THE ACCOUNTING REVIEW American Accounting Association

Vol. 89, No. 3 DOI: 10.2308/accr-50684

2014
pp. 1083–1113

Tone Management
Xuan Huang
California State University, Long Beach
Siew Hong Teoh
University of California, Irvine
Yinglei Zhang
The Chinese University of Hong Kong

ABSTRACT: We investigate whether and when firms manage the tone of words in
earnings press releases, and how investors react to tone management. We estimate
abnormal positive tone, ABTONE, as a measure of tone management from residuals of a
tone model that controls for firm quantitative fundamentals such as performance, risk,
and complexity. We find that ABTONE predicts negative future earnings and cash flows,
is positively associated with upward perception management events, such as, just
meeting/beating thresholds, future earnings restatements, SEO, and M&A, and is
negatively associated with a downward perception management event, stock option
grants. ABTONE has a positive stock return effect at the earnings announcement and a
delayed negative reaction in the one and two quarters afterward. Balance sheet
constrained firms and older firms are more likely to employ tone management over
accruals management. Overall, the evidence is consistent with managers using strategic
tone management to mislead investors about firm fundamentals.

Keywords: tone management; qualitative disclosure; earnings management; market

efficiency; behavioral finance.

I. INTRODUCTION

T
he tone of the qualitative text in earnings press releases can be too optimistic or pessimistic
relative to concurrent disclosures of quantitative performance. We call the choice of the
tone level in qualitative text that is incommensurate with the concurrent quantitative
information tone management. We investigate whether managers engage in tone management for

We thank Lucile Faurel, David Hirshleifer, Phyllis Hirshleifer, Ian Gow, Yuyuan Guan, Chansog Kim, Alex Nekrasov,
Mort Pincus, Benjamin Whipple, Liu Zheng, and workshop participants at California State University, Fullerton;
California State University, Long Beach; City University of Hong Kong; The Chinese University of Hong Kong;
University of California, Irvine; National University of Singapore; University of Colorado; Yale University; the 2012
FARS Midyear Meeting; and 2012 AAA Annual Meeting for very helpful comments. Yinglei Zhang acknowledges the
financial support from Rega Technologies Limited.
Editor’s note: Accepted by John Harry Evans III.
Submitted: November 2011
Accepted: October 2013
Published Online: December 2013

1083
1084 Huang, Teoh, and Zhang

informative or strategic purposes, and whether and to what extent the capital market discounts for
strategic motives, if any, when reacting to earnings announcements.
Quantitative information by itself provides investors with an incomplete picture of a firm’s
economic circumstances. For quantitative information to be used, investors need to first encode the
information and then process it (Fiske and Taylor 1991). The rhetoric employed in the qualitative
text of earnings press releases facilitates encoding and processing of the quantitative disclosures and
generally informs the reader. However, when agency incentives are present, the rhetoric could
instead mislead the reader. As Rapp (2010, Section 4.2) writes, rhetoric is a value-neutral tool ‘‘that
can be used by persons of virtuous or depraved character. This capacity can be used for good or bad
purposes; it can cause great benefits as well as great harms.’’
Earnings press releases, being voluntary, are not subject to explicit rules about the disclosure,
so management has wide latitude in the qualitative presentation of the quantitative information. We
are interested in studying how the tone of the press release affects readers’ response to the
communication, and whether and how tone can be used as a tool to affect investors’ perception
about the firm. As the old adage goes, ‘‘It’s not what you say; it’s how you say it.’’
In a neutral presentation of the press release, tone will vary with the quantitative content of the
disclosure, with optimism in tone increasing in firm performance. We decompose net positive tone,
the difference in the frequency of positive minus negative words in earnings press releases, into two
components. The normal component reflects a neutral tone that is commensurate with concurrent
information about current and expected future firm quantitative performance.1 The residual
component, abnormal positive tone, is the main variable of interest in our study and is intended to
capture the discretionary component of tone.
The tone of the earnings press release can be a tool for managers either to improve
understanding of, or to obscure, firm fundamentals. Concurrent quantitative information may not
fully reflect all available information about future cash flows, partly owing to limitations in
generally accepted accounting principles (GAAP). When economic fundamentals are better than
indicated by the quantitative information, an abnormal positive tone of the press release can be
used to signal better future prospects. In this case, abnormal positive tone helps inform investors
about the anticipated improved future prospects. On the other hand, tone management may be
employed opportunistically to mislead investors by being unduly positive or negative relative to
the reported quantitative information, even when this leads to a less accurate perception of
fundamentals.
In sum, a key goal for this paper is to test whether tone management in earnings press releases
informs or misinforms investors. We examine how abnormal positive tone relates to future firm
performance, whether abnormal tone is more likely used in situations where managerial strategic
incentives to manipulate investor perception are present, and whether and how investors react to
tone management at the time of, and subsequent to, the earnings announcements.
Most accounting capital markets research studies quantitative information reported by firms.
However, there is growing interest in the qualitative aspects of various types of firm
communications with investors, such as in particular sections of 10-K reports, earnings press
releases, and conference calls.2 We select earnings press releases to study tone management for
several reasons. Earnings press releases make up a large proportion of news events about the firm
that are consistently more timely and significant than financial reports to the SEC. Furthermore,

1
We control for concurrent quantitative information using data from Compustat, CRSP, I/B/E/S, and First Call
CIG. Details of the tone model are in the ‘‘Abnormal Positive Tone Measure in Earnings Press Releases’’ section.
2
Li (2011) provides an excellent comprehensive review of recent tone-related papers in the accounting literature,
including some very early papers. Section II discusses previous literature.

The Accounting Review

May 2014
Tone Management 1085

greater discretion about content and format is afforded by these voluntary disclosures than by the
mandatory 10-K reports. Finally, the trading volume and stock price reactions are generally larger
around earnings announcements than at any other time in the year except for special event
announcements.
Section II describes how tone optimism and pessimism have been measured in various ways in
the existing literature. Our main tests use the classification of words with positive and negative tone
by Loughran and MacDonald (2011; hereafter, LM) because this dictionary list was developed
specifically for accounting reports and business purposes. We also consider robustness of results
with respect to two alternative word lists, Henry (2008) and Harvard’s General Inquiry (GI).
We expect that various economic factors drive optimism in the tone of earnings press releases
and therefore we use a benchmark model for tone that controls for firms’ current and expected
future performance, growth, risk, and complexity. Abnormal positive tone therefore captures effects
that are orthogonal to the underlying quantitative fundamentals.3
We first examine whether abnormal positive tone in the earnings press release contains
incremental information about future firm accounting performance. In our sample, we find that
abnormal positive tone is associated with poor future earnings and operating cash flows in each of
one-year to three-year-ahead periods. The negative relation between abnormal positive tone and
future performance is incremental to the effect of abnormal accruals.
The sign of the relation between abnormal tone and future performance is crucial for
distinguishing between whether abnormal positive tone informs or misinforms investors. The
finding that abnormal positive tone predicts future negative performance is sufficient to reject the
hypothesis that discretionary tone informs investors. It would be counterintuitive for managers to
use an abnormally optimistic tone to signal poor future earnings or cash flows to investors. To
corroborate whether abnormal positive tone misinforms, we further examine whether abnormal
positive tone is associated with the presence of strategic incentives and whether abnormal positive
tone misleads investors.
The incentive to mislead investors may stem from a manager’s desire for prestige, or from
pecuniary motives associated with agency problems. Therefore, we investigate whether tone
management is used in settings where managerial incentives to manipulate perceptions are present.
We find that abnormal positive tone in earnings press releases increases the likelihood that the
disclosed earnings of that period just meet or beat past earnings and analysts’ consensus forecast, as
well as the likelihood of future earnings restatements, new equity issuances, and mergers and
acquisition activities, consistent with managers using abnormally optimistic tone when incentives to
bias perceptions upward are present. We also find evidence for the opposite situation for stock
option grants where the incentive to bias perceptions is downward. The strike price is often set at
the prevailing market price at the time of grant. Our evidence suggests that abnormal positive tone
decreases with the likelihood of stock option grants.
Next, we examine whether investors understand that abnormal positive tone contains negative
information about future fundamentals. If investors cannot see through the managerial opportunism
that is driving discretionary tone or if they discount insufficiently for the strategic motives, then
abnormal positive tone will incite investor optimism beyond the level warranted by future
fundamentals and cause an immediate positive market reaction at the earnings announcement. In

3
There is an obvious analogy between normal/abnormal tone and non-discretionary/discretionary accruals.
Abnormal positive tone may capture managerial discretion on tone, including managers’ biased estimation of
future fundamentals, or simply noise. Therefore, analogous to the literature on discretionary accruals, we test
whether abnormal positive tone predicts future firm performance, and whether it is associated with strategic
incentives.

The Accounting Review

May 2014
1086 Huang, Teoh, and Zhang

such cases, because information about poor future earnings or cash flows arrives either in
subsequent financial reports, analysts’ reports, or the business press, there should be a return
reversal so that the delayed reaction to abnormal positive tone is negative.4
Our evidence indicates that abnormal positive tone is associated with a more positive
immediate market response to the earnings announcement and a more negative market response in
one and two quarters subsequent to the announcement. The return reversal in the post-announce-
ment period is strong evidence of an over-reaction to abnormal positive tone at the earnings
announcement, and is in sharp contrast with the evidence of a continuation of the market response
to extreme earnings news or post-earnings announcement drift in the literature.
The past literature on discretionary accruals suggests that accruals management is a tool to
manipulate investor perception about a firm (Teoh, Welch, and Wong 1998; Xie 2001). Therefore,
we control for abnormal accruals in all of our tests to extract the incremental effect of abnormal
positive tone to avoid spurious inferences. In additional analyses, we explore the relation between
accruals management and tone management as alternative or complementary tools to manipulate
perceptions. Among firms that use both accruals and tone management in a consistent direction, we
find that tone management is more likely in older firms and firms facing higher balance sheet bloat,
as proxied by lagged assets scaled net operating assets, and so are more constrained in further
upward accruals management.
Finally, our main results are generally robust to the use of alternative dictionary lists,
alternative abnormal positive tone model specification that includes publicly disclosed managers’
expectation of future performance, alternative test methods that consider a seemingly unrelated
regressions (SUR) system for various groups of strategic actions, and additional regression controls
such as normal tone and investment opportunities. Overall, the evidence suggests that managers use
tone management to mislead investors and other financial statement users.
Section II next discusses the background of the tone literature and contribution of our paper.
Section III describes the sample and the estimation procedures for our variables, and discusses the
sample descriptive statistics. Section IV documents the relation between abnormal positive tone and
future financial performance. Section V examines whether abnormal positive tone presents in
various strategic settings. Section VI investigates immediate and delayed market reactions to
abnormal positive tone. Section VII presents additional analyses and robustness checks. Section
VIII concludes.

II. BACKGROUND
There is growing research in the empirical capital markets area in accounting and finance using
the textual analysis of qualitative information. These papers vary by the disclosure medium, the
measure for the qualitative characteristic, and outcomes that are investigated. The disclosure
medium include media news (Tetlock 2007; Tetlock, Saar-Tsechansky, and Macskassy 2008),
annual report/10-K/10-Q filings (Li 2008 and 2010), earnings press releases (Davis, Piger, and
Sedor 2012; Demers and Vega 2011), analyst reports (Lehavy, Li, and Merkley 2011; Hsieh and
Hui 2011; De Franco, Hope, Vyas, and Zhou 2013; Huang, Zang, and Zheng 2013), and conference
calls (Larcker and Zakolyukina 2012; Frankel, Mayew, and Sun 2010). The different approaches to
measure qualitative information include computational linguistics such as a naı̈ve Bayesian
algorithm (Li 2010; Huang et al. 2013), psychological dictionaries such as General Inquirer and

4
See Hirshleifer, Lim, and Teoh (2011) for a model of immediate and delayed investor response to news when
investors have limited attention. Investors underreact to news so that the initial response is muted and there is a
post-announcement drift in the same direction as the initial response.

