Incentive compensation for bank directors: The impact of deregulation ❶

Transcription

1 Incentive compensation for bank directors: The impact of deregulation ❶ David A. Becher a, Terry L. Campbell II b,*, Melissa B. Frye c a Department of Finance, Northern Illinois University, DeKalb, IL 60115, USA b Department of Finance, University of Delaware, Newark, DE 19716, USA c Department of Finance, University of Central Florida, Orlando, FL 32816, USA April 2003 Abstract Although industry deregulation leads to changes in the scale and scope of the duties of the board of directors, little is known about the changes in incentives for directors surrounding such events. The deregulation of the U.S. banking industry and associated technological and regulatory changes during the 1990s lends itself to a natural experiment. These industry shocks forced bank boards of directors to face expanded opportunity sets, increased competition, and a rapidly expanding market for corporate control. While bank directors receive significantly less equitybased compensation throughout most of our sample period, by the end of the decade their use of equity-based compensation is indistinguishable from a matched sample of industrial firms. Moreover, banks utilizing a high degree equity-based compensation for directors are associated with higher performance and higher growth without a similar increase in risk. The increase in the use of equity-based compensation for bank directors is not due to a fundamental shift in bank boards, as board size and independence have remained static. Overall, our results suggest that firms respond to deregulation by improving internal monitoring through aligning directors incentives with those of shareholders. ❶ We are grateful to Tammy Berry, Helen Bowers, Ivan Brick, Peter DaDalt, Jacqueline Garner, Jeff Harris, Albert Madansky, Sattar Mansi, David North, Robert Schweitzer, Stan Smith, seminar participants at the Financial Management Association, Southern Finance Association meetings, the University of Delaware, Northern Illinois University, and an anonymous referee for their invaluable comments. We also thank Vladimir Kotomin for research assistance. Any errors or omissions remain our own. Campbell acknowledges partial support from the Center for Corporate Governance in the Lerner College of Business and Economics at the University of Delaware and the KPMG Audit Committee Institute. *Corresponding author. Tel: ; Fax: ; address: campbelt@be.udel.edu

2 I. Introduction Numerous studies have focused on the role of corporate governance in reducing agency problems, generally focusing on how industrial firms should be governed and monitored. 1 However, less is known about the dynamics of governance structures or how governance structures are altered to react to changing business environments. Notably, altered incentives may be essential for firm success following the increased sensitivity of firm value to officer and director decision making associated with an exogenous industry shock such as deregulation. We address this shortcoming by examining how director incentives changed in response to deregulation in the U.S. banking industry. Certainly, less attention has been placed on the role of corporate governance in the U.S. banking industry than other non-regulated industries. Existing literature posits numerous factors that could cause corporate governance issues at banks to be different from industrial firms. While all firms potentially face conflicts of interests among various stakeholders, this situation is exacerbated in banking due to the nature of the industry. The existence of deposit insurance, high debt to equity ratios, and asset-liability issues among others may lead to greater conflicts. In fact, Macey and O Hara (2001a, 2001b) suggest bank directors are more important to firm success than counterparts at industrial firms. Despite greater potential for agency problems in banks, it has been argued executives and directors of firms in regulated industries such as banking should not need (or need less) incentive compensation as a portion of their remuneration. Specifically, for regulated firms, executive decision-making may be more transparent and opportunities to invest in projects severely limited (Demsetz and Lehn, 1985; Smith and Watts, 1992). Consistent with this, Bryan, Hwang, Klein, 1 See Shleifer and Vishny (1997) for a review. 1

3 and Lilien (2000) find for boards of directors both the level of stock option awards and ratio of stock to cash compensation are lower for regulated than non-regulated firms. Likewise, Houston and James (1995) find bank CEOs are paid less equity-based compensation than non-bank CEOs. However, the role of boards of directors at banks has begun to change because of a myriad of regulatory and technological changes during the past decade. As a result of this change in industry structure, banks began to face both expanded opportunity sets and increased competition as well as increased corporate control activity. Therefore, the role of boards of directors may have become significantly more important. Houston and James (1995) note that the eroding distinction between banks and non-banks implies that banks will face pressure to adopt compensation policies to those found in non-banks. Empirical work suggests that deregulation leads to the need for incentive based compensation for executives. For example, Crawford, Ezzell and Miles (1995) and Hubbard and Palia (1995) find a stronger link between CEO pay and firm performance for banks following interstate deregulation events. While banks clearly still differ from non-banks, these findings suggest that differences between banks and non-banks may be diminishing as the banking industry is further deregulated. As a result, banks are turning toward the use of equity-based compensation to align shareholder and executive interests and compensation structures at banks are becoming more like those at non-banks. While differences between bank and non-bank executive compensation have been well documented, no prior studies that we are aware of have examined differences between director compensation plans. Directors clearly serve an important monitoring role and how they are compensated should also be important due to agency problems between boards and shareholders. Perry (1999) finds a positive relation between CEO turnover and the use of incentive-based compensation for directors. The author concludes that firms adopt such plans to provide directors 2

