Determinants and consequences of timely asset impairments during the financial crisis

• Author: Joshua L. Gunn,Inder K. Khurana,Sarah E. Stein
• Date: 01 January 2018
• DOI: http://doi.org/10.1111/jbfa.12287
• Published date: 01 January 2018

DOI: 10.1111/jbfa.12287

Determinants and consequences of timely asset

impairments during the financial crisis

Joshua L. Gunn1Inder K. Khurana2Sarah E. Stein3

1KatzGraduate School of Business, University of

Pittsburgh,Pittsburgh, Pennsylvania, USA

2RobertJ. Trulaske Sr.,College of Business,

Universityof Missouri, Columbia, Missouri, USA

3PamplinCollege of Business, Virginia Tech,

Blacksburg,Virginia, USA

Correspondence

SarahE. Stein, Pamplin College of Business,

VirginiaTech, 3007 PamplinHall, Blacksburg,

VA24061, USA.

Email:sestein@vt.edu

Abstract

U.S. firms recorded an unprecedented number of asset impairments

during the recent financial crisis. We investigate the timing of these

lossesin the context of two competing views on how firms use discre-

tion over asset impairments. The first view posits that firms record

impairments to convey private information as part of their commit-

ment to a conditionally conservative reporting strategy. The sec-

ond view argues that firms use their discretion to report oppor-

tunistically by delaying the recording of bad news. Consistent with

the first view, we find that firms recorded timelier asset impair-

ments during the financial crisis if theyreported more conservatively

in the five years preceding the crisis. Further tests show this rela-

tion is greater for firms with strong corporate governance, industry-

specialist auditors, and high leverage, indicating the importance of

monitoring mechanisms in determining how firms handle the discre-

tion involved in impairment decisions. We also test for the conse-

quences of timely asset impairments during the financial crisis and

find that firms reporting conservatively both before and during the

crisis were able to acquire more debt financing, and their publicly

traded bonds suffered smaller increases in illiquidity. Collectively,

our study highlights the role of asset impairments in firms’ account-

ing choices over time.

KEYWORDS

asset impairments, bond illiquidity, conservatism, debt contracting,

monitoring

1INTRODUCTION

The recent financial crisis is characterized as the most serious economic downturn since the Great Depression. The

crisis originated in the financial sector and spread quickly throughout the U.S. economy.From a financial accounting

perspective, the crisis created a large amount of uncertainty about the value of firms’ assets and required managers to

re-examine expectationsabout future cash flows. As a result, this time period experienced an unprecedented number

of asset impairments. In the fourth quarter of 2008 alone, non-financial firms recorded almost $311 billion in asset

impairments, which is greater than the total recorded over the four years leading up to the crisis. All told, U.S. firms

J Bus Fin Acc. 2018;45:3–39. wileyonlinelibrary.com/journal/jbfa c

2017 John Wiley & Sons Ltd 3

4GUNN ET AL.

recorded over $742 billion in asset impairments between 2007 and 2010.1We examine the determinants and con-

sequences of timely asset impairments during the financial crisis. Specifically,we assess whether management's deci-

sions about the timing of impairment losses during the crisis are consistent with reporting behaviors in prior periods

and whether firms benefit from timely asset impairments.

Prior research argues that asset impairments can be used by managers to implement two distinct reporting strate-

gies. On the one hand, the conservatism literature cites asset impairments as a primary example of asymmetric loss

recognition and a keymechanism ensuring conditional conservatism (André, Filip, & Paugam, 2015; Ball & Shivakumar,

2005; Basu, 1997; Nikolaev, 2010). The argument here is that some firms commit to accounting conservatism as an

efficient contractingmechanism that reduces agency costs to better align incentives of the contracting parties (Ahmed,

Billings, Morton, & Harris, 2002; Armstrong, Guay,& Weber, 2010). As a result, management could use the discretion

inherent in accounting standards to implement a conditionally conservative reporting strategy.Under this view, firms

with conservative reporting in the pre-crisis period will record timelier impairments during the crisis period.

