1
CEO Compensation and Financial Crisis:
Evidence from US Banks
Master Thesis in Finance Author: Jennifer*
Supervisor: Dr. J.J. (Jakob) Bosma January 2018
Abstract
This paper investigates the link between CEO compensation structure on financial crisis and risk taking by using CEO cash-based and equity-based compensation as the explanatory variables and analyzing the non-linear relationships. Only a handful of literature is using this approach. By taking a sample of 93 US banks in the period of 2004-2009, I find that although not to be fully blamed, CEO equity-based compensation package, with stock option in particular, contributed to the financial crisis. Banks which took on more risk before the crisis performed worse during the crisis. Moreover, banks that had more capital ratio before crisis suffered less during the crisis. There is no different effect between banks that received TARP funding and those who did not. US banks that practiced CEO duality performed worse compared to banks that did not use CEO duality approach.
Keywords: CEO Compensation, Banks, firm performance, agency theory JEL Classification: G21, G30, G34, G35, J33
2
Table of Contents
1. INTRODUCTION ... 4
2. LITERATURE REVIEW AND THEORETICAL FRAMEWORK ... 6
2.1 Agency Theory ... 6
2.1.1 Moral Hazard in Banks ... 7
2.2 Corporate Governance System... 7
2.3 US Financial Crisis 2007/2008 ... 8
2.3.1 Troubled Asset Relief Program (TARP) ... 8
2.4. CEO Compensation Structure ... 9
2.4.1 Cash-Based Compensation ... 9
2.4.2 Equity-Based Compensation ... 9
2.4.3 Other Compensation ... 10
2.4.4 Compensation and Financial Crisis 2007/2008... 10
2.5. Firm Performance and CEO Compensation ... 11
2.6. Firm Performance and CEO Duality ... 13
2.7. Firm Performance and Firm Size ... 13
2.8 Firm Performance and Leverage ... 14
2.9 Endogeneity ... 14
2.10 Hypotheses ... 14
3. DATA AND METHODOLOGY ... 16
3.1 Data ... 16
3.2 Methodology ... 16
3.3. Dependent Variables ... 16
3.3.1 Buy-and-Hold Returns ... 16
3.3.2 Return on Asset (ROA) ... 16
3.3.3 Return on Equity (ROE) ... 17
3.4 Control Variables ... 17
3.4.1 Market Value ... 17
3.4.3 Tier 1 Capital Ratio ... 17
3.4.4 Lagged Returns ... 18
3.4.5 Book-to-Market ... 18
3.4.6 CEO Duality ... 18
3
3.4.8 TARP Recipient ... 18
3.4.9 Firm Fixed Effects (FIRM) ... 19
3.4.10 Time Effects (YEAR) ... 19
3.5. Independent variable ... 19
4. RESULTS ... 21
4.1 Descriptive Statistics ... 21
4.2 Correlation Analysis ... 24
4.3 Regression Analysis ... 26
4.3.1 Buy-and-hold Return as Bank Performance ... 26
4.3.2. Return on Assets (ROA) as Bank Performance ... 28
4.3.2. Return on Equity (ROE) as Bank Performance ... 28
4.4 Further Analysis ... 31
4.4.1 TARP Recipients and CEO Compensation ... 31
4.4.2 CEO Duality and CEO Compensation ... 31
4.5 Robustness Check ... 34
5. Discussion and Conclusion ... 34
6. Limitation and Further Research ... 35
REFERENCES ... 37
4
1. INTRODUCTION
An attractive compensation package is a way to motivate CEOs, who are responsible for high-level strategies, decision-making and day-to-day management, to make value-enhancing decision for shareholders and limit agency problem (Jensen and Murphy, 1990). It is essential to align the shareholder and CEO’s interest to avoid agency problem or moral hazard.
Many argue that the level of pay-for performance compensation CEOs receive is too high, not really tied to firm performance, and the arrangement is not focusing on long-term firm performance1. Fahlenbrach and Stulz (2011) said that this may cause more systemic risk. Although there are many different theories on what really caused financial crisis 2007-2008, one of the argument is the executive compensation arrangement that encourage them to take risk excessively and ignore future risk. Even in 2008, US banks executives, still received bonus despite the still ongoing crisis situation, with a total of US$1.6 billion of banks bail-outs2, which came from taxpayers, rewarded to the banks executives3. This is mainly due to the bonus package, so called “Golden Parachute” agreement, was agreed to before the financial crisis occurred, hence there is no way to reconsider the agreement or stop the bonus pay-out4. As Bebchuk, Cohen and Spamann (2010) found in their study, compensation arrangement provided to US bank executives that caused the credit crisis because it is correlated to excessive-risk taking. Moreover, according to Bebchuk and Spamann (2009), compensation packages protected the banks’ executives from a substantial part of potential losses caused by excessive risk taking, that could cause a negative effect on the bank’s value.
Based on the existing literature around CEO Compensation topic, it can be argued that executive compensation should be aligned to company performances (Holmström, 1979; Jensen and Murphy, 1990). If it is adequate, it can reduce the conflict of interest between the top management
1 Why corporate CEO pay is so high, and going higher; 18 May, 2015.Source: https://www.cnbc.com/2015/05/18/why-corporate-ceo-pay-is-so-high-and-going-higher.html
2 $1.6B of Bank Bailout Went to Execs; 21 December, 2008, obtained from: https://www.cbsnews.com/news/16b-of-bank-bailout-went-to-execs/
3 For example, CEO and Chairman of Goldman Sachs from 2006 until present, Llyod Blankfein, received US$ 42.9 million compensation in 2008. Goldman Sachs received a $10 billion bail-out money and an additional of $43.4 billion later on. Source: “9 Wall Street Execs Who Cashed In on the Crisis”, January/February 2010 issue, obtained from http://www.motherjones.com/politics/2010/01/wall-street-bailout-executive-compensation/
5 and shareholders by aligning their interest, so-called agency theory (Grossman et al, 1983; Ozkan, 2011). For instance, a manager’s interest may be more inclined to defend his personal power rather than pursuing profit maximizing strategies (Bebchuk and Fried, 2003).
Beyond corporate performance, other studies discussed executive compensation relationships with variable such as firm size (Veliyath, 1999). Based on the empirical evidence, firm size as a variable shows the consistency of a positive relationship with executive compensation (Tosi et al, 2004), and Kostiuk (1990) found the relationship to be stable overtime. Hartzell and Starks (2003) analyzed the relationship between institutional ownership and firm performance.
Two main components of executive compensation, which are equity-based compensation and cash-based compensation. According to Balachandran et al. (2010), it is equity-based compensation that may induce excessive risk-taking behavior as it may generate higher return, as cash-based compensation level mostly depend on historical performance.
The aim of this study is to understand better the relationship between CEO incentives and financial crisis in 2007/2008, which the existing literatures generate mixed conclusion.
This paper contributes to the existing literature by going a few steps further than Fahlenbrach and Stulz (2011) paper. Firstly, I use broader year observation from period 2004 to 2009 of the firms and applying fixed effect approach. Secondly, Fahlenbrach and Stulz (2011) mainly examined the linear relationships of the variables, thus as for the improvement, this paper tries to analyze if there is a non-linear relationship between the relevant variables. This is essential because non-linear relationship may present in some of the variables in the analysis, mainly stock options due to its convexity, which will be discussed in more detail in later section. Third, I use more control variables that I believe are relevant in the analysis, hence making the analysis result more robust. By using different types of compensation, I can identify which feature of compensation that mainly lead to excessive risk taking. Lastly, I tried to address the possible endogeneity issue that may arise by conducting several robustness checks, which Fahlenbrach and Stulz (2011) did not really address.
