An ecology of agency arrangements: mortality of savings and loan associations, 1960-1987
Hayagreeva Rao
Press.
Hannan, Michael T., and John Freeman 1977 “The population ecology of organizations.” American Journal of Sociology, 82: 929-964.
1989 Organizational Ecology. Cambridge, MA: Belknap Press of Harvard University.
Haveman, Heather 1992 “Between a rock and a hard place: Organizational change and performance under conditions of fundamental environmental Economic activity is characterized by a multitude of agency relationships: Individuals and organizations known as principals delegate resources and tasks to other individuals and organizations known as agents, since they themselves lack time and expertise (Jensen and Meckling, 1976; Shapiro, 1987), as for example, when individuals delegate to their accountants the task of filing tax returns. Agency arrangements enhance role specialization in a society, connect actors across group boundaries and physical distances, and improve collective action by principals (Luhman, 1979; Zucker, 1986).
Agency relationships may also be collectivized: Anonymous principals can reduce their risks by banding together and entrusting authority to agents who accomplish the tasks delegated to them (Mitnick, 1984). For example, stockholders of a joint-stock company collectively entrust their capital to be deployed efficiently by the managers of the company. Other examples of collectivized agencies include commercial banks, savings and loan associations, mutual savings banks, credit unions, and mutual funds. To assure themselves of reliable performance by agents, principals use diverse ownership arrangements, monitoring systems, and incentives; thus, collectivized agency relationships can take on a wide variety of organizational forms (Jensen and Meckling, 1976; Fama, 1980). In this paper we ask whether organizations with different ways of organizing collectivized agency relationships encounter different survival prospects.
The paper is motivated by three considerations. First, studies of the differential survival of collectivized agency relationships extend the ecological perspective on organizational diversity. Ecological theory frames organizational change as a population-level process consisting of the replacement of existing organizations by new organizations and depicts selection as the mechanism that regulates organizational diversity (Hannan and Freeman, 1977, 1989; Aldrich, 1979; McKelvey, 1982). Despite a vigorously expanding body of research, ecological research on diversity has been criticized for emphasizing issues of specialism-generalism and for failing to examine differences in the survival of different ownership structures (Aldrich and Marsden, 1988: 58; Meyer and Zucker, 1989: 71). Two studies, by Barnett and Carroll (1987)and Barnett (1990), examined the survival of mutual and commercial telephone companies; however, they ascribed the survival of these ownership arrangements to their specialized roles in a technical system. By contrast, we seek to trace the survival of different ownership structures to differences in their governance systems and capital structures.
Second, attempts to connect the survival of agency arrangements to differences in their governance arrangements and capital structures provide an opportunity to link ecological theory with organizational economics. Transaction cost economics, property rights theory, and agency theory also hold that selection processes shape the survival of organizational forms. However, all three perspectives emphasize efficient monitoring and incentives as central to the survival of organizations (Alchian and Demsetz, 1972; Fama and Jenson, 1983a; Williamson, 1975). By contrast, ecological theory holds that although efficiency issues affect organizational change, their impact is constrained by institutional processes such as legitimacy (Hannan and Carroll. 1992). Nonetheless, studies of how population-level change is jointly shaped by efficiency considerations and institutional processes are absent in the literature and sorely needed (Hannan and Freeman, 1989: 339).
Finally, connecting ecological and agency perspectives on organizations is useful because it directs attention to the hierarchical nature of the selection process. To the extent that agency theory focuses on monitoring arrangements in organizations, it emphasizes the selection of strategies and projects within organizations. Thus, agency theory holds that the selection of organizations depends on the efficacy of selection processes within organizations. Further, since selection mechanisms such as accounting and budgeting rules underlie variations in monitoring arrangements, studies of the dissolution of monitoring arrangements require a discussion of micro-level processes even as they draw on larger institutional processes (Zald, 1986: 327-329). Thus, conjoining the agency and ecological perspectives focuses attention on the linkage between micro-level and macro-level selection processes.
Agency Costs and the Survival of Organizations
Agency models treat organizations as sites of contracts between principals (the owners of the factors of production) and agents (those who manage these factors to create goods and services for consumers). They suggest that the most important contracts define the nature of residual claims by principals and allocate the rights to initiate, ratify, monitor, and implement proposals among principals and agents. Because contracts are costly to write, monitor, and enforce among agents with divergent interests, agency costs significantly affect the survival of organizations. The contract structure of an organization in conjunction with the production technology and legal constraints determine the costs at which an organization can meet the needs of consumers. The lower the agency costs of an organization, the more efficient it is and the more likely it is to survive (Fama and Jensen, 1983a).
Agency costs are especially significant in collectivized agency relationships, which are characterized by the separation of ownership from control. This is especially so in joint stock corporations, savings and loan associations, mutual banks, mutual funds, and nonprofit organizations. Since managers as agents have specialized skills and knowledge, they initiate and implement proposals and are compensated for their performance. Principals or owners bear residual risks and enjoy access to the residual left after all payments to agents. They control decisions by ratifying initiatives and monitoring their implementation by managers. However, when many principals exist, residual claims become diffuse, and it is costly for all claimants to be involved in decision control; hence, intermediaries, such as boards of directors, exist to facilitate decision control. As a result, control over decisions is further separated from the people whose resources are at risk, and problems between principals and agents proliferate.
Because agents tend to have more information than principals, they have opportunities to misuse, embezzle, and divert the latter’s resources. Principals defuse these hazards by limiting participation in agency relations, internalizing expertise, eliminating agents, spreading their risks, and writing contracts. Moreover, principals seek to control agents by constructing monitoring arrangements and incentive mechanisms to elicit fidelity and cooperation (Fama and Jensen, 1983a, 1983b; Jensen, 1983; Shapiro, 1987: 629-633). More importantly, the characteristics of ownership rights determine whether principals are able to monitor and motivate managers, mitigate conflicts of interest between managers and other stakeholders, and reduce agency costs (Fama and Jensen, 1983c).
When there is a separation of ownership and control, the survival of organizations is determined by the ability of monitoring arrangements and incentive mechanisms to discipline managers and to reduce conflicts of interest between managers and other constituencies. The more efficient the structure of monitoring and incentives, the lower the agency costs of an organization, the greater its comparative advantage over other organizations and the higher its survival prospects when subjected to competition. Conversely, the higher the agency costs of an organizational form, the lower its survival prospects.
