Equity ownership and thrift failures during the S&L crisis.

Williams, Thomas G.E. ; Her, M. Monica

INTRODUCTION

We explore the role of equity ownership as a mechanism for protecting the interest of shareholders during the crisis within S&L industry. According to Fama (1980) equity owned by corporate insiders helps to alleviate agency problems for shareholders if it aligns the interests of insiders and outside shareholders. However, Gorton & Rosen (1995) and Stulz (1988) suggest that equity owned by insiders may also serve to entrench managers, thereby creating a shield against discipline by outsiders. Outside shareholders interest may also be protected by outside holders of large chunks of a firm's shares. These large shareholders are usually institutional investors who are more sophisticated and have a greater incentive to expend the resources necessary to monitor the managers of a firm. To the extent that shareholders in publicly traded thrifts suffered huge losses during the crisis, an interesting question arise as to whether the heterogeneity of equity ownership within the S&L industry had any impact on the survival of these institutions.

To address this question we focus on the period of turmoil within the S&L industry extending from 1983 through 1994, which represents drastic changes and that provides several events for which we can examine the influence of equity ownership on firm survival. Other factors such as fluctuations in interest rates, deregulation, and the condition of regional economies may have contributed to failures but are beyond the control of managers. However, other factors that may have contributed to the failures such as decisions to expand assets and liabilities, the selection and mix of these assets and liabilities, and the decision to leverage up the firm are well within the bounds of managerial prerogatives. As such, this period provides an exceptional opportunity to investigate the relationship between inside equity ownership and the success or failure of savings and loan associations, which may be attributed to the quality of managerial decisions.

We explore the following questions: How effective was equity ownership in resolving shareholder-manager conflicts? Was the presence of unaffiliated blockholders an effective mechanism for representing outside shareholders' interest? To provide answers to these questions, we analyzed a sample of publicly traded S&Ls that should provide opportunities to trigger the intervention by equity owners to preserve the welfare of the firm.

Our analysis reveals that independent outside directors owned less equity in failed S&Ls than they did in non-failed institutions. Similar comparisons for affiliated outside directors show that affiliated outside directors owned more equity in failed than they did in non-failed institutions. The presence of unaffiliated blockholders among the owners of sample firms appears to be related to the probability that an S&L failed. These findings suggest that the distribution of equity ownership between insiders and outside shareholders should be incorporated among the factors to be addressed before we can be confident that the S&L problems are behind us.

These findings contribute to our understanding of the relationship between equity ownership and firm performance, thereby enriching our knowledge of the dynamics within the S&L industry. The sample consists of firms that are actively traded on national exchanges and, therefore, are exposed to the full range of corporate control mechanisms. In addition, the sample is limited to one industry in order to avoid cross industry differences. However, this restriction in no way limits the value of this study, as the role of equity ownership in resolving agency problems in non-financial firms is already well documented (Himmelberg, Hubbard & Palia, 1999; Agrawal & Knoeber, 1996; McConnell & Servaes, 1990; Morck, Shleifer & Vishny, 1988; Shleifer & Vishny, 1986). Furthermore, the use of S&Ls allows us to examine the role of the equity ownership as an alternate explanation for the failures in that industry. Finally, we differentiate between three classes of directors in an industry where dealings with certain outsiders such as attorneys, accountants, investment bankers, and financiers have often been suspected to be less than arms-length. In fact, some accountants and investment bankers have been found culpable in contributing to S&L losses. For example, the nation's four largest accounting firms were the subjects of government inquiries, with each accused of professional misconduct related to S&L failures. According to The Wall Street Journal (November 24, 1992, A3), Ernst & Young paid $400 million to settle government claims and the others have settled similar claims. The former president and chief executive officer of the failed California thrift, Columbia Savings & Loan Association, was indicted for improperly obtaining stock-warrants from the investment banking firm, Drexel Burnham Lambert, Inc. The thrift would eventually be one of Drexel Burnham Lambert's best customers, holding over $4 billion in junk bonds (2).

LITERATURE REVIEW AND DEVELOPMENT OF HYPOTHESES

The separation of the ownership from the direct control of corporate resources in a firm creates the basis for the agency problems identified by Berle and Means (1932). The problems arise because managers may have incentives to augment their welfare at the expense of the firm's owners. Therefore, mechanisms have evolved that resolve these agency conflicts. Among these are the structure of managerial compensation (Baker, Jensen & Murphy, 1988; Lewellen, Loderer & Martin, 1987), the managerial labor market (Fama, 1980), managerial stock ownership, the presence of blockholders, and the takeover market (Jensen & Ruback, 1983; Martin & McConnell, 1991). However, in this paper we will focus on equity ownership as a means to resolving manager-shareholder conflicts.

