Do say on pay votes affect the pay for performance link? : results from S&P 500 firms

(1)

Do Say on Pay Votes affect the Pay for Performance link?

Results from S&P 500 firms Annette de Graeff University of Amsterdam July 1, 2016 Florencio Lopez de Silanes Supervisor ABSTRACT

In this study, I examine the effect of the Say on Pay legislation on the compensation of CEOs of the S&P 500 firms. Based on a sample of 324 firms listed on the S&P 500 throughout the period from 2007 up until 2014, I find evidence that the Say on Pay law leads to an increase in the Pay for Performance relationship. Furthermore, I find that the regulation results in to a significant higher equity ratio with a significant lower cash ratio to total compensation. The effect between failed or passed Say on Pay votes on these ratios is not significant, nor is the amount of votes in favour of the proposed CEO compensation plan.

Keywords: CEO compensation, Say on Pay, Shareholder Voting, Pay for Performance

(2)

2

Statement of Originality

This document is written by Annette de Graeff, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

CONCLUSION ... 26 DISCUSSION AND LIMITATIONS ... 27 FURTHER RESEARCH ... 28 REFERENCES ... 29 APPENDIX ... 35 Figure 1 - The outcome of the Say on Pay votes (2011-2014) ... 35 Figure 2 - Public Float distribution ... 36 Figure 3 - Compensation Ratio’s ... 37 Table 1 – Variable description ... 38 Table 2 - Summary statistics 2008-2014 ... 39 Table 3 – Correlation Matrix independent variables ... 40 Table 4 – Total Compensation OLS Regression ... 41 Table 5 – Effect Say on Pay on Pay for Performance ... 42 Table 7 – RESET test ... 44 Table 8 – Hausman Test ... 45 Table 9 – Fisher Unit Root test ... 46 Table 10 – Effect Say on Pay on Pay for Performance with additional controls ... 47 Table 11 - Effect Say on Pay on Compensation Ratios industry fixed effects ... 48

(4)

4

Introduction

“We are adopting amendments to our rules to implement the provisions of the Dodd-Frank Wall Street Reform and Consumer Protection Act relating to shareholder approval of executive compensation (SEC, 2011, p. 1) 1_”

Compensation of Chief Executive Officers (CEOs) has been subject of discussion for quite some time, and the criticism on this topic has become more severe after the 2008 Financial Crisis. Bertrand and Mullainathan (2001) showed that before the 2008 financial crisis CEOs rewarded themselves irrespective of firm performance. This was not in line with classic principal-agent theories, which state that the owner of a firm should design compensation packages to maximize the value of the firm (Jensen & Meckling, 1976). These excessive or poorly structured compensation arrangements have been blamed for the US Financial Crisis of 2008, exposing executive compensation as an area ready for new regulation (Walker, 2010).

As a result, the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, hereby called the Dodd-Frank Act, was initiated. This act includes several executive compensation and corporate governance requirements. One of these requirements consists of a Say on Pay regulation, which states that shareholders have a non-binding vote on the compensation package for a company’s named executive officer (Anand, 2011). The aim of this regulation is to improve ‘‘accountability, transparency, and performance linkage of executive pay” (Baird & Stowasser, 2002, p. 32). Thomas, Palmiter, and Cotter (2011) believed that by increasing shareholder power, the Say on Pay vote would reduce the CEO pay spiral and would help to link CEO pay to firm performance. Additionally, researchers argued that the regulation could facilitate closer alignment between shareholder interest and the interest of managers, which in turn could lead to a reduction of managerial agency costs (e.g., Yermack, 2010; Correa & Lel, 2013). 1_{Shareholder Approval of Executive Compensation and Golden Parachute Compensation.} 2_{See annual reports on Say on Pay votes at http://www.semlerbrossy.com/say-on-pay/}

(5)

5 Problem Statement

As the vote is non-binding and managers are not obliged to follow the advice of the shareholders, the question is whether the Say on Pay law actually achieves its aim to link pay to performance. It is not apparent that the Securities and Exchange Commission (SEC) considered the usefulness of the Say on Pay vote when it was implemented in the US in January 2011. Also, it is not clear whether the SEC contemplated how Say on Pay would influence pay practices or the level of compensation. It is essential to explore this relationship as the compensation of CEOs and the link to firm performance is a measure of the effectiveness of the CEO contract (Bertrand & Mullainathan, 2001). Until now, little research on this subject has been performed. Therefore, there is need to investigate the effects of the Say on Pay law on pay for performance link of CEOs.

Relevance

This research is relevant because it provides insight into the effect of the non-binding Say on Pay votes on the compensation of CEOs in the US and will contribute to the on-going debate about the level of impact the Say on Pay law has had on linking performance to pay. It contributes to the literature on the influence of shareholder voting on executive compensation (e.g., Jensen & Murphy, 1990; Gillan & Starks, 2000) and adds to the discussion of whether shareholder voting is a weak or strong influential mechanism (e.g., Gillan & Starks, 2000). The findings of this research have the potential to build upon the asymmetries between the relationship of performance and pay that have been previously identified (e.g., Bertrand & Mullainathan, 2001; Conyon & He, 2011). Most importantly, this is the first research to focus on the effect of the Say on Pay law on pay for performance sensitivity in the US and therefore provides insight on the effectiveness of this law in the US.

The findings of this research may have implications for several groups. First, this research is important for the board members since through their votes, shareholders can directly question choices of the directors. A high responsiveness to Say on Pay votes is likely to generate concerns for their

(6)

6

reputation. Second, this research may show that monitoring by shareholders does matter for CEO compensation, and that therefore, boards would be interested to learn that shareholders can be enhancing the value of the firm due to this monitoring mechanism (Ferri & Maber, 2013; Hartzell & Starks, 2003). Third, shareholders may be interested to learn about the potential of Say on Pay votes to act as an activism tool (Cai et al. 2009). Finally, US policymakers may find this research informative, as there is an on-going legal discussion on pay for performance policy in the US. Aim and Structure The ultimate aim of this thesis is to show whether or not the Say on Pay regulation has a significant effect on the pay for performance sensitivity of US firms. It intends to give a broader understanding of Say on Pay regulation and the intrinsic motivation for companies to respond to non-binding shareholder votes. The conclusion can hopefully be useful in the generalization of the effectiveness of these Say on Pay votes.

The structure of my research is as follows. I will start by providing an in-depth literature review on the US adoption of the Say on Pay regulation. This literature review also includes an extensive overview of prior research on the Say on Pay regulation in countries aside from the US. In addition, an overview of the most relevant literature on the pay for performance relationship will be submitted, as well as the effect of shareholder voting on pay practices

Based on this literature review two hypotheses will be raised that will subsequently be tested. I then detail my methodology, as well as the reasoning behind the control variables used for this research, followed by a discussion on the sample and results. To examine this effect, I selected a sample of Standard & Poor’s 500 (S&P 500) firms within a time period of 2007 until 2014. By selecting S&P 500 firms, I narrowed my sample to comparative firms. Using Ordinary Least Square (OLS) and panel regressions, I will identify the effect of the Say on Pay rule, comparing CEO compensation in the periods before and after the introduction.