With above information, the research question for this paper is below:
6 The empirical results in this study provide evidence that confirmed CEO compensation, particularly equity-based compensation (mainly stock option), induced risk-taking and in turn contributed to the financial crisis event in 2007/2008. Banks which took on more risk before the crisis performed worse during the crisis. Moreover, banks that had more capital ratio before crisis suffered less during the crisis. I then re-estimate the results to let the variables interact with TARP recipient and CEO duality variables. There is no different effect between banks that received TARP funding and those who did not. Also, US banks that practiced CEO duality performed worse compared to banks that did not use CEO duality approach.
The structure of this paper is as follows. Section 2 presents the previous literature studies and theoretical framework on the relationship between CEO compensation and firm performance. It also provides a theoretical framework consisting of corporate governance, agency theory, literature review of the existing studies. Based on those, hypotheses are also made. Section 3 explains the methodology used and the description of the data. It also covers all the variables included in the model for regression analysis. Section 4 offers the results. The discussion and conclusion are provided in section 5. Section 6 will discuss the limitations and future research.
2. LITERATURE REVIEW AND THEORETICAL FRAMEWORK 2.1 Agency Theory
The agency theory is a concept that describes the contractual relationship between principals and agents. Principals (i.e. shareholders) are parties that mandate other parties, so-called agents (i.e. managers), to perform all activities on behalf of the principals in their capacity as decision makers (Jensen and Smith, 1984).
7 with assets. This measurement was used mainly for companies with a high turnover ratio, reflecting the asset utilization and how productive the assets are used to create shareholder value. This conflict between principals and agents can be reduced by aligning interests between them. Moreover, a conflict of interest in the company may arise because of excess cash flow, since it tends to be invested in things that have no connection with the company’s main activities and lead to sub-optimal return for the principals. According to Ozkan (2011), compensation package is an important factor to reduce this agency problem, one of the way is to relate pay for company performance.
In principal-agent relationship, there is also asymmetric information problem. Jensen and Meckling (1976) defined moral hazard as one of the types of information asymmetry problem. 2.1.1 Moral Hazard in Banks
In general, moral hazard appears if the agent does not align his behavior with the principals, as stated in the employment contract. Due to its unique nature, Bebchuk and Spamann (2009) argued that moral hazard in banks is usually related to equity.
Banks usually consists of a vast number of small depositors, that most likely have no resource or motivation to monitor the banks. Bebchuk and Spamann (2009) argued that this is due to deposit insurance by government, that insured the deposits in case the banks have issues, which is why the government has incentive to monitor and regulate banks’ activities. Taking risk excessively is enticing to the banks executive and shareholders, because they will enjoy the upside gains but only lose the capital they have in banks, even if the losses are larger than their capital (Bebchuk and Spamann, 2009). They may also rely on the government if things go south and the bank goes bust. Due to the above conditions, it is necessary for the government to closely monitor and regulate banks’ activities, to ensure they are prudent and not taking risk more than they can. However, this does not mean tight regulation can fully eliminate the moral hazard problem for banks.
2.2 Corporate Governance System
Corporate Governance is defined as a set of rules of relationships between shareholders, managers, creditors, government, employees and other relevant internal and external stakeholders.
8 directors. While on Two Tier System, the composition of members Board of Commissioners (hereafter “BOC”) and Board of Directors (hereafter “BOD”) are hold by different people. BOD oversee the daily firm management and taking strategic decisions for the firm. BOC, which consist of commissioner and independent commissioner, supervise overall management duties and ensure it is aligned with GCG principles. Independent commissioner is essential to reduce conflict of interest and protect stakeholders’ interest. US firm uses the one-tier system model.
According to Carrasco (2005), companies that use two-tier system has to spend more effort and cost on monitoring, but their target is not as demanding as compared to one-tier system.
2.3 US Financial Crisis 2007/2008
Even though the crisis started with a credit crunch, which made obtaining funds from banks more difficult, it impacted the economy worldwide, resulting in a economic activity downturn globally. Only a handful of people foresaw the financial crisis event in 2007/2008, and surprised the world, even the US bank CEOs. This crisis is said to be the worst after the Great Depression. It started with the credit crisis in subprime mortgage market in the US, but quickly developed to financial crisis after the collapse of Lehman Brothers, an investment bank in September 2008. On October 2008, the US congress authorized TARP (Trouble Asset Relief Program) of US$700 billion for bank bailout. Vemala et al. (2014) stated that researches are still ongoing to further investigate the cause and effect of this crisis. Apart from financial deregulation, Crotty (2009) stated that financial crisis was, in a big picture, caused by financial boom that caused by financial innovation.
According to the US Financial Crisis Inquiry Commission, they found that the warning signs were ignored and concluded the financial crisis could be avoided. They also found that the key crisis reason is the failure of both corporate governance and risk management, especially in those systematically important banks5.
2.3.1 Troubled Asset Relief Program (TARP)
On October 2008, US Congress authorized the Troubled Asset Relief Program (TARP), worth of $700 billion, mainly for bank bailout. The main purpose was to infuse cash to systematically important banks so they could keep operating. This was managed by the US Treasury Department, who bought shares and bonds from failing banks and companies. The idea was for banks to give 5% dividend to the government, and would be increased to 9% in 2013, which would incentivize
9 banks to buy their stocks back. Nguyen and Enomoto (2009) found that after Capital Purchase Program (CPP) of TARP was announced, stock returns volatility was reduced. One of the rules for TARP receiver is that the limits are place for executive compensation and forbid giving bonuses to the executives6. The US Department of Treasury argued that TARP is essential to avoid a second Great Depression and actually only cost around $34.5 billion compared to the authorized initial amount, which is $700 billion7. According to Black and Hazelwood (2013), large banks that received government support via TARP funding actually had higher risk taking but no increase in lending amount, suggesting the present of moral hazard.
2.4. CEO Compensation Structure
In US, CEO compensation mainly consists of equity-based and cash-based compensation.
For cash-based compensation component, which usually consists of base salary and cash bonus mainly depends on historical firm performance. For equity-based compensation, such as stock options, stock ownership, the amount may change relatively depending on the stock market value. According to Bebchuk and Fried (2003), the level of compensation is determined by the company’s board.
2.4.1 Cash-Based Compensation
For cash-based compensation component, which usually consists of base salary and cash bonus. Base salary is a fixed compensation, usually predetermined and does not linked to performance. While cash bonus mainly depends on historical firm performance.
2.4.2 Equity-Based Compensation
Aside from the cash-based compensation, a compensation package that aligns shareholder’s interest with CEO’s interest are also commonly rewarded, especially after deregulation of banks. Cuñat and Guadalupe (2007) found that fixed compensation amount declined and variable compensation amount and pay-to-performance sensitivity rose after banks deregulation. Although each company may have different equity compensation package, restricted stock and stock option are the most common ones.
6 Executive compensation. US Department of The Treasury. Source: https://www.treasury.gov/initiatives/financial-stability/TARP-Programs/executive-comp/Pages/default.aspx
10 Restricted stock is stocks rewarded as compensation, but not transferable. The terms will depend on the company’s contract, for example the stock can be transferred few years after the CEO has left the company.
As for stock options, the CEO may gain profit depending on the stock’s price increases by exercising his right to buy the stock for a predetermined price. Before maturity, stock option may contain convexity effect. This is, the option may derive the value not only from the underlying asset, but also time value of money and volatility. It is impossible for CEOs to influence the time value of money, as it is mainly affected by the risk-free asset, hence they may try to influence the volatility. According to Rajgopal and Shevlin (2002), stock price, which they referred to as stock option slope effect, gives incentives to CEO to invest in projects with positive NPV, and stock return volatility, which they referred to as stock option risk incentive effect, gives incentives to CEOs to take more risky projects in order to increase stock return volatility. If stock return volatility increases, the value of option will increase as well. This generates a convexity in the pay-performance relation. They also argued that the greater the convexity, the greater the CEOs incentives to make decisions that may increase firm risk.