Stock ownership can play an important role in resolving agency problems between managers and shareholders. However, the exact relationship between stock ownership and the alignment of stockholders' and managers' interests is unresolved. According to Baysinger and Butler (1985), Morck, Shleifer and Vishny (1988), McConnel and Servaes (1990), and Byrd and Hickman (1992), the effect of equity ownership on firm performance appears to be non-linear. These findings support the idea of an optimal distribution of equity ownership, with levels of ownership over which interests are aligned or over which managers become entrenched. Alignment of interests is consistent with the widely examined moral hazard hypothesis for insured institutions, but this study focuses on managerial entrenchment and shareholder losses, issues not given much attention in other studies that analyze problems in the S&L industry (3).

Regulators continue to have a great impact on the banking industry, as was quite evident throughout the S&L crisis of the 1980s. The shortcomings of government agencies and their complicity in the S&L crisis is extensively documented by several scholars including Strunk and Case (1988), Barth, Bartholomew and Bradley (1990), Kane (1990), Cordell, MacDonald and Wohar (1993), and Cole and Eisenbeis (1996). Regulatory intervention usually comes after the stage where internal controls such as the influence of equity ownership should have intervened to restrain managerial excesses. Furthermore, there are no regulatory edicts that should prevent equity owners from ensuring that managers pursue policies that protect their interests. The role of these equity ownership is, therefore, quite relevant to a more complete understanding of the problems in the S&L industry.

Moral hazard explanations for the S&L problems assume an alignment of interest between managers and shareholders and so most of the prior studies focuses on losses suffered by claimants other than shareholders. The notion that entrenched managers could expose outside shareholders to excessive risks has been overlooked. In this study, we extend the research that address problems experienced in the S&L industry by exploring the effect of equity ownership and what role it may have played in the crisis. Many of the earlier studies focus on causes not directly related to the internal governance of the institutions. Strunk and Case (1988, 15) list fifteen causes of thrift failure covered in the literature, none of which is related to the role of equity ownership.

The major causes attributed to S&L problems relate to regulation, deregulation, fraud, the economic environment, and supervisory bungling by the regulators. As a consequence, most of the focus is on losses incurred by the insurers who ultimately are the taxpayers. However, based on the sample used in this study, the total market capitalization of the 58 institutions that failed between 1987 and 1994 was approximately $5 billion at the end of 1985, indicating that shareholders also suffered considerable losses, even though these losses are dwarfed by the costs incurred by the insurers. Furthermore, since not all S&Ls failed, it is important to explore whether the structure of equity ownership could have protected shareholders. Thus, it may be possible to associate equity ownership with the survival outcome of S&Ls.

Mitigation of manager-shareholder agency problems through stock ownership is achieved when the level of managerial ownership is within a range where maximizing shareholders' wealth also maximizes the manager's utility. The relationship between equity ownership and firm performance is explored in several studies, including Demsetz and Lehn (1985), Stulz (1988), Morck, Shleifer and Vishny (1988), McConnell and Servaes (1990). Higher concentration of ownership should make the monitoring of managers more economical. Therefore, the shareholders of firms with concentrated ownership are more likely to actively monitor the managers. Demsetz and Lehn (1985) specify a linear relationship between firm performance and ownership concentration, but fail to detect any correlation. Brickley and James (1987) also, using concentration as a measure of ownership, find no evidence of substitution between ownership and regulation for monitoring managers in acquisition and non-acquisition states for a sample of banks. Most of the other studies that link stock ownership to firm performance use non-linear models with managerial, director, and block holder equity ownership to assess the effect of ownership on firm performance.

Stulz (1988), Morck, Shleifer and Vishny (1988), and McConnell and Servaes (1990) have shown managerial ownership structure to be effective in reducing conflicts between managers and shareholders. The relationship between firm performance and equity ownership is non-linear, but the form of the non-linearity is still unclear (4). The non-linear relationship is consistent with alignment of interest over certain ranges of ownership and entrenchment over the range where managers can reap private benefits at the expense of other shareholders.

Prior research on the effect of outside stock ownership on managerial control focuses on institutional investors and blockholders. The availability of public information for this group of investors is one reason they receive so much attention. Institutional investors are often included as blockholders because of their size. Another reason for the interest in blockholders is the strong incentive they have to actively monitor managers. The size and investment policies, in some instances, make monitoring the most cost effective means of controlling managers by these investors. According to Shleifer and Vishny (1986), unaffiliated blockholders can monitor and control managers through direct negotiation or by facilitating acquisition by outsiders. Even when internal efforts to control managers fail, external control can be made easier by the presence of a blockholder. In a study of the announcement of block trades, Barclay & Holderness (1991) associate increased stock prices with these announcements and increased turnover of top managers following the transaction. This evidence is also consistent with the hypothesis that blockholders can be active monitors, as most of these trades did not involve increased concentration of ownership, but only a change of the block holder.