4 with financial incentives to monitor management. Jensen (1993) argues that paying directors with stock or options will force board members to recognize their decisions will not only affect their own wealth but also the wealth of shareholders. We use the deregulation of the U.S. banking industry as a vehicle to ascertain the impact of deregulation on the incentives provided to boards of directors. Deregulation serves as an exogenous shock that changes the need for effective director monitoring and provides a natural experiment to explore the dynamics of corporate governance structures. We use data from ExecuComp on director compensation to compare banks and non-banks from The increase in competition and additional corporate control decisions required suggests that the oversight by bank boards of directors became much more important by the end of the decade. During the early 1990s, banks have significantly lower levels of equity-based compensation as a percentage of total compensation than non-banks over the same period. In addition, yearly changes in this percentage lag significantly behind changes for non-bank directors. In the late 1990s, the level of equity-based compensation is still lower for banks; however, the year-to-year changes in the percentage of equity-based compensation do not differ from non-banks. Moreover, when we compare our bank sample with a size-matched industrial firm sample, we find no difference in the level of director equity-based compensation. While bank directors were systematically lagging behind their counterparts at industrial firms with regards to incentive compensation in the early 1990s, we do not find evidence of this continuing into the 21 st century. During our sample period, banks utilizing high degrees of director equity-based compensation are associated with higher performance and growth without a related increase in risk. Moreover, these same banks are associated with subsequent higher growth, higher levels of engagement in the market for corporate control, weak evidence of increased performance with no 3

5 increase in risk. Finally, the increase in the use of equity-based compensation for bank directors during the 1990s is not mirrored by a change in board structure. That is, board size and independence of banks remain static while a matched sample of industrial firms show material increases in board independence. These changes indicate non-bank boards are moving toward board structures similar to those of banking firms. Overall, our results are consistent with the idea that as the operating environment in which banks compete changes, the incentives provided to the board of directors changes as well. Firms recognize the importance of effective director monitoring and alter their incentive to monitor following significant economic shocks. The remainder of the paper is organized as follows. In section II we review literature regarding boards of directors and the changing environment in the banking industry. When then motivate our testable implications. In section III we describe our sample and data collection. In section IV we present results of tests of differences in incentives for boards of directors in banks versus industrial firms. We offer a conclusion and implications for future research in section V. II. Motivation A. Deregulation of the Banking Industry The 1990s was a dynamic time period for the U.S. banking industry dominated by deregulation, changing technology and rapid consolidation. These changes have dramatically reshaped the landscape of the industry. This restructuring has led to an environment that by the end of the 1990s was materially different than the beginning of the decade. During the 1980s, the practicality of the existing bank regulatory structure began to decline. With the increased incidence of holding companies and improved technology, prior legislation proved to be a burden as banks attempted to cut costs while facing challenges from new firms providing banking services. As a result, banks have undergone a long period of 4

6 deregulation. For example, the Federal Deposit Insurance Corporation Improvement Act of 1991 mandated a least-cost resolution method and prompt resolution approach to problem and failing banks and ordered the creation of a risk-based deposit insurance assessment scheme. In 1994, the Riegle-Neal Interstate Banking and Branching Efficiency Act was signed into law, which effectively undid 70 years of state and federal restrictions on interstate banking and branching. Brook, Hendershott, and Lee (1998) determine this legislation s passage led to an $85 billion increase in value for the banking industry. In that same year, the Riegle Community Development and Regulatory Improvement Act reduced the bank regulatory burden and paperwork requirements. Another deregulatory change in the 1990s was the Economic Growth and Regulatory Paperwork Reduction Act of This Act modified the regulations impeding the flow of credit from lending institutions to businesses and consumers and streamlined the mortgage lending process. While research concludes deregulation affected bank market values, relatively little research has focused on the affect on governance structure. 2 Parallel to the changes in the regulatory environment, technological progress has also kept pace in the 1990s. Notably, advances in financial engineering, payment technologies, and delivery methods for depositor services have impacted the competitive environment (Berger, Demsetz, Strahan, 1999). Risk management has evolved as derivatives and other risk-shifting methods have increased in use. Enhancements in delivery services have taken place via ATMs and on-line banking. Increased efficiencies in payment technology have occurred in terms of electronic payments and standardization. Rajan and Zingales (2001) refer to these changes as a financial revolution, which has affected financial institutions and industrial firms substantially. 2 Adams and Mehran (2001) note the deregulation of the banking industry raises the question of whether we can expect market discipline and internal mechanisms to play an increasing role in bank governance. Also, Mayers, Shivdasani, and Smith (1992) note that compensation studies to date have focused largely on manufacturing firms and ignored regulated firms where incentives may differ. 5

