CEO pay, firm risk and performance : evidence from UK-listed companies.

(1)

Master Thesis

CEO Pay, Firm Risk and Performance: Evidence from UK-listed companies

Submitted in partial fulfillment

Of the requirement of the degree of

Master in International Finance

Amsterdam Business School

University of Amsterdam

September, 2013

©Qiujin(Lisa) Bergé-Lu 2013

(2)

CEO Pay, Firm Risk and Performance: Evidence from UK-listed companies

Abstract:

This paper examines the link between CEO compensation, firm performance, and firm risk for a sample of UK-listed companies over the period of 2002–2009. Our key findings can be summarized as follows: First, CEO compensation is positively linked to accounting based firm performance (ROA), firm performance/value (Tobin’s Q), firm size (net sales), and firm value (MV). Interestingly, CEO pay-performance sensitivity is significantly more pronounced before rather than during the recent financial crisis. Our findings also support a positive relationship between firm risk (i.e. standard deviation of ROA) and CEO compensation. Finally we present evidence showing that UK firms react to changes in the firm’s risk environment by increasing in CEO annual compensation, and in particular, the level of stock options awarded to CEOs.

(3)

1. Introduction

The past decade has witnessed numerous financial scandals occurring all around the world, such as Enron, WorldCom, AIG, Lehman Brothers, and Merrill Lynch. Such events have profoundly changed public’s perception of corporate finance and corporate governance. In addition, the magnitude of those events have raised question of whether the level and the structure of CEO compensation have contributed to the corporate scandals and the spread of the recent financial crisis (Faulkender, Kadyrzhanova, Prabhala, and Senbet; 2010).

Prior empirical studies that focus on the link between CEO pay and performance provide inconsistent results. On the one hand, a number of studies provide evidence supporting the view that executive compensation serves as a means of resolving the agency problem between managers and shareholders (Lewellen and Huntsman, 1970; Chambers 2008; Hall and Liebman, 1998; Mcknight 1999, Farmer et al., 2010). Agency problem was generated by the separation of ownership and control in a firm (Berle and Means, 1932, Jensen and Meckling, 1976, Hubbard and Palia, 1995). The downside of such separation is creating information asymmetry and insufficient oversight. In other words, managers may not act in the best interest of the shareholders. Instead, they act in their own best interest. In order to avoid such agency problem, shareholders must provide managers with incentives to stimulate them to take actions that maximize the value of the firm, which at end maximize shareholders’ wealth. Therefore, CEO compensation should be designed in a way that aligns shareholders’ interest with the CEOs. As such, the CEO compensation should be positively related to the firm performance and any increase in CEO pay level should reflect increased demand on CEO skills, leadership competencies, and talents.

On the other hand, opponents express that CEO’s compensation is not sufficiently linked to the long-term corporate performance (Ciscel and Carroll, 1980; Agrawal et al., 1991; Murphy, 1999; Gregg et al., 1993; Conyon et al., 1995). More specifically, it is documented that while CEO pay has risen dramatically, corporate performance and firm value have been at best mediocre.

Murphy (1999) indicates that during the early 1990s, stock options replaced base salaries as the single largest component of compensation (in all sectors except utilities). Stock options offer CEO the right but not the obligation to buy company stocks. The key components of stock

(5)

options are the vesting period and the strike price. Stock options link CEOs compensation directly to the share-price appreciation as the payout from exercising options increases when stock price increases. Apparently, CEO’s incentive is aligned with shareholders’ value. As such, the agency problem between company’s executive and shareholders could diminish as stock options serve them the same goal which is company value maximization. However, this is not always the case. Since the value of options increase with stock-price volatility, executives with options have incentives to engage in riskier investments (Murphy 1999). Hall and Liebman (1998) conclude that since 1980, there has been a large increase in the level of CEO pay driven by a dramatic increase in stock-option grants. As a result, it has led to a large increase in the equity holdings of top executives, and this in turn has led to a dramatic increase in the responsiveness of executive wealth to firm performance. Rocco Huang (2010) summaries the following traits of CEO pay: first, CEO compensation is disconnected from the firm performance; second, stock options are an important component of CEO compensation and it is correlated with more risk-taking.

Instead of studying CEO compensation causing firm risk, De Angelis et al (2013) disclose an alternative explanation by examining whether firm risk affects the change in executive compensation. Their experiment is based on the relaxation of short-selling constrains which was implemented in 2004. They find evidence that relaxation of short-selling constrains exogenously increased downside equity risk (stock price volatility), and companies take downside risk into consideration when designing CEO incentive compensation contracts. Moreover, they find that the firms affected by this exogenous shock include relatively more stock options in the compensation packages for their CEO and other top managers (De Angelis et al, 2013). Namely, firms increase the convexity of compensation payoff in response to an increase in firms’ downside risk.

In this study, we aim to examine empirically if there is a relationship between CEO compensation and firm performance of UK-listed companies over the period of 2002-2009. Moreover, we would like to focus on the strength of the pay-performance relationship and its development during the period 2002-2009. In addition, we would like to examine whether firm risk affects CEO compensation.

In the master thesis three research questions have been defined:

(6)

2. Whether the CEO pay-performance sensitivity weakened/strengthened during the recent financial crisis?

3. Whether firm risk affects CEO compensation?

This study makes several contributions to the literature on CEO compensation. First, existing research on the subject mainly focuses on the US market (De Angelis et al., 2013; Bebchuk, 2012; Bhattacharyya and Purnanandam, 2011; Ekrens et al., 2009; Murphy et al., 1999). Data on actual top pay-setting institutions and structures in the UK are rare; as a result, the CEO pay level and structure are not well understood for British companies (Benito and Conyon, 1999). Second, assembled evidence so far indicates that if a link between pay and performance can be established then its magnitude is quantitatively quite small relative to the effect of company size (Gregg et al., 1993; Conyon, Gregg and Machin, 1995; Conyon and Peck, 1998; Conyon and Benito, 1999). Same view has been shared by Buck et al., (2005), which contended that the link between pay, performance and shareholder return is weak, but the importance of firm size rather than performance has been a dominant feature. Third, it is also practical relevant to conduct the research for the UK. Since UK was a pioneer in introducing an advisory shareholder vote on executive remuneration in 2002 (High Pay Centre, 2013). The vote is supposed to give institutional investors the rights to vote against management-proposed incentive plans. Therefore, it is interesting to examine how CEO pay-performance sensitivity has developed over time. Furthermore, there are a number of prior studies which investigate the effect between firm risk and CEO compensation suggest that CEO compensation structure induces excessive risk taking, whereas there are only a limited number of studies available which indicate that firm risk affects CEO compensation (De Angelis et al., 2013). In addition, this study also looks at the crisis period and compares pay-performance sensitivity before and after crisis, which can be an added value to the existing literature.

The results of this study reveal several interesting findings. First, the findings suggest that CEO compensation is positively linked to firm size (net sales and firm market value) and firm performance/value (Tobin’s Q). Second, the findings suggest that CEOs annual and direct compensation is positively linked to accounting based firm performance (ROA). Third, if accounting based firm performance explains CEO compensation, we would expect that the two elements would move together all the time. However, the study indicates that those two elements have stopped move together during the crisis. In addition, the study finds evidence which support

(7)

the view from Di Angelis, Grullon, and Michenaud that firms’ risk (standard deviation of ROA) affects CEO compensation.