(7)

7

I analyze the effect of the non-binding votes of these firms on total CEO compensation, the pay for performance link, and the effect on the equity and cash ratio to total compensation. In the findings, I discuss how the Say on Pay law in the US has led to a significantly higher CEO compensation and that this compensation is more linked to firm performance compared to the period before the introduction of the law. Furthermore, I document that Say on Pay votes have led to significantly lower cash and a significantly higher equity ratio. These findings complement earlier findings on Say on Pay votes (e.g., Ferri & Maber, 2013; Iliev & Vitanova, 2015).

Before coming to the final conclusion of my hypotheses, I will engage in robustness checks and discuss identified endogeneity issues. Lastly, I discuss the results and review the limitations of my thesis providing also suggestions for

(8)

8

Literature Review and Hypotheses

In this literature review, I describe the most relevant literature related to this research. After extensive review, I also explain the two hypotheses tested for this thesis. A New Regulation

The Say on Pay regulation in the US is relatively new. This law became compulsory at all US firms in 2011, though it was introduced earlier in other countries. The Say on Pay regulation differs across varying countries (Cai & Walkling, 2011). For example, in some countries the Say on Pay law is binding, while in others it is non-binding. In recent years, there have been multiple studies examining the impact of Say on Pay. Ferri and Maber (2013) study the effect of the Say on Pay vote in the UK and find an improvement in the relationship between firm performance and CEO salary. The researchers report that even though only a small number of Say on Pay resolutions fail to pass, negative vote totals tend to increase when executive compensation is higher than expected. In the aftermath of high negative votes, firms tend to reduce the amount of total compensation and increase the performance sensitivity of top managers’ pay (Yermack, 2010).

Iliev and Vitanova (2015) investigate the effect of Say on Pay vote in the US, which is closely tied to this research. However, they focus mainly on the market reaction of the Say on Pay vote, while the focus of this study is on the pay for performance sensitivity. The work of these researchers gives background for my sample construction and provides more understanding of the Say on Pay rule in the US. Lastly, Correa and Lel (2013) use a large cross-country sample to examine the effects of Say on Pay, finding that the Say on Pay law increases pay for performance sensitivity.

(9)

9 Say on Pay vote The Say on Pay vote is best defined as shareholders having a non-binding vote on the compensation package for a company’s named executive officers [the CEO, Chief Financial Officer (CFO) and top three most other highly compensated executive officers (Anand, 2011)]. The holding of such a vote is required under Section 951 of the Dodd-Frank Act. Shareholder resolutions for the Say on Pay votes must be included in a public company’s 8-k proxy materials for its first annual shareholders’ meeting occurring on January 21, 2011 and thereafter.

One group exempted from the Say on Pay regulation was “smaller reporting” companies, which generally are those with worldwide public equity floats of less than $75 million. This exemption only holds for the first two years of the law, 2011 and 2012. As of January 21, 2013, all firms are obligated to hold a Say on Pay vote, irrespective of their public float. Say on Pay in the US: A Debate Since the votes of the Say on Pay regulation in the US are non-binding, the Say on Pay regulation initiated a debate: How effective is the Say on Pay regulation in aligning executive pay practices with shareholders’ interests? Prior research shows that non-binding shareholder votes generally fail to convey shareholder thoughts, and thus have a limited advisory role (Levit & Malenko 2011). This suggests that shareholder votes are typically a weak governance mechanism. However, other recent researchers have shown that boards become increasingly responsive to non-binding shareholder proposals (Thomas & Cotter, 2007; Ertimur et al., 2010). For example, Ertimur et al. (2010) focusing on non-binding, majority-vote shareholder proposals, find that the frequency of implementation is increasing. In addition to Say on Pay votes, there are several determinants of a board’s decision to implement shareholder proposals, such as shareholder pressure, governance structure, and type of proposals.

(10)

10 Pay for Performance Murphy (1985) depicts the link between pay and performance, as well as the major influence of regulatory changes on patterns in executive compensation (Fernandes et al., 2013). Additionally, Murphy and Jensen (1990a) argue that it is not a matter of how much CEOs get paid, but how they get paid. The focus, as they argue, is too much on the excessiveness of CEO pay. The focus should shift from the total amount of compensation to the link of compensation with performance of the company. The compensation of a CEO is usually viewed through a lens of principal-agent models, meaning that the owner of a firm designs compensation packages to maximize the value of the firm (Jensen & Meckling, 1976). This approach is called the “optimal contracting approach” (Bebchuk & Fried, 2003, p.72) However, research has shown that this approach does not always hold. The Say on Pay votes were introduced to minimize this separation of ownership and control (Bertrand & Mullainathan, 2001). Theory argues that firms should control agency costs that come with this separation (e.g., Jensen & Meckling, 1976; Bebchuk & Fried, 2003). By tying the managerial compensation directly to compensation through Say on Pay votes, the agency costs are reduced and align the interests of the shareholders with the interests of the managers.

The Say on Pay law can have an effect on the pay for performance sensitivity in several ways. In order for non-binding votes to impact compensation practices, several incentives must be attached to the realization of an adverse outcome. Since CEOs are concerned about shareholder support, they are likely to act on behalf of the shareholder input through the Say on Pay votes (Ferri & Maber, 2013). Furthermore, Say on Pay laws are a monitoring mechanism for shareholders and these mechanisms have proven to have an impact on CEO compensation (Hartzell & Starks, 2003). These factors influence the boards’ decisions to implement the shareholder proposal.

Prior research has shown that pay is sensitive to firm specific performance and that the pay for performance link is higher when shareholders are diffuse and arguably more passive (Bertrand & Mullainathan, 2001). Therefore, if the findings of this research suggest that shareholder voting leads

(11)

11 to stronger pay for performance sensitivity, it has the potential to enhance firm value even if it does influence the level of CEO compensation. Shareholder Voting

Implementing the Say on Pay rule meant that regulators took a strong interest in shareholder voting. Bebchuk and Fried (2004) urged that shareholders receive an advisory vote on executive pay at US companies, concluding that such a vote would allow shareholders to express their views on flawed pay practices, particularly pay unconnected with company performance.

Moreover, prior research shows that shareholders care about the link between pay and performance for executives, and that the market reacts positively to the announcement of equity-based proposals that tie pay with performance, particularly when those plans involve executives (DeFusco, Johnson, & Zorn, 1990). By tying compensation to the profitability of the firm and the wealth of the shareholders, shareholders encourage managers to work harder and take appropriate risks and, in general, align managerial incentives with shareholder well-being (Morgan & Poulson, 2001).