7 Given the changes in the regulatory and technological environment, the banking industry has undergone an unparalleled wave of corporate control activity throughout the 1990s. Becher (2000) documents bank merger activity increased over 200% from the early 1980s until the 1990s as the number of banks or bank holding companies decreased by over 35%. Furthermore, the number and value of public bank mergers increased considerably from 1994 to 1997 (i.e., about 50 mergers a year worth, on average, over $1 billion). In fact, Mulherin and Boone (2000) document that the banking sector leads all others in terms of merger frequency in the 1990s. Thus, it appears the external market for control mechanisms has intensified during the 1990s. Kole and Lehn (1997) argue deregulation leads to increased managerial discretion and greater sensitivity of firm value to the quality of managerial decisions. They contend that deregulation increases the severity of agency problems and the need for more efficient governance structures. To examine how firms react to exogenous shocks in their competitive environment, Kole and Lehn (1999) study deregulation in the airline industry and find airline firms react to such shocks by altering their governance systems. Specific to banking, empirical evidence on internal corporate governance mechanisms is consistent with banks receiving additional monitoring following deregulation. Research suggests banks have increased their use of equity-based compensation to align shareholder and executive interests. Crawford et al. (1995) show a stronger link between CEO pay and performance after deregulation (DIDMCA, 1980; Garn-St. Germain Act, 1982) of the banking industry. Moreover, Hubbard and Palia (1995) find stronger pay-for-performance sensitivity after deregulation permitting interstate banking through state laws or interstate regional compacts. While Houston and James (1995) find bank executives receive a smaller percentage of equity-based compensation than non-bank executives; deregulation appears to influence the compensation packages at banks and move them toward 6

8 those of other industrial firms. In fact, while Houston and James (1995) do not examine deregulation, they do note that as banks become more like non-banks, banks will face pressure to adopt compensation policies to conform to those of non-banks. Finally, Kole and Lehn (1999) find after deregulation of the airline industry, governance structures gravitate toward systems used by non-regulated firms. Post-deregulation, CEO pay as well as stock options increase. Thus, the changing regulatory and technological environment of the banking industry in the 1990s fundamentally altered the structure of banks by removing regulatory barriers and increasing the investment opportunity set of banks. Research shows this deregulation has led to an increase in the use of executive incentive-based compensation. However, no empirical study that we are aware of has examined how deregulation affected director compensation. B. Director Compensation One way to align executive and shareholder interests is through ex ante contracting, where agency costs are mitigated through the use of equity-based compensation. Studies using executive compensation conclude firms reduce agency costs by using equity-based compensation (Jensen and Murphy 1990, Hall and Liebman 1998). Further, Mehran (1995) demonstrates firm performance is positively related to the proportion of equity-based executive compensation. Empirical studies on executives are consistent with firms using equity-based compensation as a means to reduce agency conflicts and that the use of such compensation is increasing over time. Aligning the interests of directors and shareholders, however, should also be important. Directors are charged with the responsibility of managing and supervising the business and affairs of a corporation on behalf of the shareholders. They are responsible for hiring, firing, evaluating, and monitoring the management team. In addition, Mayers, Shivdasani, and Smith (1997) show directors can exert control over expenses that are more likely to involve a conflict 7

9 between managers and shareholders. John and Senbet (1998) argue the proportion of independent directors and board size influences board effectiveness in its monitoring function. Increasing attention is now focusing on how directors are compensated. Incentive compensation leading to share ownership may make directors better monitors. To illustrate, Perry (1999) finds the use of incentive compensation by boards results in improved monitoring by directors. He shows the probability of CEO turnover following poor performance increases when directors receive equity-based compensation. Bhagat, Carey, and Elson (1999) find a link between CEO replacement and the equity-holdings of the directors. Shivdasani (1993) finds the likelihood of a hostile takeover increases when outside target directors own small percentages of shares. He suggests board monitoring may substitute for monitoring from the corporate control market. Furthermore, Hermalin and Weisbach (1998) and Gillette, Noe, and Rebello (2003) develop models where incentive compensation for directors increases their monitoring efforts. In addition to providing motivation to monitor, equity-based compensation may be needed to attract and retain quality directors. Brickley, Linck, and Coles (1999) suggest firms consider merit and ability when appointing outside directors, making compensation packages important in attracting quality directors. This is also consistent with Zingales (2000), who notes human capital is emerging as the most crucial asset to many firms. Attracting, motivating, and retaining talented employees and directors may, therefore, be critical to the success of today's firms where human capital is such a valuable asset. A focus on compensation plan design is also consistent with prior literature on executive effectiveness. Morgan and Poulsen (2001) find equity-based compensation is the most efficient way to align shareholder and manager interests. Given the small ownership by outside directors and efficiency of equity-based compensation, firms may turn toward equity-based compensation 8