The remainder of this paper is structured as follows. Section 2 provides an overview of related literature and research motivation. Hypothesis and methodology is outlined in section 3. Section 4 presents data analysis and findings. Section 5 summarizes the main conclusions of the findings and explains the limitations of this study.

2. Related Literature and Motivation

An important theoretical perspective on the design of management incentives is provided by the concept of agency theory, which focuses on conflicts of interest and incentives among different corporate stakeholders, notably between management and its shareholders (Faulkender et al., 2010). Based upon agency theory, the goal of executive compensation is designed to increase alignment between managers and shareholders’ interests. Namely, executive compensation should be linked to the firm’s performance, which is so-called pay-for-performance. In other words, cash compensation should be structured to provide big rewards for outstanding performance and meaningful penalties for poor performance (Jensen and Murphy, 1990; Rost and Osterloh, 2009). However, both the earlier studies and the most recent studies of CEO pay-for-performance have revealed controversial results (Murphy et al., 1988; Jensen and Murphy 1990; Hall and Liebman, 1998; Gregg et al., 1993; Conyon et al., 1995; Mcknight, 1999; Balafas and Florackis, 2013). Moreover, some evidences have been found that CEO compensation is more dominated by firm size instead of firm performance (Bruce et al., 2005; Schaefer, 1998). Furthermore, a number of studies provide striking results that pay-performance sensitivity has deteriorated over time (Jensen and Murphy, 1990) and it is inversely correlated to the firm size (Schaefer, 1998). However, such inverse correlation does not applicable for all the industry sectors. Hubbard and Palia (1995) disclosed that the level of CEO compensation is positively related to the size of the bank according to the data gathered from US banking industries. In addition, there is widespread concern that remuneration policies may have been a contributory factor to the market crisis (FSA 2009), though it is difficult to prove a direct cause link. In the following part of literature review we will expand discussions over executive compensation and firm performance, pay-performance sensitivity, and executive compensation with firm risk. At the end of this section we will explain why UK is an interesting country to study.

(8)

2.1 Executive compensation and firm performance

Prior studies on executive pay-for-performance relationship become more complicated over time, as the number and the variety of variables included in the models increase; such as company nature (e.g. industry sector, cross-country), performance indices, ownership structure, demographics on executives and directors (e.g. age, function, nationality), and time effects (Zhou et al., 2011). Until now, there is no consensus have been reached about the relationship between executive pay and firm’s performance.

Earlier studies by Lewellen and Huntsman (1970), along with other authors argue that executive earning level was closely tied to the corporation financial performance through ownership by, or profit-related deferred income payment to, the CEO. They find that the CEO’s yearly income can be traced to property income which reflected that managers’ interest fully aligned with owners’ interest. They measure firm’s performance by its sales, assets, profits, and rate of return, and they find that CEO compensation (salary and bonus) served as an excellent proxy for total remuneration (Chambers, G.L., 2008). Hall and Liebman (1998) support such finding; they conclude that since 1980, there has been a large increase in the level of executive pay driven by a dramatic increase in stock-option grants. As a result, it has led to a large increase in the equity holdings of top executives, and this in turn has led to a dramatic increase in the responsiveness of executive wealth to firm performance. Furthermore, a first empirical study of top executive pay in the UK conducted by Mcknight (1999) presents the similar finding, which suggests that a pronounced link does exist between performances and pay over both the short and long term. In addition, using a panel of 204 large UK companies and data from FTSE-350 over period 2003 – 2007, Farmer et al., (2010) find that performance-realized short-term pay is determined relative to FTSE-350 market performance and performance-realized long-term pay is determined relative to FTSE-350 sector performance.

In contrast to the finding that there is a positive relation between CEO pay and firm performance proposed by above mentioned authors, Murphy et al. (1988) summarized Lawler’s earlier six separate studies of the relationship between pay and performance as follows: “the studies suggest that many business organizations do not do a very good job of tying pay to performance. It is surprising in particular that pay does not seem to be related to performance at the managerial level”. Moreover, Murphy pointed out that the Medoff and Abraham evidence seems to be indicative of general performance measurement and compensation systems, and the forces

(9)

responsible for these practices is unclear. Many studies focused on UK suggest that in large UK companies the relationship between directors’ salaries and a firm’s performance was weak (Gregg et al., 1993, Conyon et al., 1995). A recent study conducted by Balafas and Florackis (2013) find that CEO incentive pay is negatively associated with both short-term and medium term subsequent returns. Their study is based on a sample of 1787 UK listed companies over the period 1998-2010 and results refer to firms that are in the bottom (portfolio 10%) and top (portfolio 90%) deciles of CEO compensation. They find that firms in the bottom deciles rank earn a statistically significant CAPM alpha of 14.31% p.a., while firms in the top deciles earn an insignificant CAPM alpha of 2.6% p.a. in the three months after portfolio construction. In addition, they find one year after sorting firms on the basis of their incentive pay, firms with the lowest pay earn average risk-adjusted returns between 9.78% and 10.68% p.a., depending on the asset pricing model used. The firms with high pay earn much lower return between 2.48% and 3.79% p.a..

Furthermore, there are some researches suggest the CEOs are rewarded for luck in stead of good performance. For example, Gabaix and Landier (2006) conclude that the six-fold increase of CEO pay between 1980 and 2003 can be fully attributed to the six-fold increase in market capitalization of large US companies during that period; Bertrand and Mullainathan (2001) find that CEOs are rewarded for luck is as large as pay for performance, and pay for luck also appears on the most discretionary components of compensation, salary and bonus. Pay for luck is mainly attributed by option grants, which contain a gift component because even if the CEO does nothing they have value from the intrinsic volatility of the stock according to Black-Scholes model.

In addition, there are some studies find that CEO pay is related to firm size. Bruce et al. (2005) suggested that empirical studies regarding executive pay in the UK have produced mixed results, but the importance of firm size, rather than performance, has been a dominate feature which is in line with the US literature. Another empirical analysis performed by Schaefer (1998) for large American firms between 1991 and 1995 provides evidence that marginal return to executive effort varies with firm size. More interesting finding from Schaefer (1998) is that CEO pay-performance sensitivity decreases with firm size arise.

(10)

Overall, studies of executive compensation focused on whether pay was more closely tied to company performance, the answer has not reached consensus (Ciscel and Carroll, 1980; Agrawal et al., 1991; Murphy, 1999).

2.2 Pay-performance Sensitivity

An earlier study concerning the relationship between executive compensation and firm performance from Jensen and Murphy (1990) found that annual changes in executive compensation do not reflect changes in corporate performance. Using sample of 2,505 CEO compensation data in 1,400 public companies from 1974 through 1988, Jensen and Murphy (1990) estimate coefficients from the following least-squares regression:

• (CEO salary + bonus) t = • + • * • (shareholder wealth) t (1)

Their estimations as follows: first, for each $1,000 increase in shareholder wealth corresponds to an increase in this year’s and next year’s salary + bonus of about 2 cents! Second, CEO’s dismissal related wealth consequences for a $1,000 shareholder’s loss (equivalent to negative 50% of net-of-market returns for two consecutive years) is about 30 cents. Third, the value of CEO’s stock goes up by $2.50 for a $1,000 increase in shareholder wealth. What they have found more striking is that CEO compensation is getting worse rather than better due to the fact that CEO stock ownership for large public company was ten times greater in 1930s than in the 1980s (Jensen and Murphy, 1990). Overall, their study concludes that the degree of pay-for-performance sensitivity for cash compensation does not create adequate incentives for executives to maximize corporate value (Jensen and Murphy, 1990).