Combining the involvement of shareholders with the pay for performance sensitivity with the increasing responsiveness to non-binding shareholder proposals, one would predict that the Say on Pay regulation leads to a higher pay for performance linkage. However, the evidence on the effect of the Say on Pay regulation in different countries is ambiguous. Davis (2009) shows that the Say on Pay regulation links firm performance to pay more closely, but in contrast there is also evidence that Say on Pay regulations might lead to suboptimal pay practices (Gordon, 2009). This research provides more insight on the effect of the Say on Pay rule in the US and combines literature about executive compensation with literature about the effects of shareholder voting (Jensen & Murphy, 1990; Atanasov et al., 2010).

(12)

12 Hypotheses Development Hypothesis 1: The Say on Pay regulation leads to an increase in the pay for performance sensitivity

The aim of the Say on Pay regulation is to improve ‘‘accountability, transparency, and performance linkage of executive pay’’ (Baird & Stowasser, 2002, p. 32). Shareholders’ Say on Pay votes send a signal to the management board about their opinion on the compensation package. Jensen and Murphy (1990) show the relationship between CEO pay and the change in firm performance. These authors first introduced the pay for performance relationship. The main question of this research is whether the pay for performance relationship influences Say on Pay outcomes.

Based on research done in other countries with a non-binding Say on Pay vote, I predict an increase in the pay for performance sensitivity in the US (Ferri & Maber, 2013). Even if the Say on Pay vote fails, i.e. shareholders vote against the executive compensation package, it would be most likely because the firm has underperformed. Therefore, I still expect an increase in the pay for performance sensitivity.

Hypothesis 2: Firms that passed the Say on Pay voting have a higher equity to total compensation ratio than firms that failed the Say on Pay vote

Excessive executive compensation can be part of agency problems, and with that agency costs arise. CEO Compensation can be viewed as an instrument for addressing agency problems, as well as a part of the agency problem itself (Bebchuk & Fried, 2003). The agency costs associated with these problems can be mitigated by incentive based payments, such as equity payments (Hartzell & Straks, 2003). These so called long-term incentive payment parts of the total compensation can align interests of shareholders with the managers, reducing monitoring costs and thus reducing failure in the Say on Pay. Therefore, I predict a higher equity to total compensation ratio for firms that passed their Say on Pay

(13)

13

Methodology

To test whether the Say on Pay law has influenced the CEO compensation, I follow prior research done by Ferri and Maber (2013) and use a panel regression. This is not the most common method used on testing the effects of Say on Pay votes, which is the Difference-in-Difference method (e.g., Iliev & Vitanova, 2015). I reached the conclusion that the Difference-in-Difference method is excluded for different reasons. The main reason is that the Say on Pay vote in the US was required as of January 21, 2011 for companies with a $75 million public equity float or more. The group exempted from the Say on Pay regulation, called the “smaller reporting” companies, are those with public equity floats of less than $75 million. As previously mentioned, none of the firms in this sample are a “smaller reporting” company. Therefore, there are no control groups and the Difference-in-Difference method is thus excluded. It is therefore that I use an panel regression and I specified the following OLS panel regression:

Log (CEOComp)it = β1Say-on-PayVOTE + log(total assets) + β2SalesGrowth +

β3ROA+ β4Market-to-Book + β5Classified + βxCEOcharacteristics + εi

Using log to adjust for the non-normality of compensation distribution, the dependent variable is the log of total CEO compensation. Research has shown that compensation to firm performance is a key measure of the effectiveness of the CEO pay contract (Bertrand & Mullainathan, 2001). The total compensation of CEOs consists of several elements, such as cash, bonus, equity payment etc. If the shareholder votes of the Say on Pay lead to a stronger pay for performance sensitivity, then the law can influence firm value without changing the actual level of the CEO pay. Following common practice, I used lagged variables as the effect of the law is most probably lagged (Ferri & Maber, 2013).

The term Say-on-PayVOTE is a dummy variable equal to one if the period is after the introduction of the Say on Pay regulation (2011-2014), the opposite holds for the period before the introduction (2007-2010). By comparing the before and after period, I am able to draw conclusions on the effect of the Say on Pay law, especially when it turns out that my outcomes are in line with the findings of Peter and Vitanova (2015).

(14)

14

For the second hypothesis, I used the following panel regression:

𝐸𝑞𝑢𝑖ty-ratio = β1Say-on-PayVOTE + log (total assets) + β2SalesGrowth + β3ROA+

β4Firmreturn+ β5Classified + βxCEOcharacteristics + εi

Independent Variables

Following prior research, the control variables are log of total assets, Sales Growth, Return on Assets (ROA), Stock Return, Market to Book ratio, and CEO characteristics (Yermack, 1995; Fernandes et al., 2013). I used the accounting based measure of firm performance, ROA, because of the direct correlation between performance and compensation as shown in previous research (Jensen & Meckling, 1976). Similar to prior studies, I also included price-based Stock Return (Jensen & Murphy, 1990; Murphy, 1985). Stock Return is necessary to consider because CEO compensation is sensitive to returns (Leone, Wu, & Zimmerman, 2006).

I used Sales Growth as a proxy of size. It was necessary to include this variable as it is shown that larger firms are more complex, and therefore, require better CEOs and thus, a higher compensation package (Gabaix & Landier, 2008). I also included log of total assets because of the correlation with compensation (Barth, Lin & Wihlborg, 2012). The Market to Book ratio functioned as a proxy for growth opportunities. Smith and Watss (1992) find that pay performance sensitivity is related to proxies for growth options. Additionally, I included CEO characteristics, such as age, duality, and tenure.

Finally, I controlled for year and industry fixed effects and clustered the error term on firm level. By doing so, I accounted for multiple observations per firm, and thus controlled for omitted firm specific characteristics that are constant over time (Murphy, 1985; Froot, 1989). This mitigated the possibility of

(15)

15

Sample Construction and CEO Compensation

The sample covered the fiscal years of 2007 through 2014, which makes for an equal amount of years before and after the introduction of the Say on Pay law. By including the years 2013 and 2014, the first two years are studied in which all firms of the S&P 500 were obligated to hold a Say on Pay vote, irrespective of their public float.

