0
A (Pay-) Slice of Basel III for European Banks
Abstract
In this paper I investigate the effects of Capital Requirement Directive IV (CRD IV) implementation on the remuneration of senior managers in European banks. Specifically, I analyze if CEOs’ pay slice,
defined as the fraction of CEO pay compared to the aggregate pay of the top 5 executives of a company, changed after CRD IV. The new regulation aims at reducing variable remuneration and I
therefore investigate if CEOs’ with a high pay slice maintain higher amounts of variable remuneration relative to the CEOs with low pay slice. This paper brings early and important evidence
on the remuneration effects of the most recent large-scale regulatory development in the European banking industry.
Student Chen Xie Pan
s1816926 Supervisor Vlad Porumb
Word count 8189
1
1 Introduction
The level of management compensation in the banking sector has attracted the attention of the media especially during recent years. The CEOs’ pay has been increasing in the recent years and some even argue it was one of the reasons for the 2008 financial crisis.
Even though new regulations came into existence the CEO pay continuously rose. According to the Wall Street Journal:” analysis found the top 10% by pay earned 23% of the total compensation, while
the bottom 30% accounted for just 13% of the total”(Francis and Lublin 2014). Aside from total
compensation increasing for CEOs’, the pay gap between them and senior executives increased as well. Previous research found evidence that the CEOs’ pay compared to the top 5 executives
increased in the 1997-2007 period (Peyers et al. 2007). For example, the CEO of Disney earns up to 1950 times the firm’s median salary while the average is about 204 to 1 (Che, 2014). This pay gap has a positive relationship with risk taking (Cooper et al. 2014) which was one of the reasons of the financial crisis in 2008. The difference in remuneration (further defined as CEO pay slice or CPS) is a measure that is likely to reflect the CEO’s relative power, significance or ability compared to the other top five executives.
In 2013, the European Union adopted a legislative package to strengthen the regulation of the banking sector and to implement the Basel III agreement in the EU legal framework.
This regulation contains the requirements for credit institutions and investment firms. The new regulation package applies as of 1 January 2014 and contains:
- Capital Requirements Regulation (CRR) - Capital Requirements Directive IV (CRD IV) - Q&As on CRD IV/CRR
- European Parliament webpage on CRDIV/CRR
This paper will focus on the implementation of CRD IV based on Basel III. It specifically targets the management who can leave a material impact on risk profile. Furthermore, Friedman (2008) argues that the high compensation of CEO’s does not reflect the marginal effect on income. Additionally, CEOs’ are often paid high salaries even though the firm’s performance is ordinary or below average. This paper analyzes the effect of the implementation of Basel III in the EU, through the adoption of CRD IV in Europe. Moreover, this paper analyzes if CEOs’ with a high CEO pay slice (further denoted as CPS), which is the CEO pay compared to the aggregate pay of the top 5 executives, experience the same effects of CRD IV compared to CEO’s with a low CPS. I analyze the effects on total remuneration as well as on the structure of remuneration. Consequently, my research question will be:
Did the implementation of CRD IV affect the CEO pay slice?
The implementation of CRD IV is intended to increase the resilience of the banking sector in
absorbing shocks from the economy. The CRD IV increases the amount of regulations in the banking sector and is specifically targeted at the remuneration of top management, personnel with high remuneration and employees who alter the risk profile. Because CRD IV has only been recently implemented in EU law, not much of its effect are known. Consequently, the effect of CRD IV has on CEOs’ remuneration is not researched and I intend to close this gap in the literature.
To explore this question, I will use the analytical model which analyzes if the CRD IV
implementation had an impact on the CEO pay slice and on the structure of CEO compensation. This papers contribution is fourfold. First, the empirical model extends the current literate in the following way: articles that have studied the changes in CEO compensation and CEO characteristics often focused on the CEO pay level but ignored the changes in the compensation structure. For example, Bebchuk et al (2011) only looks at how pay slice influences Tobin’s q defined by:
2 to the CEO, CEO turnover. Second, a lot of papers focus on the pay disparity of CEO’s as a whole. By splitting the CPS in a strong CPS and weak CPS it is possible to investigate whether strong CEO’s behave differently to the new regulation compared to weak CEO’s. Third, as the CRD IV has only been recently implemented, it’s effects are mostly unknown. This paper is one of the first to show how CRD IV influences the CPS and the structure of pay. Lastly, this paper has practical implications for policymakers and rule setters, as this paper researches the variable remuneration and CPS. It provides insight for policymakers if the most controversial topics of CRD IV reached its intended effect or whether they need to adjust their policies.
The rest of the paper is structured as follows. In section 2 I will discuss all relevant literature related to CPS. Section 3 shows the development of the hypothesis. In section 4 I discuss my methodology and statistical models. Section 5 presents the results for the analytics and finally section 6 summarizes my study and provides recommendation for futures studies and limitations of this study.
2. 1Literature Review
Scholars have argued that in the context of executive compensation that the power at the top of top executives often plays a significant role into determining in the pay setting process (Bebchuck et Fried 2004). Top manager’s power can be measured from different sources, such as the CEO’s formal position (CEO’s who are also chair members) and composition of the board (executives may have personal connections to management). In addition, CEO’s often have significant influence on
nominating who become director affect the board composition and the power CEO can exert over the board (Bebchuk et Fried 2004). A new way to measure the CEO power was developed by Bebchuk et al (2011) by measuring the CEO compensation as fraction of the aggregate compensation of the top 5 executives. One of the big advantages of this measurement is that CEO power becomes easily
quantifiable.
When you look at similar implementations of law that target executive remuneration than say on pay (SoP)is similar to CRD IV. Evidence from Correa and Lel (2014) show multiple results of the adaptation of SoP. First after the adoption of SoP laws the total amount of CEO remuneration compared to top management is lower which indicates a reduction in CEO pay slice. Secondly the firms in which CEO’s had a more equal pay slice compared to the top management had better firm valuations. And lastly they found evidence that with the adoption of SoP CEO remuneration was more linked to performance thus the pay sensitivity was increased. This evidence was even stronger for firms whom had problematic pay practices and weak governance before the adaptation of SoP laws.
