Accounting Conservatism and Corporate Governance

Accounting Conservatism and Corporate Governance

Juan Manuel García Lara

Universidad Carlos III de Madrid

Beatriz García Osma

Universidad Autónoma de Madrid

Lancaster University

Fernando Penalva^*

IESE Business School, University of Navarra

September 2005

                                                 

^* Corresponding author. IESE Business School, University of Navarra, Av. Pearson, 21, 08034 Barcelona, Spain. E-mail: penalva@iese.edu. Tel. (+34) 93 253 4200, Fax. (+34) 93 253 4343. 

We appreciate the comments and suggestions from Antonio Davila, Joachim Gassen, Christian Leuz, Flora Muino, Ivana Raonic, William Rees and from seminar participants at the EAA 2005 Annual Meeting, ACCID 2005 Annual Conference, University of Alicante, and IESE Business School.

Accounting Conservatism and Corporate Governance

Abstract

Accounting conservatism produces earnings that reflect bad news faster than good news. We study whether firms with strong corporate governance exhibit a higher degree of accounting conservatism. We assess governance quality using a composite measure that incorporates several governance characteristics. We find that the sensitivity of earnings to bad news is significantly higher (lower) for strong (weak) governance firms. We also study the impact of managerial discretion on the sensitivity of earnings to bad news across governance structures. We show that good governed firms do not use abnormal accruals opportunistically. Rather, they use them to keep investors informed about bad news in a very timely manner.

Keywords:                    Conservatism, corporate governance quality, managerial discretion.

Data Availability:         Data are available from the sources identified in the paper.

1. Introduction

Conservatism in accounting imposes stronger verification requirements for the recognition of gains than for losses, and produces earnings that reflect bad news in a timelier fashion than good news. Explanations for the existence of conservatism posit that it benefits the users of financial reports, as it increases firm value by constraining management’s opportunistic payments to themselves or other parties. The increase in value is then shared among all parties to the firm, increasing their welfare (Watts, 2003a). Given that conservatism increases firm value by reducing agency and litigations costs, managers of good governed firms are expected to comply with accounting regulatory frameworks and choose accounting procedures that delay good news recognition in earnings and capture bad news in a very timely basis.

As developed in Watts (2003a), the contracting and litigation explanations for the existence of conservatism stem from the fact that the different parties to the firm have asymmetric information, asymmetric payoffs, limited liability and different time horizons. Conservatism produces accounting numbers that can be used in contracts among the different parties to reduce the moral hazard problems created by these asymmetries. In addition, conservative accounting, on average, defers earnings and generates lower net assets, which are more likely to reduce expected litigation costs for the firm than overstatement of net assets and cumulative earnings.

Corporate governance plays an important role in the implementation of conservatism. Governance is the set of mechanisms in place (1) to ensure that the assets of the firm are used efficiently and (2) to prevent the inappropriate distribution of the assets to managers or to other parties at the expense of the rest of stakeholders. Hence, strong governance results in better monitoring of management, produces more timely accounting information, accelerates the recognition of bad news to provide the board of directors with early-warning signals to investigate the reasons for the bad news, and reduces the likelihood of incurring litigation costs. Conservative choices result in all these beneficial outcomes, which leads us to the main prediction of this paper.

We propose that firms with strong corporate governance will exhibit a higher degree of accounting conservatism. That is, we expect that the sensitivity of earnings to bad news will be higher for firms with strong corporate governance than for firms with weak governance. To measure the quality of corporate governance, we develop a composite index that takes into account the exposure to the market for corporate control through antitakeover measures and several characteristics of the functioning of the board of directors, after controlling for the economic determinants of governance. 

Using a large sample of US firms during the period 1992-2003, we find that firms with low (high) antitakeover protection and where the CEO has a low (high) influence on governance decisions exhibit a higher (lower) degree of accounting conservatism. That is, the earnings of potentially better-governed firms are significantly timelier in recognizing bad news than the earnings of firms with more antitakeover protection and where the CEO has more influence on governance decisions. We also study the impact of earnings management on the timeliness of earnings to bad news across governance structures. Using several accruals models, we decompose reported earnings into a discretionary part and a nondiscretionary part. We find that the increase in accounting conservatism in strong governance firms is driven by the discretionary component of reported earnings. However, we do not find a significant difference in the sensitivity of unmanaged earnings to bad news between strong and weak governance firms. This evidence is consistent with governance determining the use of accruals to accelerate the recognition of bad news in earnings. Finally, given the evidence in Bushman et al. (2004) who find that the informativeness of earnings is associated with governance structures, we try to shed some light on whether governance influences conservatism or vice versa. Our results are consistent with the direction of causality flowing from governance to conservatism.

There is no previous in depth research on the association between accounting conservatism and governance. Only the work by Beekes et al. (2004) for the UK is in a similar line. They examine the link between accounting quality, measured by earnings timeliness and earnings conservatism, and the proportion of outside directors on the board. Their results indicate that firms with a higher proportion of outside directors are more likely to recognize bad news in earnings on a timely basis. Our research differs from Beekes et al. in several dimensions. First, we use a more complete proxy to measure the level of governance, taking into account several external and internal governance provisions. It is important to do so because both types of governance have a complementarity effect (Cremers and Nair, 2005). Second, we control for the economic determinants of governance and focus on what could be described as discretionary, endogenous or idiosyncratic governance¹. Third, we examine how managers use accruals to communicate good and bad news across firms with different governance quality. Fourth, we take into account the firms’ investment opportunity sets in assessing the difference in asymmetric timeliness across governance structures; otherwise, variation in asymmetric timeliness may simply indicate variation in investment opportunities rather than variation in conservatism (Roychowdhury and Watts, 2004). And fifth, we investigate whether governance influences conservatism or vice versa.

The rest of the paper is organized as follows. Section 2 discuses the impact of corporate governance and develops a metric of governance quality. Section 3 contains the research design, describing the measurement of conservatism and the estimation of discretionary accruals. Section 4 introduces the sample and presents summary univariate statistics. Section 5 analyzes the results: the difference in conservatism across governance structures, the influence of earnings discretion on conservatism across governance structures, the influence of the investment opportunity set on the level of conservatism, and the assessment of whether governance affects conservatism or vice versa. Section 6 concludes.

2. Corporate Governance

2.1 Governance structure

Corporate governance is a combination of external and internal mechanisms. Takeover provisions and the market for corporate control are part of the first group, whereas the board of directors and the presence of blockholders make up the second group. Both types of external and internal governance mechanisms are required for the well-functioning of the firm as they have a complementary effect. Jensen (1993) states that internal governance mechanisms are insufficient to achieve appropriate governance without the presence of the market for corporate control. Recent research indicates that external governance reinforces internal governance and vice versa (Cremers and Nair, 2005). 

Governance design, through mechanisms such as corporate governance provisions (which include antitakeover protection mechanisms like poison pills or golden parachutes), and board of directors’ characteristics, determines the power of stakeholders vis-à-vis management. The literature on corporate governance provides mounting evidence on how governance structures are associated to firm performance and how the distribution of power affects the allocation of rents. Cremers and Nair (2005) show that firms with high takeover vulnerability and with large blockholders, a proxy for strong internal governance, generate abnormal returns of 10% to 15%. Core et al. (1999) find that board of directors’ characteristics associated with weak governance —including CEO holding the chairman position, board size, directors appointed by the CEO, gray outside directors, old directors, and busy directors— are correlated with higher levels of CEO compensation after controlling for economic determinants of compensation; moreover, they find that predicted excess compensation, based on the governance structure of the firm, is negatively correlated with stock returns one, three and five years ahead.

