CEO compensation and bank stability during the great recession
Marvin Altena S3841456
University of Groningen, Faculty of Economics and Business
Supervisor: dr. J.J. Bosma Co-supervisor: dr. A. Dalò
Number of words: 5074 Abstract:
Stability of financial institutions is important to society. Systemic risk can be a threat to stability in times of crises. Research shows that CEO’s who are incentivised to take more risk through compensation packages are not necessarily correlated to worse stock performance of the institutions they lead during the great recession (Fahlenbrach and Stulz, 2011).
Findings in this thesis show that risk-stimulating incentives are significantly correlated to instability in some instances, but the results are ambiguous. Future research could use other measures of stability or look at non-CEO executives’ compensation incentive structures as well.
1
Introduction
‘The relentless focus on how much CEOs are paid diverts public attention from the real problem—how CEOs are paid.’ Jensen & Murphy (1990)
During and after the credit crises many financial institutions experienced financial distress.
Banks had large exposures to assets that lost their value quickly. This threw a spanner in the works of the financial system and led to the bankruptcy of Lehman Brothers and a
slowdown of the world economy.
Very few people correctly predicted the great recession. Bank executives perhaps were best positioned to prevent it from happening. In this thesis the relation between executive remuneration and bank stability in the USA is explored.
CEO compensation
CEO compensation peaked in the year 2000 with dips following the burst of the dot-com bubble and the great recession.1 The CEOs included in the sample for this thesis had an average annual compensation package of 7,2 million in 2006, the last complete fiscal year before the crisis started.
Executives might take more or less risk for short term gains based on the type of
compensation they receive. Compensation packages generally consist of multiple forms of remuneration. Compensation committees indicate what the best mix of these different kinds of pay is in order to maximise shareholder value without taking excessive risks.2 This aligns the incentives of the executives with those of the shareholders.
The Basel Committee on Banking Supervision (2015), known for the Basel accords, published as a corporate governance principle that “the bank’s remuneration structure should support sound corporate governance and risk management.” This highlights that executive
compensation is linked to their risk-taking behaviour and functioning.
Executive compensation generally consists of a base salary, a yearly bonus based on accounting performance, stock options and long-term incentive plans (Murphy, 1999).
These four components of compensation bind the executive to the company in different ways. For example, restricted stock holdings incentivise the executive to create value in the long-term whereas stock options give the incentive to have a high stock price at the exercise date. Yearly bonuses incentivise the executive to show short-term growth.
In the 1990’s stock options became the largest part of most executives’ compensation packages where the base salary was largest before. Bonuses for exceptional performance can have an adverse effect when risky trades are made in order to seek outperformance, as was the case with a rogue trader at Société Générale (Kirkpatrick, 2009). In the sample for this thesis the average annual base salary was just over 10% of the total annual
compensation package in 2006. For the average bonus this was around 26% that same year.
It is mandated that “under Dodd-Frank Section 956, banks with assets greater than $50 billion must ensure half of total incentives granted (annual and long-term) are deferred”3.
1 Epi.org, CEO compensation 2018. Consulted in June 2020.
2 Deloitte.com, Compensation Committee Guide. Consulted in June 2020.
3 meridiancp.com, Executive compensation in the banking industry. Consulted in June 2020
2
Research by Fahlenbrach and Stulz (2011) has found that banks with CEO’s whose interests were better aligned with those of shareholders did not perform better during the great recession than other banks. In the setup of their research, as seen from the principal-agent lens, the agent was the CEO and the shareholders were the principal. Principal-agent theory suggests that a CEO whose incentives are better aligned with those of shareholders will act in a way that is more beneficial to shareholders. The lack of evidence for this claim suggests that bank executives did not see the crisis coming. In this research the principal is not the shareholder, but society as a stakeholder.
Stability and systemic risk
Society as a whole is a stakeholder of a robust financial system. A lack of stability can cause increased default risk of a financial institution or panic amongst the public.4 Executives manage and lead the banks activities and are incentivised by the compensation packages they receive. This thesis examines the relationship between bank CEO compensation and the stability of banks.
There is no single correct measure for stability of a financial institution. Methodologies that try to measure stability involve the idea of systemic risk. Systemic risk is impairment of the entire financial system due to distress somewhere in the financial system, possibly having negative effects on other sectors (Adrian & Brunnermeier, 2016).
Banks do not only carry their individual risks, but they also depend on other banks for their business model (Lehar, 2005). Systemic risk is caused by interlinkage between firms directly due to contractual obligations and indirectly through the market price channel and
information spillovers (Clerc et al. 2016). Distress in the financial sector can lead to a compound negative effect on other sectors, as Gray, Merton & Bodie (2007) have shown.