2.4.3 Other Compensation
Few other compensation or benefits CEOs are pension plan, retirement benefits, company cars and insurance premium, which are determined by each company. These are not linked to performance.
2.4.4 Compensation and Financial Crisis 2007/2008
The US Financial Crisis Inquiry Commission highlighted in their report, that gigantic bonus is given on the upside, but limited consequences on the downside, which encourage “gambling”8. Many scholars also believe that compensation structure is also to be blamed for the financial crisis in 2007/2008. This is not without some good arguments. In their paper, Bebchuk and Spamann (2009) stated that in theory, the bank executives should be more conservative and better align with shareholder when they hold shares in the company, or given equity-based compensation. Ever since the bank deregulation, stock option as a form of equity-based compensation became popular, and the bank executives received a huge portion of stock option as their compensation. However, Chen et al. (2006) used data for period 1992-2000 and stated that stock option-based compensation induce risk taking. Bebchuk and Spamann (2009) stated that having stock options induces even
11 more risk taking than having common shares. Risk may get badly skewed for stock options. As in, when shares value increases, the CEO gets some profit from having options, but when it decreases in value, the CEO has no loss, while the investors suffer from losses. Balachandran et al. (2010) also supported that equity-based compensation increases the firm default probability, while cash based compensation decreases the probability. This is why many argue that stock option as an incentive makes CEOs focus only on short-term performance, as their incentives are to make their options to remain “in-the-money” by keeping the share price upwards.
Even though there are extensive literatures investigated the direct relationship between CEO compensation and firm performance, only a handful has examined the financial crisis period. Vemala et al. (2014) stated that since most of the papers support the positive relationship between CEO compensation and firm performance, it should also be the case during financial crisis period. The existing studies that examines the CEO compensation and the financial crisis have mixed results. Bebchuk, Cohen and Spamann (2010), focused on Bear Stearns and Lehman Brothers data for the period 2000-2008, found that CEO compensation induced excessive risk taking. In 2011, Suntheim (2011) also showed that during crisis, compensation had an impact on bank performance. Similar result was also concluded in Bhagat and Bolton (2014) paper, which used 14 US financial institution data. In contrast, Fahlenbrach and Stulz (2011) concluded that CEO incentive package is not to be blamed for the bank performance or financial crisis and the executives, even if they took into account the possibility of negative outcome, they did not realize how extreme the negative outcome would be, because if they did, they should have exercised their stock options before the bank’s collapse. However, they also suffered from substantial losses. This can be stated as Unforeseen Risk, As Bhagat and Bolton (2014) said in their paper.
2.5. Firm Performance and CEO Compensation
12 Kaplan (1994) stated that accounting based measure is more suitable than market-based measure in analyzing the CEOs’ performance.
There have been extensive literatures on the pay-performance relationship and have been going on since 1980s. Murphy (1985) published one of the first and significant paper with executive compensation topic. Most existing researches yield mixed results for the link between executive compensation and firm performance. Barclay et al. (2005) found that accounting measures are used to determined cash based compensation, making it directly linked to decisions taken by the CEO, whereas there are other unobservable factors that can determine equity based compensation, such as economic trends or interest rates.
Ozkan (2011) used panel data set of 390 UK non-financial firms from 1999-2005, and found positive relationship between cash compensation and firm performance. Cheng & Farber (2008), found a positive relation between equity incentives and the firm performance by using 289 firms for period 1997-2001. In China, Conyon and He (2011) found positive pay-performance relationship. Gao and Li (2015) found positive and significant pay-performance relationship, but it is weaker for private firms compare to public firms. In addition, Lilling (2006), concluded that employment contracts that includes incentive-based compensation is effective since he found the positive pay-performance link. There is also a higher ROE result in companies that implement long-term incentive plans found by Leonard (1990). Banker et al. (2013) found salary is positively related with past performance. In emerging market, Raithatha and Komera (2014) studied the pay-performance relationship by using Indian firms’ data and Unite et al. (2008) used Philippines data, and they found a significant positive relationship between executive compensation and firm performance.
Meanwhile, Boyd (1994) found no correlation between executive compensation and firm performance. The hypotheses were tested using 193 firms as sample in cross-section industries for the year 1980. Mäkinen (2007) used Finnish Panel data from 1996-2002 and found no relationship between change in compensation and change in ROA, except for the lagged performance. Using Indian data, Raithatha & Komera (2014) found an absence of pay-performance for small sample firms.
13 and size is negatively related to compensation. Basu et al. (2007) found presence of an agency problem in Japan by examining 174 Japanese corporation during 1992-1996. Analysing Dutch firms, Duffhues and Kabir (2008) found significant negative pay-performance relationship.
2.6. Firm Performance and CEO Duality
In US, since implements one-tier system is implemented, as explained in point 2.2, CEO can also be a chairman in the company, this refers as CEO duality. Some studies have been done on this topic and they generated mixed results. Rechner and Dalton (1991) found that companies with non-CEO duality system outperformed companies that had CEO duality system; and Lorsch and Young (1989) argued that non-CEO duality system helps companies to avoid crisis. However, Baliga and Moyer (1996) did not find any significant result that CEO duality affects performance, and the market seems to be indifferent of this. Brickley et al. (1997) pointed out that the cost of CEO duality approach is higher than its benefits for the firms. In contrast, and Yan Lam and Kam Lee (2008) found that CEO-duality system is good for non-family firms and vice versa. Abobakr and Elgiziry (2017) found that CEO-duality is positively correlated with credit risk.
In Carty and Weiss (2012) paper, it was stated that CEO duality is not forbidden according to the US banking code, confirmed by several regulators. Despite of lack in analysis, they mentioned that there are strong arguments that indicate CEO duality may contribute to financial crisis 2008. So far, there are mixed arguments about this. For example, while there are some arguments that CEO duality brings efficiency, CEO of Norges Bank Investment Management stated that CEO duality is unsustainable and US banks should end this practice, especially after what happened in 20089.
2.7. Firm Performance and Firm Size
Firm size has been used extensively as a control variable in research studies. According to Fahlenbrach and Stulz (2011), firm size is correlated to firm returns and bigger firm may able to take on more risk. Kurshev and Strebulaev (2006) argued that bigger firms have lower level of information asymmetry and have broader financing source compared to smaller firms. Boyd and Runkle (1993) also predicted that bigger banks are likely less prone to failure and can be more cost-effective compared to smaller banks. Moreover, bigger firms tend to be more complex, as argued by Hall et al. (1967).
14
2.8 Firm Performance and Leverage
For non-financial firms, leverage gives an indication how the activities of the company are dependent on the debt owed and the higher the financial leverage, the higher the risk of the company, which includes bankruptcy costs (Jensen, 1986). In contrast, for financial firms, they do prefer leverage as it generates profit for financial firms. As DeAngelo and Stulz (2015) stated, high leverage is good for banks as it generates higher profit. However, the downfall can be destructive since the losses is multiplied and may potentially cause bankruptcy, as seen from the situation during financial crisis 2008. Furthermore, Flagg (1991) also found a positive relationship between bankrupt firms and leverage ratio. To prevent future credit crisis event, Basel II requires banks to maintain justified leverage ratio10.
2.9 Endogeneity
Endogeneity is mainly a big issue in corporate governance research. One endogeneity issue that Aebi et al. (2011) pointed out in this kind of study, is due to lagged bank performance as a control variable, since bank performance is dependent variable. Thus, there may autocorrelation issue as the dependent variable may be correlated with the error term, causing in biased results. Another thing they pointed out is the existence of unobservable factor that may correlate with the variables used in this study. I follow their approach in addressing this by motivating the inclusion of the bank characteristics I use in my analysis, since it is unknown what instrument is suitable to tackle this potential issue. Moreover, I will conduct robustness check that will address
endogeneity issue. This will be discussed in later section.
