The primary objective of this study is to examine the effect of the interaction between two earnings smoothing devices, accrual management and hedging with derivatives, on firm value. Lambert (1984) suggests that devices for smoothing can generally be classified as real actions (e.g., hedging by derivatives) and artificial techniques (e.g., discretionary accruals). The former reduces earnings volatility by directly affecting cash flows, whereas the latter directly affects volatility of publicly reported income. While there is evidence that managers use hedging and discretionary accruals as substitutes to control earnings volatility (e.g., Barton, 2001), no study thus far has addressed the more fundamental question of how this trade-off affects firm value. This question is important because it provides insight into a manager’s incentives for trading off hedging and discretionary accruals at the margin. Moreover, it sheds light on the possible effect on firm value of Statement of Financial Accounting Standards (SFAS) No. 133, Accounting for Derivative Instruments and Hedging Activities.
To address this research question, we examine the interaction between accrual management and hedging with derivative instruments for income smoothing in a sample of 477 non-financial corporations with 1,117 firm-years in the period 1994 through 1996. We measure the interaction as the relative use of discretionary accruals in the earnings-smoothing program that uses both derivatives and discretionary accounting techniques to dampen earnings volatility. We refer to this measure as proportional accrual management (PAM). We estimate the cross-sectional modified Jones’ (1991) model (see Dechow et al., 1995) to obtain our measure of discretionary accruals. Consistent with previous studies (e.g., Guay and Kothari, 2003), we use a gross notional principal of derivatives holdings to capture a firm’s derivatives positions. Lagged total assets scale both discretionary accruals and derivatives positions.
We control for the endogeneity of management’s decision to choose between discretionary accruals and hedging with derivatives in evaluating the effect of the choice on firm value. Previous evidence (e.g., Haushalter, 2000; Klein, 2002) suggests that firm-specific characteristics affect the intensity of accrual management and derivatives use. The failure to control for firm characteristics may wrongly attribute the valuation effect to the choice between real and artificial smoothing tools instead of the underlying characteristics. Therefore, we use a two-stage least squares estimation to model the endogeneity of the choice in analyzing its effect on firm value. Particularly, we model PAM as a function of corporate governance mechanisms and firm fundamentals in the first stage and then use the fitted value of PAM with other characteristics to explain firm value in the second stage. Similar to Billett et al. (1995), Nohel and Tarhan (1998), and many others, we capture firm value by Tobin’s Q.
We document a negative relation between PAM and Tobin’s Q. This relation suggests that choosing to have a higher proportion of accrual management relative to real smoothing tools leads to lower market valuation. When we replace PAM with separate variables for discretionary accruals and derivatives, again using a two-stage least squares procedure, we find that firm value is negatively related to discretionary accruals and is positively related to the level of derivatives use. Overall, our findings imply that SFAS No. 133 could reduce firm value if, as Barton (2001) argues, the imposition of SFAS No. 133 reduces hedging and increases accrual management. Furthermore, the first-stage results suggest that the intensity of PAM declines with the percentage of outside directors, audit committee meeting frequency, and the percentage of outsiders sitting on the audit committee. These results suggest that boards and audit committees structured to be more independent of the management, and audit committees with more frequent meetings are more effective in monitoring the choice of the earnings smoothing devices. Taken together, these findings lend support to the Blue Ribbon Committee’s (BRC) recommendation and Sarbanes Oxley Act of 2002 (SOX 2002) on the independence of audit committees.
Since our sample firms represent 45 different industries according to the two-digit SIC classification, we also investigate whether the failure to control for industry effects masks the relation between firm value and PAM. We perform two analyses to control for industry effects. First, we construct industry-adjusted Qs by computing the log difference between raw Qs and the median Q for each industry to which the firm belongs and then re-estimate the two-stage least square regressions. Second, we use industry indicator variables based on the two-digit SIC classification in the system of equations. The results from both analyses are similar to those obtained without controlling for industry effects. We find that firms that undertake higher levels of proportional accrual management in their income-smoothing programs have lower firm values. Additionally, independent boards, and audit committees that are more independent or holding more meetings induce managers to use relative lower levels of discretionary accruals versus derivative instruments as means of reducing earnings fluctuations.