Firms directly and indirectly interact with each other, meaning distress in the market increases the chance of distress for the individual firm (Sergoviano & Goodhart, 2009;
Adrian & Brunnermeier, 2011). Besides individual firm characteristics, the degree of
interconnectivity in the network of financial institutions is also relevant for the formation of systemic risk (Acemoglu, Ozdaglar, & Tahbaz-Salehi, 2015).
A firm’s contribution to systemic risk increases more than linearly with an increase in size (Tarashev, Borio & Tsatsaronis, 2009). This relatively increased contribution of large firms to systemic risk can be caused by lower capital, less-stable funding, more engagement in market-based activities or organisational complexity (Laeven, Ratnovski, & Tong, 2014). The sample used in this thesis contains large institutions, which means they are important contributors to systemic risk.
An institutions systemic importance is related to its individual risk, exposure to common risk factors and size concentration in the market (Tarashev, Borio & Tsatsaronis, 2009). A large financial institution can be individually systemically important and smaller firms can be systemically important as a herd (Adrian & Brunnermeier, 2016).
4 theatlantic.com, The looming bank collapse. Consulted in June 2020.
3
The amount of risk an executive is willing to take can be an influence on the aforementioned characteristics. The relationship between risk-incentivising compensation and risk-taking behaviour in the banking industry is confirmed by Chen, Steiner, & Whyte (2006).
Since the great recession, there has been increased attention for banks and regulation on banks. Regulators realise that distress and defaults in one financial institution can have a cascading negative effect throughout the sector. The Basel accords have aimed at reducing risk in banks and the banking system and the same is true for the Dodd-Frank Act. Both have provisions on executive pay.
Idier, Lamé & Mésonnier (2014) found that the tier one capital ratio and share of non- performing loans are most useful for predicting equity losses during the great recession for their sample of banks. They compared these simple balance sheet ratios to a complex measure of systemic risk called the Marginal Expected Shortfall and conclude that it’s not the best suited option for supervisory purposes. As a measure of stability in this thesis, the conditional value at risk (CVaR) is used. This is an improvement on value at risk (VaR): a measure that banks have to report on to comply with the Basel ll and lll accords (Sharma, 2012).
VaR is defined as a percentile of largest losses in the historical return distribution. This gives an indication of the left tail, but it gives a plain value. The CVaR returns the average of losses below this percentile. This gives insights into the distribution of the tail beyond the given percentile.
The question that is answered in this thesis is: was risk incentivising CEO compensation correlated to higher instability for banks during the great recession?
The hypothesis that is tested in this thesis is whether CEO compensation packages that are more risk stimulating were correlated to lower stability of banks during the great recession.
This thesis will continue with a methodology section. After that the data used for this research are discussed and some summary statistics are presented. The part that follows contains the results of the analysis. The conclusion is the last part of the thesis. Limitations can be found here as well.
Methodology
This thesis utilises applied game theory in combination with statistical analysis of data on banks and their CEO’s compensation structure. The setup of the research is modelled after that of Fahlenbrach & Stulz (2011). They used stock returns during the crisis as the explained variable. This thesis uses the stability of the financial institutions as the explained variable.
Incentives
Every publicly traded company has to deal with the principal-agent problem: desires of owners may be different from those of the managers. Owners of companies typically want a share in a company that will rise in value over time. Directors running this company have a fiduciary responsibility towards shareholders to work towards this goal.
To align incentives of the CEO to the company’s shareholders, the compensation package of the CEO can be optimized. This is done using a combination of different forms of
compensation, the main four being: salary, bonuses, stock grants and option grants.
4
Different forms of compensation can incentivise the CEO in different ways:
- The amount to which salary aligns incentives goes as far as the threat of not receiving it if the CEO is replaced.
- A yearly bonus can incentivise a CEO to work on the metric that the amount of the bonus is based on. This is usually some measure of accounting performance. Since a bonus cannot be negative, a CEO may be willing to take risky decisions due to the limited downward effect (Fahlenbrach & Stulz, 2011).
- Compensation in the form of stock options will incentivise the CEO to have a high stock price on the exercise date of these options as options have more value when they are deeper ´in the money´.
- Stock with a long minimum required holding period aligns the incentives of a CEO with that of long-term shareholders.
Stability measures
Examples of systemic risk measures are the CoVaR by Adrian & Brunnermeier (2016) and marginal expected shortfall measures by Acharya et al. (2017) and Brownlees & Engle (2011).
