation and educa-
tion, which has been researched mainly by sociologists, concerns the availability
of potential students. Indeed, urbanisation was driven at least partly by better
income and job opportunities for migrants. The population inflow to cities
contributed significantly to the expansion of educational resources in urban
areas, particularly for relatively higher-​level education and in private education.
Taking Nanjing as an example, local students accounted for only 52 per cent
of the total number of pupils attending primary schools established by the local
municipal government in 1930. Around 15 per cent of these students came
from other regions of Jiangsu Province, and the remaining students (approxi-
mately 30 per cent) came from other provinces. However, in private general
secondary schools, only 17–​20 per cent of the students were locals (Nanjing
Education Bureau, 2010). After the 1930s, as noted above, refugees started to
account for a disproportionate share of rural–​urban migrants, resulting in a large
influx of low-​skilled, cheap labour into urban areas and leading to a further
decline in the share of locals in urban schools (Liu, 2009).
The third channel through which urbanisation affected education, which
has been discussed in detail by economists, is the creation of a completely
 169

Spatial-economic development in China 169

new, Western-​style education system. From the late-​19th century onwards, the
Chinese education system moved slowly away from the traditional “civil exam-
ination system” based on Confucian canons, towards a more modern system.
This shift was a long and relatively slow process but was led by cities because
of the demands of modern industry and commerce in terms of the knowledge
and skills of the labour force. According to contemporary curricula recorded in
Zhongguo Jindai Xuezhi Shiliao (Zhu, 1983), courses that were suitable for urban
industrial and commercial development were dominant in urban schools. In
contrast, in the less-​developed regions, although required by the government
to update their curricula to those of modern schools, most traditional schools
did not modernize because of a lack of teachers and teaching equipment. The
resulting gap in quality between urban and rural education motivated even more
people to send their children to schools in urban [Link] effect was aggravated
by various government policies, which, despite being designed to support and
encourage the development of modern education overall, were neverthe-
less directed mostly at cities.6 Although educationalists were excited about the
substantial progress in the development of modern education, sociologists and
economists argued that the scale and structure of education developed faster than
did the demand arising from economic growth. Particularly in pre-​1949 China,
the job market in some fields quickly became saturated. For example, according
to the statistics recorded in the first education yearbook of Republican China,
a total of 553, 986 teachers and staff members were working in all primary schools
in 1930, but the number of graduates from secondary normal schools in 1929
and 1930 combined was 580, 304 (Wang, 1934), meaning that the new graduates
for only two years outnumbered the total number of existing staff [Link]
shortage had been replaced by overproduction.

The 1949–​1979 period
In the first phase of the New China period, debates on the relationship between
urbanisation and education changed fundamentally, from being the domain of
historians, sociologists and economists to being that of contemporary polit-
ical scientists. First, as discussed in Section 2, the definition of urbanisation
transformed from a static historical understanding to a more flexible Hukou-​
based definition. Second, the government’s goal moved away from balancing
educational training and economic demand towards a focus on enlarging the
industrial base. The First Five-​year Plan (1953–​1957) established that urban
areas should be developed as bases for implementing the national industrialisa-
tion strategy and that the primary function of cities should be production (The
National Planning Commission, 1995). All of the new industrial projects in
the First Five-​year Plan were distributed among 120 cities, with a focus on 18
key cities (Kirkby, 1985), resulting in strong urban demand for skills. However,
because of the resulting rapid expansion of urbanisation (Cao and Chu, 1990),
in 1955 the government decided to decentralize industrial production units to
small cities to take full advantage of the existing infrastructure in these settings.
170

170  Meimei Wang and Bas van Leeuwen

Third, there were mixed benefits of the migration that occurred in 1949–​
1979. To ensure the smooth progress of its large-​scale industrial projects, the
government transferred economic surpluses from rural areas to urban areas via
“price scissors”. Agricultural products were sold forcibly at low prices, causing
farmers to become poorer. At the same time, urban residents had their food
supply guaranteed by the government and also benefited from social welfare
benefits such as health security (Liu and Wu, 2006). The large gap in living
conditions between rural and urban areas attracted many rural migrants to the
cities. The decentralisation of industrial projects beginning in 1955 was bene-
ficial to rural migrants because they did not need to move far away from their
hometowns. Consequently, urbanisation accelerated after 1955. During 1957–​
1960, the net increase in the urban population was 30–​40  million, and the
number of workers in industrial enterprises more than doubled (Cao and Chu,
1990). Peak migration from rural areas occurred in 1958, and rural migration
contributed at least half of the increase in the urban population during that
period (Kang, 1966).
As urbanisation placed demands on education, the expansion of education
was also included in the government’s strategy. The first step was to increase
its scale, and the second step was to modernize higher education to match the
demand from industry. Indeed, from 1951 to 1953, three-​quarters of all uni-
versities were converted to exclusively technological universities (The Editorial
Office of the China Education Yearbook, 1984).
Despite the rapid growth of education, still only around 30–​40 per cent of
junior secondary school graduates had the opportunity to attend senior sec-
ondary school, and after graduation from senior secondary school, only 5 per
cent had the possibility to enrol in higher education (Gu, 2009). Consequently,
a large number of secondary school graduates required placement in various
economic sectors after leaving the education system. Despite its rapid growth,
urban economic development was insufficient to offer enough job opportun-
ities for these graduates. The urban situation was aggravated after 1960, when
a large famine struck the entire country and it became increasingly difficult
for rural areas to supply cities with sufficient food. Due to the oversupply of
graduates and the effects of the famine, the government decided to reduce the
urban population, imposing restrictions on rural–​urban migration using the
Hukou system. As well as restricting migration from rural areas to cities, a large
number of young, educated students, as well as staff members and managers of
government agencies and state-​owned enterprises, were moved to rural areas
(known as the “Rusticated Youth” programme). However, de-​urbanisation did
not stop at the end of the Great Famine because the Cultural Revolution
swiftly followed, lasting until 1978. The programme was most active during
1966–​1972, when 41.8 per cent of secondary school graduates in urban areas
(7.15 million) were sent to rural regions with the aim of increasing rural pro-
duction and reducing the pressure on the cities. In addition, secondary and
higher education was severely limited, and regular higher education stopped
recruiting altogether in 1966–​1970 (see Figure 9.1). It was not until 1969 that
 171

Spatial-economic development in China 171

100
90
80
70
60
50
40
30
20
10
0

Figure 9.1 Number of students in higher education (10,000 persons)

Source: China Education Yearbook ca. 1949–​1981

the economy began a gradual recovery, and higher education resumed. This
improvement can be attributed mainly to the industrial sector, which had a
growth rate of 30.6 per cent (Xi and Jin, 2006). The urban economic recovery
offered many opportunities to educated urban youths who had been sent to the
countryside in the 1960s, and many returned to the cities (Gu, 2009).
Although the debate on the 1949–​1979 period has mostly been approached
from a policy point of view, economists, sociologists and educationalists have
recently joined the debate, focusing on the situation of the many students
involved in the Rusticated Youth project. Li, Rosenzweig and Zhang (2010)
found a negative relationship between the number of years these people spent
in rural areas and their total years of education. Giles et al. (2015) argued that,
in addition to the direct impacts of not attending school for several years, the
delay in schooling caused by their transfer to rural areas under this programme
increased the opportunity cost of continuing in school, because “children
were older, prior learning depreciated with time out of school, and returning
students faced increased competition for placement in the educational system
because of accumulated cohorts of students competing for a limited number
of slots”. This negative effect was mitigated for students whose parents were
party cadres. For example, Zhou and Hou (1999) argued that children whose
parents were party cadres were usually sent to better rural locations, and then
returned to cities earlier, compared with other children. Furthermore, these
children had a background more conducive to self-​study, which increased their
chances of passing the college entrance exam. Finally, some of these parents
172

172  Meimei Wang and Bas van Leeuwen

might even have been able to directly influence college admissions decisions
from 1972 to 1977, when such decisions were based on political, rather than
academic, considerations.

The Reform Period
In many ways, the migration in the post-​1978 Reform Period signalled a return
to pre-​1949 patterns. First, people with agricultural Hukou were increasingly
allowed into urban areas. Second, after 1978, improved agricultural productivity
due to the establishment of the Household Contract Responsibility System7
freed up more rural labourers to go to urban areas to pursue non-​agricultural
job opportunities. Third, the development of self-​employment and the pri-
vate economy, as well as the infrastructure construction boom in urban areas,
absorbed many rural labourers (known as the Nongmingong, 农民工) and pushed
up the urbanisation ratio. In the following discussion on the effects of these
urban developments on education, we distinguish between two types of migra-
tion: active migration, of a person moving between locations; and government-​
induced migration, consisting mostly of administrative re-​designation of regions
from rural to urban, without physical relocation of the person.