10 to align director and shareholder interests. Further, Jensen (1993) proposes that director equitybased compensation will increase awareness about how their decisions affect shareholder wealth. Perry (1999) shows the use of director equity-based compensation has increased dramatically in the 1990s, suggesting shareholders see benefits to aligning director interests with their own. Non-academic research also supports equity-based compensation for directors. For example, The National Association of Corporate Directors Blue Ribbon Report (1995) encourages the use of equity-based compensation as a means of aligning shareholder and director interests. CalPERS U.S. Corporate Governance Principles recommends that company stock be a significant proportion of total director compensation. Consistent with this, Bryan et al. (2000) find board member compensation is designed to reduce agency conflicts between shareholders and directors. While Gerety, Hoi, and Robin (2001) do not find a significant market reaction to director incentive compensation plans, they do conclude that incentive compensation can enhance the alignment of director and shareholder interests. 3 C. Boards of Directors at Banks While the regulatory environment in which banks operate may partially serve to mitigate agency problems, bank directors still serve important monitoring functions. For example, all directors owe fiduciary duties to the firm and its shareholders. Macey and O'Hara (2001a), however, note that the nature of banking makes it susceptible to greater moral hazard issues than industrial firms. In fact, they contend bank directors should be held to a broader standard of care than non-bank directors. Banks provide stability to the financial system and directors thus serve 3 Gerety, Hoi, and Robin's (2001) sample includes new incentive compensation plans and modifications to existing ones. They analyze stock returns over a three-day period surrounding the mailing of the proxy. However, Brickley, Bhagat, and Lease (1985) report that proxy mailing practices differ. Some firms mail proxies before the proxy date, while in other cases proxies are mailed after. They conclude that considerable uncertainty exists about when information in proxy statements becomes publicly available. 9

11 a critical role. Macey and O Hara (2001a, 2001b) note FDIC insurance eliminates the need for depositors to monitor the risk-taking behavior of executives and thus places a greater burden on directors to safeguard the interests of society. They conclude directors of banks may actually be more important to the financial health of the firm than counterparts at industrial firms. Macey and O Hara (2001b) argue that in most publicly held firms, excessive risk taking is controlled by devices designed to protect fixed claimants (i.e., bond covenants), the introduction of incentive contracts, and the natural inclination of managers to be more riskaverse than stockholders. In banking, however, the doctrine of too-big-to-fail and the concept of federal deposit insurance weaken the need for uninsured and insured creditors to carefully monitor banks. The authors stipulate that creditors of banks may be limited in their ability to influence the firm due to the opaqueness of their balance sheets. Finally, the introduction of incentive contracts in banking may increase agency costs. If managers are more risk averse than stockholders and deposit insurance removes the incentive for creditors to monitor excess risk taking, aligning managers interests with that of shareholders may create incentives for banks to take on too much risk. In fact, Macey and O Hara (2001a, 2001b) suggest, from a social welfare perspective, bank executives generally should not be paid with incentive compensation because it encourages risk taking. While they do not explicitly mention director incentive compensation, similar arguments can be extended to bank directors as well. Thus, bank directors may not receive any, or at least receive substantially less, equity-based compensation than non-bank directors and executives. These arguments are not necessarily consistent with empirical evidence on executive compensation at banks. As indicated, studies on executive compensation at banks show an increasing use of equity-based compensation following deregulation. However, no study has 10

12 empirically examined bank director compensation. This analysis therefore, examines how changes in the regulatory structure and technological advances have altered director incentives. The need for director monitoring most likely increased substantially following such events. Banks face increased competition, greater uncertainty, as well as expanded opportunity sets, which provide substantial challenges to directors. Likewise, directors face a heightened market for corporate control, which impacts the very survival of the bank. These are arguably one the most important decisions directors make. Given these changes in the banking industry, it seems likely that need for incentive compensation has changed at banks as well. Specifically, banks may use more equity-based compensation following economic shocks as they begin to operate in an environment that is more similar to the environment facing non-regulated industrial firms. 4 III. Sample and Summary Statistics A. Overall sample We construct a sample consisting of all firms for which director compensation data are reported on Standard & Poor's (S&P) ExecuComp database. The ExecuComp database includes information on all non-employee director compensation over the period for the S&P 500, S&P Midcap 400, and S&P Smallcap 600. To identify banking firms, we examine proxy statements and 10Ks for all financial firms (SIC code in 6000s) to determine the firm's primary business. As a result, our sample consists of 13,847 observations and banks account for 700 of these. Table 1 provides a breakdown of the number of bank and non-bank companies by year. Non-employee directors are typically paid an annual retainer, fees for attending meetings, as well as equity-based incentives, including stock options and shares of stock. To value the 4 In November 1999, the Gramm-Leach-Bliley Act passed allowing banking firms, insurance companies, investment banks, and others to merge under a financial services holding company. This new legislation should further expand opportunities for banks and will likely affect the structure of incentive compensation. 11