In contrast to Jensen and Murphy’s findings, Hall and Liebman (1998) find that CEO wealthy is strongly affected by firm performance by looking at the responsiveness of salary and bonus and option grants to firm performance. Thus, in their model they use ln (salary and bonus) and ln (direct compensation) as dependent variables and firm return in current year, firm return in previous year, S&P 500 return in current year, S&P 500 return in previous year as independent variables. The regression model they used in the study is like this:

ln (salary + bonus) = •0 + •1*firm return current year + •1*firm return previous year (2)

Their findings can be summarized as follows: first, there is a strong relationship between firm performance and CEO compensation. Second, they suggest that in improving firm performance from median (50thpercentile) to 70thpercentile, the CEO’s compensation (with stock) increases by 70.8%. Third, they also find that as firm performance declines from 50th to 30th percentile,

(11)

CEO’s compensation declines by 50.7%. In addition, they also find that changes in stock value and stock options completely outweigh changes in salary plus bonus. Namely, they indicate that salary plus bonus are not sensitive to performance. Similar view also can be found from the most recent study of Balafas and Florackis (2013). They select a sample of 1787 UK listed companies over the period 1998 - 2010 and they find that in contrast to incentive pay, cash pay is not found to affect future shareholder returns in a statistically significant way.

The conclusions of Hall and Liebman do contradict the claim made by Jensen and Murphy which they find that CEO contracts are inefficient because of the lower pay-performance sensitivity. Though, Jensen and Murphy’s study is quite influential, the econometric specification employed in their model is questioned to be inappropriate (Rosen, 1982). Schaefer (1998) noted that Jensen and Murphy specification implicitly assumes that pay-performance sensitivity is independent of firm size, while his research discovered that pay-performance sensitivity appears to be approximately inversely proportional to the square root of firm size where size is measured by either market capitalization or assets. In Schaefer’s study, he assumes the shareholders’ objective is to maximize the firm’s market value. If the firm’s value is a signal of executive effort, then one way to align the interests of CEOs and shareholders is to pay the CEO a fixed wage plus a share of the value of the firm (Schaefer, 1998),

• WAGE = •+ • * • Value (3)

Basically, Schaefer follows pay-performance measure as proposed by Jensen and Murphy (1990). In Schaefer’s model, he uses change in salary plus bonus and change in CEO wealth as dependent variables, and change in shareholder’s value as independent variable. Schaefer’s model produces a pay-performance sensitivity of 0.0125, which means a change in CEO wealth of $12.50 for each $1,000 change in shareholder’s wealth. However, this estimate is not significantly different from zero at the 5% level (Schaefer, 1998). In addition to Jensen and Murphy’s method, Schaefer also verifies that pay-performance sensitivity depends on size by partitioning the sample based on market value and assets. He performs one-tailed tests of the hypothesis that pay-performance sensitivity for the set of larger firms is less than or equal to that of the smaller firms. Using salary plus bonus as a dependent variable, then Schaefer’s result is consistent with Jensen and Murphy’s finding – that pay-performance sensitivity appears to be inversely related to firm size.

(12)

Besides the researches of Hall and Liebman, Schaefer, there are other researches found other views than the view from Jensen and Murphy. For example, Boschen and Smith (1995) estimated the dynamic response of executive compensation to a firm performance over a long time series (1948 – 1990). Their result shows that the contemporaneous response of compensation to firm performance is less significant. However, compensation responses in its subsequent periods are significant, which give rise to a cumulative response that is much larger than the contemporaneous response. Moreover, they find that a 10% increase in real returns is corresponded with a contemporaneous compensation response of only 0.3%-0.5%; however, a cumulative response is 3% - 5%. Thus, their findings reveal that the cumulative response of pay to performance is approximately 10 times of the contemporaneous response. Furthermore, Boschen and Smith describe that the compensation response pattern is more “hump-shaped” pattern – large significant responses is seen 4 years after the small contemporaneous response. In other words, on average, the most important time horizon with respect to compensation encompasses the 5 years following a performance event (Boschen and Smith, 1995). In addition, Boschen and Smith investigate changes in the pay-performance sensitivity over the long horizons (1948 – 1990). First, they find the cumulative compensation response has strengthened over this time period. Second, they find that the cumulative response of cash compensation to a 10% change in real firm return is 0.5% in 1950s versus 5.2% in the 1980s. Third, some evidence shows that the contemporaneous response of total compensation has weakened considerably in the 1970s and 1980s comparing to the years earlier. In sum, their result indicates that the contemporaneous-only pay-performance sensitivity is an inadequate characterization and the interaction of pay with firm performance over several years should be considered in order to capture the complete pay-performance relationship (Boschen and Smith, 1995).

To sum up, the understanding of the relationship and the response between firm performance and executive compensation has improved substantially since Jensen and Murphy (1990).

2.3 Executive compensation and firm risk

A number of studies have provided a possible explanation of increases in CEO compensation which is driven by excessive risk taking. It is in the interest of CEOs to boost the firm’s short-term earnings by taking on greater levels of systematic risk in a booming market as their compensation depended heavily on EPS (Bhattacharyya and Purnanandam, 2011). Rocco Huang (2010) suggests that CEO compensation affects corporate decision-making, including the risk of

(13)

the firm’s operating and financial decisions and the firm’s accounting policy. In addition, Rocco Huang (2010) summaries the following traits of CEO pay: first, CEO compensation is disconnected with CEO performance; second, stock options are an important component of CEO compensation and it is correlated with more risk-taking. Similar views are shared by other critics who argue that CEOs have incentives to take excessive risk as they are compensated exclusively with equity-like instruments, such as stock and options (Murphy, 1999; Edmans, 2010; Bolton et al. 2011). Indeed, the value of the stock for the levered firm is like the value of a call option and is increasing in the volatility (risk) of the assets held by the firm (Bolton et al. 2011). If a risk project pays off, the shareholders capture the gains, but if things go wrong their loss are capped by limited liability (Edmans, 2010). Bebchuk and Spamann (2010) use a simple example to explain the basic argument: suppose a bank has $100 of assets which financed by $90 of deposits and $10 of capital, of which $4 are debt and $6 are equity; the bank’s equity is in turn held by a bank holding company, which is financed by $2 of debt and $4 of equity and has no other assets; and the bank manager is compensated with some shares in the bank holding company. On the downside, the manager’s wealth is protected by limited liability from the consequences of any losses beyond $4. In contrast, the benefits to the managers of upside gains are unlimited. More specifically, Bebchuk and Spamann (2010) argue that the incentives are even more skewed if the managers own options on its stock instead of stock in the holding company. To simplify the reasoning Bebchuk and Spamann extend the previous example. Let’s suppose a bank with $100 in assets financed by $90 of deposits, $2 of other debt, and $8 of equity. The bank managers do not own company stocks; rather they own options on shares with a strike price corresponding to the current market price of the shares; and they are considering whether to take the risky strategy which would result in a 50 percent of probability of decreasing assets value by $20 or a 50 percent of probability of increasing assets value by $X. The options will be valuable if the market valuation of equity is beyond the strike price, but will be worthless otherwise. Due to the fact of information asymmetry, the current market price does not reflect the risk which is associated with the company’s assets. Thus, the market value of the bank’s equity remains as $8 and assumes $8 is also the strike price. The consequence is severer if the bank managers are in favor of taking excessive risk since they will gain a positive payoffs for any positive value of $X. Without take such risky strategy, their options become worthless. Though, taking risky strategy could be still worthless if strategy fails, managers gamble for the 50 percent of chance that

(14)

strategy succeeds (Bebchuk and Spamann, 2010). Overall, the example illustrated above suggests that to some extent, the options are insulated from losses the common shareholders have to bear due to asset value declines. Thus, options encourage even more risk-taking (Bebchuk and Spamann, 2010; Gormley et. al, 2012).

