In this paper I examine the relationship between systematic risk and accounting conservatism. Conservatism is the asymmetrical verification requirements for the recognition of economic losses and gains (Basu, 1997; Watts, 2003a). While prior literature has suggested various factors that create demand for and explain the cross-sectional differences in accounting conservatism, the effect of systematic risk on conservatism has not been studied. Systematic risk, the non-diversifiable risk that is attributable to market factors, is one of the fundamental concepts in asset pricing. The risk disclosure literature in accounting has related accounting numbers to systematic risk (Ryan, 1997), which assumes the role of accounting as providing information for risk assessment. Few papers have examined whether systematic risk affects managers’ financial reporting incentives and consequently, earnings properties such as timeliness or conservatism. This paper may help fill this gap, providing an examination of the effect of systematic risk on accounting conservatism.
I argue that systematic risk provides managers with 1) more incentives to delay the recognition of bad news in the hope of future good news and 2) less constraint from investors and auditors to report conservatively. Managers’ tendency to withhold bad news can stem from a standard agency problem where managers possess proprietary information not available to outsiders and their preferences are not aligned with those of shareholders. Managers have incentives to bias this information favorably to extract excess compensation (Jensen and Meckling, 1976). By delaying bad news recognition to the future when things improve, managers can realize currently exercisable options that might otherwise be worthless (Imhoff, 2003; Benmelech et al., 2010). Holding other things constant, firms with higher systematic risk have a wider range of possible future performance, giving managers higher incentives to delay the recognition of bad news in the hope of future good news. On the contrary, the future prospects for low-risk firms vary less and current bad performance is more likely to be firm specific and permanent in nature, leaving managers in low risk firms with less incentive to delay the recognition of bad news.
Systematic risk may also lessen the constraints managers face to report conservatively. Bad news coming from co-movements with the market may not be interpreted as problems specific to the business of the firm while firm-specific bad news may be highly scrutinized due to litigation and other concerns. Since observed conservatism is the equilibrium resulting from the managers’ incentives to report conservatively subject to the constraints posed by auditors and outside investors, I predict that systematic risk is negatively associated with conservatism.
I test my predictions using a sample of 141,550 observations over the period 1964-2008. Conservatism is measured using the asymmetric timeliness of earnings in the Basu (1997) model that regresses earnings on returns, a negative return dummy, and the interaction between returns and the dummy. Systematic risk is measured using the CAPM beta (BETA). To control for the effects of omitted correlated variables that may potentially confound the relationship between BETA and conservatism, I employ both a single-stage and two-stage regression approach. In the single-stage regression, I include BETA and its interactions with all three explanatory variables in the Basu (1997) model, together with the control variables and their interactions with the explanatory variables. In the two-stage regression, I first regress BETA on the set of control variables, and use the orthogonalized BETA in the Basu regression.