Sample

I began sampling by retrieving all of the S&P 500 companies from COMPUSTAT, and by merging the index constituents with the annual fundamentals. I selected these particular firms to retain a sample of similar firms in terms of market value. This is important because there is a correlation between the market value of a company and the compensation of CEOs (Larcker, 1983). To reduce potential biases, I only included the companies listed on the S&P 500 throughout the designated time period (2007 through 2014). After collecting data from Semler Brossy on the failed Say on Pay votes, I manually labelled all of the companies that failed the Say on Pay votes as of 2011.2_Votes from 2012 concern the 2012 compensation packages of CEOs and so on. Thus, I included the proxy seasons 2011 through 2014 to have an equal amount of years before and after the introduction of the law, namely four years. Figure 1 shows the outcome of the Say on Pay votes. The exact percentile outcomes of the Say on Pay votes can also be retrieved from the 8-k forms for the years following 2010 from the database proxy monitor.3_{Through my sampling, I selected 168 firms}

with exact data on the Say on Pay vote outcomes.

I retrieved information about Executive Compensation from the EXECUCOMP database and merged this file in my database. The total compensation of CEOs includes salaries, bonuses, restricted stock and option

2_{See annual reports on Say on Pay votes at http://www.semlerbrossy.com/say-on-pay/} 3_{The website http://www.proxymonitor.org/ provides direct outcomes of the 8-k forms on Say} on Pay votes

(16)

16

awards, long-term incentive plans, changes in pension plans, and all other compensation measured in US dollars. Option awards and other equity based compensation (i.e., restricted stock awards) are as reported by companies, and the option values are based on share prices at year-end. I created variables to capture the CEO pay structure; the portion of total CEO pay in the form of equity-based compensation (equity pay/ total compensation), as well as the cash based compensation (cash pay/total compensation). I manually excluded companies with extremely low total compensation.4

Furthermore, I collected the public float manually from the first page of the companies’ 10-k filings on Electonic Data Gathering, Analysis, and Retrieval system (EDGAR). 5_{The EDGAR database is established by the SEC and it provides}

an electronically collection of all the required disclosure in forms and documents. Collecting the public float information was necessary because companies that did not have a public float larger then $75 million as of the last business day of its second fiscal quarter received a two-year exemption from the Say on Pay vote. 6_{In 2013, this two-year exemption granted to “smaller}

reporting” companies under the Dodd-Frank Act expired, making companies 4_{The companies deleted are Alphabet Inc, Citigroup Inc, Yahoo Inc and Apple Inc due to a total} compensation as reported on the SEC filing 10-k of 0.001$ in one or multiple years. 5_{A Form 10-k is an annual report that gives a comprehensive summary of a company’s financial} performance 6_{Smaller reporting company: As used in this part, the term smaller reporting company means} an issuer that is not an investment company, an asset-backed issuer (as defined in § 229.1101 of this chapter), or a majority-owned subsidiary of a parent that is not a smaller reporting company and that: (1) Had a public float of less than $75 million as of the last business day of its most recently completed second fiscal quarter, computed by multiplying the aggregate worldwide number of shares of its voting and non-voting common equity held by non-affiliates by the price at which the common equity was last sold, or the average of the bid and asked prices of common equity, in the principal market for the common equity; or (2) In the case of an initial registration statement under the Securities Act or Exchange Act for shares of its common equity, had a public float of less than $75 million as of a date within 30 days of the date of the filing of the registration statement computed by multiplying the aggregate worldwide number of such shares held by non-affiliates before the registration plus, in the case of a Securities Act registration statement, the number of such shares included in the registration statement by the estimated public offering price of the shares; or (3) In the case of an issuer whose public float as calculated under paragraph (1) or (2) of this definition was zero, had annual revenues of less than $50 million during the most recently completed fiscal year for which audited financial statements are available.

(17)

17 with a public float of less than $75 million subject to existing Say on Pay voting requirements. As aforementioned, some firms held a Say on Pay vote even if they were not obligated to do so. 7 However, this does not hold for the S&P 500 firms. Figure 2 shows that for all the S&P 500 firms of my sample, the public float is much bigger than $75 million. Thus, none of these firms were excluded from the regulation and they all had to hold the Say on Pay vote since none of these companies were classified as ‘smaller reporting’ companies. I also collected data about CEO characteristics from proxy statements on form DEF14a, also retrieved form the EDGAR database. 8_{These characteristics included the following: CEO tenure, CEO duality,}

classified board, and age of the CEO. I also created a dummy variable equal to 1 if it was the CEO’s first year. Lastly, I created the independent variables, which were: Market to Book and Sales Growth. Market to Book is defined as plus common shares outstanding times the closing price plus total assets minus total stockholders’ equity and deferred taxes normalized by total assets and will function as a proxy for growth opportunities. Sales Growth is the log of the ratio of current sales to lagged sales and functions as a proxy of firm growth opportunities. I merged the file of the CEO characteristics with the date on CEO compensation and the general S&P 500. Finally, I dropped all duplicates within the file resulting in 324 unique firms and 2590 observations. Descriptive Statistics and Correlation Matrix The sample for this study consists of 324 firms. Its variable descriptions are presented in Table 1. Table 2 provides the descriptive statistics, showing that the S&P 500 firms have an average asset value of $69 million. The mean return on assets is positive, just as the mean of Sales Growth, which is defined as the log of the ratio of current sales (SALE) to lagged sales. Table 2 also provides several CEO characteristics, such as age, tenure, duality and whether the board of the company is a classified board. On average, the CEOs are 57 years old.

7_{Determination: Whether or not an issuer is a smaller reporting company is determined on an} annual basis.

(18)

18

Approximately 7% of the CEOs are in their first year, while the mean tenure is 6.25 years. The average compensation CEOs received during the period of 2007 through 2014 was around $12 million per year, consisting of an average equity ratio of 53% and a mean cash ratio is 15%. The average votes for the proposed CEO compensation plan is 88.92%. Table 3 provides the correlation matrix for the independent variables. From this table, the conclusion can be drawn that the explanatory variables appear to be independent. All the correlations are relatively low; the highest correlation is between total assets and Market to Book and is “only” 20%.

(19)

19

Empirical Results

Do Say on Pay votes lead to significant differences in CEO compensation? This section tests the effects of this law, starting the analysis by reporting the effect of the Say on Pay regulation on the CEO compensation in general. Subsequently I explore the effect of the law on the pay for performance linkage. This section concludes with the effect of the Say on Pay votes on the equity and cash part of the total compensation of CEOs. Vote Outcomes

The Say on Pay votes granted a two-year exemption for “smaller reporting” companies, i.e. a public float of smaller than $75 million. The S&P 500 sample that was used for this thesis were required to hold Say on Pay votes because none of these firms had a public float smaller than the $75 million benchmark as is shown by Figure 2. Figure 1 shows the percentile of failed and passed votes for my sample. On average around 2% of the firms fail their Say on Pay votes annually. For a firm to fail their Say on Pay votes it must receive less than 50% of votes in favor of the proposed CEO compensation plan. This equates to approximately six companies failing their Say on Pay Votes each year. Therefore, the conclusion can be drawn that failure of Say on Pay votes is rare. Semler Brossy examines the reasons why these particular companies failed their Say on Pay votes; one of the reasons was always the lack of performance linkage with compensation. On average firms receive 88.92% votes in favor of the compensation package, as is shown by Table 2.