Research from Frydman and Saks (2010) shows that there was a weak relationship between CEO pay and aggregate firm growth from 1940-1970. By contrast there is a significant larger correlation during the past 40 years. Their research also suggests that compensation arrangements for the CEO have often helped to align the CEO’s interests with the interest of the shareholders. Because their findings show that executive wealth is sensitive to firm performance in most of their sample. After the crisis, regulators aimed to implement regulations that would prevent another economic downfall.
The implementation of the Capital Requirements Directive IV (CRD IV) in the European Union in 2014 represents the implementation of Basel III capital accords. But there are also differences with Basel III. CRD IV has a strong focus on the remuneration of top management teams and people who take risk and that can influence the economic status of the firm. The most controversial aspects of CRD IV are that the variable remuneration is capped at 100% of the fixed remuneration. And that the new requirements only target the staff whose activities could have a material impact on risk profile. The variable remuneration can be raised to 200% if there is shareholders’ approval.
3 decreases. While this can be used as a proxy to measure CEO strength there are also differences between SoP and CPS. The main difference is SoP measures the CEO strength relatively towards the equity holders whilst CPS measures the CEO strength relatively towards the management. But there are also similarities, both measurements use the variable CEO compensation to measure the CEO strength.
Advocates of Say on Pay (the ability to vote on CEO remuneration) argue that the ability for shareholders to vote on CEO compensation empowers the board power to negotiate CEO
remuneration (Burns and Minnick 2013). This can potentially increase the alignment between firm performance and CEO remuneration as well as increased remuneration disclosure. Results from a meta-analysis performed by Mason et al (2014) show that the implementation of SoP did not influence the total amount of CEO remuneration which is also supported by research from Ferri and Maber (2013). But the meta-analysis from Mason et al. (2014) shows it does influence the composition of the CEO remuneration.
2. 2Theoretical Framework
The exorbitant remuneration of CEO’s can be best explained using the tournament model from Lazear and Rosen (1981). The basic model of Lazear and Rosen (1981) relies on increased pay at each hierarchical level of the organization with at the highest level the grand prize. These differences in pay are an incentive for the employees to perform better and win a tournament prize (higher pay). The higher the complexity of the tournament, the higher the payout will be for winning the tournament. One of the advantages of using a tournament model is that the high prizes are inherently motivating, the high rewards at the top end of the tournament model will result in reduced shirking, improved efforts and higher alignment with organizational goals (Henderson and Frederikson 2001).
But not all people play according to the rules. Because of the tournaments design that the winner takes all, there is a large incentive to reach the top. There is an increased incentive to sabotage the other contestants instead of increasing one’s own effort level. The theory of Münster (2007) says that contestants in an equilibrium who choose to for a higher performance level are sabotaged more heavily by their competitors. Retaining information, spreading gossips/rumors or transferring false information are one of the many ways to sabotage the competitors.
Tournament theory gives insight as to why there are such large gaps between pay levels, but according to tournament theory a manager who barely won a tournament is still deserving of the large pay gap. This study will look if there is an actual power difference between the different pay levels and to what extent. By extending on the CEO pay slice measurement from Bebchuck et al (2011) we can measure if there is an power difference through looking at the power of how they negotiate a variable pay. An alternative approach of understanding the CEO pay slice can be derived from the steam of literature dealing with managerial power. This approach has less emphasis on the needs of an organization and how that reflects on the pay packages. Instead it emphasis on CEO power. The greater the power of the CEO, the more he/she is able to influence her/his pay (Döscher and Friedl 2011). Another factor that influences the CEO power is the level of entrenchment that has taken place. The longer a manager stays in a complex business environment the more valuable and firm specific knowledge the manager obtains. Hence, the chance of replacing a manager decreases (Shleifer and Vishny 1989). Thus managers that are deeply entrenched have more leverage in dictating ones pay.
This study will mainly draw upon the agency theory to gain a better insight into the relationship between governance and CEO compensation. The conflict of interests between CEO’s and the
shareholders of co-operations is a classic example of agency theory. Jensen and Meckling (1976) state that there is an inherent struggle between the agents and the shareholders. The shareholders, hire managers as agents to increase the shareholders maximum wealth but the managers have incentives to
4 maximize their own personal wealth through perquisite consumption. The costs of perquisite
consumption are bared by the principals and are together with the cost associated with monitoring the agent’s part of agency costs.
“If shareholders had absolute and continual knowledge concerning the investment opportunities available to the firm and the corresponding CEO actions required to maximize firm value, they could design a contract which would specify such actions” (Pissaris et al 2010).
However, since it is impossible to have perfect knowledge about which investment opportunities are available to shareholders or to perfectly monitor the agent’s actions, agency theory suggests the principals should utilize compensation packages that align the agent’s interests with the principals interests and maximize shareholders value. One of the systems to control the companies is corporate governance. Laszlo et al (2015) say that corporate governance is used as a mechanism by shareholders to ensure that the interests of the managers are aligned with the shareholders. The board of directors are responsible for the governance of a company and the role of shareholders to appoint the board of directors and auditors. One of the roles of the board of directors is to supervise the management of the business and report to their shareholders on their stewardship.
“The board’s actions are subject to laws, regulations, and the shareholders in general meeting” (Laszlo et al 2015). Thus shareholders voting power is part of the corporate governance.
3 Hypothesis Development
To investigate whether the pay slice is affected by the implementation of CRD IV, I build a series of hypothesis. First, I investigate if the implementation had an overall impact on the CEO pay slice than I differentiate between CEOs’ with a large pay slice and small pay slice and analyse if the composition of their compensation is impacted differently by the new regulations.