In analyzing the effect of stock ownership on firm survival, the stock owned by managers is included as part of inside directors' equity, so our main concern is with insider-controlled equity and the stock ownership by outsiders represented by unaffiliated blockholders. Inside directors' and affiliated shareholders' equity bolsters the managers' standing and so serves as a shield against monitoring and control by outside shareholders. Failed institutions are, therefore, more likely to be the ones with entrenched managers, if poor managerial decisions contributed to the failures. This would be reflected by higher failure rates over a certain range of equity owned or controlled by inside directors.

Unaffiliated blockholders are potentially the single most powerful group of outside shareholders that could influence the behavior of insiders. First, the size of their holdings usually allows these investors to make or influence board appointments. Owners of large blocks of shares may also gain greater access to managers between regular shareholder meetings. Finally, holders of a large portion of the firm's shares can be a catalyst for takeovers, so managers have additional incentives to heed the concerns of these investors. The presence of unaffiliated blockholders among the investors in a firm should, therefore, be a positive force for outside shareholders, thus reducing the probability of failure.

METHODOLOGY

We employ the logistic regression technique to analyze the relationship between the probability of failure and the equity ownership variables. For the regression models the dependent variable is set equal to one if the institution failed between 1983 and 1994 and zero otherwise. The model is:

Prob(FAIL=1) = [e.sup.bx]/(1 + [e.sup.bx])

where:

b is the vector of parameters estimated and

x is the data matrix of explanatory variables that are hypothesized to explain S&L survival outcome.

The ownership variables include the equity owned by inside and affiliated outside directors to capture the effect of insider-controlled equity on S&L failure, equity owned by independent outside directors, and a binary variable to account for the presence of unaffiliated block holders in the ownership structure of the firm. The control variables include the log of total assets, a binary variable that is equal to one if the institution is located in any of five states (CA, FL, LA, OH, TX) with less restrictions on asset and liability powers of S&Ls and zero otherwise, a binary variable that is equal to one if an institution had a federal charter and equal to zero if it had a state charter, a binary variable that is equal to one if an institution had been established for five years or more in 1985 and equal to zero otherwise, the annual rate of change in the number of new housing permits issued in the 50 states, D.C., or Puerto Rico where the institution is located. Borrowing from Cebenoyan, Cooperman and Register (1995), we define the five states of California, Florida, Louisiana, Ohio, and Texas with less restrictive asset and liability powers as less regulated.

Size has been shown in several studies to impact firm performance and may be relevant since larger institutions may have less risk exposure, are subject to more public scrutiny, and regulators may be more inclined to delay the closure of larger institutions. We include a binary variable, to differentiate between states with less stringent asset and liability powers and other states. In addition to the difference in the regulation between states, savings and loan associations also differ as to their supervisory agency as the state banking departments supervise institutions with state charter and the Federal Home Loan Bank Board (FHLBB) supervise those with federal charter. Federally chartered institutions also controlled a disproportionate amount of total industry assets, which suggest that on average they are larger than state chartered institutions. One additional firm-specific characteristic, firm age, is captured with a binary variable that differentiates between established firm, that is, those that were five years and older in 1985 and those institutions that were less than five years old. To control for the effect that local economic conditions may have on the failure of an institution, we include a variable that is equal to the annual growth rate in the number of new housing permits issued in any of the 50 states, D.C., or Puerto Rico where the institution is located. The number of new housing permits issued ([N.sub.t]) in each state, D.C., and Puerto Rico is extracted from the U.S. Statistical Abstract microfiches for each year from 1982 through 1994. The annual growth rate in the number of new housing permits issued is computed as ([N.sub.t] - [N.sub.t-1])/[N.sub.t-1], for t = 1984, ..., 1994.