Active migration
The active movement of large numbers of migrants to urban areas affected
education in various ways. First, most public primary and secondary schools
prioritized local urban students because of limited urban educational resources.
The children of migrant workers could access these schools only if seats
remained after enrolling all of the local students, and only after paying high
additional fees. In certain large cities, schools created mainly to accommodate
the children of migrant workers were of relatively poor educational quality.
Some educationalists found that these students felt ignored or discriminated
against by teachers and by fellow students because of their rural family and
educational background (Yuan, 2011). In cases where an urban school was open
to the children of migrant workers, it was common for local urban students
to leave the school because of a reluctance (of the student and the family) for
them to attend school with the children of migrant workers. Hence, according
to educationalists, migrant children were confronted with a “ceiling” in upward
social mobility via education (Xiong, 2015). The government allocated special
funding to alleviate these problems, but initiatives were focused mainly on how
many and what proportion of children from migrant families were absorbed
into urban schools. These children’s educational results and whether they were
able to adapt to the new environment after entering these urban schools were
ignored. Moreover, these children were allowed to attend urban schools only
until the end of senior secondary school. Taking the entrance examinations
required to enrol in universities was possible only at the registered residence
recorded in their Hukou. It has been shown in Shanghai, for example, that this
 173

Spatial-economic development in China 173

kind of educational institutional inequality can lead to the re-​production of
social classes (Zhang, 2012).
Second, the large number of migrant families in urban areas led sociologists
to also consider the distribution of educational resources between rural and
urban [Link] (2017) found a linear relationship between the level of urban-
isation and the numbers of schools and students at the compulsory education
stage in both urban and rural areas in Ningxia Province. From 2002 to 2014,
the total number of schools and the number of pupils per school expanded in
urban areas. However, the increase in the number of schools was far outpaced
by demand, causing a serious shortage of educational resources (Li, 2013). In
contrast, the number of schools and students in rural areas decreased rapidly.
Hence, for rural children, the average distance to a school increased.
Third, a large proportion of children of migrant workers did not move to
cities with their parents.8 These children were educated in rural areas and were
cared for by their grandparents or other relatives, and there is an intense debate
on the effects of this situation on their education. The majority of scholars,
mainly educationalists, have concluded that leaving children in rural areas under
the care of others has obvious negative effects on the quality of their educa-
tion. For instance, Ye et  al. (2006) found that these students received worse
examination results, compared with their results before their parents left to
work in the city. However, economists and demographers have disagreed with
the assessment that leaving children in the countryside when their parents go
to the city to work inevitably leads to negative effects for the children. For
example, Hu (2012) showed that income sent back to rural areas by migrant
workers improved the educational status of their children who remained in
these areas. Likewise, Duan and Yang (2008) found that the enrolment ratio
of these children was higher than that of children who moved to the city
with their parents. Providing a more nuanced picture, Zheng and Wu (2014)
combined both the advantages and the disadvantages, finding that the absence
of the father and the mother had various effects on their children’s educa-
tion. They performed probability proportional to size sampling among primary
school students in the fourth and sixth grades, and in first-​and third-​year junior
secondary school students, in the provinces of Gansu, Ningxia,Yunnan, Sichuan
and Guangxi for the years 2006 and 2008. They found that the positive effect
of a father moving to the city and generating more family income exceeded
the disadvantage caused by the lack of companionship and effective super-
vision of the child in terms of education. However, if a mother left her chil-
dren to work in the city, the net effect was the reverse, partly because women
generally earn less compared with their male counterparts. In addition, if the
mother went to the city, the role of decision maker would often be played by
the father, who, it is argued, would attach less value to education, especially for
girls. In cases where both the mother and the father left for the city, most of
these children lived with their grandparents. Most existing research related to
intergenerational education is from a psychological or pedagogical perspective
(Huang and Chen, 2007) and tends to agree that, although grandparents have
174

174  Meimei Wang and Bas van Leeuwen

more parental experience, more spare time and more patience, children’s educa-
tion is harmed in these situations because of the grandparents’ low educational
level, outdated educational experience and indulgence of their grandchildren.
These arguments are supported by the fact that children who were raised by
their grandparents have been shown to be more likely to receive worse exam-
ination results and to suffer psychological problems such as being self-​centred,
being psychologically fragile, having poor self-​control and being introverted.
A fourth effect of urbanisation on education in terms of active migrant
workers concerns decision-​making within the household. Some sociologists
have compared decision-​ making across generations of migrant workers,
showing that urbanisation had different influences for different groups. Xia
and Chu (2014) conducted a binary logistic regression analysis using a database
on more than 1,000 rural households in north Anhui Province. Their research
showed that compared with their parents, the younger generation of migrant
workers (born in the 1980s and 1990s) was more willing to settle in [Link]
prime motivation was to pursue a better education for their children. Yang
(2007) explained that well-​educated labourers, compared with the migrants
of past generations, have steadier, more sustainable and higher incomes. They
also face fewer institutional barriers to rural–​urban migration and are able to
promote and help many other family members, relatives or even neighbours to
move to urban areas.
As a fifth effect, consumer theorists have researched the impact of urban-
isation on education from the perspective of consumer decision-​making. Yu
(2015) conducted a field investigation in Qingdao and found that migrant
workers who had never enjoyed the benefits of education found it difficult
to correctly value education. Their decisions about whether to educate their
children thus depended on external factors, such as following urban residents’
high educational investment, rather than on economic [Link] diffi-
culty causes irrational investments in the field of education (Niu, 2017); accord-
ingly, the contrast between high educational cost and low economic return
has resulted in the increasingly popular belief in rural areas of the uselessness
of study. In addition, He (2014) found that the influence of urban lifestyle and
consumerism has resulted in the spread of hedonic, conspicuous and competi-
tive consumption throughout rural areas, and that this change in expenditure
pattern has sharply reduced the amount of money available for educational
investment. Moreover, the government’s efforts to activate rural consumption
and the rural real estate market have stimulated consumerism, and expenses
related to marriage and social networks have [Link] shifts in spending
patterns have increased the financial burdens for migrant families and weakened
their capability to invest in education.

Government-​induced migration
In addition to the migrant workers who participated actively in urbanisation,
many people have been urbanized by the government. As cities expand, rural
 175

Spatial-economic development in China 175

areas have increasingly been merged into cities as new districts, where their
inhabitants obtain a non-​agricultural Hukou. Following the government’s desire
to increase the urbanisation ratio and consolidate areas of farmland to stimu-
late large-​scale farming, disenfranchised famers have been forcibly relocated
to resettlement areas, usually in rural–​urban fringe zones, to live in high-​rise
buildings that have urban [Link] policy, along with increases in the numbers
of migrants to cities, have caused expansion in both the urban population and
the urban area.9 Economists have criticized this rapid urbanisation forced by
land conveyance and concentration of land, and suggested that the urbanisation
is merely an illusion because without the support of industrial development,
there are insufficient non-​agricultural job opportunities for rural workers who
appear to have been simply re-​categorized by the government.
Many scholars have directed their attention to rural workers who became
absorbed into urban populations because of the loss of their farmlands, and to
the education of their children. Distinct from “active migrants”, these people
were not outsiders; rather, recently urbanized people. Their economic situation
seemed to have improved, if only temporarily, as they received compensation
from the government. However, unlike the active migrants who entered cities
with good preparation, clear plans and the support of their farming families in
their hometowns, government-​induced migrants were relocated to urban areas
without any preparation or means of escape. Liu (2009) conducted a survey in
a new district of Huainan city in Anhui Province, which was converted from
a rural area to an urban district in 2005. More than 80 per cent of the land
was expropriated by the government. Farmers received either cash or a job in
return, but 70.8 per cent of those surveyed were dissatisfied with the compen-
sation. Ultimately, 55 per cent of these landless rural workers had to survive
by working as manual labourers or engaging in small business. Having lost
the stable income from their farmland, they were forced to spend more time
working and therefore had less time to take care of their children. Because they
no longer had a source of guaranteed survival (i.e. farmland),10 63.7 per cent
considered education the only way out for their children, and began to pay
more attention to their children’s education. They were willing to invest, but if
they reached a point when they felt that their children were not academically
promising, they would stop their support and allow their children to enter the
labour market. Yao’s (2008) comparative research showed that this phenom-
enon occurred only many years after a region was urbanized, because the initial
mission for peasants who had just lost their farmland was to adapt to a new
lifestyle and new methods of making a living. Therefore, they initially spent less
money and time on their children’s education.

