Board Independence and Compensation Policies in Large Bank Holding Companies.

James F. Nielsen [*]

We use a sample of large bank holding companies to empirically examine the association between financial performance and organizational structure. We regress firm accounting performance on measures of board independence, CEO pay-performance sensitivity, the product

of board independence and CEO pay-performance sensitivity, and other organizational features and control variables. We find that both CEO pay-performance sensitivity and the relative tenure of independent outside directors have a positive effect on accounting performance. Their interactive effect tends to be negative. Thus, the marginal value of each mechanism for accounting performance declines as the use of the other mechanism increases. These results are robust in a simultaneous equations framework that accounts for endogeneity issues. We also find a positive relation between the percentage of independent outside directors and CEO pay-performance sensitivity.

The board of directors is considered an important internal corporate control mechanism (Fama, 1980, and Fama and Jensen, 1983). Among other things, the directors are responsible for evaluating the chief executive officer (CEO) and other top-executives, determining the level and structure of top-executive compensation, and replacing poorly performing CEOs. Recent studies find that the internal control mechanism, which operates through the board of directors, is generally effective. [1] Because the board of directors is ultimately responsible for determining the compensation package for top executives, the effectiveness of executive compensation in reducing shareholder-manager agency problems can depend on the board composition and on the level of board independence.

Our paper investigates the association between board-independence variables related to independent outside directors, pay-related incentives, and financial performance. We focus on the substitution-monitoring hypothesis, which posits

that the two monitoring variables, board independence, and pay-related incentives, are substitutes. Under this hypothesis, board independence effects should reduce the need to develop compensation contracts that align the interests of managers and shareholders.

We choose the banking industry for three major reasons. [2] First, by focusing on a single industry (Collins, Blackwell, and Sinkey, 1995), we minimize the variability in executive compensation that results purely from industry differences. Industry factors, such as competitive intensity or risk, can account for a large proportion of performance and executive compensation variability in a sample of firms (Smith and Watts, 1992), and might obscure the relation between board independence and executive compensation.

Second, not all firms have the same level of internal monitoring by the board (Lippert and Moore, 1995). Lorsch and Maclver (1989) find that board members exercise more power when firms are facing external threats, and that some of these threats might be the result of a high level of regulatory supervision or surveillance. Moreover, Kosnik (1990) and Johnson, Hoskisson, and Hitt (1993) found that board characteristics might have a greater influence on director involvement during critical times or decisions.

In the last two decades, the banking industry has had to deal with more pressure in both the legislative and economic arenas than any time since the 1930s. By necessity, all banking boards of directors have had to take a more active role in management. Thus, we are able to focus on compensation policies in firms that generally have a high level of internal monitoring by the board. [3]

Third, not all firms experience the same level of agency conflicts or have the same need for board monitoring. Since the 1980s, the deregulation of the financial services industry has opened up investment opportunities for all banks and allowed them to offer nonbanking financial services. Smith and Watts (1992) find that the level of agency conflict can vary across firms' investment opportunity sets: the greater the investment opportunity set, the greater the level of agency conflict. By focusing on the banking industry, we minimize the variability in executive compensation that results from differences in the level of agency conflicts, as a consequence of differences in investment opportunity conditions.

To examine how board independence and compensation policies influence bank performance, we use two separate regression models. The first model uses an ordinary least squares approach to investigate the extent to which various board independence measures interact with CEO pay-performance sensitivity. This analysis controls for firm size, growth opportunities, ownership structure, leadership structure, and market takeover activity. The second model estimates a simultaneous equation system of performance, board independence, and pay-related incentives. In this second model, we use a two-stage least squares (2SLS) method to address potential endogeneity issues.

In Section I, we discuss the related literature on the effectiveness of the board of directors and compensation policies in monitoring top management to reduce shareholder-manager agency problems. In Section II, we describe the sample and the variables used in the study. In Section III, we develop the research hypotheses, provide the empirical design for testing these hypotheses, and report the results. Section IV concludes.

I. Related Literature

Some of the earliest literature that deals with the effectiveness of boards of directors provides empirical support for the importance of outside directors. For example, Fama (1980) argues that outside directors are professional referees whose task is to ensure the effectiveness of the firm's top management. Fama and Jensen (1983) suggest that outside directors have incentives to develop reputations for being directors of well-run companies.

More recently, Kaplan and Reishus (1990) and Gilson (1990) investigate the market for outside directorships. Their results confirm that the external labor market for directorships is effective in disciplining directors for poor performance. Brickley, Linck, and Coles (1999) and Hoi and Coles (2000) report that poorly performing CEOs are less likely to hold board seats after they leave office, which suggests that firms consider merit and ability in selecting their outside directors.

Studies dealing with the market for corporate control include Shivdasani (1993), who finds that if board members have additional outside directorships, it decreases the probability of a takeover. Whidbee (1997) finds that more outside directors are appointed to the board when outside investors have significant control over the firm. Byrd and Hickman (1992) show that bidder firms with a majority of outside directors make better acquisitions. Brickley, Coles, and Terry (1994) provide strong evidence that the fraction of outside directors on a board is positively related to the abnormal returns at the announcement of poison pill adoptions. Cotter, Shivdasani, and Zenner (1997) find that outside directors enhance shareholder gains in tender offers.

Studies that deal with board composition and firm performance include Rosenstein and Wyatt (1990), who report positive abnormal returns at the announcement of the addition of outside directors. When they examine the factors affecting the supply of outside directors, Booth and Deli (1996) find a negative relation between the number of outside directorships and the firm's growth opportunities. Agrawal and Knoeber (1996) investigate various corporate control mechanisms empirically, and find that firm performance is actually reduced when more outsiders serve on the board. Subrahmanyam, Rangan, and Rosenstein (1997) also find that abnormal returns are negatively related to the proportion of independent outside directors on the boards of bidding banks.