The Accounting Review

May 2014
Tone Management 1087

Diction (Kothari, Li, and Short 2009), and financial-customized word lists (Loughran and
McDonald 2011; Henry 2008). The various qualitative dimensions of the disclosures that have been
studied include positive versus negative tone (Davis et al. 2012; Demers and Vega 2011; Frankel et
al. 2010), readability (Li 2008; Hsien and Hui 2011), and self-reference bias (Larcker and
Zakolyukina 2012).
With regard to outcomes investigated, several papers examine analysts’ response to the
qualitative dimension of disclosures. Lehavy et al. (2011) document that analyst following and the
informativeness of their reports are greater for firms with less readable 10-Ks. Kravet and Muslu
(2013) find that increases in risk disclosures are associated with increases in the number of analyst
earnings forecasts and revisions, and in the dispersion of forecasts.
Another group of studies investigates the stock market reactions to disclosure characteristics.
Davis et al. (2012) and Demers and Vega (2011) document a positive relation between increase in
tone optimism of earnings press releases and the immediate stock price response to earnings
announcements. Bonsall, Bozanic, and Fischer (2013), however, find a positive relation only if
quantitative earnings guidance is not provided in the earnings release. Hsieh and Hui (2011) find
that the market reacts more favorably toward analyst reports that are easier to read. In a similar vein,
De Franco et al. (2013) find that stock trading volume is higher for firms with more readable analyst
reports. Campbell, Chen, Dhaliwal, Lu, and Steele (2014) document that the market incorporates
information in disclosures about risk factors. The relation between tone and future stock returns is
also found to be positive for news media articles (Tetlock 2007), and MD&A section of 10-K/10-Q
(Feldman, Govindaraj, Livnat, and Segal 2009). Demers and Vega (2011) report a positive relation
between future returns and change in tone of earnings press releases.
Some studies examine whether the manager uses tone in qualitative disclosures to convey
information about firm fundamentals. For example, Li, Lundholm, and Minnis (2013) find that, on
average, the qualitative information in 10-K filings is very useful in assessing a firm’s competition
environment. Li (2008) finds that firms with lower earnings have less readable annual reports, and
that readability increases with earnings persistence in firms that are profitable. He concludes that
managers report tone strategically, consistent with an obfuscation incentive to mask a lower level or
lower persistence of earnings. Our study considers managerial opportunistic behavior, and so is
similar in spirit. Larcker and Zakolyukina (2012) find that a linguistics-based deceptiveness
measure in corporate executive answers to questions during quarterly earnings conference calls
predict accounting manipulations better than a model based on discretionary accruals. Tama-Sweet
(2010) finds that managers increase optimism in the tone of an earnings press release prior to
exercising options when litigation risk is low. Davis and Tama-Sweet (2012) report greater tone
pessimism in the MD&A than in the earnings press release when managers have strong incentives
to report strategically and that tone pessimism in the MD&A predicts poor future earnings.
This paper differs from the above studies in the following ways. We view disclosure tone as
jointly determined by economic fundamentals and managerial incentives. In other words, both
truthful and strategic disclosures co-exist. Accordingly, we decompose tone into a non-discretion-
ary component based on economic fundamentals, and a discretionary component that could reflect
managerial incentives, manager’s private information about future fundamentals, manager’s biased
estimation of fundamentals, or noise. Therefore, we test whether abnormal positive tone reveals
managerial incentives to inform or misinform investors about future performance. Furthermore, we
examine whether abnormal positive tone is related to various events that past literature has
identified as associated with the presence of managerial incentives to bias investor perception.
These events include firms just meeting or beating various earnings benchmarks, inflating earnings
so that they subsequently had to be restated, issuing new equity, undertaking mergers and
acquisitions, and granting stock options to managers. Third, in keeping with our emphasis on the

The Accounting Review

May 2014
1088 Huang, Teoh, and Zhang

strategic use of tone, we provide evidence suggesting that abnormal positive tone misleads
investors. We contribute to the market efficiency literature by systematically studying investors’
contemporaneous and delayed responses to abnormal positive tone. Whereas several studies
discussed above find that tone in various venues positively predicts future returns (e.g., Feldman et
al. 2009; Demers and Vega 2011), we find that abnormal positive tone at the time of an earnings
press release negatively predicts abnormal future returns. Our evidence is consistent with investors
being temporarily misled by tone management and with subsequent market correction.5
Our study is also related to the strategic disclosure literature that documents opportunistic
managerial disclosure choices. Schrand and Walther (2000) find evidence that managers
strategically disclose a prior performance benchmark that the firm is able to beat in the earnings
announcement, whereas we examine both positive and negative incentive effects using
discretionary tone of the qualitative text in earnings announcements. While the tools differ, the
goal of the disclosure choices reflects managerial incentives to affect investor perception differently
than is warranted by the fundamentals. Finally, our study contributes to the literature on earnings
management with a pilot exploration of how quantitative earnings management relates to qualitative
tone management.
In sum, we examine the information content of abnormal positive tone, investor response to
abnormal positive tone, and the relation between abnormal positive tone and various settings
associated with the presence of perception management. By systematically and collectively
studying these settings, we provide evidence that abnormal positive tone is used to facilitate
managerial incentives to mask weak future fundamentals and mislead investors.

III. SAMPLE AND DESCRIPTIVE STATISTICS

Sample and Data

We obtain the text of annual earnings press releases from PR Newswire and Business Wire,
historical financial data from Compustat, stock returns from CRSP, analysts’ earnings forecasts data
from I/B/E/S, seasoned equity offering (SEO) and merger and acquisition (M&A) effective dates
from SDC, and option grants data for CEOs from ExecuComp. We first match earnings press
releases with the CRSP/Compustat merged database by company name and announcement dates.
The availability of earnings press release text data determines our sample period, 1997–2007. We
eliminate observations without sufficient accounting and financial-market variables or with stock
prices below $1. Each year, all financial variables except returns are winsorized at the 1 percent
level. We obtain 14,475 observations of firm-year abnormal positive tone as described in the
‘‘Abnormal Positive Tone Measure in Earnings Press Releases’’ section for details. Since we do not
require firms to have future earnings, returns, or restatement data to estimate abnormal positive
tone, the sample sizes vary across different test specifications and are noted in the tables.6

Variable Measurements
Discretionary Accruals
Following prior literature (Dechow, Sloan, and Sweeney 1995), we measure discretionary
accruals using the cross-sectional modified Jones model. The sample period of 1997–2007 permits

5
Footnote 18 suggests some reasons why our findings differ from the prior literature.
6
Our sample has more unique firms than past studies because we use an additional text source from Business Wire
in addition to the PR Newswire to obtain earnings press releases. For comparison, Davis et al. (2012) examine
23,017 firm-quarterly observations from 1998–2003 and Demers and Vega (2011) examine between 14,649 and
20,899 firm quarterly observations from 1998–2006, which averages to less than 1,000 firms per quarter,
whereas our annual sample averages about 1,300 firms per year. Our sample period is also longer.

The Accounting Review

May 2014
Tone Management 1089

us to use SFAS No. 95 statement of cash flow data in the following model to estimate accruals
rather than using balance sheet data that Hribar and Collins (2002) suggest is less accurate:
TAccjt ¼ EBEIjt  ðCFOjt  EIDOjt Þ ð1Þ
where:
TAcc ¼ total accruals;
EBEI ¼ income before extraordinary items;
CFO ¼ cash flows from operations; and
EIDO ¼ extraordinary items and discontinued operations included in CFO for each firm j in
year t.
We then run the following regression for each two-digit SIC-year combination with at least
twenty observations:
TAccjt ¼ b0 ð1=Assetsj;t1 Þ þ b1 ðDSalesjt  DARjt Þ þ b2 PPEjt þ mjt ð2Þ
where:
Assets ¼ total assets;
DSales ¼ annual change in sales;
DAR ¼ change in accounts receivable from operating activities; and
PPEjt ¼ gross property, plant, and equipment, all scaled by lagged total assets.
Discretionary accruals (DA) are the regression residuals.

Abnormal Positive Tone Measure in Earnings Press Releases

Previous literature measures qualitative characteristics of financial reports using various
software packages, such as Diction (Davis et al. 2012), General Inquirer (Tetlock 2007; Tetlock et
al. 2008), and Bayesian machine learning algorithms (Li 2010). Loughran and McDonald (2011)
argue that word classifications developed for general purposes are not appropriate for evaluating
business communications. Based on a large sample of 10-Ks, they find that many words classified
as negative in the Harvard Psychological Dictionary (IV-4) using the General Inquirer software are
not typically negative for financial reports.7 They compile an alternative word list that they show is
more suitable for describing positive and negative tone in financial communications. Therefore, we
use their word list to classify the frequency of optimistic versus pessimistic words appearing in the
earnings press release. Again following Loughran and McDonald (2011), if there are negation
words (no, not, none, neither, never, and nobody) immediately before a positive word, we count the
positive word as negative.8 We create the variable TONE as the frequency difference between the
positive and the negative words scaled by total words in an earnings press release.9
Positive disclosure tone can arise for several reasons. It may merely be an expression of good
current and expected financial performance. Alternatively, tone can be upwardly biased for multiple
reasons. A positive bias in tone may be used by managers to signal to investors the private
information about positive future performance that current quantitative disclosures fail to reveal,
owing perhaps to GAAP constraints. Alternatively, positive bias may result from managers’

7
Words like tax, liability, or foreign are defined as negative words in the Harvard Psychological Dictionary, but
have few negative connotations in financial reports.
8
The frequency of instances of double negatives, i.e., negation words immediately before other negative words, is
2 percent. The results are the same whether we ignore double negatives or count them as positive.
9
The results are qualitatively similar in terms of sign, magnitude, and statistical significance when we use the sum
of positive and negative words as the denominator.

