Do High Ability Managers Smooth Earnings to Signal Private

Preparing to load PDF file. please wait...

0 of 0
100%
Do High Ability Managers Smooth Earnings to Signal Private

Transcript Of Do High Ability Managers Smooth Earnings to Signal Private

Do High Ability Managers Smooth Earnings to Signal Private Information? *
Bok Baik Seoul National University
Sunhwa Choi Lancaster University
David B. Farber The University of Texas at El Paso
March 2014
Abstract: In this study, we investigate whether managerial ability is related to income smoothing and if so, whether smoothing associated with managerial ability is informative about future earnings. Using a sample of firms for the period of 1991-2009, we find that discretionary smoothing due to high ability managers is greater than that due to low ability managers and that this relation is more pronounced for firms with high information asymmetry. More importantly, we show discretionary smoothing is only useful when employed by high ability managers in firms that face high levels of information asymmetry. Overall, our results suggest that high ability managers smooth earnings to signal private information. Our study should be of interest to researchers, practitioners, and others concerned with understanding the determinants and usefulness of smoothing.
*We thank Lee-Seok Hwang, Jae Yong Shin, and seminar participants at Seoul National University and University of Texas at El Paso for their helpful comments and suggestions.

1. Introduction
A primary purpose of financial reporting is to provide information that is useful for
decision making (FASB 2010). SFAC No. 8 (FASB 2010) suggests that because accruals smooth fluctuations in the timing of cash payments and receipts, income smoothing1 can potentially
enhance users’ ability to assess firms’ future performance. Anecdotal evidence suggests that
corporate executives believe that smoothing is useful for decision making. In their survey paper,
Graham et al. (2005) document that the vast majority of financial executives have a strong preference for smooth earnings.2 Given its importance, it is surprising that such little research
exists about the usefulness of smoothing. Our study contributes to the literature by providing
evidence about whether managerial ability is related to smoothing and if so, whether smoothing
due to managerial ability improves the usefulness of future earnings.
We employ a powerful setting for informing the debate about the usefulness of
smoothing because it likely varies cross-sectionally with managerial ability (Schipper and
Vincent 2003; Kirschenheiter and Melumad 2004). Compared to low ability managers, high
ability managers have a superior understanding of the environment in which their firms operate
and are therefore in a better position to anticipate changes in their firms’ economic prospects.
High ability managers have incentives (e.g., equity-based compensation) to use discretion to
signal their private information about their firms’ prospects via smoothing because doing so increases firm value (Barth et al. 1999; Tucker and Zarowin 2006; Louis and White 2007).3 Thus,
compared to low ability managers, high ability managers are more likely to use their discretion to
1 We define income smoothing as the intentional dampening of earnings fluctuations, consistent with Beidleman (1973) and others. For the remainder of the paper, we interchangeably use “income smoothing” and “smoothing”. 2 Nearly all (96.9%) of the financial executives responded that they prefer a smooth earnings path, while 80% of the financial executives indicated that smoother earnings help analysts and investors predict future earnings. 3 Smoothing is a credible signal because managers would be irrational to report higher earnings than what is expected to persist because their firms would likely incur negative capital market consequences and managerial reputation would likely decline. This argument is consistent with that in Ronen and Sadan (1981), who adopt Spence’s (1973) model to predict that only firms with good prospects smooth.
1