Other research focuses on the need to regulate board composition to increase outside representation and overall board effectiveness (American Law Institute, 1982). However, MacAvoy, Cantor, Dana, and Peck (1983) do not find evidence to support the fact that firms with a majority of outside directors on the board perform better than those firms whose boards fail to meet American Law Institute's proposed guidelines. Demsetz (1983) argues against the need of regulating board composition, maintaining that executive compensation contracts are sufficient to guide top management in serving the interests of shareholders. Cahen and Wilkinson (1999) find that the proportion of outside directors in a sample of New Zealand companies increased by about 5% after the 1993 Companies Act. Their study suggests that the value of outside directors actually increases as a result of the political process.

Another major body of literature deals with CEO compensation issues and firm performance. In examining the effectiveness of executive compensation policies, Murphy (1985) finds that executive compensation is strongly and positively related to shareholder returns. Coughlan and Schmidt (1985) also find that salary and bonus changes are significantly related to shareholder returns. Gerhart and Milkovich (1990) show that future profitability is positively related to the level of incentives in the compensation mix. On the other hand, Jensen and Murphy (1990a), find that pay-performance alignment is weak in large US corporations. [4] McConaughy and Mishra (1996) find that increasing pay-performance alignment increases performance in low-performance firms, but has little impact on high-performance firms.

Agrawal and Knoeber (1996), Beatty and Zajac (1994), and Mehran (1992) focus their research on managerial equity holdings. Their results are consistent with Demsetz and Lehn (1985), who find no relation between managerial equity holdings and firm performance. Loderer and Martin (1997) and Cho (1998) find that managerial equity holdings do not affect firm performance, but firm performance does affect managerial equity holdings. Implicit in these results is the fact that other factors could be influencing the pay-performance relation.

Studies on the role of the board compensation committee include Jordan (1999), who argues that the firm's compensation committee does not control the level of managerial equity holdings. Rather, managers themselves use their private, firm-specific information to determine the level of their shareholdings. Newman and Mozes (1999) examine the composition of the board compensation committee under different performance conditions, and find that there is no difference in compensation practices between firms when performance is strong. However, when performance is weak, the insider-influenced firms place less weight on the relation between CEO compensation and firm performance than do the outsider-influenced firms.

Baliga, Moyer, and Rao (1996) and Brickley, Coles, and Jarrell (1997) investigate the leadership structure issue. They present arguments both for and against a combined leadership structure (when the CEO is also the board chair). However, both studies find that combining leadership positions consolidates authority in a single manager's hands and might hinder the board's ability to perform its monitoring role. Baliga, Moyer and Rao find little evidence that leadership structure affects firm performance, and Brickley, Coles, and Jarrell find that the effect of the combined leadership on firm performance is positive.

Pi and Timme (1993) find that banking firms with a separate leadership structure outperform those with a combined leadership structure. On the other hand, it could be more efficient to combine leadership positions. Thus, in many firms, CEOs who perform well are sometimes granted both titles. Even though the evidence is inconclusive, McWilliams and Sen (1997) and Coles and Hesterly (2000) suggest that monitoring by outside independent directors is important when the CEO is also the chairman of the board.

II. Data

We construct our data set from the 100 largest commercial bank holding companies, as measured by their December 31, 1990 total asset levels. All performance and board independence measures come from the banks' 1990 annual reports and proxy statements. [5] Since 11 of the top 100 banks were not publicly held and were, in fact, wholly owned by foreign corporations, our sample was reduced to 89 firms. Our sample was further reduced by another 22 banks, for which we were unable to obtain pay-performance sensitivity data. We enlarged the sample later to include 40 randomly chosen nonfinancial firms.

We obtain our pay-performance sensitivity data from Jensen and Murphy (1990b), who report sensitivity as the dollar change in CEO pay per $1,000 change in shareholder wealth. Jensen and Murphy report sensitivities for various types of incentives, such as salary and bonus, other pay, which includes performance-based rewards, loss of pay due to the increased probability of dismissal, stock options, and the CEO's stock holdings. We call the sum of these "Total Incentives."

We note that there is a potential shortcoming in using the Jensen-Murphy data set. We estimate these coefficients with a minimum of eight and a maximum of 15 observations over a 15-year time frame. Earlier authors who have used this data set have pointed out the measurement error problem as well. [6] In a multivariate setting, if the variables are not measured precisely, the measurement error can either work against finding a relation (even if there is one) or in favor of it, depending on how the sensitivity coefficients are correlated with the other regressors.

In our study, we calculate the pay-performance sensitivities from the pay and incentives that are under the direct control of the board. These sensitivities include the CEO's salary, bonus, accounting-based performance rewards, and stock options. We use the Forbes executive compensation surveys to calculate the pay-performance sensitivities over two separate periods. The first period covers the entire 15-year period from 1975 through 1989. We obtain this data from Jensen and Murphy (1990b). The second period comprises the eight years from 1982 through 1989. Using this period allows us to account for the change in the pay-performance relation of bank compensation contracts resulting from the deregulation moves of the early 1980s, as documented by Crawford, Ezzell, and Miles (l995). [7]

Our two main indicators for measuring board independence are %OUTDIR and RELTENURE. %OUTDIR is the percentage ratio of outside directors to total directors who actively sit on the board. We define an outside director as a board member whose only relationship with the firm involves their directorship (Byrd and Hickman, 1992). We calculate RELTENURE by dividing the average tenure of the outside directors by the number of years the CEO has held his/her position.

We use a three-way classification scheme developed by Baysinger and Butler (1985) and Byrd and Hickman (1992). Based on the affiliations and transactions noted in a firm's proxy statement, directors fall into one of three categories: inside directors, affiliated outside directors ("gray" directors), and independent outside directors. Inside directors include employees, retired employees, or members of their immediate families. Affiliated outside directors include investment bankers, commercial bankers, lawyers, consultants, officers and directors of the firm's suppliers and customers, and interlocking directors. Independent outside directors include business executives, academicians, public policy makers, and private investors.

Our study defines an outside director as an independent outside director. Gray directors are excluded. Because outside directors are usually independent of the CEO, it follows that the greater the percentage of outside directors, the more likely it is that significant checks and balances will be maintained in board deliberations (Vance, 1983), and the greater the degree of overall board independence.

However, a board's ability to carry out its legal role of representing shareholder interests over the interests of management can also be greatly influenced by the power the CEO has over the board (Pearce and Zahra, 1991). Since the CEO has the ability to shape board membership over time (Alderfer, 1986), the CEO can gain power the longer he/she holds the position. By measuring the average tenure of outside directors relative to the CEO, we attempt to measure the quality of board independence or the degree of influence the outside directors have in corporate governance (Singh and Harianto, 1989).