2.10 Hypotheses
Based on the information above, I analyze if the financial crisis is related CEO compensation package by using US banks data for the period 2004-2009. Following Fahlenbrach and Stulz (2011), buy-and-hold returns and accounting measures will be used.
According to Fahlenbrach and Stulz (2011), buy-and-hold returns is suitable to assess bank’s performance in long-run, and generates same results as using a constructed portfolio abnormal performance and investigates the elements which affects banks returns.
H1: Buy-and-hold returns has a positive relationship with CEO cash-based compensation H2: Buy-and-hold returns has a positive relationship with CEO equity-based compensation
15 Penno (1990) argued that accounting indicators are commonly used to measure employees’ performance. Hence, two accounting measures, ROA (Return on Asset) and ROE (Return on Equity) will be used as the firm performance indicators. ROA is the most common indicator for banks to assess their performance and compare it with other banks. It indicates the effectiveness of a firm by showing how much profit is earned for every dollar of the firm’s assets. According to Merchant (2011), most companies calculate the bonus compensation by using accounting-based measures. Research by Carter et al (2010), Adams and Ferreira (2009), Hull and Rothenberg (2008), Levi et al (2014) and Ross et al. (2017), found that ROA indicates the company's ability to allocate and manage resources effectively. On the other hand, ROE indicates the effectiveness of a firm by showing how much profit is earned for every dollar of the firm’s capital. ROE becomes more popular among larger banks, because it’s independent of the asset, thus it can be applied to any business sectors. By using both ROA and ROE, it gives clearer indicators of the firm management’s effectiveness in managing its resources. Thus, I use both measures in my analysis. Since in this paper, we examine the relationship between past performance and CEO compensation, it is expected for the annual cash-bonus compensation amount to be based on one of the common performance indicators, which is ROA. Nulla (2013) found a correlation between CEO bonus and ROA.
H3: ROA has a positive relationship with CEO cash-based compensation
Similarly, annual-cash bonus compensation is also determined based on the ROE of the firm. Tosi et al (2000) found a positive relationship between CEO pay and ROE.
H4: ROE has a positive relationship with CEO cash-based compensation
Equity-based compensation, which is a long-term incentive plan, is largely used to reward the management to better link with shareholder interest. Frye (2004) found a positive relationship between ROA and equity-based compensation. Similarly, Pyo and Abedin (2017) found a positive relationship between ROA and pay-to-performance sensitivity.
H5: ROA has a positive relationship with CEO equity-based compensation
16
3. DATA AND METHODOLOGY 3.1 Data
The data is mainly generated from Execucomp. The sample covers 5-year period, from 2005 to 2009, and the final raw sample contains 93 firms. The whole sample is an unbalanced panel with a total of 461 firm-year observations. For details on the sample firms, please refer to Appendix Table A.1.
3.2 Methodology
The basic model for the regression of this paper based on the hypotheses are below:
Bank Performance i,t = β0 + β1Cash bonus/salary+ Equity incentive measures + Control
Variables + FIRMi + YEARt +
ε
i,tAll the variables will be explained in details in the next sections.
Following Fahlenbrach and Stulz (2011) and Raithatha & Komera (2016), I use Ordinary Least Squares (OLS) for the regression of panel data of past CEO compensation and performance. Kaplan (1992) argued that regression result from OLS method is easy to explain, and it produced same results as with logit regression method.
3.3. Dependent Variables
Three performance ratios are used in this paper, namely buy-and-hold returns, ROA and ROE. While there are not many papers use buy-and-hold returns, ROA and ROE are commonly used in extensive literatures. Following Aebi et al. (2011), buy-and-hold return, ROA and ROE measures are winsorized at the 1st and 99th percentile. The main reason for this is to exclude outliers that may greatly influence the regression and making the results less accurate.
3.3.1 Buy-and-Hold Returns
The first bank performance I use is buy-and-hold returns. This measure is also used by Aebi et al. (2011) and Beltratti and Stulz (2012). The measure is also useful as an indicator of how bank performed before the crisis.
3.3.2 Return on Asset (ROA)
17 in 2007 is calculated from net income at the end of year 2007 divided by the total assets value at the end of 2006. The motivation to do this is to be able to really capture the financial crisis condition.
ROA =
Net Income Total Assets 3.3.3 Return on Equity (ROE)
Aside from ROA, ROE is also used to test the link between executive pay-performance. ROE indicates the efficiency of management in utilizing its equity and generating better return to investors. Kuo et al. (2013), Sigler (2011) and Raithatha & Komera (2014) used return on equity (ROE) as their performance measure in their researches. Following Fahlenbrach and Stulz (2011) approach, the ROE measure takes net income in the respective period and divided with the previous period total assets value. For example, the ROE in 2007 is calculated from net income at the end of year 2007 divided by the total assets value at the end of 2006. The motivation to do this is to be able to really capture the financial crisis condition.
ROE =
Net Income Total Equity
3.4 Control Variables
There may be a lot of factors that can contribute to compensation aside from firm performance, hence we need to control it in our model to ensure we test our hypotheses correctly.
3.4.1 Market Value
Controlling for firm size is important in order to really capture the relationship between dependent and independent variables. There are many variable options for size control variable, but following Fahlenbrach and Stulz (2011), I use market value as the size variable. Since the market value in my dataset is rightly skewed, the natural logarithm value will be used.
3.4.3 Tier 1 Capital Ratio
18
3.4.4 Lagged Returns
In respective regressions to bank performances, lagged buy-and-hold returns, lagged ROA and lagged ROE are also part of the control variables. According to Aebi et al. (2011), this is essential to see the bank performance from the previous period in order to assess the performance before and during the financial crisis.
3.4.5 Book-to-Market
Dhatt et al (1999) stated that book to market ratio is one of the best determinant of stock returns. Pontiff and Schall (1998) found that unlike interest yield spreads and dividend yields, book-to-market ratio can predict future returns. Aebi et al. (2011) pointed out that book-to-book-to-market measure can help to check the relationship between growth expectation and bank performance. To remove outliers, I winsorized book-to-market at the 2nd and 98th percentile, following Fahlenbrach and Stulz (2011).
3.4.6 CEO Duality
CEO duality means that the CEO is also a chairman in the board of the company. I will use a dichotomous variable, that is valued 1 if a CEO is also a chairman, and 0 otherwise.
3.4.7 Leverage (LEV)
Financial leverage is one of the control variables in this paper. I followed Matolscy (2012) for the financial leverage calculation, which is below:
Leverage = Total Debt Total Assets
3.4.8 TARP Recipient
19
3.4.9 Firm Fixed Effects (FIRM)
According to Allison (2009), firm fixed effect is used to increase efficiency. As there may be some unobservable factors that may influence the variables in the regressions, firm-fixed effect helps to omit this issue. However, since our main focus is not on firm-fixed effects, the coefficients is not reported in the results.
3.4.10 Time Effects (YEAR)
Many literatures, such as Basu et al (2007) and Darmadi (2011) controlled for time effect, in order to check and control any potential differences across the research years in compensation. Especially since I use data from 2005 to 2009, year dummy is essential as a control variable. Moreover, as Fahlenbrach and Stulz (2011) pointed, in 2006, there is a new disclosure requirement implemented by the Securities and Exchange Commission (SEC), and compensation was one of the items that were changed. However, since our main focus is not on time-effects, the coefficients is not reported in the results.