There are two benefits for offering more compensation towards CEOs’ and increasing the pay disparity in top executive teams. First, the firm is able to attract more talented top executives and star players. From a tournament perspective, CEOs’ that have a larger pay slice are considered to be more talented, or are stronger compared to the other competitors as they have won the tournament and beat their competition (Henderson and Frederikson 2001). By offering a higher CPS the other members of the executive team are attracted to win the tournament and in turn more meritorious pool of CEO candidates are attracted to the firm. Second, the agency theory argues that by offering a large
compensation the interests of the principals and agents get aligned and acts as an effective governance mechanism.
The intended effect of CRD IV was to limit the growth of the remuneration try to keep the total amount of remuneration in boundaries. It immensely affects the variable remuneration as this is now limited to 100% of the fixed remuneration and may only be raised to 200% with shareholders approval. With the limitation of growth and the increased disclosure for high remuneration I expect the pay disparity to disappear between CEO and the top management team. As the increased disclosure of the remuneration of the CEO and top management team will show clear evidence of pay disparity, the social standing of top management team will also become more evident. Ridge et al (2014) also found evidence that individuals evaluate their rewards to compare their social standing. Thus I derive the following hypothesis.
H1: CPS decreases after the implementation of CRD IV.
By implementing the CRD IV regulation CEOs’ with a larger pay slice should be more heavily affected than CEOs’ with a smaller pay slice. Correa and Lel (2013) investigated what the effect was of implementing the SoP laws on CPS. They found that the total compensation decreased. But their
5 research didn’t say if CEOs’ with previously a large CPS were impacted differently than CEO’s with a small CPS. By analyzing if strong CEOs’ are impacted differently we expand on the literature of Correa and Lel(2013) if regulations impact CEO’s differently. I argue that strong CEO’s will be less impacted by the CRD IV proportionally by the implementation than weak CEOs’ as strong CEOs’ are more capable to argue that their pay is justified by using their talent.
H1.1: CEO’s with a high CPS pre CRD IV, are less affected by the implementation than CEO’s with a
low CPS pre CRD IV. Ceteris paribus
H1.2: CEO’s with a low CPS pre CRD IV, are more affected by the implementation than CEO’s with a
high CPS pre CRD IV. Ceteris paribus
The basic tournament model of Lazear and Rosen (1981) relies on increased pay at each hierarchical level of the organization with at the highest level the grand prize. These differences in pay are an incentive for the employees to perform better and win a tournament prize (higher pay). From a tournament perspective, contestants who are vying for a CEO position (grand prize) enter in a tournament like competition in which contestants compete for a prize and expand increasing effort to increase the likelihood of winning a prize. In the tournament the importance isn’t put on the absolute level of performance but the importance is put on how well you perform in relation to your other competitors (Lazear and Rosen 1981). A weak CEO will be less likely to enter a tournament in which there is a lot of strong competition as this will diminish his chance of winning a prize. I theorize when the CEO has a smaller pay slice he has entered a tournament with a weaker competition. To avoid drawing more competition to a new tournament he will be less likely to use methods to increase his pay by a significant amount. Or alter their compensation that relies more on variable as this is often contingent on the CEO and firm performance. But a weak CEO is less likely to perform and obtain all of its variable pay.
Furthermore a weak CEO will be less likely to be able to exert his power and the top management team will be more likely in close pursuit in terms of power and remuneration. A strong CEO will be able to show there is a clear difference in power between the top management team and the CEO. A strong CEO will be less likely to collaborate, or collude with the top management team as this will decrease his chances of staying at the top of the tournament. Thus a strong CEO will be more likely to influence its pay. But because of the implementation of CRD IV the mandatory variable remuneration is 100% of the fixed remuneration and if the shareholders agree the cap can be raised to 200%. A strong CEO will be more likely to influence the shareholders to raise the cap.
Research from Burns and Minnick (2013) shows that one of the effects of implementing say on pay is that CEO’s compensation is altered. The pay is altered in such a way that it shifts from monetary compensation towards more incentive compensation.
H2.1: CEOs with high pay slice in the pre-CRD IV period are going to have higher levels of
percentage variable pay in the post-CRD IV period.
H2.2: CEOs with low pay slice in the pre-CRD IV period are going to have lower levels of percentage
6
4 Methodology
Sample construction
My original sample comes from the ExecuComp database. The ExecuComp database includes detailed historical data on remuneration for executives. It includes information on total pay as well as a
breakdown of its components such as salary, bonus, and equity pay. My sample consists of banks that are registered in the EU and are subject to the Basel III law and are required to implement the CRD IV accords from 1 January 2014 onwards. I use the compensation data retrieved from the database to at the remuneration of personnel who can have a material impact on risk management, as well as the manually compute the CPS. CPS is calculated as follow: First, I compute the total annual
compensation of the executives as the sum of fixed compensation and variable compensation (including stock options). Next, I divide the total annual compensation of the CEO by the aggregate compensation of the top 3 or top 5 executives. I eliminate firms whose accounting and financial variables are not available on COMPUSTAT, furthermore I eliminate observations of management that show no remuneration. Since I’m primarily interested on the effects of CRD IV on the CEO pay slice, I only keep around where the top executive of the firm is identifiable and excludes observations that show 2 top executives for a given year. I than merge this dataset with the dataset from Bankscope. This provide commonly used firm identifiers such as ROA, total assets, total debt and sales. The final sample consists of 2320 firm-year observations with 63 unique firms from 19 countries ranging from 2008-2014. Countries included are countries from Scandinavia and West-Europe who have a strong economic performance, countries from South-Europe who experienced large reforms due to the financial crisis and countries from East-Europe who are developing their economy. My sample is one of the most comprehensive and most recent in the literature in this area.