Wharton (2010) argues that the recent financial crisis is primarily trigged by bad bets in the financial sector, which has brought the reality of corporate failure to the fore. Financial Services Authority (FSA 2009) concluded that “although it is hard to prove a direct cause link, there is widespread concern that remuneration policies may have been a contributory factor to the market crisis. The policies in common use during the period leading up to the crisis, mainly but not exclusively in investment banking, tended to reward short-term revenue and profit target”. In response to those discussions, Walker (2009) provided several recommendations regarding CEO remuneration in his review of corporate governance in UK banks and other financial industry entities. Those recommendations mainly reflect alignment of compensation and risk taking, transparency of the process and levels of executive pay; deferral of incentive payments; and performance criteria related to long term profitability (Gregg et al., 2011).

There are some other researches looked at executive’s pay-performance in relation to firm risks, such as credit risk and downgrade risk. The first empirical research to focus on CEO compensation and realized credit risk was done by Moody’s in 2005. For each of the three major compensation components – salary, bonus, and stock option awards – they derive estimates of “unexplained” compensation as pay that deviates substantially from expected pay based on firm size, past performance, and other variables. Then, they relate these measures of unexplained compensation to the risk of default and large rating downgrades between 1993 and 2003 (Moody’s 2005). Moody’s study focus exclusively on non-financial firms in the US with senior unsecured bond ratings of B3 or higher from 1993 and 2003. One of the interesting results from Moody’s study occurs in the top 10% unexplained bonus and unexplained options quintile. Firms with high unexplained bonuses and high unexplained option grants experienced dramatically higher default rates and dramatically higher downgrade rates than did the middle 70% of the distribution. Another interesting result presented by Moody’s study is the firms in the upper tail of unexplained compensation distribution experienced operating income growth of 3% in the previous year and 13.1% in the current year. A superficial analysis would not likely flag these companies as being in trouble and yet their default rates were between two and three times

(15)

higher than the middle ranked firms. Downgrade rates were almost two times higher (Moody’s, 2005). In sum, Moody’s study suggests that large, positive, unexplained bonus and option awards are predictive of both default and large rating downgrades.

In response to the public debate that banking bonuses are excessive and banking-bonuses cause or contribute to the financial crisis, Fahlenbrach and Stulz conducted a banking industry research in 2009. They use compensation data from S&P Exeucomp database in fiscal year 2006, accounting data from Compustat, additional banking data from Compustat Bank, insider trading data from Thomson Financial, and stock return data from CRSP. The filter sample of their study contains 98 banks which excludes pure brokerage houses. Their analysis state that Bank CEOs incentives cannot be blamed for the credit crisis or for the performance of banks during that crisis as there is no evidence that banks with CEOs whose incentives were less well aligned with the interests of their shareholders performed worse during the crisis. However, they find that performance of banks does vary in the cross-section. An evidence has been found that banks where CEOs had better incentives in terms of the dollar value of their stake in their bank performed significantly worse than banks where CEOs had poorer incentives. They do not find that stock options have adverse impact on bank performance during the crisis (Fahlenbrach and Stulz, 2009). In addition, Fahlenbrach and Stulz provide possible explanation regarding bank CEOs with better incentives performed worse during the crisis than the bank CEOs with low incentives as following: bank CEOs with better incentives do act in the interest of their shareholders in order to maximize their wealth by taking risks initially, however these risks had unexpected poor results. The supporting evidence for such possible explanation is backed by Fahlenbrach and Stulz‘s examination of the wealth consequence of the crisis for Bank CEOs. If CEOs took risks that they knew were not in the interests of their shareholders, they would have sold shares ahead of the crisis. But, this did not happen. As a matter of fact, CEO holdings of shares on net increased (Fahlenbrach and Stulz, 2009). Unsurprisingly, CEOs made large losses both on their holdings of shares and options during the crisis.

A similar study was carried out by Murphy in 2012 to explore the banking bonus culture and its role in inducing risk-taking. Murphy argued that the attacks on banking bonuses are driven primarily by anger, jealousy and envy, and not by evidence that the bonuses are set in a non-competitive market. Murphy select companies from S&P 500 firms and divide the sample into three sectors by their SIC codes: Broker-dealers (SIC between 6200-6212), Banks (SIC between

(16)

6000-6199), and Industrials (SIC below 6000 and above 7000), excluding utilities (SIC between 4900-4999). The compensation data used in Murphy’s study is from ExecuComp between period 1992 and 2010. The most focus of Murphy’s study has been on the top executives instead of lower-level traders and managers. The results of his study can be summarized as following: (1) first, actually the pay level, equity incentives, and equity ownership in traditional banks are more similar to that in industrial firms than to that in broker-dealer firms; (2) second, until the market collapse during the financial crisis, equity incentives and equity ownership for broker-dealer executives were substantially higher than their counterparts in banking and industry; (3) third, during the financial crisis, the realized pay, equity incentives, and the value of equity ownership plummeted across all sectors, the decline was especially pronounced for executives in broker-dealer firms; (4) fourth, after crisis, the realized pay, equity incentives, and the value of equity ownership for executives in large broker-dealer firms has largely converged to general industry practices; (5) fifth, bonuses for banking executives are effectively linear with small salaries, low bonus thresholds, and no caps. A major part of the bonuses are paid in the form of unvested stock or un-exercisable stock options, which reduces the value of bonuses based on subsequent performance effectively. In sum, Murphy’s study suggests that the incentive plan for banking executives focus on long term value creation rather than short term gains. Moreover, financial services firms provide significant penalties for failure in their cash bonus plans by keeping salaries below competitive market levels, so that earning a zero bonus represents a penalty. As such, those plans reduce incentives for risk taking rather than increase excessive risk taking (Murphy, 2012).

In sum, contradictive views regarding CEO compensation toward firm performance and firm risk has been emerged in the prior studies. The allegation that CEOs compensations are inconsistent with firms’ performance and CEOs compensation are insufficiently linked to firm risk which causes economic crisis remains debatable.