The longer the outside directors have served on the board, the greater their organization influence (Singh and Harianto, 1989), and the less likely they are to succumb to management pressures for conformity (Kosnik, 1990). However, the longer the members of a board of directors have worked together, the more likely they are to be committed to the status quo (Katz, 1982), to resist change (Goodstein and Boeker, 1991), and to tolerate poor performance on the part of senior management (Vance, 1983). As a result, as far as tenure is concerned, heterogeneous boards are more likely to do a more effective job of monitoring senior management and interceding on the part of shareholders (Kosnik, 1990). Finally, one of the strongest ways in which outside directors can influence senior management is through membership on the compensation committee.

We use two variables to measure bank performance, return on average total assets (ROA) and return on average common equity (ROE). Both of these variables represent accounting measures of financial performance and have been widely used in performance-based studies. We calculate these variables based on year-end 1990 data.

We also gather data on three additional variables: ownership structure, as measured by the percentage of stock held by directors (inside, affiliated outside, independent outside), institutions, and blockholders (Brickley, Lease, and Smith, 1988); leadership structure, as measured by CEO tenure and CEO/chairman duality (Brickley, Coles, and Jarrell, 1997); and market takeover activity, as measured by the restrictiveness of interstate branching laws (Brickley and James, 1987). Information for these measures comes from Disclosure, legal statutes that took effect in 1990 in the states in which the bank holding companies are operating, and 1990 annual reports and proxy statements.

III. Empirical Tests

In principle, the two monitoring variables, board independence and pay-related incentives, could be either substitutes or complements. If independent boards provide the necessary monitoring of senior management, the need for compensation contracts to align the interests of managers and shareholders can be reduced. On the other hand, if independent boards and compensation contracts complement each other, both would play a role in the monitoring function. Since the substitution-monitoring hypothesis is more solidly grounded in theory (Beatty and Zajac, 1994; Mehran, 1992), we examine the association between the two corporate control mechanisms on the basis of this hypothesis. As part of this analysis, we control for firm size, growth opportunities, ownership structure, leadership structure, and market takeover activity.

In Table I, we report descriptive statistics for the sample of 67 banks. The mean board size is 16, of which 64% are outsiders. This result is consistent with Kosnik (1990) who found that, on average, 62% of board members are outsiders. Jensen and Murphy (1990a) report that the median pay-performance sensitivity for all firms is $3.25 (CEO wealth changes an average of $3.25 for every $1,000 change in shareholder wealth). We find that the median level of pay-performance sensitivity is $3.90 (mean = $4.27), suggesting a slightly higher level of incentive alignment in banking firms.

A. Ordinary-Least-Squares (OLS) Regression Results

In Table II, after controlling for firm size, growth opportunities, and other organizational features, we examine the relation between pay-performance sensitivity, board independence, and financial performance. We measure firm size by the natural logarithm of total assets and growth opportunity by the ratio of market value per share to book value per share as of yearend 1990 (Smith and Watts, 1992; Gaver and Gaver, 1995).

Due to space limitations, the regression results reported in Table II use only ROA as the dependent variable. We find similar results when we use ROE as the dependent variable. We also use (but do not report) three additional measures of board independence: the average length of time the independent outside directors have served on the board, the tenure heterogeneity of independent outside directors, and the percentage of independent outside directors on the board's compensation committee. In all cases, the results are similar to the results obtained with the tenure-related board independence variable (RELTENURE).

Model 1 includes a leadership structure variable (CEO duality). This is a zero/one variable, which equals one when the CEO and chairman of the board positions are combined. Model 2 adds three ownership variables, percent of ownership by inside directors, percent of ownership by affiliated outside directors, and percent of ownership by independent outside directors.

The percent of ownership by inside directors might have two offsetting effects, a positive effect due to the incentive alignment and a negative effect due to the increased ability to entrench by resisting a value-increasing takeover. If the percent of ownership by outside directors enhances board monitoring, it should have a positive effect on performance.

Models 3, 4, and 5 build on Model 2. In Model 3, we interact the independence variable with the incentives variable to determine if board independence really is a substitute for pay-related incentives. Under substitution monitoring, we expect the coefficient of the interactive variable to be negative. In Model 4, we interact the independence variable with the CEO duality variable to determine if the marginal effect of board independence is reduced by CEO duality. If it is, we also expect the interaction coefficient to be negative. To complete our tests, Model 5 includes both interaction variables.

Our approach is analogous to a production function with one output and two inputs, where the output is firm performance, and the inputs include board independence and pay-performance sensitivity. In Model 3, when we measure board independence by the relative tenure of independent outside directors (RELTENURE), the derivative of accounting performance with respect to board independence ([partial](Performance)/[partial](RELTENURE)) is 0.2783, the derivative of accounting performance with respect to pay-related incentives ([partial](Performance)/[partial](Pay-Related Incentives) is 0.0630, and the "marginal rate of technical substitution," holding accounting performance constant, is negative [[partial](Pay-Related Incentives)/[partial](RELTENURE) = - ([partial](Performance)/[partial](RELTENURE)) / ([partial](Performance)/[partial](Pay-Related Incentives) = -0.2783/0.0630 = -4.42]. Thus, the marginal products of pay-related incentives and the relative tenure of independent outside directors are positive, and these inputs can be substituted for one another along a negatively sloped isoquant to maintain accounting performance.

In terms of their economic significance, when we maintain both variables at their mean levels, we find that 4.42 units of pay-related incentives are needed to substitute for the effect of one unit of relative tenure of outside directors on accounting performance. In examining the regression coefficient elasticities and the percent changes, we find that a 1% increase in pay-related incentives increases accounting performance by 0.53% (4.27*0.0630/0.51 = 0.53), but a 1% increase in the relative tenure of outside directors increases accounting performance by 0.57% (1.05*0.2783/0.51 = 0.57). The numbers 4.27, 1.05, and 0.51 are the mean values of the pay-related incentives, RELTENURE, and ROA variables shown in Table I.

