A re‐examination of accruals quality following restatements

• Author: Jan Bartholdy,Marie Herly,Frank Thinggaard
• Date: 01 July 2020
• DOI: http://doi.org/10.1111/jbfa.12445
• Published date: 01 July 2020

DOI: 10.1111/jbfa.12445

A re-examination of accruals quality following

restatements

Marie Herly Jan Bartholdy Frank Thinggaard

Department of Economics and Business

Economics, Aarhus University

Correspondence

MarieHerly, Aarhus University,Department of

Economicsand Business Economics, Fuglesangs

Alle4, DK-8210 Aarhus V.

Email:mherly@econ.au.dk

Fundinginformation

FSR’sStudie- og Understøttelsesfond

Abstract

Empirical research from the first years following the Sarbanes-Oxley

Act (SOX) in the US suggests that firms improve accruals qual-

ity following restatements, but the materiality of restatements has

declined since then. This decline may affect firms’ responses to

restatements, and hence we re-examine whether restatements are

associated with subsequent improvements in accruals quality in a

more recent sample. We compare the changes in accruals quality

of firms restating between 2000 and 2014 with that of a control

group. We do not find that firms improve accruals quality more than

the control group following a restatement, even when we isolate

the types of restatements considered to be most material. However,

we do find that restatements followed by the most negative stock

market reactions are associated with a relative increase in accru-

als quality,indicating that only restatements deemed very severe by

investors lead to subsequent improvements in accruals quality.Our

results suggest that firms’ responses to restatements have changed

concurrently with the trend of fewer and less material restatements

in recent years.

KEYWORDS

accrual accounting, capital markets, consequences of earnings

misstatements, difference-in-differences, earnings quality, enforce-

ment, materiality,propensity score matching, restatement of corpo-

rate earnings, Sarbanes-Oxley Act

JEL CLASSIFICATION

M41, G14, M48

1INTRODUCTION

Financial restatements are serious events, and prior research suggests a number of negative consequences to them,

such as large drops in market prices (Palmrose & Scholz, 2004; Wilson, 2008), an increased cost of capital (Hribar &

Jenkins, 2004; Kravet & Shevlin,2010), worse contract terms (Graham, Li, & Qiu, 2008; Karpoff, Lee, & Martin, 2008),

and managerial turnover (Desai, Hogan, & Wilkins, 2006; Hennes, Leone, & Miller, 2008). Wiedman and Hendricks

882 c

2020 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jbfa J Bus Fin Acc. 2020;47:882–909.

HERLYET AL.883

(2013) find evidence suggesting that firms respond to these negative consequences by improving accruals quality to

signal higher reporting quality.

These papers investigate restatements occurring before 2003, and the nature of restatements has changed sub-

stantially since then. Scholz (2014) and Whalen, Usvyatsky, and Tanona (2018) describe how both the number and

severityof restatements have declined since the mid-2000s. In particular, recent restatements tend to be less material,

involve fewer accounting issues, restate fewer periods, and are less likely to reduce net income. This trend maystem

from improved internal control following SOX, an increase in the number of revision restatements (“little r” restate-

ments), and a number of bulletins from the SEC clarifying the threshold for restating.1As a result of the changed

nature of restatements, investors’ and firms’ responses to restatements have changed in recent years (Burks, 2011;

Chen, Elder,& Hung, 2014; Hirschey, Smith, & Wilson, 2015). This change calls for a re-examination of accruals quality

following restatements using a more recent sample.

We compare the accruals quality of 1,534 firms that have restated once between 2000 and 2014 with a con-

trol group matched based on propensity scores. Our measure of accruals quality follows Dechow and Dichev (2002),

estimated cross-sectionally in accordance with Jones, Krishnan, and Melendrez (2008) and Wiedman and Hendricks

(2013). The results indicate that single-restatement firms do not improveaccruals quality more than the control group.

We therefore cannot attribute anyimprovement to the restatement and do not find evidence of a disciplining effect of

a restatement. We divide our sample into groups with severe restatements, namely restatements involvingfraud, SEC

investigations, and restatements of earnings. Even for these groups, we do not observe a subsequent improvementin

accruals quality.

Finally, we isolate the restatements with verynegative stock market reactions to the restatement announcement,

an indication that investors deem those restatements material. Our results suggest that the restating firms experienc-

ing the most negative market reactions improve accruals quality significantly more than the control group. Thus, our

evidence implies that restatements per se do not motivate firms to change accruals quality,only restatements deemed

very severe by financial markets induce measurable corrective actions. These results are robust to several alterna-

tive specifications.

Our evidence contributes to the restatement literature by extending our knowledge about the consequences of

restatements (e.g., Kravet and Shevlin(2010); Graham et al. (2008); Wilson (2008)) to include the long-term develop-

ment in earnings quality after a restatement. It also complements Farber (2005), Hennes, Leone, and Miller (2014),

and Chakravarthy,de Haan, and Rajgopal (2014), who provide evidence that firms attempt to restore the credibility

of financial statements after restatements by changing corporate governance and external auditors. In addition, our

evidence that restatements do not necessarily improve accruals quality may explain the large frequency of repeated

restatements described by Files, Sharp, and Thompson (2014).

Our study also contributes to the literature by illustrating how the declining materiality of restatements in recent

years has changed firms’ responses to them. Wiedman and Hendricks (2013) find that firms improve performance-

matched accruals quality following restatements between 1997 and 2003. In contrast, our results indicate that firms

do not improve accruals quality,on average, after restatements, which suggests that firms’ responses to restatements

have changed since then. Our results thus corroborate He, Sarath, and Wans’s(2019) recent evidence. They find that

restatements associated with material weakness disclosures announced between 2003 and 2013 are associated with

lower subsequent accruals quality. Combined with evidence that investors’ responses to restatements are declining

(Burks, 2011; Scholz, 2008), our results suggest that firms’ incentives to improve accruals quality have fallen concur-

rently with investors’ responses to them. Our paper also adds to prior research suggesting that firms are more sensi-

tive to the restatement severity in the post-SOX era: Chen et al. (2014) find that the relation between severemarket

reactions and asymmetric timeliness is more pronounced in the post-SOX era, while Burks (2010) provides evidence

1Nagy(2010) finds fewer restatements after the passing of SOX Section 404 Management Assessment of Internal Controls. The declining number of restate-

mentsmay be offset by an increasing number of non-4.02 item revisions, caused partly by the SAB 108 issued in September 2006 (Scholz, 2014; Tan & Young,

2015).