Conclusion
In this chapter, we assessed the connection between historical and regional
studies in China using the effect of urbanisation on education as a case study.
Three threads run through this chapter. First, how urbanisation is estimated
176

176  Meimei Wang and Bas van Leeuwen

depends to a great extent on the particular academic field of the author and
the current political situation. Second, because of the significance of the gov-
ernment throughout the history of urbanisation, connections with policy and
political perspectives need to be [Link], different effects of urbanisation on
education are discussed in various academic fields, including history, education,
sociology and economics.
From the 1890s to the 1930s, the development of the modern urban
economy brought more educational funding and increased the number of
migrant students. More importantly, this development promoted the transition
from an education system that was based on the humanities and dominated
by the civil examination towards the modern system of education, which
emphasizes modern technology and sciences. However, the structural imbal-
ance of the modern economy led to the uneven development of different types
of schools and different specialties. Hence, discussions began about the over-
production of students, which was aggravated by refugees during the wars in
the 1930s and 1940s.
In the period from 1949 to 1979, the situation changed fundamentally.
Before the 1960s, the government’s strategy of rapid industrialisation was dom-
inant. Strong demand from the modern industrial sector led to a boom in
educational specialties related to mining, mechanism, construction and manu-
facturing. However, during the 1960s and continuing through 1979, economic
stagnation caused de-​urbanisation to become the main focus of the govern-
ment. Access to urban education, especially higher education, became severely
restricted. Although some people gradually returned to urban areas beginning
in the early 1970s, political factors continued to have a strong effect on educa-
tion through the regulation of rural migrants’ access.
The Reform Period was, in many ways, a return to the pre-​1949 situation.
Generally speaking, the new requirements of the labour market, awareness
regarding the significance of education and increasing income helped in
promoting the expansion of education. However, the final results are more
complicated due to many other factors, which include the difficult situations
faced by children left behind by their parents in rural areas, land-​deprived
rural workers and the children of migrant workers; the increasingly wide gap
between rural and urban education caused by population mobility; and the
labour market saturation and consequent class solidification.

Notes
1 The research leading to these results received funding from the European Research
Council under the European Union’s Horizon 2020 Programme (ERC-​ StG
637695 –​HinDI) as part of the project “The historical dynamics of industrialization
in Northwestern Europe and China ca. 1800–​2010: A regional interpretation”.
2 According to this distribution, the second-​largest city in a country has about half the
inhabitants of the largest city, the third-​largest city has about one-​third the inhabitants
of the largest city, and so on.
 177

Spatial-economic development in China 177

3 In 2010, the sixth census showed that 49.68 per cent of the total population lived
in urban areas. However, the demographic data published by the Ministry of Public
Security recorded only 34.17 per cent as non-​agricultural [Link] National
Bureau of Statistics (2010) acknowledged that around one-​third of the urban popu-
lation consisted of migrant rural workers and their family members. These people
were engaged in non-​agricultural production and lived in urban areas but did not
have the same access to social welfare and public services as did urban households.
Some of these migrants migrated regularly between urban areas and rural areas.
Some scholars have argued that this part of the population was only semi-​urbanized
(see, e.g. Li, 2013).
4 In 2017, 42.35 per cent of the population had non-​agricultural Hukou (see National
Bureau of Statistics of China, 2017). This implied that “real urbanisation” was much
lower. However, other scholars argued for different ways of estimating urbanisation.
For example, Li (2010) argued that urbanisation should be calculated based on the
occupational structure, where the total non-​agricultural population made up the
urban share. If this approach were adopted, urbanisation would have been more than
60 per cent because, at that time, many people with agricultural Hukou made their
living in urban areas or worked in secondary or tertiary industries in rural areas.
5 In addition to educationalists, economic historians also noticed the role of educa-
tional funds, but these scholars focused mostly on modern factories in urban areas,
which also started an apprentice education system to train labourers (see, e.g. Wang
and Xue, 2016). On average, the number of apprentices in Beijing accounted for
roughly 25 per cent of the industrial workforce (Gamble, 1921).
6 For instance, to cultivate more teachers for modern education, the government
offered a tuition exemption and subsidies, as well as guaranteed jobs, to normal
school students.
7 This system gave farmers the right to make independent decisions on the use of a
small plot of contracted land. The production and management were entrusted to
individual farming households through long-​term contracts. During the contract
period, the farmers paid taxes to the state and kept all the other proceeds for them-
selves. The system resulted in a steep increase in agricultural productivity.
8 According to the results of the sixth census, conducted in 2010, there were a total
of 69.7 million children who were left behind by their parents. Of these children,
61  million were in rural areas. Children who were left behind by their parents
accounted for 28.52 per cent of all rural children (Duan et al., 2013).
9 From 1990 to 2000, urban land expanded by 90 per cent, and the urban popula-
tion increased by 52 per cent. In the following ten years, urbanisation expanded at
a slower pace, but urban land increased by 83.41 per cent (Huang, 2016).
10 Before they became landless, 47.2 per cent of the peasants believed that, if their chil-
dren could not continue to move forward in education, their survival would still be
guaranteed by their farmland.

References
Bai, N. S. (白南生), 2003. The urbanization in China (中国的城市化). Management
World (管理世界), 11, pp. 78–​97.
Bairoch, P., 1988. Cities and economic development: From the dawn of history to the present.
Chicago: University of Chicago Press.
178

178  Meimei Wang and Bas van Leeuwen

Brezis, E., 2008. Bairoch, Paul (1930–​1999). In:  S. N. Durlauf, and L. E. Blume, eds.
The new Palgrave dictionary of economics. London: Palgrave Macmillan. Pp. 332–​333.
[Link]
Cao, S. J. (曹树基), 2000a. China’s population history (中国人口史), 4.  Shanghai
(上海): Fudan University Press (复旦大学出版社).
Cao, S. J. (曹树基), 2000b. China’s population history (中国人口史), 5.  Shanghai
(上海): Fudan University Press (复旦大学出版社).
Cao, H. T. (曹洪涛), and Chu, Ch. H. (储传亨), 1990. Urban construction in contemporary
China (当代中国的城市建设). Beijing (北京): China Social Sciences Press (中国
社会科学出版社).
Chandler, T., 1987. Four thousand years of urban growth: A historical census. Lewiston: The
Edwin Mellen Press.
Chi, Z. H. (池子华), 2001. History of refugees in China (中国流民史), the modern period
(近代卷). Hefei (合肥): Anhui Renmin Press (安徽人民出版社).
Department of Comprehensive Statistics of National Bureau of Statistics
(国家统计局国民经济综合统计司), 2010. China compendium of statistics 1949–​
2009 (新中国六十年统计资料汇编). Beijing (北京):  China Statistics Press
(中国统计出版社).
Duan, Ch. R. (段成荣), and Yang, G. (杨舸), 2008. A study on the situation of left-​
behind children in rural areas of China (我国农村留守儿童状况研究). Population
Research (人口研究), 32, pp. 15–​24.
Duan, Ch. R. (段成荣), Lü, L. D. (吕利丹), Guo, J. (郭静), Wang, Z. P.
(王宗萍), 2013. The general situation of survival and development of
left-​behind children in rural areas of China:  An analysis based on the
data of the 6th census (我国农村留守儿童生存和发展基本状况—​—​
基于第六次人口普查数据的分析). Population Journal (人口学刊), 35, pp. 37–​46.
Gamble, S. D., 1921. Peking:  A social survey conducted under the auspices of the Princeton
University Center in China and the Peking Young Men’s Christian Association.
New York: George H. Doran Company.
Giles, J., Park, A., and Wang, M.Y., 2015. The great proletarian cultural revolution, disruptions
to education, and the returns to schooling in urban China. The Institute for the Study of
Labor. Discussion paper No. 8930.
Gu, H. Zh. (顾洪章), 2009. The “rusticated youth” program (中国知识青年上山下乡始末).
Beijing (北京): China Daily Press (人民日报出版社).
He, X. F. (贺雪峰), 2014. The urbanization path of China (城市化的中国道路). Beijing
(北京): Oriental Press (东方出版社).
Hu, F., 2012. Migration, remittances, and children’s high school attendance: The case of
rural China, International Journal of Educational Development, 32, pp. 401–​411.
Huang, X. F. (黄晓芳), 2016. Ministry of Natural Resources of the People’s
Republic of China:  Permanent basic farmlands around 89 cities have been
delimited (国土资源部：89城市周边永久基本农田划定). Economic Daily
(经济日报),  05–​26.
Huang, Sh. (黄姗), and Chen, X. P. (陈小萍), 2007. Literature review of intergenerational
education (隔代教育研究综述). Modern Education Science (现代教育科学), 2,
pp.  63–​65.
Kang, Ch., 1966. Industrialization and urban housing in communist China. The Journal
of Asian Studies, 25(3), pp. 381–​396. [Link]
Kirkby, R. J., 1985. Urbanization in China: Town and country in a developing economy 1949-​
2000AD. New York: Columbia University Press.
 179