Measuring the effect of CRD IV on CPS (Hypothesis 1)
Following Bebchuck et al (2011) I define the CEO pay slice as the percent of total CEO compensation compared to the total compensation of the top 5 executives as well as the top 3 executives. This is different from Bebchuck et al that only looked at the top 5 executives. Using CPS measured by the top 3 executives will show a more clear difference between power at top management level. For the total compensation we use the data provided by EXECUCOMP and can be defined as the sum of salary, bonus plus other non-equity-based long-term incentive payments, stock-based compensation including restricted shares and stock options granted during the year, and other compensation payout not
separately disclosed anywhere else. In addition, I also decompose the compensation of the CEO and other top executives into the part that is variable based (i.e., from options, bonuses, stock grants, everything that isn’t included in the base salary) and the remainder (base salary that isn’t variable) component. I measure hypothesis 1 using an OLS regression. I measure hypothesis 1.1 and 1.2 by using an OLS regression as well. To measure hypothesis 2 I perform the same analysis as for hypothesis 1. But instead of using CPS as a dependant variable I use the variable pay of CPS as the dependant variable. I also include an interaction term (CRD IV * CPS).
c. Statistical Model
In order to investigate whether CEOs with a high CPS are affected differently by the implementation of CRD IV, I regress CPS on the implementation of CRD IV, firm-specific variables and the CEO tenure. I use model 1 (where CPS is measured for top three executives in model 1 and top 5 executives in model 2) to predict H1:
7
CPS 3 = β*CRD + β*LOG+ β*LEV+ β*PERF (1)
CPS 5 = β*CRD + β*LOG+ β*LEV+ β*PERF (2)
Where CPS is the CEO pay slice measured as the percentage of the total compensation of top-three or top-five executives that is captured by the CEO. CRD is a dummy variable that equals one for the time period in which CRD IV has been implemented and 0 otherwise. LEV is the firms ratio of long term debt to its total assets. PERF is a proxy for the firms performance measured as the return on average assets. If my prediction of the implementation of CRD IV holds, I expect CPS to be negatively associated with CRD IV, because this new regulation will put stricter disclosure and requirements on remuneration.
Model 1 and 2 measures the effect of the implementation of CRD IV on CPS and uses firm specific characteristics as control variables. I expect to find that the implementation of CRD IV has a
significant negative influence on the CPS when measured by the top 5 executives. I expect to find that this influence is even stronger when looking at CPS for the top 3 executives.
I rerun this model again for hypothesis 1.1 and 1.2 but change CPS by removing all values when CPS is > than the median and leaving the other CPS values intact. Afterwards I rerun the same model but remove all values when CPS < than median and leaving the other values intact. Thus I’m adding a dummy-variable for CPS. Which takes 1 for CPS that are higher than the median and 0 for CPS that is lower than the median.
To measure hypothesis 1.1 and 1.2 I will use model 2 and 3 respectively. Model 2 and 3 measures the different effects that the implementation has on CEOs’ with a large pay slice and CEOs’ with a small pay slice. I expect that CEOs’ with a large pay slice are affected differently by the implementation of CRD IV. As strong CEOs’ with a large CPS have more ability to exert power and dictate their pay. I expect CRD IV to have a more significant negative influence on CEO’s with a large pay slice compared to CEOs’ with a small pay slice.
DCPS 3 = β*CRD + HCPS+ β*LOG+ β*LEV+ β*PERF (3)
DCPS 5 = β*CRD + HCPS+ β*LOG+ β*LEV+ β*PERF (4)
Variable pay = β*CPS+ β*CRD + CPS*CRD + β*LOG+ β*LEV+ (5) Total pay
Finally when measuring hypothesis 2 I use model. Model 5 measures the effect that CRD IV has on the composition of compensation. It also measures the effect CRD IV has on the composition of compensation as well if the composition of compensation is affected differently by CRD IV for CEOs’ with a large pay slice compared to CEOs’ with a small pay slice. I expect that the variable
compensation for CEOs’ with a large pay slice will be more negatively affected by the
implementation of CRD IV than CEOs’ with a small pay slice. As the implementation of CRD IV largely influences variable compensation and strong CEOs’ are more likely to demand variable compensation based on performance.
8
Control variables
Firm size
Main et al (1993) investigated the pay structures of CEO’s and found that within their sample the pay disparity increases with the amount of contestants between organizational levels. This is consistent with the tournament theory that predicts that a larger amount of players increases the grand prize. Higher levels of internationalization, diversification and sheer firm size increases the complexity of a firm. As the complexity of a firm increases the tasks of top management come more to rely on each other and interdependent (Siegel and Hambrick 2005). But to run a complex firm successfully there is a need for top management teams to collaborate. However with the increased collaboration the task interdependence and join-decision making also increases which makes it more difficult to monitor the top management teams (Eisenhardt 1989). Furthermore, complexity makes it difficult to accurately measure the performance of individuals of a top management team as tasks become intertwined. As a result pay spread increases due to ambiguity of responsibilities and tasks of top management teams.
Firm performance
Bebchuck et al 2011 found evidence that CPS is higher when the ROAA of the prior year is higher. This could be explained due to the fact that the CEO compensation is more sensitive toward
performance compared to the top management team. Also the study from Main et al (1993) shows that there is a link albeit a weak link between the firm’s performance with wage dispersion and positively associated with the mean salary of top management. This is in line with Lazear (1989).
Firm leverage
While literature is missing on why firm leverage impacts CPS. Bebchuck et al 2011 did find a positive correlation between firm leverage and CPS. A possible explanation that they give is while leverage is viewed as costly to the members of top management teams. A default on these leverages could result in disproportionate loss in reputation. This might be even more costly to the CEO. Thus the CEO might need more compensation to compensate for the increased risk.
9
5.1 Results
Summary Statistics
Table 1 displays the descriptive statistics.. The mean CEO pay slice for top 5 executives is 0.301
which is close to the average CPS of 0.357 in Bebchuk et al. (2011). This suggests that CEO’s, on average, represents 30% of the total pay for the top 5 executives and 41% for the top 3 executives. The average salary is comprised for 43,5% of variable salary and 56,5% for fixed salary. The average firm has 55% liabilities compared to the assets and has an ROAA of -2% which indicates that firms performance wasn’t that good.Table 1 Descriptive statistics.