When extended to collectivized agencies, the agency perspective predicts that the more efficient the monitoring and incentive systems in a collectivized agency, the lower its agency costs and the higher its survival prospects when exposed to competition. Thus, agency arrangements are selected on the basis of their efficiency, and selection regulates diversity in agency arrangements in a given organizational field (Fama and Jensen, 1983c).
The Survival of Collectivized Agencies
Ecological theory uses the concept of a niche to detail how competition shapes the survival of organizations and regulates organizational diversity (Hannan and Freeman, 1977, 1989; Hannan and Carroll, 1992). In the ecological framework, the fundamental niche of a collectivized agency relationship can be said to consist of the social, economic, and political conditions necessary to sustain the agency arrangement in question. If two forms of collectivized agencies rely on similar resources and institutions for support, then their fundamental niches intersect, and the scope for potential competition is directly proportional to the overlap of their fundamental niches. In turn, niche overlap shrinks the resource space and lowers the carrying capacity available to a given population of collectivized agencies.
Ecological researchers conceptualize interpopulation competition and intrapopulation competition by building on the idea that the addition of a competitor in the environment lowers the capacity of the environment to sustain organizations. When numbers are small and resources are relatively abundant, then the likelihood and intensity of competition diminish. When the density of organizations increases, competition is intensified. Rising density increases direct and diffuse competition. In contrast to direct competition, which obtains when pairs of organizations engage in head-to-head competition, diffuse competition is indirect and obtains among anonymous organizations. If direct competition is strategic and conscious, diffuse competition is passive and even unconscious. As the number of organizations grows, organizations cannot avoid direct head-to-head competition with other organizations. Since there are |N.sup.2~/2 – N/2 possible pairwise competitive interactions in a population of N organizations, it becomes difficult for any organization to devise and sustain a coherent strategy. Rising density also multiplies diffuse competition at an increasing rate because variations in the higher ranges of density have a greater effect than changes in the lower ranges. The potential for competition increases geometrically as density increases linearly (Hannan and Carroll, 1992). Because intense competition exhausts the supplies of patrons, resources, members, and trained organization builders, it increases the mortality of organizations and reduces the formation of new organizations. Competition thus constrains organizational diversity and heightens structural homogeneity (Hannan and Freeman, 1989).
Ecological models of organizations also hold that environmental variability, or the variance of a series around its mean, constrains organizational diversity by complicating the problem of reliable performance (Freeman and Hannan, 1983). Organizations need excess resources to sustain reliable performances in the face of environmental variability. When excess resources are absorbed in the form of excess costs and perquisites, it is hard to recover them, because any attempt to recover them entails a disturbance of organizational routines (Hannan and Freeman, 1989: 106). Routines are also truces among contending interest groups in an organization, and any attempt to alter them intensifies conflict within organizations (Nelson and Winter, 1982). By contrast, unabsorbed excess resources enable an organization to respond easily to changing conditions because they are uncommitted to specific uses.
Whether excess resources are absorbed or unabsorbed depends in part on the structure of incentives and monitoring in organizations. When there is negligible monitoring and managers have minimal incentives to act as entrepreneurs but considerable incentives to extract resources in the form of indirect benefits such as perquisites, the absorption of slack increases, impeding reliable performance in the face of environmental variability. When the structure of incentives and monitoring encourages the absorption of slack resources, organizations are more likely to succumb to environmental variability, and organizational diversity is more likely to be diminished. Conversely, when the structure of incentives and monitoring discourage the absorption of slack resources, organizations can use excess slack to sustain reliable performance when faced with environmental variability.
In this paper we examine whether differences in ownership rights and capital structure translate into differences in the survival of collectivized agencies when they are exposed to competition and environmental variability. Specifically, we discuss how two forms of savings and loan associations, mutuals and stocks, differ in their ownership arrangements and capital structures. We then investigate whether these two forms encounter different survival prospects when they are faced with competition and environmental variability. The savings and loan industry is an attractive research site because the structural differences between mutual and stock savings and loan associations are institutionalized in the chartering procedures of federal and state-level regulators. Additionally there are a growing number of failures in the industry, which threatens to undermine confidence in these financial systems (Brumbaugh and Carron, 1988).
ORGANIZATIONAL FORMS IN THE SAVINGS AND LOAN INDUSTRY
Savings and loan associations (SLAs) are specialized financial intermediaries whose assets consist mainly of loans for home building and whose liabilities comprise funds deposited by consumers. SLAs could be chartered by regulatory agencies in individual states or by the Federal Home Loan Bank Board (FHLBB) up to 1988, and, from 1989 onward, by the Office of Thrift Supervision. SLAs may be insured by insurance funds in individual states or by the Federal Savings and Loan Insurance Corporation (FSLIC) up to 1988, and, from 1989 onward, by the Savings Association Insurance Fund, which is part of the Federal Deposit Insurance Corporation. SLAs can also be structured as mutual or stock companies. These two organizational forms differ in their ownership rights and the allocation of control between managers and owners.
Traditionally, SLAs were organized as mutuals. They originated as “friendly societies” built by networks of friends and acquaintances as a mechanism to borrow money from acquaintances to build houses and as an alternative to borrowing money from strangers (Rasmusen, 1988:415). Hence, the mutual was embedded in a community in that its formal structure was based on a substructure of close interpersonal relationships. As a result, the mutual enjoyed access to resources and was characterized by informal and low-cost monitoring (Bodfish, 1931). As the community gave way to a society of strangers, the formal structure of the mutual, which reflected its cooperative origins, persisted even though the social surround was transformed. The result was that informal monitoring decayed and the opportunities for malfeasance increased. The mutual acquired fieflike characteristics, especially after the Depression, when regulators encouraged attorneys, insurance specialists, accountants, and real-estate professionals to form mutuals with a view toward stimulating the financing of home building (Barth and Regalia, 1988).
