Stockholder Wealth Maximization during the Troubled Asset Relief Program Period: Is Executive Pay Harmful?

: This study investigates governance mechanisms and their relation to firm value, i


Introduction
Prior studies have scrutinized executive compensation as it is considered one of the effective mechanisms that can reduce agency conflict in firms (Jensen and Meckling 1976).The literature suggests that this mechanism may yield two opposing consequences.According to the optimal contracting theory (Mirrlees 1976), compensation plans might reduce the agency costs of publicly-held firms, thus increasing firm performance and ultimately enhancing shareholder wealth (Ke et al. 1999;He et al. 2022).Meanwhile, the managerial power hypothesis conjectures that executive compensation may underlie agency problems in companies (Jiraporn et al. 2005).Such circumstances indicate that executive compensation could create a bond between managers and shareholders if properly used but might engender managerial entrenchment and moral hazard if misused or dysfunctional.Equity-based compensation has been widely used since the 1980s (Frydman and Saks 2010) and positively affects firm performance (Mehran 1995).In examining whether compensation for bank CEOs became more sensitive to performance after a series of deregulations during the 1980s, several studies found that compensation levels rose slightly after deregulation (Crawford et al. 1995;Cuñat and Guadalupe 2009).Another study found that the relationship between executive compensation and firm performance only occurred in public insurance companies among 209 insurance companies (Ke et al. 1999), indicating that public companies with diffuse ownership need to utilize performance-based compensation to reduce agency problems.
However, the mortgage crisis sparked researchers' interest in executive compensation.A string of studies has examined the effect of executive compensation on bank performance during the mortgage crisis since it is considered one of the critical factors that triggered corporate scandals.In testing 112 Troubled Asset Relief Program (TARP) recipients, Phillips (2010) found that several performance measures were positively related to compensation in 2007 but not observable in 2006.Surprisingly, Phillips (2010) also reported that some performance indicators were negatively related to CEO compensation, suggesting that earnings-related fundamentals are less able to explain the compensation practices of TARP recipients.While banks and business media expressed concerns that TARP would eliminate the best talents, Nwaeze et al. (2018) showed that the TARP program improved bank performance by inducing CEO resignations as the TARP recipients enjoyed better performance after their CEOs exited.Their study also reports that TARP firms that experienced CEO resignation enjoyed higher returns in the year following their CEOs' exit, whereas other banks that provided pension benefits to compensate for the wealth-reduction effect of bank executives also enjoyed higher returns in the long run (Nwaeze et al. 2018).In addition, Fahlenbrach and Stulz (2011) found that executive compensation schemes, whether cash bonuses or stock options, were disadvantageous to bank performance during the crisis, even in banks with executives who owned more shares.Conservely, a more recent study documented that better performance was shown by companies that paid above the Treasury limit compared to companies that did not (Vega et al. 2020).Such findings demonstrate that the relationship between executive compensation and firm performance during the TARP period yields mixed evidence and remains inconclusive.Accordingly, this paper delves deeper into the relationship between executive compensation and firm performance amid more stringent regulations by examining ten events pertaining to executive compensation for the executives of TARP recipients.
This paper contributes to the literature in several ways.First, although the existing literature has studied executive compensation in the context of TARP, these studies focused on the relationship between executive compensation and post-deregulation firm performance (Chhaochharia and Grinstein 2009;Hubbard and Palia 1995;Phillips 2010;Nwaeze et al. 2018).In addition, several existing studies, such as Vega et al. (2020) and Fahlenbrach and Stulz (2011) examined the impact of executive compensation regulations on bank performance, yet those studies investigated the relationship between executive compensation and firm performance in more general events, e.g., the pre-TARP period and the post-TARP period.In contrast to prior studies, our paper investigates the relationship between executive compensation and firm performance in ten events to highlight market reactions toward TARP.Employing the Patell t-test, we find that the cumulative abnormal return (CAR) was significantly negative in the first three events, suggesting that the public was still unsure regarding the effectiveness of the government's stringent policies and legislation in limiting executive compensation.Conversely, we find that the market positively reacted to the IV, V, VI, VIII, IX, and X events, demonstrating that the public concurred that executives should not take bonuses as the firms struggled under the TARP, and was more optimistic about the more stringent and comprehensive executive regulations.Our evidence corroborates Nwaeze et al. (2018) as the market reacted positively to the more stringent and comprehensive governance regulations, suggesting that market participants perceived that the regulations could improve governance without harming the quality of managerial talents.These findings are consistent when using the standard event study method, ordinary least squares (OLS), and multivariate regression model (MVRM) for both TARP firms and 11 industry samples.
Second, we investigate the relationship between the changes in executive compensation, proxied by equity-based and cash-based pay, and shareholder wealth over the regulatory period.Utilizing pooled OLS, fixed effects, two-stage least squares, and threestage least squares, we find that an increase (decrease) in total compensation led to a decrease (increase) in shareholder wealth over the regulatory period.These results are consistent for both equity-based and cash-based compensation.Third, Vega et al. (2020) found that changes in CEO compensation for firms that left TARP produced negative and significant coefficients for both short-run and long-run estimates.Their findings indicate that investors perceived that those firms wanted to be free from restrictions under TARP.With pooled OLS, fixed effects, two-stage least squares, and three-stage least squares, we further examine this notion and find that the change in total compensation was negatively related to firm value for TARP firms that repaid the bailout, thereby supporting Vega et al. (2020).However, we find that this relationship was weaker than that for TARP firms that had not repaid the bailout funds, suggesting that the market was less responsive to firms that repaid bailout funds since investors might perceive these firms as safer and more capable of recovering quickly from financial distress.
We organize this paper as follows.Section 2 discusses the literature review and hypotheses.Research methods are presented in Section 3. Section 4 presents the findings and analysis, and we conclude this paper in Section 5.