Variable Mean Std. Dev Min Max
CPS 3 0,410 0,120 0,029 0,809 CPS 5 0,301 0,112 0,026 0,755 DCPS 3 0,460 0,499 0,000 1,000 DCPS 5 0,410 0,494 0,000 1,000 Vcom 0,435 0,353 0,000 1,000 Assets 12,848 0,897 10,576 14,401 LEV 0,558 0,213 0,007 1,230 ROAA -0,206 9,703 -149,106 17,471 CRD IV 0,111 0,314 0,000 1,000
CPS 3 and CPS 5 is the CEO pay Slice measured as the percentage of the total compensation of top-three, top-five executives respectively that is captured by the CEO and CPS ; DCPS 3 and DCPS 5 is a dummy variable measured as the value 1 for values higher than the median CPS and 0 for values lower than the median respectively.Vcom is the variable based compensation measured as the total variable compensation divided by the total compensation. Assets is the natural logarithm of the total assets. Lev is the total debt to total asset ratio. ROAA is the return on average assets measured as the net income divided by the net total assets. CRD IV is a dummy variable that takes one for the years that CRD IV is implemented (2014) and, zero otherwise.
Table 2 shows the pair wise relation between the variables. Implementation of CRD IV is negatively related to CPS which is in accordance with my expectation. As I expected that CRD IV would negatively impact CPS due to increased regulation for high remuneration. The size of the firm is negatively related to CPS. Firm performance is significantly and positively related to CPS which is expected as powerful CEO’s are more capable of leading a firm and influencing firm performance. These powerful CEO’s are thus also more likely to influence their pay. Finally leverage is positively related to CPS which is in accordance with my expectations as higher leverage shows a higher risk for reputational damage as the firm is more likely to get in financial distress.
10 Table 2 Correlation matrix
1 2 3 4 5 6 7 8 9 CPS 3 1 CPS 5 0,918** 1 Vcom 0,166 0,122 1 Assets -0,06 -0,061 -0,079** 1 LEV 0,039 0,023 -0,079** -0,150** 1 ROAA 0,188** 0,156* 0,159** -0,446** -0,016 1 CRD IV -0,004 -0,040 0,006 0,028 0,024 0,057** 1 DCPS 3 ,732** ,713** ,162** -,025 -,064 -,009 ,031 1 DCPS 5 ,685** ,775** ,120 -,065 -,075 -,044 -,068 ,759** 1
***, **, * indicate coefficient is significant at 1, 5, and 10% levels respectively
5.2 Effects of CRD IV on CPS|
In this section I investigate whether CRD IV impacts the CPS as a whole or not. The issue of CPS has been investigated in previous studies for example Chongwoo et al (2014) found evidence that the pay of CEO is related to the power of the CEO by measuring the CPS. Following the model of Bebchuk et al (2011) I use CPS as a dependent variable and assets, leverage and firm performance as a control variable. I add CRD IV as a dummy variable to the model. Table 3 panel A shows the results of the linear regression model. The data was subjected to multiple linear regression for CPS against the independent variables. I used Regression analysis because provides the relationship between two or more variables and also information on the strength of the relationship.
Table 3 Panel A Effects of CRD IV on CPS
Dependent variable: CPS Model 1 Model 2 CRD IV -,046 -,042 (,022) (,027) Assets ,010 -,007 (,008) (,010) LEV -,001 ,002 (,034) (,040) ROAA -,010 ,001 R2 (,001) ,002 (,001) ,002
11 In both models CRD IV has a negative coefficient but the impact is insignificant. This implies that when CRD IV was implemented CEOs’ salary decreased relative to other top executives. As the results are insignificant we can reject hypothesis 1 and hypothesis 2. I find that once CRD IV is introduced CEO remuneration share declines with 4,6% and 4,2% for CEOs’compared to the aggregate compensation of the top 3 and top 5 executives respectively. When holding remuneration for other top 3 and 5 executives constant, CEO remuneration decreases with 13,8% and 22% respectively. A possible explanation why CRD IV didn’t have a significant impact on CPS is that CPS measures the relative difference in compensation. CRD IV could have affected the compensation of all executives proportionally which would indicate no change in CPS. The control variables show no significant impact on CPS which is in accordance with the study of Bebchuck et al (2011) it is also in line with findings from Al-Najjar et al (2016) whom also didn’t find any significant evidence for said control variables.
Because CRD IV is aimed at the remuneration of not only the CEO but of everyone in a firm that can leave a material impact on the risk profile. I replace the dependent variable CPS for total
compensation paid and test if the implementation of CRD IV has an influence on the remuneration. Table 3 panel B shows the results of the regression.
Table 3 panel B Effects of CRD IV on
remuneration
Dependent variable: total
compensation Model 1 CRD IV ,127*** (,042) Assets ,158*** (,015) LEV -,657*** (,064) ROAA -,003* R2 (,001) 0,111
***, **, * indicate coefficient is significant at 1, 5, and 10% levels respectively Total compensation is measured as the logarithm of compensation received in a year.
Contrary to the findings of measuring CPS, we see that this model has significant results for all variables used. However this model only explains 11,1% (R=0,111) The coefficient of CRD IV is positively significant (,127 t=3,035 p<0,01) which is against my expectations. This shows that when CRD IV is implemented, remuneration increases on average with the natural log(,127). I would expect that CRD IV has a negative impact on the remuneration as it subjects the remuneration to certain rules. A possible explanation could be that CRD IV doesn’t restrict the height of the remuneration but restricts the composition of the remuneration as 200% variable remuneration is only permitted with
12 the shareholders’ approval. The coefficient of firm size is positively significant with remuneration (,158 t=10,366 p<0,01) and shows that remuneration increases with log(,157) for every increase in assets measured by the natural logarithm. this relationship has been extensively researched and has evidence from multiple studies Volker (2007). And it can be explained through tournament theory, as firm size increases, the competitors increases, as well as the tournament prizes. The coefficient of leverage is negatively significant with total remuneration (-,657 t = -10,340 p<0,01). When looking at the firm performance on the total remuneration we see that firm performance has a negative
coefficient (-0,003 t=-,198 p<0,1). An increase in ROAA of 1% would decrease remuneration with -0,003%. While this coefficient is negative, it barely influences the total amount of remuneration and can almost be neglected. The results from this variable are comparable to the study from Haeney et al (2010) who also didn’t find any evidence of a positive relation between CEO remuneration and following year performance.