2.4 The UK institutional setting

The UK was a pioneer in introducing an advisory shareholder vote on executive remuneration in 2002 (High Pay Centre, 2013). Ferri and Maber (2012) examine the effects of “say on pay” legislation. They find a positive market reaction to the announcement of say on pay regulation for firms with controversial CEO pay practices which specifically reflect weak penalties for poor performance. Furthermore, they find that firms respond to high voting dissent by removing

(17)

controversial provisions criticized as rewards for failure, such as long notice periods and retesting provisions for option grants. In addition, they find the sensitivity of CEO pay to poor performance has increase significantly, particularly among firms that experience high dissent at the first vote and firms with excess CEO pay before the regulation (Ferri and Maber, 2012). Overall, Ferri and Maber support shareholders’ view that “say on pay” is a value enhancing monitoring mechanism. However, a recent study conducted by Conyon et al., (2011) find that after controlling for the risk premium, median risk-adjusted pay for U.S. CEOs is not consistently higher than that for U.K. CEOs. Specially, they find that risk-adjusted pay to be higher for U.S. CEOs in 1997, but higher for U.K. CEOs in 2003. Likewise, Fernandes et al., (2012) present that after controlling for sales, industry, and firm, ownership, and board characteristics, the estimated 2006 CEO pay in the UK is the highest among 14 countries according to mandatory pay disclosures. Furthermore, some recent researches on UK listed firms provide evidences of weak relationship between executive pay and company performance for both short term and long term (Balafas and Florackis, 2013; Ian Tonks, 2012). In particular, Ian Tonks (2012) reports that firm size has a larger effect on executive pay than firm return: a 10% increase in firm size measured by total assets increases CEO pay by 2% (this translates into a £11,815 increase in CEO pay at the median level of £543,200), while a 10% additional increase in company share price performance leads to a 0.68% increase in the CEO pay (this translates into a £3,726 increase in CEO pay at the median level of £543,200). It seems that “say on pay” alone is ineffective due to a matter of fact that voting is controlled by wealthy investment fund managers who also benefit from a culture of high pay (High Pay Centre, May 2013). According to High Pay Centre report 2012, in the experience of “say on pay” to date, in the UK there are only two FTSE 100 companies lost the vote on their remuneration report in 2011. These facts suggest that the introduction of “say on pay” does has intention to enhance shareholders’ value creation, however, complementary measures are needed to fulfill public’s hope of more proportionate executive pay.

The main purpose of this study is to investigate how pay-performance sensitivities of UK listed companies have been developed over time and how CEO compensation is related to firm risk-taking. If CEO compensation is to be blamed for causing global economic crisis, then we would expect to see that executive compensation and firm performance do not move in the same direction during the crisis. If evidences have been found that CEO compensation is positively

(18)

related to firm risk taking, then, it is necessary to change the structure of executive compensation and add additional pay constraints based on the study outcome, such as pay ration, pay cap for bailed-out banks, and employee “say on pay” which have been adopted by EU countries (France, Germany, Switzerland) and US (High Pay Centre report, 2013).

3. Hypothesis

3.1 Hypothesis development

Agency theory

Agency problem was generated by the separation of ownership and control in a firm (Berle and Means, 1932, Jensen and Meckling, 1976, Hubbard and Palia, 1994). The downside of such separation is creating information asymmetry and insufficient oversight. In other words, managers may not act in the best interest of the shareholders. Instead, they act in their own best interest. For example, managers would like to take risk projects. Managers capture huge gains if the projects go well, if not, then, the shareholders and debt holders have to bear the losses. In this kind of situation, the behavior of managers is called high risk tolerance. The alternative situation is called high risk aversion. As we all know that in general, the shareholders own a well diversified portfolio with different assets classes and they prefer to take risky investment decisions. However, managers have limited ownership within the company and their wealth is not diversified, thus they are risk averse.

According to agency theory, CEO compensation plans should be designed to resolve the conflict of interest between shareholders and management in order to maximize shareholders’ wealth and reduce agency costs. Thus, if CEO compensation package reduces agency costs, then it should result in an increase in shareholders’ wealth. CEO compensation-governed decisions should lead to firm performance improvement. Moreover, a rise in CEO compensation should reflect on the increased demands for CEO managerial and entrepreneurial skills as the firms grow in size.

Managerial power theory

An alternative theory is called managerial power theory, which means that CEOs have more power in many ways to influence the board and shareholders to obtain attractive pay packages which do not reflect on firm value maximization. This view is exemplified recently by Bebchuk and Fried. They state that such excess pay constitutes and economic rent, an amount greater than necessary to get the CEO to work in the firm. In addition, Bebchuk and Fried point out that

(19)

meanwhile CEOs have to bear constrains such as facing reputation loss and embarrassment if indeed they are caught for rent extraction (Bebchuk and Fried, 2003; Conyon, 2006). Bebchuk and Fried also note that, at end the compensation package should be pushed toward the directions of shareholders value creation due to market influence and directions favorable for managers due to managerial influences (Bebchuk and Fried, 2003; Bootsma, 2010).

To summarize the theatrical discussions before, agency theory suggests that performance related executive compensation can be used as a tool to resolve the agency conflict between shareholders and management as it aligns the interests from both sides. However, managerial power theory point that management have power to influence board and compensation committee to agree on the excessive compensation which is not aligned with shareholders interest and not linked to firm performance. The controversial views have been reflected in the current debate in the UK that CEOs compensation is inadequately tied to corporate performance (Gregg et al., 1993; Conyon et al., 1995; Conyon and Benito, 1999). Sahlman (2009) stated that the macroeconomic problems were caused by the micro-economic decisions with which executives put their stakeholders at risk. Thus, the first objective of this study is exploring whether there is a positive relationship between pay and firm performance and I would like to propose the following hypothesis:

A positive relationship exists between CEO compensation and firm performance in UK (H1)

Most recent study conducted by Fernandes et al (2010) report that in the UK, salary and pensions constituted 51% of the total compensation package in the year 2006, with equity incentives comprising 30% and bonuses comprising 19%, while Conyon and Murphy (2000) document that in the 1997 fiscal year UK CEOs’ total compensation was made up with 64% base salary and pensions, 19% equity-based incentive plans, and 18% bonuses. This indicates that over the period 1997-2006 there has been a significant decrease in the base salaries and a significant increase in the equity-based incentive pay for executive (Gregg et al., 2011). By effectively granting the CEO an ownership stake in the firm, equity-based compensation creates incentives to take actions that benefit shareholders (Frydman and Jenter, 2010). Thus, it seems that change in compensation structure reduces agency conflicts. However, this assumption is not complete true. It has been now well recognized that equity-based compensations have provided executives

(20)

with incentives to focus excessively on short-term results by taking excessive risk at the expenses of shareholders (Bebchuk and Spamann, 2010). The form of equity-based pay can be regarded as the value of a call option and is increasing in the volatility (riskiness) of the assets held by the firm (Bolton et al., 2011). Edmans (2010) points that equity compensation is a “one-way bet” for managers at companies close to bankruptcy – if the risk pays off, the value of equity shoots up, otherwise zero. An article published by Wharton in 2010 states that the recent financial crisis, triggered primarily by bed bets in the financial sector, has brought the reality of corporate failure to the fore, adding momentum to the idea that executive compensation should be tied more closely to corporate debt rather than equity (Wharton, 2010). We argue that if CEO pay structure were to blame for the crisis, we would expect to find evidence that CEO pay and firm performance those two elements do not move together in the same direction during the crisis. In this context, the second research hypothesis can be formulated as follow:

Pay-performance sensitivity becomes stronger during the recent financial crisis. (H2)

In addition, some recent studies disclose that CEO compensation structures are shifted from direct compensation (cash and bonus) to equity based compensation over past years (Murphy, 1999; Hall and Liebman, 1998; Fernandes et al., 2010; Conyon and Murphy, 2000; Gregg et al., 2011). Equity based compensation has been considered as a mechanism to tied the interests of managers to the shareholders (Jensen and Meckling, 1976, Angie Low, 2006). Conyon (2006) suggests that the standard agency model predicts greater expected compensation when incentives are greater. This increase is required to compensate the CEO for the imposition of greater risk and the increased effort induced by higher incentives. The rationale can be explained as follows. In risky companies, where there is a high volatility in profitability and stock price, CEOs are more likely to be paid in options or LTIPs (Long term incentive plans). This may due to convexity of compensation payoff and the power of CEOs to influence their own compensation package through being involved on who sits on the compensation committee. This rational leads to the following testable hypothesis:

(21)

3.2Data and Methodology

3.2.1 Data

The research uses data from two different sources that cover the period 2002-2009. The first source is BoardEx, which provides on CEOs compensation of UK listed companies, corporate board structure, several other board, director characteristics, and director’s compensations. BoardEx data is matched at a company level with accounting and market data obtained from Thomson DataStream with selection criteria LFTALLSH. These data are supplemented by information on market value, net sales, return on assets, return on equity, leverage, and industry type. Starting with 1567 companies extracted from BoardEx and 3774 companies extracted from Thomson DataStream, after merging CEO compensation data with firm performance data by ISIN code, only 321 observations where information on the individual CEO payment and firm performance were complete are included in the data sample1.