Various arguments support these findings. The most cited one is to attribute them to an increase in agency costs associated with weak governance. Agency costs appear because of the lack of alignment among the interests of the different parties to the firm, which seek to maximize their own welfare instead of the value of the firm. These costs translate, among other things, into lower operational performance, excess rents to managers in the form of compensation, excess distributions to shareholders at the expense of bondholders, and are consistent with the existence of inefficient contracts. Contracts written using conservative accounting numbers contribute to reduce agency costs. Thus, we expect that strong governance structures will tend to favor accounting conservatism more than weak governance structures.

2.2 Measurement of governance quality

In Section 2.2.1, we develop a measure of unconditional total governance that incorporates attributes of internal and external governance. In Section 2.2.2, we take into account the economic determinants of governance and build a measure of idiosyncratic governance.

2.2.1 Measurement of unconditional governance quality

We measure the quality of unconditional governance using an approach similar to the one in Bertrand and Mullainathan (2001). Specifically, we develop a composite governance variable (Totgov) that incorporates the level of antitakeover protection (external governance) and several characteristics of the board’s structure (internal governance). This variable combines the following four governance proxies:

(a)         We use as an external governance proxy the takeover protection index developed by

Gompers et al. (2003). Their index can be interpreted as a measure of shareholders’

rights and, as Cremers and Nair (2005), also as a measure of takeover vulnerability. Using data compiled by the Investors Responsibility Research Center (IRRC) and state takeover law data, Gompers et al. construct an index for each sample firm by adding one point for every provision that reduces takeover vulnerability.² Higher values of this index are associated with more protection against takeovers. Cremers and Nair (2005) also use a narrower alternative takeover index that only accounts for the three components of the IRRC data that are critical to takeovers. They report that their results do not change and conclude that there are no systematic biases in the Gompers et al. index, and that it can be correctly interpreted as a measure of takeover protection.

(b)        The Gompers et al. index does not capture information on internal governance, such as board characteristics. Hermalin and Weisbach (1998, 2003) argue that the main factor affecting board’s effectiveness is its independence from the CEO. Expanding this argument, we include an indicator variable that takes on the value of one if the CEO is also the chairman of the board and zero otherwise. The CEO has more influence on governance when the same person holds the CEO and chairman titles.

(c)         Previous research finds that independent directors positively influence board decisions. Weisbach (1988) shows that the presence of outside directors is positively related to CEO removal decisions. Byrd and Hickman (1992) find that bidding firms on which independent outside directors hold at least 50% of the seats have significantly higher announcement-date abnormal returns than other bidders. As a second proxy for internal governance, we include the proportion of top executives that serve on the board. Higher proportions of executives on the board are associated with higher CEO influence on governance. 

(d)        Finally, Adams (2000) and Vafeas (1999) suggest that the number of board meetings is a good proxy for the directors’ monitoring effort. We include the inverse of this variable, where a higher value is associated with lower board effectiveness. 

Following Bertrand and Mullainathan (2001), we define the composite governance variable (Totgov) by taking the unweighted average of the standardized variables.³ The standardization is performed to take into account the different scales of the variables that make up the composite measure. Higher values of Totgov are expected to be associated with governance structures with higher antitakeover protection and high CEO influence on board decisions. For brevity, we refer to these structures as weak governance. Conversely, governance structures with low antitakeover protection and low CEO involvement in board decisions are referred to as strong governance. This is the meaning that we attach to weak and strong governance throughout the paper.

2.2.2 Measurement of idiosyncratic governance

Governance quality is partly determined by factors that do not reflect management’s choices. We focus on the idiosyncratic component of corporate governance. Predetermined firm characteristics, such as industry sector, growth opportunities, risk, etc., will lead to widespread adoption of certain governance provisions across similar firms. However, managers enjoy wide discretion as to how to implement and enforce these provisions, and whether to use others. In this sense, previous studies that analyze governance using these predetermined economic factors are only able to explain a reduced percentage of the total governance quality (Ashbaugh et al., 2004; Bushman et al., 2004). We adopt the view that the quality of governance structures should be analyzed leaving aside these predetermined economic factors. To exclude these factors from our measure of corporate governance quality and create a measure of idiosyncratic governance, we develop a model of hypothesized economic determinants of governance. Based on the findings of previous literature, we predict that governance will be partly determined by the following variables:

(a)         Size. Larger firms are more complex and place higher demands on governance structures. Demsetz and Lehn (1985) find that size is significantly associated with ownership concentration. We measure size as the three-year average of the natural logarithm of the market value of equity, measured at the end of the fiscal year, Log(market value of equity), and predict a negative association with our proxy for unconditional governance, Totgov. 

(b)        Growth opportunities. Previous research documents an association between growth opportunities and governance. Following Smith and Watts (1992), our (inverse) proxy for growth is the three-year average of the annual book-to-market value of assets ratio, measured at the fiscal year end, Book-to-market. The market value of assets is defined as the market value of equity plus the book value of liabilities. 

(c)         Firm age. We hypothesize that firm age is related to governance structure. Following Bushman et al. (2004), our proxy is the natural logarithm of the firm’s age at the end of the fiscal year, Log(firm age), measured as the number of years the firm has been public.

(d)        Free cash flow problem. High free cash flow poses a problem for firms with low growth opportunities, because managers may invest excess cash in negative net present value projects or engage in empire-building acquisitions. Jensen (1986) suggests that governance structures can mitigate this agency problem. Following Lang et al. (1991), our proxy to capture this determinant of option incentives, Free cash flow problem, is the three-year average of [(operating cash flow minus preferred and common dividends)/total assets] if the book-to-market ratio is greater than or equal to one; and zero otherwise. Firms with book-to-market ratios greater than one are expected to have low growth opportunities. The free cash flow problem demands better governance; therefore, we expect to find a negative association between Free cash flow problem and Totgov.

(e)         Idiosyncratic risk. Demsetz and Lehn (1985) suggest that the amount of noise in the firm’s operating environment is expected to increase the costs of direct monitoring, which in turn, increases the demands on governance structures. These costs are expected to grow at a decreasing rate with the difficulty in monitoring. Hence, we use the logarithmic transformation of the firm’s idiosyncratic risk. Idiosyncratic Risk is defined as the natural logarithm of the standard deviation of the residual return from a

36-month market model regression of the firm’s monthly returns on the returns to the CRSP value-weighted market portfolio, imposing a minimum of 12 observations. We predict a negative association between Idiosyncratic risk and our proxy for unconditional governance, Totgov.

(f)         Leverage. Cremers and Nair (2005) find that internal and external governance mechanisms are stronger complements in firms with low leverage, because higher debt reduces the probability of a takeover, making the target less attractive to the prospective acquirer. This fact reduces the governance usefulness of antitakeover mechanisms. Our proxy for leverage is the ratio of short and long term debt to total common shareholders equity, and we expect to find a positive association between Leverage and Totgov as higher leverage requires lower antitakeover protection.

(g)        Industry concentration and geographic concentration. Bushman et al. (2004) argue that organizational complexity increases with industry and geographic diversification. These authors hypothesize and find that the complexity associated with diversification causes costly governance responses, because the inherent additional managerial difficulties generated by more complex firms place higher demands on the governance structures. To control for the level of diversification, we employ the same proxies used by Bushman et al. (2004). Industry concentration is defined as the three-year average of the sum of the squares of the firm sales in each industry segment divided by total firm sales. Geographic concentration is defined as the three-year average of the sum of squares of the firm sales in each geographic segment divided by total firm sales. Higher values of these two proxies indicate more industry/geographic concentration. These proxies are inverse measures of

diversification; therefore, we expect to find a positive association with Totgov.

