Debiasing the Measurement of Conditional Conservatism

• Published date: 01 September 2021
• Author: MARC BADIA,MIGUEL DURO,FERNANDO PENALVA,STEPHEN G. RYAN
• Date: 01 September 2021
• DOI: http://doi.org/10.1111/1475-679X.12366

DOI: 10.1111/1475-679X.12366

Journal of Accounting Research

Vol. 59 No. 4 September 2021

Printed in U.S.A.

Debiasing the Measurement of

Conditional Conservatism

MARC BADIA,∗MIGUEL DURO,∗FERNANDO PENALVA,∗

AND STEPHEN G. RYAN†

Received 18 February 2020; accepted 2 April 2021

ABSTRACT

Basu’s [“The Conservatism Principle and the Asymmetric Timeliness of Earn-

ings.” Journal of Accounting and Economics 24 (1997): 3–37] measurement of

conditional conservatism as the asymmetric timeliness of earnings underlies

hundreds of studies. However, many subsequent studies cast doubt on the

extent to which Basu’s measure captures conditional conservatism versus sta-

tistical biases or alternative constructs (collectively, “biases”), thereby ques-

tioning the validity of the inferences that empirical researchers draw from

analyses using the measure. We modify Basu’s measure in four simple ways

to remove these biases. Our key modification is the inclusion of interactive

∗IESE Business School, University of Navarra; †Leonard N. Stern School of Business, New

York University

Accepted by Regina Wittenberg Moerman. Marc Badia, Miguel Duro, and Fernando

Penalva (corresponding author) acknowledge financial assistance from research projects

ECO2016-77579-C3-1-P and PID2019-111143GB-C31 funded by the Spanish Ministry of Eco-

nomics, Industry and Competitiveness, and the Ministry of Science and Innovation, respec-

tively. The authors thank participants of the accounting seminars at Boston College, IESE

Business School, INSEAD Business School, London Business School, Temple University,

University of Texas at Dallas, Universidad Carlos III de Madrid, and WHU Otto Beisheim

School of Management, as well as FARS 2020 participants, an anonymous FARS confer-

ence reviewer, an anonymous reviewer, Ashiq Ali, Sudipta Basu, Darren Bernardt, Matthias

Breuer, Dmitri Byzalov, John Core, Aytekin Ertan, Rebecca Files, Umit Garun, Juan Manuel

Garcia Lara, Alastair Lawrence, Ningzhong Li, Maria Loumioti, Stanimir Markov, Sam Me-

lessa (discussant), Max Mueller, Gaizka Ormazabal, Beatriz Garcia Osma, Eddie Riedl, Sug-

ata Roychowdhury, Gil Sadka, Catherine Schrand, Ahmed Tahoun, and Florin Vasvari for

helpful comments or discussions. An online appendix to this paper can be downloaded at

http://research.chicagobooth.edu/arc/journal-of-accounting-research/online-supplements.

1221

© 2021 The Chookaszian Accounting Research Center at the University of Chicago Booth School of

Business

1222 m. badia, m. duro, f. penalva, and s. g. ryan

controls for return variance, a volatility proxy that captures Patatoukas and

Thomas’ [“More Evidence of Bias in Differential Timeliness Estimates of Con-

ditional Conservatism.” The Accounting Review 86 (2011): 1765–1794] return

variance effect and various sources of economic optionality and adjustment

costs. This inclusion captures volatility-related effects on both the level of earn-

ings and the sensitivity of earnings to returns, and it allows the magnitudes of

these effects to vary with the sign of returns. We conduct validation analyses

using placebo-dependent variables, synthetic returns, and nonconditionally

conservative earnings components that show our modified Basu measure is

largely free of known biases. We further show that our measure is associated

with contracting and other economic variables as predicted by theory. Our

findings suggest that researchers can rely on our modified Basu measure to

identify the determinants and effects of conditional conservatism.