The Accounting Review

May 2014
1090 Huang, Teoh, and Zhang

strategic attempts to mask poor current performance or to hype investors’ perception about the
future performance so as to mislead investors.
We decompose TONE into a normal component, NTONE, to reflect a neutral description of
current available information about fundamentals, and an abnormal component ABTONE, our key
independent variable that proxies for managerial strategic choice of tone either to inform or
misinform investors. We run annual cross-sectional regressions of TONE on tone determinants
suggested in Li (2010) that are generally available to investors at the time of the press release to
avoid the look-ahead bias. The determinants are measures for current available fundamental
information, growth opportunities, operating risks, and complexity. Specifically, the regression
is:10,11
TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt
þb5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt
þb10 DEARNjt þ b11 AFEjt þ b12 AFjt þ ejt ; ð3Þ
where:
EARN ¼ earnings before extraordinary items scaled by lagged total assets;
RET ¼ contemporaneous annual stock returns calculated using CRSP monthly return data;
SIZE ¼ logarithm of market value of equity at fiscal year-end;
BTM ¼ book-to-market ratio measured at fiscal year-end;
STD_RET ¼ standard deviation of monthly stock returns over the fiscal year;
STD_EARN ¼ standard deviation of EARN calculated over the last five years, with at least three
years of data required;
AGE ¼ log(1 þ age from the first year the firm entered the CRSP dataset);
BUSSEG ¼ log(1 þ number of business segments), or 1 if item is missing from Compustat; and
GEOSEG ¼ log(1 þ number of geographic segments), or 1 if item is missing from Compustat.
We include three performance benchmarks because managers’ assessment of current
performance is often framed relative to benchmarks. The benchmarks are LOSS, an indicator
variable set to 1 when EARN is negative, and is 0 otherwise; DEARN change in earnings before
extraordinary item scaled by beginning total assets; and AFE, analyst forecast error, defined as I/B/
E/S earnings per share minus the median of the most recent analysts’ forecasts, deflated by stock
price per share at the end of the fiscal year. Earnings, forecasts, and stock prices are all split-
adjusted. We also include AF, analyst consensus forecast for one-year-ahead earnings per share,

10
Our specification differs from Li (2010) in several ways. We do not include variables related to managerial
discretionary behavior, such as special items, seasoned equity offering (SEO), and mergers and acquisition
(M&A) variables, specifically so that the residual as a measure of abnormal tone can reflect these strategic
incentives. These variables may also not be known to investors at the time of the earnings press release. Our data
pertain to the annual earnings press releases, unlike Li’s (2010) sample of 10-Q reports, so quarter indicator
variables are not used. These differences contribute to differences in sign and significance of some of the
coefficients, and to our smaller adjusted R2 as compared to Li (2010), in part reflecting the wider latitude in tone
of earnings press releases than in the MD&A section of 10-Q reports. Our R 2 improves, but remains smaller than
Li (2010), when instruments for strategic incentives are included. The improvement in R2 supports our
perspective that ABTONE is indeed related to strategic motives. See also the ‘‘Alternative Tone Models’’ section
for other specifications.
11
We also estimate Regression (3) at the industry-year level to obtain abnormal tone akin to the cross-sectional
Jones model for estimating discretionary accruals. Our results are quantitatively similar but statistically weaker.
This is because the tone sample is considerably smaller than the Compustat population and there are 13
independent variables in Regression (3) versus only three in the modified Jones model. Requiring at least 20
degrees of freedom for each annual cross-sectional regression for proper estimation significantly reduces sample
size.

The Accounting Review

May 2014
Tone Management 1091

scaled by stock price per share at the end of the fiscal year to control for managerial assessment
about future performance.12
The tone model determinants are selected to control for information about firm fundamentals to
obtain abnormal positive tone. Dechow, Ge, and Schrand (2010) suggest that the three determinants of
firm fundamental value for a firm existing over multiple periods are cash flows generated during the
current period, the present value of cash flows that will be generated in the future as a result of current
actions, and the present value of the change in liquidation value as a result of current actions. We
include profitability (EARN) and three earnings performance benchmarks (LOSS, DEARN, and AFE)
to capture the cash flows generated during the current period. We exploit the forward-looking property
of market variables, stock returns (RET) and book-to-market ratio (BTM), to capture information about
growth and the present value of consequent future cash flows beyond what is conveyed by current
accounting numbers. In addition, we include analyst earnings forecasts (AF) as a direct proxy for
expectations of future financial performance. We also include a loss indicator because it conveys
information about liquidation value (Hayn 1995). We further include volatility of stock returns
(STD_RET) and volatility of earnings (STD_EARN) to proxy for the operating and business risk
environment of the company. AGE captures life cycle stage of the company. The number of business
segments (BUSSEG) and geographic segments (GEOSEG) proxy for operating complexity of the firm.
Table 1 reports the estimation results of Regression (1). We find that TONE is more positive
when the firm is small, profitable, growing, and has more volatile stock returns, fewer business
segments, and strong performance relative to analyst earnings forecast. Normal positive tone,
NTONE, is the predicted value of Regression (3). ABTONE, abnormal positive tone, is the residual
of Regression (3). By construction, ABTONE is therefore designed to be unrelated to firm
fundamentals and business environment such as current market and financial performance, growth
prospects, and firm operating risk and complexity.

Summary Statistics
Each year, we obtain the mean, median, standard deviation, 1st , 25th, 75th, and 99th percentile
of the variables in our sample. We then report the annual average of the cross-sectional statistics for
the variables in Table 2. Mean (median) TONE is 0.43 percent (0.42 percent), indicating disclosure
tone in earnings press releases is generally relatively optimistic. In contrast, Loughran and
McDonald (2011) report higher mean negative words than positive words in 10-K filings. The
greater net optimism in disclosure tone in earnings press releases than in 10-K filings is consistent
with that managers being more likely to choose earnings press releases over the 10-K filings as a
venue to hype the firm. The higher salience of the more timely earnings announcement and the
lower litigation concern for the earnings press release likely increases the net benefits to tone
management in the earnings press release relative to tone management in the 10-K.13 By

12
We thank a referee for suggesting the performance benchmarks and expectations of future performance variables as tone
model determinants. Inclusion of analyst forecast variables reduces our sample and biases the sample toward larger
companies with analysts’ data. The results of an earlier version of our study show that the full set of analyses is robust with
respect to a tone model that excludes analysts’ data variables, AFE and AF. Furthermore, Table 10 reports robustness to
including managerial forecasts to proxy for managerial expectations of future performance in the tone model. It is
debatable whether expectations of future performance variables are tone determinants or proxies for strategic incentives to
manage tone. Excluding these variables from the tone model likely biases abnormal positive tone upward and, therefore,
biases toward finding a positive relation between abnormal tone and future performance and against our results that
abnormal tone predicts negative future accounting and returns performance.
13
The 10-K report is audited, its form and format are dictated to a large extent by accounting rules and regulations,
and it is more subject to evidentiary use in litigations. As suggested in Li (2011), managers are reluctant to be
optimistic in 10-K filings because of litigation concerns. Davis and Tama-Sweet (2012) report greater pessimism
in the MD&A than in the earnings press release, which they attribute to investors reacting less to MD&A
disclosures than to disclosures in earnings press releases.

The Accounting Review

May 2014
1092 Huang, Teoh, and Zhang

TABLE 1
Expected Tone Modela
TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt
þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt þ b11 AFEjt
þ b12 AFjt þ ejt :

Dependent Var. TONE Dependent Var. TONE

Indep. Var. Coefficient t-stat Indep. Var. Coefficient t-stat
a 0.0057*** (7.02) AGE 0.0003 (1.63)
EARN 0.0011** (2.47) BUS_SEG 0.0006*** (4.44)
RET 0.0000 (0.01) GEO_SEG 0.0002 (0.79)
SIZE 0.0002*** (3.34) LOSS 0.0013*** (4.48)
BTM 0.0013*** (4.52) DEARN 0.0012 (1.19)
STD_RET 0.0690*** (7.58) AFE 0.0008*** (3.10)
STD_EARN 0.0000 (0.05) AF 0.0001 (0.30)

**, *** Indicates p , 0.05 and p , 0.01, respectively.

a
Number of observations: 14,475; Adjusted R2: 4.41%.

Variable Definitions:
TONE ¼ (#positive words  #negative words)/total non-numerical words;
EARN ¼ earnings before extraordinary items/beginning total assets;
RET ¼ buy-and-hold monthly returns for 12 months ending three months after the fiscal year-end;
SIZE ¼ log(market value of equity at the fiscal year-end);
BTM ¼ book-to-market ratio measured at the fiscal year-end;
STD_RET ¼ standard deviation of RET over the last 12 months ending three months after the fiscal year-end;
STD_EARN ¼ standard deviation of EARN over the last five years;
AGE ¼ log(1 þ #years since a firm appears in CRSP monthly file);
BUSSEG ¼ log(1 þ # of business segments);
GEOSEG ¼ log(1 þ # of geographic segments);
LOSS ¼ 1 if EARN is negative, 0 otherwise;
DEARN ¼ change in earnings before extraordinary items/beginning total assets;
AFE ¼ (I/B/E/S actual EPS  median of most recent analysts’ forecasts)/stock price at the fiscal year-end; and
AF ¼ analyst consensus forecast for one-year-ahead EPS/stock price at the fiscal year-end.

construction, the mean of ABTONE is 0. More importantly, ABTONE shows considerable variation
within sample.
The summary statistics for the remaining variables are similar to those from previous literature.
In the sample period, 13.13 percent of the sample beat or meet analysts’ forecasts, 4.84 percent
restate earnings in three years after the earnings announcement, 9.84 percent and 11.36 percent of
firm-year observations engage in seasonal equity offering (SEO) and merger and acquisition (M&A)
activities, respectively, and between 15 percent and 16 percent of firm-year observations award
above median-sized stock option grants to CEOs as compensation.
The ABTONE measure as a proxy for discretionary tone is new to the literature. There are no
official rules mandating specific words for voluntary disclosures such as the earnings press release,
so total TONE itself could be regarded as discretionary. As a practical matter, words in the press
release need to be truthful (Rule 10-b5 and litigation) and SEC Regulation G imposes restrictions
on non-GAAP terms such as giving equal prominence to and reconciling non-GAAP with GAAP
earnings. However, even words that are literally true can convey other implicit connotations. If
TONE follows a simple random walk, then the change in tone, DTONE, is also another potentially
reasonable proxy for discretionary tone. Our purpose is to control for the tone components that are

The Accounting Review

May 2014
Tone Management 1093

TABLE 2
Descriptive Statistics
Variable Mean Median Std. Dev. P1 P25 P75 P99
TONE 0.0043 0.0042 0.0071 0.0135 0.0003 0.0090 0.0217
ABTONE 0.0000 0.0000 0.0068 0.0170 0.0042 0.0044 0.0167
DA 0.0140 0.0051 0.1082 0.4059 0.0518 0.0385 0.2717
EARN 0.0056 0.0379 0.1957 0.8613 0.0255 0.0852 0.3446
CFO 0.0629 0.0800 0.1723 0.6506 0.0128 0.1474 0.4494
RET 0.1726 0.0345 0.6460 0.7335 0.2228 0.3918 2.7812
SIZE 6.3041 6.1741 1.7431 2.8636 5.0360 7.4296 10.5974
BTM 0.5729 0.4563 0.5010 0.2113 0.2646 0.7371 2.6471
STD_RET 0.0340 0.0317 0.0147 0.0121 0.0227 0.0428 0.0786
STD_EARN 0.0704 0.0380 0.0875 0.0000 0.0148 0.0899 0.4336
AGE 2.5860 2.4348 0.7780 1.3863 1.9705 3.1570 4.3599
BUSSEG 1.1418 0.9512 0.6199 0.6931 0.6931 1.4354 3.0725
GEOSEG 0.9542 0.6931 0.3971 0.6931 0.6931 1.0618 2.4935
LOSS 0.3013 0.0000 0.4557 0.0000 0.0000 0.8182 1.0000
DEARN 0.0058 0.0053 0.1501 0.4820 0.0292 0.0380 0.6087
AFE 0.0168 0.0003 0.1670 0.3126 0.0037 0.0020 0.0625
AF 0.0404 0.0565 0.1455 0.3693 0.0315 0.0793 0.2045
JMBE_change 0.0697 0.0000 0.2531 0.0000 0.0000 0.0000 1.0000
JMBE_analyst 0.1313 0.0000 0.3375 0.0000 0.0000 0.0000 1.0000
RESTATEtþ1 0.0193 0.0000 0.1146 0.0000 0.0000 0.0000 0.6364
RESTATEtþ2 0.0364 0.0000 0.1534 0.0000 0.0000 0.0000 0.7273
RESTATEtþ3 0.0484 0.0000 0.1756 0.0000 0.0000 0.0000 0.7273
SEOtþ1 0.0984 0.0000 0.2954 0.0000 0.0000 0.0000 1.0000
M&Atþ1 0.1136 0.0000 0.3163 0.0000 0.0000 0.0000 1.0000
GRANTt 0.1455 0.0000 0.3226 0.0000 0.0000 0.0000 0.9091
GRANTtþ1 0.1550 0.0000 0.3308 0.0000 0.0000 0.0000 0.9000
SUE 0.1351 0.1246 1.6015 3.9172 0.7044 1.1037 3.7571
CR(1, þ1) 0.0025 0.0002 0.0597 0.1360 0.0269 0.0279 0.1889
CR(þ2, þ61) 0.0419 0.0255 0.2634 0.5386 0.1079 0.1646 0.8732
CR(þ2, þ121) 0.0607 0.0088 0.4606 0.7069 0.1869 0.2290 1.5891
NOA 0.6561 0.6560 0.3746 0.1323 0.4392 0.8201 1.9759