The owners of the modern mutual are its depositors, but they enjoy limited ownership rights. During 1966-1982, cash and asset distributions to depositors were legally precluded by FHLBB interest-ceiling regulations. Depositors have nontransferable claims to the current earnings of the association, and the earnings are retained rather than paid out as dividends to the depositors (Masulis, 1987: 31). Moreover, depositors also have weak voting rights: Each depositor is entitled to one vote for every $100, up to a maximum of 50 votes (O’Hara, 1981: 318). Hence, the ownership of the mutual is diffuse. Additionally, the board of directors controlling the mutual has perpetual proxies signed by a preponderance of depositors. In a study of proxies in 1965, Herman (1969) reported that 55.5 percent were made out to the incumbent management, 20 percent were made out to a single individual, and only 10 percent were given to the board of directors. Although the proxies are revocable, the voting rights of depositors are attenuated by limits on outside nominations to the board, restricted disclosure requirements, and the ability of managers to redeem a member’s savings account and eliminate voting rights (Masulis, 1987: 31). Additionally, depositors have minimal incentives to monitor the mutual because their deposits are insured by the FSLIC for up to $100,000, and they are also paid the prevailing interest rate.
The managers of mutuals enjoy enormous control over the mutual because no individual can concentrate ownership of the mutual by purchasing shares. Moreover, the manager’s autonomy is not constrained by a possible proxy fight, because uncooperative depositors can be expelled by the managers (Rasmusen, 1988: 398). Although mutuals are legally required to hold meetings to elect managers and directors, the meetings are poorly attended and ritualistic. Nichols (1972: 75) reported that according to the president of a San Diego mutual, no one had turned up for the annual meetings in 30 years. Herman (1969: 654) observed that only three depositors of a large California mutual had attended the annual meeting in 13 years and that only one depositor had attended the annual meeting in over 20 years at another mutual. The only way depositors can discipline the managers of the mutual is to redeem their claims and withdraw their deposits. Such redeemable claims are a high-cost disciplinary mechanism, however, because the assets of mutuals also include nonfinancial assets that can only be traded at high transaction costs.
Although the managers of mutuals enjoy considerable autonomy, they have little incentive to act as entrepreneurs, mostly because of the way mutuals are organized. The organizers of mutuals do not enjoy legally recognized rights in the mutual. Moreover, the FSLIC rules state that the organizers must contribute 20 percent of withdrawable savings or $250,000, whichever is less. In return, the organizers determine the initial composition of the board of directors, which is likely to persist, because the organizers can eject troublesome depositors by returning their deposits (Masulis, 1987: 32). While the salaries of mutual managers are limited by regulators, they have opportunities and incentives to extract resources in the form of perquisites, favorable loans extended to relatives and friends, and the purchase of mortgage-related services at noncompetitive prices (O’Hara, 1981 ). Some studies note that mutuals suffer not only from higher administrative expenses than stock companies (Verbrugge and Jahera, 1981; Woerheide, 1984) but also from higher pension liabilities on behalf of executives (O’Hara, 1981 ). Nichols (1972: 1-4) portrayed the mutual as a “self-perpetuating aristocracy” and observed that the management is as “external to the firm as any supplier. The mutual is a quasi-firm because it lacks the objectives of the firm, i.e., profits, and has objectives not internal to the firm.”
In addition, the mutual form of a savings and loan organization also suffers from a financial weakness flowing from the structure of ownership rights. It cannot issue stock to augment its financial resources. A mutual’s current earnings or losses thus directly affect net worth; specifically, if operating losses are incurred, the mutual’s financial plight worsens unless it is able to dispose of those of its assets that have a high market value in comparison with their book value and thereby retire its debt. This is difficult for most mutuals because their assets are mortgages booked at face value but earning below-market interest rates (Masulis, 1987: 33).
By contrast, stock SLAs emerged as a means of protecting the interests of strangers acting in concert for economic benefits. Unlike the mutual, the stock SLA does not merge the roles of owners and customers but differentiates between stockholders and depositors. The managers of stock SLAs may or may not own shares, but, in either case, they are beholden to shareholders for a return on their investment and are offered financial incentives for meeting profit objectives. Therefore, managers have significant incentives to keep costs to a minimum. Moreover, the stock form promotes efficient risk bearing and provides depositors with a safety net, because shareholders rather than depositors bear the risk of failure. Stock SLAs are owned by their stockholders, whose residual claims are freely alienable, since they can sell their stock at any time they desire. External monitoring thus exists in the form of a stock market for pricing and transferring stock at low cost. Another external monitor of the stock SLA is the market for corporate control: Because the residual claims are alienable, incumbent managers can be displaced by other outside managers through a proxy fight or a tender offer. The boards of directors of stock SLAs also serve as a means of recruiting, monitoring, and rewarding managers according to their performance. Outside members of boards act as arbiters between board members who are internal managers and stockholders, and they thereby ensure that the discretion of internal managers can be curtailed. Finally, potential problems between managers and stockholders are further reduced when managers also own stock (Fama and Jensen, 1983c; Masulis, 1987). These governance arrangements induce executives of stocks to incur lesser operating costs than mutuals (Verbrugge and Jahera, 1981) and to take prudent risks and increase profitability (Verbrugge and Goldstein, 1981).
The preceding comparison of mutual and stock SLAs suggests that stocks enjoy agency cost advantages over mutuals. Superior monitoring devices and stronger definition of ownership rights exist in stocks. Moreover, mutuals are impaired by their inability to issue stock. A contrary argument, proposed by Mayer and Smith (1986), is that mutuals enjoy an agency advantage over stocks because they reduce the conflicts between risk-averse depositors and risk-prone owners by combining the roles of depositor and owner. Moreover, the managers of mutuals have little incentive to take risks because they can jeopardize their perquisites if depositors withdraw from the mutual. By contrast, executives of stock SLAs, which are often closely held (Kane, 1989: 58), have incentives to pursue risky investments because of the opportunity for the stockholders and themselves to profit. As Masulis (1987: 32) observed, however, this advantage of mutuals has been negated by deposit insurance. After the advent of deposit insurance in 1934, the incentives for managers of mutuals to take risks increased because the depositors were protected. One can argue, therefore, that the incentives for managers of mutuals to take risks when faced with failure or competition are as strong, if not stronger than the incentives for managers of stocks.
Survival Prospects of Mutuals and Stocks
The preceding assessment of the differences between mutual and stock SLAs invites the following general argument about the survival prospects of the two forms. Since mutual SLAs suffer from inferior monitoring capabilities and the inability to issue common stock, they are likely to have higher agency costs than stock SLAs and experience higher mortality when exposed to competition and environmental variability.