Literature Review
The separation of ownership and control has provided management with power and discretion.The discretionary power could be abused by the agent as they might not act in the best interests of the principal (Jensen and Meckling 1976).This agency relationship has indeed become the starting point as well as the focal point in the study of corporate governance.To minimize agency costs, shareholders can utilize internal and external governance mechanisms (El-Chaarani et al. 2022).Internal mechanisms purport to improve the alignment between shareholders' interests and those of management by empowering the board of directors, setting value-maximizing compensation packages, employing leverage, and using other internal policies.On the other hand, external mechanisms rely on financial markets (stock and debt markets), the market for corporate control (mergers and acquisitions), the market for executive jobs, and other external influences.
Extensive discussion on corporate governance and regulations has motivated financial experts to conduct research on governance mechanisms and their relations to firm value.Previous work has highlighted this relationship using various aspects of corporate governance, i.e., compliance practices, board characteristics (Jaffar et al. 2023), ownership structures (El-Chaarani et al. 2022), political pressure (El-Chaarani et al. 2022), and shortselling mechanisms (Mai and Hamid 2021).Our study, however, focuses on executive compensation as a corporate governance mechanism.
The study on executive compensation during the financial crisis is compelling.The financial crisis put compensation structure at the forefront of many policy debates on the root causes of the crisis.Guo et al. (2015) investigated the impact of executive compensation in the 2007 banking crisis on larger banks, especially the too-big-to-fail banks.They found that bank risk and stock return volatility increased with the percentage of short-term and long-term incentive compensation.Interestingly, the too-big-to-fail banks actually had greater risk taking (lower Z-scores) and were more likely to experience financial distress than the smaller ones.Moreover, incentives generated by the compensation programs are correlated with excessive risk taking by banks (Bhagat and Bolton 2014).As a result, a series of compensation restrictions were imposed in October 2008 and revised later in February 2009: limiting the tax deductibility of compensation for senior executives, requiring bonus claw-backs, and restricting golden parachute payments (Berger et al. 2020).However, some experts claim that the regulations made are simply a product of anger by politicians, the press, media, trade unions, and the general public, but not shareholders.Therefore, the best way the government can fix executive compensation is to stop trying to fix it and undo the damage already caused through existing regulations that have caused enormous costs to organizations, shareholders, and social welfare (Murphy and Jensen 2018).
Following the contracting theory (Ke et al. 1999) and managerial power hypothesis (Jiraporn et al. 2005), executive compensation may alleviate the agency problem or may even be part of the agency problem per se.In this scenario, the level and structure of executive compensation are misused by the executives to indulge their desires for self-maximization of wealth irrespective of firm performance.When the board of directors is strong, the executive compensation structure is more likely to reflect performance-based systems.However, when the CEO has stronger bargaining power, the compensation structure will lean toward cash-based pay, which is preferred by the CEO who does not like the pay-at-risk system (Elsaid et al. 2009).Despite the arguments from both sides (proponents and opponents of the high level of executive compensation) and the mixed empirical findings (e.g., Fernandes 2008;Sun et al. 2009;Chhaochharia and Grinstein 2009;Berger et al. 2020), it is important to justify compensation in relation to value creation.If the compensation level that is very high and growing at a skyrocketing level is a reward to management for enhancing shareholder wealth, then the executive pay is fair and justifiable.
Since previous research has examined in various settings and in many ways how firm performance and firm value affect executive compensation (see Benson and Davidson 2010;Bizjak et al. 2008;Cuñat and Guadalupe 2009;Frydman and Saks 2010), this study intends to confirm the relation within the subset of public firms that received TARP funds.TARP is a U.S. government program intended to restore the U.S. economy by purchasing assets and equities from troubled institutions.As of 31 December 2011, 926 firms had received TARP funds from the government, and 297 were publicly listed firms (Department of Treasury 2011;ProPublica 2011).Of the 297 public companies, 128 firms received at least $50 million.Total TARP funds disbursed were $411.54 billion, excluding bailout money for Fannie Mae and Freddie Mac, whereas $279.42 billion was repaid by the recipients (Department of Treasury 2011;ProPublica 2011).Executive compensation rules and guidance were made under Section 111 when President Bush signed the Emergency Economic Stabilization Act (EESA) to launch TARP on 3 October 2008.
However, the EESA is abstract and less straightforward as a guide to executive compensation limits.For instance, the definition and signs of compensation that encourage excessive risk taking might raise various arguments.The public was appalled that Wall Street firms in general and TARP recipients were still paying their executives huge 2008 salaries and bonuses.For example, Merrill Lynch, which was facing bankruptcy and going to be sold to the Bank of America, awarded more than $3 billion in bonuses to its executives.Meanwhile, in early 2009, Citigroup still had plans to procure executive aircraft, which would cost around $50 million.Wall Street executive bonuses totaled $19 billion in 2008 (Stolberg and Labaton 2009).
Since the EESA had not been able to curtail the executive compensation of troubled firms, the House of Representatives passed another bill, the American Recovery and Reinvestment Act (AARA), on 28 January 2009.On 10 February 2009, the Senate passed AARA.On 13 February 2009, the amended version of ARRA was promulgated in both the House (246 vs. 183) and the Senate (60 vs. 38 votes).President Obama then signed it into law on 17 February 2009.Title VII regulates the executive compensation limits under this law.
On 4 February 2009, because EESA and AARA had not been straightforward about the amount of executive compensation limits, the Obama Administration issued Financial Guidelines on executive compensation limitations for companies that received extraordinary assistance through TARP.Accordingly, the Guidelines were issued approximately two weeks prior to the AARA attestation.While these Guidelines did not apply retroactively to compensation plans previously announced before February 2009, they certainly put pressure on firms participating in TARP to rein in compensation packages amid media scrutiny and public outrage.The Pay Czar issued his first rulings on 22 October 2009, to follow up on the Interim Final Rule.The first rulings regulated compensation for the 25 highest-compensated executives in firms that had received exceptional assistance under TARP, including AIG, Bank of America, Chrysler, GM, GMAC, and Chrysler Financial (Solomon and Fitzpatrick 2009).On 11 December 2009, the Pay Czar announced his second ruling to regulate the 26-100 most highly paid executives of firms receiving exceptional assistance under the TARP (particularly AIG, Citigroup, GM, and GMAC).
Phillips (2010) investigated 112 TARP recipients and found that only some performance measures were positively related to compensation in 2007 but not in 2006.She also reported that some performance measures were negatively related to CEO compensation, indicating that earnings-related fundamentals were less able to explain the executive compensation of TARP recipients.Crawford et al. (1995) tested the pay-for-performance sensitivity of bank CEOs after a series of deregulations during the 1980s and found support for the deregulation hypothesis, i.e., the sensitivity of CEO compensation to the increase in shareholder wealth was significantly higher in the post-deregulation period.Koo (2015) examined the managerial power hypothesis during the financial crisis and found that powerful CEOs took more managerial compensation due to agency costs.However, he pointed out that these powerful CEOs were more likely to have access to TARP funding during the crisis and to increase firm values for TARP firms compared to non-TARP firms.
Similarly, Hubbard and Palia (1995) found that compensation increased sharply in the deregulated banking industry during the 1980s.However, Hubbard and Palia's (1995) results on the pay-for-performance sensitivity are weaker as they only found a significant finding in the univariate analysis but not in the multivariate setting.In examining executive compensation during the period of strengthened regulation from 2000 to 2005, Chhaochharia and Grinstein (2009) found that firms that did not comply with the Sarbanes-Oxley Act (SOX), and the new exchange rules tended to decrease CEO compensation in the post-regulation period, especially firms with low institutional and shareholder ownership.In studying the effect of compensation restrictions introduced by the TARP, Nwaeze et al. (2018) documented a significant performance improvement among banks that experienced CEO resignations after they accepted TARP funds, and this improvement became more significant in the year following the CEOs' exit.In contrast, TARP banks that retained their CEOs showed a significant increase in CEO pensions, a form of post-TARP executive compensation.Finally, Fan et al. (2020) found a significantly positive impact of TARP on the earnings management of recipient banks, compared with their non-recipient peers.Their finding supports the benefits of tighter regulations following corporate scandals.However, these studies focused on the relationship between executive compensation and post-deregulation firm performance.
Meanwhile, the stock market does not always react positively to binding compensation with performance.For instance, the markets react negatively to firms that apply a higher percentage of performance-based compensation to their new CEOs (Elsaid et al. 2009).Accordingly, we examine the market reactions to a series of news and events regarding the Administration's perception, policies, and regulations in 2009 on curbing executive compensation for the executives of TARP recipients.We compile news from the Wall Street Journal, New York Times, and Treasury Department websites.The selected events are presented in detail in Figure 1.
We expect positive market reactions to events 1, 2, 3, 4, 5, 8, 9, and 10 since these events signify the government's intention to enforce the pay-for-performance sensitivity with a long-term perspective.On the contrary, we expect a negative market reaction to the seventh event, which illustrates the TARP recipient firms' desire to exit the program and be free from compensation provisions.Meanwhile, the market reaction to event 6 is an empirical question as there were two opposing pieces of news: (1) the ratification of a new bill on compensation restrictions and (2) four TARP banks paid back the bailout money, and thus they were not bound by the compensation rules.We expect positive market reactions to events 1, 2, 3, 4, 5, 8, 9, and 10 since these events signify the government's intention to enforce the pay-for-performance sensitivity with a long-term perspective.On the contrary, we expect a negative market reaction to the seventh event, which illustrates the TARP recipient firms' desire to exit the program and be free from compensation provisions.Meanwhile, the market reaction to event 6 is an empirical question as there were two opposing pieces of news: (1) the ratification of a new bill on compensation restrictions and (2) four TARP banks paid back the bailout money, and thus they were not bound by the compensation rules.