Marginal expected shortfall is a measure used by Idier, Lamé & Mésonnier (2014) to predict the collapse of banks ex post. It performed worse than simple balance sheet measures like the tier 1 capital ratio or the non-performing loan ratio.
In this thesis, the conditional value at risk is used. CVaR is a measure of the tail risk in the distribution of returns of a particular asset. It is an improvement on the value at risk (VaR) measure (Rockafellar & Uryasev, 2000).
The VaR measure gives the worst expected loss given a certain confidence level. For example, a 5% VaR gives a value that the loss will only exceed 5% of the time based on historical data. The VaR is a widely used idiosyncratic risk measure. It is prescribed in the Basel ll and lll accords and needs to be calculated over a period of at least twelve months (Sharma, 2012). Artzner et al. (1999) conclude that the VaR has nonsubadditivity problems and therefore reject the method.
The CVaR (see equation 1) is defined as the average of losses below a certain percentile, taking into account possible outliers in the tail of the distribution and the shape of the tail.
In this thesis this is the method that is used as an indicator of the stability of a bank.
𝐶𝑉𝑎𝑅𝑖𝑡𝑎 = −𝐸(𝑅𝑖𝑡 |𝑅𝑖𝑡 ≤ 𝑞𝑎 ) (1)
Conditional value at risk for company i over a given time period is -1 times the average of observed daily returns as long as these returns are below predetermined quantile a. This is given in figure 1. Commonly used quantiles are 1%, 5% and 10%. In this research, the 5%
leftmost observations on the profit distributions are used to calculate the CVAR for all included firms over the period of 1-7-2007 to 31-12-2008.
The CVaR used in this thesis is based on the individual returns of a financial institution during a set period and does not incorporate the difference between normal times and during event of distress like Adrian & Brunnermeier’s (2016) CoVaR measure. Because of
5
that, the CVaR is an imperfect measure of systemic risk. This is one of the limitations of this thesis.
Data
Data required for this research includes detailed information on the compensation that CEO’s of banks receive and data on performance of banks’ stock for fiscal years 2004 up to 2009.
The U.S. Securities and Exchange Commission (SEC) requires that public companies disclose how much their executives get paid.5 Furthermore, CEO compensation is under scrutiny from regulators and shareholders.
Fahlenbrach and Stulz (2011) used the Compustat ExecuComp database for their research.
This database contains executive compensation data from 1992 onward for companies in the S&P 1500 index.6 This thesis utilises data from that same source for details on
compensation. Data on firm characteristics are sourced from the Compustat databases as well. Time series data of stock performance for banks in the sample are sourced from Eikon – Thomson Reuters.
The data on CEO compensation are an unbalanced panel of variables at the end of the fiscal year for the period 2004 to 2009 for 93 banks. The institutions included in the sample are listed in appendix 1. The sample includes only banks and thus excludes firms that operate in sectors like financial advice. All variables except tier 1 capital ratio have 93 observations for the summary statistics in table 1. The reason that tier 1 capital ratio has 80 observations is that this variable is not applicable for investment banks.
As can be observed in table 1, the sample consists of large companies with a median market capitalization of 2.685,52 million. Tier 1 capital ratio and tangible common equity ratio are measures of capital strength. The regulatory minimum tier 1 capital ratio was 4%
(Fahlenbrach & Stulz, 2011) and with a minimum of 6,35% every bank complies to this rule.
None of the banks in the sample had a negative income in 2006. It can also be seen that there is a small number of banks that are much larger than the others from the standard deviation being higher than the mean and the distribution of the quartiles.
Variables in table 1 describe the sample of institutions. In table 2 summary statistics on the compensation of the CEO’s of these banks are presented. These data are observed at the end of fiscal year 2006 and are sourced from the same ExecuComp database as the firm characteristics.