Competition. SLAs take short-term deposits and make long-term mortgage loans. In doing so, they not only compete among themselves but also with other financial intermediaries whose niches intersect the resource space of SLAs. In particular, commercial banks, life insurance companies, mutual savings banks, and other financial intermediaries, such as money-market mutual funds and mortgage banks, compete with SLAs for securing deposits or originating and servicing mortgages. As the density of any of these organizational forms increases, competition is likely to intensify and resources become exhausted. As a result, mutuals are likely to suffer significantly higher mortality than stocks.
Competition among SLAs can be localized to a state or county or also be nonlocal, to the extent that an SLA based in one state may compete for deposits with financial intermediaries in other states. During the postwar years up to 1963, many SLAs on the West Coast attracted deposits from eastern markets through small brokers (Strunk and Case, 1989: 81). This practice was curtailed when the FHLBB limited an association’s brokered funds to 5 percent of its total deposits. In 1982 the FHLBB withdrew restrictions on brokered funds, however, and this led to an increase in nonlocal competition.
SLAs compete with commercial banks for savings deposits and residential mortgage loans. In 1985 commercial banks held 30 percent of all financial assets in the U.S. By contrast, SLAs accounted for 16 percent of all assets (Brumbaugh, 1988: 13). Although SLAs were allowed to pay higher interest rates than banks to secure funds, this interest rate differential (Regulation Q) was rescinded in 1984. This, in turn, intensified competition between SLAs and commercial banks. Commercial banks accounted for 14.1 percent of nonfarm residential mortgage holdings, compared with the 33.4 percent held by SLAs in 1985 (Brumbaugh, 1988: 164). Competition with commercial banks is especially problematic because the number of banks is far larger than the number of SLAs: More than 13,000 banks exist, as compared with 3,000 SLAs (Balderston, 1985). Diffuse competition between SLAs and commercial banks can be localized to a state or Standard Metropolitan Statistical Area (SMSA). Although there are restrictions against branch banking under the McFadden Act, nonlocal competition is possible because an SLA based in one state may compete with commercial banks based in other states for brokered deposits. Moreover, commercial banks have also found ways to circumvent branch banking restrictions by creating nonbank corporations and employing communications technology.
Life insurance companies also compete with SLAs for private savings and for residential mortgages. In 1985 life insurance companies accounted for 12 percent of all financial assets and 1.9 percent of nonfarm residential holdings (Brumbaugh, 1988: 13, 164). Competition between SLAs and life insurance companies can be confined to a state or county or extend beyond, because life insurers based in other states may also compete in a given geographic area for funds and mortgage business.
Mutual savings banks compete with SLAs to garner deposits and residential mortgages in the northeastern region. They are similar in structure to mutual savings and loan associations, but the depositors lack even the right to vote (Rasmusen, 1988). Mutual savings banks also depend on deposits but extend consumer loans as well as mortgage loans; their savings accounts are flexible and they serve lower-income groups. As of 1985 mutual savings banks held 5.4 percent of all nonfarm mortgages, but their share of financial assets declined from 8 percent in 1955 to 3 percent by 1985 (Brumbaugh, 1988: 13, 164).
In addition to these traditional sources of competition, SLAs have had to compete with money-market mutual funds and mortgage bankers, private organizations specializing in housing finance from 1979-1980 onward. Money-market mutual funds pool the resources of investors in Treasury bills, short-term debt of high-quality corporations, and large-denomination short-term certificates of deposit offered by banks and thrifts that were free of the limitations imposed by Regulation Q. Despite being uninsured, money-market mutual funds had high-quality assets and offered depositors restricted check-writing privileges. Although they had not existed until 1972, money-market funds controlled $9.5 billion in assets in 1978, but their cumulative assets exploded from $42.5 billion in 1979 to $310.7 billion in 1987 (White, 1991:68-69). Finally, private mortgage bankers also began to make incursions into the traditional domain of thrifts by exploiting the securitization of mortgages. In particular, the development and substantial growth of collaterized mortgage obligations, beginning in 1983, has led to an increasing prominent role for mortgage bankers in the origination and servicing of mortgages (Brumbaugh, 1988: 169). Given mutuals’ inferior monitoring capabilities and their inability to issue common stock, the effects of competition should be greater for them than for stocks:
Hypothesis 1: Competition has stronger positive effects on the mortality of mutuals than of stocks.
Environmental variability. Environmental variability complicates the problem of reliable performance for mutuals and stocks. Both types of organizations require excess resources to be able to sustain reliable performances when faced with altered circumstances. Whether organizations are able to use excess resources depends on whether they are absorbed in other uses or unabsorbed (Williamson, 1963). If excess resources are absorbed, it is difficult to recover them and, as a result, organizations become vulnerable to environmental variability. By contrast, if excess resources are unabsorbed, organizations are able to access them easily and project reliable performances even when faced with altered circumstances. Mutuals are likely to be more vulnerable than stocks to environmental variability because their excess resources are often absorbed. Since the managers of mutuals have little incentive to be entrepreneurial and have little scope to benefit from the growth of the organization, they are liable to extract resources in the form of perquisites and loans to friends (O’Hara, 1981; Masulis, 1987). By contrast, the monitoring and incentive systems in stocks are more likely to lead to unabsorbed excess resources because managers are rewarded for performance and also because many stocks, as Kane (1989: 58) noted, are closely owned. Given these differences, we hypothesize:
Hypothesis 2: Environmental variability has stronger positive effects on the mortality of mutuals than of stocks.
Factors Reducing Structural Differences between Mutuals and Stocks
The hypotheses above presume that structural differences exist between mutuals and stocks. Although transaction and technological differences between mutuals and stocks are endorsed and safeguarded by regulatory agencies, the boundaries between mutual and stock SLAs can be made less distinct by blending processes. Hannan and Freeman (1989) discussed several blending processes that erase the differences between organizational forms. In the SLA population, two blending processes appear to be important: coercive isomorphism and deregulation.