3.5. Independent variable
20 Table 1. Definitions and Sources for Variables
Definitions/Calculations Description Source Dependent Variables
Buy-and-hold Retruns Stock Returns Execucomp
Return on Assets Net Income/Total Assets Execucomp Return on Equity Net Income/Total Equity Execucomp Control Variables
Market Value Natural Logarithm of Market Value Execucomp
Tier 1 Capital Ratio In percentage Execucomp
Leverage Total Liabilities/Total Assets Execucomp
Book to Market In percentage Execucomp
Lagged Returns Lagged buy-and-hold returns, lagged ROA
or lagged ROE Execucomp
Duality 1 if CEO is also a chairman, 0 otherwise Execucomp TARP recipient 1 if a bank receives TARP funding, 0
otherwise CNN Money
11
Independent Variables
Cash bonus/salary Cash bonus to cash salary ratio Execucomp Dollar gain from +1% value change in CEO’s portfolio of equity
by 1% change in the stock price Execucomp Equity Risk ($) dollar percentage change in value of equity
portfolio for a 1% change in stock volatility Execucomp
Ownership (%)
percentage total of restricted and unrestricted shares, as well as the stock options held by the CEO divided by the number of outstanding shares is also used (ownership)
Execucomp
Equity Risk (%) Percentage change in value of equity
portfolio for a 1% change in stock volatility Execucomp
21
4. RESULTS
4.1 Descriptive Statistics
Table 2 and Table 3 show the descriptive statistics of all the relevant variables in this study. Table 2 represents an overview of the sample banks summary statistics. The total number of observation for all observations, except for Tier 1 Capital Ratio, is 461. Based on the size variables (total assets, total liabilities, and market capitalization), large and small banks are included in the sample with median of total asset $13.5 billion. Tier 1 Capital Ratio and Tangible Common Equity Ratio are measures for bank’s financial health. On average at 10.11%, the sample banks have sufficient capital. Throughout 2004 to 2007, all banks kept their Tier 1 capital ratio above 4%, which was the regulatory requirement in 2006.
Table 2
Descriptive Statistics for period 2004-2009
The table shows descriptive statistics for key variables for a sample of 93 banks for period 2004-2009. The list of sample banks is in the Appendix 1. Tangible common equity ratio is tangible common equity divided by tangible assets. All accounting variables are reported in millions of dollars.
Mean Median Maximum Minimum Std. Dev. Obs.
Total assets 135,029.70 13,503.36 2,223,299.00 1,004.81 349,498.20 461
Total liabilities 125,099.60 12,074.64 2,074,033.00 918.38 324,266.50 461
Market Capitalization 14,626.16 1,883.51 273,691.20 17.30 36,911.11 461
Net Income/total assets 0.68% 0.99% 3.69% -7.70% 1.39% 461
Net Income/book equity 4.49% 9.55% 29.05% -70.42% 19.33% 461
Cash/Total Assets 2.45% 2.13% 26.90% 0.25% 2.19% 461
Dividend per share 0.74 0.53 3.54 -0.40 0.69 461
Book-to-Market ratio 1.19 0.59 73.63 -18.72 4.28 461
Tier 1 Capital Ratio 10.11% 9.83% 21.50% 1.50% 2.51% 390
Tangible Common Equity
Ratio 5.99% 5.82% 34.52% 0.00% 3.13% 461
Leverage 0.906 0.908 1.017 0.655 0.034 461
22 dramatic changes. The equity holding of CEO increased after 2008, this is in line with Sonenshine et al. (2016) that found the compensation composition shifted more towards equity after crisis.
Figure 1A. Average and Median CEO Total Compensation and Total Bonus in the years 2004-2009
Figure 1B. Average and median CEO Equity Holding in the years 2004-2009
Table 3 reports the overview of CEO compensation for period 2004 to 2009. From the statistics, it can be seen that the shares value that CEOs had in the companies they worked was about ten times larger than their annual compensation. Based on this, CEOs’ interests should be well aligned with shareholders’ interests. According to Fahlenbrach and Stulz (2011), with such a big difference in the shares value owned by CEOs and the annual compensation, it is hardly reasonable for them to
USD 0 USD 1,000,000 USD 2,000,000 USD 3,000,000 USD 4,000,000 USD 5,000,000 USD 6,000,000 USD 7,000,000 USD 8,000,000 2004 2005 2006 2007 2008 2009
Average Total Compensation Average Total Bonus Median Total Compensation Median Total Bonus
Years USD 0 USD 10,000,000 USD 20,000,000 USD 30,000,000 USD 40,000,000 USD 50,000,000 USD 60,000,000 USD 70,000,000 USD 80,000,000 2004 2005 2006 2007 2008 2009
Average Equity Holding Median Equity Holding
23 take actions that decrease shareholders’ wealth consciously. For example, in 2006 (before the crisis in 2007), CEO of Lehman Brothers held approximately $1.5 billion of equity holdings, Bear Stearns’s CEO followed with approximately $1 billion of equity holdings. One of the possible argument they stated that might influence CEO’s action to differ from shareholder’s interest, is that in good times, CEOs will receive huge cash bonuses, however, during bad times, the cash bonus may just be zero since it is impossible for it to have negative value.
Table 3
CEO compensation and equity ownership for period 2004-2009
The table shows descriptive statistics for main compensation variables for a sample consists of 93 banks for period 2004-2009. Values are in thousands of dollars.
Mean Median Maximum Minimum Std. Dev.
Annual Compensation
Total Compensation 5,745.59 2,196.28 46,375.35 32.17 8,894.07
Salary 752.27 737.50 2,866.67 0.00 338.24
Cash Bonus 983.67 0.00 19,557.36 0.00 2,834.98
Dollar value of annual stock grant 1,012.50 0.00 36,179.92 0.00 3,620.47
Dollar value of annual option grant 668.23 0.00 19,868.00 0.00 2,271.95
Other compensation 252.31 92.84 6,482.06 0.00 657.38
Cash Bonus/Salary 1.47 0.00 68.28 0.00 6.04
Equity Portfolio Value
Value of total equity portfolio 49,909.86 14,628.16 1,415,853 0.00 125,074.5
Value of shares 30,855.55 2,189.79 871,852.9 0.00 92,475.46
Value of exercisable options
(Black-Scholes) 10,426.96 2,066.14 308,172.3 0.00 25,589.60
Value of unexercisable options
(Black-Scholes) 1,618.05 34.80 47,273.66 0.00 3,989.43
Value of unvested restricted stock 7,009.30 209.69 527,988.2 0.00 32,894.39
Value of total equity portfolio/total
annual compensation 15.45 4.63 1,236.18 0.00 61.47
Value of shares/salary 57.28 4.09 3,487.41 0.00 282.22
Equity portfolio incentives
Percentage ownership from shares 1.46 0.22 36.71 0.00 4.25
Percentage ownership 1.91 0.66 39.12 0.00 4.48
Dollar gain from +1% 613.17 193.94 27,401.43 0.00 1,628.20
Equity portfolio risk exposure
Percentage equity risk 0.14 0.07 1.11 0.00 0.16
24
4.2 Correlation Analysis
25
Table 4
Pairwise Correlation Matrix
No Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1 Buy-and-Hold Return 1.000
2 Return on Assets (ROA) -0.137 1.000
3 Return on Equity (ROE) -0.179 0.952 1.000
4 Cash bonus/Salary -0.034 0.241 0.283 1.000
5 Dollar gain from +1% -0.103 0.048 0.074 0.271 1.000
6 Ownership -0.018 -0.068 -0.063 -0.028 0.381 1.000 7 Equity Risk ($) -0.072 0.105 0.121 0.371 0.525 -0.062 1.000 8 Equity Risk (%) 0.017 -0.081 -0.093 -0.027 -0.060 -0.209 0.174 1.000 9 Lagged ROA -0.362 0.314 0.334 0.200 0.176 0.063 0.142 -0.098 1.000 10 Lagged ROE -0.347 0.313 0.359 0.230 0.198 0.084 0.153 -0.093 0.958 1.000 11 Book-to-Market 0.073 -0.282 -0.331 -0.212 -0.229 -0.084 -0.203 0.115 -0.517 -0.545 1.000
12 Tier 1 Capital Ratio -0.144 0.017 0.024 -0.164 -0.156 -0.038 -0.238 -0.029 -0.017 -0.055 0.132 1.000
13 Market Capitalization -0.232 0.267 0.299 0.416 0.403 -0.267 0.550 0.014 0.412 0.402 -0.249 -0.267 1.000
14 CEO Duality -0.085 0.068 0.064 0.058 0.275 0.091 0.214 -0.065 0.001 -0.006 0.076 -0.161 0.332 1.000
15 Leverage 0.174 -0.039 0.000 0.101 0.102 0.170 0.072 -0.001 -0.152 0.009 -0.044 -0.475 -0.091 0.055 1.000
26
4.3 Regression Analysis
One measure that is taken in reducing the biased result issue that may be caused by omitted factors that may influence the variables in OLS estimation, is to apply firm-fixed effect and time-fixed effect. I apply both in all the regressions. Prior to this, I tested the all the regressions, to see if firm-fixed effect is suitable to be applied to the regressions by conducting Redundant Fixed Effects and Hausman test. If the results for both test are inconsistent, Lagrange Multiplier Test (LM Test) is conducted.