On the other hand, impairments can be considered “movable” expenses used to achieve earnings targets (Watts,

2003a). This view predicts managers use their discretion over asset impairments to report opportunistically.Oppor-

tunistic impairments can manifest in two ways: (1) delaying the recording of impairment losses in the hope that con-

ditions will improve in time to avoid taking the write-off,and (2) recording a larger impairment than is economically

justified in order to increase reported income in future periods (i.e., a “big bath”). These two choices are not mutually

exclusiveand can be used in conjunction with each other to report opportunistically, as big bath reporting could occur

before or after the period in which the firm experienceseconomic losses. Since the extent and severity of the crisis was

unknown, managers may prefer to delay the recording of bad news in the hope that conditions will improve; however,

once the impairment is no longer avoidable,the manager may write off more than is necessary to avoid future negative

earnings surprises. Empirical research provides evidence consistent with recent increases in both delayedreporting of

impairment losses and big bath behavior (Beatty & Weber,2006; Filip, Jeanjean, & Paugam, 2015; Ramanna & Watts,

2012; Riedl, 2004). Further,“overly” conservativer eportingin the boom phase of the business cycle provides an oppor-

tunity to smooth earnings and delay the recording of bad news during the crisis period (Kothari, Ramanna, & Skinner,

2010). If firms primarily used asset impairments to report opportunistically during the crisis period, we would expect

to observe a negativeor an insignificant association between the timing of crisis-period impairments and conservatism

in the pre-crisis period.2

Toperform our tests, we develop a new measure of the timing of asset impairments. Over a sufficient time frame,

all of a firm's long-term assets are eventually expensed (ignoring the portion disposed of for cash). Arguably then, the

more interesting question is not whether an asset was written off, but when it was written off.As the economic reces-

sion unfolded, most firms were forced to revise their expectationsof future cash flows downwards at roughly the same

time and for similar macroeconomic reasons. Thus, the economic downturn serves as a common source of bad news for

all firms.3This unique setting allows us to create a measure to capture the timing of asset impairments (also referred to

as “impairment speed”). The intuition for our measure is that earlier recording of impairments during the crisis period

is consistent with management imposing a lower verification standard to record losses as information about the full

scope of the recession and extent of the recoverywas not available at the beginning of the crisis period.4By examining

firms that record impairments during the crisis period, we can condition the analysis on firms we know, ex post, had bad

1Wecalculated these values by summing the quarterly values for pre-tax impairments of goodwill and other indefinite-lived intangibles (GDWLIPQ) andpre-

taximpairments of finite-lived intangibles and other long-lived assets (WDPQ) for all non-financial firms in Compustat over the relevant periods. These values

arelarger than those reported in Figure 1 ($508 billion) due to data restrictions for our sample.

2Ourprimary sample only includes firms with impairment losses recorded during the crisis period. This design removes the risk that firms could be so conser-

vativethat they anticipated and recorded all crisis-period losses during the pre-crisis period.

3Tohighlight the point that a downturn can be “bad news” in and of itself,a recent Wall Street Journal article states the following: “There is a risk, of course, that

arun-up in interest rates or a drop in the stock market could spark an increase in goodwill write-downs” (Chasan & Murphy, 2013).

4We make no claim that earlier recording of impairments is “better” or more “optimal,”only that early recording of impairments is consistent with a lower

verification threshold for recording losses, and, thus, consistent with conditionally conservative accounting (Basu, 1997). In this way,our measure is well

suitedto address whether or not impairments are recorded earlier by firms that choose conditionally conservative reporting strategies.

GUNN ET AL.5

news to report (via an impairment charge). Thus, we can focus on the timing of the impairment charge, as opposed to

whether an impairment charge was required or recorded. In Section 3, we provide additional discussion of our impair-

ment speed measure.

