This action might not be possible to undo. Are you sure you want to continue?
American Accounting Association DOI: 10.2308 / accr.2010.85.1.343
Group-Based Compensation in Professional Service Firms: An Empirical Analysis of Medical Group Practices
Mina Pizzini Southern Methodist University
ABSTRACT: This study investigates factors inﬂuencing the use of group-based compensation in professional service partnerships and the relation between group-based compensation and performance. I use data on 11,971 physicians in 935 medical groups to expand the extant literature on group incentives by providing some of the ﬁrst largesample, empirical evidence on the role of task interdependence, income risk, mutual monitoring, and group size in a ﬁrm’s choice of compensation contract. Consistent with agency theory, I ﬁnd that group incentives are more prevalent in medical partnerships that practice highly task-interdependent specialties and those that face greater malpractice risk. Group-based incentives are also more common in relatively small groups in which homogeneity in training, experience, and gender facilitate mutual monitoring. Tests relating individual physician productivity to compensation method suggest that productive beneﬁts induced by group incentives offset reductions in output associated with free-riding and effort devoted to monitoring. Keywords: group incentives; risk sharing; teams; partnerships; compensation. Data Availability: Data are available from sources identiﬁed in the article.
I. INTRODUCTION his study identiﬁes factors that inﬂuence the use of group incentives in labormanaged professional service ﬁrms and compares the performance effects of group incentives with those of individual incentives. Partnerships are the predominant form of organization in human-capital intensive professions such as accounting, law, consulting, and medicine. Professionals form partnerships to exploit productive interactions, share resources, and spread risk. In practice, partnerships employ a wide variety of compensation schemes or ‘‘sharing rules,’’ with some partnerships rewarding members based solely on individual performance, while others base compensation on group performance and divide
I thank J. Harry Evans III, Christopher Ittner, David Larcker, Wendy Wilson, and seminar participants at the University of Rochester, Southern Methodist University, Tilburg University, The Ohio State University, the Lone Star Accounting Conference, and the Global Management Accounting Research Symposium for their comments. I appreciate the constructive comments of two anonymous referees and the suggestions of Steven Kachelmeier (editor). I also thank the Medical Group Management Association (MGMA) for granting me access to their data. Editor’s note: Accepted by Steven Kachelmeier, with thanks to Dan Dhaliwal for serving as editor on a previous version.
Submitted: April 2007 Accepted: February 2009 Published Online: January 2010
the group’s net proceeds equally among partners (i.e., group-based or team-based compensation). As compensation moves away from individual performance to group performance, the incentive to free ride on the effort of others increases. Traditional agency theory contends that ﬁrms weigh the beneﬁts arising from group-based compensation, which include greater cooperation among group members and risk sharing, against losses in output from diminished incentives (e.g., Alchian and Demsetz 1972; Newhouse 1973; Holmstrom 1982). More recent work indicates that in certain settings, group incentives induce mutual monitoring and thereby lead to higher levels of effort (e.g., Kandel and Lazear 1992; Che and Yoo 2001; Huddart and Liang 2005). Taken together, this research suggests that ﬁrms will use group-based compensation under the following conditions: (1) work tasks are highly interdependent, (2) income risk is relatively high, (3) ﬁrm members can cost-effectively monitor each other, and (4) ﬁrm size is relatively small. An extensive theoretical literature analyzes compensation arrangements and free-riding in groups; yet, there is relatively little empirical research on the determinants and performance effects of group-based compensation (see Prendergast  for a review). I investigate the determinants and performance effects of group-based compensation using a database of 11,971 physicians in 935 medical group practices. This setting and data offer several advantages. First, physician compensation contracts tend to provide a relatively simple expression of the link between pay and performance. Second, compensation contracts vary by practice even though medical group practices operate in the same service sector (thereby controlling for confounding industry effects). Groups generally base compensation primarily on individual performance or primarily on group performance, with all members sharing equally in the group’s net proceeds. Third, physician-level performance measures provide a unique opportunity to identify the effects of group incentives on individual performance. Prior research primarily examines group-level performance. Fourth, physician-level data measure factors that potentially inﬂuence both the use of group incentives and the effects of group incentives on individual performance. Variations in these factors afford for powerful tests of hypotheses suggested by theoretical models. Finally, rising expenditures for physician and clinical services, which totaled $447.6 billion in 2006, have drawn critical attention to the structure of physician incentives (Centers for Medicare & Medicaid Services 2008). This study expands the extant