Spatial-economic development in China 179

Li,A. M. (李爱民), 2013.A study into semi-​urbanization in China (中国半城镇化研究).
Population Research (人口研究), 37, pp. 80–​89.
Li, H. Q. (李焕勤), 2013. An analysis of the effect of new urbanization factors and
counter-​measures on the balanced allocation of resources in the urban and rural
compulsory education: A case of Henan Province (新型城镇化对城乡义务教育
资源均衡配置的影响及对策探析). Value Engineering (价值工程), 10, pp. 240–​242.
Li, X. L. (李迅雷), 2010. Urbanization in China is underestimated (中国城市化率被低
估了). New Fortune (新财富), 11, pp. 36–​39.
Li, H. B., Rosenzweig, M. and Zhang, J. S., 2010. Altruism, favouritism, and guilt in the
allocation of family resources: Sophie’s choice in Mao’s mass send-​down movement.
Journal of Political Economy, 118(1), pp. 1–​38. [Link]
Liu, E. M. (刘恩民), 2009. The strategy and the education plight of the landless farmers’
children in the development of the urbanization（城市化进程中失地农民子女的教
育困境及对策研究—​—​以安徽省淮南市三和乡为个案. Southwest University
(西南大学).
Liu, Sh. (刘胜), 2009. Correlations of modern refugees and urbanization in Kuomingtang
Regime (浅析民国时期流民与城市化关系). Journal of Wuhan Polytechnic (武汉职
业技术学院学报), 8, pp. 15–​17.
Liu,Y. T. and Wu, F. L., 2006. The state，institutional transition and the creation of new
urban poverty in China. Social Policy & Administration, 40(2), pp. 121–​137. [Link]
[Link]/​10.1111/​j.1467-​9515.2006.00480.x
Lu, Ch. (卢川), 2011. Urban education in Jingzhou during the Qing Dynasty (略论清
代荆州的城市教育), Journal of Yangtze University (Social Sciences) (长江大学学报社
会科学版), 34(6), pp. 6–​8.
Malanima, P., 2009. Decline or growth? European cities and rural economies. In: M.
Cerman and E. Landsteiner, eds. Zwischen land und stadt. Jahrbuch für geschichte des
ländlichen raumes, 6, pp. 2–​29.
Modelski, G., 2003. World cities: -​3000 to 2000. Washington: FAROS.
Nanjing Education Bureau (南京特别市教育局), 2010. Educational statistics of Nanjing
City in 1929, 1930 (民国十八年度、十九年度南京特别市教育概况统计). In: Y.
Wang, ed. Compilation of archives and statistics of education in Republican China (民国教
育资料统计汇编), 13. Beijing (北京): National Library of China Publishing House
(国家图书馆出版社).
National Bureau of Statistics of China (国家统计局), 2008. Provisions on the statistical
division of urban and rural areas (关于统计上划分城乡的规定).
National Bureau of Statistics of China (国家统计局), 2018. Statistical communique of The
People’s Republic of China on 2017 (2017年国民经济统计公报).
National Bureau of Statistics of China (国家统计局), 2018. China statistical yearbook
2018. Beijing(北京): China Statistics Press (中国统计出版社).
Niu,W. X. (牛文学), 2017. Analysis of individual decision-​making mechanism in urban-
ization (城镇化中个体决策机理分析). Special Zone Economy Issue (特区经济), 343,
pp.  70–​73.
Ouyang, D. X. (欧阳德修), 1936. Problems and comparison of urban and rural popu-
lation in China (中国都市人口和乡村人口问题及比较). South Wind (南风), 1,
pp.  5–​6.
Skinner, W., 1977. The city in late imperial China. Stanford CA: Stanford University Press.
Sun, B. W. (孙本文), 1946. Social problems in China (现代中国社会问题), 2.  Beijing
(北京): The Commercial Press (商务印书馆).
180

180  Meimei Wang and Bas van Leeuwen

The editorial office of the China Education Yearbook (中国教育年鉴编辑部), 1984.
China education yearbook ca 1949–​1984(中国教育年鉴1949–​1984). Beijing
(北京): Encyclopedia of China Publishing House(中国大百科全书出版社).
The National Planning Commission. (国家计划委员会), 1955. The 1st five-​year plan
of China (中华人民共和国发展国民经济的第一个五年计划). Beijing (北京),
People’s Publishing House (人民出版社).
United Nations, Department of Economic and Social Affairs, Population Division,
2018. World urbanization prospects: The 2018 revision. United Nations.
Wang, Sh. J. (王世杰), 1934. The first education yearbook of Republican China (第一次中
国教育年鉴). Shanghai (上海): Shanghai Kaiming Book Store (上海开明书店).
Wang, T. (王涛), 2013. Missionary schools and modernization of Chinese education
(教会学校与中国教育近代化). Journal of Ningxia University (Humanities and Social
Sciences) (宁夏大学学报（人文社会科学版）), 35(2), pp. 178–​180.
Wang, Y. Zh. (王远征), 2001. The options and obstacles in China’s urbanization path
(中国城市化道路的选择和障碍). Strategy and Management (战略与管理), 1,
pp.  31–​37.
Wang, J. D. (王季点) and Xue, Z. Q. (薛正清), 2016. An investigation report of factories
in Beijing (调查北京工厂报告). In:  Ch. Zheng (郑成林), ed. The third collection
of economic data of Republican China (民国时期经济资料调查三编), 20. Beijing
(北京):  National Library of China Publishing House (国家图书馆出版社), pp.
167–​208.
Xi, X. (席宣), and Jin, Ch. M. (金春明), 2006. A brief history of the Cultural Revolution
(文化大革命简史). Beijing (北京): Chinese Communist Party History Publishing
House (中央党史出版社).
Xia, Y. J. (夏永久), and Chu, J. L. (储金龙), 2014. An inter-​generational comparison
of peasants’ willingness to migrate to urban areas:  Evidence from northern
Anhui Province (基于代际比较比较视角的农民城镇化意愿及影响因素—​—​
来自皖北的实证). Urban Development Studies (城市发展研究), 21(9), pp. 12–​17.
Xiong, Y. H. (熊易寒), 2015. The education of children of migrant workers is still
affected by institutional discrimination (农民工子女教育仍面临制度性歧视).
China Youth Daily (中国青年报), 5(4), pp. 10.
Xu, Sh. L. (许仕廉), 1930. The population issue in China (中国人口问题). Beijing
(北京): The Commercial Press (商务印书馆).
Xu, Y., van Leeuwen, B. and Van Zanden, J. L., 2018. Urbanization in China, ca. 1100–​
1900. Frontiers of Economics in China, 13(3), pp. 639–​664. [Link]
s060-​007-​018-​0018-​9
Yang, G. K. (杨公科), [Link] transfer of the rural labour force and the career orienta-
tion of persons with rural education (谈农村劳动力转移与农村教育的职业取向).
Education and Vocation (教育与职业), 24, pp. 173–​174.
Yao, B.T. (姚宝亭), 2008. Analysis on the educational dilemma and accordingly causes of the children
of landless farmers in the process of urbanization (城市化进程中失地农民子女的教育
困境及成因分析). Shanghai Normal University (上海师范大学).
Yao, Y. Zh. (姚银枝等), 2017. An empirical study of the impact of urbanization on
the scale of compulsory education in urban and rural areas in Ningxia Province
(城镇化对宁夏义务教育城乡办学规模影响的实证研究), Journal of Ningxia
University Humanities & Social Sciences Edition (宁夏大学学报（人文社会科学版）),
39(1), pp. 149–​158.
Ye, J. Z. (叶敬忠等), Wang, Y. H. (王伊欢), Zhang, K. Y (张克云), Lu, J. X. (陆继霞),
2006. The influence of parents’ going out to perform labour on the study of
 181

