An agency theory investigation of medical contractors versus member physicians *.

Professional organizations increasingly utilize contingent labor such as independent contractors and employees of contract companies to provide professional services to clients (Matusik and Hill, 1998). Although there has been a growing research interest in contingent work, it has not been well studied among highly skilled occupations (Masters and Miles, 2002). In addition, the literature has not examined how organizations that use contingent labor manage their contingent workers (see Lepak and Snell (2002) for an exception). Agency theory suggests that organizations use governance mechanisms, such as monitoring and incentives, to ensure that workers act in the organizations' interests (Jensen and Meckling, 1976). Agency theorists specifically include professionals among the types of workers to which the theory applies (Fama and Jensen, 1983). Although agency theory has not been tested in the context of managing contingent versus core professional workers, the theory suggests that professional firms would use more intense governance mechanisms with contract professionals than with member professionals (partners or professional employees).

It is important to understand the governance arrangements that professional organizations use with the professionals with whom they contract. These contractors provide critical services at the technical core of the professional firm where they can strongly influence the organizations' performance (Thompson, 1967). Furthermore, any incompetent performance (malpractice) on the part of contract professionals could damage the firm's reputation and create legal liability for the firm. In addition to having an economic impact on professional firms, the services that professionals provide are critical to their clients and society. It is therefore important for professional organizations to ensure that the people who provide professional services on the firm's behalf perform appropriately. Because contract professionals are not subject to the same selection processes, in-house training, socialization in the organization's culture, and long-term mutually committed relationships as are partners and employees, one might expect professional organizations to use stronger governance mechanisms with contract professionals.

This study examines whether the performance monitoring, decision ratification and incentive compensation that professional organizations use with contract professionals differ from those used with member professionals. We develop hypotheses based on agency theory and test them in a sample of California medical groups. In doing so, we contribute to the agency theory literature, where much of the research involves CEOs and the boundary conditions of the theory have not been sufficiently explored. In addition, we extend the literature on contingent work to include the governance arrangements used with contingent versus member physicians.

THEORY

Professionals

There is a large body of literature on professional organizations and the sociology of the professions (Freidson, 1984; Larson, 1977; Parsons, 1968; Scott, 1982). Much of this work has conflicting perspectives on the degree to which the professions are different from or the same as other occupations, and the degree to which professionals are altruistic and committed to service versus self-interested agents who exploit opportunities to maximize their powerful positions for their own benefit (Sharma, 1997; Van Maanen and Barley, 1984). Much of the sociological theory on professionals has pertained to the mechanisms of control over professionals, including socialization in the profession, a sincere desire to serve others, control by professional peers within and outside of the employing firm, reputation in the community, bureaucratic controls from the hierarchical supervision within firms, and client control (Hodgson, 2002; Sharma, 1997; Tolbert and Stern, 1991). Although collegial control is thought by many researchers to be of utmost importance in the professions, research questions its effectiveness (Van Maanen and Barley, 1984). Professional associations also provide control by creating and enforcing codes of ethics, and because they can revoke the memberships and licenses of opportunistic or incompetent professionals (Sharma, 1997). However, reports of professionals who have repeatedly engaged in misconduct and harmed clients suggest that professional associations alone do not provide sufficient discipline (Crane, 2003; Lindsay, 2003).

Governing professionals is complex for a number of reasons (Freidson, 2001; Ouchi, 1979; Raelin, 1986). Professionals are strategically central and perform key functions in the technical core of firm operations where they are required to exercise discretion and independent decision making based on specialized knowledge of varying situations (Clegg, 1981; Jones, 1983). The high cost of transferring case-specific knowledge and the potential for ambiguity about the correct action make the evaluation of professional performance difficult. In addition, it is difficult for consumers of professional services to provide the type of customer control that is important in some other fields (Eisenhardt, 1988). This is because professionals' clients usually lack the training necessary to evaluate the quality of professional services and are generally submissive in their relationships with professionals (Van Maanen and Barley, 1984). As a result, client control is based more on interpersonal attraction and trust rather than an objective evaluation of performance (Sharma, 1997).

In addition to the difficulties just discussed, professional culture, which values self-control and autonomy, makes it challenging for professional organizations to govern professional workers (Jones, 1983; Van Maanen and Barley, 1984). Most professionals have internalized a set of norms as a result of their lengthy and rigorous education and socialization, and govern themselves by social-regulative rules more than other types of workers (Clegg, 1981; Hodgson, 2002). Professionals tend to resist organizations' attempts to exert control over them or in any way limit their freedom to make important decisions independently (Freidson, 2001; Raelin, 1986). When an organization's interests, standards or procedures conflict with the opinions of peers in the profession, which is common, professionals usually side with the profession (Bolton, 2004). This is because most professionals' identity is more closely associated with the profession than with the organization (Alvesson and Willmott, 2002). In spite of the difficulties associated with governing professionals, research suggests that organizations do utilize controls to limit the autonomy of their professional workers (Osterman, 1994).

Contingent Workers

Although most professionals, such as physicians, lawyers, and accountants, once practiced as solo practitioners or partners, the majority now work for organizations (Greenwood et al., 1997). Today's professional organizations must manage a variety of core and contingent employment arrangements. According to a 1999 survey by the United States Bureau of Labor Statistics (BLS), 29% of college and university instructors, 12.3% of physicians, and 11.8% of biological and life scientists were contingent workers (Hipple, 2001). Overall, 6.7% of people working in the professional specialty category of occupations work in contingent arrangements (Hipple, 2001). The BLS defines contingent work as "Any job in which an individual does not have an explicit or implicit contract for long-term employment" (Hipple, 2001: 4).

Although the management implications of contingent employment are not yet well understood, research has provided some insight into its growing popularity (Kochan et al., 1994). Organizations use contingent workers to save benefits costs, keep total labor costs low, increase flexibility, including the flexibility to terminate workers, screen candidates for permanent positions, and access knowledge that contingent workers have gained through their experiences at other organizations (Autor, 2003; Houseman, 2001; Matusik and Hill, 1998). Organizations also use contingent workers to meet seasonal needs, work on special projects, and fill in when regular employees are unavailable (Hipple, 2001). Organizations tend to use contract workers to meet general needs and access skills that are widely available in the marketplace, while maintaining an employment relationship with individuals who perform work that is more critical to firm competitiveness (Lepak and Snell, 2002). A recent study found that the use of contingent labor was associated with higher organizational performance with no increase in systematic risk or standard deviation of returns (Nayar and Willinger, 2001). This suggests that using contingent labor may be an effective strategy for maximizing shareholder returns.

Research suggests that contingent work also has advantages for highly skilled workers, who may intentionally pursue career paths working temporarily for various organizations (DiNatale, 2001; Marler et al., 2002). They find security in developing their skills so that they can easily find other jobs when the current assignment ends. The market demand for their skills gives some professional workers greater control and flexibility in their career. Contingent work can also help professionals to balance work and family demands, whereas professionals in employment relationships are often pressured to work very long hours (Landers et al., 1996). In addition, contract professionals enjoy the variety, flexibility and autonomy of contingent work (Oberle, 1989).