H2:
The market reacted negatively to event 7.
Furthermore, we formally investigate the effect of changes in executive compensation on shareholder wealth during the 2009-2010 regulatory period, which shows the efficacy of new regulations on compensation and governance based on firm value.Shareholders, the market, and the government expected significant reductions in TARP's corporate executive compensation due to the enforcement of restrictions and other governance provisions under strict regulations, which were expected to benefit the company by increasing its value during the regulatory period.Thus, if the regulation was effective, compensation changes must have been negatively related to firm value during the regulatory period.

H3: Change in shareholder wealth was negatively related to changes in executive compensation
during the regulatory period for TARP firms.

Data and Sample
The sample in this study includes a subset of public firms that received TARP funds from the U.S. government.Following Phillips (2010), we began by using public firms that received $50 million or more in bailouts for two main reasons.First is the unavailability of data on companies receiving TARPs below $50 million since lots of them were over-the-

H2:
The market reacted negatively to event 7.
Furthermore, we formally investigate the effect of changes in executive compensation on shareholder wealth during the 2009-2010 regulatory period, which shows the efficacy of new regulations on compensation and governance based on firm value.Shareholders, the market, and the government expected significant reductions in TARP's corporate executive compensation due to the enforcement of restrictions and other governance provisions under strict regulations, which were expected to benefit the company by increasing its value during the regulatory period.Thus, if the regulation was effective, compensation changes must have been negatively related to firm value during the regulatory period.
H3: Change in shareholder wealth was negatively related to changes in executive compensation during the regulatory period for TARP firms.