Our primary findings reveal a positive and significant association between asset impairment speed and several

proxies for pre-crisis conservatism based on prior literature. These results hold after controlling for firm character-

istics known to explain the variation in accounting conservatism across firms. Importantly,we find that our results are

robust to inclusion of controls for both the timing and level of firm performance, as firm performance could explain

both pre-crisis conservatism and the timing of crisis-period impairments. These controls are particularly important in

our setting to differentiate firms’ reporting decisions from firms’ poor operating performance. In additional tests, we

use instances when the firm's book-to-market ratio is greater than one as a proxy for non-discretionary impairments

as in Lawrence, Sloan, and Sun (2013). We find the positive association between pre-crisis conservatism and crisis-

period impairments is driven by discretionary asset write-downs, suggesting that firms use their discretion over asset

impairments to implement a conservative financial reporting strategy.

Toaddress certain limitations of our impairment speed measure and research design, we also use an empirical model

based on Basu's (1997) differential timeliness measure for a broader sample of impairment and non-impairment firms.

Similar to Banker,Basu, and Byzalov (2017), we adjust the Basu model to incorporate impairment losses as the depen-

dent variable. This test complements our primary results by addressing two important assumptions in our impairment

speed measure: (1) it allows for analysis of both impairment and non-impairment firms, and (2) it incorporates stock

returns to measure timeliness rather than interpreting earlier impairments as more conservative impairments. Our

results indicate that the association between the magnitude of impairments and negative stock returns during the cri-

sis period is significantly larger when firms report conservatively in the pre-crisis period.

These findings suggest that firms with a revealed preference for conservatism in the pre-crisis period recorded

crisis-period impairments earlier, presumably to communicate bad news in a timely manner as part of their commit-

ment to a conditionally conservative reporting strategy. An alternativeexplanation is that crisis-period impairments

were “big baths,” intended to improve earnings in future periods ratherthan convey private information. Ex ante,there

are two reasons to expect that our main findings are not driven by “big bath” behavior. First, Riedl (2004) finds an

increase in both delayed asset impairments and “big bath” behavior after a change in impairment accounting stan-

dards, suggesting that the two behaviors go hand in hand. Second, pre-crisis conservative reporting would create slack

in firms’ balance sheets (Barton & Simko, 2002), lessening the need to take a “big bath” at the onset of the financial

crisis. Nonetheless, we include total impairment size in our model as a proxy for the likelihood that an impairment is

a “big bath” as the true level of impairment is unobservable. We find impairment size exhibits anegative relation with

impairment speed, especially in the models focusing on discretionary impairments. This result suggests that the posi-

tive relation between pre-crisis conservatism and impairment speed is not the outcome of “big bath” behavior as firms

with relatively large impairments were not recording those losses earlier. Our differential timeliness model adds fur-

ther evidence on this point, as the impairment magnitudes of firms reporting conservatively in the pre-crisis period

correlate more strongly with contemporaneous negative stock returns.

Existing research suggests that firms choose conservative accounting practices based on the mix of incentives and

agency problems unique to the individual firm. Therefore, we also examine whether the relation of pre-crisis conser-

vatism and crisis-period impairments varies cross-sectionally based on the quality of internal and external monitoring

mechanisms. This empirical exerciseis motivated by studies that identify corporate governance, auditors, and lenders

as key monitors associated with the level of conditional conservatism in earnings (Nikolaev, 2010; Roychowdhury,

2010; Vyas,2011; Zhang, 2008). Using severalfeatures of corporate governance, industry-specialist auditors, and firm

leverage, we find that the positive relation between asset impairment speed and pre-crisis conservative reporting is

increasing with the strength of the firm's monitoring mechanisms. This evidence adds credence to the agency theory

explanations for the commitment to conservatism suggested by our earlier tests and indicates that impairments are

more likely to be used to report opportunistically in settings with weak monitoring mechanisms.

Last, we assess the consequences of timely asset impairments during the financial crisis. Existing theory posits that

conditional conservatism benefits the firm by making debt contractingeasier and less costly. During the financial crisis,