Effect of Say on Pay Regulation on CEO Compensation

The Say on Pay regulation was introduced with the aim to more closely link CEO pay with company performance (Baird & Stowasser, 2002). Shareholder votes were expected to give shareholders input on this issue so that they might have effect on CEO compensation.

(20)

20

Table 3 reports the results of the two panel regressions, comparing the period before the Say on Pay regulation was mandatory with the period after the introduction of the regulation. The first column reports the regression of the effect of the Say on Pay vote dummy variable on the log of CEO compensation. The second column also controls for a variety of CEO characteristics, such as tenure and age, as well as a variety of firm characteristics, for example ROA and Market to Book ratio. It shows that the Say on Pay regulation has a significant effect on the CEO compensation. The regression shows a statistically significant increase of 21% comparing the years before the introduction with the years after the introduction. It is important to note that the estimated effect in the panel regression with control variables is 14% lower, meaning that firms have different characteristics. However, the effect remains positive on the total amount of compensation, meaning that after the introduction of the law, CEOs earned more.

This means that the Say on Pay vote in the US has so far led to a higher total compensation of CEOs.). Iliev and Vitanova (2015) argue that this result is because of the additional risk that shareholder votes generate for CEOs and that CEOs should have more compensation for this risk. However, this result does not mean that the Say on Pay law has also led to a performance linkage of CEO compensation; it only states something about the compensation in general. Effect of Say on Pay Regulation on Pay for Performance Sensitivity Next, I ask whether the Say on Pay regulation has led to a higher pay for performance sensitivity. My first hypotheses, related to this, stated the following: The Say on Pay regulation leads to an increase in the pay for performance sensitivity. Prior research has focused on the sensitivity of CEO compensation to firm performance as a measure of pay contract efficiency (Bertrand & Mullainathan, 2001). If non-binding Say on Pay votes strengthen the link between CEO compensation and firm performance, then Say on Pay regulation enhances firm value.

Table 5 reports the results of regressing the log of CEO compensation on firm returns, defined as the 12 month return over the previous fiscal year in

(21)

21

percentages, interacted with pre and post regulation dummy’s. I find that CEO compensation sensitivity significantly increases with performance after the Say on Pay rule became effective. This result is supported by the findings of Ferri and Maber (2013) and Iliev and Vitanova (2015).

Moreover, before the Say on Pay law was in place, the pay for performance sensitivity was significantly lower. This finding further emphasizes that after the introduction of the law, the pay for performance sensitivity increased. Thus, one can conclude that the Say on Pay regulation achieved its aim to create a stronger link between pay and performance. I therefore conclude that my first hypothesis holds. Effect of Say on Pay Regulation on Compensation Ratios Knowing that the Say on Pay votes influence the CEO compensation level as well as its link with company performance, I then explored the effect on the different ratios of the compensation. The second hypothesis states: Firms that passed the Say on Pay voting have a higher equity to total compensation ratio than firms that failed the Say on Pay vote. I expect that the equity compensation aligns the interests of shareholders with the interests of managers, reducing monitoring costs and thus reducing failure in the Say on Pay.

Table 6 provides an overview on whether the introduction of the Say on Pay votes has significantly changed the cash and equity ratio, as well as the influence of passed proposed compensation plan and the effect of a higher passing range on these ratios. I began my analysis by regressing the ratios on a dummy variable equal to one for the period after the introduction of the Say on Pay regulation. The first and second columns provide the results of this regression. I found a significant increase of 0.094 in the equity ratio of the compensation due to the introduction of Say on Pay votes. I also found a significant decrease of 0.042 in the cash ratio of CEO compensation due to the introduction of the Say on Pay votes.

The third and fourth columns depict the effect of passed Say on Pay votes on the compensation ratios. I determined a non-significant decrease in the equity ratio of 0.021 as well as a non-significant increase of 0.0046 in cash ratio if the

(22)

22

firm passed their Say on Pay vote. These results do not align with my second hypothesis. Therefore, one cannot conclude that long-term incentive payment aligns the interests of shareholders with those of the managers, reducing monitoring costs and thus reducing failure in the Say on Pay. In the discussion, I will further elaborate on possible theories for this result.

The results of columns 3 and 4 are in line with my findings of columns 5 and 6, which investigate the impact of the voting rate in favor of the proposed compensation plan. I found that a higher rate of votes in favor lead to a small non-significant decrease in equity ratio of 0.00039 and a non-significant increase of 0.000079 in the cash ratio. This result is supported by Figure 3, showing that the equity as well as the cash ratio both not significantly increase or decrease over time. Both of the ratios show a decline after the introduction of the Say on Pay law in 2011, which supports the significant decreases of the first and second columns.

The lack of significance in the third, fourth, fifth, and sixth columns in Table 6 are difficult to explain within the theoretical framework of hypothesis 2. Although I expected that firms with a higher passing rate would also have a higher equity to total compensation ratio because the incentives would be even more aligned, it seems, though not significant, that the opposite holds; a higher passing rate leads to a lower equity ratio. It is worth noting that the Say on Pay votes itself led to a significantly higher equity ratio. However, one cannot conclude that a higher rate also leads to a higher equity ratio.

(23)

23

Robustness Check

This section includes discussion of the robustness checks used to ensure the quality and credibility of this study. It also explores some possible endogeneity issues. Some of these issues discussed were addressed in the sample construction and the regressions for data analysis. Endogeneity and Fixed Effects For robust inference, I have already used the clustered option throughout the thesis. If I do use the robust standard errors, the results do not differ. This means that the function is homoscedastic and that the endogeneity assumption is not violated. I also tested the multicollinearity by examining the respective correlation coefficients between the independent variables, as well as between the independent variables and the dependent variable. As Table 3 shows, there is no perfect correlation between any of the independent variables, which means that there is no perfect collinearity.

To ensure the quality of this study, I also tested for omitted variable bias using the Ramsey Regression Specification Error Test9_{. Omitted variable bias is a}

common endogeneity issue that arises when using an OLS regression. Omitted variables are variables that are not in the regression model, but that have an effect on the outcome of the dependent variable. By not including these variables the problem of omitted variable bias arises. The approximation of the dependent variable without these variables would be less precise than an approximation of the dependent variable with these omitted variables taken into account. Table 6 shows the outcome of the Ramsey Regression Specification Error Test, from which the conclusion can be drawn that the regression does not have an omitted variable bias.