In coercive isomorphism, a powerful external agency imposes organizational structure and reduces structural variety among potentially divergent forms (DiMaggio and Powell, 1983). Coercive isomorphism is likely in the SLA population because all SLAs are chartered and regulated by either the FHLBB (reconstituted in 1989 as the Office of Thrift Supervision) or by state-level regulatory agencies. The potential therefore exists for regulatory agencies to require both mutuals and stocks to adopt practices that compensate de facto for any differences in the viability of their monitoring arrangements and capital structures. Thus, the disbanding rate of SLAs could be a function of the agency that regulates them rather than their internal organizational structure. Moreover, regulatory agencies, as guardians of trust, differ in their ability to monitor organizations and prevent failures (Mitnick, 1984; Shapiro, 1987).
Federal and state-level regulators not only differ in their administrative capacity but also vary in the asset powers they grant to SLAs. State regulators and federal regulators compete in a market for charters, and state regulators are more likely to be lax in their regulation and confer broader asset powers to all SLAs (Balderston, 1985; Strunk and Case, 1989). In a study of 205 insolvencies, Barth, Bartholomew, and Labich (1989) found that a disproportionate number of failed thrifts consisted of state-chartered SLAs and that initial investigations also uncovered significant contraventions of regulations and malfeasance. Therefore, the following hypothesis needs to be examined:
Hypothesis 3: State-chartered SLAs will encounter higher mortality than federally chartered SLAs.
Another blending process was the deregulation of the SLA industry from 1980 onward, which dramatically altered the asset powers, opportunities, and incentives of SLA managers to take risks. Prior to 1980, SLAs had restricted asset powers, limited opportunities to secure resources because of Regulation Q, which enabled them to pay a higher rate of interest than banks, and limited incentives to take risks because of stable interest rates and the absence of competitors such as mortgage bankers and money-market mutual funds. New SLAs were founded only in unprovided areas (Strunk and Case, 1989). The regulatory edifice began to crumble slowly, beginning in 1974, with the chartering of stock SLAs by federal regulators, the growth of a secondary market for mortgages, the diffusion of Negotiable Order of Withdrawal (NOW) accounts, and the erosion of the tax advantages enjoyed by SLAs (Woerheide, 1984). Moreover, interest rates were relaxed in 1978 and as a result, SLAs were able to offer money-market certificates with a minimum denomination of $10,000 at market interest rates. By 1979, these money-market accounts constituted 20 percent of all the total deposits at thrifts.
From 1980, legislation, interest-rate variability, and the appearance of new intermediaries dramatically altered the asset powers and opportunities available to SLAs and their incentives to take risks (White, 1991). First, the Depository Institutions Deregulation and Monetary Control Act (DIDCMA) of 1980 and the Garn-St. Germain Act of 1982 extended the asset powers and opportunities of SLAs. Both pieces of legislation promoted the use of adjustable rate mortgages and authorized SLAs to issue credit cards, disburse commercial and consumer loans, and take equity positions in ventures through service corporations. Additionally, the DlDCMA ordered the phasing out of Regulation Q and permitted SLAs to offer interest-paying checking accounts to individuals nationwide. The Garn-St. Germain Act speeded the phaseout of Regulation Q and also authorized SLAs to offer checking accounts to commercial businesses. Finally, the DIDCMA also raised the insurance coverage from $40,000 per owner for insured accounts at insured institutions to $100,000 per owner and thereby increased the incentives for managers to take risks with the funds at their disposal. Second, several states also extended the asset powers of state-chartered thrifts because they were concerned that state-chartered SLAs would switch to a federal charter to exploit widened asset powers. Florida and California were among the notable states that enacted legislation to extend the discretion of SLA managers. Third, the Economic Recovery Act of 1981 permitted shortened depreciation periods for real estate investments and enhanced their profitability. Favorable tax treatment persuaded several thrifts, especially in the Southwest, to fund commercial real estate projects. Finally, SLAs were motivated to take risks because steep increases in interest rates eroded their net worth during 1980-1982. Additionally, the appearance of money-market mutual funds and mortgage bankers using collateralized mortgage obligations intensified competition for deposits and further accentuated the incentives for managers to take risks. As a result of deregulation, the structural differences between mutuals and stocks may have been attenuated, and the mortality of all SLAs may have increased:
Hypothesis 4: Deregulation increases the mortality of SLAs.
Other than structural similarity caused by blending processes, the survival of SLAs is also likely to be influenced by their size and age. Smaller organizations are more likely to disband than larger organizations because they lack internal resources and external supports (Hannan and Freeman, 1989). Similarly, the survival of SLAs is also likely to be influenced by their age, although the specific nature of the relationship between age and mortality is controversial. Some writers suggest that organizations face a liability of newness at their inception and that the risk of failure declines as a function of age (Hannan and Freeman, 1989). Others propose that organizations face a liability of adolescence and, as a result, the relationship between age and mortality is nonmonotonic (Bruderl and Schussler, 1989). We thus made no predictions about the effects of age and size but examined these factors in our analysis.
METHOD
Data
Event history data were collected to explore the mortality of SLAs. An event history is a record containing information on the timing of the founding of an organization, its significant characteristics, and its mortality. Event history data were collected retrospectively for 900 SLAs insured by the FSLIC. Data collection was confined to SLAs founded after January 1, 1960 and up to August 31, 1987. This was done to eliminate left-censored data, i.e., on SLAs that were founded before the period under study. Each organization’s event history was broken into yearly spells, and the values of all covariates were updated in each year. All spells other than the year in which an organization disbanded were treated as right-censored. Data were collected from the annual directories of membership of the FSLIC and cross-checked with the merger and insolvency records maintained by the FHLBB. During the 28-year period, 900 SLAs were founded; 240 were mutuals and 660 were stock SLAs. At the end of the observation period, 315 SLAs had disbanded: 140 were mutuals and 175 were stocks.
Figure 1 shows the relationship between age and the mortality of mutuals and stocks. It clearly suggests a nonmonotonic relationship between age and mortality: As age increased, the probability of mortality also increased and, after a point, the relationship between age and mortality was reversed. Thus, mutuals and stocks were afflicted by a liability of adolescence.
Measurement
Foundings were defined as the chartering of an SLA by the FHLBB or state-level regulators. Charters manifest the commitment of founders and powerful external agencies (Singh, Tucker, and House, 1986). The founding date was defined as the date on which an SLA was chartered. Many thrifts grew by merger, and some of them tended to change their founding dates to signal stability: Acquiring thrifts would report the charter date of an acquired SLA if it was older. In such cases, we used the charter date of the acquiring thrift and not the older date of the acquired thrift.
