Effect of 2013 Independence Standards on Board-level Governance:

of the 2013 listing standards, there will be no impact for calendar-year reporting firms on their 2013 proxy season. However, prior studies show that, even though firms were not formally required to comply with the listing standards until the effective date, the publication of the proposed recommendations itself probably led many firms to start complying long before the effective date.

For example, Chhaochharia and Grinstein’s (2009) uses year 2003 as the breaking point for examining differences in CEO compensation before and after the listing standards despite the fact that the amended standards were approved on November 4, 2003. This is consistent with the notion that many firms began to change their board structure once the recommendations were put forth by the national exchanges. Following this notion, I choose 2013 as the cut-off and regard 2013~2014 as the years after the new listing standards. Second, I do not consider years 2008~2009 because the variations in CEO compensation and turnovers tend to be substantially high during this financial crisis periods.

I hypothesize that if the 2013 independence standards imposed on the compensation committee affect board-level governance, boards that did not comply with these standards should respond more to the standards than boards that did comply with them. The measure of the level of compliance is whether the firm has a fully independent compensation committee in 2010~2012 (before the new independence standards were announced). The dependent variable of interest is board-level governance (denoted by GOVERNANCE). Following Adams and Ferreira (2009), GOVERNANCE can be captured by CEO compensation (denoted by CEO_COMP) and CEO

turnover (denoted by CEO_TURN). To test my hypothesis, I run the following regression (1) over an unbalanced panel of firm-observations in the years 2010~2014:

)

CEO_COMP = (a) Natural log of CEO’s total compensation or (b) the ratio of CEO’s equity-based compensation to total compensation in year t;

CEO_TURN = A dummy variable that equals one if the CEO leaves the CEO position in the following year (I also define CEO_TURN to be missing in the final year of the sample);

NONCOMPLY = A dummy variable that equals one if the firm did not have a fully independent compensation committee on the board in 2010~2012 and zero otherwise (a director is defined as an independent director if the director was not an employee of the firm during the previous 3 years, if the director does not have family affiliation of the officers of the firm, and if the director does not have any business transactions with the firm);

POST = A dummy variable that equals one if the firm-observation is in the period 2013~2014 and zero otherwise;

SALES = Natural log of firm’s net sales (in $millions) in year t−1;

SEGMENTS = Number the firm’s business segments in year t−1;

ROA = Natural log of one plus net income before extraordinary items and discontinued operations divided by the book value of assets, all measured in year t−1;

RETURN = The firm’s raw stock return for year t−1 net of the CRSP value-weighted index, both compounded continuously;

VOLA = Standard deviation of monthly stock returns from CRSP over the previous five years from t−5 to t−1;

BOARDSIZE = Number of directors on the board in year t−1;

BOARDIND = Percentage of independent directors on the board in year t−1;

AGE = The age of the CEO in year t−1;

HOLDINGS = Percentage of outstanding shares held by the CEO in year t−1;

GENDER = A dummy variable that equals one if the CEO is female and zero otherwise;

CHAIR = A dummy variable that equals one if the CEO is also the chairman in year t−1 and zero otherwise;

TENURE = Natural log of (1 + number of years CEO served in the firm by year t−1);

MEETING = Number of meetings held by the compensation committee in year t−1;

OVERLAP = Percentage of compensation committee members who also sit in the audit committee in year t−1;

experts with work experience as CPAs, CFOs, vice presidents of finance, financial controllers, or any other major accounting positions;

Industry-year fixed effects = The Fama–French (1997) 48-industry dummies interacted with year dummies;

 = Residual term.

Following prior studies (e.g., Chhaochharia and Grinstein 2009; Adams and Ferreira 2009), I use a natural logarithmic transformation of total compensation (including base salary, bonuses, options, restricted stocks, and other compensation) and the ratio of equity-based compensation to total compensation as two measures of CEO_COMP. Because the ratio of equity-based compensation is bounded between zero and one, I use its log transform (i.e., the log odds ratio) as the dependent variable. If the ratio is z, my dependent variable is ln(z/(1-z)+), where  is a very small number I add to ensure I do not attempt to take the logarithm of zero. In the above regression model (1), if the dependent variable is CEO_COMP, coefficient 3 represents the change in each of the two CEO compensation measures of the noncomplying firms in the post-regulation period (years 2013-2014) compared to the CEO compensation of the complying firms. A significantly negative 3 suggests that the 2013 independence standards improve compensation committee independence, leading to reduced CEO compensation. Similarly, if the dependent variable is CEO_TURN, coefficient 3 represents the change in CEO turnover of the noncomplying firms in the post-regulation period compared to the CEO turn of the complying firms. A significantly positive 3 suggests that the 2013 independence standards improve compensation committee independence, leading to higher CEO turnover. Note that I restrict my sample to turnover events that are not classified as turnover due to CEO death or retirement.