signal their private information through smoothing.4 Furthermore, investors are likely to have difficulty processing and pricing information for firms with high information asymmetry (Bushee and Miller 2012). Thus, given a positive relation between managerial ability and income smoothing, the relation will be likely be more pronounced when there is a high level of information asymmetry between managers and investors since managers face pressure to alleviate information asymmetry.
We also assess whether discretionary smoothing due to managerial ability enhances the informativeness of future earnings. To the extent that high ability managers have superior skills in assessing their firms’ future performance, we expect that their discretionary accounting choices to smooth earnings will be more accurate, and thus of higher quality, than those of lower ability managers. We therefore expect that compared to smoothing by low ability managers, smoothing by high ability managers will make earnings more informative, particularly for firms with higher information asymmetry.
We execute our tests using a common factor for firm-level smoothing based on three measures of smoothing widely used in the literature (e.g., Leuz et al. 2003; Tucker and Zarowin 2006; Dou et al. 2013): (i) standard deviation of earnings divided by the standard deviation of cash flows from operations; (ii) correlation between changes in accruals and changes in cash flows from operations; and (iii) correlation between changes in discretionary accruals and changes in pre-managed earnings. We use the model from Lang et al. (2012) to partition our smoothing measure into its discretionary and fundamental components. Our main proxy for
4 However, high ability managers have more to lose in terms of compensation (Falato et al. 2012; Graham et al. 2011) and reputation (Fudenberg and Tirole 1995) if unexpected negative shocks in the future lead to a poor mapping of current period smoothing to future earnings realization. Thus, higher ability managers would likely have less of an incentive to smooth if the costs of doing so outweigh the benefits.
2

managerial ability is MA-Score (Demerjian et al. 2012), which is a manager-specific measure derived from Data Envelope Analysis.
We use a large sample of U.S. firms over the period 1991-2009 and find that discretionary smoothing increases in managerial ability. We also find that this relation is more pronounced for firms with higher information asymmetry, consistent with a signaling story.
To examine whether smoothing is informative about future earnings, we use models from Collins et al. (1994), Lundholm and Myers (2002), and Tucker and Zarowin (2006). We first confirm the finding from Tucker and Zarowin (2006) that current stock returns of firms with smoother earnings reflect more information about future earnings than the stock returns of firms with less smooth earnings. Importantly, when we partition our smoothing measure into its discretionary and fundamental components, we find that only discretionary smoothing is significant for the incorporation of future earnings into current returns. Moreover, we show that these results only obtain for firms with higher levels of information asymmetry.
Overall, these results are consistent with the characterization of income smoothing as a credible signal about firms’ future earnings, supporting the argument that discretionary smoothing due to high ability managers is useful to equity investors.
We perform a battery of robustness tests and find that results are insensitive to alternative measures of managerial ability (i.e., historical industry-adjusted stock returns and return on assets (ROA)), to firm fixed effects models, to using the individual components of our smoothing measure, to including additional control variables, and to the use of fractional ranks for our smoothing measures.
Our study makes several important contributions to the literature. First, we contribute to the literature examining the usefulness of smoothing, and, more broadly, the literature on the use
3

of financial reporting to communicate managers’ private information (Subramanyam 1996; Louis and Robinson 2005; Tucker and Zarowin 2006). In one of the few U.S-based studies on the impact of smoothing on earnings informativeness, Tucker and Zarowin (2006) report that smoothing improves the degree to which the information in future earnings is reflected in current stock returns. While Tucker and Zarowin (2006) take smoothing as given and do not attempt to model cross-sectional differences in the underlying motivation to smooth, we identify an important new source of smoothing - managerial ability - and link it to the market’s incorporation of information in future earnings. By doing so, we identify an important factor that affects the usefulness of smoothing. Our approach is consistent with Dechow et al.’s (2010, p. 391) call for more research that uses a complete path approach, which provides deeper insights than research that only examines either determinants or consequences of smoothing. Our study therefore extends and complements Tucker and Zarowin (2006) by enhancing our understanding of smoothing as a measure of earnings quality.
Our study is also related to the emerging stream of research on the role of managerial ability in the determination and consequences of earnings quality (Demerjian et al. 2012; Demerjian et al. 2013a, b).5 We complement this line of research by showing that discretionary smoothing due to high ability managers improves earnings informativeness. We also extend research on determinants of smoothing (Dascher and Malcom 1970; Barnea et al. 1976; McNichols and Wilson 1988; Chaney et al. 1998; Kanagaretnam et al. 2004) by identifying managerial ability as an important determinant of discretionary smoothing. Findings from our
5 In a contemporaneous study, Demerjian et al. (2013b) also examine the role of managerial ability in smoothing. Our study differs from theirs in that we assess whether discretionary smoothing helps equity investors impound information about future earnings into current returns, while Demerjian et al (2013b) assess the impact of smoothing on future earnings and managerial turnover. Our studies also significantly differ in terms of sample size and smoothing measures, either or both of which may lead to different inferences. Overall, our results complement those in Demerjian et al. (2013b).
4