3.2.2 Variables

This section outlines the dependent, explanatory variables, and control variables used in the empirical models. Detailed description for variable definitions and codes are provided in Table 1. The dependent variables used in this study are “total annual compensation”, “total direct compensation”, and “total equity compensation” respectively. “Total annual compensation” is the sum of salary, annual bonus, and valuation for stock option, performance plans, time-vesting options, and restricted stock. The sum of salary and annual bonus is regarded as “total direct compensation”. Stock options are treated as total equity compensation.

Explanatory variables selected for measuring the relationship between payment and performance as follows: net sales, firm value/performance, return on assets (ROA), return on equity (ROE), market value of the firm, standard deviation of ROA, and standard deviation of ROE.

Net Sales: Balsam et al. (2011) and Gill et al. (2009) use logarithm of sales as a proxy for firm

size. Same method has been applied in this study.

1_{Among 321 observations, a couple of them have missing values in “Total direct compensation”, “Total equity}

compensation”, and “Net sales”. In order to apply logarithm of those variables into the regression model, I

transformed those “zeros” into positive numbers by adding a start value. For instance, variable “net sales” has a start value of 10, so I add 10 into each of the observation. As such, if the original observations have “Net sales” of zero, after transformation, their “Net sales” will show as 10.

(22)

Firm Value/Performance: the majority of previous studies uses Tobin’s Q2as a proxy for firm performance/value (see, for example, Morck et al., 1988; McConnell and Servaes, 1990; Short and Keasey, 1999, Davies et al., 2005; Florackis et al., 2013). In this study, Tobin’s Q is measured as the ratio of the book value of assets minus the book value of equity plus the market value of equity to the book value of assets.

Return on Assets: prior studies have used ROA as a proxy for company performance indicator,

such as Buck et al (2008) and Balsam et al. (2011). For this study, I use ROA as well as a performance indicator.

Return on Equity: alternatively, I also utilize ROE as a proxy of firm performance in the study as

ROE indicates basically how effectively managers use shareholders money. Using ROE as independent variable to explain CEO compensation can be found in the prior studies, such as Core et al. (1998), Zhou et al. (2011), and Bootsma (2010).

Standard deviation of ROA: is the standard deviation of annual percentage of return on assets for

the previous five years. In the prior study, standard deviation of ROA is used as risk proxy (J.E. Core et al, 1998).

Standard deviation of ROE: alternatively, I also utilize standard deviation of ROE as a proxy for

firm risk.

Market Value (MV): the logarithm of the market value of firm’s common equity is utilized as

well as one of the firm performance indicators to test the strength of pay-performance relationship. Similar method can be found in prior studies (J.E. Core et al, 1998 and Balsam et al. 2011).

Controls: A set of controls are incorporated in the model. Industry-indicator is used as control

for industry differences in the demand for managerial talent. Leverage is a risk control variable as it measures company risk and influences the return on existing equity capital (John and John, 1993; Bührer M.S., 2010) and the voting power of managers’ equity stakes. Managers may therefore increase the debt-assets in response to equity compensation (Hengartner, 2006; Bührer M.S., 2010). Furthermore, Time-indicator is utilized as control variable which allows us to

2_{Tobin’s Q ration is calculated as the market value of a company divided by the replacement value of the firm’s}

assets: Q Ratio = Total Market Value of Firm / Total Asset Value. Tobin’s Q ration is between 0 and 1. A low Q ratio means that the cost to replace a firm’s assets is greater than the value of its stock. This implies that the stock is undervalued. Conversely, a high Q (greater than 1) implies that a firm’s stock is more expensive than the

replacement cost of its assets, which implies that the stock is overvalued.

(23)

measure strength of pay-performance relationship whether changes over the periods. In addition, interaction terms (ROA * Crisis dummy3 and ROE * Crisis dummy3) are created to test sensitivity whether changes prior to the crisis and during the crisis.

3.2.3 Research design

I start with examining the pay-performance relationship for the whole period from 2002 to 2009 using a basic model from Jensen and Murphy (1990). Then, I investigate whether pay-performance sensitivity has changed over the time. I construct the latter test in two ways. The first way is I create extra interaction terms (ROA * crisis dummy and ROE * crisis dummy) as explanatory variable and include them into the regression models. The second way is simply estimate the basic model for different sub periods in order to check whether the magnitude (significance) of the coefficients varies differently. Separate regressions were run using Total Annual Compensation, Total Direct Compensation, and Total Equity Compensation in turn. Sub-periods mean period before the crisis (up to 2007) and period after the crisis (from 2008). Year 2008 is chosen to distinguish pre and after crisis is because the U.S. sub prime mortgage crisis was a set of events and conditions that led to a financial crisis and subsequent recession that began in 2008. Comparing those two methods, the second way is more appropriate when there is a significant amount of observations for all sub-periods considered. However, in this study, the estimation for the crisis period is based on a very small number of observations (only 35). Thus, it is worth noting that the results might have limited predictions. For the first method the regression model can be described as follows:

Model 1: Ln(compensation)it=a + •1*ROAit + •2*Std dev of ROAit + •3*Ln(Sales)it +

•4*ROA*Crisis Dummyit + •5*Industry controlsit + •6*Leverage controlsit + •it

In addition to using ROA as firm performance indicator, we also run the regression by using ROE as “return on equity” indicates how effectively a company’s management uses investors’ money.

So replace ROA with ROE in Model 1:

(24)

Ln(compensation)it=a + •1*ROEit + •2*Std dev of ROEit + •3*Ln(MV)it +

•4*Tobin’Qit+ •5*ROE*Crisis Dummyit + •6*Industry controlsit + •7*Leverage

controlsit + •it

Where Ln(compensation)it denotes the natural logarithm of CEO compensations respectively, for

company i in year t; Ln(Sales)it and LN(MV)it are expressed as firm size indicators; ROAit ,

ROEit , Std dev of ROAit , Std dev of ROEit , and Tobin’s Qitare expressed as firm performance

indicators; Year controls it, Industry controls it, Leverage controls it , and interaction terms

(ROA*crisis dummy and ROE*crisis dummy) are the control variables which enables us to provide evidence regarding the relative importance of these variables in explaining CEO pay; •it

is the error term; •1,•2, •3,… are coefficients which reflect pay-performance sensitivity.