I also control for characteristics related to the firm, the board, and the CEO that have been reported in prior studies to have impacts on CEO compensation and turnover. For firm characteristics, I control for firm size (SALES), operation complexity (SEGMENTS), performance (ROA and RETURN), and firm risk (VOLA). I use the natural log of a firm’s net sales (in $millions)

to control for firm size. I also use two measures to control for firm’s performance: the natural log of the gross annual stock return and the natural log of the gross return on assets, which is defined as net income before extraordinary items divided by the book value of assets. Both performance measures are lagged 1 year to avoid measuring the effect of compensation on performance. For board characteristics, I consider board size (BOARDSIZE) and the percentage of independent directors on the board (BOARDIND). For CEO characteristics, I control for age, share ownership, gender, and whether CEO is also the chairman of the board. To control for CEO seniority, I use the natural log of one plus the number of years the CEO has served in the firm. Finally, I control for industry shocks that may affect the supply and demand for CEOs in different industries in different years by interacting industry dummies with year dummies. My industry classification follows that of Fama and French (1997). I also include firm fixed effects to control for any unobservable fixed firm characteristics that can affect compensation (e.g., the complexity of the tasks that the CEO faces) and turnover. In all the regressions I cluster the standard errors at the firm-period level for potential heteroskedasticity. For comparison purposes, I adjust for inflation by converting all compensation variables into 2010 dollars using the Consumer Price Index-All Urban Consumers (CPI-U), produced by the Bureau of Labor Statistics. By construction, there is a one-year lag between GOVERNANCE proxies and all control variables.

In addition to the above firm-specific control variables, I also include three variables that may influence the effectiveness of the compensation committee. This procedure is important because improving compensation committee effectiveness is the main purpose of the 2013 independence standards. First, because there is no prior study that has ever adopted proxies for compensation committee’s effort, I borrow from the audit committee literature (e.g., DeZoort et al. 2002; Engel et al. 2010; Menon and Williams 1994) and use the number of meetings held by the compensation committee in a given year (MEETING) to capture committee’s diligence. More diligent compensation committees are more likely to meet more frequently.

Second, Laux and Laux (2009) argue that due to the conflicting interests between the audit and

(OVERLAP) will tradeoff the costs between providing incentives to the CEOs and preventing CEOs’

earnings management. This tradeoff leads to a preference for offering less incentive compensation contracts to mitigate CEOs’ incentives in managing earnings. Hoitash and Hoitash (2009) provide empirical evidence consistent with the above prediction. Finally, even though the SEC’s 1992 rules and the 2013 independence standards allow firms to hire outside independent compensation advisors or consultants to help the compensation committees determine CEO compensation, the final decisions are made by the members of the compensation committees. Because shareholders prefer the CEOs to be compensated by their performance and CEOs’ operation performance is usually reflected in firms’ financial statements, compensation committee members who have accounting expertise shall be able to better evaluate CEOs’ true performance from accounting numbers and determine the appropriate level of CEO pays. Following prior research, I measure compensation committee’s accounting expertise (EXERTISE) as the percentage of compensation committee members who are accounting experts with work experience as CPAs, CFOs, vice presidents of finance, financial controllers, or any other major accounting positions (e.g., Krishnan and Visvanathan 2008; and Dhaliwal et al. 2010).

3.2 Effect of Equity-based Compensation on Compensation Committee Independence:

Even though the 2103 independence standards intend to enhance the independence of compensation committees, the efficacy of these standards may be enhanced or impaired if firms pay more equity-based compensation to their compensation committee members. On the one hand, corporate governance literature indicates that it is beneficial to have outside directors own stocks to align their interests with the shareholders (e.g., Fama and Jensen 1983; Monks and Minow 2001;