Data and Sample
The sample in this study includes a subset of public firms that received TARP funds from the U.S. government.Following Phillips (2010), we began by using public firms that received $50 million or more in bailouts for two main reasons.First is the unavailability of data on companies receiving TARPs below $50 million since lots of them were overthe-counter (OTC) and pink-sheet companies.Second, the government was more lenient to TARP firms receiving below $50 million with respect to their payments of non-cash compensation.
In this study, we focused on 297 TARP recipients, with 128 firms receiving a minimum of $50 million in funds.Out of these 128 firms, 84 were identified as a clean sample, meticulously matched among Compustat, ExecuComp, Center for Research in Security Prices (CRSP), and proxy statements.This rigorous matching process ensured complete data integrity, resulting in a dataset comprising 168 firm-year unbalanced observations.
During the regulatory period from 2009 to 2010, we observe a significant shift in governance dynamics.This period was characterized by the implementation of strict governance rules and restrictions on executive compensation, which signaled an evolution toward performance-based compensation systems as part of the governance mechanism.For this period, we gathered data for our clean sample, which includes 84 companies drawn from diverse stock exchanges-37 from the NYSE, 43 from NASDAQ, and four from the OTC markets.These companies were categorized into 11 distinct industries based on their four-digit SIC codes, encompassing various sectors, such as commercial banks, insurance companies, and financial services firms.In addition, our analysis reveals that the clean sample firms exhibit significantly larger sizes, which is substantiated by the independent sample t-test and Wilcoxon rank-sum test.
The bailed-out firms' data were collected from the ProPublica dataset, the Department of Treasury's website, and USA Today.We obtained financial statements and SIC data from the Compustat database, whereas stock return, market return, stock price, and market capitalization data were collected from the CRSP database.We obtained executive compensation, director compensation, executive age, managerial ownership, and options data from the ExecuComp database.Since board and governance variables (e.g., board size, board independence, CEO/Chairman duality, and entrenchment index) are not available on Compustat and ExecuComp, we hand-collected them from the sample firms' statements, 10-Ks, bylaws, articles of incorporation, and other related documents provided by the SEC's EDGAR.
In the data integration process, this study initially merged TARP recipient data (receiving at least $50 million) with Compustat data based on ticker symbols, ensuring accuracy through manual verification of firm names and addresses.Subsequently, this merged dataset was combined with ExecuComp data using global vantage keys (GVKEY) and then merged with CRSP data based on CUSIP numbers.Additional governance variables from proxy statements and data from sources like the SEC's EDGAR website were incorporated.Supplementary data included inflation data from the Federal Reserve Bank of St. Louis' Economic Research Data and Fama-French's three factors and risk-free return data from Kenneth French's data library.

Variables
Our main variables are described as follows: 1.

Total Executive Compensation
We employ total executive compensation (TDC1) from ExecuComp, which consists of salary, bonus, other annual, total value of restricted stock granted, total value of stock options granted (using Black-Scholes), and long-term incentive payouts.According to Bebchuk and Grinstein (2005), the aggregate of the top five highest-paid executives' compensation growth is more able to highlight the nexus between compensation and performance.Accordingly, we define total compensation as the average of the five highest-paid executive compensations of each company in each year, as follows: where: i = firm i, t = year t, and j = executive of firm i

Cash-Based Pay
Cash-based compensation is measured as salary, bonus, and other compensation (TOTAL_CURR + OTHCOMP) from ExecuComp.Equation ( 1) is also applied to cash-based pay in order to get the average per company per year.

Equity-Based Compensation
Equity-based pay is defined as the total value of stock awards (STOCK_AWARDS) and the total value of option awards (OPTION_AWARDS).Equity-based compensation also applies Equation (1) to get the average compensation per company per year.

Firm Market Performance
Firm market performance is measured as Tobin's Q, defined as: In our analysis, if the book value of equity (TEQ) from Compustat falls below −$35 billion, it is treated as a missing value.To proxy firm market performance, we use stock return (the geometric average of monthly returns for firm i in calendar year t), calculated as follows: Return nsi,t = e Log Return nsi,t − 1 nsi = non-sample firm i in a certain industry, t = year t, m = month m Accordingly, the cross-sectional arithmetic mean of the geometric mean of monthly returns for each industry within that specific year is calculated as follows: Return nsi,t y = industry y, nsi = non-sample firm i in industry y, t = year t Outliers and missing values are handled by excluding return data from CRSP that fall below −1 or exceed 2. It is important to note that sample firms are excluded from industry-specific calculations.

Shareholder Wealth
We proxy shareholder wealth with market capitalization, as follows: Market capitalization = Stock price * Number of shares outstanding (3) We employ several control variables, including: a.

Firm Size
Most previous studies concur a positive relationship between compensation and firm size (e.g., Duffhues and Kabir 2008;Gao 2010;Jiraporn et al. 2005;Ke et al. 1999), although several studies have also reported a negative relationship (e.g., Benson and Davidson 2010).Boards tend to be larger and slower to make decisions as a company grows, allowing the CEO to exercise power over compensation design.We proxy firm size with total assets.

b. Firm Growth Opportunities
Firms with high growth opportunities are more likely to award higher pay to their executives (e.g., Baber et al. 1996;Bizjak et al. 2008;Chourou et al. 2008;Core et al. 1999;Cuñat and Guadalupe 2009;Gao 2010).Following Baber et al. (1996), we measure firm growth opportunities as the ratio of capital expenditures to total assets, as follows: Two hypotheses regarding the use of debt and its effect on executives are (1) the equity agency cost hypothesis and (2) the debt agency cost hypothesis.Both suggest a negative relationship between leverage and compensation (Ortiz-Molina 2007).Leverage is measured as the debt ratio to total assets.

d. Managerial Ownership
Managerial ownership helps align the shareholders' and management interests, in that higher ownership results in a higher executive's sense of belonging.Thus, increasing managerial ownership can reduce the compensation's effectiveness as a mechanism to alleviate agency problems (Chourou et al. 2008;Core et al. 1999).We utilize the percentage of shares owned by the executive, including exercisable options within 60 days (SHROWN_TOT_PCT) from the ExecuComp, to measure managerial ownership.

e.
Accounting Performance Accounting performance is proxied by return on assets (ROA).Prior findings suggest a positive link between ROA and compensation (Benson and Davidson 2010;Bizjak et al. 2008).These studies also suggest adjusting ROA to the industry average.Therefore, we calculate each industry's ROA for a particular year simply as the arithmetic mean of ROAs of firms belonging to that industry in that year.Similar to the estimation of industry returns, the calculations for sample firms are conducted separately from their respective industries.f.
Total Risk Previous studies have indicated that firm volatility might affect compensation positively and negatively (Benson and Davidson 2010).From a positive prediction, executives will find equity-based payments less attractive when volatility is higher (Benson and Davidson 2010;Gao 2010).Meanwhile, the negative perspective suggests that more volatile firms give higher discretionary power to management (Core et al. 1999;Ortiz-Molina 2007).Hence, the relationship between compensation and risk is an empirical question.We measure risk as the standard deviation of monthly returns.g.
Board Size Core et al. (1999) reported a negative relation between compensation and board size, where board size is industry-adjusted.We collect all directorial data from ExecuComp for the 11 industries of sample firms to estimate each industry's board size each year, as measured by the arithmetic mean of company board sizes in those industries.Again, we exclude sample firms from the calculation of industry board size.