The guiding rule in defining mortality was that the organization ends when “it ceases to carry out routine actions that sustain its structure, maintain flows of resources and secure the allegiance of its members” (Hannan and Freeman, 1989: 149). Accordingly, we did not treat SLAs deemed to have negative or zero net worth according to Generally Accepted Accounting Principles (GAAP) as disbandings, because they were in existence and continued operations. Similarly, we did not treat SLAs under the Management Consignment Program as disbandings, because although their management was changed, the organization continued to operate in the marketplace. Likewise, SLAs subject to regulatory intervention, such as supervisory agreements about the conduct of business or those under orders to cease and desist from certain activities, were also not treated as having disbanded because they were still in existence. Similarly, reorganizations of mutuals into stocks were not defined as disbandings. First, such reorganizations did not mark the end of the organization as an actor, only the cessation of a certain mode of organizing. Second, few mutuals founded during 1960-1987 converted into stocks.
Disbandings were defined as the liquidation, forced merger, or voluntary merger of SLAs. The distinction between forced merger and liquidation is somewhat thin in the SLA industry because regulators discouraged liquidations to preserve confidence and reduce insurance costs (Brumbaugh and Carron, 1988). Moreover, many voluntary mergers were only a few steps away from becoming forced mergers and involved the absorption of small SLAs by large SLAs (Woerheide, 1984: 183). However, managers of mutual SLAs have equally strong incentives to resist voluntary and forced mergers: Mergers by absorption deprive the managers of mutuals of their control over resources. Unlike their counterparts in stocks, they cannot sell stock or gain stock in the new entity created after the merger.
Differences in chartering were measured by a dummy variable (FEDERAL): 1 stood for federal charter and regulation and 0 for state-level charter and regulation by state agencies. To account for differences in asset powers among state-chartered associations (Benston, 1985), we created a dummy called LIBERAL to test whether state-chartered SLAs with liberal asset powers based in California, Texas, Florida, and Louisiana had a higher mortality than SLAs with narrower asset powers based in other states. An interaction term, LIBERAL X STOCK, was created to examine whether stocks based in liberal states had different survival prospects than other SLAs. A period dummy called DEREGULATION was created to account for the advent of deregulation from 1980 on. Another dummy variable, STOCK, was constructed to account for structural differences: 1 stood for stocks and 0 for mutuals.
Environmental variability was defined as variation in six-month T-bill rates. Our measure of environmental variability (ENVIRONMENTAL VARIABILITY) was the coefficient of variation for each year from 1960 to 1987; V = |sigma~/x, where x is the mean net savings inflows over the period and |sigma~ is the standard deviation of the time series over the period. Thus, our operationalization of environmental variability was similar to that of Freeman and Hannan (1983) in their study of restaurants.
Density, or the number of existing organizations in a population, was used as a measure of competition (Hannan and Carroll, 1992). Local and nonlocal densities of SLAs, commercial banks, life insurance companies, and mutual savings banks were computed. The local density of each financial intermediary was the number of organizations of that type in existence in a state in which a focal SLA was based. The nonlocal density of each financial intermediary was the number of organizations in states other than the one in which the focal SLA was located.(1)
Data on the size of SLAs were fragmentary and could not be used in the general models of mortality. We collected size data for a subsample of SLAs in order to test, at least partially, for the effect of size on mortality. This subsample consisted of SLAs founded after 1979. Size was defined as the assets of an SLA. Organizational age was treated as an independent covariate, and first- and second-order terms were constructed, in view of the nonmonotonic relationship between age and mortality.
MODELING STRATEGY AND RESULTS
Our dependent variable was the probability of mortality, and we chose to study it by the logit regression model that is expressed as follows:
|Mathematical Expression Omitted~
where |alpha~ is the intercept parameter, |beta~ is a vector of parameters, and x is a vector of time-constant and time-varying covariates.
Table 1 displays the results obtained from modeling the mortality of SLAs in our sample. Model 1 shows a significant nonmonotonic relationship between age and mortality: The first-order effect is positive and the second-order effect is negative. Federally chartered SLAs were significantly more prone to fail than state-chartered SLAs. Deregulation significantly increased the mortality of SLAs. Environmental variability significantly raised the mortality of SLAs. Local competition from commercial banks and life insurance companies also significantly boosted the mortality of SLAs. Nonlocal competition from SLAs, commercial banks, life insurance companies, and mutual banks was also significantly hazardous for SLAs. Model 2 adds the effects of organizational form. Stock SLAs significantly outlasted mutual SLAs. The effects of all other variables remained unchanged. The -2 log likelihood leaps from 631.30 to TABULAR DATA OMITTED 661.40, indicating that the organizational form is an important variable that explains the mortality of SLAs.
In model 3, we explored whether state-chartered SLAs based in liberal states had shorter lives than other SLAs and if state-chartered stocks in liberal states had significantly different mortality rates than other SLAs. The effects of LIBERAL are insignificant, showing that there were no differences in the mortality of SLAs based in liberal states and state-chartered stocks based in other states. The interaction term LIBERAL X STOCK is also insignificant and indicates that there were no differences in the mortality of state-chartered stocks based in liberal states and other SLAs. One reason why state-chartered SLAs based in liberal states did not have higher mortality than other SLAs is that the other SLAs were primarily located in the southwestern states, where oil-price declines led to the mortality of SLAs that had invested in commercial real estate (White, 1991). Nonetheless, stocks in model 3 continue to outlive mutuals significantly. Further, competition from nonlocal SLAs, local and nonlocal commercial banks and life insurance companies, and nonlocal mutual banks significantly increased the mortality of SLAs.
In model 4, we dropped the FEDERAL, LIBERAL, and LIBERAL X STOCK dummies because of their insignificance. Stocks still significantly outlast mutuals, and the local commercial banks and life insurance companies and nonlocal SLAs, mutual savings banks, commercial banks, and life insurance companies significantly raise the mortality of SLAs. Deregulation also significantly increases the mortality of SLAs. A comparison of models 4 and 2 indicates that the explanatory power of model 4 is comparable to model 2, even after dropping the FEDERAL, LIBERAL, and LIBERAL X STOCK dummies.