SAMPLE SELECTION AND DESCRIPTION OF DATA

We rely heavily on Barth, Beaver and Stinson (1991) to develop our initial sample of 165 publicly traded S&Ls. We tracked these firms from 1983 through 1995 and found that only 49 of these S&Ls survived, either in their original form or as a holding company. The remaining 116 S&Ls either failed, or were acquired by or merged into other firms. Based on information gathered from Lexis/Nexis and the Wall Street Journal Index for each of the 116 S&Ls that did not survive we classified 58 as acquired or merged and 58 as failed. Inspection of all the articles that reported acquisitions and mergers of institutions indicated these institutions were in sound financial condition at the time of the acquisition or merger. For all analysis we combined the acquired and merged institutions and then added these to the institutions that survived. Therefore, non-failed institutions include a total of 107 S&Ls. We classify as failed all institutions that became bankrupt, were taken over by regulators, or were taken over by another firm with the assistance of a government insurance or regulatory agency. None of the sample firms failed from 1983 through 1986 or during 1994. The majority of the failures occurred between 1989 and 1991. The sample was reduced because of availability of ownership data in proxy statements, financial data from the Center for Research in Security Prices (CRSP), Compustat, Compact Disclosure databases, and Moody's Banking & Finance Manual. These sources provide a total of 355 firm-years of data used in our analysis; 286 observations for 73 firms that did not fail, and 69 observations for the 37 failed institutions that remain in the data set. For the failed firms we collected data for only the three most recent years available before failure to capture characteristics of the firms, which are mostly likely to be associated with the failure.

All the directors are classified as inside, affiliated outside, or independent outside directors. Inside directors include all directors who are current and former employees of the institution, and the immediate relatives of these officers. The affiliated director classification was suggested by Baysinger and Butler (1985) and has been used in subsequent research such as Weisbach (1988), Gilson (1990), Hermalin and Weisbach (1991), Byrd and Hickman (1992), Lee, Rosenstein, Rangan and Davidson (1992), and Shivdasani (1993). Some authors used the classification of affiliated outsiders to differentiate these from independent outside directors. Even though the exact definition of affiliated outside and affiliated director differ among the studies, they all seem to capture the essence that this group of directors are somehow different from independent directors and should be analyzed separately. We include as affiliated directors officers of firms or individuals having major business relationships with the institution, financiers and financial professionals, management and financial consultants, and lawyers. All other directors including professional directors, private investors, educators, government officials, members of the clergy, and medical practitioners are classified as independent outside directors. The owners of 5 percent or more of the company's stock are classified as either affiliated or unaffiliated blockholders. We define a block holder as unaffiliated if the holder is not an inside or affiliated director, or has no substantial business relationship with the firm.

We read each proxy statement and recorded equity ownership as the percent of equity held by each category of director. We also recorded the stock holdings of block holders, the age and type of charter for each institution. For those institutions that were acquired or that failed, we include the total assets for the three most recent years prior to the year of acquisition or failure. All total assets data are adjusted by the consumer price indices (CPI) published in the U.S. Statistical Abstracts to reflect constant 1983 dollars. Due to the limitations of our data sources the number of observations for each firm are not equal for either failed or non-failed institutions. The numbers of observations range from one to three for failed institutions and from one to eleven for non-failed institutions. Also the annual data for each firm do not always represent consecutive years.

The summary of the descriptive statistics for the sample is presented in Table I. The average size of sample firms is $4.44 billion, with a range from $16.36 million to $39.03 billion in total assets. Independent outside directors held very little equity in these firms, they owned 2.46% of the equity compared to the 9.72 % that is owned by insiders. Based on evidence provided by Williams (1998), we classified as insider-controlled equity, the sum of the equity owned of inside and affiliated directors. Of the stakeholders examined in this study, block holders held the largest equity stakes with an average of 21.02%.

DISCUSSION OF RESULTS

The comparisons between failed and non-failed firms are reported in Table II. It appears the largest institutions were among the ones most likely to survive. Even though the sample is restricted to publicly traded institutions, it includes some relatively small firms. Only the equity holdings of independent outside directors, affiliated directors, and unaffiliated block holders reflected any significant difference between the failed and non-failed institutions. Outside directors held 1.54% of the equity of failed institutions compared to 2.68% for non-failed institutions. These differences are significant at better than the 1 percent level. It appears that outside directors were more effective monitors when they had a direct economic interest, represented by equity ownership in the firm. For affiliated directors the situation is reversed. Affiliated directors of failed institutions held three times the amount of equity in their institutions, as do affiliated directors of non-failed institutions, 2.99% compared to 0.98%.

The effect of inside directors' equity ownership on the failure of S&Ls appears less important than hypothesized, even though insiders appear to hold slightly larger stakes in failed institutions. This finding is not surprising since most of the studies that have linked inside equity ownership to firm performance found the relationship to be non-linear. Only when affiliated directors' equity is included as a part of insider-controlled equity does statistical significance appear in the difference between the ownership of insiders at failed and non-failed institutions. In this case, insiders' equity averaged 12.34 percent (median of 8.62 percent) at failed institutions compared to 9.09 percent (median of 5.34 percent) at non-failed institutions. Support for the equity ownership hypothesis is, therefore, provided by both insider-controlled equity (inside plus affiliated directors' equity).