Spatial-economic development in China 181

left-​behind children in rural areas (父母外出务工对农村留守儿童学习的影响).
Rural Economy (农村经济), 7, pp. 119–​123.
Yin, M. X. (殷梦霞), and Li, Q. (李强), 2009. A compilation of economic investigations along
the railway lines in the Republican China (民国铁路沿线经济调查报告汇编),Vol. 14.
Beijing (北京): National Library of China Publishing House (国家图书馆出版社).
Yu, J. (于洁), 2015. Rural migrant workers’ decision-​ making behav-
iour on education expenditure:  Based on field work in city of Qingdao
(进城农民工家庭教育支出决策行为分析—​—​基于山东省青岛市的实地调查),
Education Research Monthly (教育学术月刊), 2, pp. 56–​61.
Yuan, L. X. (袁立新), 2011. A comparative study of discrimination against
middle school students in public schools and schools for migrant workers
(公立学校与民工子弟学校初中生流动儿童受歧视状况比较). Chinese Journal of
School Health (中国学校卫生), 32(7), pp. 856–​857.
Zhang,Y. H. (张银海), 2012. Children of migrant workers: How to escape from status
reproduction (农民工子弟：如何逃离阶层再生产). South Reviews (南风窗), 1,
pp.  64–​66.
Zhao,Y. Q. (赵燕青), 1988. A review of theories of the urbanization path in China (中
国城市化道路理论述评). In: Ye, W., ed. A preliminary study on the urbanization path
in China. Beijing (北京): China Prospect Publishing House (中国展望出版社).
Zheng, L. (郑磊), and Wu, Y. X. (吴映雄), 2014. The impact of parental migra-
tion on educational development of left-​behind children: Evidence from western
rural area of China (劳动力迁移对农村留守儿童教育发展的影响—​—​
来自西部农村地区调查的证据). Journal of Beijing Normal University (Social Sciences
edition) (北京师范大学学报社会科学版), 2, pp. 139–​146.
Zhou, X. G., and Hou, L. R., 1999. Children of the Cultural Revolution: The state and
the life course in the People’s Republic of China. American Sociological Review, 64, pp.
12–​36. [Link]
Zhu, L. P. (朱利培), 2009. A brief analysis of the impact of missionary schools on the
modernization of Chinese education (浅析教会学校对中国教育近代化的影响).
Knowledge Economy (知识经济.), 10, pp. 176–​177.
Zhu, Y. H. (朱有瓛), 1983. Historical materials of modern educational system in China
(中国近代学制史料), 1(1). Shanghai (上海):  Eastern China Normal University
Press (华东师范大学出版社).
182
 183

Conclusions
184
 185

10 
Setting an agenda for a
“New Regional History”
Marijn Molema

Introduction
You might recognize the critical voice in your head which accompanies you
as you work on a project, repeatedly asking the same irritating, basic questions.
For me it was Bugs Bunny –​the Warner Bros cartoon rabbit who outsmarted
everyone in animated films such as Looney Tunes  –​who delivered a running
commentary in my head for a number of years. Every now and then he would
stop rushing around and provokingly ask his famous question: “Eeh… What’s
up, doc?” If we translate this question to the theme of this edited volume, we
would ask: what is the use of spanning the gap between regional studies and
history? Luckily, the essays collected in this book helped me to find the straight-
forward answer needed to convince an impatient and restless character like
Bugs Bunny. First, bridging the gap revitalizes an older theme in the study of
history, that of regional development, and also increases its societal relevance.
Second, pulling history into regional studies strengthens the explanatory power
of this multidisciplinary field.
This is the short answer to the Bugs Bunny question. However, the last
chapter of this edited volume is not the script of a cartoon movie. Here, I take
the opportunity to elaborate again on the interrelationship between history and
regional studies and construe some research perspectives which emerge from
that by using the individual contributions to this book. My work is also based
on the presentations and discussions from five workshops which I co-​organized
in 2017 and 2018 as the key contact person from the RSA Research Network
on [Link] workshops had an explorative character and yielded (too) many
themes and questions (Molema et al., 2017; Parrish et al., 2017). It would have
been completely unfeasible to design a conceptual framework which included
all the phenomena and questions that we touched upon. Instead, this chapter is
more of a personal vision of a fruitful direction for future research, inspired and
strengthened by the ideas of others.
To begin with, let us imagine a field of research in which regional historians
exchange knowledge with colleagues from the social sciences, such as
geographers, economists, political scientists and sociologists. Imagine this field
to be a football field demarcated by four corner flags. Each corner flag stands
186

186  Marijn Molema
for a basic principle which is shared by both historians and their social scien-
tific colleagues. The field’s demarcation causes the interplay between historians
and social scientists to be more effective, because it offers an overview of the
area within which the two scientific domains can interact. I will sketch the four
corner flags in the four following sections, in order to shape a multidisciplinary
field of future research. My contribution will end by naming this field “New
Regional History”.

Societal challenges
The first “corner flag” argues that every multidisciplinary exchange should
begin from a societal problem or challenge. Closer inspection of regional studies
and history reveals the phenomenon of “academic drift” at work in both scien-
tific domains. With this, I mean an inborn tendency of scientists to seek inspir-
ation for new research questions within the academic field of their research
itself. Discussions with and reading the work of colleagues are the sources of
scientific renewal. There is nothing fundamentally wrong with the drift away
from the more practically oriented world outside academia. Criticism of the
utilitarian view of science is in many cases perfectly understandable (Collini,
2017). Many innovative ideas and most ground-​breaking discoveries were made
in isolated academic systems.
However, this begins to become problematic when fields of research ignore
the societal problems that call for academic reflection and new knowledge pro-
duction, and which help societies to deal with complex situations and challenges.
When it comes to regional development, the field of regional studies cannot
be accused of paying no attention to societal constraints. The Regional Studies
Association (RSA) was founded by engaged academics in 1965. Its founding
members wanted to assist policymakers by discovering new opportunities for
regional development, and they contributed to regional policies by assessing
its instruments and theories. In the early years of its existence, the RSA fre-
quently engaged in public debates and was heard as a collective voice which
praised or criticized political decisions. Critical observers claim that the RSA
has evolved from its beginnings in the 1990s into an international community,
primarily based on the services, products and networks that it provides towards
its members (Hopkins, 2015). As a result, the association’s “collective voice”
function diminished. Nevertheless, it is without doubt that regional studies
as a field still concentrates on the problems discussed among policymakers,
politicians and citizens involved in regional development.
The historical sciences are also great suppliers of public [Link]
many historians avoid speaking about the future because they “understand only
the past”, there are a lot of scholars who endeavour to help the public grasp
current events with their historical knowledge. However, the core of this dis-
cipline consists of academics with a nationalist view of history. The nation state
has been a safe haven of historians for more than 150 years now, only recently
being defied by “global history”. Regional perspectives lie more towards the
 187

Setting an agenda 187
fringes of the discipline, although there are countries, such as Germany, where
federal structures support the production of conferences, journals and books
about regional history. Regional history is generally regarded as a piece of a
national mosaic. Regional history in most cases relates to national history by
presenting a different view, or by explaining why national developments took
the course they did. Again, there is nothing wrong with this kind of regional
history –​it enriches our knowledge and provides inspiration for many. Be that
as it may, many societal challenges take place at the regional level. The signifi-
cance of these regional phenomena is not grounded in some higher level of
scale; it is the very regional nature of these challenges themselves that legitimizes
a regional view. Current challenges within regions, however, are almost off the
radar of regional and other historians.
What then are these “problems and challenges”? In order to answer this
question, it is instructive to look at how our ideas about regional develop-
ment have been changed from the post-​war period to the present day. In the
1950s and 1960s, most Western countries embraced industrialization as the
impulse for economic development. Regional industrialization policies were
established to give vulnerable regions an impulse and spread welfare across
the country. The industrial development perspective lost its dominance in the
1970s however, because of factors including market saturation, competition
from low-​wage countries and emerging environmental awareness. From the
1980s onwards, innovation became the key word  –​innovation which could
occur in all sectors and across economic activities, and not only in industry.
The consequence was that regions had to find their strengths within their own
economic structures. A new “endogenous growth” perspective also emphasized
other economic opportunity factors, such as the local work ethic, favourable
knowledge institutions, effective cooperation between regional stakeholders,
etc. Growth incentives had to come from within, instead of the exogenous state
subsidies for industrial development.
It is however quite difficult to define and identify clearly the aspects of eco-
nomic potential which might serve as starting points for future development.
These aspects of potential are deeply interwoven into an area’s ecosystem (Stam,
2015). As something which is part of an evolutionary and path-​dependent pro-
cess, regional economic potential is not always visible on the surface. The dif-
ficulty is to first recognize the opportunities for regional development, then to
acknowledge them and finally to find ways to strengthen their economic poten-
tial. Policymaking on the basis of the endogenous development opportunities
is even more difficult for regions which are vulnerable and/​or are situated at
the periphery of a country’s economic and political core areas. The economic
structure of such regions might be historically weak and their “potential” could
therefore be harder to find. The weaknesses of such areas are a reality which
cannot be washed away with a discourse of bottom-​up development. Moreover,
these areas quite often lack the expertise and human resources to recognize,
acknowledge and strengthen their economic [Link] resources are abso-
lutely necessary:  reflecting on a regional economy and formulating strategic
188