Model 5 pertains to the subsample of SLAs for whom size data were available and is similar to model 2 except that we excluded the DEREGULATION dummy because our subsample with size data consisted of SLAs founded after the advent of deregulation. Model 5 shows a significant nonmonotonic relationship between age and mortality and also reveals that stocks significantly outlasted mutuals. Both results conform to the pattern observed in earlier models. The effects of environmental variability, however, are insignificant. Moreover, the effects of local and nonlocal SLA density and life insurance company density are negative and significant and also contrary to expectations. Only the effects of local commercial bank density and nonlocal commercial bank density and mutual savings bank density are significant and positive and in accord with expectations. In model 6, we added the size control. Stocks still significantly outlast mutuals, and the effects of all other variables remain unchanged. Our analysis of the subsample with size data thus suggests that stocks significantly outlasted mutuals and that only the effects of local commercial bank density and nonlocal commercial bank and mutual savings bank density were significantly harmful to SLAs. Despite the foregoing analysis, the differences in the mortality of mutuals and stocks could be due to size differences, because the effects with size controlled may be the result of sampling error. There is always the possibility that with full-sample size controls, the differences between mutuals and stocks might disappear.
We separately modeled the mortality of mutuals and stocks to test if mutuals were more vulnerable to environmental variability and competition. We tested for differences between the individual regression coefficients for mutuals and stocks by using the comparison-of-means test (Wonnacott and Wonnacott, 1970: 214; Miner, Amburgey, and Stearns, 1990). Table 2 shows the results.
Models 6 and 7 show the estimates for mutuals and stocks. Both forms were characterized by a significant nonmonotonic relationship between age and mortality. Moreover, deregulation significantly increased the mortality of mutuals and stocks, and a comparison-of-means test did not show any differences in their vulnerability to deregulation. Thus, the agency cost advantages of stocks were significantly attenuated by deregulation. In the mutual and stock populations, there are no differences between federal and state-chartered SLAs. Additionally, the effect of environmental variability is insignificant. Local SLA density and life insurance density and nonlocal SLA density and life insurance density significantly increase the mortality of mutuals but not stocks. The effects of local mutual savings bank density are insignificant for both populations, but the nonlocal mutual savings bank density significantly raises the mortality of stocks, while mutuals are unaffected. Local and nonlocal commercial bank density, however, significantly increases the mortality of mutuals and stocks. A comparison-of-means test indicates that mutuals are more vulnerable than stocks to competition from local and nonlocal commercial banks.
TABULAR DATA OMITTED
DISCUSSION AND CONCLUSIONS
The findings indicate that mutuals tended to have significantly higher mortality than stocks in models with and without size controls. The hypothesis that mutuals were more vulnerable to environmental variability than stocks received weak support. Although environmental variability increased mortality in models of all SLAs at the .10 level of significance, it ceased to be significant in analyses of the subsample with size data and separate analyses of mutuals and stocks.
The hypothesis that mutuals were more vulnerable than stocks to competition was unambiguously supported in the case of commercial banks. Competition from local and nonlocal commercial banks was significantly hazardous in models including all SLAs and in analyses of the subsample with size data. Moreover, separate analysis of mutuals and stocks indicated that mutuals were more vulnerable than stocks to competition from local and nonlocal commercial banks. In unreported analyses, we analyzed involuntary mortality and found that stocks significantly outlived mutuals and that competition from local and nonlocal commercial banks significantly increased involuntary mortality.
By contrast, there was less unequivocal evidence showing that mutuals were more vulnerable than stocks to competition from local and nonlocal SLAs, life insurance companies, and mutual savings banks. Local mutual savings bank density exerted insignificant effects. Although nonlocal mutual savings bank density significantly increased mortality in models with and without size, it had no significant effects when the mortality of mutuals was separately analyzed. Local SLA density significantly increased mortality only when the mortality of mutuals was separately studied, and it reversed direction in models with size. Similarly, although nonlocal SLA density and local and nonlocal life insurance company density significantly raised mortality in models 1-4 and in separate analysis of mutuals, they reversed direction in models with size controls.
Together these findings suggest that the positive significant effects of local life insurance company density and nonlocal SLA, life insurance company, and mutual savings bank density may be the result of specification error. Favorable conditions may lead both populations in question to expand, but there is neither mutualism nor competition. Moreover, many of the competing SLAs were founded before 1960, and the density-dependence literature makes no predictions under such conditions. A study of the complete history of the population is necessary to adjudicate whether the positive effects of local life insurance company density and nonlocal SLA, life insurance company, and mutual savings bank density were the results of specification error.
Our results indicate that there were no differences in the mortality of federal and state-chartered SLAs. One reason why regulators might not be able to affect involuntary dissolutions is that regulatory agencies are concerned with conformity rather than efficiency (O’Hara, 1981; Meyer and Scott, 1983). Additionally, understaffing, undersupply of information, and jurisdictional conflicts may also have led to regulatory impotence (Balderston, 1985; White, 1991). The results also show that stocks significantly outlasted mutuals. The findings show weak support for the hypothesis that mutuals were more vulnerable than stocks to environmental variability. But the hypothesis that mutuals are more susceptible than stocks to competition was unambiguously endorsed in the case of local and nonlocal commercial banks. Together, these results lend support to the thesis that mutuals had higher mortality rates than stocks because of their weaker monitoring, incentives, and capital structure; however, the findings also show that deregulation attenuated the agency-cost advantages of stocks. Deregulation created new capabilities, opportunities, and incentives for SLA managers to take risks that overstretched the monitoring and incentive systems in stock SLAs. Another study, by White (1991: 115-116), analyzed a sample of soon-to-fail thrifts in 1985 and observed that a large fraction was accounted for by stock organizations that imprudently used their new capabilities to pursue new and risky opportunities.
The finding that mutual SLAs are more vulnerable than stock SLAs to competition requires careful generalization. One reason is that our study analyzed a sample consisting of SLAs founded after 1960 and up to 1987 and therefore did not follow density-dependent processes of legitimation and competition over the complete history of the SLA industry. Additionally, we were able to analyze only a subsample with size data on SLAs founded after the advent of deregulation. Further, we did not analyze the effects of diversification on survival (Haveman, 1992) and did not investigate whether reorganizations of business domains were differentially hazardous to mutuals and stocks. Moreover, the mutual form may have agency-cost advantages over the stock form in financial industries devoid of deposit insurance. Mayer and Smith (1986) suggested that mutual life insurance companies are more efficient than stock life insurance companies because they avoid the costs of conflict between policyholders and stockholders. Similarly, Fama and Jensen (1983c) suggested that the mutual form is suited for financial organizations whose assets are the securities of other organizations, because the costs of expanding or contracting assets and securing indices of asset values are low. Thus, the mutual form is the dominant way of organizing money-market funds and investment funds because administrators purchase publicly traded assets, disclose their expenses, and the depositors are able to sell their shares at a price that reflects the value of the assets held by the mutual.