In Models 3 and 5, the signs of the board independence and pay-related incentives interactive variables are negative. In further reviewing the size of the coefficient estimates in Model 3, it appears that the added return from increasing the relative tenure variable (RELTENURE) is outweighed by the reduction in return attributed to the interactive variable. For example, in Model 3, we have 0.2783 - (0.0727*4.27) = -0.0321, which suggests a substitution relation between the relative tenure of outside directors and CEO's pay-related incentives.

When we observe the magnitude of the coefficient on the proportion of outside directors in Model 3, we see that a 1% increase in the proportion of outside directors reduces the accounting performance by 0.0077%. The negative sign is consistent with the findings of Agrawal and Knoeber (1996) and Subramanyam, Rangan, and Rosenstein (1997) that bank performance is negatively related to the percentage of outside directors on the board.

In Models 4 and 5, we find that the CEO duality interactive coefficients are also significant. The negative interaction indicates that board independence is blunted when the CEO is also the chairman of the board. The positive coefficient on the board independence variable suggests that accounting performance would actually be improved if board independence increases and the CEO does not serve as chair.

Other Table II results show that the growth opportunity variable (market-to-book ratio) is the most significant factor in bank performance. As expected, the coefficient for this variable is both positive and significant in all regression equations. The coefficient signs on the ownership variables indicate that inside director ownership actually reduces bank performance and affiliated outside director ownership and independent outside director ownership increase bank performance. The negative coefficient on the CEO duality variable indicates that bank performance is lower when the CEO is also the chairman of the board. The results in Table II are robust to the inclusion of other organizational features, such as CEO tenure, state takeover restrictions, and institutional and blockholder ownership.

B. Two-Stage Least Squares (2SLS) Regression Results

We analyze the association between board independence, pay-related incentives, and bank performance by using a two-stage least squares approach within a simultaneous system of equations (Agrawal and Knober, 1996; Loderer and Martin, 1997; Cho, 1998). We estimate the following system of three equations:

Performance = f (Size, Growth Opportunity, CEO Duality, %Ownership by Outside Independent Directors, Board Independence, Pay-Related Incentives, Board Independence*Pay-Related Incentives, Board Independence*CEO Duality) (1)

Board Independence = f(Size, Growth Opportunity, CEO Duality, CEO Tenure, Pay-Related Incentives, Performance) (2)

Pay-Related Incentives = f(Size, Growth Opportunity, % Ownership by CEO, %Ownership by Inside Directors, %Ownership by Outside Affiliate Directors, Board Independence) (3)

Equation (1), the performance equation, includes CEO duality, board independence, pay-related incentives, their interactions, and other control variables.

Equation (2), which examines board independence, contains the CEO leadership attributes (CEO duality, CEO tenure, pay-related incentives), and performance. CEO duality tends to reduce board independence. If the substitution-monitoring hypothesis holds, we would expect to see a negative sign on the coefficient of pay-related incentives. Moreover, if high performance were associated with a high level of board independence, the sign on the performance coefficient should be positive.

Equation (3), the pay-related incentives equation, includes three-insider ownership structure variables (%ownership by the CEO, %ownership by inside directors not including the CEO, and %ownership by outside affiliated directors) and board independence. The greater the insider ownership and the more independent the board, the greater the pay performance sensitivity (pay-related incentives). However, if the substitution-monitoring hypothesis holds, the sign of the board independence variable should be negative.

All three equations include firm size and market-to-book ratio variables. We also use as additional variables size, market-to-book ratio, sales growth, % ownership by inside directors, % ownership by outside affiliated directors, % ownership by outside independent directors, % ownership by CEO, CEO tenure, and CEO duality. The total system includes three equations and nine exogenous variables. To identify the system of equations, we must exclude at least two exogenous variables from each equation. In our case, the system becomes overidentified, and a two-stage least squares approach is more appropriate (Pindyck and Rubinfeld, 1991).

In Table III, we see several interesting results. In the performance equations, the signs of the board independence, pay-related incentives, and the interactive variables are consistent with those in Table II. In the board independence equations, the sign of the performance variable is positive and statistically significant in the relative tenure equation, indicating that high-performance banks generally have high levels of board independence.

In the pay-related incentives equations, we note a negative coefficient (-0.0953) for the board independence variable when we measure board independence by RELTENURE. In the RELTENURE equation, we also note a negative coefficient for pay-related incentives (-0.0075). Even though these coefficients are not statistically significant, the signs support a substitution relation between the relative tenure of outside directors and CEO's pay-related incentives.

However, when we measure board independence by %OUTDIR, we note a statistically significant positive coefficient (0.1674) for board independence in the pay-related incentives equation, and also a statistically significant positive coefficient for pay-related incentives (2.0900) in the %OUTDIR equation. These results suggest that a complementary relation exists between %OUTDIR and CEO's pay-related incentives.

We examine several other specifications for the three equations, and the signs and the significance of the seven coefficients (board independence, pay-relayed incentives, CEO duality, their products, and performance) are similar in all specifications. For example, when we included the percentage of stock held by institutions and the percentage held by blockholders in the board independence equations (Whidbee, 1997), it did not change the signs or significance of the performance and the pay-related incentives variables.

When we tried using two lagged measures of performance (average ROA over the prior two years and sales growth over the prior five years) in the pay-related incentives equation, it did not change the reported results for the board independence variables. The coefficient on the lagged ROA variable is negative but insignificant, and the coefficient on the prior sales growth variable is negative and significant in all pay-related incentives regressions, indicating that lower growth (performance) firms use higher levels of pay-related incentives.

To test the hypothesis that the sensitivity of performance to the organizational features is different for banks relative to nonfinancial firms, we collected data on 40 nonfinancial and nonutility firms, which were randomly chosen from the Jensen and Murphy (1990b) list. We obtained board and CEO information from proxy reports, and the accounting and ownership data from disclosure statements. All the data are for 1990.