(h)        CEO tenure. Hermalin (2005) develops a model in which a trend towards more board diligence leads to shorter CEO tenures. Bushman et al. (2004) find that the number of years the CEO has been a director is positively associated with the presence of more inside directors in the board. Hermalin and Weisbach (1988) find that board independence declines over the course of the CEO’s tenure. We hypothesize that the number of years the CEO has been in office, CEO tenure, is another determinant of governance as longer tenures increase the likelihood of having more insiders in the board. We predict a positive association between CEO tenure and Totgov. 

(i)          Earnings timeliness. Bushman et al. (2004) find that the informativeness of earnings is associated with governance structures. They argue that firms operating in less transparent accounting environments will try to balance the lower quality of the accounting numbers by placing more demands on the governance structures. However, they are unable to establish the direction of the causality, as they cannot rule out the possibility of governance affecting the quality of accounting information. We use their measurement of earnings timeliness as a control variable. Earnings timeliness is the average of the variables Rank_REV_SLOPE, Rank_REV_R2 and

Rank_ERC_R2 as described in Bushman et al.: equation (2) below is estimated annually for each firm using a time series of ten years (with a minimum of eight years) where R (return) is defined as 15-month (ending three months after fiscal year end) stock return; the percentile ranking of the estimated coefficient β₂ is referred to as Rank_REV_SLOPE. The percentile ranking of the R² of the previous regressions is referred to as Rank_REV_R2. The percentile ranking of the R², referred to as Rank_ERC_R2, of the firm-specific regression of 15-month (ending three months after fiscal year end) stock return on the level of and on the change of earnings before extraordinary items and discontinued operations deflated by market value of equity at the beginning of the period. Each regression has ten years of data, with a minimum of eight years. We predict a positive association between Earnings timeliness and Totgov. Notice that this proxy is different from Basu’s (1997) measure of conservatism, namely the asymmetric timeliness of earnings.

(j)          Performance. Previous research documents the association between certain governance attributes and past firm performance. Hermalin and Weisbach (1988) find that the likelihood of independent directors being added to the board increases following poor firm performance. Similar to Demsetz and Lehn (1985), to control for past firm performance we use the three-year stock return measured as the continuously compounded monthly CRSP return over 36 months, ending at fiscal year end.

(k)        Regulation. The additional monitoring provided by regulators may systematically affect the governance characteristics of firms operating in regulated environments. Following Demsetz and Lehn (1985) and Bushman et al. (2004), we include an indicator variable that takes on the value of one if the firm is a utility, and zero otherwise. As explained below, we do not control for financial firms because our sample excludes these firms.

We estimate annually the following cross-sectional model of the economic determinants of governance:

              Totgov _t = α₁ Log(market value of equity) _t-1 + α₂ Book-to-market _t-1

                               + α₃ Log(firm age) _t-1 + α₄ Free cash flow problem _t-1 

                               + α₅ Idiosynchratic risk _t-1 + α₆ Leverage _t-1

                               + α₇ Industry concentration _t-1 + α₈ Geographic concentration _t-1

                               + α₉ CEO tenure _t + α₁₀ Earnings timeliness _t-1 + α₁₁ Performance _t-1

                               + α₁₂ Regulation + α₁₃ Constant + ε _t                                                                             (1)

Finally, we define our composite idiosyncratic governance variable (Totgov^*) as the firm-year residuals from equation (1) and use Totgov^* as our proxy for the level of governance after controlling for firm characteristics. These residuals capture the variation in governance orthogonal to firm characteristics. This is, variation unaccounted for by these characteristics. Notice that in the model, CEO tenure is measured contemporaneously to Totgov. The reason is to avoid losing one year of observations as CEO tenure is only available as of 1992. The results are not affected by this choice.

Firms with low (high) values of Totgov^* are those that, with respect to their peers, have low (high) antitakeover protection and low (high) CEO involvement in board decisions, which we refer to as firms with strong (weak) idiosyncratic governance. Differences in the quality of governance across firms with similar characteristics and contracting environments are explained by (1) economic determinants only influencing governance structures. The governance structure is largely determined endogenously by the firm; and (2) the fact that governance is sticky and changes occur slowly. It takes time to put in place, modify or remove antitakeover provisions, or to alter the composition and functioning of the board of directors. Firm characteristics and the contracting environment can change at a fast rate whereas governance structures are adapted at a slower pace and respond with a certain lag to the new demands.

3. Research Design

3.1 Measurement of conservatism

We adopt Basu’s (1997) measure of conservatism. Under conservative accounting, earnings capture bad news faster than good news, because of the asymmetric standards of verification of losses and gains. Basu uses stock returns to proxy for good or bad news. Stock prices incorporate all the information arriving to the market from multiple sources, including reported earnings, in a timely fashion. Therefore, stock price changes are a measure of news arrival during the period. Because earnings are timelier in recognizing bad news than good news, Basu expects to find a higher association of earnings with negative returns (his bad news proxy) than with positive returns (the good news proxy). We use Basu’s regression as follows:

                 X_t =β₀ +β β β₁D_t + ₂R_t + ₃D R_{t t} +µ_t                                                                              (2)

where X_t is earnings per share before extraordinary items and discontinued operations deflated by share price at the beginning of the period. R_t is the stock rate of return of the firm, measured compounding twelve monthly CRSP stock returns ending the last day of fiscal year t.⁴ D_t is a dummy variable that equals 1 in the case of bad news (negative or zero stock rate of return) and 0 in the case of good news (positive stock rate of return). The coefficient β₃ measures the level of asymmetric timeliness—the level of conservatism—and it is expected to be positive and significant.

 In our first set of tests, using (Totgov^*), we perform several partitions of the sample to compare whether there are significant differences across governance groups. We hypothesize and find that the asymmetric timeliness coefficient β₃ is significantly higher for firms with strong governance. In our partitions, we also control for differences in the investment opportunity set. It is important to do so, because variation in the investment opportunity set may cause variation in asymmetric timeliness not related to differences in conservatism, as pointed out by Roychowdhury and Watts (2004). Following these authors, our proxy for the investment opportunity set is the ratio of market-to-book value of equity (MTB) measured at the end of the fiscal year. The aim of our tests is not to estimate precisely the level of conservatism but to assess whether there are significant differences in conservatism across governance structures, after controlling for the investment opportunity set. The market-tobook ratio is also a potential proxy for conservatism in the balance sheet, this is, understatement of net assets due to the non-recognition of intangible assets or to the use of historical cost. According to Pope and Walker (2003) and Beaver and Ryan (2005), firms with a greater understatement of assets are likely to have a less pronounced asymmetric timeliness of earnings, as they are anticipating bad news. Conservatism in the balance sheet is no more than an extreme form of conservatism in earnings. Bearing this in mind, it is prudent to control for the understatement of assets so that comparisons across partitions of firms with strong and weak corporate governance are not influenced by balance sheet conservatism.

 In a second set of tests, we investigate the reason for the observed differences in asymmetric timeliness. Extant research on corporate governance finds that firms with weak corporate governance structures engage in more earnings manipulation, that is, they have lower quality earnings and accruals (e.g. Dechow et al., 1996; Becker et al., 1998; Peasnell et al., 2001; Klein, 2002). However, Bowen et al. (2004) find that, on average, variation across governance structures in the use of discretionary accruals is not driven by opportunistic reasons; rather, accruals are used as a signaling device to convey information to the market. This is consistent with managers using discretionary accruals to make accounting information more relevant, aligning earnings and returns (Guay et al., 1996). Based on these findings, we hypothesize that stronger governance structures provide managers with incentives to make more conservative accounting choices by using discretionary accruals. To test this prediction, we run equation (2) taking into account the possible effect of earnings discretion on asymmetric timeliness. 