JEL codes: C23, D21, G32, M4

Keywords: conditional conservatism; asymmetric timeliness

1. Introduction

Basu [1997] (hereafter “Basu”) is the first study to empirically implement

the construct, now referred to as conditional conservatism, that account-

ing rules require stronger verification of economic gains than of economic

losses to record these items in earnings, so that earnings reflect bad news

in a timelier fashion than good news.1Using equity returns as the proxy for

news based on theory and empirical evidence that returns capture changes

in market expectations of firm value, Basu measures conditional conser-

vatism using a piece-wise linear regression of earnings on favorable returns

(good news) and unfavorable returns (bad news). Basu refers to the in-

cremental coefficient on unfavorable returns relative to the coefficient on

favorable returns as asymmetric timeliness (“AT”). The accounting central-

ity, straightforward intuition, and empirical simplicity of Basu’s AT measure

generated a new field in accounting research that now comprises hundreds

of papers. The results of these studies generally support the view that ac-

counting is conditionally conservative and that conditional conservatism

varies across firms and time consistent with theory (Ball, Kothari, and Niko-

laev [2013b]).

Starting with Dietrich, Muller, and Riedl [2007], however, numerous

studies provide evidence that Basu’s AT measure is subject to statistical bi-

ases or captures economic constructs other than conditional conservatism

(collectively, “biases”). The authors of many of these studies cast doubts

1This conceptualization of conditional conservatism excludes timelier recognition in earn-

ings of bad news than of good news that is attributable to economic fundamentals or to inter-

actions between fundamentals and the application of accounting rules; see Ball et al. [2013b]

for a more expansive conceptualization and our discussion of curtailments in footnote 3. The

excluded forms of AT may also be of interest to accounting researchers; for example, lenders

may factor any source of AT into their determination of debt covenants (Schrand [2014]).

debiasing the measurement of conditional conservatism 1223

on Basu’s AT measure and the validity of the inferences drawn regarding

conditional conservatism using the measure, and they recommend that re-

searchers use the measure with caution or abandon it altogether.2Despite

this skepticism, many of the identified biases are directly related to a sin-

gle and readily controllable source: the volatility of returns or profitability

(hereafter, return variance). Return variance differs across firms based on

their scale, lifecycle stage, and other factors, and thus captures key sources

of sample heterogeneity. Return variance is necessary for economic options

and adjustment costs to have consequences for managerial decision mak-

ing and valuation. Prior research shows that sample heterogeneity and eco-

nomic options and adjustment costs can yield concavity (and in certain

cases convexity) in the earnings-return relation that mimics (offsets) the

effect of conditional conservatism.

Specifically, Patatoukas and Thomas [2011] provide evidence that scale-

related differences across firms in the frequency and magnitude of losses

and the variance of returns yield concavity in the estimated earnings-return

relation for the pooled sample. At least four prior studies provide evidence

of biases attributable to economic options and adjustment costs. Beaver and

Ryan [2009] show that the muting of negative equity returns by the limited

liability of equity along with the pervasive use of amortized cost accounting

for risky debt generates a concave earnings-return relation. Banker et al.

[2016] show that cost stickiness in bad times generates a convex earnings-

return relation, whereas Lawrence, Sloan, and Sun [2018] show that curtail-

ments (e.g., asset retirements and business sales and restructurings) in bad

times generate a concave earnings-return relation.3Breuer and Windisch

[2019] show that firms’ exploitation (mitigation) of positive (negative)

profitability shocks generates a concave earnings-return relation.4In addi-

tion, their modeling of short-run adjustment frictions allows endogenously

for both cost stickiness as examined by Banker et al. [2016] and curtailment

as examined by Lawrence, Sloan, and Sun [2018]. The earnings-returns

asymmetries that result from these economic options and adjustment costs

2Some of these authors recommend measuring conditional conservatism using a similar

piecewise linear regression but with a substantially less comprehensive measure of news than

returns, such as the innovation in cash flow from operations, and with (conditionally con-

servative) accruals rather than earnings as the dependent variable (e.g., Dutta, Patatoukas,

and Wang [2020]). Others recommend using considerably less direct proxies for conditional

conservatism, such as the skewness of earnings or accruals (e.g., Patatoukas and Thomas

[2011,2016], Dutta and Patatoukas [2017]).

3We note that it is difficult or impossible to distinguish curtailments from conditional con-

servatism, for two related reasons. First, curtailments often span multiple accounting periods.

Second, generally accepted accounting principles (ASC 410 and 420) generally require that

the remaining costs of curtailments be recognized at fair value. Hence, initial curtailment

write-downs essentially are (required) manifestations of conditional conservatism.

4Although not explicitly framing their analysis in terms of economic options or adjustment

costs, similar to the studies cited in this paragraph, Dutta and Patatoukas [2017] show that

right skewness in equity returns owing to the right skewness of cash flow news yields concavity

in the earnings-returns relation.