Nevertheless, the finding that stock SLAs were less vulnerable than mutual SLAs to competition but that the agency-cost advantages of stocks were eroded by deregulation suggests that building bridges between agency and ecological perspectives is necessary to enhance our understanding of the selection process. When agency theory proposes that the efficiency of monitoring and incentive alignment determines the survival of organizations, it posits that the selection of organizations depends on the efficacy of micro-level selection processes within organizations, thus extending ecological models of organizations. By the same token, ecological theory delineates the scope of predictions derived from agency theory when it proposes that institutional processes constrain efficiency advantages.
One way of expanding the dialogue between ecological and agency theories of organizations is to focus on how stocks and mutuals are embedded in networks of interpersonal relations (Granovetter, 1985) and study the effects of differential embeddedness on survival. Thus, one question for future research is whether there are significant differences in the mortality of closely owned stocks and diffusely owned stocks. A related question is whether mutuals that are connected to other enterprises by dense social ties encounter different mortality than mutuals that are less connected to other enterprises by interpersonal ties. A final extension of this avenue of research is to study differences between the mortality of mutuals and stocks in the early history of the SLA industry. Mutuals arose as organizations built for the purpose of borrowing from friends, whereas stocks originated as a means of raising capital from strangers for business purposes (Kendall, 1962). As a result, mutuals were more dependent on networks of social relations for monitoring and resources, whereas stock SLAs were less dependent on networks of friends for monitoring and resources. Thus, it would be interesting to discern whether mutuals, as organizations requiring high embeddedness, thrived when they were located in stable neighborhoods and failed when neighborhoods were increasingly populated by strangers. The other side of the coin is whether stocks as organizations with little need for embeddedness flourished in areas populated by strangers and decayed in areas characterized by dense social ties. Studies of how variability in networks of interpersonal ties underlies differential exposure to selection processes would not only connect ecological theory with embeddedness theory but would also qualify the predictions of agency theory. The result promises to be an enhanced social history of organizations in which the survival of organizations becomes contingent on specific historical and social processes rather than a functionally inevitable outcome.
1 We did not compute second-order terms for these densities because we were studying an established population late in its history and felt that the competitive effects of density were likely to overwhelm the legitimation effects of density (Hannan and Freeman, 1989; Carroll and Hannan, 1989).
REFERENCES
Alchian, Armen, and Harold Demsetz 1972 “Production, information costs and economic organization.” American Economic Review, 62: 777-795.
Aldrich, Howard 1979 Organizations and Environment. Englewood Cliffs, N J: Prentice-Hall.
Aldrich, Howard, and Peter V. Marsden 1988 “Environments and organizations.” In Neil J. Smelzer ed.), Handbook of Sociology: 361-392. Beverly Hills, CA: Sage.
Balderston, Fred E. 1985 Thrifts in Crisis: Structural Transformation in the Savings and Loan Industry. Cambridge, MA: Ballinger.
Barnett, William P. 1990 “The organizational ecology of a technological system.” Administrative Science Quarterly, 35: 31-60.
Barnett, William P., and Glenn R. Carroll 1987 “Competition and mutualism among early telephone companies.” Administrative Science Quarterly, 30: 400-421.
Barth, James R., Philip F. Bartholomew, and Carol J. Labich 1989 “Moral hazard and the thrift crisis: An analysis of 1989 resolutions.” In Banking System Risk: Charting a New Course: 344-384. Chicago: Federal Reserve Bank of Chicago.
Barth, James R., and Martin A. Regalia 1988 “The evolving role of regulation in the savings and loan industry.” In Catherine England and Thomas Huertas (eds.), The Financial Services Revolution: Policy Directions for the Future: 113-161. Boston: Kluwer Academic.
Benston, George F. 1985 “An analysis of the causes of savings and loan association failures.” Monograph, Graduate School of Business Administration, New York University.
Bodfish, Morton 1931 History of Building and Loans in the United States.
Chicago: U.S. Building and Loan League.
Bruderl, Josef, and Rudolf Schussler 1989 “Organizational mortality: Liabilities of newness and adolescence.” Administrative Science Quarterly, 36: 530-547.
Brumbaugh, R. Dan 1988 Thrifts under Siege. Cambridge, MA: Ballinger.
Brumbaugh, R. Dan, and Andrew S. Carron 1988 “The thrift industry crisis: Causes and solutions.” Brookings Papers on Economic Activity, Washington, DC: Brookings Institution.
Carroll, Glenn R., and Michael T. Hannan 1989 “Density dependence in the evolution of newspaper populations.” American Sociological Review, 54: 524-541.
DiMaggio, Paul, and Walter W. Powell 1983 “The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields.” American Sociological Review, 48: 147-160.
Fama, Eugene 1980 “Agency problems and the theory of the firm.” Journal of Political Economy, 88: 288-307.
Fama, Eugene, and Michael Jensen 1983a “Organizational forms and investment decisions.” Managerial Economics Research Center, Working Paper no. MERC 83-03, Graduate School of Management, University of Rochester.
1983b “Separation of ownership and control.” Journal of Law and Economics, 26: 301-323.
1983c “Agency problems and residual claims.” Journal of Law and Economics, 26: 327-349.
Freeman, John, and Michael T. Hannan 1983 “Niche width and the dynamics of organizational populations.” American Journal of Sociology, 88: 1117-1145.
Granovetter, Mark 1985 “Economic action and social structure: The problem of embeddedness.” American Journal of Sociology, 91: 481-51O.
Hannan, Michael T., and Glenn R. Carroll 1992 Dynamics of Organizational Populations: Density, Legitimation and Competition. New York: Oxford University
COPYRIGHT 1992 Cornell University, Johnson Graduate School
COPYRIGHT 2004 Gale Group