5 sec.gov, Executive compensation. Consulted in May 2020
6 rug.nl, ExecuComp. Consulted in May 2020
6
Variable Number Minimum
Lower
quartile Median
Upper
quartile Maximum Mean
Standard deviation Total assets (mln $) 93 1885.96 6431.80 14669.73 57064.90 1884318.00 141342.61 337641.13 Total liabilities (mln $) 93 1753.69 5760.72 13365.04 50783.81 1764535.00 131221.98 314958.14 Market capitalization (mln $) 93 64.24 1190.28 2685.52 12290.66 273691.19 19222.07 45425.05
Net income/total assets 93 0.03% 0.88% 1.16% 1.45% 2.91% 1.19% 0.46%
Net income/book equity 93 0.33% 9.88% 13.01% 16.63% 26.65% 13.44% 5.36%
Cash/total assets 93 0.35% 1.63% 2.24% 2.75% 8.17% 2.30% 1.23%
Dividend per share ($) 93 0.00 0.40 0.81 1.28 3.54 0.97 0.72
Book-to-market ratio 93 0.27 0.42 0.50 0.64 31.57 0.87 3.22
Tier 1 capital ratio 80 6.35% 8.57% 9.56% 11.26% 19.04% 9.83% 1.98%
Tangible common equity ratio 93 0.00% 4.73% 6.07% 7.08% 22.42% 6.18% 2.84%
Table 1. Summary statistics of the bank included in the sample at the end of 2006.
The values displayed for total compensation are the figures that were reported in SEC filings.
The cash bonus variable consists of the bonus plus nonequity incentives. The dollar value of both the annual stock grant and the annual option grant are as reported following the FAS 123R guideline and discloses the costs that were charged to the company for that year7. The value of total equity portfolio is the total of shares owned by the CEO with options included. Value of exercisable options is the estimated value of exercisable options that a CEO holds and that are in the money. Value of unexercisable options is the estimated value of unexercisable options that are in the money. Value of unvested restricted stock holdings is the part of the portfolio of a CEO that cannot be sold freely. Value of total equity
portfolio/total annual compensation is the value of all equity holdings including options divided by the total amount of annual compensation. Value of shares/salary is an indicator of stock holdings over salary.
Percentage ownership shares is different from percentage ownership in that the former reflects ownership only as a part of shares owned and the latter also includes option if they were to be exercised. Dollar gain from +1% is the amount with which a CEO’s portfolio will increase if the stock price goes up by 1%.
Percentage equity risk is the percentage return a CEO may have on his total portfolio if volatility of the stock price increases with one percent. Dollar equity risk is the dollar amount that this return would amount to.
7 takming.edu.tw, ExecuComp data definitions by table. Consulted in June 2020.
7
(dollar amounts x $1000) Mean Median
Annual compensation
Total compensation 7211.61 2657.21
Salary 755.67 741.6
Cash bonus 1883.1 636.77
Dollar value of annual stock grant 2049.41 293.07
Dollar value of annual option grant 1785.69 431.96
Other compensation 254 119.49
Cash bonus/salary 2.21 0
Equity portfolio value
Value of total equity portfolio 71722.46 20614.23
Value of shares 43173.50 3104.38
Value of exercisable options 15946.01 4723.73
Value of unexercisable options 1767.49 264
Value of unvested restricted stock 10835.46 186.49
Value of total equity portfolio/total annual compensation 15.31 6.19
Value of shares/salary 82.75 4.57
Equity portfolio incentives
Percentage ownership from shares 1.48 0.25
Percentage ownership 2.07 0.7
Dollar gain from +1% 1107.46 322.67
Equity portfolio risk exposure
Percentage equity risk 0.12 0.09
Dollar equity risk 154.68 18.03
Table 2. Summary statistics of CEO’s compensation in 2006.
For most of the variables the mean is much higher than the median. This is caused by either outliers or a broader distribution of observations above the median. It suggests that there is a small number of firms that pay their CEOs significantly more. The percentage stock
ownership by the CEO is also much larger for a small number of firms. From the large differences between the mean and median it can be concluded that there are large
differences between the values of annual compensation packages for CEOs, as was the case for the institutions themselves.
As can be deducted from table 2, the median salary makes up 28% percent of the total compensation package and for the mean this is slightly higher than 10%. The average cash bonus is a multiple of the average salary.
The average total stock and options holdings of a CEO are 71,7 million dollars with a median of 20,6 million. Notably, the value of this portfolio is on average 15 times what is earned in a year. The largest part of the portfolio consists of freely tradable shares and exercisable options. The dollar gain from a stock price increase of 1% gives the average CEO a dollar return of 1,1 million dollar. A 1% increase in the volatility of the stock price gives the average CEO a dollar return of 155.000.