Compared to banking firms, the sample of nonfinancial firms has lower average pay-related incentives ($2.32 vs. $4.27), lower average %OUTDIR (58% vs. 64%), lower average CEO tenure (4.76 years vs. 12.08 years), greater average RELTENURE (2.59 years vs. 1.05 years), lower average CEO duality (87% vs. 95%), lower average ownership by inside directors (1.31% vs. 2.29%), greater average ownership by outside affiliated directors (1.46% vs. 0.94%), greater average ownership by outside independent directors (1.63% vs. 1.32%), and a greater average market-to-book ratio (2.78 vs. 0.88).

In the regression analysis, we include a dummy variable to differentiate banking firms (1) from nonbanking firms (0). In the two-stage least squares regressions, we use the same set of instruments that we used in Table III, p1us two-digit SIC industry dummies. These results are shown in Table IV.

Based on the significance of the 2SLS coefficients of the interactive variables, there does not appear to be a significant difference in the effect of board independence variables on accounting performance between banks and nonfinancial firms. However, in the case of 2SLS estimates of pay-related incentives, the coefficient is insignificant and negative for nonfinancial firms and significantly positive for banks. For example, the coefficient for nonfinancial firms is -0.0476, and the coefficient for banks is 0.0425 (-0.0476 + 0.0901). The positive significant coefficient for banks suggests that pay-related incentives play a more positive role in banking firms.

Furthermore, in the case of CEO duality, the effect on firm performance was negative for banks and positive for nonfinancial firms. For example, with 2SLS estimates, the coefficient for nonfinancial firms is 4.6822, and the coefficient for banks is -0.9462 (4.6822 - 5.6284). As a result, in nonfinancial firms, combined leadership structure (CEO duality) appears to play a more positive role in the performance equation. This finding is consistent with Brickley, Coles, and Jarrell (1997), who find that in seven of the eight specifications (in their Table 10), the estimated coefficient on CEO duality is positive.

These results are also consistent with the notion that, in high-performance firms, the CEOs are granted both titles. Overall, we agree with Brickley, Coles, and Jarrell (1997) that more detailed analysis of the costs and benefits of alternative leadership structures and compensation structures (CEO compensation sensitivity and CEO duality) are needed before we can reach a definitive conclusion on their efficiency in financial and nonfinancial firms.

IV. Summary and Conclusions

In this paper, we investigate the association between board independence, bank compensation policies, financial performance, and other organizational features. We focus on independent outside directors and the use of the substitution-monitoring hypothesis. Under this hypothesis, board independence effects should reduce the need to develop compensation contracts to align the interests of managers and shareholders.

We conduct tests by using both ordinary least squares and two-stage least squares regression models and various measures of board independence. Throughout the analysis, we control for firm size, growth opportunities, ownership structure, leadership structure, and market takeover activity.

In the ordinary-least-squares tests, we find that a substitution relation exists between pay-related incentives and the tenure of independent outside directors relative to the CEO. We find that the marginal products of pay-related incentives and the relative tenure of independent outside directors are positive and that their cross product is negative, which suggests that these variables can be substituted for one another along a negatively sloped isoquant to maintain accounting performance.

Our results continue to be robust when we use a two-stage least squares approach and when we add a control sample of nonfinancial firms to avoid selection bias. However, under two-stage least squares, we find that there is a positive or complementary relation between pay-related incentives and the percentage of independent outside directors, indicating that firms with a greater percentage of independent outside directors also have higher levels of pay-related incentives.

The different results obtained for the two board independence measures have important implications for the role of outside directors in corporate governance. Along with compensation policies, both the percentage of outside directors and the relative tenure of those directors are important considerations when it comes to aligning the financial interests of managers and shareholders. A greater proportion of independent outside directors could be needed to ensure that CEO compensation programs are properly aligned with the goals of the firm's shareholders. However, once the compensation programs are aligned through pay, inside directors who know the company and markets might be more valuable than outside directors. Furthermore, pay-related incentives appear to be especially important when the relative tenure of the firm's independent outside directors is low.

In addition to these issues, we find that inside director ownership actually reduces bank performance, and that both affiliated and independent outside director ownership increase bank performance. We find that board independence is blunted when the CEO serves as chairman of the board. When we add a separate sample of nonfinancial firms to control for selection bias, we find that pay-related incentives play a more positive role on the accounting performance in banking firms. We find no significant differences in the effect of board independence variables on accounting performance between banks and nonfinancial firms. We also find that when the CEO is the chairman of the board, this combined leadership structure (CEO duality) plays a more positive role in the performance equation of nonfinancial firms than in banks.

We want to thank participants at the 1997 and 1998 meetings of the Financial Management Association International, seminar participants at Oregon State University and Agder College, Norway, and especially two anonymous reviewers and the Editors of Financial Management, whose comments and suggestions greatly improved our paper.

(*.) Chandra S. Mishra is an Associate Professor of Finance at Oregon State University. James F. Nielsen is a Professor of Finance at Oregon State University.

(1.) Several authors have examined the extent of board effectiveness. For example, see Baysinger and Butler (1985), Brickley and James (1987), Kerr and Bettis (1987), Warner, Watts, and Wruck (1988), Weisbach (1988), Singh and Harianto (1989), Gilson (1990), Kaplan and Reishus (1990), Rosenstein and Wyatt (1990), Kosnik (1990), Hermalin and Weisbach (1991), Pearce and Zahra (1991), Byrd and Hickman (1992), Lee, Rosenstein, Rangan, and Davidson (1992), Jensen (1993), Johnson, Hoskisson, and Hitt (1993), Shivdasani (1993), Brickley, Coles, and Terry (1994), and Cotter, Shivdasani, and Zenner (1997).

(2.) We later add a control sample of nonfinancial firms to avoid any selection bias.

(3.) We note that with the passage of FIRREA in 1989, bank directors have become even more actively involved, since they are now expected to supervise as well as direct the affairs of their banks.

(4.) Many researchers find this assessment of Jensen and Murphy (1990a) too strong. Haubrich and Popova (1998) show that reasonable parameterizations of the standard agency problem result in optimal contacts with pay-performance sensitivities similar to those documented in Jensen and Murphy (1990a).

(5.) This data set was obtained from Fiegener, Nielsen, and Sisson (1996).

(6.) Bizjak, Brickley, and Coles (1993) were the first to explain the cross-sectional variation in the Jensen and Murphy (1990b) sensitivity coefficient. Lippert and Moore (1994, 1995) and McConaughy and Mishra (1996) have also used these data.