13 stock options and shares of stock, we follow Mehran and Tracy's (2001) use of the values reported by ExecuComp for executives. ExecuComp values options using a modified Black- Scholes approach. Mehran and Tracy (2001) note that company handbooks on stock compensation plans generally do not distinguish between executive and non-executive equity compensation plans. 5 Furthermore, for a random sample of 50 firms, we examine individual proxy statements to verify data reported by ExecuComp and to ensure the integrity of the data. Total cash compensation is the sum of total meeting fees and the annual retainer. Total equity compensation is the sum of the value of the stock options granted and stock shares granted. Total compensation is the sum of total cash and total equity compensation. The percentage of equity-based compensation is the total equity compensation divided by the total compensation. This percentage provides information about the relative importance of equitybased compensation in the firm's director compensation plan. To examine bank characteristics, we collect data from Bank Compustat. We use performance and risk measures identified in prior banking studies (Cornett and Tehranian, 1992; Cornett, Mehran, and Tehranian, 1998; Cornett, Ors, and Tehranian, 2002). We examine two profitability indicators: return on asset and return on equity. These capture a bank s overall performance. Capital risk indicators measure a bank s ability to meet regulated capital standards and continue to attract deposits and loans. They include core capital to assets (shareholders equity as a percent of total assets), loans to total capital (total loans as a percent of total capital) and deposits to total capital (total deposits as a percent of total capital). We use yearly growth rate in asset to assess the banks growth. Bank Compustat reports data only for surviving banks. 5 Perry (1999) values options using a Black-Scholes method with assumptions about price and time to maturity. For the restricted shares of stock, he uses the price of the firm's stock at the end of the preceding calendar year. The results in Table 2 demonstrate that using the ExecuComp values produce qualitatively similar values to Perry's. 12

14 Finally, in order to examine how growth occurs, we examine the merger activity of the banking firms in our sample. In particular, using the CRSP delisting files combined with Lexis- Nexis, MergerStat Review, and the Wall Street Journal Index, for each bank in our sample, we collect all completed public bank mergers undertaken in the 1990s. For all these mergers, we collect the dollar amount of assets acquired. B. Matched samples To ascertain possible differences over time in board and incentive structure between an industry undergoing regulatory and other economic shocks and other non-regulated firms, we match a sample of banks to non-banks. Similar to Kole and Lehn (1999) and Flannery, Kwan, and Nimalendran (2002), we exclude financial firms (SIC ) and other regulated firms (SIC ) from our potential match firms. From the potential match pool, we select final matches based on market values in We examine board characteristics at the beginning of our sample period (1992) and the end (1999). The matching procedure allows us to examine shifts in board composition for the 86 banking firms in 1992 relative to a comparable non-regulated, non-financial firm. Note not all banks and not all matched firms survive until Given the market value of the surviving firms could have shifted dramatically over the eight years, we construct a second match in 1999 (new match) based on market values in We report results using both sets of matches in 1999, but the results are qualitatively similar. To examine board structure, we obtain board size as well as number of inside, outside, and gray directors from proxies and 10Ks. We follow Mayers, Shivdasani, and Smith (1997), Servaes and Zenner (1996), and Baysinger and Butler (1985) in classifying directors. Inside directors are those currently serving as executives. To capture gray or affiliated directors, we use two measures. Grey1 directors include all former or retired employees as well as directors related 13

15 to current or previous employees. Grey2 directors include directors who are or were associated in some business or financial capacity (i.e., lawyer, consultant, investment banker). Grey2 directors have a professional affiliation and may not serve in a true monitoring role. Outside directors are those with no relationship with the firm, past or present, other than their position on the board. Outside directors are most likely to serve as monitors on behalf of shareholders. We collect data for approximately 2,200 bank directors and 2,000 non-bank directors. IV. Results A. Banks versus Non-banks In Table 2, we examine bank versus non-bank director compensation by year to ascertain the dynamics of director compensation. Annual retainers for banks and non-banks increase over the sample period slightly; however, the two do not differ significantly. Total meeting fees paid to directors shows banks paying significantly higher fees in the early 1990s; however, the differences disappear in the latter part of the decade. Total cash compensation (annual retainer plus meeting fees) remains reasonably constant over the sample period for both banks and nonbanks. Total equity-based compensation dramatically increases for both banks and non-banks. To illustrate, bank directors average $2,680 in equity-based compensation in 1992 compared to over $26,000 in For non-banks, the average director receives $7,630 in 1992 compared to $49,090 in Total compensation also increases over the sample and banks generally lag significantly behind non-banks. 6 The percentage of equity-based compensation also shows a trend toward equity-based compensation for directors; however, banks use significantly less than 6 We also examine inflation-adjusted compensation figures (adjusted to 1999 dollars). Annual retainers, meeting fees, and total compensation remain relatively stable or fall slightly in the 1990s. Equity and total compensation increased substantially in terms of real dollars, which is consistent with the information in Table 2. To illustrate, real total compensation for banks (non-banks) increases from $29,061 ($35,990) in 1992 to $51,810 ($74,440) in Also, differences between banks and non-banks are consistent to those using inflation-adjusted figures. 14

