Optimal versus suboptimal choices of accounting expertise on audit committees and earnings quality

Abstract

By employing a Heckman two-stage selection model, we identify whether employing a financial expert with or without accounting expertise on the audit committee is optimal and how earnings quality varies across these optimal and suboptimal choices. Using four earnings quality measures (informativeness, timely loss recognition, earnings persistence, and accruals quality), we find no differences in earnings quality between firms optimally choosing an expert with or without accounting expertise, consistent with Demsetz and Lehn (J Polit Econ 93:1155–1177, 1985) and others who argue that when firms optimize their choice (i.e., accounting expertise), there should be no difference across the characteristic (i.e., earnings quality) being examined. We do find, however, earnings quality is significantly higher for firms that optimally choose an accounting expert relative to firms that choose (with/without accounting expertise) suboptimally. Finally, firms suboptimally choosing an accounting expert exhibit no improvement, or even lower earnings quality, than firms that optimally choose no accounting expert. Our results provide important evidence of the impact accounting expertise has on earnings quality when considering the firm’s choice.

Appendix

Variable definitions

Determinants of an accounting expert on the audit committee
AEXP i,t	An indicator variable set to one if the firm selects an accounting expert on the audit committee, otherwise zero;
SAEXP i,t	An indicator variable set to one if the firm selects a single accounting expert on the audit committee, otherwise zero;
MAEXP i,t	An indicator variable set to one if the firm selects multiple accounting experts on the audit committee, otherwise zero;
BDSZ i,t	Natural logarithm of the number of members on the board of directors;
ACSZ i,t	Natural logarithm of the number of directors on the audit committee;
BDIND i,t	Percentage of directors on the board who are independent, where a director is considered independent if he or she is not a former employee of the firm, is not a relative of the company’s executives, and does not have business relationship with the company;
OTHBD i,t	Average other directorships held by outside directors;
INSTOWN i,t	Percentage of the firm’s aggregate common stock owned by institutional investors, averaged over the year;
DIROWN i,t	Percentage of the firm’s aggregate common stock owned by outside directors;
DUAL i,t	An indicator variable set to one if the CEO serves as the chairman of the board, otherwise zero;
CEOOWN i,t	Percentage of the firm’s common stock held by the CEO;
SIZE i,t	Natural logarithm of the firm’s total assets for year t;
MTB i,t	The firm’s market-to-book ratio, measured at the beginning of year t;
SEG i,t	Natural logarithm of the firm’s number of business segments;
FRGN i,t	An indicator variable set to one if the firm has foreign operations, otherwise zero;
LEV i,t	The firm’s short-term and long-term debt divided by total assets, measured at the beginning of year t;
LOSS i,t−1	An indicator variable set to one if the firm reported a loss for year t − 1, otherwise zero;
LITIIND i,t	An indicator variable set to one if the firm operates in biotechnology (SIC 2833-2836 and 8731-8734), computers (SIC 3570-3577 and 7370-7374), electronics (SIC 3600-3674), or retail (SIC 5200-5961) industries, otherwise zero;
MKRET i,t−1	Market-adjusted 12-month stock return for year t − 1;
STDRET i,t−1	Standard deviation of the firm’s 12-month return for year t − 1;
SKEW i,t−1	Skewness of the firm’s 12-month return for year t − 1;
SALEGRW i,t−1	Year t − 1 sales less year t − 2 sales scaled by the beginning of year t − 1 total assets; and
TURNOVER i,t−1	Trading volume accumulated over the 12-month period for year t − 1, scaled by beginning of year t − 1 shares outstanding.
Inverse mills ratio, measures of earnings quality, and controls
$W_{i,t}^{k}$	Inverse Mills ratio (kε{1, 0}), measured as φ(β′Z)/Ф(β′Z)) for firms choosing an accounting expert to serve on the audit committee and −φ(β′Z)/(1 − Ф(β′Z)) for firms not choosing an accounting expert to serve on the audit committee, where: φ(·) represents the probability density function (pdf) of a standard normal distribution; Ф(·) represents the cumulative density function (cdf) of a standard normal distribution; β’Z is the prediction from the probit model in Eq. (1);
ARET i,t	Buy-and-hold return compounded from the day after year t − 1’s earnings announcement date through year t’s earnings announcement date, less expected returns estimated by Fama and French’s (1993) three-factor model;
UE i,t	Unexpected earnings reported for year t, measured by the actual reported earnings less the consensus analyst forecasts of annual earnings measured by the first forecast from I/B/E/S after year t − 1’s earnings announcement date;
PREC i,t	One divided by the number of analyst forecasters, measured in the month after year t − 1’s earnings announcement;
LTG i,t	Consensus analyst forecasts of long-term growth from I/B/E/S, measured in the month after year t − 1’s earnings announcement;
DECPRC i,t	An indicator variable set to one if the change in price from year t − 1 to year t was negative, otherwise zero;
EARN i,t	Income before extraordinary items (Compustat item IB) deflated by the market value of equity at the beginning of year t;
RET i,t	Buy-and-hold daily return compounded over year t;
DR i,t	An indicator variable set to one if RET i,t  < 0, otherwise zero;
IB i,t	Income before extraordinary items deflated by average total assets for year t;
TCA i,t	Income before extraordinary items minus cash flow from operations (Compustat item OANCF), plus depreciation and amortization expenses (Compustat item DP), deflated by average total assets for year t; and
CFO i,t	Cash flow from operations deflated by average total assets for year t