Variable Definitions (not previously provided in Table 1):

DA ¼ discretionary accruals calculated using the two-digit SIC industry cross-sectional modified Jones model;
Accruals ¼ (income before extraordinary items  cash flow from operations net of extraordinary items and discontinued
operations included in cash flows)/beginning total assets;
CFO ¼ operating cash flows/beginning total assets;
JMBE_change ¼ 1 if 0  change in earnings/beginning market value of equity , 0.005, and 0 otherwise;
JMBE_analyst ¼ 1 if 0  firm’s analysts’ consensus forecast error AFE , 0.01 (one cent), and 0 otherwise;
RESTATEtþi ¼ 1 if the firm restates its earnings due to irregularities within year tþi, i ¼ 1, 2, and 3, after earnings
announcements, and is 0 otherwise;
SEOtþ1 ¼ 1 when the Sale of Common and Pref. Stock (SSTK) one year after the earnings press release is greater than 10
percent of beginning total assets, and is 0 otherwise;
M&Atþ1 ¼ 1 if the amount of acquisition (AQC) in one year after the earnings press release is greater than 10 percent of
beginning total assets, and is 0 otherwise;
GRANTtþi ¼ 1 when the reported Black-Scholes fair value of stock options (from ExecuComp) granted to the CEO in
year tþi, i ¼ 0, 1 is greater than the median grants of that year in the sample, and is 0 otherwise;
SUE ¼ standardized unexpected earnings, calculated as the change from same quarter last year’s earnings scaled by its
standard deviations, calculated over previous 20 quarters data, with at least ten observations available. SUE is
winsorized at the value of 5;
(continued on next page)

The Accounting Review

May 2014
1094 Huang, Teoh, and Zhang

TABLE 2 (continued)
CR(1, þ1) ¼ three-trading day cumulative stock returns one trading day before to one trading day after earnings
announcements;
CR(þ2, þ61) ¼ 60-trading day cumulative stock returns starting the second day after earnings announcements;
CR(þ2, þ121) ¼ 120-trading day cumulative stock returns starting the second day after earnings announcements; and
NOA ¼ asset-scaled net operating assets at the beginning of the fiscal year. NOA is calculated according to Hirshleifer et
al. (2004).

dictated by quantitative performance, leaving ABTONE to represent only the discretionary

component. Therefore, a good discretionary tone proxy should have little or no correlation with
quantitative measures of firm fundamentals. We also prefer the discretionary tone to reflect strategic
incentives and so would like the measure to be unrelated to managerial style. Therefore, we want a
proxy that has low time-series persistence.14
Table 3, Panel A, reports the Spearman correlations of 16 firm fundamental characteristics with
ABTONE in column (1), TONE in column (2), and DTONE in column (3). The 16 fundamental
characteristics represent Dechow et al.’s (2010) three exhaustive dimensions of firm fundamental
values, and investment growth opportunity proxies, which are beginning asset-scaled R&D, and
beginning asset-scaled capital expenditure (CAPEX). The results show that ABTONE has much less
correlation with fundamentals. ABTONE is significantly correlated with six of the 16 characteristics,
whereas TONE and DTONE are significantly correlated with 13 and 11 firm characteristics,
respectively. Further, the statistically significant ABTONE correlations are much smaller in absolute
magnitude than the corresponding correlations involving TONE and DTONE.
Table 3, Panel B, reports the cross-sectional average of each individual firm’s time-series
persistence of ABTONE, TONE, and DTONE. To calculate average persistence, we regress
ABTONE, TONE, or DTONE on its respective one-year lagged variable for all cases with at least
five years of observations for each firm and then average the coefficients on the lagged variable
across all firms in the sample. The average persistence of ABTONE, TONE, and DTONE is 0.0509,
0.1016, and 0.3470 respectively.
In sum, we find that ABTONE has the smallest persistence and the smallest absolute
correlations with firm fundamentals when compared to TONE and DTONE. These results support
our choice of ABTONE as a better proxy for discretionary tone than the other two measures.

IV. DOES ABNORMAL POSITIVE TONE PREDICT FUTURE EARNINGS AND CASH
FLOWS?
By construction, ABTONE is unrelated to the current financial performance and other firm
characteristics. We next investigate whether it can identify the effects of strategic managerial
behavior by testing its ability to predict future financial performance incremental to the reported
financial numbers and controls. If ABTONE predicts positive future earnings and cash flows, then it
contains incremental managerial private information that cannot be conveyed through reported
earnings, owing to GAAP constraints. On the other hand, if ABTONE predicts no/negative future
earnings and cash flows, then managers likely use tone to simply hype or mask poor future
performance to mislead investors.
We examine the relation between ABTONE and future one- to three-year-ahead financial
performance as measured by either earnings or cash flows from operations in the following

14
We thank an anonymous referee for suggesting these tests. Unlike our research objective, some prior studies
discussed in Section II examine whether tone conveys fundamental information, so TONE or DTONE is sufficient
for their purposes.

The Accounting Review

May 2014
Tone Management 1095

TABLE 3
Abnormal Positive Tone Correlations with Firm Characteristics and Persistence
Panel A: Spearman Correlation between Tone and Firm Characteristics
Variable ABTONE TONE DTONE
EARN 0.0389 0.0561 0.0464
RET 0.0013 0.0057 0.0374
SIZE 0.0040 0.0928 0.0057
BTM 0.0262 0.0989 0.0343
STD_RET 0.0061 0.1663 0.0186
STD_EARN 0.0045 0.0345 0.0046
AGE 0.0001 0.1065 0.0153
BUSSEG 0.0026 0.1062 0.0015
GEOSEG 0.0084 0.0420 0.0008
LOSS 0.0035 0.0085 0.0385
DEARN 0.0210 0.0389 0.0750
AFE 0.0344 0.0365 0.0321
AF 0.0266 0.0507 0.0091
CFO 0.0083 0.0092 0.0191
R&D 0.0903 0.1492 0.0178
CAPEX 0.0205 0.0265 0.0101

Panel B: Persistence of Tone

ABTONE TONE DTONE
PERSISTENCE 0.0509 0.1016 0.3470

Bold numbers indicate significance at less than the 5 percent level.

This table reports the Spearman correlation coefficients of abnormal positive tone with firm characteristics and
persistence.
All other variables are as defined in Tables 1 and 2.

Variable Definitions:
R&D ¼ R&D expenditure scaled by beginning total assets; and
CAPEX ¼ capital expenditure scaled by beginning total assets.

regressions:

EARNjtþn ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt

þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð4Þ

where n ¼ (1, 2, or 3),

CFOjtþn ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt

þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð5Þ

where n ¼ (1, 2, or 3).

Table 4 presents the estimation results of Regression (4) in Panel A and Regression (5) in Panel
B. In addition to controlling for the two-digit SIC industry and year dummies, we present clustered

The Accounting Review

May 2014
 1096
TABLE 4
Abnormal Positive Tone and Future Financial Performance

Panel A: Future Earnings and Abnormal Positive Tone

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Adj. R2
EARNtþ1 0.0226 0.3021** 0.2617*** 0.6214*** 0.0030*** 0.0045 0.0170*** 1.5643*** 0.0569* 12,640 54.21%
t-stat (0.82) (2.42) (14.07) (13.87) (3.05) (1.05) (2.98) (4.95) (1.67)
EARNtþ2 0.0566 0.3935* 0.2363*** 0.4596*** 0.0035** 0.0023 0.0049 1.5719*** 0.0645* 11,542 38.21%
t-stat (1.25) (1.68) (5.81) (8.69) (2.00) (0.46) (0.68) (5.01) (1.93)
EARNtþ3 0.0208 0.6020* 0.2355*** 0.3782*** 0.0038*** 0.0056 0.0005 1.2872*** 0.0681 9,920 30.51%
t-stat (0.76) (1.69) (9.37) (7.93) (3.32) (1.06) (0.08) (4.69) (1.63)

Panel B: Future Cash Flows and Abnormal Positive Tone

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Adj. R2
CFOtþ1 0.0338 0.3609*** 0.3717*** 0.5680*** 0.0062*** 0.0033 0.0073* 0.9005*** 0.0336 12,624 54.19%
t-stat (1.16) (3.16) (20.45) (12.76) (7.18) (0.81) (1.74) (4.81) (1.33)
CFOtþ2 0.0498 0.6159*** 0.2965*** 0.4568*** 0.0044*** 0.0061 0.0046 1.0463*** 0.0331 11,528 43.49%
t-stat (1.26) (2.74) (18.60) (10.52) (3.94) (1.32) (1.12) (5.45) (1.00)
CFOtþ3 0.0713*** 0.5725* 0.2741*** 0.3803*** 0.0033*** 0.0103** 0.0008 0.9543*** 0.0373 9,908 35.63%
t-stat (3.08) (1.92) (9.23) (9.31) (4.18) (2.04) (0.16) (4.67) (0.89)

*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.

t-statistics based on two-way clustering at both the firm level and the year level are reported in parentheses.
This table reports the results of the regression of future performance on ABTONE and other control variables. The dependent variables are EARN in one to three years ahead in Panel
A, and CFO in one to three years ahead in Panel B. ABTONE is abnormal positive tone, measured as the residual from the annual cross-sectional regression in Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Industry and year fixed effects are included in the regressions, but are not reported.
All variables are as defined in Tables 1 and 2.

May 2014
The Accounting Review
Huang, Teoh, and Zhang
Tone Management 1097

t-statistics by firm and year to correct for cross-sectional and time-series dependence of errors in all
the relevant tests throughout the paper (Peterson 2009; Gow, Ormazabal, and Taylor 2010).15 For all
horizons from one to three years ahead, the coefficients of ABTONE in future earnings and cash flows
regressions are negative and significant. The ABTONE coefficients are 0.30, 0.39 and 0.60 for
one- to three-year-ahead EARN regressions, respectively. A one standard deviation increase (0.0068)
in ABTONE implies a decrease of 0.20 percent, 0.27 percent, and 0.41 percent in the one- to
three-year-ahead earnings (EARN) respectively. For comparison, the 0.41 percent decline amounts to
about 11 percent of the median earnings (EARN), which is 3.79 percent for the sample.16
For CFO regressions, the coefficients on ABTONE in Table 4, Panel B are 0.36, 0.62, and
0.57 for one- to three-year-ahead horizons, respectively. A one standard deviation increase in
ABTONE therefore translates to a decrease in asset-scaled CFO of 0.24 percent, 0.42 percent, and
0.39 percent, respectively. The 0.39 percent decline amounts to 5 percent of the median CFO of 8
percent for the sample.