16 non-banks in each year. While there have been some changes in bank director compensation over time, bank directors still receive a lower percentage of equity-based compensation than non-bank directors and receive less total compensation. B. Effects of Deregulation In Table 3, we compare compensation for bank and non-bank directors in levels and changes over three time periods. While these time periods are somewhat arbitrary, they reflect the notion that a great deal of deregulation took place during the 1990s and allow us to explore how banks responded over time. The early period includes data for , the middle period includes data for , and the late period includes data for Several interesting results emerge. First, while the change in total cash compensation is significantly greater for banks in the early period, it is not in the middle and late periods. Next, while the amount of total meeting fees for banks are significantly higher in early period (p-value of 0.00), the difference is only marginal (p-value of 0.10) in the late period. It seems the use of short-term compensation for bank directors versus others is indistinguishable by the end of the 1990s. In addition, the levels of equity and total compensation are significantly lower for banks than non-banks for all three periods. Since cash compensation levels did not differ significantly, the difference in total compensation is attributable to stock compensation. However, there are no differences in the changes in equity and total compensation for any of the three periods. In addition, while banks experience significantly lower changes in the percentage of equity-based compensation in the early period, the change is no longer significantly different in the latter two periods and is actually insignificantly higher for banks than non-banks. 7 Results are qualitatively the same if we alter the time periods by a year. 15

17 To illustrate, banks on average increase the percentage of equity-based compensation during the late (middle) period by 4.54% (4.34%) while non-banks increase use by 3.99% (4.14%). These results suggest while the level of equity-based compensation is significantly lower for banks than non-banks, the change beginning in the middle of the decade is not. Again, this is consistent with banks following non-banks in terms of increasing their use of equity-based compensation. The levels are thus lower since banks started at significantly lower levels. This may suggest both banks and non-banks perceive value from aligning shareholder and director interests. In addition, the deregulation results are consistent with studies on banks executive compensation (Crawford, Ezzell, and Miles, 1995, Hubbard and Palia, 1995), which show banks alter compensation structures following regulatory changes. It is also consistent with Kole and Lehn (1999) who show an increase in equity-based incentives for CEOs following deregulation of the airline industry. Firms alter monitoring incentives following major economic shocks. C. Deregulation and Other Determinants of Director Compensation We also analyze the determinants of the level of and changes in the percentage of equitybased compensation (Tables 4). We focus on equity-based compensation as this is the primary form of incentive compensation for directors and since the literature suggests equity-based compensation can be used to align shareholder and director interests. The multivariate approach allows us to examine deregulation effects after controlling for other potential determinants of director equity-based compensation. Since we have cross-sectional time-series data, a panel model is appropriate. Thus, we analyze the data with a random effects regression model. 8 Using 8 A fixed-effects specification assumes a fixed component across observations and is equivalent to adding a dummy variable for each firm in the sample. A random-effects specification assumes the model residual is comprised of a fixed component (fixed over time) and a component that varies both cross-sectionally and over time. Hausman (1978) specification tests for the consistency of the random effects estimates are unable to reject such specifications. 16

18 panel data allows us to control for unobservable or neglected firm-specific characteristics that may be correlated with the compensation and control variables in the models. By including random effects, we control for endogeneity associated with neglected firm heterogeneity. 9 A dummy variable equal to 1 for banks is used to determine if banks differ from nonbanks. In addition, we control for other variables used in prior literature to explain the use of equity-based compensation. 10 Specifically, we include controls for market-to-book ratio, leverage, firm size, and firm risk. We also include additional dummy variables for other nonfinancial regulated companies and financial companies. The market-to-book ratio is included to capture the effects of growth or investment opportunities. As investment opportunities expand, so does the information asymmetry. Shareholders will rely less on direct monitoring and attempt to reduce the agency costs by increasing the use of equity-based compensation. Debt-to-asset ratios (leverage) may be positively or negatively related to the use of equity-based compensation. Core and Guay (2001) argue that since option grants do not require a cash outlay, firms will rely more on such compensation when their ability to borrow is constrained. However, Jensen (1986) and Stulz (1990) show increased leverage can increase firm value by preventing managers from taking poor projects, suggesting debt serves to alleviate agency problems making equity-based compensation less necessary. Larger firms may be more difficult for shareholders to monitor, suggesting an increased need for equity-based incentives. However, it is also possible smaller firms may be more cash constrained and thus have a preference for equity-based compensation. Since firm risk may alter director incentives, we control for the riskiness of the firm's operations. 9 Cornwell and Trumbull (1994) note that failure to control for neglected variables will result in inconsistent coefficient estimates. They highlight that a panel model approach with fixed or random effects specifically controls for this endogeneity problem. 10 In general, we follow Bryan, Hwang, Klein, and Lilien (2000). However, several of their hypothesized determinants cannot be calculated for banks or are not applicable to banks (e.g., free cash flow). 17