h. Board Independence
We measure board independence as the percentage of independent directors on the board.More independent directors on a board are expected to improve performance-based compensation systems by emphasizing equity-based payments as well as curbing discretionary managerial power to embezzle corporate wealth and enjoy excessive compensation (Core et al. 1999;Mehran 1995).

i. Executive Entrenchment
Entrenchment is proxied by CEO/Chairman duality.A previous study confirmed the positive link between compensation and CEO/Chairman duality (Core et al. 1999).(Bebchuk et al. 2009) developed an entrenchment index as an alternative measurement.They showed that only six provisions are significantly associated with lower firm value and negative abnormal returns from 24 Gompers, Ishii, and Metrick (GIM) governance provisions.The six provisions are (1) a staggered board system, (2) the presence of poison pills, (3) the presence of golden parachutes, (4) limits to bylaws amendments, (5) limits to charter amendments, and (6) supermajority requirement to approve a merger.Each provision has one point; thus, the total entrenchment index points range from zero to six each year.The higher the total points, the worse the corporate governance.A provision will be treated as a missing value if the required documents are not available.

j. Executive Age
Executives might focus more on short-term performance and make as much money as possible when approaching retirement age (Davidson et al. 2007).Meanwhile, another study documented that the compensatory incentive component is lower for CEOs approaching retirement (Harvey and Shrieves 2001).Therefore, executive age may have different effects on cash-based and equity-based compensations.

k. Director Compensation
Directors may be less motivated to be good monitors since their salaries are lower than executive compensation or their own compensation packages at home firms.Therefore, director fees are expected to be negatively related to executive compensation.We measure director fees as the total director compensation reported in SEC filings (TOTAL_SEC) from ExecuComp.

Testing Procedures
We conducted an event study to investigate the market reaction toward news and events during the restriction period of executive compensation and stricter governance rules, especially for TARP recipients.We followed the procedures shown in Campbell et al. ( 1998) and Boehmer et al. (2002) to conduct the event study.This study uses an estimation window of −180 to −31 days before each announcement date.To ensure that the regression has at least 30 observations if a particular event does not have historical returns reaching −180 days, then the closest historical returns to day −180 are used as long as the observations have at least −70 days.We utilized the market model to estimate expected returns across multiple event windows, i.e., Day 0, Day (−1, +1), and Day (−2, +2).
To further examine the stock return effect of the 10 announcements, we ran OLS regressions of stock return on market return and the 10 announcement dummies as follows: where R i,t is the return on security i on day t; R m,t is the market return on day t; R m,t+1 is the market return on day t + 1; D is a dummy variable for days (0, +1) surrounding an event; β and g are regression coefficients; e is error.The regression equation above was applied to both sample firms and all other firms in the 11 industries.For each event's dummy variable, we used the event window (0, +1) surrounding the event as it produces the most meaningful results from the standard event study methodology in the previous section.We also utilized the event window (0, 0), but the results were less significant than those from the event window (0, +1).The analysis period spanned from 25 June 2008 (150 days before the first announcement) to 20 July 2010 (150 days after the tenth event).
For comparison, we also employed a multivariate regression model (MVRM) based on a system of equations with seemingly unrelated regressions (SURs) following Cornett et al. (1996) and Cornett et al. (1998).We formed two simple portfolios, the first comprising sample firms and the second non-sample firms in the 11 industries.To get an average return for each date, we simply calculated the cross-sectional mean of daily stock returns within each portfolio.Similar to the OLS regressions above, the time span was between 25 June 2008 and 20 July 2010.We ran the MVRM test utilizing the equation as follows: where: R S,t is the mean return on sample firms on day t; R NS,t is the mean return on all other firms in the 11 industries on day t; D is a dummy variable for days (0, +1) surrounding an event; B and g are regression coefficients; e = error.Furthermore, to highlight the direct benefits of the new regulations, we investigated how the change in executive compensation affected shareholder wealth during the regulatory period.If the new compensation and governance regulations were effective, we expect the changes in TARP firms' compensation to be negatively associated with firm value gain or loss during the regulatory period.To test the hypothesis, we employed the following model: where ∆SW is firm value gain or loss; ∆C changes in executive compensation; LSIZE is log total assets or log total sales; GR is firm growth opportunities, defined as capital expenditures to total assets; LEV is total debt to total assets; MOW is the percentage of shares owned by the executive; ADJROA is industry-adjusted ROA; RISK is the volatility of monthly returns; ADJBS is the industry-adjusted number of directors on the board; BI is the percentage of independent directors on the board; DUAL is a dummy variable taking the value of 1 if the CEO is also the Chairman and 0 otherwise; EI is entrenchment index; AGE is executive age; LDIR is log total director compensation; i = firm i; t = years 2009 and 2010; α = subject-specific intercept; δ = year fixed effect; β = regression coefficient; and e = error.
The change in compensation and the change in shareholder wealth might have an endogenous relationship.While changes in executive compensation might affect firm value, firm performance in the previous period may also influence the compensation level.Thus, we employed the following equation to address possible endogeneity: where ∆C is change in executive compensation; Q is Tobin's Q; ∑ N−1 i=1 α i DF i,t is to create equivalent firm fixed effects in the simultaneous equations; ∑ T−1 t=1 δ t DT i,t is to create equivalent year fixed effects in the simultaneous equations.We utilized two-stage (2SLS) and three-stage least squares (3SLS) methods to estimate coefficients, with the previous year's Tobin's Q being the instrument for the change in executive compensation.
Lastly, many TARP companies repaid the bailout funds, possibly to avoid compensation restrictions and further government involvement in governance under TARP, such as appointing an executive.Therefore, we are interested in examining whether the effect of compensation changes on shareholder wealth is different for TARP firms that repaid bailouts before 11 December 2009, the day Pay Czar announced his second ruling.The following model was applied to test the relationship: where DPAY is a dummy variable for TARP firms that repaid the bailout money before 11 December 2009.