Firm leverage is negatively associated with total remuneration (-,657 t=-10,340 p<0,01). For every percentage increase in firm leverage, remuneration decreases with log(-,00657). This is against my expectation and against the results of studies such as Chemmanur et al (2013) who found evidence that suggests that firm leverage is positively related to the cash, equity and total compensation for CEO’s and average employees. They explain that employees ask for a higher wage in firms with a high leverage due to increased financial risk for the firm and thus a higher chance that the employee loses his job. A different theory given by Jensen (1993) provides a possible explanation why remuneration decreases when leverage increases. He argues that the leverage of the firm determines the available cash flow of the firm and its ability to meet contractual obligations, therefore if the leverage is high the executives remuneration will be reduced.
Overall my findings for CPS are in accordance with my expectations even though they are
insignificant, while the findings for total compensation are at odds with the intentions of CRD IV.
5.3 Effects of CRD IV on strong and weak CEO’
In this section I investigate whether CRD IV impacts CEO’s with a large CPS differently compared to CEO’s with a small CPS. Table 4 reports the results for the regression analysis used to measure the differences of effects between strong and weak CEO’s. The analysis performed is the same as in table 3 panel A and B. I change the dependent variable for CPS but I make a separation between CEO’s with high CPS and low CPS. Model 3.1 will report the regression of low CPS captured by the top three executives and model 3.2 the high CPS.
Table 3 Effects of CRD IV on CPS Dependent variable CPS Model 3.1 Model 3.2 Model 4.1 Model 4.2 CRD IV -,031 ,025 -,016 0,022 (,025) (,021) (,021) (,038) Assets ,001 ,009 -,009 -,023* (,009) (,008) (,008) (,013) LEV -,078* 0,40*** ,023 ,101** (,040) (,030) (,030) (,042) ROAA -,001 ,002 ,000 ,000 (,002) (,000) (,000) (,001) R2 ,018 ,112 -,063 ,133
13 executives and model 3.2 the high CPS. The above model shows that the explanatory factor for model 3.1 and model 4.1 are neglectable small (R2 = 0,18 and R2 = -,063). These two models use low CPS as dependent variable (CPS below the median). The explanatory factor for CPS is stronger when looking at high CPS (model 3.2 R2 = ,112 and model 4.2 R2 = ,133). The coefficient for CRD IV and low CPS is negative but insignifant (-,031 and -,016) which would suggest that CEO’s that have a smaller pay slice are negatively influenced by the implementation of CRD IV. This implies that the implementation of CRD IV reduces weak CEOs’share of the aggregate compensation with -3,1% and -1,6% respectively. When holding the compensation of the other top 5 executives constant, CEO compensation decreases on average with 9,3% and 8% respectively. This is in accordance with my expectation and with the results from table 3 panel A. When looking at the coefficient for CRD IV and high CPS, the results are different. I notice a positive coefficient (0,025 and 0,022) even though it is insignifant. The implementation of CRD IV actually increases strong CEOs’ share of the aggregate compensation with 2,5% and 2,2% when measured for top 5 executives and top 3 executives
respectively. This result would suggest two things. First CEO’s who have a large CPS are affected differently by CRD IV than CEO’s who have a low CPS. Second because the coefficient is positive this would suggest that CRD IV doesn’t impact the power balance at all, in fact it actually suggests that strong CEOs’ have a way to circumvent the new regulation and increase his share of aggregate compensation.
The coefficient for leverage is positively significant (,040 t = 3,915 p < 0,01 and ,101 t = 2,395 p < 0,05) for CEO’s with a high CPS (measured for aggregate top 3 and top 5 executives respectively). For every percentage increase in leverage, variable compensation increases with 0,04% and 0,1% respectively. This finding differs largely from my findings when regressing for the whole CPS population where I found no significant evidence for leverage and CPS. Furthermore the relation between leverage and CEO’s with a weak CPS is negative (-,078 and ,023) and significant at the p <0,1 level (t = -1,825) and p < 0,01 level ( t = 3,915). For every percentage increase in leverage, variable compensation decreases with 0,078% and increases with 0,023%. This suggests that CEO’s with a weak CPS behave differently than CEO’s with a high CPS depending on the capital structure of the firm. A possible explanation could be derived from the model of Titman (1984). He created a model in which he argues that when a firm has increased risk to bankruptcy than the employee will demand a higher wage as compensation. I argue that in the case of highly leveraged firms, only the more skillful employees will stay behind. Due to the increased risk of the firm going bankrukpt, more capable people will stay at the firm to steer away from bankruptcy. Because these people are more skillful the CEO’s is less able to exert power and prove that he deserves a larger pay. Overall my findings are not outside my expectation. While my findings are insignificant they do show that CEO’s with a high CPS behave differently than CEO’s with a low CPS. My findings suggests that CRD IV barely impacts strong CEO’s, and that strong CEO’s are able to demand a higher wage in highly leveraged firms due to their increased risk of losing their job and because they perform better
5.4 Effects of CRD IV on variable remuneration
In this section I investigate the relation between CPS and the variable remuneration. I also investigate whether the implementation of CRD IV influences the variable compensation of strong CEO’s differently compared to weak CEO’s. I add a new dependent variable: variable remuneration which is the total variable compensation divided by the total overall compensation. I also add a dummy variable for strong CEO’s which takes the value of 1 for CPS which were higher than the median and 0 for CEO’s with CPS lower than the median. I also add an interaction term that measures the interaction between CRD IV and the dummy variable CPS. Results of this regression can be found in table 5. Model 5.1 uses a dummy variable for CPS measured by the top three executives and model 5.2 uses a dummy variable for CPS measured by the top five executives.