8
CVaR Cash
bonus/salary Dollar gain from +1%
Ownership (%)
Equity risk ($)
Equity risk (%)
Stock return in 2006
Book- to- market
Log (market value)
Tier 1 capital ratio
CVaR 1
Cash bonus/salary 0.2276* 1
Dollar gain from +1% 0.1451 0.4104* 1
Ownership (%) 0.0286 0.0322 0.1021 1
Equity risk ($) 0.0309 0.1409 0.9045* -0.0273 1
Equity risk (%) 0.1035 -0.1049 0.0572 -0.2973* 0.2191 1
Stock return in 2006 0.0356 0.2670* 0.0349 -0.199 0.0948 0.042 1
Book-to-market 0.3112* -0.0213 -0.0356 -0.0348 -0.0157 -0.1209 0.0674 1
Log (market value) 0.2211 0.2713* 0.4131* -0.2371* 0.3293* 0.2547* 0.2041 0.2768* 1
Tier 1 capital ratio 0.3708* -0.0754 -0.1169 0.1998 -0.1201 -0.0192 -0.194 0.0423 -0.4187* 1
Table 3. Correlation between regression variables. * p<0.05
Table 3 shows the variables that are used for the regression in the results section. Market value is significantly correlated to most variables with CVaR and 2006 stock return being the exceptions. Some of the other variables appear to be related to each other too. Notably, stability measure CVaR is significantly correlated to cash bonus/salary, book-to-market and Tier 1 capital ratio.
Results
In this part of the thesis the relation between CEO pay/incentive and bank stability is explored. The period examined to calculate the CVaR matches the period Fahlenbrach &
Stulz (2011) used for the buy and hold returns. The period starts on 1-7-2007 and ends on 31-12-2008. Arguably, the crisis did not end in 2008, but in order to compare findings with those of Fahlenbrach & Stulz (2011) this is still the most relevant period.
The CVaR can be calculated using different period lengths and different confidence levels.
For the purpose of this paper the 5% confidence level is used and the minimum required period of one year of historical daily data (Sharma, 2012) is also satisfied.
The difference between long term and short-term incentives is represented by different variables, which can be seen in table 4. For this regression the explained variable is the conditional value at risk during the financial crisis and the explanatory variables are all observations from fiscal year 2006.
The Cash bonus / salary variable in table 4 is an indicator for short term incentives. A higher bonus relative to salary should be correlated with a higher willingness to take on risk
according to agent-principal theories. The dollar gain from +1% variable is the increase in the CEO’s portfolio in dollars when the stock price rises 1%. Ownership % is the percentage of company stock that is owned by the CEO including options as part of all outstanding stock. A higher percentage of ownership of shares in the company by the CEO may indicate that a CEO’s incentives are more aligned with those of shareholders. Equity risk ($) is the change in value of a CEO’s portfolio when volatility rises with 1%. Equity risk (%) is the percentage change in value of a CEO’s portfolio when volatility rises with 1%. The other four variables are firm characteristics that are controlled for in multiple regressions 6 through 8.
9
Dependent variable:
CVaR
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Cash bonus/salary 0.2301*** -0.1331 0.1526** -1.5156 -0.6703**
(0.0481) (0.1302) (0.0735) (1.3588) (0.3041)
Dollar gain from +1% 0.0004 0.0026** 0.0012
(0.0005) (0.0012) (0.0012)
Ownership (%) 0.0490 0.1005 0.1638
(0.0951) (0.1063) (0.1408)
Equity risk ($) -0.0003 -0.0096** -0.0023
(0.0002) (0.0037) (0.0049)
Equity risk (%) -7.9989 -7.4948 -3.5834
(8.9747) (8.7220) (7.0867)
Stock return in 2006 0.0233 -0.0443 0.0346 0.0249
(0.0482) (0.0812) (0.0628) (0.0589)
Book-to-market 0.9930*** 1.0075*** 10.8659** 10.5035*
(0.0910) (0.1142) (5.1817) (5.4013)
Log (market value) 1.4597** 1.8385** 0.1994 0.6856
(0.6760) (0.7686) (0.6191) (0.5954)
Tier 1 capital ratio -1.5286** -1.5348**
(0.7141) (0.6970)
Observations 78 79 79 79 79 77 77 67 67
R-squared 0.0518 0.0211 0.0008 0.0010 0.0107 0.3490 0.2556 0.2403 0.2473
Table 4. Regression of variables on the conditional value at risk. *** p<0.01, ** p<0.05, * p<0.1 Table 4 contains the results of OLS regressions between variables and the CVaR. The CVaR is the average of the 5% worst daily stock price declines in the examined period of July 2007 to December 2008, as a positive number. A higher CVaR indicates higher average losses in the left tail of a bank’s performance. The row variables concern observations at the end of 2006, the last fiscal year end before the crisis started.