(7.) We are grateful to Kevin Murphy for providing us the compensation data to re-estimate their sensitivities over the 1982-89 period. The average pay-performance sensitivity for banks over the 1982-89 period, using the Jensen-Murphy (1990b) method, was 2.37. This average is lower than the estimate of Crawford, Ezzell, and Miles (1995), who report an estimate of 3.17 (not including CEO stockholdings). The difference between the two estimates is mostly due to the difference in the sensitivity of CEO option wealth. Using data provided by Kevin Murphy, we estimated the pay-performance sensitivity of CEO option wealth equal to 2.05. Crawford et. al., using their data, estimated the same at 2.95. There were some other differences between the Jensen-Murphy method and the method used by Crawford et al. (for example, see their Footnote 13). When we re-estimate all our regressions with the 1982-1989 pay-performance sensitivities, we find that the coefficients and signs on the board independence variables and the pay- related incentives variable are similar to those reported with the 1975-1989 data.

References

Agrawal, A. and C.R. Knober, 1996, "Firm Performance and Mechanisms to Control Agency Problems between Managers and Shareholders," Journal of Financial and Quantitative Analysis 31(3), 377-397.

Alderfer, C. P., 1986, "The Invisible Director on Corporate Boards," Harvard Business Review 64(6), 38-50.

American Law Institute, 1982, Principles of Corporate Governance and Structure: Restatement and Recommendations, Tentative Draft No. 1, Philadelphia.

Baliga, B.R., R.C. Moyer, and R.S. Rao, 1996, "CEO Duality and Firm Performance: What Is the Fuss?" Strategic Management Journal 17(1), 41-53.

Baysinger, B.D. and H. Butler, 1985, "Corporate Governance and the Board of Directors: Performance Effects of Changes in Board Composition," Journal of Law, Economics, and Organizations 1(1), 101-124.

Beatty, R.P. and E.J. Zajac, 1994, "Managerial Incentives, Monitoring, and Risk Bearing: A Study of Executive Compensation, Ownership, and Board Structure in Initial Public Offerings," Administrative Sciences Quarterly 39(2), 313-36.

Bizjak, J., J. Brickley, and J. Coles, 1993, "Stock-Based Incentive Compensation and Investment Behavior," Journal of Accounting and Economics 16(1), 349-372.

Booth, J.R. and D.N. Deli, 1996. "Factors Affecting the Number of Outside Directorships Held by CEOs," Journal of Financial Economics 40(1), 81-104.

Brickley, J.A. and C. James, 1987, "The Takeover Market, Corporate Board Composition and Ownership Structure: The Case of Banking," Journal of Law and Economics 30(1), 161-180.

Brickley, J.A., J.L. Coles, and G. Jarrell, 1997, "Leadership Structure: Separating the CEO and Chairman of the Board," Journal of Corporate Finance (3), 189-220.

Brickley, J.A., J.L. Coles, and R.L. Terry, 1994, "The Board of Directors and the Enactment of Poison Pills," Journal of Financial Economics 35(3), 371-390.

Brickley, J.A., R.C. Lease, and C.W. Smith, 1988, "Ownership Structure and Voting on Antitakeover Amendments," Journal of Financial Economics 20(1), 267-291.

Brickley, J.A., J.S. Linck, and J.L. Coles, 1999, "What Happens to CEOs After They Retire? New Evidence on Career Concerns, Horizon Problems, and CEO Incentives," Journal of Financial Economics 52(3), 341-378.

Byrd, J. and K. Hickman, 1992, "Do Outside Directors Monitor Managers? Evidence from Tender Offer Bids," Journal of Financial Economics 32(2), 195-221.

Cahen, S.F. and B.R. Wilkinson, 1999, "Board Composition and Regulatory Change: Evidence from the Enactment of New Companies Legislation in New Zealand," Financial Management 28(1), 32-42.

Cho, M., 1998, "Ownership Structure, Investment, and the Corporate Value: An Empirical Analysis," Journal of Financial Economics 47(1), 103-21.

Coles, J.W. and W.S. Hesterly, 2000, "Independence of the Chairman and Board Composition: Firm Choices and Shareholder Value," Journal of Management 26(2), 195-214.

Collins, M.C., D.W. Blackwell, and J.F. Sinkey, 1995," The Relationship Between Corporate Compensation Policies and Investment Opportunities: Empirical Evidence for Large Bank Holding Companies," Financial Management 24(3), 40-53.

Cotter, J., A. Shivdasani, and M. Zenner, 1997, "Do Independent Directors Enhance Target Shareholders Wealth During Tender Offers?" Journal of Financial Economics 43(2), 195-218.

Coughlan, A. and R. Schmidt, 1985, "Executive Compensation, Management Turnover, and Firm Performance: An Empirical Investigation," Journal of Accounting and Economics 7(1-3), 43-66.

Crawford, A., J. Ezzell, and J. Miles, 1995, "Bank CEO Pay-Performance Relations and the Effects of Deregulation," Journal of Business 68(2), 231-256.

Demsetz, H., 1983, "The Monitoring of Management," in Statement of the Business Roundtable on the American Law Institute's Proposed Principles of Corporate Governance and Structure: Restatement and Recommendation, Business Roundtable, New York.

Demsetz, H. and K. Lehn, 1985, "The Structure of Corporate Ownership: Causes and Consequences," Journal of Political Economy 93(6), 1155-1177.

Fama, E., 1980, "Agency Problems and the Theory of the Firm," Journal of Political Economy 88(2), 288-307.

Fama, E., and M.C. Jensen, 1983, "Separation of Ownership and Control," Journal of Law and Economics 26(2), 301-325.

Fiegener, M., J. Nielsen, and J.R. Sisson, 1996, "Tenure Characteristics of Outside Directors and Financial Performance: Results from the Banking Industry," American Business Review 14(1), 89-96.

Gaver, J. and K. Gayer, 1995, "Compensation Policy and the Investment Opportunity Set," Financial Management 24(1), 19-32.

Gilson, S., 1990, "Bankruptcy, Boards, Banks, and Blockholders: Evidence on Changes on Corporate Ownership and Control when Firms Default," Journal of Financial Economics 27(2), 355-387.