4.3.1 Buy-and-hold Return as Bank Performance
27
Table 5
Buy-and-hold returns and CEO annual compensation in the period 2004-2009
This table shows the results of OLS estimations with inclusion of firm-fixed and time-fixed effects. Columns 1, 2 and 3 show results from return on assets, which is regressed on CEO annual cash bonus, ownership incentives and equity risk sensitivity. "Cash bonus/salary" is the ratio of cash bonus to salary. "Dollar gain from +1%" is the change in dollar value of CEO's equity portfolio respective to 1% change in stock price. "Ownership (%)" is the sum of all shares, restricted and unrestricted, and delta-weighted options, exercisable and unexercisable, which is divided by total amount of outstanding shares. "Equity Risk ($)" is the value change (in dollar) of CEO's equity portfolio respective to a 1% increase in stock volatility. "Equity Risk (%)" is the percentage change in CEO's equity portfolio respective to a 1% increase in stock volatility, and is measured based on all options held by the CEO. Control variables are natural logarithm of market capitalization, Tier 1 Capital Ratio, book-to-market ratio, lagged buy-and-hold returns, CEO duality and leverage. CEO duality is a dichotomous variable, which is valued 1 if the CEO is also a chairman in the firm, 0 otherwise. Stock returns is respective to the data year. All variables are measured at the end of the year. Standard errors are reported in parentheses. Significances at 0.01, 0.05 and 0.1 level are indicated by ***, ** and * respectively.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Cash bonus/salary -0.959* -0.930 -0.670 0.774 1.802
[0.519] [0.502] [0.623] [2.048] [3.024]
Dollar gain from +1% 0.003 0.006 0.014*
[0.006] [0.007] [0.007] Ownership (%) -2.133 -0.692 -1.560 [2.731] [2.940] [3.154] Equity Risk ($) 0.179** 0.160* 0.094 [0.085] [0.091] [0.109] Equity Risk (%) 1.150*** 0.935** 0.887** [0.358] [0.366] [0.390]
Lagged buy-and-hold returns 0.012 -0.070 -0.008 -0.095 [0.105] [0.115] [0.114] [0.130]
Book-to-Market -3.395** -6.475*** -1.576*** -8.305***
[6.262] [7.169] [7.227] [7.870] Log (market value) -23.720*** -25.314*** -30.669*** -27.322***
[5.888] [6.880] [7.329] [7.772]
Tier 1 capital ratio 0.801* 0.459*
28
4.3.2. Return on Assets (ROA) as Bank Performance
The table below shows results for ROA. Column 1 and 2 in Table 6 use all banks in the regression, while in Column 3 and 4, tier 1 capital ratio is added as a control variable, so all investment banks and non-depository banks are automatically excluded. One of equity incentive measure, dollar gain from +1%, has a positive and significant relation with ROA with small coefficient. The coefficient is even more significant when only banks that are required to have capital ratio are included. This means that when dollar gain from +1% measure increases, the ROA also increases. However, the other measures such as the short-term incentives, ownership and equity risk have non-significant relationship with ROA. Those are also in-line with Fahlenbrach and Stulz (2011) found. As for the control variables, market value, which measures the firm’s size is statistically significant for all the regression results, while for book-to-market appears to be significant while excluding investment banks. On the other hand, leverage is positively significant while the regression includes investment banks. This is in line as what Beltratti and Paladino (2015) found, which there is a significantly positive relationship between leverage and profitability.
4.3.2. Return on Equity (ROE) as Bank Performance
29
Table 6
Return on assets (ROA) and CEO annual compensation in the period 2004-2009
This table shows the results of OLS estimations with inclusion of firm-fixed and time-fixed effects. Columns 1, 2 and 3 show results from return on assets, which is regressed on CEO annual cash bonus, ownership incentives and equity risk sensitivity. "Cash bonus/salary" is the ratio of cash bonus to salary. "Dollar gain from +1%" is the change in dollar value of CEO's equity portfolio respective to 1% change in stock price. "Ownership (%)" is the sum of all shares, restricted and unrestricted, and delta-weighted options, exercisable and unexercisable, which is divided by total amount of outstanding shares. "Equity Risk ($)" is the value change (in dollar) of CEO's equity portfolio respective to a 1% increase in stock volatility. "Equity Risk (%)" is the percentage change in CEO's equity portfolio respective to a 1% increase in stock volatility, and is measured based on all options held by the CEO. Control variables are natural logarithm of market capitalization, Tier 1 Capital Ratio, book-to-market ratio, lagged ROA, CEO duality and leverage. CEO duality is a dichotomous variable, which is valued 1 if the CEO is also a chairman in the firm, 0 otherwise. All variables are measured at the end of the year. Standard errors are reported in parentheses. Significances at 0.01, 0.05 and 0.1 level are indicated by ***, ** and * respectively.
1 2 3 4
Cash bonus/salary 0.005 0.002 0.033 0.080 [0.025] [0.049] [0.077] [0.119] Dollar gain from +1% 0.001* 0.001**
[0.000] [0.000] Equity Risk ($) 0.002 0.002 [0.004] [0.004] Ownership (%) 0.023 -0.001 [0.119] [0.124] Equity Risk (%) 0.004 0.005 [0.016] [0.016] Lagged ROA -0.057 -0.013 -0.115 -0.052 [0.122] [0.150] [0.130] [0.159] Book-to-Market -0.376 -0.427 -0.534** -0.470 [0.237] [0.282] [0.268] [0.296] Market Value 0.963*** 1.113*** 1.270*** 1.193*** [0.247] [0.327] [0.315] [0.352]
Tier 1 Capital Ratio 0.101 0.121**
30
Table 7
Return on equity (ROE) and CEO annual compensation in the period 2004-2009
This table shows the results of OLS estimations with inclusion of firm-fixed and time-fixed effects. Columns 1, 2, 3 and 4. show results from return on assets, which is regressed on CEO annual cash bonus, ownership incentives and/or equity risk sensitivity. "Cash bonus/salary" is the ratio of cash bonus to salary. "Dollar gain from +1%" is the change in dollar value of CEO's equity portfolio respective to 1% change in stock price. "Ownership (%)" is the sum of all shares, restricted and unrestricted, and delta-weighted options, exercisable and unexercisable, which is divided by total amount of outstanding shares. "Equity Risk ($)" is the value change (in dollar) of CEO's equity portfolio respective to a 1% increase in stock volatility. "Equity Risk (%)" is the percentage change in CEO's equity portfolio respective to a 1% increase in stock volatility, and is measured based on all options held by the CEO. Control variables are natural logarithm of market capitalization, Tier 1 Capital Ratio, book-to-market ratio, lagged ROE, CEO duality and leverage. CEO duality is a dichotomous variable, which is valued 1 if the CEO is also a chairman in the firm, 0 otherwise. All variables are measured at the end of the year. Standard errors are reported in parentheses. Significances at 0.01, 0.05 and 0.1 level are indicated by ***, ** and * respectively.