Consistent with the hypothesis that unaffiliated block holders help to protect the interest of outside shareholders, we find that unaffiliated block holders held significantly less equity in failed institutions. These block holders held 24 percent less equity in failed institutions than their counterparts at non-failed institutions. The difference is significant for both means and medians.

Next, we conduct a more complex analysis using the logistic regression technique to obtain maximum likelihood estimates. The results are presented in Table III. This model relates the probability that an institution failed between 1983 and 1994 to variation in equity ownership, and a series of control variables. The levels of significance for the estimated parameters in the regressions are based on robust standard errors. These standard errors are obtained from an estimator of the variance-covariance matrix that is associated with White (1980) and Huber (1967). The version of the estimator applied here is presented in Stata and groups the observations by firms; the groups are treated as independent while the observations within each group are not independent.

Based on the findings of Morck, Shleifer and Vishny (1988) for the relationship between inside equity and firm performance, we explore the possibility of a non-linear relationship between inside equity and S&L failure. The data for our sample does not fit their specification for the relationship between insiders' equity and S&L failure. A study by Knopf and Teall (1996) also found no support for Morck, Shleifer, & Vishny's specification of the relationship between inside equity ownership and S&L performance. It should be noted, however, that Morck, Shleifer, & Vishny use total board ownership to represent insiders' equity. Similarly, Knopf and Teall (1996) define insider ownership as only the equity owned by managers. The variation in how insider ownership is defined may, in part, explain the differences in the findings. Following McConnell and Servaes (1990) and Gorton and Rosen (1995) we include a quadratic term for insider-controlled equity in the model to capture any non-linearity in equity ownership.

The evidence from this sample is consistent with earlier studies that report a non-linear relationship between insider equity and firm performance. The estimated coefficients for the inside equity variables have signs consistent with entrenchment over a certain range of ownership. The estimated coefficient for the insider-controlled equity variable is 21.5876 with a p-value of 0.000 and for the squared term is -42.089 with a p-value of 0.011. This is consistent with the prediction of the entrenchment hypothesis that the probability of failure increases with the level of insider-controlled equity and then declines. We use the estimated coefficients from Model 1 to plot the relationship between insider-controlled equity and the probability of failure at the median for all the other variables. The resulting relationship depicted in Figure 1, shows greater probability of fail when insider controlled equity is between10% and 40%.

The regressions also reveal an interesting relationship between outside directors' equity holdings and the probability of failure. Consistent with lower outside directors' equity at failed institutions, the estimated coefficient is significant and negative which is similar to Shivdasani (1993), who reports that increased ownership by unaffiliated outside directors benefited shareholders. The estimated coefficient for the binary variable representing the presence of unaffiliated blockholders is negative and consistent with the prediction that unaffiliated blockholders should reduce the probability of failure.

The overall evidence from the relationship between equity ownership and the probability of S&L failure reinforces the prior evidence that directors and managers should not be grouped together and classified as insiders. This study there shows that valuable information would be lost, if a distinction is not made between affiliated outside and independent outside directors.

[FIGURE 1 OMITTED]

The evidence from the regressions also suggests that the probability that an institution failed was correlated with the control variables. Liberal asset and liability powers could have provided managers with the opportunity to take on excessive risk, and could also have created the opening for managers to indulge in inappropriate behavior. However, the evidence indicates that such activities were not widespread. Instead, the estimated coefficient is negative and significant in both models, suggesting that the probability of failure was lower in states with less restrictive asset and liability powers. This is inconsistent with prior studies (White, 1991, 102; Esty, 1993) that associate S&L problems with increased investment in non-traditional assets and liabilities. Apparently, the opportunity to engage in non-traditional activities, by itself, was not a sufficient condition for thrifts to fail.

Larger institutions were more likely to fail if they were state chartered. The probability of failure at federally chartered institutions was also much less sensitive to differences in firm size. In interpreting the different effect of firm size between federal-chartered and state-chartered institutions it would be useful to note that like for the industry, sample federal institutions were larger than state institutions. In addition, all federal-chartered institutions were insured by the FSLIC while some state-chartered institutions were not, so any postponement of failure, particularly of larger institutions by federal regulators, would be more likely to affect federal-chartered institutions.

Consistent with the notion that age is a good proxy for firm risk, we find that institutions that were established for at least five years in 1985 were less likely to fail. These institutions were 0.01 percent less likely to fail than institutions that were established for less than five year, significant at better than the 1 percent level.