Goodstein, J. and W. Boeker, 1991, "Turbulence at the Top: A New Perspective on Governance Structure Changes and Strategic Change," Academy of Management Journal 34(2), 306-330.

Haubrich, J.G. and I. Popova, 1998, "Executive Compensation: A Calibration Approach," Economic Theory 12(3), 561-581.

Hermalin, B.E. and M.S. Weisbach, 1991, "The Effects of Board Composition and Direct Incentives on Firm Performance," Financial Management 20(4), 101-112.

Hoi, C.K. and J.L. Coles, 2000, "New Evidence on the Managerial Labor Market: Outside Directorships and Pennsylvania Senate Bill 13l0," Journal of Financial Economics (forthcoming).

Jensen, M.C. and K. Murphy, 1990a, "Performance Pay and Top Management Incentives," Journal of Political Economy 98(21), 225-264.

Jensen, M.C. and K. Murphy, 1990b, "A New Survey of Executive Compensation: Full Survey and Technical Appendix," Harvard Business School Working Paper No. 90-067.

Jensen, M.C., 1993, "The Modern Industrial Revolution, Exit, and the Failure of Internal Control Systems," Journal of Finance 48(3), 831-880.

Johnson, R.A., R.E. Hoskisson, and M.A. Hitt, 1993, "Board of Director Involvement in Restructuring: The Effects of Board versus Managerial Controls and Characteristics," Strategic Management Journal 23(14), 33-50.

Jordan, J.S., 2000, "Bank Managers' Opportunistic Trading of their Firm's Shares," Financial Management 28(4), 36-51.

Kaplan, S. and D. Reishus, 1990, "Outside Directorships and Corporate Performance," Journal of Financial Economics 27(2), 389-410.

Kerr, J. and R.A. Bettis, 1987, "Boards of Directors, Top Management Compensation, and Shareholder Returns," Academy of Management Journal 30(4), 645-664.

Kosnik, R.D., 1990, "Effects of Board Demography and Directors' Incentives on Corporate Greenmail Decisions," Academy of Management Journal 33(1), 129-150.

Lee, C.I., S. Rosenstein, N. Rangan, and W.N. Davidson III, 1992, "Board Composition and Shareholder Wealth: The Case of Management Buyouts," Financial Management 21(1), 58-72.

Lippert, R. and W. Moore, 1994, "Compensation Contracts of Chief Executive Officers: Determinants of Pay-Performance Sensitivity," Journal of Financial Research 17(3), 321-332.

Lippert, R. and W. Moore, 1995, "Monitoring versus Bonding: Shareholder Rights and Management Compensation," Financial Management 24(3), 54-61.

Loderer, C. and K. Martin, 1997, "Executive Stock Ownership and Performance: Tracking Faint Traces," Journal of Financial Economics 45(2), 223-255.

Lorsch, J.W. and E. MacIver, 1989, Pawns and Potentates. The Reality of America's Corporate Boards. Boston: Harvard Business School Press.

MacAvoy, P.S., S. Cantor, J. Dana, and S. Peck, 1983, "ALI Proposals for Increased Control of the Corporation by the Board of Directors," in Statement of the Business Roundtable on the American Law Institute Proposed Principles of Corporate Governance and Structure: Restatement and Recommendations, Business Roundtable (New York).

McConaughy, D.L. and C.S. Mishra, 1996, "Debt, Performance-Based Incentives, and Firm Performance," Financial Management 25(2), 37-51.

McWilliams, V.B. and N. Sen, 1997, "Board Monitoring and Antitakeover Amendments," Journal of Financial and Quantitative Analysis 32(4), 491-505.

Mehran, H., 1992, "Executive Incentive Plans, Corporate Control and Capital Structure," Journal of Financial and Quantitative Analysis 27(4), 539-560.

Murphy, K., 1985, "Corporate Performance and Managerial Remuneration: An Empirical Analysis," Journal of Accounting and Economics 7(1-3), 11-42.

Newman, H.A. and H.A. Mozes, 1999, "Does the Composition of the Compensation Committee Influence CEO Compensation Practices?" Financial Management 28(3), 41-53.

Pearce, J.A. and S. Zahra, 1991, "The Relative Power of CEOs and Boards of Directors: Associations with Corporate Performance," Strategic Management Journal 12(2), 135-153.

Pi, L. and S.G. Timme, 1993, "Corporate Control and Bank Efficiency," Journal of Banking and Finance 17(2-3), 515-530.

Pyndick, R.S. and D.L. Rubinfeld, 1991, Econometric Models & Economic Forecasts, New York: McGraw Hill.

Rosenstein, S. and J.G. Wyatt, 1990, "Outside Directors, Board Independence, and Shareholder Wealth," Journal of Financial Economics 26(2), 175-192.

Shivdasani, A., 1993, "Board Composition, Ownership Structure and Hostile Takeovers," Journal of Accounting and Economics 16(1-3), 167-198.

Singh, H. and F. Harianto, 1989, "Management-Board Relationships, Takeover Risk, and the Adoption of Golden Parachutes," Academy of Management Journal 32(1), 7-24.

Smith, C. and R. Watts, 1992, "The Investment Opportunity Set and Corporate Financing, Dividend, and Compensation Policies," Journal of Financial Economics 32(3), 263-292.

Subrahmanyam, V., N. Rangan, and S. Rosenstein, 1997, "The Role of Outside Directors in Bank Acquisitions," Financial Management 26(3), 26, 23-36.

Vance, S.C., 1983, Corporate Leadership: Boards, Directors, and Strategy, New York: McGraw Hill.

Warner, J.B., R.L. Watts, and K.H. Wruck, 1988, "Stock Prices and Top Management Changes," Journal of Financial Economics 20(1,2), 461-492.

Whidbee, D.A., 1997, "Board Composition and Control of Shareholder Voting Rights in the Banking Industry," Financial Management 26(4), 27-42.

Weisbach, M.S., 1988, "Outside Directors and CEO Turnover," Journal of Financial Economics 20(1,2), 431-460.