Lastly, in order to test the effect of the Say on Pay law on CEO compensation, I used fixed effects in all of my regressions. Even though the

9_{The statistical test as proposed by J.B. Ramsey}

(24)

24

random effects have become increasingly popular to use, the Hausman test, as reported in table 8, is significant and thus the fixed effects estimator is to be preferred over the random effects (Wooldridge, 2002). So these effects are used to mitigate endogeneity issues. However using fixed effects does not necessarily eliminate endogeneity (Liu et al., 2015).,although these fixed effects do allow for time and/or company specific differences in the panel data and so leaving these firm fixed effects out could affect the outcome. Unit Root There are several tests that can be used to control for a unit root, such as the Levin–Lin–Chu (2002), Harris–Tzavalis (1999), Im–Pesaran–Shin (2003), and Fisher Unit root test (Fisher-type) (Choi 2001). The null hypothesis for all of these tests is that all of the panels contain a unit root. Besides these tests, the Augmented Dickey Fuller (ADF) test is excluded, since this test is more common to use in time series when the other tests are being use for panel data.

For this research, I used the Fisher test which is based on the ADF test. The Fisher test was best suited to my study because the panel data had gaps within and did not have strongly balanced data, causing me to rule out the other tests10_{. Table 9 show the results of the Fisher-type of the log of CEO}

compensation. Since all the four p-values are smaller than 0.01, I strongly rejected the null hypothesis that all the panels contain unit root. This indicates that the log of CEO compensation is stationary.

Control Variables

I reproduced Table 5 and included more control variables to verify if adding more control variables had a significant influence on my previously presented results. I also added Market to Book ratio because it is a proxy for growth opportunities. Furthermore, I added the leverage (debt to equity) ratio because some studies suggest that there is a direct relationship between the

10_{The Harris-Tzavlis test as well as the Levin-Li-Chiu test require strongly balanced data,} whereas the Im-Pesaran-Shin test cannot have gaps in the data.

(25)

25

leverage ratio of a firm and the compensation of a CEO (Liu & Mauer, 2011). I define leverage as total liabilities divided by shareholders’ equity. I winsorized the outcome at a 1% level to ensure that the extreme outcomes were confined. Table 10 provides the outcomes with these two additional control variables. Table 10 shows that in the period before the introduction of the Say on Pay law there is little difference when including these two variables. However, the result of the period when the Say on Pay law became active differs: the results remain positive. However, the result is not significant.

Industry Fixed Effects

Having already controlled for omitted firm specific characteristics that are constant over time by clustering the error term on firm level, I then controlled for year fixed effects at industry level for potential omitted time varying variables. Controlling for these fixed effects ensured that my results were not the result of changes at industry level in the period of the introduction of the law. This is essential because Mosem and Ng (2013) suggest that certain industries receive higher voting dissents. Table 11 shows that my results are robust to this control for omitted time varying variables. It shows that introducing industry fixed effects does not lead to a change in the key indicators of the effect of the Say on Pay law. Therefore, the effect of the Say on Pay law on my key variable of interest, the compensation ratios, is not industry dependent.

(26)

26

Conclusion

This thesis shows the effect of the Say on Pay regulation on the pay for performance sensitivity of CEOs of S&P 500 firms. With the use of OLS and panel regressions, I identified the effect of these votes on CEO compensation, comparing the periods before and after the introduction of the Say on Pay law. These results contribute to the literature by showing the effect that non-binding shareholder votes can have on compensation packages and, additionally, whether Say on Pay regulation can limit the problem of excessive compensation. Furthermore, if my findings correspond with findings in other countries such as the UK, it could also provide a generalization of findings of non-binding Say on Pay votes.

In order to answer whether Say on Pay impacts pay for performance sensitivity, I stated two hypotheses. The first hypothesis was the closest related to the main question and focussed on the pay for performance link of CEO compensation. The second hypothesis focuses on the compensation package structure aiming to prove that firms that passed the Say on Pay vote have a higher equity to total compensation ratio than firms that failed.

For the first hypothesis I found that Say on Pay votes result not only in a significant increase in the compensation of the CEO, but also align pay with performance more. This holds for my sample, consisting of 324 firms over a period of eight years. These findings correspond with findings in the UK, which adds to the credibility of my results and provides a generalization of non-binding Say on Pay votes.

For my second hypothesis I found that the introduction of the Say on Pay vote has led to a significantly lower cash and a significantly higher equity ratio. However, I did not find that a higher percentage of votes for the proposed compensation plan led to a higher equity ratio, nor did I find that a passed Say on Pay proposal had a higher equity ratio. This means that the second hypothesis does not hold. In contrast, the results of this thesis show a decrease in the equity part of the compensation when the votes for the proposed CEO compensation plan increase, yet this result is not significant. To conclude, it seems as if Say on Pay votes lead to an increase in the pay for performance sensitivity, even though the votes are non-binding.

(27)

27 Discussion and Limitations This section discusses the limitations and disputable points of this study. Identifying the effect of a particular law is inherently difficult. The Say on Pay law was introduced with other regulations that were included in the Dodd-Frank Act. To capture the sole effect of the Say on Pay regulation using a Difference-in-Difference in method would have been ideal. Unfortunately, due to the lack of a control group, the Difference-in-Difference method was not possible in this research. This endogeneity issue is one of the significant limitations of this study. For further research, I recommend taking a larger sample and ensuring that the sample consists of firms with a public float smaller than $75 million.

Additionally, I followed prior research and set the benchmark for failed Say on Pay votes at 50% (Iliev & Vitanova, 2015; Ferri & Maber, 2013). Since the votes are non-binding, there is no official failure benchmark. It could be argued that 50% is a relatively low benchmark. It is a lot to have more than 50% of the shareholders to vote against a proposed compensation package. Another benchmark could be used in further research, such as 70%, to control for the effect of the Say on Pay votes. Another limitation of this study is the relatively small sample that could limit the possible generalization of these outcomes. The sample consists of 324 firms, while the sample that had to comply with the law is much bigger. It could be that firms with smaller capitalization show different results. However, given that my results are in line with other research, it is highly assumable that these results also apply for smaller capitalization companies. Yet another limitation of my research is the lack of ownership as a control variable. Several prior researchers obtained a negative relationship between CEO compensation and CEO ownership (Goldberg & Idson, 1995; Core et al., 1999). Unfortunately, the databases provided by the University of Amsterdam did not include CEO ownership, nor was it provided in the EDGAR database. Although, I do consider this as a limitation, I believe that because my results are in line with prior research, including CEO ownership as an independent variable would only influence the significance and R-squared, but not the outcomes itself. Moreover, it is unclear whether the effect of CEO ownership is negatively or

(28)

28

positively correlated to CEO compensation. On the one hand, CEO ownership could lead to powerful and entrenched CEOs, which leads to a lower degree of monitoring (Boone et al, 2004). On the other hand, CEO ownership could also lead to the presence of block holders, which may be associated with a higher degree of monitoring (Correa & LEl, 2013).