V. ABNORMAL POSITIVE TONE IN STRATEGIC SETTINGS

The evidence above that abnormal positive tone is higher in firms with negative future
fundamentals raises the question of whether managers exploit tone opportunistically to misinform
investors about the negative future fundamentals instead of revealing useful private information.
Therefore, we investigate next whether abnormal positive tone is associated with the presence of
strong incentives to influence investor perceptions. Because incentives are not directly observable, we
use events that the past literature has documented as being associated with the presence of incentives
to either hype or depress the firm’s image. Events associated with incentives to bias perceptions
upward include just meeting or beating earnings benchmarks (JMBE), future earnings restatements
because earnings were previously manipulated, and major corporate transactions, such as mergers and
acquisitions (M&A) and seasoned equity offerings (SEO). For an event associated with incentives to
bias perceptions downward, we consider stock option grants. The purpose of these tests is not to show
that tone management causes these strategic events, but rather to demonstrate whether firms facing
incentives to manipulate perceptions resort to opportunistic tone management.17

15
As a robustness check, we also include R&D in the regression because Table 3 shows that ABTONE is correlated with
R&D. The results are qualitatively similar in sign, magnitude, and statistical significance of the estimated coefficients.
16
These results differ from the evidence that tone optimism predicts positive future earnings in Davis et al. (2012)
and Demers and Vega (2011). We note the following differences. Our sample uses annual earnings press releases
from both PR Newswire and Business Wire, whereas the other two studies examine quarterly earnings press
releases only from PR Newswire. Our larger sample size contains almost twice as many distinct firms. We read
every report headline to ensure that each observation in our sample is an announcement about earnings, not other
events. Demers and Vega (2011) delete observations where announcements of dividends or mergers/acquisitions
occurred within two weeks of the earnings press release. We do not rule out such firms because we are interested
in studying the relation between tone management and managerial strategic incentives. However, we do remove
earnings announcements with other concurrent news announcements. Finally, instead of TONE or DTONE, we
focus on ABTONE because Table 3 results indicate that TONE and DTONE contain useful information about firm
fundamentals and so are less suitable proxies for discretionary tone. We offer additional results about normal
tone, NTONE, the tone component that is related to fundamentals in the ‘‘Inclusion of Normal Tone in Test
Regressions’’ section. Table 11 reports that NTONE has a positive relation with future performance.
17
Our conversations with investor relations (IR) professionals and analysts indicate that the IR staff prepares the
initial draft of the earnings press release using information supplied by the finance department, key messaging
notes from the CEO and CFO, and guidance from the legal compliance team. When there are major messaging
issues, either an internal or external marketing/PR team is consulted to refine the message. The draft is then sent
up the management chain. After consultation with the IR director and the CFO, the CEO approves the draft
before the release. Therefore, the tone of the press release very probably reflects the CEO’s position. This is
analogous to past studies relating accruals management with CEO incentives, which generally assume that CEOs
influence the reported numbers without being the direct party ‘‘doing the accounting.’’

The Accounting Review

May 2014
1098 Huang, Teoh, and Zhang

Abnormal Positive Tone and Contemporaneous Just Meet or Beat Earnings Events
We consider settings in which managers may have manipulated earnings to just meet or beat
prior year’s earnings or the analysts’ consensus forecasts (e.g., Burgstahler and Dichev 1997;
Degeorge, Patel, and Zeckhauser 1999). In a related paper, Frankel et al. (2010) do not find that the
tone of conference calls is more negative for firms that just miss the analyst forecast by a penny than
firms that just meet or beat the analyst forecast by a penny. Some sections of conference calls are
spontaneous conversations, and so may be less likely to be strategically manipulated than the tone
in earnings press releases that could be strategically designed.
We run the following logistic regression to examine whether abnormal positive tone is
associated with a higher likelihood of just meeting or beating earnings (JMBE) benchmarks.18
JMBEjt ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt
þ b6 STD RETjt þ b7 STD EARNjt þ ejt : ð6Þ
The dependent variable, JMBE, is either an indicator variable, JMBE_change or JMBE_analyst,
defined as follows. JMBE_change is set to 1 if the change in net income from year t1 to t, scaled
by the beginning market value of equity is nonnegative, but less than 0.005, and is 0 otherwise.
JMBE_analyst is set to 1 if a firm’s consensus analyst forecast error is nonnegative, but smaller than
0.01 (one cent), and is 0 otherwise. The control variables are defined previously in Section III.
Table 5 presents estimation results in Panel A for the prior year’s earnings threshold and in
Panel B for the analysts’ forecast threshold. In both panels, ABTONE is associated with
significantly higher likelihood of JMBE, suggesting that tone management complements beating or
meeting earnings benchmarks to affect investor perception. The magnitudes of the economic effects
are significant. A one standard deviation increase in ABTONE increases the odds of reporting a very
small earnings by 9.2 percent and the odds of just beating or meeting analyst forecasts by 7.3
percent. Consistent with prior studies, we also find that bigger firms, growth firms, and those with
recent poor stock returns are more likely to just meet or beat thresholds.

Abnormal Positive Tone and Future Earnings Restatements

Next, we test whether tone management predicts future earnings restatements incrementally to
accruals management. It is well accepted that discretionary accruals are poor proxies of earnings
management, and the literature continues to debate whether firms manage accruals to just meet/beat
thresholds and during corporate transactions. Earnings restatements (irregularities), however, are
clear indicators of prior earnings manipulation using accruals. Therefore, the restatement sample is
especially pertinent to test whether tone management is used when firms have incentives to manage
perceptions. We run the following logistic regression:
RESTATEj;tþn ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt
þ b5 RETjt þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð7Þ
where n ¼ (1, 2, or 3).
The restatement data are from the GAO database from 1997 to 2006.19 Hennes, Leone, and
Miller (2008) classify restatements into innocuous accounting errors versus irregularities that likely

18
A reverse regression where ABTONE is the dependent variable and the incentive variable is the main independent
variable yields similar inferences as the regressions in Equations (6) through (10). A regression on ABTONE
when all incentive variables are added jointly as independent variables shows significance for all incentive
variables except those for option grants.
19
Our sample of observations in Table 6 stops in 2004 because we relate the abnormal positive tone to future
restatements up to three years subsequent to the earnings press release and the data for restatements end in 2006.

The Accounting Review

May 2014
 TABLE 5
Abnormal Positive Tone and Just Meeting or Beating Earnings Thresholds

May 2014
Panel A: Earnings Target 1—Scaled Earnings Changes
Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
Tone Management

JMBE_change 1.7958*** 12.9044** 0.6728 1.2302*** 0.1039*** 0.7969*** 0.4626** 13.3532** 3.4214*** 13,545 7.73%

The Accounting Review

p-value (0.0010) (0.0268) (0.1413) (0.0011) (0.0006) (0.0000) (0.0300) (0.0232) (0.0063)

Panel B: Earnings Target 2—Analysts’ Forecast

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
JMBE_analyst 3.9457*** 10.3595*** 0.0286 0.6097*** 0.0990*** 0.3153*** 0.1257** 4.4505 0.2211 13,569 3.39%
p-value (0.0002) (0.0031) (0.8645) (0.0021) (0.0000) (0.0015) (0.0323) (0.1232) (0.2881)

**, *** Indicates p , 0.05 and p , 0.01, respectively.

p-values are reported in parentheses.
This table shows the logistic regression results of just meeting or beating performance thresholds on ABTONE and control variables. In Panel A, the dependent variable is an
indicator variable for whether current earnings just meet or beat the prior earnings benchmark, JMBE_change. In Panel B, the dependent variable is an indicator variable for
whether current earnings just meet or beat the analysts’ consensus forecast, JMBE_analyst. The key independent variable, ABTONE, is measured as the residual from the annual
cross-sectional regression in Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Industry and year fixed effects are included in the regressions, but are not reported. Standard errors are based on two-way clustering at both the firm level and the year level.
All other variables are as defined in Tables 1 and 2.
1099
1100 Huang, Teoh, and Zhang

stem from an earlier earnings manipulation. Therefore, for a cleaner test for the association with
earnings manipulation, we focus only on the irregularities sample. Since not all restatements occur
shortly after earnings press releases, we study the likelihood of restatements in one-year, two-year,
and three-year horizons after the earnings press releases, labeled RESTATEtþ1, RESTATEtþ2, and
RESTATEtþ3, respectively. After matching restatement data with our original dataset, we have
samples of 182, 342, and 460 firms restating their financial statements due to irregularities in the
one-, two-, and three-year horizons after earnings press releases, respectively.
We present the estimation results of Regression (7) in Table 6. The coefficients on ABTONE are
positive in all three horizon regressions, and statistically significant at conventional levels for the two-
year and the three-year horizon regressions.20 To measure the economic effect of ABTONE on future
earnings restatement, we calculate the marginal effect of the probability of restating earnings when
abnormal positive tone changes by one standard deviation, holding all other independent variables at
their means. We find that a one standard deviation increase in ABTONE increases the odds of a future
restatement in two years by 11.3 percent and in three years by 9.8 percent. With regard to the control
variables, we find that larger firms and value firms are more likely to restate earnings. Bigger firms are
more likely to attract scrutiny from regulators (Lee, Li, and Yue 2006). In sum, our results show that
firms with higher abnormal positive tone are more likely to restate earnings due to reporting
irregularities in the two to three years after the earnings press release.

Abnormal Positive Tone and Corporate Transactions

We next examine whether ABTONE is positively related to two major corporate transactions: a
seasoned equity offering (SEO) and mergers and acquisitions (M&A). Prior studies have
documented that these transactions create settings where firms have manipulated investor
perceptions upward through earnings numbers (e.g., Teoh et al. 1998; Erickson and Wang
1999). We estimate the following two logistic regressions:
SEOtþ1 ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt
þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð8Þ

M&Atþ1 ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt

þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð9Þ
where:
SEOtþ1 ¼ a dummy variable that is set to 1 when the Sale of Common and Preferred Stock
(SSTK) in one year after an earnings press release is greater than 10 percent of lagged total
assets, and is 0 otherwise; and
M&Atþ1 ¼ 1 if the amount of acquisition (AQC) in one year after an earnings press release is
greater than 10 percent of lagged total assets, and is 0 otherwise.
When the Compustat items are missing, we set them to 0. There are 1,476 SEOtþ1 observations and
1,547 M&Atþ1 observations that are equal to 1.
Panel A of Table 7 presents the results for SEO. The ABTONE coefficient is positive and
significant. The effect of ABTONE is also economically significant with a one standard deviation
increase in ABTONE increasing the probability of an SEO by 9 percent. These results are consistent
with the hypothesis that managers deploy tone management when announcing earnings prior to a

20
The weaker one-year results suggest that restatements take time to manifest from when perceptions are first
managed upward. The sample contains only 186 irregularities in the one-year horizon compared with 460
restatements in the three-year horizon.