A positive growth in new housing construction is one sign of a healthy economy. This, however, does not preclude the possibility that growth in the housing sector could represent over-capacity construction. In any case, it is unlikely that over-capacity construction could systematically drive the new housing permits issued growth rate in a less than robust economic environment. The negative estimated coefficient, therefore, suggest that S&Ls were less likely to fail in a thriving economy. The new housing permits issued growth rate appears to be a good proxy for capturing the effect of the local economy on the probability of S&L failure. The role of the regional agricultural, oil and gas, and construction sectors on bank and S&L problems is well documented elsewhere, including Gunther (1989) and Barth (1991).

CONCLUSION

Our investigation of the relationship between S&L failure and the distribution of equity ownership reveals interesting results and expands the number of factors that are associated with S&L failures. We find evidence that insider-controlled equity defined as equity owned by inside directors with plus the equity owned by affiliated outside directors was related to the incidence of failure at publicly traded S&Ls. In addition, independent outside directors held less equity in failed institutions. There is also some support for the notion that unaffiliated block holders represent the interest of all outside shareholders.

Similar to Morck, Shleifer and Vishny (1988), McConnell and Servaes (1990), and Gorton and Rosen (1995), we find a non-linear relationship between insider-controlled equity and the probability of failure for S&Ls. This result is consistent with the entrenchment hypothesis and indicates that the probability of failure was greatest when insiders controlled between 20% and 35% of the outstanding equity. A number of factors beyond the control of the managers such as the type charter and age of the firm and local economic conditions also had some impact on the incidence of failure within the S&L industry.

Overall the evidence suggests that the distribution of equity among the different categories of directors was an important factor in the S&L crisis. The impact of less than arms length dealings represented by the role of affiliated directors also appeared to aggravate S&L problems. Therefore, by incorporating these factors in the analysis we can improve on the prescriptions already in place to deal with regulatory, fraud, and moral hazard problems in the S&L industry.

REFERENCES

Agrawal, A. & C. R. Knoeber (1996). Firm performance and mechanisms to control agency problems between managers and shareholders, Journal of Financial and Quantitative Analysis 31(3), 377-397.

Baker, G. P., M. C. Jensen & K. J. Murphy (1988). Compensation and incentives: practice vs. theory. Journal of Finance 43(3), 593-616.

Barclay, M. J. & C. G. Holderness (1991). Negotiated block trades and corporate control. Journal of Finance 46(3), 861-878.

Barth, J. R., P. F. Bartholomew & M. G. Bradley (1990). Determinants of thrift institution resolution costs. Journal of Finance 45(3), 731-754.

Barth, J. R. (1991). The Great Savings and Loan Debacle. Washington, D.C.: The AEI Press.

Barth, M. E., W. H. Beaver & C. H. Stinson (1991). Supplemental data and the structure of shrift share prices. Accounting Review 66(1), 56-66.

Baysinger, B. D. & H. N. Butler (1985). Corporate governance and the board of directors: performance effects of changes in board composition. Journal of Law, Economics, and Organization 1(1), 101-124.

Berle, Jr., A. A. & G. C. Means (1932). The Modern Corporation and Private Property. New York, NY: MacMillan.

Brickley, J. A. & C. M. James (1987). The takeover market, corporate board composition and ownership structure: The case of banking. Journal of Law and Economics 30(1), 161-180.

Byrd, J. W. & K. A. Hickman (1992). Do outside directors monitor managers? Journal of Financial Economics 32(2), 195-221.

Cebenoyan, A. S., E. S. Cooperman & C. A. Register (1995). Deregulation, reregulation, equity ownership structure, and S&L risk-taking. Financial Management 24(3), 63-76.

Cole, R. A. & R. A. Eisenbeis (1996). The role of principal-agent conflicts in the 1980s thrift crisis. Real Estate Economics 24(2), 195-218.

Cordell, L. R., G. D. MacDonald & M. E. Wohar (1993). Corporate ownership and the thrift crisis. Journal of Law and Economics 36 (2), 719-756.

Demsetz, H. & K. Lehn (1985). The structure of corporate ownership: Causes and consequences. Journal of Political Economy 93(6), 1155-1177.

Esty, B. C. (1993). Ownership concentration and risk-taking in the S&L industry. Working Paper, Harvard University.

Esty, B. C. (1997a). Organizational form and risk taking in the savings and loan industry. Journal of Financial Economics, 44(1), 25-55.

Esty, B. C. (1997b). A case study of organizational form and risk shifting in the savings and loan industry. Journal of Financial Economics, 44(1), 57-76.

Fama, E. (1980). Agency problems and the theory of the firm. Journal of Political Economy 88(2), 288-307.

Gilson, S. (1990). Bankruptcy, boards, banks, and blockholders: Evidence on changes in corporate ownership and control when firms default. Journal of Financial Economics 27(2), 355-387.