Classical Assumption Tests
We checked several classical assumptions in regression analysis.Potential heteroskedasticity was remedied using Arrelano's cluster corrected robust estimators for panel data, which take into account White's heteroskedastic-robust estimators for each firm, and then computed the average of the n estimates of covariance matrixes (Ajmani 2011).
Although data are asymptotically approaching normality for a large sample size (Wooldridge 2005), we improved data normality and results interpretation by transforming several variables into log values.Those variables are executive compensation, total assets, total sales, and director compensation.Then, we conducted a regression of log total compensation on the lagged value of Tobin's Q, industry-adjusted return, log total assets, log total sales, growth opportunities, leverage, managerial ownership, industry-adjusted ROA, risk, industry-adjusted board size, board independence, CEO/Chairman duality dummy, entrenchment index, executive age, and log director compensation.A plot and histogram are reported in Figure 2A,B.The plot and the histogram depict improved data normality.Skewness decreases to 0.13 while kurtosis is 1.14.The Kolmogorov-Smirnov D-statistic drops to 0.04, barely significant at the 10% level.
data, which take into account White's heteroskedastic-robust estimators for each firm, and then computed the average of the n estimates of covariance matrixes (Ajmani 2011).
Although data are asymptotically approaching normality for a large sample size (Wooldridge 2005), we improved data normality and results interpretation by transforming several variables into log values.Those variables are executive compensation, total assets, total sales, and director compensation.Then, we conducted a regression of log total compensation on the lagged value of Tobin's Q, industry-adjusted return, log total assets, log total sales, growth opportunities, leverage, managerial ownership, industry-adjusted ROA, risk, industry-adjusted board size, board independence, CEO/Chairman duality dummy, entrenchment index, executive age, and log director compensation.A plot and histogram are reported in Figure 2A,B.The plot and the histogram depict improved data normality.Skewness decreases to 0.13 while kurtosis is 1.14.The Kolmogorov-Smirnov D-statistic drops to 0.04, barely significant at the 10% level.To check whether a multicollinearity problem exists, we regressed log total compensation on the lagged value of Tobin's Q, industry-adjusted return, log total assets, log total sales, growth opportunities, leverage, managerial ownership, industry-adjusted ROA, risk, industry-adjusted board size, board independence, CEO/Chairman duality dummy, entrenchment index, executive age, and log director compensation, to get variance inflation ratios (VIFs) for the explanatory variables.Our unreported results show that all explanatory variables have VIFs below the predetermined level (5.56).Hence, the potential multicollinearity problem is not material, and should not be a concern.

The Stock Return Effect of the 10 Announcements
Mean abnormal returns for five days surrounding the 10 events are presented in Table 1.For comparison, we analyzed the returns on all other firms in 11 industries, excluding sample firms.Panel A of Table 1 documents the mean CARs for four event windows, i.e., days (−1, +1), (0, 0), (0, +1), and (−2, +2).The findings show that event windows (0, 0) and (0, 1) generally provide more significant results than others.We also used a modified method in the event study.Instead of estimating expected returns from the market model, we employed the mean returns on all other firms in the 11 industries as the proxy for expected returns.The results in Panel B of Table 1 show that the calculated results are consistent with the standard event study methodology.
During the first three events, sample firms' mean CARs were significantly negative.This evidence indicates that the public still felt doubtful and unsure regarding the effectiveness of the government's stringent policies and legislation in limiting executive To check whether a multicollinearity problem exists, we regressed log total compensation on the lagged value of Tobin's Q, industry-adjusted return, log total assets, log total sales, growth opportunities, leverage, managerial ownership, industry-adjusted ROA, risk, industry-adjusted board size, board independence, CEO/Chairman duality dummy, entrenchment index, executive age, and log director compensation, to get variance inflation ratios (VIFs) for the explanatory variables.Our unreported results show that all explanatory variables have VIFs below the predetermined level (5.56).Hence, the potential multicollinearity problem is not material, and should not be a concern.