The Accounting Review

May 2014
 TABLE 6
Abnormal Positive Tone and Future Earnings Restatements

May 2014
Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
RESTATEtþ1 5.3597*** 5.7576 0.0462 0.0373 0.2075*** 0.2977*** 0.0635 17.4791 0.9209 8,482 5.42%
Tone Management

p-value (0.0000) (0.2890) (0.8956) (0.9369) (0.0000) (0.0022) (0.3945) (0.1351) (0.2511)
4.6959*** 15.7989** 0.3392 0.0094 0.1864*** 0.2488*** 0.0216 9.2212 0.5615 8,784 4.77%

The Accounting Review

RESTATEtþ2
p-value (0.0000) (0.0383) (0.5067) (0.9811) (0.0000) (0.0002) (0.7696) (0.3706) (0.5020)
RESTATEtþ3 4.2802*** 13.7990* 0.1498 0.2493 0.1614*** 0.1854*** 0.0841 3.0131 0.2886 8,901 5.05%
p-value (0.0000) (0.0763) (0.6806) (0.4231) (0.0000) (0.0070) (0.2491) (0.7514) (0.6943)

*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.

p-values are reported in parentheses.
This table shows the logistic regression results of future earnings restatements on ABTONE and other control variables. The dependent variables are the restatements in the next
one to three years after earnings announcements. Our sample stops at year 2004 because we examine the future restatements up to three years head. For simplicity, industry and
year dummies are included in the regression, but not reported in the table. ABTONE is abnormal positive tone, measured as the residual from the annual cross-sectional
regression Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Standard errors are based on two-way clustering at both the firm level and the year level.
All other variables are as defined in Tables 1 and 2.
1101
 TABLE 7
1102

Abnormal Positive Tone and Corporate Transactions

Panel A: Abnormal Positive Tone and SEOtþ1
Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
SEOtþ1 1.2456*** 12.7428* 1.1465*** 2.2816*** 0.2030*** 1.7589*** 0.4831*** 5.8325** 1.9857*** 13,178 24.40%
p-value (0.0007) (0.0644) (0.0020) (0.0000) (0.0000) (0.0000) (0.0000) (0.0208) (0.0000)

Panel B: Abnormal Positive Tone and M&Atþ1

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
M&Atþ1 0.5925 13.5208** 1.0933*** 2.9978*** 0.0555*** 0.3501*** 0.1335*** 18.6079*** 0.1697 13,546 7.48%
p-value (0.1164) (0.0206) (0.0061) (0.0000) (0.0074) (0.0000) (0.0000) (0.0000) (0.7679)

Panel C: Abnormal Positive Tone and SEOtþ1_121 Days

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
SEOtþ1_121 days 14.0381*** 25.2717** 0.1700 0.9147*** 0.1803*** 1.1692** 0.4755*** 15.5857 0.5106 12,177 11.70%
p-value (0.0000) (0.0120) (0.6960) (0.0054) (0.0006) (0.0150) (0.0000) (0.1029) (0.5461)

Panel D: Abnormal Positive Tone and M&Atþ1_121 Days

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
M&Atþ1_121days 3.0529*** 9.6628** 1.0175*** 2.5670*** 0.0500*** 0.2481** 0.1563*** 14.3050*** 0.3062 13,411 5.65%
p-value (0.0007) (0.0212) (0.0038) (0.0002) (0.0060) (0.0197) (0.0038) (0.0022) (0.7013)
*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.
p-values are reported in parentheses.
This table presents logistic regression results of stock issuance activities on abnormal positive tone. The dependent indicator variables in Panels A and B are for equity issuance and
M&A activities, respectively, within one year after the earnings announcement, whereas those in Panels C and D are for events occurring within 121 trading days of earnings
announcements. ABTONE is measured as the residual from the annual cross-sectional regression in Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Industry and year fixed effects are included in the regressions, but are not reported. Standard errors are based on two-way clustering at both firm level and year level.
All other variables are as defined in Tables 1 and 2.

May 2014
The Accounting Review
Huang, Teoh, and Zhang
Tone Management 1103

stock issuance to incite greater excitement about the firm so as to obtain a better price for the newly
issued shares. We also find that the coefficient on discretionary accruals, DA, in Panel A is positive
and significant, consistent with Teoh et al.’s (1998) finding of accruals management prior to an equity
issuance. Small firms and firms with low earnings, more growth opportunities (low BTM), higher
momentum in stock returns over the year prior to earnings press releases (high RET), and firms with
higher stock return or earnings volatilities are also more likely to issue stock in the subsequent year.
Panel B of Table 7 presents the estimation results of Regression (9) when the dependent variable
is M&Atþ1. The results are similar to those for SEOs because ABTONE is significantly positively
associated with undertaking M&A activities in the immediate future. A one standard deviation
increase in ABTONE is associated with an increase in the frequency of M&A of 10 percent. The
results suggest that tone management often accompanies acquisition activities. In sum, the SEO and
M&A test results are both consistent with the hypothesis that managers strategically use disclosure
tones to influence investors’ perception positively prior to major corporate transactions.

Abnormal Positive Tone and Stock Option Grants

This subsection examines stock option grants, a setting in which executives have incentive to
manipulate perceptions downward. Prior studies show that firms strategically disclose bad news
prior to stock option grants to reduce the stock price to ensure a lower option strike price at the grant
date (Aboody and Kasznik 2000; Baker, Collins, and Reitenga 2003; McAnally, Srivastava, and
Weaver 2008). We test whether abnormal positive tone is negatively related to stock option grants
by estimating the following logistic regression:
GRANTtþi ¼ a þ b0 ABTONEjt þ b1 DAjt þ b2 EARNjt þ b3 SIZEjt þ b4 BTMjt þ b5 RETjt
þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð10Þ
for i ¼ 0, 1.
GRANT is set to 1 when the reported Black-Scholes fair value of stock options granted to the
CEO that year is higher than the median of the year in the sample. We select larger grants because
the incentives to bias perceptions downward are expected to be stronger for large grants than for
small continuous grants. Because grants are not awarded by every firm or in every year, we set
missing values to 0.
There are 2,008 firm-year observations with contemporaneous-period large grants, and 1,460
large grants in the following year. Aboody and Kaznik (2000) report that 40 percent of all option
grants are awarded in the months of December, January, and February, which may be just before or
contemporaneously announced in the earnings press release. We examine both contemporaneous
and future stock option grants because we do not have the specific grant date.
The results for the contemporaneous relation between ABTONE and GRANTt are in Panel A of
Table 8 and for the one-year-ahead GRANTtþ1 in Panel B. The coefficient on ABTONE is negative
and significant in both panels and the effects are economically significant. A one standard deviation
increase in ABTONE increases the likelihood of option grants in the contemporaneous (next) year of
9.1 percent (8.2 percent). These results are consistent with the hypothesis that managers
strategically bias perceptions downward using tone in the earnings press release prior to an option
grant. We also find that the coefficient on discretionary accruals is negative and significant in Panel
B, consistent with Baker et al. (2003).

VI. IMMEDIATE AND DELAYED MARKET REACTIONS TO ABNORMAL POSITIVE

TONE
Thus far, we have established that managers manipulate tone incrementally to manipulating
accruals when facing various strategic incentives. The ultimate goal of such strategic actions is to

The Accounting Review

May 2014
 TABLE 8
1104

Abnormal Positive Tone and Option Grants

Panel A: Abnormal Positive Tone and Contemporaneous Option Grant
Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
GRANTt 23.6171*** 12.7638*** 0.5838 1.4071*** 0.7170*** 0.1413 0.0648 0.6060 0.1095 12,623 24.10%
p-value (0.0000) (0.0043) (0.1673) (0.0000) (0.0000) (0.1068) (0.2124) (0.8724) (0.7976)

Panel B: Abnormal Positive Tone and Future Option Grant

Dep. Var. a ABTONE DA EARN SIZE BTM RET STD_RET STD_EARN No. Obs. Pseudo R2
GRANTtþ1 22.4082*** 11.6028*** 1.8387*** 1.7984*** 0.7339*** 0.0988 0.0141 0.5593 0.0382 8,435 25.70%
p-value (0.0000) (0.0040) (0.0000) (0.0000) (0.0000) (0.4235) (0.7593) (0.8422) (0.9411)
*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.
p-values are reported in parentheses.
Standard errors are based on two-way clustering at both firm level and year level. This table presents logistic regression results of stock option grant event in the same year in Panel
A and in the following year after the earnings announcement in Panel B on abnormal positive tone. ABTONE is measured as the residual from the annual cross-sectional regression
in Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Industry and year fixed effects are included in the regressions, but are not reported.
All other variables are as defined in Tables 1 and 2.

May 2014
The Accounting Review
Huang, Teoh, and Zhang
Tone Management 1105

influence stock valuations. Therefore, we next test whether investors can see through tone
management. We examine the stock price reactions to ABTONE in the short window around
earnings announcements, and in the longer horizon of one and two quarters after earnings
announcements.

Market Reaction to ABTONE at the Time of Earnings Announcements

Section IV documents that ABTONE is negatively related to future earnings and cash flow from
operations. If investors anticipate even partially strategic tone hyping, then they would discount for
ABTONE, and so we would predict a negative return response to ABTONE at the earnings
announcement. In contrast, if managers succeed in misleading investors by inciting over-optimism
using tone manipulation, then we would expect a positive response at earnings announcements. We
run the following regression:
CR½1; þ1 ¼ a þ b0 RABTONEjt þ b1 RDAjt þ b2 RSUEjt þ b3 SIZEjt þ b4 BTMjt
þ b5 RETjt þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð11Þ
where SUEjt ¼ firm j’s current quarterly earnings minus earnings of same quarter last year, scaled by
market value of the beginning of the quarter. The dependent variable is the three-day cumulative
returns from one trading day before to one trading day after the earnings announcement. We include
DA and SUE as proxies for the quantitative news. To gauge economic significance more easily, we
use annual decile ranks for the key independent variables, RDA, RSUE, and RABTONE (Bernard
and Thomas 1990). The decile rankings (1 to 10) are reduced by 1 and then divided by 9 so as to
range between 0 and 1. Thus, the slope of coefficients can be viewed as abnormal returns to zero-
investment portfolios.
We present the estimation results of Regression (11) in Panel A of Table 9. Consistent with the
literature on earnings response coefficients, we find that the contemporaneous return response to
earnings news (RSUE) is significantly positive (t ¼ 10.01). Stock returns are higher for small, value,
and less volatile returns firms. The response to abnormal accruals is significantly negative. For our
key test variable, RABTONE, the coefficient of 59 basis points is significantly positive (t ¼ 2.19),
and economically meaningful.
Thus, investors do not appear to discount for the negative information about future
performance contained in ABTONE. This raises the possibility that the higher stock prices
associated with tone management reflect over-valuation at the time of earnings announcements that
will subsequently be reversed. We test this possibility in the next subsection.

Delayed Market Reaction after Earnings Announcements

Next, we formally test whether investors are misled by abnormal positive tone at the time of
earnings announcements. If this is the case, then we expect that ABTONE negatively predicts future
stock returns as investors correct their initial pricing errors gradually as more information about
fundamentals is released over time. We run the following regressions:
CR½þ2; þ61 ¼ a þ b0 RABTONEjt þ b1 RDAjt þ b2 RSUEjt þ b3 SIZEjt þ b4 BTMjt
þ b5 RETjt þ b6 STD RETjt þ b7 STD EARNjt þ ejt ; ð12Þ

CR½þ2; þ121 ¼ a þ b0 RABTONEjt þ b1 RDAjt þ b2 RSUEjt þ b3 SIZEjt þ b4 BTMjt

þ b5 RETjt þ b6 STD RETjt þ b7 STD EARNjt þ ejt : ð13Þ
The dependent variable is the cumulative returns one quarter and two quarters after the
earnings announcements for Regression (12) in Panel B and Regression (13) in Panel C of Table 9,

The Accounting Review

May 2014
 1106
TABLE 9
Market Immediate and Delayed Reactions to Abnormal Positive Tone
Panel A: Immediate Stock Return Reactions during the Earnings Announcements
Dep. Var. a RABTONE RDA RSUE SIZE BTM RET STD_RET STD_EARN No. Obs. Adj. R2
CR (1, þ1) 0.0303 0.0059** 0.0084*** 0.03005*** 0.0010** 0.0094*** 0.0165*** 0.3426*** 0.0190 13,060 3.76%
t-stat (1.59) (2.19) (3.03) (10.01) (2.53) (3.84) (5.26) (3.37) (1.36)

Panel B: Delayed One Quarter Return Reactions after Earnings Announcements

Dep. Var. a RABTONE RDA RSUE SIZE BTM RET STD_ RET STD_EARN No. Obs. Adj. R2
CR (þ2, þ61) 0.0262 0.0212*** 0.0325*** 0.0025 0.0061* 0.0487*** 0.0476*** 1.1566 0.1088 13,062 6.51%
t-stat (0.36) (2.63) (4.43) (0.27) (1.91) (4.98) (3.46) (0.98) (1.67)