Descriptive Statistics for Banks

ROA is return on average total assets. ROE is return on average common equity. Size is the logarithm of total assets. Market-to-Book Ratio is the ratio of market value per share to book value per share. Pay-Related Incentives (PRI's) are Jensen and Murphy's (1990b) estimates of pay-performance sensitivities. We compute the PRI's from the Jensen and Murphy compensation equation, using 15 years of Forbes' compensation data prior to 1990. These data comprise the CEO's salary, bonus, stock options, and accounting-based performance rewards. %OUTDIR is the percentage ratio of independent outside directors to total directors sitting on the board. We define independent outside directors as board members whose only relationship with the firm involves their directorship (Byrd and Hickman, 1992). RELTENURE represents average tenure of the independent outside directors divided by the number of years the CEO has held his/her position. Ownership by Inside Directors is the percentage of equity owned by inside directors (exc luding the CEO). Ownership by Outside Affiliated Directors is the percentage of equity owned by affiliated ("gray") outside directors. Ownership by Outside Independent Directors is the percentage of equity owned by independent outside directors. CEO Duality is 1 if the CEO is also the chairman of the board, 0 otherwise. State Takeover Restrictions equal 1 for states with the least restrictive interstate branching laws, and 4 for states with the most restrictive laws (Brickley and James, 1987). CEO Tenure is the tenure as CEO in years. CEO Ownership is the percentage of equity owned by the CEO.

Ordinary Least Squares (OLS) Estimates of the Association Between Board Independence, Pay-Related Incentives, and Bank Performance

The dependent variable is ROA. We run five different specifications for each board independence variable. We report White's t-statistics in parentheses for the regression models. ROA is return on average total assets. ROE is return on average common equity. Size is the logarithm of total assets. Market-to-Book Ratio is the ratio of market value per share to book value per share. Pay-Related Incentives are Jensen and Murphy's (1990b) estimates of pay-performance sensitivities. We compute the PRIs from the Jensen and Murphy compensation equation, using 15 years of Forbes' compensation data prior to 1990. These data comprise the CEO's salary, bonus, stock options, and accounting-based performance rewards. %OUTDIR is the percentage ratio of independent outside directors to total directors sitting on the board. We define independent outside directors as board members whose only relationship with the firm involves their directorship (Byrd and Hickman, 1992). RELTENURE represents average tenure of the independent ou tside directors divided by the number of years the CEO has held his/her position. Ownership by Inside Directors is the percentage of equity owned by inside directors (excluding the CEO). Ownership by Outside Affiliated Directors is the percentage of equity owned by affiliated ("gray") outside directors. Ownership by Outside Independent Directors is the percentage of equity owned by independent outside directors. CEO Duality is 1 if the CEO is also the chairman of the board, 0 otherwise.

Two-Stage Least Squares (2SLS) Estimates of the Association Between Board Independence, Pay-Related Incentives, and Performance for Banks

The system consists of three equations: ROA, Board Independence, and PAY For each equation, we use the following variables as instruments: Size, Market-to-Book Ratio, Sales Growth, % Ownership by Inside Directors, % Ownership by Outside Affiliated Directors, % Ownership by Outside Independent Directors, % Ownership by CEO, CEO Tenure, and CEO Duality. We report t-statistics in parentheses for the regression models. ROA is return on average total assets. ROE is return on average common equity. Size is the logarithm of total assets. Market-to-Book Ratio is the ratio of market value per share to book value per share. Pay-Related Incentives are Jensen and Murphy's (1990b) estimates of pay-performance sensitivities. We compute the PRIs from the Jensen and Murphy compensation equation, using 15 years of Forbes' compensation data prior to 1990. These data comprise the CEO's salary, bonus, stock options, and accounting-based performance rewards. %OUTDIR is the percentage ratio of independent outside directors to tota l directors sitting on the board. We define independent outside directors as board members whose only relationship with the firm involves their directorship (Byrd and Hickman, 1992). RELTENURE represents average tenure of the independent outside directors divided by the number of years the CEO has held his/her position. Ownership by Inside Directors is the percentage of equity owned by inside directors (excluding the CEO). Ownership by Outside Affiliated Directors is the percentage of equity owned by affiliated ("gray") outside directors. Ownership by Outside Independent Directors is the percentage of equity owned by independent outside directors. CEO Duality is 1 if the CEO is also the chairman of the board, 0 otherwise. CEO Tenure is the tenure as CEO in years. CEO Ownership is the percentage of equity owned by the CEO.

OLS and 2SLS Estimates of the Association Between Board Independence, Pay-Related Incentives, and Performance for Banks and Nonfinancial Firms

The dependent variable is ROA. We report t-statistics in parentheses for the OLS regression models. In the 2SLS models, we use the following variables as instruments: Size, Market-to-Book Ratio, Sales Growth, % Ownership by Inside Directors, % Ownership by Outside Affiliated Directors, % Ownership by Outside Independent Directors, % Ownership by CEO, CEO Tenure, CEO Duality, and two-digit SIC dummies. ROA is return on average total assets. Size is the logarithm of total assets. Market-to-Book Ratio is the ratio of market value per share to book value per share. Pay-Related Incentives are Jensen and Murphy's (1990b) estimates of pay-performance sensitivities. We compute the PRIs from the Jensen and Murphy compensation equation, using 15 years of Forbes' compensation data prior to 1990. These data comprise the CEO's salary, bonus, stock options, and accounting-based performance rewards. %OUTDIR is the percentage ratio of independent outside directors to total directors sitting on the board. We define independen t outside directors as board members whose only relationship with the firm involves their directorship (Byrd and Hickman, 1992). RELTENURE represents average tenure of the independent outside directors divided by the number of years the CEO has held his/her position. Ownership by Inside Directors is the percentage of equity owned by inside directors (excluding the CEO). Ownership by Outside Affiliated Directors is the percentage of equity owned by affiliated ("gray") outside directors. Ownership by Outside Independent Directors is the percentage of equity owned by independent outside directors. CEO Duality is 1 if the CEO is also the chairman of the board, 0 otherwise. CEO Tenure is the tenure as CEO in years. CEO Ownership is the percentage of equity owned by the CEO. Bank Dummy is 1 if it is a bank, 0 otherwise.