study should be of interest to researchers, practitioners, and others concerned with understanding the determinants and usefulness of income smoothing.
Our study proceeds as follows. In section 2, we review the relevant literature and develop our hypotheses. In section 3, we provide our research design and data. Section 4 contains our empirical results. We summarize and conclude our study in section 5.
2. Literature Review and Hypotheses Our study is related to research that investigates determinants and consequences of
smoothing, and to research assessing the role of managerial ability in financial reporting. Below, we briefly review this literature and develop our hypotheses.
Managerial Ability and Discretionary Smoothing Following Beidelman (1973) and others, we define income smoothing as the use of
managerial discretion to dampen fluctuations in earnings streams. There is a fairly well developed empirical literature on determinants of smoothing (e.g., Dascher and Malcom 1970; Barnea et al. 1976; McNichols and Wilson 1988; Chaney et al. 1998; Kanagaretnam et al. 2004). We extend this literature by examining whether a managerial characteristic, namely ability, helps to explain discretionary smoothing and its relation to earnings informativeness.
Prior research supports a plausible link between managerial ability and discretionary smoothing. Schipper and Vincent (2003) suggest that superior managers are in a better position to smooth earnings due to their ability to forecast future earnings changes. Kirschenheiter and Melumad (2004) provide a model in which higher ability managers smooth transitory components of earnings.
5

We argue that high ability managers have incentives to use smoothing to signal their private information. Conceptually, Chaney and Lewis (1995) provide a model based on Spence (1973) to argue that “high-quality” firms use smoothing to signal their type. Prior empirical research suggests that smoother earnings are associated with higher firm value, thereby making managers’ equity-based compensation more valuable. Barth et al. (1999) show that the market attaches a premium to firms with year-over-year earnings increases. Tucker and Zarowin (2006) document that smoother earnings increase the informativeness of future earnings. Consistent with this view, Graham et al. (2005) reports that 80% of financial executives believe that smoother earnings help analysts and investors predict future earnings. In addition, CEOs’ concerns about their job security also create an incentive for managers to smooth earnings. Fudenberg and Tirole (1995) provide a model which suggests that concerns about job security create an incentive for managers to smooth earnings; DeFond and Park (1997) provide some empirical support for this model. The preceding analysis suggests a positive relation between managerial ability and discretionary smoothing, and leads to our first hypothesis, stated in the alternative form: H1a: There is a positive relation between discretionary smoothing and managerial ability.
We may not find evidence consistent with our hypothesis if the costs of smoothing outweigh the benefits of doing so. High ability managers might suffer a loss of compensation (Falato et al. 2012; Graham et al. 2011) and reputation (Fudenberg and Tirole 1995) if unexpected negative shocks in the future lead to a poor mapping of current period smoothing to future earnings realization. Additionally, Ronen and Sadan (1981) argue that smoothing is costly due to actions by auditors, legal liability or regulatory intervention (e.g., SEC enforcement). Moreover, potential costs associated with the revelation of proprietary information also likely
6