Panel C: Delayed Two Quarter Return Reactions after Earnings Announcements

Dep. Var. a RABTONE RDA RSUE SIZE BTM RET STD_ RET STD_EARN No. Obs. Adj. R2
CR (þ2, þ121) 0.1232 0.0384*** 0.0411* 0.0068 0.0183** 0.0506*** 0.0485** 0.0295 0.1904 12,978 9.27%
t-stat (0.87) (2.79) (1.88) (0.64) (2.02) (2.68) (2.13) (0.01) (1.51)

*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.

t-statistics based on two-way clustering at both the firm level and the year level are reported in parentheses.
This table presents regression results of the market immediate and delayed reactions to abnormal positive tone. RABTONE, RDA, and RSUE are annual decile ranks for ABTONE,
DA, and SUE, respectively. In Panel A, the dependent variable CR (1,þ1) is the three-trading-days cumulative returns surrounding the earnings announcements. In Panel B, the
dependent variable CR (þ2, þ61) is the 60-trading-days cumulative stock returns starting the second day after earnings announcements. In Panel C, the dependent variable CR (þ2,
þ121) is the 121-trading-days cumulative stock returns starting the second day after earnings announcements. ABTONE is measured as the residual from the annual cross-sectional
regression in Table 1:

TONEjt ¼ a þ b0 EARNjt þ b1 RETjt þ b2 SIZEjt þ b3 BTMjt þ b4 STD RETjt þ b5 STD EARNjt þ b6 AGEjt þ b7 BUSSEGjt þ b8 GEOSEGjt þ b9 LOSSjt þ b10 DEARNjt
þ b11 AFEjt þ b12 AFjt þ ejt :
Industry and year fixed effects are included in the regressions, but are not reported.
All other variables are as defined in Tables 1 and 2.

May 2014
The Accounting Review
Huang, Teoh, and Zhang
Tone Management 1107

respectively. In both panels, RSUE coefficients are positive but insignificant and RDA coefficients
are negative and significant. The general signs for the coefficients on these variables are consistent
with previous literature, and the varying statistical significance is also consistent with weak post-
earnings-announcement drift (PEAD) and discretionary accruals anomalies in more recent periods.
Consistent with the broad asset-pricing literature, SIZE is negatively related, and BTM and return
momentum are positively related with future stock returns over various horizons.
Turning to our key variable, Table 9 shows that RABTONE significantly predicts negative stock
returns both one quarter ahead (Panel B, t ¼ 2.63) and two quarters ahead (Panel C, t ¼
2.79).21,22 The magnitudes of the coefficients are also economically significant, and comparable in
magnitude with the strength of the discretionary accruals and PEAD anomalies. The abnormal
returns from tone management are 2.12 percent for one quarter (8.48 percent annualized) and 3.84
percent for two quarters (7.68 percent annualized). The evidence is consistent with the hypothesis
that abnormal positive tone misleads investors at the time of earnings announcements to temporarily
over-value the firm and the market subsequently corrects the mispricing.

VII. ADDITIONAL ANALYSIS

We first offer several robustness checks on all of the previous results. Then, we examine how
tone management and accruals management may be related.

Alternative Dictionary Lists

We examine robustness with respect to using two alternative dictionary lists to classify words
with optimistic or pessimistic tone, Henry (2008) and Harvard’s GI modified by removing
accounting terms from the positive and negative word list. Columns (1) and (2) of Table 10 show
that the majority of the results are generally robust. Note that an insignificant relation between
ABTONE and future performance is inconsistent with the hypothesis that discretionary tone
informs, but is still consistent with the hypothesis that discretionary tone misinforms investors. For
the Henry list, the coefficients have the expected signs in all regressions, and are statistically
significant in some regressions, but not others. For the modified GI list, ABTONE significantly
negatively predicts future earnings and cash flows. The coefficients in the other regressions again
have expected signs, and some are statistically significant and some are not. In sum, the inference
that abnormal tone misinforms is robust to using alternative word lists.23

Alternative Tone Models

In our main tone model (1), we include current returns (RET), the book-to-market ratio
(BTM), and analysts’ forecasts of one-year-ahead earnings to capture concurrent quantitative
information about expected future performance in the earnings press release. As a robustness
check, columns (3) and (4) of Table 10 report results using abnormal positive tone estimated

21
RABTONE does not significantly predict abnormal returns that are three- and four-quarters ahead. The shorter
predictability period for future abnormal returns than for future accounting performance is consistent with
evidence in the literature that stock returns anticipate fundamentals (Beaver, Lambert, and Morse 1980).
22
Demers and Vega (2011) report a positive relation between DTONE and the 60-trading day post-announcement-
period abnormal returns. In untabulated tests, we find that DTONE is negatively related to future restatements,
positively related to SEO, and incrementally unrelated to accounting performance, stock returns, JMBE, M&A,
and option grants. However, analogous to why, in studies of accruals, DACCRUALS is a poor measure of
discretionary accruals, DTONE, being confounded with firm fundamentals (correlations reported in Table 3), is
less suitable than ABTONE as a proxy for tone management.
23
Since neither the LM nor Henry list is perfect, we conduct the full set of analyses also using a combined LM and
Henry list. All inferences are robust (untabulated).

The Accounting Review

May 2014
1108 Huang, Teoh, and Zhang

TABLE 10
Regression Results Summary of Alternative Word Lists and Alternative Positive Tone
Models
(4)
(3) Include
Include Management
(1) (2) Management Forecast
Tests Expected Sign Henry Modified GI Forecast Dummy
Future Performance
EARNtþ1 0/ 0.0436 0.1943*** 0.1153 0.3070**
EARNtþ2 0/ 0.1492 0.2217*** 0.1237 0.3948*
EARNtþ3 0/ 0.1347 0.3638*** 0.0316 0.6015*
CFOtþ1 0/ 0.0028 0.2377*** 0.0885 0.3698***
CFOtþ2 0/ 0.0552 0.3161*** 0.4322* 0.6214***
CFOtþ3 0/ 0.1173 0.2312** 0.0627 0.5747*
JMBE
JMBE_change þ 8.9547*** 4.1989 16.4002 13.1827**
JMBE_analyst þ 14.2192*** 6.2672*** 16.4215** 10.1705***
Restatement
RESTATEtþ2 þ 14.5136* 4.8147 28.3767 14.9071**
RESTATEtþ3 þ 11.9562* 2.9318 19.8684 13.2451*
Financial Transactions
SEOtþ1 þ 5.8115* 3.0563 16.4343 12.6447*
M&Atþ1 þ 6.7901** 4.6273* 30.3183*** 13.4740**
Option Grants
GRANTt  8.5586*** 4.3391* 6.5185 13.0589***
GRANTtþ1  4.9362 1.7691 15.2248* 12.0947***
Returns
CR (1, þ1) þ 0.0045** 0.0021 0.002 0.0059**
CR (þ2, þ61)  0.0084 0.0085 0.0031** 0.0227***
CR (þ2, þ121)  0.0301* 0.0122 0.0323 0.0404***

*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.

t-statistics are based on two-way clustering at both the firm level and the year level.
This table presents a summary of regression results using alternative word lists and alternative tone models. All controls,
industry, and year fixed effects are included in the regressions, but are not reported.
All other variables are as defined in Tables 1 and 2.

from a tone model that also further includes managerial forecasts of one-year-ahead earnings as a
direct proxy of managerial expectations of future performance. Managerial forecasts are from the
First Call CIG dataset, and their inclusion drastically reduces the sample size by about 75
percent. We use either the point forecast or the midpoint if the manager issues a range forecast,
or else an indicator variable for whether the manager issued a forecast. As before, all coefficients
are in the expected direction. Only some test regression coefficients are statistically significant
when managerial forecast values are included in the tone model, whereas all test regression
coefficients are statistically significant when the managerial forecast indicator variable is used
instead in the tone model. Therefore, the general inference that abnormal positive tone
misinforms remain robust to alternative tone models.

The Accounting Review

May 2014
Tone Management 1109

TABLE 11
Regression Results Summary of Including Normal Tone
Expected Sign of
Tests ABTONE ABTONE NTONE
Future Performance
EARNtþ1 0/ 0.0021** 0.0064*
EARNtþ2 0/ 0.0031** 0.0008
EARNtþ3 0/ 0.0046** 0.0007
CFOtþ1 0/ 0.0026*** 0.0090**
CFOtþ2 0/ 0.0047*** 0.0056
CFOtþ3 0/ 0.0044** 0.0063
JMBE
JMBE_change þ 0.0866** 0.3610***
JMBE_analyst þ 0.0730*** 0.1534***
Restatement
RESTATEtþ2 þ 0.1046* 0.1082
RESTATEtþ3 þ 0.0915 0.0746
Financial Transactions
SEOtþ1 þ 0.0792* 0.2412***
M&Atþ1 þ 0.0846** 0.1321**
Option Grants
GRANTt  0.0822*** 0.0246
GRANTtþ1  0.0738*** 0.0117
Returns
CR (1, þ1) þ 0.0065** 0.0030
CR (þ2, þ61)  0.0223*** 0.0048
CR (þ2, þ121)  0.0354*** 0.0075

*, **, *** Indicates p , 0.10, p , 0.05, and p , 0.01, respectively.

t-statistics are based on two-way clustering at both the firm level and the year level.
This table presents regression results when expected tone, NTONE, is included as an additional regressor to all
regressions in Tables 4 through 9. Only coefficients for NTONE and ABTONE are reported for comparison. NTONE is
the expected value and ABTONE is the residual from the annual cross-sectional regression in Table 1 of TONE on a set of
determinants. All controls, industry, and year fixed effects are included in the regressions, but are not reported.
All other variables are as defined in Tables 1 and 2.

Inclusion of Normal Tone in Test Regressions

We add normal tone, NTONE, to the earlier test regressions to examine whether normal tone and
abnormal positive tone affect the dependent variables differently. Table 11 tabulates the coefficients of
NTONE and ABTONE and indicates their statistical significance. The results show that ABTONE
remains statistically significant in all test regressions, which is expected since NTONE and ABTONE
are orthogonal by design. The NTONE coefficient is positively significantly related to future one-year-
ahead earnings and cash flows, consistent with normal tone informing about future fundamentals.
Recall that NTONE is a linear combination of current reported quantitative items, and so these results
suggest that earnings and cash flows are persistent, on average, in our sample. NTONE is also
significant and positive in the JMBE, SEO, and M&A regressions but not for the option grant, future
restatements. Incremental to other quantitative fundamentals included directly in the test regressions,

The Accounting Review

May 2014
1110 Huang, Teoh, and Zhang

NTONE is unrelated to immediate announcement returns and future returns, suggesting that investors
price NTONE and ABTONE differently.24

Seemingly Unrelated System Limited Dependent Regressions

The wide variety of strategic events that we examine is likely to be correlated in a given firm.
For instance, firms likely to grant an above-average number of stock options to CEOs are also more
likely growth firms and to issue new equity. Therefore, the regression residuals are likely to be
correlated across the various regressions of strategic events that we consider. A SUR method that
takes advantage of the correlation in residuals can improve estimation efficiency.25 Therefore, as a
robustness check, we run a SUR system of probit regressions to allow for both the cross-sectional
and the time-series dependence of residuals for the following groups of event regressions: SEO and
M&A, M&A and option grants, and SEO and option grants. The results using the SUR system of
equations are all quantitatively and qualitatively similar to the results from individual probit
regressions; thus, our results are robust to residual dependence across regressions.