To disentangle the effects of earnings discretion and conservatism, we start from the simple accounting equality that earnings equal cash flows plus total accruals (X_t = CFO_t + TACC_t). Cash flows are expected to capture news symmetrically. Thus, accountants will use accruals to make earnings timelier.⁵ Accruals can be further decomposed into nondiscretionary (normal) and discretionary (abnormal) components. Several discretionary accruals models are used in the literature and there is currently much debate on the appropriateness of the different methods. It is beyond the scope of this paper to enter this controversy. We estimate discretionary accruals using four different methodologies as a robustness check. In this way, we expect to minimize the likelihood of our results being driven by the particular choice of discretionary accruals estimation method.

3.2 Estimation of discretionary accruals

To obtain the expected accruals model for all firms in each industry j for year t, we use the modified Jones (1991) model defined by Dechow et al. (1995). We delete industry-year combinations with less than six observations. We estimate the following model crosssectionally for all remaining industry-year combinations (in parenthesis, we provide the Compustat data items):

                TACC _{j t,}                            ⎡ 1 ⎤              ⎡∆REV_{j t,} ⎤          ⎡PPE_{j t,} ⎤

                   =α_{j t,} ⎢  ⎥+β_{j t,} ⎢  ⎥+γ_{j t,} ⎢  ⎥+ε_{j t,}                                       (3)

                  TAj t, −1                      ⎢⎣TAj t, −1 ⎥⎦         ⎢⎣ TAj t, −1 ⎥⎦        ⎢⎣ TAj t, −1 ⎥⎦

where TACC are total accruals, calculated as the difference between net income before extraordinary items and cash flows from operations (Compustat item #18 - Compustat item #308).⁶ TA are total assets (Compustat item #6), ∆REV is the change in net sales (Compustat item #12), and PPE is gross property, plant and equipment (Compustat item #7). 

TACC j t,        ⎛       ⎢⎡ 1 ⎥⎤+βˆ j t, ⎡⎢∆REVj t, −∆ARj t, ⎥⎤+γˆj t, ⎢⎡PPEj t, ⎥⎤⎞⎟⎟                      DAXt = TAj t, −1 −⎜⎜⎝αˆ j t, ⎢⎣TAj t, −1 ⎦⎥        ⎣⎢        TAj t, −1                 ⎥⎦        ⎢⎣ TAj t, −1 ⎥⎦⎠

(4)

Next, for each firm i, we calculate its discretionary accruals as:

where αˆ,βγ^ˆ, ˆ are the fitted coefficients from equation (3) and ∆AR is the change in accounts receivable (Compustat item #2). From this model we obtain our first discretionary accrual proxy (DAX1).

As current research indicates that management has the most discretion over current accruals, and that manipulation of long-term accruals is unlikely due to its high visibility (Becker et al., 1998; Young, 1999), we define a second measure of discretionary accruals using a measure of current accruals manipulation:

                                WCA_{j t,}                             ⎡ 1 ⎤              ⎡∆REV_{j t,} ⎤

                                   =α_{j t,} ⎢  ⎥+β_{j t,} ⎢  ⎥+ε_{j t,}                                                     (5)

                                 TAj t, −1                    ⎢⎣TAj t, −1 ⎥⎦         ⎢⎣ TAj t, −1 ⎥⎦

where WCA are working capital accruals, calculated as net income before extraordinary items (Compustat item #123) plus depreciation and amortization (Compustat item #125) minus operating cash flows (Compustat item #308). Using the parameter estimates from equation (5) we calculate discretionary (abnormal) working capital accruals (AWCA) as follows:

                                          WCAj,t                  ⎛⎜αˆ ⎡ 1 ⎤                 ⎡∆REV −∆AR ⎤⎞

AWCAj,t = TA ⎜ j,t ⎢⎢TAj,t−1 ⎥⎦⎥+βˆ j,t ⎣⎢⎢  TAj,t j,t−1                        j,t ⎦⎥⎥⎟⎟⎠                          (6) −         j,t−1      ⎝           ⎣

Our second measure of discretionary accruals (DAX2) is the abnormal working capital accruals of each firm obtained from this equation.

Prior research recognizes the need to control for the effect of performance in tests of earnings management (e.g. Teoh et al., 1998a; 1998b) and to take into account the correlation between cash flows and accruals (Dechow, 1994). We follow recent literature and calculate discretionary accruals using the Kasznick (1999) model (DAX3). Finally, we take into account the developments suggested by Kothari et al. (2002) and adjust our measures of abnormal accruals using the firms' lagged return-on-assets (lagROA) to obtain our final discretionary accruals measure (DAX4).

Thus, we obtain four metrics of discretionary accruals: DAX1 through DAX4. To make these measures of discretionary accruals consistent with our definition of earnings in equation (2), we multiply DAX by lagged total assets and divide it by lagged stock price. 

To perform our tests on the influence of earnings discretion on conservatism across governance structures, we replace the dependent variable in equation (2) as in García Lara et al. (2005), first for a measure of discretionary accruals (DAX_t), and then, for earnings before discretionary accruals (X_t^* = X_t – DAX_t), using our four alternative DAX measures. If discretionary accruals are one of the tools used by management to achieve a higher level of conservatism in strong governance firms, we expect to find significant differences in the asymmetric timeliness coefficient β₃ across governance structures when the dependent variable in the Basu regression is discretionary accruals (DAX_t). However, we do not expect to observe these differences when the dependent variable is earnings before discretionary accruals (X_t^*).

4. Sample Description and Estimation of the Governance Model

Accounting data are taken from the 2003 version of Compustat. Industry and geographic concentration variables are constructed with segment data extracted from the Compustat Business Information File. Market return data are taken from CRSP. Board characteristics and CEO data come from the 2003 version of Execucomp. The antitakeover protection index constructed by Gompers et al. (2003) with IRRC data was downloaded from Andrew Metrick’s web page⁷. The Execucomp and the IRRC data cover approximately the 1,500 firms that make up the S&P 500, MidCap and SmallCap indices. We eliminate firms with negative book value of equity, and firms in the financial sector (SIC 6000-6999) because the discretionary accrual methods are not appropriate for these firms. To reduce the adverse effect of outliers, observations in the top or bottom 1% of stock returns (R_t), deflated earnings (X_t) and discretionary accruals (DAX_t) are truncated. In addition, after visual inspection, the variables market-to-book value of equity (MTB), Leverage, and CEO tenure are winsorized at the top percentile of their distributions. The intersection of these databases and the additional data requirements yield a sample that contains 9,209 firm-year observations for the period 1992-2003, corresponding to 1,623 different firms.

Table 1 contains descriptive statistics of the data used to estimate the governance model described in equation (1). In our sample, on average, firms have 9 antitakeover provisions, the board meets 7 times per year, 32% of the board members are executives, and the CEO is also the chair of the board 73% of the times. The average book-to-market value of assets ratio over the sample period is 0.63, the average size in market-value-of-equity terms is $1,317 million and the average CEO tenure is close to 8 years. The sample firms exhibit similar industry and geographic concentration, and the mean annualized market return for the previous three years is 0.18 (or 0.66 for the three-year period). Table 2 depicts the Pearson correlation matrix of the variables in the governance model. 