19 Results from random effects model specifications in which the dependent variable is the percentage of equity-based compensation are in Table We provide six models with data in both levels and changes. The first two columns are for the early period ( ); the third and fourth columns are the middle period (1995 to 1996), while the final two columns are the late period (1997 to 1999). In each case, the first column (columns 1, 3, and 5) presents the results for levels, while the second columns (columns 2, 4, and 6) presents the results for changes in equity-based compensation. We split the data in order to examine whether the determinants of director incentive compensation have changed due to regulatory and technological shocks. Results from Table 4 indicate banks have significantly lower levels of equity-based compensation. 12 The dummy variable for banks is significantly negative for all three periods. Thus, deregulation has not led to banks increasing the level of equity-based compensation to that of all non-banks. Bank directors continue to receive lower levels of equity-based compensation. However, the bank dummy is significantly negatively related to the change in the percentage of equity-based compensation in the early period, but that difference disappears in both the middle and late period. This suggests prior to the significant regulatory and technological shocks of the 11 Note the percentage of equity-based compensation is bounded at 0 and 1 and changes in the percentage of equitybased compensation at -1 and 1, suggesting a Tobit model may be appropriate. To control for this, we originally tried to utilize a random effects Tobit model. However, our data lacked variation within the panel because of the presence of various dummy variables with the bounding caused the quadrature approximation used in the random effects Tobit estimators to be suspect. The quadrature is used to compute the log likelihood and its derivatives. Note that a conditional fixed effects model cannot be estimated for Tobit models because there does not exist a sufficient statistic allowing the fixed effects to be conditioned out of the likelihood. Since a random effects regression model does not bound the data, we are able to utilize this approach with the dummy variables. However, all reported results are qualitatively and quantitatively similar using a non-panel Tobit model. In addition, we follow Bryan, Hwang, Klein, and Lilien (2000) by bounding the Tobit models with only one limit (0 for levels and 1 for changes), which captures the majority of bounded observations. The results are again qualitatively to those reported. 12 Observations in Tables 4 are less than those reported in Table 1, attributable to missing stock valuations in ExecuComp (473 firm years) and Compustat data (614 firm years or roughly 77 of nearly 1,800 firms per year). Operating income data, used to measure firm risk, are the most common missing data. However, most lost observations are missing more than one control variable. The number of observations further decreases from levels to changes primarily because observations for 1992 are lost. ExecuComp also added companies during our sample period. We are able to compute changes in stock-based compensation for 10,720 firm years (changes cannot be calculated the first year a firm is included). The remaining 522 firm years lost are due to missing Compustat data. 18

20 1990s, banks increased their use of stock compensation at a significantly lower rate than nonbank firms. However, subsequent to the economic shocks, this difference disappears and the change in the percentage of equity-based compensation for bank directors is no different than non-banks. While banks may not be catching up with non-banks in terms of levels of equitybased compensation, they do appear to be keeping pace with changes in such compensation. The other hypothesized determinants of director equity-based compensation show varying degrees of statistical significance and economic inference. Both the level and change in the market-to-book ratio is statistically significant and positive during the early period and insignificant during others. This suggests that expanding growth opportunities in the early 1990s motivated firms to increase their use of equity-based compensation. However, in the late 1990s, most firms increased their reliance on such compensation, not just firms with expanding growth opportunities. Leverage is negative and statistically significant for levels and changes in the early period and levels in the middle period. This is consistent with the notion highly leveraged firms may not need additional monitoring in the form of equity-based compensation, as leverage itself may serve as a monitoring mechanism. This relationship is consistent with Bryan et al. (2000), who find a negative relationship between equity-based compensation and leverage for CEOs in the insurance industry. However, this relation disappears by the late period. Size appears to be important in the early years of our sample, suggesting smaller firms were more likely to use equity-based compensation given the negative and significant coefficient. If smaller firms tend to be more cash constrained, they may prefer equity-based compensation since there is no direct cash outlay for the company. However, by the late period, the size variable loses its significance. The change in firm size is positive and significant in all model specifications. This is consistent with the hypothesis that as firms become larger, they are 19