Restricting SEOs to within 121 Trading Days of Earnings Announcements

The earlier results show that abnormal positive tone effects on future abnormal returns last
about 121 trading days after the earnings announcement. If managers manipulate tone to facilitate
strategic events, then we would expect managers to undertake these events within the 121-day time
frame before the tone effects dissipate. We therefore refine our earlier tests and restrict SEO and
M&A events to those occurring within 121 days of the earnings announcements as a robustness
check. We obtain SEO and M&A transaction date data from Thompson’s SDC database. Table 7,
Panels C and D report that the ABTONE results are robust to this alternative specification.26

A Pilot Analysis of the Joint Use of Tone Management and Earnings Management
Throughout the paper, we consider qualitative and quantitative management as two separate
perception management tools, but they may be related either as complements or substitutes. In our
sample, we find that the correlation between DA and ABTONE is a statistically significant 0.02,27
suggesting that managers often use both tools simultaneously.
We regress the ratio ABTONE/DA on the set of 16 firm characteristics for cases when tone
management and accruals management are in the same direction, and the results, untabulated for
brevity, indicate that the ratio is correlated only with AGE. Thus, tone is more persuasive for
investors in older firms with more established reputations.
Finally, we examine the correlation of ABTONE/DA with asset-scaled net operating assets at
the beginning of the fiscal year (NOA), a proxy for the limits to accruals management (Barton and
Simko 2002; Hirshleifer, Hou, Teoh, and Zhang 2004; Das et al. 2011). The Spearman correlation

24
To facilitate comparison, we standardize NTONE and ABTONE when including both variables in the regression.
The results suggest that normal tone is priced appropriately, whereas abnormal tone is mispriced.
25
We thank an anonymous referee for this suggestion. Since SAS does not support a SUR system of logistic
regressions, we verify that the individual event probit regressions indeed provide similar results as the individual
event logistic regressions.
26
A similar concern exists for option grants. Unfortunately, we do not have data on dates of stock option award and
so we are unable to check robustness for stock option award.
27
The small magnitude may be because DA and ABTONE have large measurement errors. Prior literature also finds
that accruals management and expectation management have similar magnitudes of correlation (Das, Kim, and
Patro 2011).

The Accounting Review

May 2014
Tone Management 1111

is significantly positive at 0.077, suggesting that when managers are constrained in manipulating
accruals due to the balance sheet constraint, they are more likely to resort to tone management.

VIII. CONCLUSION
Earnings press releases are an important venue in which to study pricing effects of accounting
information. Beyond disclosing the quantitative information, these press releases also contain
qualitative text to help investors interpret the quantitative information and to market the firm to
investors. The tone in these qualitative disclosures is important in influencing investors’
assessments about the value of the firm.
We analyze how managers use tone in the earnings press releases either to inform by clarifying
accompanying quantitative information or signaling additional private information that cannot be
incorporated into current quantitative results according to accounting GAAP requirements, or to
misinform by masking poor future financial performance. The evidence is consistent with tone
management misinforming investors.
We find that abnormal positive tone predicts negative future earnings and cash flows. We
further find that abnormal positive tone is more positive when firms have strong incentives to bias
investor perceptions upward. Abnormal positive tone is usually higher when firms just meet or beat
past earnings or analysts’ consensus forecasts, when earnings are upwardly biased to such an extent
as to require a future restatement, and before a new equity issuance or a merger or acquisition
activity. In contrast, when firms award stock options to CEOs, with an associated managerial
incentive to reduce the share price, they prefer to manipulate abnormal tone downward.
Our evidence indicates that tone manipulation succeeds in misleading investors, and that this
effect is incremental to the effect of accruals management. An abnormally positive tone incites an
overly optimistic immediate stock price response to the earnings announcement and a subsequent
return reversal. Finally, we document that firms that engage in tone management tend to be older
firms rather than younger firms, and they tend to have more bloated balance sheets. Overall, our
evidence indicates that abnormal positive tone contains negative information about future firm
fundamentals, that firms tend to engage in tone management particularly when incentives to
manipulate investor perceptions are high, and that investors are misinformed by tone management.

REFERENCES
Aboody, D., and R. Kasznik. 2000. CEO stock option award and the timing of corporate voluntary
disclosures. Journal of Accounting and Economics 29: 73–100.
Baker, T., D. Collins, and A. Reitenga. 2003. Stock option compensation and earnings management
incentives. Journal of Accounting, Auditing, and Finance 18: 557–582.
Barton, J., and P. Simko. 2002. The balance sheet as an earnings management constraint. The Accounting
Review 77 (Supplement): 1–27.
Beaver, W., R. Lambert, and D. Morse. 1980. The information content of security prices. Journal of
Accounting and Economics 2: 3–28.
Bernard, V., and J. Thomas. 1990. Evidence that stock prices do not fully reflect the implications of current
earnings for future earnings. Journal of Accounting and Economics 13: 305–340.
Bonsall, S., Z. Bozanic, and P. Fischer. 2013. The Informativeness of Disclosure Tone. Working paper, The
Ohio State University.
Burgstahler, D., and I. Dichev. 1997. Earnings management to avoid earnings decreases and losses. Journal
of Accounting and Economics 24: 99–126.
Campbell, J., H. Chen, D. Dhaliwal, H. Lu, and L. Steele. 2014. The information content of mandatory risk
factor disclosures in corporate filings. Review of Accounting Studies 19: 396–455.

The Accounting Review

May 2014
1112 Huang, Teoh, and Zhang

Das, S., K. Kim, and S. Patro. 2011. An analysis of managerial use and market consequences of earnings
management and expectation management. The Accounting Review 86: 1935–1967.
Davis, A., J. Piger, and L. Sedor. 2012. Beyond the numbers: Managers’ use of optimistic and pessimistic
tone in earnings press releases. Contemporary Accounting Research 29: 845–868.
Davis, A., and I. Tama-Sweet. 2012. Managers’ use of language across alternative disclosure outlets:
Earnings press releases versus MD&A. Contemporary Accounting Research 29: 804–837.
Dechow, P., R. Sloan, and A. Sweeney. 1995. Detecting earnings management. The Accounting Review 70:
193–225.
Dechow, P., W. Ge, and C. Schrand. 2010. Understanding earnings quality: A review of the proxies, their
determinants and their consequences. Journal of Accounting and Economics 50: 344–401.
De Franco, G., O. Hope, D. Vyas, and Y. Zhou. 2013. Analyst Report Readability. Working paper,
University of Toronto.
Degeorge, F., J. Patel, and R. Zeckhauser. 1999. Earnings management to exceed thresholds. Journal of
Business 72: 1–33.
Demers, E., and C. Vega. 2011. Linguistic Tone in Earnings Announcements: News Or Noise? Working
paper, INSEAD.
Erickson, M., and S. Wang. 1999. Earnings management by acquiring firms in stock-for-stock mergers.
Journal of Accounting and Economics 27: 149–176.
Feldman, R., S. Govindaraj, J. Livnat, and B. Segal. 2009. Management’s tone change, post earnings
announcement drift and accruals. Review of Accounting Studies 15: 915–953.
Fiske, S., and S. Taylor. 1991. Social Cognition. 2nd edition. New York, NY: McGraw-Hill.
Frankel, R., W. Mayew, and Y. Sun. 2010. Do pennies matter? Investor relations consequences of small
negative earnings surprises. Review of Accounting Studies 15: 220–242.
Gow, I., G. Ormazabal, and D. Taylor. 2010. Correcting for cross-sectional and time-series dependence in
accounting research. The Accounting Review 85: 483–512.
Hayn, C. 1995. The information content of losses. Journal of Accounting and Economics 20: 125–153.
Hennes, K., A. Leone, and B. Miller. 2008. The importance of distinguishing errors from irregularities in
restatement research: The case of restatements and CEO/CFO turnover. The Accounting Review 83
(6): 1487–1519.
Henry, E. 2008. Are investors influenced by how earnings press releases are written? Journal of Business
Communication 45: 363–407.
Hirshleifer, D., K. Hou, S. Teoh, and Y. Zhang. 2004. Do investors overvalue firms with bloated balance
sheets? Journal of Accounting and Economics 38: 297–331.
Hirshleifer, D., S. Lim, and S. Teoh. 2011. Limited investor attention and stock market misreactions to
accounting information. Review of Asset Pricing Studies 1: 35–73.
Hribar, P., and D. Collins. 2002. Errors in estimating accruals: Implications for empirical research. Journal
of Accounting Research 40: 105–134.
Hsieh, C., and K. Hui. 2011. Analyst Report Readability in High-Technology Firms. Working paper, The
Hong Kong University of Science and Technology.
Huang, A., A. Zang, and R. Zheng. 2013. Informativeness of Text in Analyst Reports: A Naı̈ve Bayes
Machine Learning Approach. Working paper, The Hong Kong University of Science and
Technology.
Kothari, S., X. Li, and J. Short. 2009. The effect of disclosures by management, analysts, and financial press
on the equity cost of capital: A study using content analysis. The Accounting Review 84: 1639–1670.
Kravet, T., and V. Muslu. 2013. Textual risk disclosure and investors’ risk perceptions. Review of
Accounting Studies 18: 1088–1122.
Larcker, D., and A. Zakolyukina. 2012. Detecting deceptive discussions in conference calls. Journal of
Accounting Research 50: 495–540.
Lee, C., L. Li, and H. Yue. 2006. Performance, growth and earnings management. Review of Accounting
Studies 11: 305–334.
Lehavy, R., F. Li, and K. Merkley. 2011. The effect of annual report readability on analyst following and
the properties of their earnings forecasts. The Accounting Review 86: 1087–1115.

The Accounting Review

May 2014
Tone Management 1113

Li, F. 2008. Annual report readability, current earnings, and earnings persistence. Journal of Accounting
and Economics 45: 221–247.
Li, F. 2010. The information content of forward-looking statements in corporate filings—A naı̈ve Bayesian
machine learning approach. Journal of Accounting Research 48: 1049–1102.
Li, F. 2011. Textual analysis of corporate disclosures: A survey of the literature. Journal of Accounting
Literature 29: 143–165.
Li, F., R. Lundholm, and M. Minnis. 2013. A measure of competition based on 10-K filings. Journal of
Accounting Research 51: 399–436.
Loughran, T., and B. McDonald. 2011. When is a liability not a liability? Textual analysis, dictionaries, and
10-Ks. Journal of Finance 66: 35–65.
McAnally, M., A. Srivastava, and C. Weaver. 2008. Executive stock options, missed earnings targets and
earnings management. The Accounting Review 83: 185–216.
Peterson, M. 2009. Estimating standard errors in finance panel data sets: Comparing approaches. Review of
Financial Studies 22: 435–480.
Rapp, C. 2010 Aristotle’s rhetoric. Available at: [Link]
rhetoric
Schrand, C., and B. Walther. 2000. Strategic benchmarks in earnings announcements: The selective
disclosure of prior-period earnings components. The Accounting Review 75: 151–177.
Tama-Sweet, I. 2010. Do Managers Alter the Tone of Their Earnings Announcements Around Stock Option
Grants and Exercises? Working paper, California State University, Fullerton.
Teoh, S., I. Welch, and T. Wong. 1998. Earnings management and the long-run underperformance of
seasoned equity offerings. Journal of Financial Economics 50: 63–100.
Tetlock, P. 2007. Giving content to investor sentiment: The role of media in the stock market. Journal of
Finance 62: 1139–1167.
Tetlock, P., M. Saar-Tsechansky, and S. Macskassy. 2008. More than words: Quantifying language to
measure firms’ fundamentals. Journal of Finance 63: 1437–1467.
Xie, H. 2001. The mispricing of abnormal accruals. The Accounting Review 76: 357–373.

The Accounting Review

May 2014
Copyright of Accounting Review is the property of American Accounting Association and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.