14 Table 3 Effects of CRD IV on variable remuneration
Dependent variable: variable remuneration ratio Model 5.1 Model 5.2 CRD IV ,046 0,096 (,066) (0,073) Assets ,053*** ,069*** (,018) (,022) LEV -,535*** -,603*** (,074) (,091) ROAA -,002 -,002 (,002) (,002) DCPS ,094*** ,080* (,036) (,042) DCPS*CRD IV -,002 -,018 (,096) (,137) R2 ,257 ,239
***, **, * indicate coefficient is significant at 1, 5, and 10% levels respectively
Variable remuneration is measured as the total variable compensation divided by total overall compensation. DCPS is a dummy variable that takes the value 1 for CEO’s with a high CPS and 0 for CEO’s with a low CPS and DCPS*CRDIV is an interaction term.
The above models 5.1 and 5.2 explain almost 25% of the variable remuneration. When looking at one of our variables of interest CRD IV the results show no significance relation between CRD IV and variable remuneration. The coefficient is positive which is against my expectations. As CRD IV limits the variable remuneration to 100% of fixed remuneration and only allows it to reach 200% when shareholders vote and agree to it. Even though the result is insignificant it does suggest that variable remuneration isn’t largely affected by the new regulation.The coefficient for DCPS is positively significant (0,094 t = 2,608 p<0,01 and ,080 t 1,900 p<0,1) at the 1% and 10% level. This suggests that there is a positive relation between CRD IV and strong CEO’s which is in accordance with my expectations. As variable remuneration is often performance based and strong CEO’s are likely more able to exert more power or to perform better and thus obtain a larger variable remuneration. The coefficient of 0,094 indicates that a strong CEO earns on average 18,8% more variable remuneration compared to a weak CEO.
Leverage has significantly negative relation with CRD IV (-,535 t = -7,256 p<0,01 and -,603 t = --6,611 p<0,01) at the 1% level and can be explained using the same theory used for explaining
leverage in the total compensation model. An increase of 1% in the leverage of the firm decreases the variable remuneration on average with -1,1% (holding fixed remuneration constant). The relation between firm size and variable remuneration is also positively significant at the 1% level (,053 t = 2,908 p<0,01 and 0,069 t = 3,064 p<0,01). Which suggests that on average an increase of log size (1) increases variable remuneration with 10,6% and 13,8% respectively (holding fixed remuneration constant). What is notable is that variable remuneration increases more than fixed remuneration. I would expect that variable remuneration would increase proportionally to the fixed remuneration when firm size increases. A possible explanation why this isn’t the case could be that large firms are harder to monitor for shareholders and stakeholders. Offenberg (2010) argues that large firms with managers that impose excessive agency costs upon their shareholders willincur a size discount. This size discount is the negative relationship between a firm’s size and its value. I argue that large firms in
15 an effort to prevent this size discount to occur offer their managers a larger variable pay which would reduce the agency costs. Regarding the interaction term we see that the relation is insignificant thus we can conclude that the implementation of CRD IV did not significantly influence the percentage of variable compensation for strong CEO’s or weak CEO’s.
6 Conclusion
Following the recent financial crisis, the implementation of Basel III and the CRD IV accords are part of one of the most extensive regulations that have been implemented in our current times. The new regulations aims to increase the resilience of the financing sector in absorbing economic shocks. One of the constituents of CRD IV is the regulation that limits the remuneration of key personnel that can have an material impact on the risk profile. Through my analysis of remuneration data from
EXECUCOMP on key personnel from European banks, using CPS as a measurement I investigate whether CRD IV influences CEO’s differently based on their earnings as part of the top three or top 5 executives. I have found evidence that helps to paint how CRD IV influences the remuneration as well as the composition of remuneration. I found no significant evidence for the relation between CPS and CRD IV on the fixed remuneration or the variable remuneration. However I did find significant evidence that CPS significantly influences the composition of remuneration and that strong CEO’s will have their compensation based more on variable remuneration than on fixed remuneration, on average strong CEO’s earnings consists of 18,8% more variable remuneration than weak CEO’s. Furthermore I also didn’t find significant evidence that CRD IV influenced the total remuneration or total variable remuneration which begs the question whether CRD IV was influential at all and if CRD IV was only a measure to please the public.
What explains the fact that CEO’s with a high CPS have a percentage variable remuneration? First off CEO’s with high CPS are CEO’s that compared to the other executives are often more powerful or have a better performance. Second variable remuneration is often based on performance and reliant on reaching certain benchmarks. CEO’s that are strong or high performant are more capable of meeting these benchmarks and thus demand a higher compensation as variable remuneration.
My study is one of the first to show how new regulations regarding remuneration influences the power balance in top management firms measured by the CPS. However my study is not without caveats. My largest caveat is the range of my data sample, the remuneration data collected range from CEO’s to unnamed personnel, board members and even secretaries. Future studies could improve on this aspect by measuring CPS as a set measurement by comparing the CEO pay slice to set positions in the company e.g. CFO, CTO, head of the audit committee, etc. This would increase the stability of the study and improve its predictability. A second caveat is that this study only uses one year to measure the effects of CRD IV. A time series study would be able to research the effect of CRD IV to a greater extent.
In my study I also found new areas of interest that could be researched in future studies. I excluded the firms with dual CEO’s from my data sample. It would be interesting to research how the power dynamics work in firms with a dual CEO’s measured by the pay slice. Another area of interest is the variable remuneration that strong CEO’s receive compared to weak CEO’s. From my analysis I find that strong CEO’s remuneration has a significantly larger part for variable remuneration. But in my study I found no evidence that supports firm performance influencing variable remuneration. Thus I would recommend future studies to look at if strong CEO’s receive more variable remuneration but it’s not based on firm’s performance. Than what is the basis for their larger variable remuneration.