Williamson 1984) and provide incentives for outside directors’ monitoring (e.g., Beasley 1996), resulting in better firm performance (e.g., Fich and Shivdasani 2006; Hanlon et al. 2003). In addition, compensation committee members may have incentives to maintain their reputation due to potential litigation exposure. Fama and Jensen (1983) and Zajac and Westphal (1996) point out that

a primary incentive for outside directors to effectively monitor managers is to enhance outside directors’ reputation. This is consistent with Yermack’s (2004) finding that 40% of total outside director incentives are related to reputation, which basically drives the chance to get and the risk to lose directorships. Further, some research finds that outside directors with a poor track record of monitoring are disciplined by the managerial labor market (e.g., Milgrom and Roberts 1992; Gerety and Lehn 1997). Finally, equity-based compensation could help firms in attracting talented compensation committee members who can determine appropriate CEO compensation. Overall, the above observations underscore the potential economic costs borne by compensation committee members as a result of their ineffective functioning. Therefore, members of compensation committees receiving more equity-based compensation shall face strong incentives to exercise due diligence to mitigate the risk of litigation and the consequential reputation loss.

On the other hand, it is possible that equity-based compensation may impair the effectiveness of compensation committees. This argument is supported by the reasoning that, because stocks and options tie compensation committee members’ wealth to firms’ short-term and long-term financial performance, the objectivity of compensation committee members in determining appropriate CEO compensation level may decrease when their compensation create conflicts that induce them to compromise independence, leading to weak board-level governance. The National Association of Corporate Directors (NACD 2001, 2003) currently promotes the use of equity-based compensation for directors (Archambeault et al. 2008; Magilke et al. 2009) and large U.S. companies have increased the use of stocks and options to compensate their non-executive directors, of which compensation committees are to be formed. However, there is a lack in the literature that examines the association between equity-based compensation and compensation committees’ effectiveness.³ More importantly, even though the 2013 independence standards mandate that all compensation        

3Recent studies have found that equity-based compensation creates incentives for audit committee members to tolerate earnings management by managers since audit committee directors directly benefit from such actions (e.g., Magilke et al. 2009). Empirical studies also show that option compensation is associated with higher likelihood of restatements (e.g., Archambeault et al. 2008) and earnings management (e.g., Bedard et al. 2004). In addition, clients whose audit committees have larger stock ownership are more likely to dismiss their auditors following first-time going concern opinion (Carcello and Neal 2003). Firms demanding less monitoring over financial reporting pay more equity-based compensation to their audit committees (Engel et al. 2010).

requirement when they are newly appointed, but become less independent when they receive equity-based compensation during their tenure periods. Therefore, the answers to the question of whether new regulations are needed to govern how compensation committees should be compensated to ensure their effectiveness over time bear important policy implications to the regulators and national exchanges.

Given these opposite predictions, the net effect of equity-based compensation on compensation committee independence is an empirical issue. This issue becomes even more important under the 2013 independence standards because the NYSE and NASDAQ explicitly require the boards consider “… the sources of compensation of the director or other compensatory fee paid by the company to the director...” in evaluating the independence of compensation committee members.

To test if equity-based compensation enhances or harms compensation committees’ independence, I include the ratio of equity-based compensation to total compensation paid to the compensation committees (denoted by EBC) into regression model (1), leading to the following model (2):

)

EBC = Ratio of equity-based compensation to total compensation paid to the compensation committees in year t−1;

The definitions of all other variables are the same as those in model (1).

If equity-based compensation impairs the independence of compensation committees, coefficient 4 shall be positive when CEO_COMP is the dependent variable (i.e., the CEO will receive larger total compensation and higher portion of equity-based compensation) and be negative when CEO_TURN is the dependent variable (i.e., the CEO is less likely to be dismissed

when his/her performance is poor).

3.3 Data and Sample Selection:

The data source for CEO compensation, tenure, turnover, and chairmanship is the Execucomp database, which has all compensation information about firms that belong to the NYSE or that once belonged to the exchange. I do not consider firms listed at NASDAQ because its 2013 independence standards are different from those of the NYSE in three substantial ways. First, the NASDAQ standards only require firms to have a compensation committee of at least two independent directors. This requirement is not consistent with the definition of the variable NONCOMPLY, which is coded one if all members of a compensation committee are independent.

Second, the NASDAQ standards do not implement the “enhanced independence” requirement that the boards must consider relevant factors to make affirmative determination about the independence of compensation committee members. Finally, the NASDAQ standards maintain the existing exception allowing a listed firm to have inside directors serving on its compensation committee.