188  Marijn Molema
development plans is quite difficult, at least if you compare it with the one-​size-​
fits-​all model of regional policies in the post-​war era. Meanwhile, cities and
metropolitan areas do have the means to draw up attractive development plans
and assign lobbyists to raise funds for the implementation of these plans.
With an exhausting range of long-​ term statistics on income, financial
assets and purchasing power, Thomas Piketty (2013) showed how inequality
increased from the 1980s. Piketty collected his data mainly as averages at
the nation state level. From a bird’s eye view of these nation states, however,
inequality also acquires a spatial dimension. In his impactful article, the eco-
nomic geographer Andrés Rodríguez-​Pose (2018) stressed this spatial dimen-
sion of inequality and connected it to political populism. This phenomenon
has, according to Rodríguez-​Pose, strong territorial foundations and is a real
societal challenge:  “the populist wave is challenging the sources of existing
well-​being in both the less-​dynamic and the more prosperous areas” (p. 189).
Vulnerable regions are accompanied by societal risks of social and economic
deprivation, demographic decline and, as a result of this, the loss of cultural
variety. These regions might turn their backs on the establishment, and retain
popular and/​or extremist views and convictions.
To recap:  driven by structural transformations, ideas about economic
development shifted away from the primacy of industry and embraced the
endogenous, specific opportunities presented by a territory instead. However,
as a probably unintended consequence of this paradigm shift, state policies for
economically vulnerable regions lost their prominent position in most modern
countries. Because of the electoral repercussions, the problem of left-​behind
areas is regaining its significance in societal and political debates. This problem
is too complicated to be tackled by a single discipline. Instead, we need close
cooperation between history and regional studies.
The added value of historians is their holistic view and their tendency to
understand problems in an evolutionary sense (Molema and Van der Zwet,
2017). For historians, the theme of regional development also has the poten-
tial to reach a new public. A new regional history offers great opportunities to
load this old sub-​discipline with new energy and enhance its societal impact.
Regional historians gain the opportunity to come into contact with people and
organizations which need historical knowledge and insight in order to find the
right direction for the future. The consequence of this is a concentration on
themes connected to societal challenges: territorial inequality, regional popula-
tion decline and the economic renewal of cities and regions.

Construction of regions
A second point of departure or corner flag represents the idea that history
and regional studies should together raise our understanding of how regions
were constructed, by whom and on the basis of what kind of arguments. The
areas studied by most regional historians are delineated pragmatically:  they
determine the borders fixed by constitutional law. Sponsors such as regional
 189

Setting an agenda 189
governments or regional cultural funds support regional historians to write the
history of these regions or districts. This is at least the case in my country, the
Netherlands, where impressive works have been published in this way about
several of our provinces. The social, cultural and economic characteristics of
regions such as the Dutch provinces, however, do not necessarily coincide with
the geographical space of such administrative entities. Commuting to work and
economic exchanges between companies are examples of daily practices which
transcend the official borders of regions. There are many more such examples,
like participation in sports activities or the receipt of specialized healthcare,
the experience of a landscape and cultural values, collaborative initiatives to
support economic innovation –​all these illustrate activities which frequently
have a regional character and which are not confined to a particular administra-
tive district. Such insights persuaded geographers decades ago that regions are
not fixed, but have a relational character (Tomaney, 2009). The relational char-
acter of regions refers to individual phenomena connected to each other within
a given area. In other words, viewing regions as objects of research which exist
principally in the sense of their administrative function has been superseded
as an approach by the analysis of regions through studying the geographical
scales at which people, towns, cities, companies, institutions, landscapes, etc.
are interrelated. The geographical area of a region is fundamentally open and
in fact part of the research question. The only convention which needs to be
accepted is the idea that the size of regions resides somewhere between a town
and the national territory.
The relational bases for regions have been theorized by geographers from
diverse backgrounds. The German geographer Blotevogel (1996) drew a the-
oretical distinction between three types of region. The first type comprises
regions which are principally understood as analytical categories (“Analytical
Regions”), in which specific characteristics define what the region is. A region
can be perceived as a territory which is densely populated. Another example
can be seen in the idea of a region as an area in which a more or less coherent
labour market functions. Blotevogel’s second type is the Activity Region.
Activities can be economic, such as a region which presents itself as a unity,
for example because of its marketing goals. Activities can also have a political
character: regions are then delineated by their administrative boundaries or the
area on which a specific development plan focuses. A third type is the Identity
Region. The social, cultural and political identities of individuals or groups are
frequently connected to a specific territory, which forms the spatial frame-
work into which feelings of regional belonging are loaded. Anssi Paasi (1991)
elaborates on the theme of regional identity. His impactful interpretation of the
region as a sociocultural and historical category stresses the evolutionary char-
acter of the construction of regions. The institutionalization of these regions is
a result of four stages. A region needs: 1) a territorial shape; 2) symbolic aspects
connected to a specific area; and 3) an institutional shape. Stage 4 requires that
the result of the first three stages becomes embedded into a society’s political
and societal consciousness.
190

190  Marijn Molema
Blotevogel and Paasi are among the academics who have provided a termin-
ology for understanding regions as constructed by human action. This insight
gives way to a critical investigation of the agents and processes which lie at the
basis of certain regional categories: are they scientists, politicians, entrepreneurs
or a combination of the three? We should also investigate how regions
change: with Blotevogel, we can look into the dynamics within and between
Analytical, Activity and Identity Regions. Such a deconstruction of the region
is especially valuable with respect to the goals of historically informed regional
studies. These studies, as we have noted in the “societal challenges” section,
should impact on current debates about regional development. However, if
a researcher aims to inform future decisions for a specific region, they must
understand the genesis of the region. In other words: the conscious or uncon-
scious political, economic and cultural interests which affected the construction
of a region should be integral to the investigation. Problematizing the regional
level prevents the researcher from taking regions for granted, and helps them to
situate the regional level in a wider perspective.

Multidisciplinary exchange
So far, we have planted corner flags which demarcate new regional history
as a field in which research projects are connected to societal challenges, and
where the construction of regions is part of the questioning. A third corner
flag should introduce multidisciplinary exchanges between regional and his-
torical studies as a rule of conduct. Regional studies is a relatively young and
multidisciplinary field in itself which belongs to the social sciences, whereas
history has a longstanding tradition within the humanities. Molema and
Tomaney, in their contribution to this edited volume, analysed the distinctions
between the two academic fields in terms of the methodological differences in
which history, broadly understood, applies a narrative approach while regional
studies seeks general theories. They sketched how the social and historical
sciences have diverged since the late 19th century, and converged again during
the closing years of the 20th century. There is therefore a tradition of cross-​
over between the humanities and the social sciences, a tradition on which
we can build our research agenda. Moreover, the research objectives of both
regional and historical studies are situated in a regional society, which makes
the overlap obvious.
As obvious as this overlap may be, it is impossible to transfer theories, meth-
odologies and concepts from one field to the other without further exploring
its compatibility. This edited volume can be regarded as such an exploration,
which offers us insights into what we gain by connecting regional and histor-
ical studies. From an analytical perspective, we can distinguish several routes
for multidisciplinary exchange, such as the research perspective, the concepts
used or the methodology applied. Where Huggins and Thompson’s chapter
is an example of the first, Fernández-​ de-​
Sevilla and Costa’s contribution
illustrates the second route. The third and methodological perspective has been
 191

Setting an agenda 191
demonstrated by the chapters on quantitative measurement and their implicit
or explicit connection to qualitative methodologies.
In their chapter, Robert Huggins and Piers Thompson touch upon the
structure–​agency problem which smoulders in both the humanities and the
social sciences. How much agency does one person or a group of persons have?
Can they change a historical process, or is this impossible, because historical
processes are embedded in the collective, behavioural patterns that structure
human action? The British sociologist Anthony Giddens (1984) answered these
questions by finding a middle ground: human behaviour is structured by social
forces, but these structures can be changed by human action. A similar rejec-
tion of extreme structuralism can be found in the philosophy of history as well.
Studies of the past in which structures dominate historical explanations, and
in which people are reduced to background actors, remain static: they do not
reveal the experiences and indentation of our ancestors (cf. Ankersmit, 1994,
pp.  182–​238). Indirectly intervening in these broad and to a certain extent
philosophical discussions, Huggins and Thompson argue for a behavioural
approach in which we take the role of ordinary people into account. Like every
other kind of development, regional development is connected to individuals
who are able to inspire their urban or regional society to change. Illustrated
with the seminal work of Peter Hall (1999), Cities in Civilization, their examples
are manifold: from entrepreneurs and inventors, to mayors and leaders of civil
society movements. These people, who have a great impact on their societies,
become woven into urban and regional cultures. The reciprocal relationship
between leading individuals and these cultures is an interesting and important
theme for future research.
Multidisciplinary connections can be established by the transfer of concepts
from regional studies to historical studies. These concepts can offer tools to
make historical studies more analytical, and help to compare different regional
economic development histories with each other. Regional studies is a rich
field when it comes to theoretical ideas which help to grasp the regional devel-
opment process. Recent decades have seen the emergence of several concepts
which impacted on scientific reflection as well as on policymaking practice.
These include the learning region (Morgan 1997), smart specialization (McCann
and Ortega-​Argilés, 2013) and economic resilience (Martin and Sunley, 2015).
However innovative these concepts may be, the ongoing introduction of new
terms into the academic debate is not without controversy. Critical scholars
have accused academic proponents of concepts such as “learning region” of
sacrificing their academic independence and prioritizing their relationships
with politicians and policymakers over obtaining a true understanding of how
regional and other economies work (Lovering, 1999). Judgements which are
more fundamental because they criticize the heuristic value of this concep-
tual abundance, also come from scholars who point to the vagueness of many
regional studies (Markusen, 2003). Such discussions within regional studies
underline the importance of empirical research. A longer time perspective can
be very helpful here, because it helps both with theory construction based on
192