The Stock Return Effect of the 10 Announcements
Mean abnormal returns for five days surrounding the 10 events are presented in Table 1.For comparison, we analyzed the returns on all other firms in 11 industries, excluding sample firms.Panel A of Table 1 documents the mean CARs for four event windows, i.e., days (−1, +1), (0, 0), (0, +1), and (−2, +2).The findings show that event windows (0, 0) and (0, 1) generally provide more significant results than others.We also used a modified method in the event study.Instead of estimating expected returns from the market model, we employed the mean returns on all other firms in the 11 industries as the proxy for expected returns.The results in Panel B of Table 1 show that the calculated results are consistent with the standard event study methodology.During the first three events, sample firms' mean CARs were significantly negative.This evidence indicates that the public still felt doubtful and unsure regarding the effectiveness of the government's stringent policies and legislation in limiting executive compensation.Historically, the U.S. has been a market-driven country where the public has seldom (if ever) seen government involvement in honing corporate decision-making, such as executive compensation design.Thus, although investors were aware that the TARP firms were on the verge of bankruptcy due to poor governance, they might be perplexed about the efficacy and long-term impacts of executive compensation restrictions on the company's prospects.
For the first three events, all other firms in 11 industries also recorded a negative and significant mean CAR.However, the magnitude of their mean CARs was smaller than that of sample firms.Such conditions indicate that this announcement created higher market uncertainty as to whether the new government policy would apply to non-TARP firms in the industry.Panel B supports these indications.The sample firms performed significantly worse than the non-sample firms, indicating that the market reacted negatively toward EESA and AARA since they were deemed not straightforward policies.
On the contrary, the market reacted positively to the fourth event, when the President ordered the Secretary of Treasury to block the $165 million bonus that AIG would pay to its executives.The mean CAR of TARP firms was significant for the event window (0, +1), indicating that the public was more optimistic about the policy and agreed that AIG executives should not take bonuses as the firm was still struggling under TARP.
Similarly, the market reacted positively to events five and six, indicating that investors welcomed further announcements of more stringent and comprehensive governance and compensation regulations.Non-sample firms also saw positive market reactions to these events.However, their mean CARs were much smaller and statistically less significant.Results in Panel B of Table 1 show a more favorable market reaction toward the sample firms in events four, five, and six, supporting that indication.
When 10 banks were allowed by the government to pay back the TARP funds on 9 June 2009, the market responded negatively, thereby confirming hypothesis two.All other firms in the 11 industries also suffered negative mean CARs on the days surrounding this announcement, but their mean CARs were less negative than those of the sample firms.Panel B confirms that sample firms underperformed non-sample firms during this event.Such conditions again show that investors tended to agree that companies should be regulated strictly and comprehensively.
Subsequently, events VIII (the Department of Treasury's Interim Final Rule), IX (Pay Czar's First Rulings), and X (Pay Czar's Second Rulings) were reacted to positively by investors but with much weaker intensity.The mean CAR for event eight of TARP firms was only significant for the event window (0, +1), whereas event 10 only saw a significantly positive mean CAR for the event window (0, 0).In addition, the mean CARs of TARP recipients during events eight, nine, and 10 were all below 1.4%.Results in Panel B show that sample firms underperformed non-sample firms in event eight but enjoyed significantly positive abnormal returns during events nine and 10.
Overall, our evidence suggests that the market was reluctant and cautious at the beginning of the process of regulating executive compensation of TARP firms (and possibly the whole industries).As time passed, investors responded positively to more stringent and comprehensive governance regulations, which appeared to improve governance without affecting or harming the quality of managerial talents.Even the public negatively reacted to the 10 relatively large banks when they decided to leave TARP, possibly because investors suspected that those banks quit to release themselves from stringent governance and compensation rules under the TARP.Later announcements involving the Department of Treasury and the Pay Czar were still reacted to favorably by the market, but their magnitudes were much smaller since the public must have anticipated the coming of detailed rulings pertaining to the implementation of executive compensation restrictions.
To further examine the stock return effect of the 10 announcements, we ran OLS regressions of stock return on market return and the 10 announcement dummies as denoted in Equation ( 5).For comparison, we also employed a multivariate regression model (MVRM) based on a system of equations with seemingly unrelated regressions (SURs), as denoted in Equations ( 6) and ( 7).Table 2 exhibits the OLS and SURs test results.Consistent with the results from standard event study methodology, the market, in the beginning, reacted negatively to the first three events.According to Kim (2012), the market reacted negatively to compensation regulation since investors were concerned about losing talented executives due to the upper compensation limit.However, we find positive reactions in the fourth and fifth announcements, indicating that investors possibly saw the potential benefits of new stringent rules to improve governance and pay-for-performance sensitivity.Contrary to investors' concerns that the TARP program might drain top talents, a previous study reported that the TARP program was more likely to induce firm performance and was thereby less likely to harm managerial talents (Nwaeze et al. 2018).Our study provides evidence that the market reacted positively to the more stringent and comprehensive governance regulations, suggesting that these regulations were deemed able to improve governance without harming the quality of managerial talents.
Subsequently, the market responded negatively again when 10 banks were approved to quit the TARP (event 7), allegedly circumventing stricter compensation restrictions and governance provisions.However, the regression results are not significant for events VIII, IX, and X, most likely because the public anticipated the announcements of detailed rulings to implement new executive compensation laws.As shown in Models 3 and 4 of Table 2, the SUR regression results have similar parameter estimates with the OLS', but the coefficients on the event period dummies are not significant under this model.
Overall, our evidence on 10 events suggests that the managerial talent hypothesis view of market reaction, which hypothesizes that investors react negatively to more stringent executive compensation regulations as they can drain top talent, seems too excessive.On the first three events, our results suggest that the market is still unsure regarding the effectiveness of the government's stringent policies and legislation in limiting executive compensation.However, the market reacted positively to events IV, V, VI, VIII, IX, and X, indicating that the investors concurred that executives should not take bonuses as the firms struggled under the TARP, and were more optimistic about the more stringent and comprehensive executive regulations.

Do Investors Agree on Stricter and More Comprehensive Compensation Regulations?
In the previous section, we have shown evidence that the negative reactions toward compensation regulations were more likely due to investors being unsure about the effectiveness of these regulations.In contrast, positive reactions to the other six events indicate that investors agreed on more comprehensive and stringent compensation restrictions.In this section, we examine further whether the investors agreed with the implementation of stricter and more comprehensive regulations by analyzing the relationship between the change in compensation and the change in shareholder wealth.If investors believed that the new regulation could improve firm performance without harming the availability of the talent pool, we expect a negative relationship between the change in compensation and the change in shareholder wealth during the regulatory period.
We employed Equations ( 8)-( 10) to test this conjecture.We utilized two-stage (2SLS) and three-stage least squares (3SLS) methods to estimate coefficients, with the previous year's Tobin's Q being the instrument for the change in executive compensation.Tables 3-5 show that the change in shareholder wealth is negatively and significantly related to the change in executive compensation for both components of compensation, i.e., equitybased and cash-based pay.These findings indicate that an increase in TARP company executive compensation during the regulatory period would decrease shareholder value, thus corroborating our conjecture.For instance, in Model 3 of Table 3, TARP firms that lowered total compensation by $1000 during the regulatory period were likely to enjoy firm value gains of $1.21 million, suggesting that the market reacted positively to companies that complied with stricter and more comprehensive compensation regulations.
As reported by Winkelvoss et al. (2014), compensation was reduced at most TARP firms following the government bailout of the banking industry.Our findings are in the same spirit, showing that investors agreed with the government that executive compensation needed to be regulated more strictly and comprehensively.Furthermore, a recent study suggested that a fair compensation composition structure must be designed based on financial performance to reduce managerial opportunism since executives may not act in the interests of shareholders (Pathak and Chandani 2023).In this regard, Pathak and Chandani (2023) pointed out that a fair composition structure can align the interests of managers and shareholders.Accordingly, our findings also show that the new compensation and governance regulations were beneficial to TARP firms' shareholders by helping them align their interests with managers'.
With respect to control variables, our results document that TARP firms with lower leverage were more likely to experience an increase in firm value, indicating that less risky firms were more likely to increase shareholder wealth.According to Oanh et al. (2023), a higher leverage ratio will reduce ROA and ROE, whereas bank funding liquidity positively affects the performance of commercial banks.In addition, a prior study finds that a higher capital ratio exerts a positive impact on bank resilience and bank performance (Velliscig et al. 2023).Larger board size and higher board independence are positively related to the change in firm value, as larger boards may provide a wide range of skills, especially financial expertise needed to restore firm performance after the financial crisis.This finding supports Pathak and Chandani (2023), who suggested that board composition and how well it supervises its executives determine the successful implementation of corporate governance.Interestingly, the entrenchment index is positively related to the change in shareholder wealth.Probably this is because TARP firms, which had higher entrenchment index scores (poor governance), were forced by the new rules to dismiss fraudulent executives, and to improve governance mechanisms (including compensation) so as to increase firm value over the regulatory period.Subsequently, firms with fewer aged executives are more likely to gain value, supporting Harvey and Shrieves (2001) and Davidson et al. (2007).Older executives might be less interested in the firm's long-term prospects.Hence, they could press toward the goal of earning as much cash compensation as possible before quitting or retiring (Davidson et al. 2007).
Moreover, CEO compensation for TARP firms that paid over $500,000 hurt firm performance, both in the short and long terms.This is similar to the finding by Vega et al. (2020).We also find that the market reacted negatively to TARP companies that decided and were approved to repay the bailout funds.Investors viewed those firms as wanting to be free from TARP restrictions.Subsequently, we are interested in testing whether the effect of compensation change on shareholder wealth was different for TARP firms that repaid bailout funds before 11 December 2009, i.e., the day Pay Czar announced his second ruling.We employed Equation (11) to examine this notion.
Results in Table 5 report that the change in total compensation is negatively related to firm value for TARP firms that repaid the bailout money before the Pay Czar's second rulings, thereby substantiating our hypothesis.However, this relationship is less negative than that for TARP firms that had not repaid the bailout funds.Referring to Model 2, for example, −3.207 (coefficient on TC) + 2.306 (coefficient on DPAY*DTC) = −0.901,indicating that a TARP firm that repaid the bailout funds and increased its total executive compensation by $1000 during the regulatory period was likely to suffer a decline in value by $901,000.Meanwhile, shareholders of non-paid-back TARP firms doing the same increase might lose $3.207 million.Since they were freed from the TARP, investors may have perceived paid-back TARP firms as safer and more capable of recovering quickly from financial distress.Therefore, their decision to decrease (increase) executive compensation during the regulatory period was responded to less positively (negatively) by the market.These findings also apply to equity and cash compensations but are insignificant for equitybased pay.