Table 3 contains the estimation of the governance model described in equation (1). We estimate the model annually to avoid look-ahead bias. With the residuals of the annual regressions, we construct our proxy for idiosyncratic governance, Totgov^*. The table shows the average estimated coefficients of the annual regressions. The coefficient estimates are significant in most cases and have the predicted signs. Larger firms, with low growth opportunities, high free cash flow problem, and high idiosyncratic risk place more demands on the governance structures (i.e., negative association with Totgov). On the contrary, older firms, with high leverage, high geographic concentration and longer CEO tenures appear to be associated with less demanding governance structures. Additionally, regulation seems to positively affect the quality of governance. To make sure that influential observations in the dependent variable Totgov are not unduly affecting the coefficients, we also run median regressions, which fit a line through the data that minimizes the sum of the absolute residuals rather than the sum of the squares of the residuals as in an ordinary regression. The results are virtually identical up to the third decimal.

Table 4 contains the summary statistics of the variables used in our tests of the association between conservatism and governance. We partition the sample in two groups at the median of idiosyncratic governance Totgov^*. Observations below (above) the median are those firms with low (high) antitakeover protection and low (high) CEO involvement in board decisions. For simplicity, we refer to these two groups as firms with strong and weak governance, respectively. The summary statistics indicate that, on average, strong (weak) governance firms have 8 (10) antitakeover provisions, the board meets 8 (6) times per year, have 26% (38%) of the board made up of executives, and the CEO is also the chairman of the board 54% (92%) of the times. The market-to-book ratio of strong governance firms is 3.14 whereas the one of weak firms is 3.09, although the difference is not statistically significant at conventional levels. On the contrary, strong firms have lower average deflated earnings (X) and discretionary accruals (DAX) than weak firms: 0.033 and 0.003 versus 0.042 and 0.006, respectively. Consistent with the existence of conservatism in earnings, that is, with the asymmetric timeliness of earnings, earnings are negatively skewed (medians exceed means). The skewness exists both for earnings (X) and earnings before discretionary accruals (X*), and it is more pronounced in earnings for firms with strong corporate governance. Table 5 depicts the correlation matrix.

5. Empirical Results

5.1 Difference in conservatism across governance structures

Table 6 contains the results of the estimation of equation (2) when we partition the sample at the median of the total idiosyncratic governance proxy (Totgov^*). The t-statistics reported in all regressions are based on Huber-White standard errors, which are robust to both heteroscedasticity and serial correlation (Rogers, 1993).

Table 6 shows the estimation results using pooled regressions. When the dependent variable is earnings (X), the asymmetric timeliness of earnings coefficient β₃ provides an estimate of the level of conservatism. We observe that strong governance firms are more conservative than weak governance firms (0.15 vs. 0.12). The difference in conservatism is significant (p-value = 0.04), thus providing initial support to our hypothesis⁸. To rule out the possibility that our results are being influenced by possible cross-sectional dependence problems, we also use Fama and MacBeth (1973) mean annual regressions.⁹ Unreported results confirm our initial findings. It is worthwhile to mention the large size in the negative returns coefficient β₃ and the small size of the positive returns coefficient β₂. This is consistent with recent evidence (Basu, 1997; Ball et al., 2000). Interpreting this evidence, Watts (2003b) concludes that in recent years “US firms’ accounting earnings are not timely at all in reflecting good news but are timely in reflecting bad news.” Watts attributes this significant increase in conservatism to the influence of the Financial Accounting Standards Board (FASB), the US standard-setter. 

We also measure conservatism using the metric developed by Penman and Zhang (2002). They construct an index of conservatism (C-Score) that captures the effect of conservative accounting on the balance sheet.¹⁰ They define the C-Score as the level of estimated reserves created by conservatism relative to net operating assets. The C-Score, as other measures of conservatism in the balance sheet (like the market to book ratio), captures extreme forms of earnings conservatism, where bad news is not just captured on a very timely basis, but even anticipated. Unreported results show that the mean (median) C-Score for strong governance firms is significantly higher than that for weak firms: 0.38 vs. 0.27 (0.08 vs. 0.06). These differences are statistically significant at a confidence level of 0.01.

5.2 Influence of earnings discretion on conservatism across governance structures

To understand the reason for the previous findings, we investigate the possible influence of earnings management on conservatism across governance structures. The two middle columns in Table 6 show the estimation of equation (2) when the dependent variable is replaced by an estimate of discretionary accruals (DAX). For parsimony, we only report the results that use the modified Jones model to estimate discretionary accruals (DAX1). The results are not affected by the choice of accruals estimation method. The 3^rd and 4^th columns of Table 6 show that the difference in conservatism between strong and weak firms (0.06 versus 0.02) is significant (p-value = 0.05). However, when the dependent variable in the Basu regression (2) is earnings before discretionary accruals (X^*), we do not find any difference in conservatism, as depicted in the last two columns of Table 6. This evidence is consistent with strong governance firms using discretionary accruals to report more conservative earnings than weak firms. The findings in Bowen et al. (2004) also add support to this interpretation, as they document that, on average, management does not use discretionary accruals opportunistically, but rather to convey information to the market.

However, our results could be driven by our particular choice of sample partition. To reject this possibility, we also partition the sample using the quartiles of Totgov^*. We define the strong (weak) governance firms as those in the bottom (top) quartile of Totgov^*. Then, we repeat the analyses in Table 6. The unreported results yield the same inferences. 

As a further robustness test, we modify equation (2) to include the level of total idiosyncratic governance as follows:

Dep. Var_t = β₀ + β₁ D_t + β₂ Totgov^*_t + β₃ R_t + β₄ D_t Totgov^*_t + β₅ R_t Totgov^*_t 

                                        + β₆ D_t R_t + β₇ D_t R_t Totgov^*_t + u_t                                                          (7)

where the dependent variable is either X, DAX or X^*. In this way, we can use all the information in Totgov^* and avoid the possible arbitrariness of particular sample partitions. This specification presents a further advantage, because it allows to jointly analyze firms in equilibrium and firms out-of-equilibrium with respect to their governance structures. Equilibrium firms are those with zero or small values of Totgov^*. For these firms, we do not expect to find significant differences in conservatism. On the contrary, we do expect to observe differences in conservatism between strong and weak governance firms; that is, firms with low and high values of Totgov^*, respectively. In particular, we hypothesize that the asymmetric timeliness coefficient β₆ will be positive and significant and that β₇ will be negative. Thus, the total conservatism (β₆ + β₇) of weak governance firms will be smaller than that of strong firms, because higher values of Totgov^* are associated with weaker governance. 

Table 7 contains the estimation of equation (7). When the dependent variable is X or DAX, the β₆ coefficient is positive and significant and the β₇ coefficient is negative and also significant. The reduction in conservatism is of, approximately, the same size as the difference in the β₆ coefficients across governance structures that we report in Table 6. However, when the dependent variable is X^*, the coefficient β₇ becomes insignificantly different from zero, indicating that there is no difference in conservatism once discretionary accruals are controlled for.

5.3 Influence of the investment opportunity set on the level of conservatism

Roychowdhury and Watts (2004) show that, it is important to control for the investment opportunity set when estimating the level of asymmetric timeliness. Variation in growth opportunities can influence the variation in the estimate of asymmetric timeliness for reasons unrelated to conservatism. To check that our results are not driven by differences in the investment opportunity set, we repeat all the tests in Table 6 controlling for the level of the market-to-book ratio (MTB), our proxy for the investment opportunity set, which can also be understood as a proxy for conservatism in the balance sheet. Following Roychowdhury and Watts (2004), to perform this test we introduce MTB into equation (2) in the following fashion:

Dep. Var_t = β₀ + β₁ D_t + β₂ MTB_t + β₃ R_t + β₄ D_t MTB_t + β₅ R_t MTB_t 

                                                + β₆ D_t R_t + β₇ D_t R_t MTB_t +  µ_t                                                             (8)

where the dependent variable is either X, DAX or X^*. 