For the second method, the model can be illustrated like this:

Model 2: Ln(compensation)it=a + •1*ROAit+ •2* Std dev of ROAit+ •3*Ln(Sales) it +

•4*Industry controlsit + •5*Leverage controlsit + •it

Ln(compensation)it=a + •1*ROEit + •2*Std dev of ROEit + •3*Ln(MV)it +

•4*Tobin’Qit+ •5*Industry controlsit + •6*Leverage controlsit + •it

To testing the third hypothesis “firm risk affects CEO compensation in UK” I simply use model 2 without splitting periods, in addition, I also add one extra control variable “Year control” into the regression model. Thus, the third model is presented like this:

Model 3: Ln(compensation)it=a+•1*ROAit+ •2* Std dev of ROAit + •3*Ln(Sales) it +

•4*Year controlsit + •5*Industry controlsit + •6*Leverage controlsit + •it

Ln(compensation)it=a+•1*ROEit + •2*Std dev of ROEit + •3*Ln(MV)it +

•4*Tobins’Qit + •5*Year controls it + •6*Industry controls it + •7*Leverage

controlsit + •it

4. Descriptive Statistics and Results

4.1 Sample descriptive statistics

The annual averages shown in Table 2 indicate that CEO compensation did not increase in tandem with return on asset, return on equity, net sales, and market value of the company.

(25)

Namely, year-on-year averages indicate a negative relationship between executive compensation and performance and size.

The descriptive statistics shown in Table 3 indicate that that the average CEO compensations were very volatile as reflected by extremely standard deviations. The same can be said of the explanatory variables. Net Sales, Market Value, ROA, and ROE show high volatilities among the companies.

4.2 Correlation analysis

The Spearman’s correlation coefficient was used to examine the extent of linear correlation among the key variables which are utilized in the model as indicated in Table 4. Logarithm Net sales and Logarithm Market value (MV) appears to have a strong positive correlation (correlation coefficient = 0.7396). Moreover, Return on Assets (ROA) and Return on Equity (ROE) appear to have a strong positive correlation (correlation coefficient = 0.8021). In addition, Standard deviation of ROA and Standard deviation of ROE shows a relative strong positive correlation (correlation coefficient = 0.6596). However, for the variables which are strongly correlated, they were not be used in the same regression. Furthermore, there is almost no evidence of multi-collinearity among the explanatory variables and control variables since the correlations among them are not very strong. Thus, all the variables can be incorporated into the subsequent regression analysis.

4.3 Hausman test

Hausman test tests whether there is a significant difference between the fixed effects (FE) and random effects (RE) estimator. The FE model allows the individual-specific effects •i to be correlated with regressors X while RE model assumes that the individual-specific effects •i are distributed independently of the regressors. The Hausman test statistic can be calculated only for the time-varying regressors. If the Hausman test is insignificant, then RE model should be used; if the Hausman test is significant, then FE model is used.

Hausman test result can be found on Table 5. Test result on each dependent variable shows that Hausman test is insignificant.

(26)

4.4 Regression analysis

The association between CEO compensation and firm performance, firm size, firm performance/value, and firm risk is examined using a GLS regression. The regression equation includes a dependent variable which is one of the three measures of CEO compensation (either total annual compensation, total direct compensation, or total equity compensation) and includes explanatory variables which are presented in Table 1 as proxies for the CEO compensation economic determinants.

The regressions of CEO compensation on its economic determinants, industry type, leverage, and interaction control are presented in Table 6 and Table 7. The dependent variable in the second column is based on total annual compensation, whereas the dependent variables in the third and fourth column are total direct compensation and total equity compensation, respectively.

The regression results presented in the second column of Table 6 and Table 7 demonstrate that the level of total CEO compensation is related to firm size (net sales and market value), firm performance (ROA), firm performance/value (Tobin’s Q) and firm risk (standard deviation of ROA). Larger firms pay higher CEO compensation, which we interpret as reflecting company’s demand for higher talented mangers. The coefficients of LN(Sales), LN(market value), and return on assets (ROA) exhibit a positive and significant association with compensation. The coefficient of return on equity (ROE) is not significant. The coefficient on standard deviation of ROA is positive at marginal significance level, whereas the coefficient on standard deviation of ROE is negative at marginal significance level.

In terms of explanatory power, the regression model in the second column of Table 6 and Table 7 indicates that total CEO annual compensation has a significant association with firm size (net sale/market value), firm performance (ROA), firm performance/value (Tobin’s Q), and firm risk (standard deviation of ROA) (R^2 = 6.25%, F = 1.736, P < 0.05 in Table 6; adjusted-R^2 = 4.6%, F = 1.512, P < 0.05 in Table 7).

We also use total direct compensation as measure of compensation in regression models in Table 6 and Table 7. The third column of Table 6 and Table 7 present the results when direct compensation is used as the dependent variable. These results are essentially identical to those

(27)

using total annual compensation, with the exception that the variable standard deviation of ROA changes sign and loses significance.

The fourth column of Table 6 and Table 7 present the results when total equity compensation is used as the dependent variable. These results are relatively identical to those using total annual compensation, with the exception that the variable return on assets (ROA) loses significance, and variable return on equity (ROE) changes sign and becomes significant.

The results of comparing pay-performance sensitivity in two different periods are illustrated in Table 8 and Table 9. The results show that the relationships between explanatory variables and CEO total annual compensation are the same in both periods, except variable return on equity (ROE) changes sign. Furthermore, the results indicate that firm size (net sales/market value) has a significant positive association with CEO total annual compensation in both period, whereas explanatory variables return on assets (ROA) and Tobin’s Q which are used as firm performance/value indicators completely loss significance during the crisis.

The regressions of CEO compensation and firm risk are presented in Table 10 and Table 11. The regression results presented in the second column of Table 10 demonstrates that the level of total CEO annual compensation is strongly positively related to Standard deviation of ROA (coefficient = 0.0128 with t-statistic = 1.697, 0.01 < P < 0.05). Moreover, the results also reveal strong evidence that the total equity compensation is positively linked to the standard deviation for ROA (coefficient = 0.026 with t-statistic = 1.783, 0.01 < P < 0.05) in column four in Table 10. The variance or standard deviation of historical accounting returns (e.g. ROA) are commonly used by strategic management researchers to measure firm risk because, firm risk generally refers to the variability of firm’s performance (Bromiley et al., 2001; Miller and Reuer, 1996; Miller and Bromiley, 1990; Miller and Chen, 2003). Thus, the positive relationship indicates that CEO compensation protects them suffering from firm risk (standard deviation of ROA) which might be out of their control.

4.5 Additional findings

Principal of agent theory considers the CEO compensation package as an instrument to alleviate the agency problem with assumption that compensation is organized to provide agent (supervisory director) with incentives to maximize the long-run firm’s value (Jensen and Murphy 2004; Shum 2010). Thus, we would expect that size of the supervisory directors might have

(28)

positive influence on managerial performance, firm performance, and protecting shareholder’s interests.