16
References
Adhikari, H. Bulmash, S. Krolikowski, M, Sah, N (2015). “Dynamics of CEO compensation: Old is gold”. The Quarterly Review of Economics and Finance. 191-206
Al-Najjar, B. Ding, R. Hussainey, K.” Determinants and value relevance of UK CEO pay slice”. International Review of Applied Economics. 403-421
Bebchuck, L. Fried, J,M. (2004).” Pay without performance: The unfulfilled promise of executive compensation”. Cambridge, MA: Harvard University Press
Bebchuk, L. Cremers, M. Peyer, U. (2011). “The CEO Pay Slice”. Journal of Financial
Economics 199-221
Burns, N. Minnick, K. (2013). Does Say on Pay Matter? “Evidence from Say on Pay Proposals in the United States”. Financial Review. 233-258.
CFA institute: Capital Requirements Directive IV
Che, J. August (2014). “Here's How Outrageous The Pay Gap Between CEOs And Workers Is”. Huffington Post
Chemmanur, T. Cheng, Y. Zhang, T.(2013). “Human capital, capital structure, and employee pay: An empirical analysis”. Journal of Financial Economics. 478-502
Cooper, E. Uzun, H. Yudan, Zheng (2014).” Pay Gap, Risk-taking, and the Financial Crisis”.
Banking and Finance Review. 55-74
Conyon,M. Sandler, G. (2010).
“
Shareholder Voting and Directors' Remuneration Report Legislation: Say on Pay in the UK”. Corporate Governance: An International Review. 296-31217 Valuation around the World”. U.S. Federal Reserve Board’s International Fin. Issue
1083/1084
EBA Report on Remuneration and Allowances
Denis, D. J., Denis, D. K., & Atulya, S. (1997). “Ownership structure and top executive turnover”. Journal of Financial Economics, 193-221
Döscher, T. Friedl,G. (2011), “Corporate Governance, Stakeholder Power, and Executive Compensation”, OR Spectrum,309–331.
Eisenhardt, K,M. (1989), “Agency Theory: An Assessment and Review”, The Academy of
Management Review. 57–74.
Francis, T. Lublin, J (May 2014) “CEO Pay Rises Moderately; a Few Reap Huge Rewards”.
Wallstreet Journal.
Ferri, F. Maber,M. (2013). “Say on pay votes and CEO compensation: evidence from the UK”. Review of Finance. 527-563.
Frydman, C. and Saks, S. E. (2010). “Executive compensation: A New View From a Long-Term Perspective”.The Review of Financial Studies.2099-2138.
Friedman, M (2008).” Living wage and optimal inequality in a Sarkarian framework”. Review of Social Economy. 93-111.
Henderson A, Fredrickson J. (2001) . “Top management team coordination needs and the CEO pay gap: a competitive test of economic and behavioral views”. Academy of
Management Journal.96–117.
Hill, C. Phan, P. (1991).”CEO Tenure as a Determinant of CEO Pay”. Academy of
Management Journal. 707-717
Huson, M.; R. Parrino; and L. Starks(2001). “Internal Monitoring Mechanisms and CEO Turnover: A LongTerm Perspective”. Journal of Finance.2265-2297
18 Jensen, M,C. Meckling, W.H. (1976). “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure.” The Journal of Financial Economics. 305- 360.
Jensen, M.C. & Meckling, W.H. (1976), “Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure”, Journal of Financial Economics, 305-360
Laszlo, T. Scott, G. Gerard, G. (2015).”Rethinking Governance In Management Research”.
Academy of Management Journal. 1-9
La Porta, R. Lopez de Silanes, F. Shleifer, A. Vishny, R. (1998). “Law and Finance”. Journal
of Political Economy. 1113-1155
Lazear, Edward P. and Sherwin Rosen (1981). “Rank-Order Tournaments as Optimum Labor Contracts”. The Journal of Political Economy. 841–6
4
Main, B,G,M. O’Reilly, A,C. Wade,J (1993). “Top Executive Pay: Tournament or Teamwork”.
Journal of Labour Economics. 606-628
Münster, J.(2007).Selection Tournaments, Sabotage, and Participation” Journal of
Economics & Management Strategy. 943–970.
Offenberg, D. (2010).” Agency Costs and the Size Discount: Evidence from Acquisitions”.
Journal of Economics, Finance and Administrative Science. 73-93
Palomino,F. Peyrache, E. “Internal versus External CEO Choice and the Structure of Compensation Contracts”. Journal of Financial and Quantitative Analysis. 1301-1331
Peyers, U,C. Martijn-Cremers, K,J. Bebchuck, L, A.” Pay Distribution in the Top Executive Team”. American Law & Economics Association Papers. 1-52
Pissaris, S. Jeffus, W. Gleason, K,C. (2010). “The Joint Impact of Executive Pay Disparity and Corporate Governance on Corporate Performance. Journal of Managerial Issues. 306-329
19 Heaney, R. Tawani, T. Goodwin,J (2010).” Australian CEO Remuneration”. Economic
Papers. 27-109
Ridge, J, W. Aime, F. White, M, A. (2014). “When Much More Makes a Difference: Social Comparison and Tournaments in the CEO’s Top Teams”. Strategic Management Journal. 618-636
Shleifer, A. Vishny, R,W. (1989), “Management Entrenchment: The Case of Manager-Specific Investments”, Journal of Financial Economics. 123-139.
Titman, S., 1984. “The effect of capital structure on a firm's liquidation decision”. Journal
of Financial Economics 13, 137–151
Siegel, P. Hambrick, D, C. (2005), “Pay Disparities Within Top Management Groups: Evidence of Harmful Effects on Performance of High-Technology Firms”, Organization
Science 16, 259–74.
Volker, G (2007). “Firm Size, Productivity, and Manager Wages: A Job Assignment Approach” B.E. Journal of Theoretical Economics: Advances in Theoretical Economics. 1-39