reduce incentives to smooth. Managers plausibly incorporate the likelihood of these costs when deciding on smoothing decisions.
We also investigate whether cross-sectional differences in firms’ information asymmetry influence managers’ use of smoothing to signal their private information. 6 This analysis is motivated in part by Louis and White (2007), who suggest that managers’ incentive to signal private information through repurchase tender offers increases with firms’ information asymmetry. We therefore expect that high ability managers’ incentive to smooth increases when their firms face higher information asymmetry.
H1b: The relation between managerial ability and discretionary smoothing is more pronounced for firms with higher information asymmetry.
Discretionary Smoothing and Earnings Informativeness A natural question arising from our discussion about the relation between managerial
ability and smoothing is whether it is decision useful. The extant research on the usefulness of smoothing provides mixed results. Non-U.S.-based studies generally provide results which suggest that smoothing reduces informativeness. Leuz et al. (2003) show that smoothing is higher in countries with weaker investor protection. Bhattacharya et al. (2003) find that smoothing is related to a higher cost of capital and lower trading volume. Biddle and Hilary (2006) report that smoothing is related to lower investment efficiency. DeFond et al. (2007) show that more smoothing leads to higher variance in returns around annual earnings announcements.
In contrast to the cross-country studies, U.S.-based studies provide evidence suggesting that smoothing improves informativeness. Subramanyam (1996) reports that returns increase in contemporaneous discretionary accruals, while Tucker and Zarowin (2006) report that smoothing
6 Note that signaling was originally developed as a mechanism to mitigate information asymmetry (Spence 1973).
7

improves the incorporation of information contained in future earnings into current returns. We extend Tucker and Zarowin (2006) by identifying a potential source of cross-sectional variation in smoothing (i.e., managerial ability) and assessing whether it improves the informativeness of future earnings. As Dechow et al. (2010, p. 390) indicate, studies that take a complete path approach “…substantially enhance our understanding of earnings quality”. Because our assessment of the relation between managerial ability and smoothing, as well as its impact on earnings informativeness provide such a complete path, we can draw strong inferences about the usefulness of smoothing and therefore make a significant contribution to the literature.
To the extent that high ability managers have superior skills in assessing their firms’ future performance, we expect that their discretionary accounting choices to smooth earnings will be more accurate and, consequently, of higher quality than those of lower ability managers. The foregoing arguments lead to our second hypothesis, stated in the alternative form, as follows: H2a: Compared to discretionary smoothing by low ability managers, discretionary
smoothing by high ability managers is more informative about future earnings. Furthermore, to the extent that smoothing by high ability managers sends a credible
signal, we expect that the effect of improving earnings informativeness through smoothing will be greater for firms with higher information asymmetry: H2b: The informativeness of discretionary smoothing by high ability managers is more
pronounced for firms with higher information asymmetry.
8

3. Data and Methodology Data
Our initial sample includes firm-years listed in the intersection of the Compustat and CRSP for the years 1991-2009 after excluding financial services and utilities firms.7 We exclude firm-years with acquisition activity in excess of five percent of lagged assets, as major acquisition activity could unduly affect both the measure of managerial ability and income smoothing measures (McNichols 2002; Demerjian et al. 2013a). After requiring data to compute the regression variables, our final sample consists of 36,251 firm-year observations. To mitigate the effect of outliers, we winsorize variables at the 1% and 99% levels.
Managerial Ability Measure To measure managerial ability for each firm, we employ the two-step method developed
by Demerjian et al. (2012). As a first step, Demerjian et al. (2012) use DEA to estimate firm efficiency. DEA is a nonparametric method that uses multiple inputs and outputs to measure the relative efficiency of decision making units (DMUs). DEA creates an efficient frontier of observed production points from linear programming to maximize a ratio of outputs to inputs. DEA assigns a value of one to the most efficient DMUs on the frontier and a value of less than one to inefficient DMUs, with the efficiency score for inefficient units being interpreted as the distance from the frontier (See Cooper, Seiford, and Tone (2000) for more information). The DEA score represents how efficiently the firm utilizes available corporate resources to maximize outputs (Baik et al. 2013). Since the estimated efficiency scores from DEA can be attributable to both the firm and the manager, Demerjian et al. (2012) modify the DEA scores by purging them
7 Our sample begins with 1991 because we require at least three observations for changes in cash flows for SMTH2 and the Statement of Cash Flows became widely available in 1988. We also use data for 2010 through 2012 to measure future returns (Rt3) and future earnings (Xt3) for 2009.
9
Ability ManagersAbilityEarningsFirmsFuture Earnings