There are a number of prior studies have found that the supervisory board size could affects managerial pay positively or negatively. Knop and Mertens (2010) conclude that large or “overcrowded” supervisory boards are less effective which results in higher CEO compensation. Similar view has been shared by Ghose and Sirmans (2003) who points that larger board size is less effective in monitoring because of lack of interaction, and a lack of focus due to potentially wide dispersion in directors’ skills and motivations. Under these circumstances, directors are likely to be retained less for their contribution and more for their personal relationship with the CEO, resulting in lax monitoring and higher CEO pay. A controversial point of view of Fahlenbrach (2008) suggests that firms with larger boards have a significantly smaller pay-for-performance sensitivity. A one-standard-deviation increase in board size is associated with 17.4% lower equity incentives, which corresponds to a dollar amount of $260,000. Based on these arguments, we would like to examine the relationship between supervisory board size and CEO compensation of UK listed companies by using the same data sample. Variable “SD/Board size” and Age were added into Model 3 as explanatory variables to allow for the possibility that total CEO compensation is proportional to board size. The results presented in Table 12 to 13 yield several interesting findings with respect to the rest variables used as predictors of CEO compensation. Among other characteristics considered, “SD/Board size” has a positive and statistically significant association with CEO compensation. Namely, the findings suggest that, ceteris paribus, a higher number of supervisory boards results in a higher CEO pay. In addition, comparing with other types of compensation, the “total equity compensation” is more enhanced with supervisory board size which can be explained by the highest coefficient (coefficient = 2.315 in Table 12 and coefficient = 2.074 in Table 13).

5. Conclusion

This study finds significant evidences supporting hypothesis 1 that CEOs compensation is positively linked to firm’s performance (ROA/Tobin’s Q) in UK. Moreover, this study documents that the positive pay-performance sensitivity is significantly more pronounced for the firms before the crisis than during the crisis. This gives strong evidence not to support hypothesis 2 that pay-performance sensitivity becomes stronger during the recent financial crisis. In addition, this study reveals a causal effect of firm risk (standard deviation of ROA) in the design

(29)

of CEO compensation. There are strong evidences showing that UK firms react to changes in the firm’s risk environment by increasing in CEO annual compensation, and in particular, the level of stock options awarded to CEOs. This gives sufficient evidence to support hypothesis 3 that firm risk affects CEO compensation. All in all, this study has added value into the literature on UK CEO compensation by looking at pay-performance relationship at the crisis period and demonstrating a causal effect of firm risk and CEO compensation.

However, this study is subject to several limitations. First, the data available for this study is not substantial, which includes only 321 observations with complete compensation data and firm performance data. Thus, this study is rather exploratory and further examination is necessary as selected samples might be not sufficient enough to represent pay-performance relationship for the whole population of UK-listed companies. Second, this study has to recognize that there might be an endogenous problem which is similar to other existing studies relating to the topic. This means that it is unknown if estimated relations are due to measurement error, auto-regression with auto-correlated errors, omitted variables et cetera. Therefore, it has to be aware that there might be other variables and relations which could positively or negatively influence on the relationships discussed.

(30)

Reference List

Agrawal, A., Makhija, A.K., Mandelker, G.N., (1991). Executive Compensation and Corporate Performance in Electric and Gas Utilities. Working Paper. Financial Management winter 1991 Issue.

Balafas, N., and Florackis, C., (2013). CEO Compensation and Future Shareholder Returns: Evidence from London Stock Exchange. Working Paper.

Balsam, S., Fan, H., and Mawani, A., (2011). How Canadian CEO compensation is approaching U.S. levels. Working Paper.

Bebchuk, L.A., and Fried, J.M., (2003). Executive Compensation as an Agency Problem. Journal of Economic Perspectives – Volume 17, Number 3 – Summer 2003 – Pages 71-92.

Bebchuk, L.A., and Spamann, H., (2010). Regulating bankers’ pay. Harvard, John M. Olin Center For Law, Economics, and Business. Georgetown Law Journal 2010.

Bebchuk, L., (2012). Executive Pay and the Financial Crisis. The Harvard Law School Forum on Corporate Governance and Financial Regulation.

Berle, A.A., and Means, G.C., (1932). The modern corporation and private property (Macmillan, New York, NY).

Bertrand, M., and Mullainathan, S., (2001). Are CEOs rewarded for luck? The ones without principals are. The quarterly journal of economics, August 2001.

Bhattacharyya, S., Purnanandam, A., (2011). Risk taking by banks: What did we know and when did we know it? Working paper. Ross School of Business at the University of Michigan.

Bolton, P., Mehran, H., and Shapiro, J., (2011). Executive compensation and risk taking. Federal Reserve Bank of New York, Staff Report, no. 456.

Bootsma, A.A., (2010). Pay-for-performance? An Empirical Investigation of the Relationship between Executive Compensation and Firm Performance in the Netherlands.

Boschen, J.F., Smith, K.J., (1995). You can pay me now and you can pay me later: The dynamic response of executive compensation to firm performance. Journal of Business, vol. 68, nr. 4, pp. 577-608.

Bromiley, P., Miller, K.D., and Rau, D., (2001). “Risk In Strategic Management Research”, in M.A. Hitt, R.E. Freeman and J.S. Harrison (eds) The Blackwell handbook of Strategic Management, pp. 259-88. Malden, MA: Blackwell.

(31)

Bruce, A., Skovoroda, R., Fatorusso, J., Buck, T., (2005). Executive Bonus and Firm performance in the UK. Business School Occasional Papers Series, Paper 2005:3 ISBN 1 85091 1950, Loughborough University UK.

Buck, T., Liu, X., and Skovoroda, R., (2008). Top Executive Pay and Firm Performance in China. Journal of International Business Studies, (2008) 39, 833-850.

Bührer, M. S., (2010). CEO and Chairperson Compensation: The impact of the financial crisis. Dissertation, HSG.

Chambers, Gregory L., (2008). CEO compensation structure related to firm performance in the United States defense and aerospace industry. Working paper. Nova Southeastern University, United Sates Florida.

Ciscel, D.H., Carroll, T.M., (1980). The determinants of executive salary: An econometric survey. Review of Economics and Statistics (February 1980), pp. 7-13.

Conyon, M., Gregg, P., Machin, S., (1995). Taking care of business: Executive compensation in the United Kingdom. The Economic Journal, May, Vol. 105, No. 704.

Conyon, M.J., and Peck, S.I., (1998). “Board control, Remuneration Committees and Top Management Compensation.,” Academy of Management Journal, 41(2), 146-157.

Conyon, M., Benito, A., (1999). The Governance of Directors’ Pay: Evidence from UK Companies. Journal of management and Governance 3: 117-136, 1999.

Conyon, M., and Murphy, K., (2000). “The Prince and The Pauper? CEO Pay in The United States and United Kingdom”, Economic Journal, 105, 704-714.

Conyon, M.J., (2006). Executive Compensation and Incentives. Working paper.

Conyon, M., Core, J., Guay, W., (2011). Are U.S. CEOs paid more than U.K. CEOs? Inferences from risk-adjusted pay. Review of Financial Studies 24, 402-438.

Core, J.E., Holthausen, R.W., Larcker, D.F., (1998). Corporate governance, chief executive office compensation, and firm performance. Journey of Financial Economics 51 (1999) 371-406. Davies, J.R., Hillier, D., and McColgan, D., (2005). Ownership Structure, Managerial Behaviour and Corporate value. Journal of Corporate Finance 11, 645-660.

De Angelis, D., Grullon, G., and Michenaud, S., (2013). Downside Risk and the Design of CEO Incentives: Evidence from a Natural Experiment. Working paper.

Edmans, A., (2010). New thinking on executive compensation: Pay CEOs with debt. Working paper.

CEO pay, firm risk and performance : evidence from UK-listed companies.

Master Thesis