1 2 3 4
Cash bonus/salary 0.106 0.058 0.182 0.568
[0.252] [0.467] [0.722] [1.096]
Dollar gain from +1% 0.105 0.004*
[0.252] [0.003] Equity Risk ($) 0.032 0.022 [0.036] [0.042] Ownership (%) -0.460 -0.536 [1.138] [1.158] Equity Risk (%) -0.004 0.030 [0.150] [0.151] Lagged ROE -0.121 -0.088 -0.253** -0.199*** [0.110] [0.125] [0.114] [0.137] Book-to-Market -0.006 -4.564* -5.484** -5.029** [2.346] [2.676] [2.507] [2.734] Market Value 6.292** 12.866*** 15.426*** 14.286*** [2.607] [3.161] [2.985] [3.301]
Tier 1 Capital Ratio -1.141 -1.363*
31
4.4 Further Analysis
4.4.1 TARP Recipients and CEO Compensation
As Fahlenbrach and Stulz (2011) stated in their paper, the results above may differ if the analysis focuses on the banks that received Troubled Asset Relief Program (TARP) funding compared to banks that did not receive it. TARP funding was to stabilize the US financial system after 2008 financial crisis. In my sample of 93 banks, 52 banks received TARP funding. Since the TARP program has some effects to the stock return volatility, as what Nguyen and Enomoto (2009) found, as well as Fahlenbrach and Stulz (2011) argued that return could be affected for banks that received TARP funding, I use only accounting measures. I re-estimate the results in table 6 and 7 to let the incentive measures interact with TARP recipient indicator, which is equal to 1 if bank receives TARP funding. The results are reported in Table 8. Regression 1 is using ROA as the bank performance, and regression 2 is using ROE as the performance indicator. In line with Fahlenbrach and Stulz (2011) found, there is no difference between banks that received TARP funding and banks that did not receive it for both regressions.
4.4.2 CEO Duality and CEO Compensation
32
Table 8
Return on assets (ROA), Return on Equity (ROE), CEO annual compensation in the period 2004-2009 for Troubled Asset Relief Program (TARP) recipients.
This table shows the results of OLS estimations with inclusion of firm-fixed and time-fixed effects. Column 1 (2) shows results from return on assets (return on equity), which is regressed on CEO annual cash bonus, ownership incentives and equity risk sensitivity. "Cash bonus/salary" is the ratio of cash bonus to salary. "Dollar gain from +1%" is the change in dollar value of CEO's equity portfolio respective to 1% change in stock price. "Ownership (%)" is the sum of all shares, restricted and unrestricted, and delta-weighted options, exercisable and unexercisable, which is divided by total amount of outstanding shares. "Equity Risk ($)" is the value change (in dollar) of CEO's equity portfolio respective to a 1% increase in stock volatility. "Equity Risk (%)" is the percentage change in CEO's equity portfolio respective to a 1% increase in stock volatility, and is measured based on all options held by the CEO. Control variables are natural logarithm of market capitalization, Tier 1 Capital Ratio, book-to-market ratio, CEO duality and leverage. CEO duality is a dichotomous variable, which is valued 1 if the CEO is also a chairman in the firm, 0 otherwise. Lagged Returns is lagged return on assets (equity) for column 1 (2). All variables are measured at the end of the year. Standard errors are reported in parentheses. Significances at 0.01, 0.05 and 0.1 level are indicated by ***, ** and * respectively.
ROA ROE
TARP recipient -0.364 -3.521
[0.222] [2.170]
Cash bonus/salary 0.139 1.750
[0.076] [0.741] TARP recipient x Cash bonus/salary -0.143 1.765
[0.077] [0.751]
Dollar gain from +1% 0.001*** 0.008***
[0.000] [0.003] TARP recipient x Dollar Gain from +1% 0.001 0.008
[0.000] [0.003]
Equity Risk ($) 0.000 -0.010
[0.004] [0.036] TARP recipient x Equity Risk ($) 0.000 0.021
33
Table 9
Return on assets (ROA), Return on Equity (ROE), CEO annual compensation in the period 2004-2009 for Banks with CEO Duality.
This table shows the results of OLS estimations with inclusion of firm-fixed and time-fixed effects. Columns 1 (2) shows results from return on assets (return on equity), which is regressed on CEO annual cash bonus, ownership incentives and equity risk sensitivity. "Cash bonus/salary" is the ratio of cash bonus to salary. "Dollar gain from +1%" is the change in dollar value of CEO's equity portfolio respective to 1% change in stock price. "Ownership (%)" is the sum of all shares, restricted and unrestricted, and delta-weighted options, exercisable and unexercisable, which is divided by total amount of outstanding shares. "Equity Risk ($)" is the value change (in dollar) of CEO's equity portfolio respective to a 1% increase in stock volatility. "Equity Risk (%)" is the percentage change in CEO's equity portfolio respective to a 1% increase in stock volatility, and is measured based on all options held by the CEO. Control variables are natural logarithm of market capitalization, Tier 1 Capital Ratio, book-to-market ratio, CEO duality and leverage. CEO duality is a dichotomous variable, which is valued 1 if the CEO is also a chairman in the firm, 0 otherwise. Lagged Returns is lagged return on assets (equity) for column 1 (2), All variables are measured at the end of the year. Standard errors are reported in parentheses. Significances at 0.01, 0.05 and 0.1 level are indicated by ***, ** and * respectively.
ROA ROE 1 2 3 4 Duality 0.978** 0.656 8.583** 6,849* [0.420] [0,433] [4.230] 4.067 Cash bonus/salary -0.047 -0.072 0.912 0.855 [0.122] [0,212] [1.227] 1.990 Duality x Cash bonus/salary 0.055 0.117 -0.794 -0.641
[0.120] [0,201] [1.208] [1.88] Dollar gain from +1% 0.002*** 0,002*** 0.009 0.008 [0.001] [0,001] [0.007] [0.01] Duality x Dollar Gain from
+1% -0.002** -0,002** -0.005 -0.004
[0.001] [0,001] [0.007] [0.01]
Equity Risk ($) 0.013 0.007 0.095 0.127
[0.008] [0,01] [0.084] [0.10] Duality x Equity Risk ($) -0.013 -0.006 -0.079 -0.118 [0.009] [0,011 [0.089] [0.11] Lagged Returns -0.088 -0.145 -0.124 -3,09** [0.121] [0.13] [0.111] [1.22] Book-to-Market -0.333 -0,524* -0.275 -5,256** [0.235] [0.27] [2.352] [2.51] Market Value 0.912*** 1,289*** 5.856** 1,536*** [0.246] [0.31] [2.630] [2.95]
Tier 1 Capital Ratio 0.086 1,111*
34
4.5 Robustness Check
In order to test the robustness of my regression, I do some robustness check. First, following Fahlenbrach and Stulz (2011), I use tangible common equity ratio as one of the control variables, instead of Tier 1 capital ratio to measure bank’s capital strength. The results are reported in Table A2 and A3 in Appendix section.