The potential results of the votes could also be biased, leading to another limitation of this study. Shareholders might have voted for the compensation package because of the possibility of negative publicity. Firms that fail their Say on Pay votes have been in the news negatively. Shareholders might keep this in mind when they vote, instead of looking at firm performance.

The lack of proof for the second hypothesis concerning the link between votes and Say on Pay vote results can be explained by theory suggesting that stock options increase voting dissent (Kimbro & Xu, 2015). On the contrary, Sadler and Conyon (2010) find that restricted equity compensation has the opposite effect. This incentive based payment increased the compensation risk for CEOs, and thus needs this payment to be balanced with higher total compensation. Together this could lead to a u-shaped of equity compensation on Say on Pay voting dissent. Further Research For further research I recommend using the same method, but having a larger sample and including the variable of CEO ownership. The results of such future research could support the generalization of non-binding Say on Pay votes. Furthermore, having a larger sample improves the chances of including firms with a public float of smaller than $75 million. This allows for the possibility of a control group and therefore the possibility of doing a Difference-in-Difference regression, which would mean that the true effect of the law could be further captured.

Another recommendation for future research is to do a cross-country analysis on non-binding Say on Pay votes. The comparison of these results could provide more insight of the effect of these votes. Finally, I advise including

(29)

29

References

Anand, S. (2011). Essentials Of The Dodd-Frank Act. New York: John Wiley and Sons. John Wiley &Amp; Sons

Atanasov, V. A., Bernard S. B., Conrad S., & Stanley B. G. (2010). How Does Law Affect Finance? An Examination of Equity Tunneling in Bulgaria. Journal of Financial Economics, 96, 155-173.

Baird, J., & Stowasser, P. (2002). Executive compensation disclosure requirements: the German, UK, and US approaches, PracticalLaw.com, PLC Document 4-101-7960, September 23.

Barth, J. R., Lin, C., & Wihlborg C. (2012). Research handbook on International Banking and Governance. ISBN: 978-1849802932

Bebchuk, L., Fried, J. (2003). Executive Compensation as an Agency Problem. Journal of Economics Perspective, 17, 71-92.

Bebchuk, L., Fried, J. (2004). Pay without performance: The Unfulfilled Promise of Executive Compensation. Harvard University Press.

Berrington, A., Smith, P. WF., & Sturgis, P. (2006). An Overview of Methods for the Analysis of Panel Data. ESRC National Centre for Research Methods, NCRM Methods Review Papers, NCRM/007.

Bertrand, M., & Mullainathan S. (2001). Are CEOs Rewarded for Luck? The Ones without Principals Are. The Quarterly Journal of Economics, 116, 901-932. doi:10.1162/00335530152466269

Cai, J., Garner, J., & Walkling, R. (2009). Electing directors. Journal of Finance, 64, 2389–2421.

(30)

30

Cai, J. and Walkling, R. (2011), Shareholders Say on Pay: does it create value? Journal of Financial and Quantitative Analysis, 46, 299–339. Conyon, M. J., & He, L. (2011). Executive compensation and corporate governance in China. Journal of Corporate Finance, 17, 1158–1175. Core, J.E., Holthausen, R.W., & Larcker, D.F. (1999). Corporate Governance, Chief Executive Officer Compensation, and Firm Performance. Journal of Financial Economics, 51, 371-406. Core, J., & Guay W. (1999). The Use of Equity Grants to Manage Optimal Equity Incentive Levels. Journal of Accounting and Economics, 28, 151–184. doi:10.1016/S0165-4101(99)00019-1

Correa, R., & Lel, U. (2013). The Effects of Say-on-Pay Laws on Executive

Compensation, CEO Pay Slice, and Firm Value around the World (Working

paper). Retrieved from Board of Governors of the Federal Reserve System: https://www.federalreserve.gov/pubs/ifdp/2013/1084/ifdp1084.pdf

Davis, S. (2007). Does ‘Say on Pay’ work? Lessons on making CEO compensation accountable. Retrieved from Millstein Center for Corporate Governance and Performance, Policy Briefing No. 1 (2007): http://www.shareholderforum.com/op/Library/20070620_Davis-briefing.pdf DeFusco, R., Johnson, R., & Zorn, T. (1990). The effect of executive stock option plans on stockholders and bondholders. Journal of Finance, 45, 617-627.

Del Guercio, D., Wallis L., & Woidtke T. (2008). Do boards pay attention when institutional investor activists ‘just vote no’. Journal of Financial Economics 90, 84-103.

(31)

31

Ertimur, Y., Ferri, F., & Stubben, S. (2010). Board of directors’ responsiveness to shareholders: Evidence from shareholder proposals. Journal of Corporate Finance, 16, 53–72.

Ertimur Y., Oesch D., & Ferri F. (2011). Shareholder Votes and Proxy Advisors: Evidence from the Say on Pay. Journal of Accounting Research, 51, 951-996. doi:1 0.1111/1475-679X.12024

Fernandes, N., Ferreira, M., Matos, P. & Murphy , K. (2013). Are US CEOs Paid More? New International Evidence. The Review of Financial Studies, 26, 323-67. doi: 10.1093/rfs/hhs122 Ferri, F., & Maber M. (2013). Say on Pay votes and CEO compensation: evidence from the UK. Review of Finance, 17, 527-563.

Froot, K. (1989). Consistent Covariance Matrix Estimation with Cross-Sectional Dependence and Heteroskedasticity in Financial Data. Journal of Financial and Quantitative Analysis, 24, 333-355. Gabaix X., & Landier. A. (2008). Why Has CEO Pay Increased So Much? Quarterly Journal of Economics, 123, 49-100 Gillan, S. & L. Starks. (2000). Corporate Governance Proposals and Shareholder Activism: The Role of Institutional Investors. Journal of Financial Economics, 57, 275 – 305.

Goldberg, L. G., & Idson, T. L. (1995). Executive Compensation and Agency Effects. The Financial Review, 30, 313-335. Gordon, J. (2009). “Say on Pay”: cautionary notes on the U.K. experience and the case for shareholder opt-in. Harvard Journal on Legislation, 46, 323-341. Harris, R. D. F., & Tzavalis, E. (1999). Inference for unit roots in dynamic panels where the time dimension is fixed. Journal of Econometrics, 91, 201–226.

(32)

32

Hartzell, J., & Starks L. (2003). Institutional investors and executive compensation, Journal of Finance, 58, 2351-2374.