192  Marijn Molema
the new phenomena which are revealed, and with strengthening the practical
application of theoretical concepts. The chapter by the economic historians
Tomàs Fernández-​de-​Sevilla and Armando J. dalla Costa is an illustration of this
added value. They embrace the cluster concept and describe how the growth
of an automotive cluster around São Paulo also simultaneously fundamentally
transformed it. Small repair and car parts manufacturing workshops popped up
spontaneously, and formed the basis of a cluster which was gradually –​espe-
cially after the Brazilian government’s Foreign Direct Investment (FDI) policies
in the 1950s –​taken over by a few large companies. As such, the chapter shows
that the evolution of clusters is not only dependent on market circumstances;
economic policies also play a crucial role. It is therefore important to look at
those policies, and explain their origins with reference to the history of the
societies in which they were developed.
A third route towards multidisciplinary exchanges between regional and his-
torical studies has a methodological character. The combination of qualitative
and quantitative research deserves our special attention here. Economic history
has a particularly rich quantitative tradition in which theories and methodolo-
gies from the social sciences, especially economics, are a source for inspiration.
Cliometric approaches can prove the evolution of historical events, processes
and structures. However, the underlying explanation also requires a qualita-
tive route, through which the historical context, the crucial variables and the
contingent factors are explained. Multidisciplinary exchanges between regional
studies and regional history thus requires a combination of quantitative and
qualitative approaches. This edited volume therefore presents two chapters on
this subject. The economic historians Kerstin Enflo and Anna Missiaia clarify
methodologies on measuring Gross Domestic Product (GDP) in regions. The
measurement of regional GDP is a spinoff from major projects within eco-
nomic history which resulted in national accounts and provided insight into
the long-​term development of countries. The regionalization of these quan-
titative approaches provides more insight into the concrete development of
individual regions, and stimulates interregional comparisons. Enflo and Missiaia
summarize some of their results on Swedish regions, which also offer an oppor-
tunity to consider regional [Link] their time series, they show how a
decrease in regional equalities coincided with the Industrial Revolution in the
latter half of the 19th century. Only in the later decades, from 1980 onwards, did
regional inequalities arise again. Such long-​term insights stimulate a search for
explanations and thus a combination of quantitative and qualitative approaches.
Mixed methods are also applied in the chapter by Justyna Kościńska (soci-
ologist) and Mikołaj Herbst (economist). They look primarily at educational
achievements, because education is an important driver of regional develop-
ment. First, on the basis of secondary school mathematics test scores, they show
that the regions of Poland that were formerly parts of the Austro-​Hungarian
Empire performed better than those regions which had belonged to Russia,
and substantially better than the former Prussian regions. Second, they explain
these differences with the help of questionnaires and in-​depth interviews with
 193

Setting an agenda 193
the students’ parents. The differences are bound up in different forms of social
capital, differences in the levels of trust of education and variation in the per-
ception of educational goals. Combining these methods constructively, this
chapter optimally illustrates the benefits of multidisciplinary exchange applied
from a long-​term perspective.

Comparison and transfer

There is nothing wrong with a regional history that is not invoked by soci-
etal challenges, which does not question the regional borders set by con-
stitutional law and which does not strive for a multidisciplinary exchange
with the field of regional studies. As already stated, the traditional method for
conducting regional history raises our knowledge of the past and is a source
of inspiration to many. However, it is problematic if regional history remains
limited to traditional themes and does not innovate with new themes and
approaches, in spite of societal developments which revive the importance of
the regional level for dealing with questions of economic, social and cultural
sustainability. In that case, regional history runs the risk of becoming a mar-
ginal sub-​discipline entangled with itself, and thus limited to the history of
a region’s past, which can result in increasingly sophisticated stories, but can
also lack reference to a regional society’s present or future. A final corner flag,
in addition to the three already mentioned, helps regional history to avoid
these marginalizing effects. This fourth corner flag argues for a comparative
approach and a preference for analyzing the transfer of people, ideas and com-
modities between regions.
Regions and their past, present and futures must be compared for simi-
larities and differences and mechanisms of adaptation, change or lock-​in to
be analysed. The broad comparison between Europe and the United States
presented by Kevin Cox in his chapter demonstrates the merit of juxtaposing
different categories. Cox shows how different political approaches towards
regional economic development are rooted in a continent’s history and society.
Whereas European states took a leading role in economic and spatial policies
in the development of vulnerable regions, this was not the case in the United
States. There, the involvement of the centre was reduced to a minimum; local
governments, competing with each other for market locations, took the lead
instead. The bottom-​up approach in America is connected to the develop-
ment of capitalism in that country and the absence of deeply entrenched pre-​
capitalist structures, as was the case in European societies. Meimei Wang and
Bas van Leeuwen illuminate a very different tradition of state involvement with
their chapter on China. These two economic historians, like Kościńska and
Herbst in this volume, considered education, and especially its relationship with
urbanization. After a discussion of how urbanization can be measured and how
this relates to different disciplinary approaches, the two authors reveal how the
Chinese state impacted the education process through prohibitive and permis-
sive policies.
194

194  Marijn Molema
The contribution of Martin Åberg and Thomas Denk provides the meth-
odological basis needed to employ comparative methods properly, and in a way
that opens new heuristic possibilities. Comparative Sequence Design (CSD),
as they call their method, can be used in any research design, first, because the
method captures both synchronic comparisons (focused on the similarities and
differences between certain periods) and diachronic comparisons (focused on
the parallel analysis of different cases). Second, CSD provides the tools for ana-
lyzing one case study separately, but also for examining the interdependence
between case studies. As such, it backs the emancipation of regional level ana-
lysis: the logic behind CSD values regions as independent geographical entities
which can directly impact each other. This is why the emphasis on regional
interdependence releases the region from the national straitjacket. Contrary to
mosaic thinking, in which the significance of regions is closely aligned to the
national whole, regional developments stand for themselves and can impact
each other without intervening national governments. This makes the chapter
by Åberg and Denk an inspiring contribution which stimulates the discovery of
regions which impact each other across national borders.

“New Regional History”

At the beginning of these concluding remarks, I used the metaphor of a playing
field demarcated by four corner flags, which set some basic rules for the inter-
play between historical and regional studies. The first flag required that every
research question should start from a societal problem or challenge. The second
flag required that researchers view how regions were constructed critically.
The third rule embraces multidisciplinary exchange, in which long-​term ana-
lysis goes together with concepts and theories from regional studies. Fourth,
research should be comparative in nature. From a bird’s eye view, we can call
the playing field “New Regional History”.
What is new about this kind of history? New is that a region’s history is
not viewed as part of a larger (national or European) whole. We are not pro-
posing mosaic thinking, as though regions only have significance as part of a
larger whole. Instead, we want to emancipate the regional view of history as
a perspective which has considerable value in its own right. Its value resides
in the societal challenges which need to be grasped in the context of their
historical genesis.
The title of this essay was “setting” an agenda. This verb, setting, makes clear
that the agenda is not yet set. The basis of this edited volume is also not strong
enough to act as a foundation for a new approach which offers new scien-
tific opportunities and provides scholars a method for doing research which
is relevant and important with regard to societal challenges. We need more
time and experience for communities within both regional studies and his-
tory to create a research tradition of their own, with a common set of research
themes, questions and approaches. This edited volume sketched some angular
points for such a tradition. Hopefully, it will encourage research into regional
 195

Setting an agenda 195
challenges, and I hope it inspires the long-​term investigation of regional growth
and adaption, regional inequality and policies which help find ways to prevent
social, economic and cultural deprivation.