In trying to address the endogeneity issue that may appear in the previous regressions, I follow Aebi et al. (2011) approach by omitting lagged returns as one of the control variables for all the regressions. The results are reported in Table A4 to A8 in Appendix section.
For both robustness check, the results remain qualitatively similar to the previous regression results.
5. Discussion and Conclusion
The paper analyzed the if CEO Compensation package induce excessive risk taking that resulted in financial crisis 2007/2008. Previous handful literatures on this topic generated mixed result. Thus, I want to contribute to the existing literatures by trying to extend the research in various ways, such as adding some relevant variables and do some non-linear regressions. I also re-estimate all my regressions to conduct robustness check for my regressions.
According to the results, I find that although not to be fully blamed, the alignment situation between CEO compensation and shareholders’ interests during my sample period, is one of the factor that contribute to the financial crisis or bank performance during crisis. This result supports the view of Bebchuk, Cohen and Spamann (2010), and Bhagat and Bolton (2014) paper and opposes to what Fahlenbrach and Stulz (2011) found. Bebchuk, Cohen and Spamann (2010) in particular, pointed out that unfortunately, the compensation arrangement before the crisis encouraged risk taking instead of avoiding it.
In my results, the variable that is directly linked to the stock option (dollar gain from +1%) is almost positively significant to the bank performance in all my regressions. This means that in particular, the compensation in a form of stock option did influence CEOs to take on more risk and as a result, contributed to financial crisis. This has been predicted by Chen et al. (2006) that found stock option compensation plan induced risk taking.
35 regulation on compensation, or monitoring banks will not wholly eliminate the problem completely. If not, it may possibly create another problem that may not have been observed yet. Aside from regulation, the most important thing Is to monitor and ensure the banks comply to it. After letting the incentive measure variables interacted with TARP recipient variable, I found that there is no different impact on banks that received TARP funding, and those who did not, in terms of bank performance. This supported Fahlenbrach and Stulz (2011) results.
While examining CEO duality approach, I found that banks with this approach performed worse than banks with non-CEO duality approach. Hence, the system should be reviewed for the long-run sustainability and to decide what is the best approach for this. While some shareholders are really keen to end the CEO-duality system, it may actually work if there are some changes taken, for example, a senior independent director acting as a non-executive chairman should be appointed12.
Overall, it should not be surprising that CEO incentive did somehow contribute to financial crisis, since CEO enjoyed the upside of taking risk, while not suffer any losses personally during the downside. In addition, the regulation on risk taking were too relaxed during this period13.
6. Limitation and Further Research
In this study, I did not focus on corporate governance variables, such as CEO age, education, tenure, and turnover. Those characteristics may influence the results, as there are extensive literatures using those characteristics. For example, CEO who is hired from the outside may be paid more compare to promoting within the company due to a premium paid for hiring externally14. Similar to that, I also did not take into account the board composition, or the non-executive composition or their compensation in this study and mainly focus on CEO compensation. Sanders & Hambrick (2007) argued that CEO is not the only one to decide firm’s action on risk taking. Similarly, Berger et al (2016) found that shares ownership of CEO did not influence bank failure directly.
12 Should banks keep combined the role of CEO and chairman? 1 May, 2013. Source : https://www.ft.com/content/fb849930-b251-11e2-8540-00144feabdc0
13 Executive Pay and The Financial Crisis, 31 January 2012. Source:
http://blogs.worldbank.org/allaboutfinance/executive-pay-and-the-financial-crisis
37
REFERENCES
Abobakr, G, M., Elgiziry, K., 2017. The Relationship Between Board of Directors’ Characteristics and Bank Risk-Taking: Evidence from Egyptian Banking Sector. Journal of Finance and Accounting 5 (1), 24-33.
Adams, R., Ferreira, D., 2009. Women in the boardroom and their impact on governance and performance. Journal of Financial Economics 94, 291-309.
Aduda, J., (2011). The Relationship Between Executive Compensation and Firm Performance in the Kenyan Banking Sector. Research Paper, Department of Finance and Accounting, School and Business, University of Nairobi, Kenya.
Aebi, V., Sabato, G., Schmid, M., 2012. Risk management, corporate governance, and bank performance in the financial crisis. Journal of Banking & Finance 36(12), 3213-3226.
Alin, A., 2010. Multicollinearity. Wiley Interdisciplinary Reviews: Computational Statistics, 2(3), 370–374.
Allison, P.D., 2009. Fixed effects regression models (Vol. 160). SAGE publications.
Ang, J., Cole, R., Lin, J., 2000. Agency costs and ownership structure. Journal of Finance 55(1), 81-106.
Augulyte, S., 2016. Equity-Based CEO Compensation, Risk Taking and Firm Financial performance. Unpublished working paper. University of Groningen.
Balachandran, S., Kogut, B., Harnal, H., 2010. Did executive compensation encourage excessive risk-taking in financial institutions? Unpublished working paper. Columbia University.
Baliga, B.R., Moyer, R.C., Rao, R.S., 1996. CEO duality and firm performance: What's the fuss?. Strategic Management Journal, 41-53.
Banker, R. D., Darrough, M. N., Huang, R. and Plehn-Dujowich, J. M., 2013. The Relation between Ceo Compensation and Past Performance. The Accounting Review 88(1), 1–30.
Barclay, M.J., Gode, D., Kothari, S.P., 2005. Matching delivered performance. Journal of Contemporary Accounting & Economics 1(1), 1-25.
38 Bebchuk, L. A., Cohen, A. Spamann, H., 2010. The Wages of Failure: Executive Compensation at Bear Stearns and Lehman 2000-2008. Yale Journal on Regulation 27, 257-282. Harvard Law and Economics Discussion Paper 657.
Bebchuk, L. A., Fried, J. M., 2003. Executive Compensation As an Agency Problem. Journal of Economic Perspectives 17(3), 71–92.
Bebchuk, L. A., Spamann, H., 2009. Regulating Bankers' Pay. Georgetown Law Journal 98(2), 247-287. 2010, Harvard Law and Economics Discussion Paper No. 641.
Beltratti, A., Paladino, G., 2015. Bank leverage and profitability: Evidence from a sample of international banks. Review of Financial Economics, 27, 46-57.
Beltratti, A. and Stulz, R.M., 2012. The credit crisis around the globe: Why did some banks perform better?. Journal of Financial Economics 105(1), 1-17.
Berger, A.N., Bouwman, C.H., 2013. How does capital affect bank performance during financial crises?. Journal of Financial Economics 109(1), 146-176.
Berger, A.N., Imbierowicz, B. and Rauch, C., 2016. The roles of corporate governance in bank failures during the recent financial crisis. Journal of Money, Credit and Banking 48(4), 729-770. Bhagat, S. and Bolton, B., 2014. Financial crisis and bank executive incentive compensation. Journal of Corporate Finance 25, 313-341.
Black, L.K. and Hazelwood, L.N., 2013. The effect of TARP on bank risk-taking. Journal of Financial Stability 9(4), 790-803.
Boyd, J.H., Runkle, D.E., 1993. Size and performance of banking firms: Testing the predictions of theory. Journal of monetary economics 31(1), 47-67.
Boyd, B. K., 1994. Board Control and Ceo Compensation. Strategic Management Journal 15(5), 335–344.
Brickley, J.A., Coles, J.L., Jarrell, G., 1997. Leadership structure: Separating the CEO and chairman of the board. Journal of corporate Finance 3(3), 189-220.
Carrasco, V., 2005. Corporate Board Structure, Managerial Self-Dealing, and Common Agency. Stanford Institute for Economic Policy Research, 04-15.