Holmström, B. (1979). Moral Hazard and Observability. The Bell Journal of Economics, 10, 74-91. Hubbard, R. G., & Palia, D. (1995). Executive Pay and performance Evidence from the U.S. banking industry. Journal of Financial Economics, 39, 105-130. doi: 10.1016/0304-405X(94)00816-J

Iliev, P., & Vitanova S, WP. (2015). The effect of Say-on-Pay vote in the U.S (file name: SSRN-id2559181). Retrieved from SSRN: http://ssrn.com/abstract=2235064 Im, K. S., Pesaran M. H., & Shin, Y. (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics, 115, 53–74. Jensen, M. C., & Mecking W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305-360.

Jensen, M. C., & Murphy K. J. (1990). Performance Pay and Top-Management Incentives. Journal of Political Economics, 98, 225–264.

Jensen, M. C. & Murphy, K. (1990a). CEO Incentives: It is not how much you pay but how. Harvard Business Review, 68, 138-153

Kimbro, M. B., & Xu, D. (2015). Shareholders have an say in executive compensation: Evidence from the say-on-pay in the United States. Journal of accounting and Policy, 36(4), pp. Forthcoming.

(33)

33

Larcker, D. F. (1983). The assosciation between performance plan adoption and corporate capital investment. Journal of Accounting and Economics, 5(1), 3-30

Leone, A., Wu, J., & Zimmerman J. (2006). Asymmetric sensitivity of CEO cash compensation to stock returns. Journal of Accounting and Economics, 42, 167–192.

Levin, A., C. F., Lin, & C. S. J. Chu. (2002). Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics, 108, 1–24

Levit, D. & Malenko, N. (2011) Non-binding voting for shareholder proposals, Journal of Finance, 66, 1579–1614.

Liu, Y., & Mauer, D. C. (2011). Corporate cash holdings and CEO compensation incentives. Journal of Financial Economics, 102, 183-198.

Liu, Y., Miletkov, M. K., Wei, Z., & Yang, T. (2015). Board independence and firm performance in China. Journal of Corporate Finance, 30, 223-244.

Morgan, A., & Poulsen, A. (2001). Linking pay to performance – compensation proposals in the S&P500. Journal of Financial Economics, 62, 489-523.

Monem, R., & Ng, C. (2013). Australia ‘two strike’ rule and the performance link: Are shareholders judicious? Journal of Contemporary Accounting &

Economics, 9(2), 237-54.

Murphy, K. J. (1985). Corporate performance and managerial remuneration: An empirical analysis. Journal of Accounting and Economics, 7, 1142.

Roberts, M., Whited, T. (2011). Endogeneity in Empirical Corporate Finance. Unpublished Working Paper, University of Pennsylvania, University of Rochester.

(34)

34

Sadler, G., & Conyon, M. (2010) 'Shareholder Voting and Directors’ Remuneration Report Legislation: Say on Pay in the UK'. Corporate Governance - An International Review 18 (4), 296-312. Available from http://dx.doi.org/10.1111/j.1467-8683.2010.00802.x

Smith, C. W., & Watts R. L. (1992). The investment opportunity set and corporate financing, dividend and compensation policies. Journal of Financial Economics 32, 263-292.

Thomas, R. & Cotter, J. (2007). Shareholder proposals in the new millennium: Shareholder support, board response and market reaction. Journal of Corporate Finance 13, 368–391.

Thomas, R., Palmiter A., & Cotter J. (2011). “Frank-Dodd’s Say on Pay: Will It Lead to a Greater Role for Shareholders in Corporate Governance?” Working paper. Vanderbilt University Law and Economics Research Paper No. 11–49. Walker, D. I. (2010). The Challenge of Improving the Long-Term Focus of Executive Pay, 51 B.C.L. Rev. 435. Retrieved from http://lawdigitalcommons.bc.edu/bclr/vol51/iss2/3 Wooldridge, J. M. (2002). Econometric Analysis of Cross-section and Panel Data. Cambridge: M.I.T. Press.

Yermack, D. (1995). Do Corporations Award CEO Stock Options Effectively? Journal of Financial Economics, 39, 237 – 269.

Yermack, D. (2010). Shareholder voting and corporate governance. Annual Review of Financial Economics, 2, 103-125.

(35)

35

Appendix

Figure 1 - The outcome of the Say on Pay votes (2011-2014) Notes: These figures depict the outcomes of the Say on Pay votes in the years 2011-2014. The upper bar chart shows the percentages of failed Say on Pay votes in red and the percentage passed Say on Pay votes in blue. The below graph bar is the frequency graph of failed Say on Pay votes. 0 2 4 6 F re q u e n cy 2011 2012 2013 2014 2011-2014 S&P 500 Companies

Frequency failed Say on Pay vote 98.9691 1.03093 97.9167 2.08333 97.9592 2.04082 98.0198 1.9802 0 20 40 60 80 100 Pe rce n ta g e 2011 2012 2013 2014 2011-2014 S&P 500 Companies

Outcome Say on Pay vote

(36)

36 Figure 2 - Public Float distribution Notes: This figure represents the distribution of Public float in million dollar as of the last business day of its second fiscal quarter (2010) for the S&P 500 companies that were listed throughout the whole period. The highest bar depicts the frequency of firms with a public float >1 billion USD 0 100 200 300 F re q u e n cy 0 200 400 600 800 1000 publicfloat_1billion S&P 500 companies Public float distribution

(37)

37 Figure 3 - Compensation Ratio’s

Notes: This figure depicts the Equity and Cash ratio to total CEO compensation over the period 2007-2014. Equity compensation is the Ratio of stock awards plus option awards total compensation. Cash compensation is Ratio of cash plus bonus to total compensation. .1 .2 .3 .4 .5 .6 2006 2008 2010 2012 2014

Data Year - Fiscal

Equity Ratio Cash Ratio

S&P 500 companies

(38)

38 Table 1 – Variable description

Variable Variable definition Source

Total assets Total assets Compustat

ROA Return on Assets defined as income before extraordinary items divided by book value of total assets. Compustat Sales Growth Sales growth is log of the ratio of current sales to lagged sales. Compustat Firm Return Firm return is the 12 month return over the previous fiscal year in percentages. CRSP Market to Book ratio Common shares outstanding times the closing price plus total assets minus total stockholders’ equity and deferred taxes normalized by total assets Compustat Total Compensation Compensation received by CEO 10-k

CEO Age Age of the CEO DEF 14

CEO Tenure Numbers of years the CEO has been in the office DEF 14 Classified Board Dummy variable equal to 1 if just a fraction of the complete board of directors is elected each time DEF 14 Duality Dummy variable equal to 1 if the CEO has also the function of a chairman DEF 14

Cash ratio Ratio of cash plus bonus to total compensation DEF 14

Equity ratio Ratio of stock awards plus option awards total compensation DEF 14 Public Float Total Public equity float 10-k Votes For Votes in favour of the proposed total compensation of the CEO 8-k

Do say on pay votes affect the pay for performance link? : results from S&P 500 firms