The equation that is predicted in regression 1 to 5 is:
𝐶𝑉𝑎𝑅 = ß0+ ß1∗ 𝐼𝑉 (2)
The CVaR is conditional value at risk, a constant coefficient is included and IV is a coefficient for an incentive variable. For regression 6 to 9, the predicted model is:
𝐶𝑉𝑎𝑅 = ß0+ ß1∗ 𝐼𝑉1+ ß2∗ 𝐼𝑉2+ ß3∗ 𝐼𝑉3+ ß4∗ 𝑆𝑅 + ß5∗ 𝐵𝑡𝑀 + ß6∗ ln(𝑀𝑉) + [𝑓𝑜𝑟 𝑟𝑒𝑔𝑟𝑒𝑠𝑠𝑖𝑜𝑛 8 & 9: ] ß6∗ 𝑇1 (3)
where ß0 is the constant, the instrumental variables are Cash bonus/salary, Dollar gain from +1% and Equity risk ($) for regression 6 and 8 and Cash bonus/salary, Ownership (%) and Equity risk (%). The model also includes coefficients for firm characteristics stock return (RS), book-to-market ratio (BtM), the natural log of the market value and in the case of
regression 7 and 9 the tier 1 capital ratio (T1).
The first five regressions examine the individual results of CEO incentive variables. Short- term incentives are represented by the cash bonus / salary variable. A CEO is expected to act more short term focussed if his bonus makes up a large part of his compensation package. Cash bonus / salary is highly significant when no control variables are used (regression 1), and also with controls in regression 7 and 9. In regression 6 and 8 however, the control variables render the result insignificant.
Regression 1 and 7 give a positive value for the cash bonus / salary variable, indicating that a pay package with a relatively large part consisting of bonuses is correlated with higher average losses in the tail of the bank’s profit distribution. Interestingly, in regression 9 the
10
variable is statistically significant and negative. Besides cash bonus / salary none of the other incentive variables are significant on their own (regression 2 – 5).
The stock return in 2006 is not correlated significantly with the CVaR during the crisis. Other research on the topic suggest that a correlation between poor stock performance during the crisis and good stock performance before the crisis could mean “that banks that took on more exposures that the market rewarded in 2006 suffered more during the crisis”
(Fahlenbrach & Stulz, 2011, pp 19). If this was the result of market co-movements, we would have seen significant positive estimators for the 2006 performance variable due to the market performing worse during the crisis than in 2006.
Book-to-market is statistically significant and positive in all regressions. Banks with a higher book-to-market value on average suffered worse losses in the distribution of their daily stock returns during the crisis. These banks were less stable. This might indicate that investors were right on average. The higher market price compared to the book value suggests confidence of investors in the company and its future prospects and indeed these banks were more stable during the great recession.
Due to missing observations the number of banks included in the sample differs for the regressions. The tier 1 capital ratio variable causes all nondepository financial institutions to be excluded from the regression. In the regression that includes only depository banks the beta for cash bonus/salary is negative. For that sample a bank was more stable if the CEO’s bonus was larger in relation to salary. This is evidence against the claim that bonuses incentivise CEO’s to take more risk, as the bonus has a lower bound of 0 and no defined upper bound. Regressions 1 and 7 show a significant positive beta for this variable and thus support the claim.
In regression 6 and 7 the market value variable is significant and positive, but when tier 1 capital ratio is included in regression 8 and 9 market value becomes insignificant and the negative tier 1 capital ratio estimator takes on significance. For fiscal year 2006 there is a strongly significant negative correlation between tier 1 capital ratio and market value. With multicollinearity present and tier 1 capital ratio gaining significance when included suggests that this is the more correct variable. This is in line with Idier, Lamé, & Mésonnier (2014), who found that tier one capital is a good indicator for performance during the crisis.
Interestingly, in regression 6 a higher profit sourced from an increase in volatility is correlated with a more stable institution. This is not expected based on incentives, as the CEO has an incentive to take risk increasing volatility. The significant result does not hold for the sample controlling for tier 1 capital ratio, excluding investment banks.
11
Conclusion
Relationship between CEO compensation incentives and stock returns is much explored.
This thesis examined the relationship between a CEO´s compensation incentives and stability. This is done using the conditional value at risk as an indicator for stability and compensation data sourced from the ExecuComp database.
A stable banking sector is vital for society. Distress in a financial institution can have an effect on systemic risk in the financial sector, which can have compound effects on other sectors. Regulators recognise that the total risk of the banking sector is not equal to the sum of all individual institutions´ risk.