Table 8 shows the estimation results of equation (8) across governance structures when the sample is partitioned at the median of Totgov^*. Even after controlling for the investment opportunity set, strong governance firms exhibit more accounting conservatism than weak firms. When the dependent variable is earnings (X) the first two columns of Table 8 indicate that the difference in the asymmetric timeliness coefficients β₆ across governance structures is still significant at conventional levels, whereas the coefficient β₇ that captures the influence of investment opportunities on asymmetric timeliness is not significantly different from zero. The same result occurs when the dependent variable is discretionary accruals (DAX) as shown by the third and fourth column of Table 4. However, when the dependent variable is earnings before discretionary accruals (X^*) we find no difference in asymmetric timeliness across governance structures even after controlling for investment opportunities as shown in columns five and six of Table 8. Overall, the findings in Table 8 confirm that the observed differences in conservatism across governance structures depicted in Tables 6 and 7 are not driven by differences in investment opportunities. These results are not unexpected because the summary statistics in Table 4 show that the mean and median of MTB for strong and weak governance firms are not significantly different between the two groups.

5.4 Does governance influence conservatism or vice versa?

Bushman et al. (2004) document an inverse association between measures of the informativeness of accounting numbers and governance. In particular, they posit that firms that produce accounting information of limited transparency place a higher burden in governance structures in order to overcome this shortcoming. They measure the informativeness of accounting numbers using earnings timeliness, which they define as “the extent to which current accounting earnings incorporate current economic income or valuerelevant information.” They find that past earnings timeliness is negatively associated with current governance quality. However, they are unable to rule out the possibility “that governance structures also influence the properties of accounting numbers through accounting policy choices and earnings management activities” because their test is a simple association test that is not informative about the direction of the causation. They conclude that their proxy for earnings timeliness captures a firm characteristic over which management has little discretion. This evidence in Bushman et al. leads us to include their proxy for earnings timeliness among the explanatory variables in our governance model (1). Notice that our proxy for conservatism, Basu’s earnings asymmetric timeliness, is different from the measure of the usefulness of accounting numbers used by Bushman et al. Their measure captures earnings symmetric timeliness.

In our study, we are implicitly assuming the direction of causation: better governance results in more conservative accounting choices. Our findings in Section 5.1 document a positive association between governance and conservatism. This is a necessary condition but not sufficient to infer the direction of causation. In addition, the discretionary accruals results in Section 5.2 seem to provide some support for the hypothesis that governance influences conservatism by providing a plausible link between the two. Nevertheless, this evidence is clearly insufficient to draw any meaningful conclusion. To overcome this fact and to try to shed some light on whether governance influences conservatism or vice versa, we incorporate some dynamic features in our tests in order to find more evidence consistent with our implicit assumption.

In our first test, we partition the sample into strong and weak governance firms using Totgov^* at different points in time (from t–3 to t+3). To increase the power of our tests, we define as strong (weak) governance firms those in the bottom (top) quartile of Totgov^*. Then we run regression (2) at time t and compare the asymmetric timeliness coefficients β₃ of strong and weak governance firms. We begin partitioning the sample at time t–3. That is, we look at whether firms that were strong three years ago exhibit now more accounting conservatism than firms that were weak. The first two columns of Table 9 show the results when the dependent variable is X. When the sample is partitioned at time t–3, three years later we observe that strong governance firms exhibit higher accounting conservatism than weak firms (β₃ = 0.15 vs. 0.08) and the difference is statistically significant (p-value = 0.02). We obtain the same result when we partition the sample at time t–2: two years later strong governance firms exhibit more accounting conservatism than weak firms (β₃ = 0.18 vs. 0.11) and the difference is statistically significant (p-value = 0.01). If the partition is done at time t–1, the following year strong governance firms exhibit higher accounting conservatism than weak firms (β₃ = 0.18 vs. 0.13) and the difference is statistically significant (p-value = 0.02). The partition at time t basically reproduces the results of Table 6: strong governance firms are more conservative than weak firms. However, when the sample is partitioned at time t+1, the previous year there is no significant difference in the level of accounting conservatism (β₃ = 0.10 vs. 0.09, p-value of the difference = 0.20). The same occurs when the partition is done at time t+2 or t+3. The firms that are strong and weak at time t+2 (t+3), do not exhibit significant differences in conservatism two (three) years before, as shown in Table 9. This first set of results seems to be consistent with the hypothesis that governance influences accounting conservatism and not the other way around.

Notice that the number of observations decreases as we take additional lags or leads because each lag or lead implies losing one year of observations. The previous findings could be driven by the significantly different number of observations in the regressions of each partition (from 1206 at t–3 to 2303 at t, and down to 1290 at t+3). If we repeat the previous test using the same years for all the partitions (from 1995 to 2000), the number of observations in all partitions is almost constant. Unreported results show that the inferences drawn from Table 9 do not change.

We continue exploring the robustness of the previous findings by replacing the dependent variable in equation (2) and using discretionary accruals (DAX) instead of earnings. The third and fourth columns of Table 9 contain the results, which mirror the findings described above. These results are also consistent with the hypothesis that governance influences accounting conservatism and that managers use discretionary accruals to affect the level of conservatism. Finally, when the dependent variable is earnings before discretionary accruals (X^*) we do not observe statistically significant differences in conservatism across governance structures regardless of when the partition is made, as depicted in the last two columns of Table 9.

Inspired by Roychowdhury and Watts (2004), an alternative way of performing the previous tests and avoid the possible arbitrariness of the type of partition chosen, is to use all the information in Totgov^*. In particular, we run regression (7) at time t but measuring Totgov^* at different points in time (from t–3 to t+3):

Dep. Var_t = β₀ + β₁ D_t + β₂ Totgov^*_t+j + β₃ R_t + β₄ D_t Totgov^*_t+j + β₅ R_t Totgov^*_t+j 

                                                                + β₆ D_t R_t + β₇ D_t R_t Totgov^*_t+j + u_t                                            (9)

where the subindex j ranges from –3 to +3. The coefficient of interest is β₇ that captures the differential level of conservatism across governance structures. When the dependent variable is X or DAX, we expect β₇ to be significantly negative if Totgov^* is measured between t–3 and t. This means than when past or current governance is weak (i.e., high values of Totgov^*) firms exhibit less accounting conservatism. However, when governance is measured in future periods (t+1 to t+3) we expect to observe no differences in current accounting conservatism (i.e., β₇ not significantly different from zero). When the dependent variable is X^*, we do not expect to observe differences in conservatism across governance structures regardless of when governance is measured with respect to current earnings and returns. Table 10 shows the results of this analysis, which confirm our predictions and add more support for the hypotheses that governance influences accounting conservatism and that managers use discretionary accruals to capture bad news faster.

Finally, we perform a third test using changes in governance. We select a sub-sample of firms that experience an improvement in governance from time t–1 to time t (i.e., ∆Totgov^* < 0). Then, for the same set of firms, we run regression (2) at time t–1 and at time t and compare the change in β₃. If governance affects accounting conservatism, we expect to observe an increase in the size of β₃. Panel A in Table 11 shows the results of this test. When the dependent variable is X, we observe that β₃ increases from 0.08 to 0.13 and this change is statistically significant (p-value = 0.00). When the dependent variable is DAX, β₃ increases from 0.01 to 0.05 and the change is still significant (p-value = 0.09). However, when the dependent variable is X^*, the change in β₃ is not significantly different from zero (p-value = 0.38). Again, these results are consistent with the hypotheses that governance influences accounting conservatism and that managers use discretionary accruals to accelerate bad news recognition. 