References
Ankersmit, F., 1994. History and tropology: The rise and fall of metaphor. Berkeley, Los
Angeles and London: University of California Press.
Blotevogel, H. H., 1996. Auf dem Wege zu einer ‘Theorie der Regionalität’: Die Region
als Forschungsobjekt der Geographie. In: G. Brunn ed. Region und Regionsbildung in
Europa. Baden-​Baden: Nomos Verlag. Pp. 44–​68.
Collini, S., 2017. Speaking of universities. London and New York: Verso.
Giddens, A., 1984. The constitution of society. Outline of the theory of structuration. Cambridge:
Polity Press.
Hall, P., 1999. Cities in civilization: Culture, innovation, and urban order. London: Phoenix
Giant.
Hopkins, J. A., 2015. Knowledge, networks and policy: Regional studies in postwar Britain and
beyond. London: Routledge.
Lovering, J., 1999. Theory led by policy:  The inadequacies of the ‘new regionalism’
(illustrated from the case of Wales). International Journal of Urban and Regional Research,
23(2), pp. 379–​395. [Link]
Markusen, A., 2003. Fuzzy concepts, scanty evidence, policy distance:  The case for
rigour and policy relevance in critical regional studies. Regional Studies 37(6 and 7),
pp. 701–​717. [Link]
Martin, R. and Sunley, P., 2015. On the notion of regional economic resilience:
Conceptualization and explanation. Journal of Economic Geography 15(1), pp. 1–​42.
[Link]
McCann, P. and Ortega-​Argilés, R., 2013. Smart specialization, regional growth and
applications to European Union cohesion policy. Regional Studies, 49(8), pp. 1291–​
1302. [Link]
Molema, A. M. and Van der Zwet, A., 2017. Research network on regional economic
and policy history. Planning Perspectives, 32(3), 459–​466. [Link]
02665433.2017.1331753
Molema, A. M., Reeploeg, S., Åberg, M. and Van der Zwet, A., 2017. Report on the
inaugural workshop of the RSA Research Network on Regional Economic and
Policy History. Regions, 308, pp. 29–​30.
Morgan, K., 1997. The learning region: Institutions, innovation, and regional renewal.
Regional Studies, 31(5), pp. 491–​503. [Link]
Paasi,A., 1991. Deconstructing regions: Notes on the scales of spatial life. Environment and
Planning A: Economy and Space, 23(2), 239–​256. [Link]
Parrish, M., Molema, A. M., Svensson, S. and Reeploeg, S., 2017. Report on the second
workshop of the RSA Research Network on Regional Economic and Policy
History. Regions, 308, 31–​32.
Piketty, T., 2013. Le capital au XXIe siècle. Paris: Éditions du Seuil.
Rodrìguez-​Pose, A., 2018. The revenge of the places that don’t matter (and what to
do about it). Cambridge Journal of Regions, Economy and Society, 11(1), pp. 189–​209.
[Link]
196

196  Marijn Molema
Stam, E., 2015. Entrepreneurial ecosystems and regional policy:  A sympathetic cri-
tique. European Planning Studies, 23(9), 1759–​ 1769. [Link]
09654313.2015.1061484
Tomaney, J., 2009. Region. In: R. Kitchin and N. Thrift eds. International encyclopedia of
human geography, vol. 9. Oxford: Elsevier. Pp. 136–​150.
 197

Index

ABC region 119, 124, 128–​129, 131–​132 cross-​sectional design 104–​107, 110–​111
agency 14, 36–​39, 46–​47, 51–​53, 191 Cultural Revolution 170
agglomeration 69; diseconomies of 145,
147, 149, 151, 154–​155, 160 distributism 148
“Älvsborgs ransom” 90
Anchieta motorway 119, 129 education 40, 44, 63–​64, 158,
ANFAVEA 130–​133, 135 192–​193; Chinese reforms in 164,
Annales School of History 23 168–​169; educational policies 123;
anti-​urbanism  148 intergenerational 173; organizations 133;
Polish system of 14–​15, 66–​67, 74–​75
Bairoch 165, 168 evolutionary economic geography 6–​7,
behavioural approaches 191; behavioural 12, 29–​30
economics 36–​37; behavioural Executive Group of Automobile Industry
psychology 12, 36, 45 (GEIA) 128–​132
big five personality traits 38, 41
bounded rationality 37, 46 Federation of Industries of the State of São
branch plant 151–​153, 155–​156, 160 Paulo (FIESP 124–​125, 127
Brazil 16, 119–​122, 124–​134, 136–​137, 192 Ford 123–​125, 127–​131, 137
Bugs Bunny 185
business climate 154, 156, 159 Gadamer, H. G. 23
Galicia 75–​77
case study design 104–​105, 113 Geary and Stark method 84, 87, 92–​93
CEDAR Umeå Demographic database 90 General Motors (GM) 124–​125, 130
Centre of Industries of the State of São Geographical Regression Discontinuity
Paulo (CIESP) 124–​125 Design 68–​70;Vargas G. 126–​128
centre/​periphery  28 Glasgow 50–​51
China 16, 155, 164–​181, 193; civil Gravier, J. F. 148
examination system 164, 168–​169, 176; “growth machines” 154
planning 13, 28, 31; land use 158–​159;
urban and regional 16, 120, 140–​149, Hall, P. 47, 191
153, 160 historical institutionalism 12, 27, 35
clusters 6, 13, 16, 21, 30–​31, 42, 44, 102, historical sociology 25, 32
119–​129, 131–​139, 150, 165, 192 hub-​and-​spoke district 122, 133, 136, 138
comparative approaches: diachronic Hukou System 166–​167, 169–​170, 172, 175
comparisons 15, 99, 100, 104–​105,
107–​112, 194; method 14, 194; research immigration see migration
design 6, 98–​112; sequence design (CSD) industrial districts 13, 21, 121–​122, 136
15, 100, 104–​105, 109–​112, 194 inequality 15, 32, 36, 82–​83, 145, 149, 158,
competitiveness 39, 44, 52–​53, 133 173, 188, 195
computer-​assisted personal interview 70 interdependence 15, 51, 98–​99, 194
198

198 Index
Ipiranga 124, 126, 128–​129, 131 quasi-​experimental methods  66–​68
Italy 65, 86, 88, 121–​122, 153
Ranke, L. von 22
Kubitschek, J. 128, 134–​135 regional convergence 86, 155, 156
Kingdom of Prussia 15, 63, 69–​78, 192 Regional Studies Association (RSA)
186, 12–​13
Lamprecht, K. 22–​23 rent 149–​150, 152, 153, 155, 157;
learning region 13, 191 rent-​seeking behaviour 38, 40, 46, 51
lock-​in 14, 29, 51, 120, 193 resilience 13, 21, 29, 111, 191
longue durée 64; Chauvel, L. 65 rural-​urban migration see migration
Russian Empire 63
Maddison, A. 83 rusticated youth 170, 171
Manchester 48–​49, 51, 150, 154
marshallian externalities 121–​122, 135 São José dos Campos 119, 129, 130,
Mercedes-​Benz 128–​130 133, 134
Methodenstreit 23, 32 Shanghai 172
migration 38, 43, 51, 65, 71–​72, 101, Simca 130, 131
112, 125, 164–​175; migrant 40, 43, 69, Sindipeças 127, 133, 135
23, 168; migrant children 172, 173; smart specialization 191
migrant workers 172–​176; rural migrant social capital 39, 40–​41, 43, 71, 75–​76, 78,
170, 172 192–​193
moral economy 159 social cohesion 39–​41, 44, 52
Most Different Systems Designs social norms 40, 43, 64–​65, 68,
(MDSD) 103 75–​76, 79
Most Similar System Design (MSSD) 103 state, territorial structure of 15, 16, 29, 63,
105, 110, 157, 158, 188
Nanjing 168 subnational comparison 99
new institutional economics 26 Sweden 84, 88, 90–​94, 106, 153
new towns 16, 150, 154–​156, 160 synchronic comparison 105, 107, 108,
110, 194
path dependency 4, 29, 30, 64, 66, 72,
110–​111, 187 Tabellverket 90–​91
personality psychology 14, 37–​39, 41, territorial politics 27
42–​44, 52, 76; planning 13, 28, 31; civil territorial division see state, territorial
examination system 164, 168–​169, 176; structure of
land use 158–​159; urban and regional, 16, treaty ports 168
120, 146–​149
policy: analysis 99–​100, 108; development uneven development 37, 145–​146, 155, 159,
126, 146, 150, 152, 154–​155, 157; 160, 176
diffusion 15, 100, 112; implementation urban history 30–​31
67, 100, 112, 124, 132; industrial 16; urbanization 31, 71, 150, 164–​166,
regional 11–​13, 101, 108, 120, 149, 174–​175, 193
151, 157
Professional Association of the Autoparts Vemag 126–​127
Industry of São Paulo (Associaçao) 127 Volkswagen (VW) 128, 130, 138
property capital 16, 146, 149, 150, 160
proxy-​based approach 84, 85, 86, 89 wages 65, 84, 87–​93, 122, 128, 146, 156,
psychocultural behaviour 38 160, 187
psychocultural profile 14, 43–​45, 46, 51–​53 Willys–​Overland Motors (Willys or WOB)
psychological imprint 38 128–​131, 138