Conclusions
This paper examines the relationship between executive compensation and firm performance amid more stringent regulations by analyzing 10 events with respect to compensation for the executives of TARP recipients.Employing standard event study procedures, our evidence on 10 events suggests that the managerial talent hypothesis, which conjectures that investors react negatively to more stringent executive compensation regulations as they can drain top talents, seems too excessive.
On the first three events, our results report that the market was still unsure regarding the effectiveness of the government's stringent policies and legislations in limiting executive compensation.Meanwhile, the market reacted positively to events IV, V, VI, VIII, IX, and X, indicating that the investors concurred that executives should not take bonuses as the firms struggled under the TARP, and were more optimistic about the more stringent and comprehensive executive regulations.Our further testing results using the ordinary least squares (OLS) and multivariate regression model (MVRM) based on a system of equations with seemingly unrelated regressions (SURs) substantiate these findings.
We analyze further whether investors were in favor of the implementation of stricter and overarching compensation regulations by investigating the relationship between the change in compensation and the change in shareholder wealth during the regulatory period.Utilizing pooled OLS, fixed effects, two-stage (2SLS), and three-stage least squares (3SLS) methods, our findings show that the change in executive compensation is negatively related to shareholder wealth.Both types of compensation, i.e., equity-based and cashbased compensation, show significant results, suggesting that investors believed that the new regulations could improve firm performance without harming the availability of the talent pool.These findings also indicate that the new compensation and governance regulations were relatively beneficial to TARP firms' shareholders.
We also test the notion that leaving the TARP program during the regulatory period was perceived by investors as an intention to be free from restrictions under TARP.With pooled OLS, fixed effects, two-stage least squares, and three-stage least squares, we find that the change in total compensation was negatively related to firm value for TARP firms that repaid the bailout money, corroborating our hypothesis.However, this relationship was actually less strong than that for TARP firms that had not repaid the bailout funds, indicating that the market was less responsive to firms that had repaid bailout money since investors might perceive these firms as safer and more capable of recovering quickly from financial ailments.
This table reports the 10 events' mean CARs for four event windows (−1, +1; 0, 0; 0, +1; and −2, +2).Non-sample firms are all other firms in the 11 industries.t-stats are reported in first-line parentheses under CARs.Patell t-stats are reported in second-line parentheses under CARs.***, **, and * denote significances at 1%, 5%, and 10%, respectively, based on t-tests.† † †, † †, and † denote significances at 1%, 5%, and 10%, respectively, based on Patell t-tests.Panel A reports an event study with standard methodology, i.e., using the market model to estimate expected returns.Panel B shows the results of an event study where the average returns on all other firms in the 11 industries are employed as the expected returns.

Table 2 .
Event Study Using Ordinary Least Squares (OLS) Model and Multivariate Regression Model (MVRM).
This table presents the results from the OLS model and MVRM.The dependent variable is return on security i on day t (RETURNi,t) for OLS and mean return on sample firms or non-sample firms on day t (RETURNs,t or RETURNns,t) for MVRM.Independent variables are market return on day t (Rmt), market return on day t + 1 (Rmt + 1), and dummy variables representing each event period (Ds).The event period is days (0, +1) surrounding an announcement.t-stats are reported in parentheses.***, **, and * denote significances at 1%, 5%, and 10%, respectively.

Table 3 .
Regressions of Shareholder Wealth on Total Compensation.

Table 4 .
Regressions of Shareholder Wealth on Equity Compensation.

Table 5 .
Change in Shareholder Wealth to Change in Compensation of TARP Firms that Repaid Bailout Funds before 11 December 2009.LDIR), dummy for companies that paid back bailout money by 11 December 2009 (DPAY), and the interaction between DPAY and change in compensation.Models 1, 3, and 5 are based on pooled regression models whereas Models 2, 4, and 6 are fixed effects models.t-stats are reported in parentheses.***, **, and * denote significances at 1%, 5%, and 10%, respectively.