Gorton, G. & R. Rosen (1995). Corporate control, portfolio choice, and the decline of banking. Journal of Finance 50(5), 1377-1420.

Gunther, J. W. (1989). Texas banking conditions: Managerial versus economic factors. Federal Reserve Bank of Dallas, Financial Industry Studies, 1-18.

Hermalin, B. E. & M. S. Weisbach (1991). The effects of board composition and direct incentives on firm performance. Financial Management 20(Winter), 101-112.

Himmelberg, C. P., R. G. Hubbard & D. Palia, (1999). Understanding the determinants of managerial ownership and the link between ownership and performance, Journal of Financial Economics 53(3), 353-384.

Huber, P. J. (1967). The behavior of maximum likelihood estimates under non-standard conditions. Proceedings of the Fifth Berkeley Symposium in Mathematical Statistics and Probability, Berkeley, CA: University of California Press, 221-233.

Jensen, M. & R. Ruback (1983). The Market for corporate control: Scientific evidence. Journal of Financial Economics 11(1-4), 5-50.

Kane, E. J. (1990). Principal-agent problems in S&L salvage. Journal of Finance 45(3), 755-764.

Knopf, J. D. & J. L. Teall (1996). Risk-taking behavior in the U.S. thrift industry: ownership structure and regulatory changes. Journal of Banking and Finance 20(8), 1329-1350.

Lee, C. I., S. Rosenstein, N. Rangan & W. N. Davidson, III (1992). Board composition and shareholder wealth: The case of management buyouts. Financial Management 21(Spring), 58-72.

Lewellen, W., C. Loderer & K. Martin (1987). Executive compensation and executive incentive problems: An empirical analysis. Journal of Accounting and Economics 9(3), 287-310.

Martin, K. J. & J. J. McConnell (1991). Corporate performance, corporate takeovers, and management turnover. Journal of Finance 46(2), 671-687.

McConnell, J. J. & H. Servaes (1990). Additional evidence on equity ownership and corporate value. Journal of Financial Economics 27(2), 595-612.

Morck, R., A. Shleifer & R. W. Vishny (1988). Management ownership and market valuation: An empirical analysis. Journal of Financial Economics 20(1-2), 293-315.

Shivdasani, A. (1993). Board composition, ownership structure, and hostile takeovers. Journal of Accounting and Economics 16(1-3), 167-198.

Shleifer, A. & R. W. Vishny (1986). Large shareholders and corporate control. Journal of Political Economy 94(3), 461-488.

Strunk, N.& F. Case (1988). Where Deregulation Went Wrong: Look at the Causes Behind Savings and Loan Failures in the 1980s, Chicago, IL: United States League of Savings Institutions.

Stulz, R. (1988). Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics 20(1-2), 25-54.

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

White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48, 817-830.

White, L. J. (1991). The S&L Debacle. New York, NY: Oxford University Press.

Williams, T. G. E. (1998). Corporate control and failure at publicly traded savings and loan associations. Unpublished doctoral dissertation, Texas A&M University.

ENDNOTES

(1) Some of this work was completed while the authors were Ph.D. students at Texas A&M University.

(2) More than two years later The Wall Street Journal (March 11, 1994, p. A2) reported that Thomas Spiegel, the former executive of Columbia Savings, was acquitted of all 55 counts in the 1992 indictment.

(3) Evidence that could be used to support the moral hazard hypothesis is represented by the greater risk taking observed for stock compared to mutual owned institutions reported by Cordell, MacDonald, & Wohar (1993), and Esty (1997a). Esty (1997b) provides an extreme example of the alignment of shareholder and managerial interests in a case study of Twin City, a S&L where the directors and CEO owned 100 percent of the equity of the firm.

(4) Morck, Shleifer, & Vishny (1988) using a linear specification on their data set detected no relationship between ownership structure and firm performance, which is consistent with Demsetz & Lehn (1985). Rather than specify the form of the non-linearity Gorton & Rosen (1995) used a semi-parametric technique and let the data determine the form that the ownership variable enters the model. The results provide evidence consistent with entrenchment that is similar to the other studies that use the parametric specification. Their sample was also robust to the parametric specification.

Thomas G. E. Williams, William Paterson University

M. Monica Her, California State University, Northridge

Table I: Summary Statistics for Sample Data

Summary of total assets and ownership structure data for sample
of publicly traded savings and loan institutions during the
period 1983 through 1994. A total of 355 firm-years of data is
used for all computation except for unaffiliated block
holders with 269 firm-years.