The regressions in this thesis consider the effect of incentives on stability, controlling for firm characteristics profitability, book-to-market, firm size and tier 1 capital ratio. The regressions show some significant results, but there are also multicollinearity problems present.
There is some evidence of higher instability for banks whose CEO’s incentives are aligned relatively poorly with the interests of shareholders. The evidence is ambiguous and significance is not robust to all different control variables. The complete sample of banks showed significant positive correlation between instability and a large bonus relative to salary. Bonuses can incentivise a CEO to take excessive risks. In the sample with only depository banks the effect switches signs.
Fahlenbrach and Stulz (2011) found no evidence that poorer incentives lead to poorer stock returns for shareholders. They concluded that CEOs whose incentives were more aligned with shareholders took risks promising good returns, but ended up with poor outcomes due to the great recession. Adding to that, Acemoglu, Ozdaglar, & Tahbaz-Salehi (2015) conclude that features that generate stability under certain conditions may be a source of systemic risk under other conditions, although they don’t link it to executive behaviour.
The conclusion of this thesis is that there is no conclusive evidence that bank CEOs whose incentives may incline him towards excessive risk taking took excessive risks, putting the stability of the financial system in danger.
This analysis is performed on data surrounding the financial crisis as there is much pressure on financial institutions during this time. Future research could focus on the relationship between CEO incentives and stability of banks in different periods and look for other predicted variables that may be influenced by CEO incentives.
The chosen dependent variable is a limitation of this thesis. Instead of the CVaR, the CoVaR by Adrian & Brunnermeier (2016) or the marginal expected shortfall measures by Acharya et al. (2017) and Brownlees & Engle (2011) could have been used. These measures are more sophisticated in that they include market performance or deltas between normal times and times of distress. Another limitation of this thesis is that only CEO’s incentive packages were considered. Future research could include non-CEO compensation in the analysis.
12
References
Acemoglu, D., Ozdaglar, A., & Tahbaz-Salehi, A., 2015. Systemic risk and stability in financial networks. American Economic Review, 105(2), 564-608.
Acharya, V. V., Pedersen, L. H., Philippon, T., & Richardson, M., 2017. Measuring systemic risk. The Review of Financial Studies, 30(1), 2-47.
Adrian, T., & Brunnermeier, M. K., 2016. CoVaR. The American Economic Review, 106(7), 1705.
Artzner, P., Delbaen, F., Eber, J. M., & Heath, D., 1999. Coherent risk measures.
Mathematical Finance, 9(3), 203-228.
Basel Committee on banking supervision, 2015. Corporate governance principles for banks.
Brownlees, C. T., & Engle, R., 2012. Volatility, correlation and tails for systemic risk measurement. Available at SSRN, 1611229.
Chen, C. R., Steiner, T. L., & Whyte, A. M., 2006. Does stock option-based executive
compensation induce risk-taking? An analysis of the banking industry. Journal of Banking &
Finance, 30(3), 915-945.
Clerc, L., Langfield, S., Giovannini, A., Portes, R., Scheicher, M., & Peltonen, T., 2016. Indirect contagion: The policy problem. ESRB, occasional paper series, No 9/January 2016.
Fahlenbrach, R., & Stulz, R. M., 2011. Bank CEO incentives and the credit crisis. Journal of financial economics, 99(1), 11-26.
Gray, D. F., Merton, R. C., & Bodie, Z., 2007. New framework for measuring and managing macrofinancial risk and financial stability (No. w13607). National Bureau of Economic Research.
Idier, J., Lamé, G., & Mésonnier, J. S., 2014. How useful is the marginal expected shortfall for the measurement of systemic exposure? A practical assessment. Journal of Banking &
Finance, 47, 134-146.
Jensen, M. C., & Murphy, K. J., 1990. CEO incentives—It's not how much you pay, but how.
Journal of Applied Corporate Finance, 3(3), 36-49.
Kirkpatrick, G., 2009. The corporate governance lessons from the financial crisis. OECD Journal: Financial Market Trends, 2009(1), 61-87.
Laeven, M. L., Ratnovski, L., & Tong, H., 2014. Bank size and systemic risk (No. 14).
International Monetary Fund.
Lehar, A., 2005. Measuring systemic risk: A risk management approach. Journal of Banking &
Finance, 29(10), 2577-2603.
Murphy, K. J., 1999. Executive compensation. Handbook of labor economics, 3, 2485-2563.
13
Rockafellar, R. T., & Uryasev, S., 2000. Optimization of conditional value-at-risk. Journal of risk, 2, 21-42.
Segoviano, M., & Goodhart, C., 2009. Banking stability measures. IMF working papers, 1-54.