In Panel B of Table 11, we perform the same test but using a subsample of firms that experience a deterioration of governance from t–1 to time t (i.e., ∆Totgov^* > 0). When the dependent variable is X or DAX, the change in β₃ is not significantly different from zero (pvalue = 0.26 and 0.53, respectively). These results are unexpected and do not allow to draw any inference on the direction of causation. A possible interpretation is that the sub-sample of firms with negative changes in Totgov^* had “excess” governance and are simply adjusting their governance structures towards the optimum without changing the level of conservatism.

As a conclusion, the three tests performed in this subsection are consistent with the hypothesized direction of causation from governance to conservatism. Nevertheless, we stop short from concluding that good governance causes more accounting conservatism, but if this were the case we should observe results similar to the ones reported in Tables 9 through 11.

Our findings should not be interpreted as contradicting the findings of Bushman et al. (2004). On the contrary, we interpret our results as complementary to those of Bushman et al. It is plausible that firms with noisier accounting environments beyond the control of management call for enhanced governance structures and that, at the same time, better governance leads to increases in accounting conservatism for those firms with governance structures below the equilibrium level.

6. Conclusions

In this paper we assess the association between accounting conservatism and corporate governance. Accounting conservatism produces earnings that reflect bad news faster than good news. In particular, we investigate whether firms with strong corporate governance exhibit a higher degree of accounting conservatism, than firms with weak governance. Our proxy for the level of accounting conservatism is Basu’s (1997) asymmetric timeliness of earnings measure. 

We assess the quality of corporate governance using a composite measure that incorporates the level of antitakeover protection and level of CEO involvement in the decisions of the board of directors. Our governance proxy incorporates mechanisms of external and internal governance of the firm. It is important to include both as they have a complementary effect: external governance reinforces the effectiveness of internal governance, and vice versa. In addition, we control for the economic determinants of governance and use an idiosyncratic measure of governance quality. Using a large sample of US firms during the period 1992-2003, we find that the asymmetric timeliness of earnings is significantly higher for firms with low antitakeover protection and low CEO involvement in board decisions. This result is robust to controlling for the investment opportunity set, as it has been shown that differences in asymmetric timeliness can be driven by differences in growth opportunities unrelated to conservatism (Roychowdhury and Watts, 2004).

To further investigate the reason why strong governance structures seem to induce more accounting conservatism, we also study the impact of earnings management on the sensitivity of earnings to bad news across governance structures. Using several accruals models, we decompose reported earnings into a discretionary part and a non-discretionary part. We find that the increase in accounting conservatism in strong governance firms is driven by the discretionary component of reported earnings. However, we do not find a significant difference in the sensitivity of unmanaged earnings (defined as earnings before discretionary accruals) to bad news between strong and weak governance firms. This evidence is consistent with strong governance firms using accruals to accelerate the recognition of bad news in earnings. This result also holds after controlling for the investment opportunity set.

Finally, we investigate the direction of causality as our previous findings only document a positive association between governance and accounting conservatism. We find that past governance is associated with current conservatism but not vice versa, and that firms with improvements in governance also exhibit increases in conservatism. We do not dare to conclude that good governance leads to more conservative accounting numbers, but our last set of results provides evidence consistent with this hypothesis. 

                                                 

ENDNOTES

 1  We use the three terms interchangeably throughout the paper.

2

 Gompers et al. (2003) examine 24 provisions: anti-greenmail, blank check preferred stock, business combination laws, bylaw and charter amendment limitations, classified board, compensation plans with change in control provisions, director indemnification contracts, control share cash-out laws, cumulative voting requirements, director’s duties, fair price requirements, golden parachutes, director indemnification, limitations on director liability, pension parachutes, poison pills, secret ballot, executive severance agreements, silver parachutes, special meeting requirements, supermajority requirements, unequal voting rights and limitations on action by written consent.

3

 Like Bertrand and Mullainathan (2001), we use unit weights to construct Totgov following the recommendations of Grice and Harris (1998), who find that unit-weighted composites exhibit better psychometric properties than alternative weighting schemes. 

4  Basu uses the annual stock rate of return measured from 9 months before fiscal year end t to 3 months after fiscal year end t. However, most subsequent studies have used the fiscal year. Measuring returns 3 months after fiscal year end is aimed at giving time to the market to incorporate information in contemporaneous earnings. Using fiscal year returns avoids returns being distorted by new information (different from earnings) coming to the market. Our results are not affected by this choice.

5

 Managers may also manipulate the timing and level of cash flows (e.g. Roychowdhury, 2004, Bushee, 1998; Bartov, 1993), however, due to its low flexibility and high visibility, this is expected to be a residual form of earnings management (Peasnell et al., 2000).

                                                                                                                                                        

6

 TACC may be defined also indirectly as (∆CA - ∆CL - ∆Cash + ∆STDEBT - DEPN), where ∆CA is the change in current assets (Compustat item #4), ∆CL is change in current liabilities (Compustat item #5), ∆Cash is change in cash (Compustat item #1), ∆STDEBT is change in debt in current liabilities (Compustat item #34), and DEPN is depreciation and amortization expense (Compustat item #14). We estimate TACC using this second definition, and recalculate all the models. Our results are not sensitive to the use of the direct or indirect method to define total accruals, however, the existing evidence shows that the use of the indirect method (1) results in different estimations of CFO and TACC, and (2) creates backing-out problems and biases in the estimation of DAX (Drtina and Largay, 1985; Hribar and Collins, 2002; Lim and Lustgarten, 2002; Elgers et al., 2003), thus we only report the results of the direct method.

7                       Our data covers the period 1992-2003. The IRRC data is only available for 1990, 1993, 1995, 1998, 2000, and 2002. Gompers et al. (2003) report that for the majority of firms there is little time-series variation in the index. Taking advantage of this fact, like Cremers and Nair (2005), we align the index values available for 1990 with firm data for 1992, the index values for 1993 with firm data for 1993 and 1994, the index values for 1995 with firm data for 1995, 1996 and 1997, the index values for 1998 with firm data for 1998 and 1999, the index values for 2000 with firm data for 2000 and 2001, and the index values for 2002 with firm data for 2002 and 2003.

8                       This result does not change qualitatively if we partition the sample by Totgov or by the Gompers et al. (2003) antitakeover protection index.

9

 Fama and MacBeth (1973) regressions should be interpreted with caution. Basu (1999) gives a number of reasons against the use of mean annual regressions, related mainly to the parameters not being stationary.

10  We thank Penman and Zhang for providing the computer code to construct the C-Score.

References

Adams, R. B. 2000. What do boards do? Evidence from committee meeting and director compensation. Working paper, Federal Reserve Bank of New York.

Asbaugh, H., D.W. Collins and R. LaFond. 2004. Corporate governance and the cost of equity capital. Working paper, University of Wisconsin and University of Iowa.

Ball, R., S.P. Kothari, and A. Robin. 2000. The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and Economics, 29: 1-51.

Bartov, E. 1993. The timing of asset sales and earnings manipulation. The Accounting Review, 68 (4): 840-855.

Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics, 24: 3-37. 

Basu, S. 1999. Discussion of International differences in the timeliness, conservatism and classification of earnings. Journal of Accounting Research, 37 (Supplement): 89-99.

Beaver, W.H. and S.G. Ryan. 2005. Conditional and unconditional conservatism: concepts and modeling. Review of Accounting Studies, forthcoming.