The data source for board structure and director information (e.g., size, independence, compensation, memberships, age, gender, holdings, and education and prior work experience) comes from BoardEx. In particular, the database has information about whether the director is independent and about whether the director serves on the compensation and audit committees. My analysis spans the years 2010 to 2014 covering all companies listed at NYSE. To ensure that I do not capture changes in compensation due to firms entering and leaving the samples, and to ensure that the firms are subject to the 2013 independence standards, I include in the analysis only U.S.

firms that existed in these two databases for the entire period and that are members of the NYSE. I will retrieve financial information for each firm from COMPUSTAT. All variables are adjusted for inflation using 2010 as the base year. Firm observations will be deleted if they: (a) do not have complete financial and compensation data and (b) do not pay equity compensation. To control for outlier problem, I winsorize observations that fall in the top and bottom 1 percent of the empirical distribution for both the dependent and independent variables (Bulter et al. 2004; Fan and Wong

Two major compensation components will be examined: stock awards (which include common stock with and without restrictions, deferred stock units, and phantom stock units) and option grants (which include short-term and long-term stock options). The value of stocks will be

determined by multiplying the number of shares awarded by the closing price. Following Brick et al. (2006) and Core et al. (1999), I will compute the value of options using the 25 percent of their exercise price or the closing market price on the annual meeting date if exercise price is not available. I exclude meeting fees because they are often viewed as an opportunity cost of attending a meeting and, thus, are not similar to annual compensation (Adams and Ferreirs 2008).

3.4 Further Analyses:

3.4.1 Low vs. High External Monitoring

Prior research shows that the extent to which the decrease in CEO compensation is related to the existence of other external monitoring mechanisms in the noncomplying firms. I focus on two mechanisms that have been shown to affect CEO compensation. The first is the concentration of institutional holdings. Hartzell and Starks (2003) find that firms with high concentrations of institutional holdings tend to give more efficient compensation schemes to their managers.

Following Dittmar and Mahrt-Smith (2007), I first calculate the sum of all ownership positions greater than 5% held by institutional investors for each firm in each year. These blockholdings, as collected from the 13-F filings by Thomson Financial, can be considered a measure of the extent active large shareholders oversee the management. Since large shareholders have greater incentive to monitor management, firms improve their governance from within by taking policies and procedures to protect their investments in the face of potential agency conflicts (Gompers et al.

2003). I then use the median value of these sums to partition the sample into two groups and code observations above (below) the median as having high (low) monitoring. Model (1) is estimated separately for each group.

The second mechanism is the existence of a nonemployee blockholder on the board. Core et al.

(1999) show that when a blockholder sits on the board of directors, compensation to the manager is lower, consistent with Spatt’s (2006) conjecture that stronger incentives to directors are likely to make them more effective. Following Chhaochharia and Grinstein (2009), I define a director blockholder as a nonemployee director who holds 5% or more of the shares in the firm. I partition the sample into firms with and without a blockholder and re-estimate model (1) separately for each group. I hypothesize that noncomplying firms that have a large concentration of ownership by institutional investors or a nonemployee blockholder on the board should not decrease the firm’s CEO compensation as much as firms that do not have a large concentration of holdings.

3.4.2 Low vs. High Fraud Risk

Firms with a higher risk of fraud are of particular interest to investors, auditors, and regulators.

I examine whether the effect of compensation committee independence on board-level governance varies with firms’ fraud risk. I use the F-score to proxy for fraud risk (Dechow et al. 2011).

DeChow et al.’s (2011) F-score provides an ex ante indicator of a firm’s risk of having material accounting misstatements due to fraud. Specifically, a larger F-score indicates a higher probability of material misstatements. I partition the sample at the median value of the F-score and code observations above (below) the median as high (low) fraud risk and re-estimate model (1) separately for each partition. I hypothesize that noncomplying firms that have larger fraud risk should decrease the firm’s CEO compensation more than firms that have smaller fraud risk.

3.4.3 Low vs. High Managerial Ability

Demerjian et al. (2013) show that, because superior managers are more knowledgeable of their company and business, they will make better judgments and estimates, leading to higher earnings quality. Based on this conclusion, I hypothesize that firms having more capable CEOs will have less incentive to manipulate earnings because they do not need to do so to maintain their earnings

Demerjian et al. (2013) show that, because superior managers are more knowledgeable of their company and business, they will make better judgments and estimates, leading to higher earnings quality. Based on this conclusion, I hypothesize that firms having more capable CEOs will have less incentive to manipulate earnings because they do not need to do so to maintain their earnings