Sharma, M., 2012. Evaluation of Basel III revision of quantitative standards for
implementation of internal models for market risk. IIMB Management Review, 24(4), 234- 244.
Tarashev, N. A., Borio, C. E., & Tsatsaronis, K., 2009. The systemic importance of financial institutions. BIS Quarterly Review, September.
14
Appendix
Appendix 1
List of banks included in the sample
1 BANK OF NEW YORK MELLON CORP 48 PROVIDENT BANKSHARES CORP
2 BANK OF AMERICA CORP 49 CHITTENDEN CORP
3 WELLS FARGO & CO 50 TD BANKNORTH INC
4 PNC FINANCIAL SVCS GROUP INC 51 WASHINGTON FEDERAL INC
5 KEYCORP 52 EAST WEST BANCORP INC
6 SUNTRUST BANKS INC 53 FIRSTFED FINANCIAL CORP/CA
7 MORGAN STANLEY 54 STERLING BANCSHARES INC/TX
8 FIFTH THIRD BANCORP 55 UCBH HOLDINGS INC
9 FIRST HORIZON NATIONAL CORP 56 INDYMAC BANCORP INC
10 MARSHALL & ILSLEY CORP 57 FANNIE MAE
11 NORTHERN TRUST CORP 58 TRUSTCO BANK CORP/NY
12 WILMINGTON TRUST CORP 59 COLONIAL BANCGROUP
13 BEAR STEARNS COMPANIES INC 60 DIME COMMUNITY BANCSHARES
14 NATIONAL CITY CORP 61 BOSTON PRIVATE FINL HOLDINGS
15 CULLEN/FROST BANKERS INC 62 HUNTINGTON BANCSHARES
16 ZIONS BANCORPORATION 63 GREATER BAY BANCORP
17 ASTORIA FINANCIAL CORP 64 MERCANTILE BANKSHARES CORP
18 BB&T CORP 65 BANK OF HAWAII CORP
19 CITY NATIONAL CORP 66 CITIGROUP INC
20 COMERICA INC 67 HUDSON CITY BANCORP INC
21 TCF FINANCIAL CORP 68 FLAGSTAR BANCORP INC
22 POPULAR INC 69 BROOKLINE BANCORP INC
23 COMPASS BANCSHARES INC 70 UMPQUA HOLDINGS CORP
24 U S BANCORP 71 ASSOCIATED BANC-CORP
25 WASHINGTON MUTUAL INC 72 IRWIN FINANCIAL CORP
26 COUNTRYWIDE FINANCIAL CORP 73 WINTRUST FINANCIAL CORP
27 M & T BANK CORP 74 JPMORGAN CHASE & CO
28 SYNOVUS FINANCIAL CORP 75 CENTRAL PACIFIC FINANCIAL CP
29 LEHMAN BROTHERS HOLDINGS INC 76 STERLING FINANCIAL CORP/WA 30 MERRILL LYNCH & CO INC 77 PROSPERITY BANCSHARES INC
31 NEW YORK CMNTY BANCORP INC 78 CATHAY GENERAL BANCORP
32 UNITED BANKSHARES INC/WV 79 GLACIER BANCORP INC
33 SVB FINANCIAL GROUP 80 FRANKLIN BANK CORP
34 SUSQUEHANNA BANCSHARES INC 81 FIRST NIAGARA FINANCIAL GRP
35 SOUTH FINANCIAL GROUP INC 82 HANMI FINANCIAL CORP
36 MAF BANCORP INC 83 BANK MUTUAL CORP
37 ANCHOR BANCORP WISCONSIN INC 84 FIRST COMMONWLTH FINL CP/PA
38 JEFFERIES GROUP LLC 85 FIRST INDIANA CORP
39 FIRST MIDWEST BANCORP INC 86 INDEPENDENT BANK CORP/MI
40 WACHOVIA CORP 87 STERLING BANCORP/NY -OLD
41 DOWNEY FINANCIAL CORP 88 WILSHIRE BANCORP INC
42 FIRST BANCORP P R 89 CASCADE BANCORP
43 WESTAMERICA BANCORPORATION 90 FIRST FINL BANCORP INC/OH
44 WEBSTER FINANCIAL CORP 91 UNITED COMMUNITY BANKS INC
45 COMMERCE BANCORP INC/NJ 92 FIRSTMERIT CORP
46 INVESTORS FINANCIAL SVCS CP 93 CORUS BANKSHARES INC 47 GOLDMAN SACHS GROUP INC