 Mean Std Dev

Total assets $4,438.94 $7,499.53
Outside directors' equity 2.46% 2.64%
Inside directors' equity 8.35% 11.63%
Affiliated directors, equity 1.37% 3.00%
Insider-controlled equity 9.72% 12.15%
Unaffiliated block holders equity 21.02% 12.52%

 Mean Std Dev

Total assets $4,438.94 $7,499. 53
Outside directors' equity 2.46% 2.64%
Inside directors' equity 8.35% 11.63%
Affiliated directors, equity 1.37% 3.00%
Insider-controlled equity 9.72% 12.15%
Unaffiliated block holders equity 21.02% 12.52%

 Minimum

Total assets $16.36
Outside directors' equity 0
Inside directors' equity 0
Affiliated directors, equity 0
Insider-controlled equity 0.05%
Unaffiliated block holders equity 5.07%

Table II: Directors and Block Holders Equity Ownership

Ownership structure variables for sample of savings and loan
associations with proxy statements available on Q Data
Corporation microfiche for the years 1983 through 1994. The
statistics for failed institutions were computed from data
for the three most recent years prior to the year of failure
for which data was available. For non-failed institutions the
statistics include all available data for the full sample period.
We report the sample means and medians (in parentheses). N is the
number of firm-years of data used to compute the corresponding
statistics. The level of statistical significance are based on
difference of means t-test for the means and Wilcoxon rank sum
Z-test for medians, and compares the data column 2 to those column 3.

Description Non-failed Failed
 institutions institutions

 N = 286 N = 69
Total assets * $4,764.75 $3,088.50 (c)
 ($1,183.24) ($1,424.53)
Outside directors' equity 2.68% 1.54% (a)
 (1.63%) (0.82% (a))
Affiliated directors' equity 0.98% 2.99% (a)
 (0.11%) (0.80%)
Inside directors equity 8.11% 9.35%
 (4.60%) (5.43%)
Insider-controlled equity 9.09% 12.34% (b)
 (5.34%) (8.62% (a))
Unaffiliated block 16.72% 12.64% (b)
holders' equity (14.87%) (7.60% (a))

* Millions
(a) Significant at the 1% level
(b) Significant at the 5% level
(c) Significant at the 10% level.

Table III: Logistic Regressions

Logistic regressions with inside and affiliated directors' equity
combined as insider-controlled equity. The dependent variable is
equal to 1 if the institution failed between 1983 and 1994, and
zero otherwise. The explanatory variables include equity ownership
variables, asset and liability powers, firm size, type of charter,
firm age, and the effect of local economic conditions. The dollar
figures for total assets are adjusted by the consumer price indices
to reflect constant 1983 dollars. The local economic condition
variable is the percentage change in the number of new housing
permits issued or unemployment rate in all 50 states, D.C., and
Puerto Rico for each year covered by the sample period. The p-values
are in parentheses and are estimated using standard errors that are
computed assuming the observations are independent across firms but
not between years for each firm.

Variables Model 1 Model 2

Intercept -1.9048 -16.6844 (c)
 (0.229) (0.052)
Insider-controlled equity 21.5786 (a) 21.1715 (a)
 (0.000) (0.001)
Insider-controlled equity -42.089 (b) -43.2183 (a)
squared
 (0.011) (0.010)
Outside directors' equity -42.157 (a) -42.7555 (a)
 -0.003 (0.001)
Binary variable = 1 if the -0.8764 (b) -0.7245 (c)
firm has unaffiliated block
holders and = 0 otherwise
 (0.033) (0.080)
Binary variable = 1 for -1.2163 (a) -0.8513 (c)
firms located in any of 5
states (CA, FL, LA, OH, TX)
and = 0 otherwise
 (0.007) (0.070)
Logarithm of total assets 0.2554 2.3953 (b)
 (0.163) (0.046)
Binary variable = 1 if the 2.7223 (a) 19.950 (b)
institution is federally
chartered and 0 if state
chartered
 (0.000) (0.034)
Binary variable = 1 if the -4.1078 (a) -5.3198 (a)
institution has been
established for 5 or more
years and = 0 if less than
5 years
 (0.000) (0.000)
Interaction term between -2.3243 (c)
log of total assets and
type of charter
 (0.057)
Annual rate of change in -4.5149 (a) -4.4923
the number of new housing
permits issued in the home
state of each S&L
 (0.000) (0.000)
Number of firm-years of data 327 327
Likelihood Ratio -101.14 -97.26
([chi square])
 (0.000) (0.000)
Pseudo [R.sup.2] 0.2939 0.3210

(a) Statistically significant at the 1% level.

(b) Statistically significant at the 5% level.

(c) Statistically significant at the 10 % level.