21 more difficult for shareholders to monitor. Thus, they increase their reliance on equity-based compensation. Consistent with Bryan et al. (2000), we find non-financial regulated firms also use less equity-based compensation. The financial companies generally do not differ significantly from other industrial firms. This is consistent with the notion all financial firms are not the same, as banks appear to differ significantly from other financial companies. D. The impact of bank director equity-based compensation Arguably, deregulation and the changing environment facing banks in the 1990s had a differential impact within the banking industry, suggesting the shift in director incentives may not be uniform across all banks. To explore this, we examine cross-sectional differences within the bank sample. Specifically, we compare performance, risk, growth, monitoring activities, and corporate control activities for banks that changed director incentives substantially to those did not. The performance variables that we analyze are a subset taken from prior banking studies and are discussed in Section III. We also measure director monitoring by examining CEO turnover. Specifically, CEO turnover is one if the bank has a new CEO and 0 otherwise. Replacing a CEO represents one the most important monitoring function of corporate boards. Finally, we include two corporate control indicators: the natural log of average assets acquired and the natural log of total assets acquired. It is possible director equity-based incentives encourage banks to be more aggressive players in the corporate control market. Table 5 presents results for our sample of banks based on their use of equity-based compensation. 13 The first two columns split the bank sample based on their use of equity-based 13 For these analyses, we focus only on banks in our sample for the entire sample period ( ). Our focus on these banks allows us to capture and analyze how the dynamics of director incentives and performance are related. Banks that enter or leave the sample do not allow us to really capture the evolution of director incentives. To ensure robustness of our results, we repeat the analyses including all banks. Results are qualitatively similar. 20

22 compensation during the 1990s. A one indicates those banks utilizing a relatively high level of director equity-based compensation, while a zero indicates those banks utilizing a low level of director equity-based compensation. The third column presents a p-value for differences between the sample means. 14 Likewise, the right-hand columns split the bank sample into higher (lower) than average changes in the use of equity-based compensation measured by a one (zero). The cross-sectional differences highlight interesting differences for banks. Banks in which directors are compensated with high levels of equity-based compensation are associated with significantly higher overall performance as measured by either return on assets (ROA) or return on equity (ROE). Moreover, banks that increased their use of equity-based incentives the most are associated with significantly higher overall performance. This supports Mehran s (1995) link between equity-based compensation for executives and firm performance. We also test if the improved performance for banks using high degrees of equity-based compensation is associated with increased risk. As shown by the capital risk indicators, these banks are not associated with increased risk. In fact, banks using a high level of EBC have both significantly higher core capital to assets and loans to total capital. Cornett et al. (2002) indicate these ratios measure a bank s ability to meet regulated capital standards while still attracting loans and deposits. Thus, it appears banks using a high level of EBC have less capital risk. The growth rate in assets is significantly higher for firms using high levels and changes of EBC. Interestingly, this greater growth in assets is not due to being more active in the market for corporate control as measured by the corporate control indicators. These banks seem to be generating higher levels of internal growth. Finally, we observe no significant difference in CEO turnover between high and low EBC users. While on the surface this may suggest increased 14 Not included for brevity, we test for differences in medians. Results are qualitatively and quantitatively similar. 21

23 incentive-based compensation does not lead to increased monitoring by a board and contrast with Perry (1999), firms using high equity-based compensation perform better and thus may have less need to replace an existing CEO. Finally, our findings are similar to Kole and Lehn (1999) who do not find evidence the threat of dismissal increased in the airline industry post deregulation. Overall, we find cross-sectional differences in bank performance based on the use of equity-based compensation during from 1992 to Banks providing high levels of director equity-based compensation are associated with better performance without a related increase in risk. Moreover, these banks are associated with significantly greater internal asset growth rates. We next examine if the use of equity-based compensation for bank directors in our sample period (1992 to 1999) is related to subsequent bank performance ( ). Table 6 replicates the sample of high and low EBC users, but measures performance and other characteristics for We find weak evidence banks using high equity-based compensation in our sample have higher performance. Mean values for ROE and ROA are higher for banks using the most equity-based compensation and for banks increasing their use the most; however, the differences are not statistically significant. Similar to the within sample period results, we find no increase in the level of risk. In fact, loans to capital and deposits to capital are significantly higher for banks using the most equity-based compensation. Again, we observe significantly greater asset growth for banks using high degrees of equity-based compensation. However, in this out-of-sample period, the difference may be partially attributed to merger activity. As indicated by the corporate control indicators, banks using a higher than average change in equity-based compensation are marginally more active in the takeover market in the early 2000s. 22