Becker, C., M. DeFond, J. Jiambalvo, and K.R. Subramanyam. 1998. The effect of audit quality on earnings management. Contemporary Accounting Research, 15: 1-24.

Beekes, W., P.F. Pope, and S. Young. 2004. The link between earnings timeliness, earnings conservatism and board composition: evidence from the UK. Corporate Governance, 12 (1): 47-51.

Bertrand, M. and S. Mullainathan. 2001. Are CEOs rewarded for luck? The ones without principals are. The Quarterly Journal of Economics, 116 (3): 901-932.

Bowen, R.M., S. Rajgopal, and M. Venkatachalam. 2004. Accounting discretion, corporate governance and firm performance. Working paper, University of Washington.

Bushee, B.J. 1998. The influence of institutional investors on myopic R&D investment behaviour. The Accounting Review, 73 (3): 305-333.

Bushman, R., Q. Chen, E. Engel, and A. Smith. 2004. Financial accounting information, organizational complexity and corporate governance systems. Journal of Accounting and Economics, 37: 167-201.

Byrd, J.W. and K.A. Hickman. 1992. Do outside directors monitor managers? Evidence from tender offer bids. Journal of Financial Economics, 32: 195-222.

Core, J. E., R.W. Holthausen, and D.F. Larcker. 1999. Corporate governance, the executive officer compensation, and firm performance. Journal of Financial Economics, 51: 371406.

Cremers, K. J. M. and V.B. Nair. 2005. Governance mechanisms and equity prices. Journal of Finance, forthcoming.

Dechow, P.M. 1994. Accounting earnings and cash flows as measures of firm performance: The role of accounting accruals. Journal of Accounting and Economics, 18 (1): 3-42.

Dechow, P.M., R.G. Sloan, and A.P. Sweeney. 1995. Detecting earnings management. The Accounting Review, 70: 193-225.

Dechow, P.M., R.G. Sloan, and A.P. Sweeney. 1996. Causes and consequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research, 13: 1-36.

Demsetz, H. and K. Lehn. 1985. The structure of corporate ownership: causes and consequences. The Journal of Political Economy, 93 (6): 1155-1178.

Drtina, R., and J. Largay. 1985. Pitfalls in calculating cash flow from operations. The Accounting Review, 60 (2): 314-326.

Elgers, P.T., R.J. Pfeiffer, and S.L. Porter. 2003. Anticipatory income smoothing: a reexamination. Journal of Accounting and Economics, 35 (3): 405-422.

Fama, E. and J. MacBeth. 1973. Risk, return, and equilibrium: empirical tests. Journal of Political Economy, 81: 607-636.

Francis, J. and K. Schipper. 1999. Have financial statements lost their relevance? Journal of Accounting Research, 37 (2): 319-352.

García Lara, J.M., B. García Osma, and A. Mora. 2005. The effect of earnings management on the asymmetric timeliness of earnings. Journal of Business Finance and Accounting, 34 (3-4): 691-726.

Gompers, P., J. Ishii, and A. Metrick. 2003. Corporate governance and equity prices. The Quarterly Journal of Economics, 118 (1): 107-155.

Grice, J.W., and R.J. Harris. 1998. A comparison of regression and loading weights for the computation of factor scores. Multivariate Behavioral Research, 33 (2): 221-247.

Guay, W.R., Kothari, S.P. and Watts, R.L. 1996. A market based evaluation of discretionary accruals models. Journal of Accounting Research, 34 (Supplement): 83-105.

Hermalin, B. E. 2005. Trends in corporate governance. Journal of Finance, forthcoming.

Hermalin, B. E. and M.S. Weisbach. 1988. The determinants of board composition. RAND Journal of Economics, 19 (4): 589-606.

Hermalin, B. E. and M.S. Weisbach. 1998. Endogenously chosen boards of directors and the monitoring of the CEO. American Economic Review, 88: 96-118.

Hermalin, B. E. and M.S. Weisbach. 2003. Boards of directors as an endogenously determined institution: A survey of the economic literature. Economic Policy Review: 726.

Hribar, P., and D.W. Collins. 2002. Errors in estimating accruals: Implications for empirical research. Journal of Accounting Research, 40 (1): 105-134.

Jensen, M.C. 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review, 76: 323-329.

Jensen, M.C. 1993. The modern industrial revolution, exit and the failure of internal control systems. Journal of Finance, 48: 831-880.

Jones, J.J. 1991. Earnings management during import relief investigations. Journal of Accounting Research, 29 (2): 193-228.

Kasznik, R. 1999. On the association between voluntary disclosure and earnings management. Journal of Accounting Research, 37 (1): 57-81.

Klein, A. 2002. Audit committee, board of directors characteristics, and earnings management. Journal of Accounting and Economics, 33 (3): 375-400.

Kothari, S.P., A.J. Leone and C.E. Wasley. 2002. Performance matched discretionary accrual measures. Working Paper, MIT.

Lang, L., R. Stulz, and R. Walkling. 1991. A test of the free cash flow hypothesis: The case of bidder returns. Journal of Financial Economics, 29(2): 315-335.

Lim, S.C., and S. Lustgarten. 2002. Testing for income smoothing using the backing out method: A review of specification issues. Review of Quantitative Finance and Accounting, 19 (3): 273-290.

Peasnell, K.V., P.F. Pope and S. Young. 2000. Accrual management to meet earnings targets: UK evidence pre- and post- Cadbury. British Accounting Review, 32: 415-445.

Peasnell, K.V., P.F. Pope and S. Young. 2001. Board monitoring and earnings management: Do outside directors influence abnormal accruals?. Working paper, Lancaster University.

Penman, S., and X-J. Zhang. 2002. Accounting conservatism, the quality of earnings and stock returns. The Accounting Review, 77 (2): 237-264.

Pope, P.F. and M. Walker. 2003. Ex-ante and ex-post accounting conservatism, asset recognition and asymmetric earnings timeliness. Working paper. Lancaster University and The University of Manchester.

Rogers, W. 1993. Regression standard errors in clustered samples. Stata Technical Bulletin Reprints, 3: 83-94. College Station, TX: Stata Press.

Roychowdhury, S. 2004. Management of earnings through the manipulation of real activities that affect cash flow from operations. Working paper. MIT.

Roychowdhury, S. and R. Watts. 2004. Asymmetric timeliness of earnings, market-to-book and conservatism in financial reporting. Working paper, University of Rochester.

Smith, C. and R. Watts. 1992. The investment opportunity set and corporate financing, dividends, and compensation policies. Journal of Financial Economics, 32: 263-292.

Teoh, S.H., I. Welch, and T.J. Wong. 1998a. Earnings management and the under performance of seasoned equity offerings. Journal of Financial Economics, 50 (1): 63-99.

Teoh, S.H., I. Welch, and T.J. Wong. 1998b. Earnings management and the long-run market performance of initial public offerings. Journal of Finance, 53 (6): 1935-1974.

Vafeas, N. 1999. Board meeting frequency and firm performance. Journal of Financial Economics, 53: 113-143.

Watts, R. 2003a. Conservatism in accounting Part I: Explanations and implications. Accounting Horizons, 17 (3): 207-221.

Watts, R. 2003b. Conservatism in accounting Part II: Evidence and research opportunities. Accounting Horizons, 17 (4): 287-301.

Weisbach, M.S. 1988. Outside directors and CEO turnover. Journal of Financial Economics, 20: 431-460.

Young, S. 1999. Systematic measurement error in the estimation of discretionary accruals: An evaluation of alternative modeling procedures. Journal of Business Finance and Accounting, 26 (7-8): 833-862.