literature on group incentives by providing some of the ﬁrst large-sample, empirical evidence on the role of task interdependence, income risk, mutual monitoring, and group size in a ﬁrm’s choice of compensation contract. Medical groups that practice highly task-interdependent specialties (e.g., emergency medicine) are signiﬁcantly more likely to use group incentives than physicians who practice specialties with little opportunity for cooperation (e.g., psychiatry). The use of group incentives is positively associated with malpractice risk, suggesting that physicians use group-based compensation to insure themselves against income risk arising from malpractice lawsuits. Group incentives are also more prevalent in groups with relatively low mutual monitoring costs, as is the case with single-sex groups composed of members with similar training and experience. Finally, small- to medium-sized groups are more likely to use group incentives than larger groups. This study also contributes to the literature by examining the individual performance effects of group incentives. Medical practices using group incentives exhibit signiﬁcantly less variation in productivity than those utilizing individual incentive plans. More importantly, in contrast to the ﬁndings of prior empirical research (Encinosa et al. 1997), physicians whose compensation is based upon group performance are not less productive than
The Accounting Review American Accounting Association January 2010
Group-Based Compensation in Professional Service Firms
those whose compensation is based upon individual performance. Thus, in physician ﬁrms, the productive beneﬁts induced by group incentives offset reductions in output associated with free-riding and effort devoted to mutual monitoring. These ﬁndings are relevant for other labor-managed professional ﬁrms that share key organizational features with medical practices as well as corporations that increasingly require employees to work in teams. Section II develops the research hypotheses. Section III discusses the sample and measures. Empirical tests are reported in Section IV. Results are discussed in Section V and conclusions are presented in Section VI. II. BACKGROUND ON PHYSICIAN GROUP PRACTICES The majority of medical doctors work in group practices owned by the doctors themselves. Non-owner doctors in physician-owned practices are typically new hires who will be granted an ownership interest after a few years of service (Lee 1990). Medical group practices range in size from three physicians to upward of 500, although most contain between ﬁve and 50 members (American Medical Association [AMA] 1999). I use the terms ‘‘partnership’’ and ‘‘group’’ interchangeably to refer to the provision of health care services by three or more physicians who are formally organized as a legal entity governed by physicians and who share business, clinical, and administrative facilities (AMA 1999, 1). Physician-owned medical group practices are normally organized as limited liability corporations (LLC), limited liability partnerships (LLP), professional associations (PA), or partnerships.1 The health care literature attributes the predominance of group practices in the medical profession to the ability to exploit economies of scale (Pauly 1996), smooth work schedules, and internalize referrals (Gaynor and Gertler 1995). In addition, groups have the potential to structure and coordinate care so as to achieve higher quality, greater patient satisfaction (Pauly 1996), and reputational economies of scale (Getzen 1984). Outside of the health care literature, Levin and Tadelis (2005) show that partnerships assure clients of higher levels of quality when quality is difﬁcult to observe. Although most physicians belong to group practices, the majority of physician outpatient services are provided in one-to-one interactions between a physician and a patient (Pauly 1996, 26). Differences in work habits, work pace, specialty, interpersonal skills, and other factors produce intra-group variations in the quantity, perceived quality, and cost of care provided, resulting in different contributions to the group’s net earnings. Physician groups distribute earnings based upon individual performance or group performance. In the data set used for this study, 84 percent of practices use individual plans and 16 percent use group plans. Under individual performance plans, each physician receives some combination of salary and productivity-based pay determined by his/her individual performance. Physician ﬁrms normally use accounting metrics such as gross charges or adjusted charges to measure productivity. Under a group performance plan, each physician receives an equal share of the entire net proceeds of the group, regardless of individual output and effort. I refer to such incentive plans as equal-sharing, team-based compensation, and group-based compensation throughout the article.
In all four organizational forms, the net proceeds of the entity pass through to its owners who are taxed at the individual level. These organizational forms also offer limited liability, which prevents a group member from being legally responsible for the acts of another group member. Practices that are not owned by physicians are typically owned by hospitals and foundations. Such practices are excluded from the analysis because these physicians are employees of the entity that owns the practice.
The Accounting Review
January 2010 American Accounting Association
it is optimal for principals to use simple group incentive plans to encourage this activity. Doctors also typically share ‘‘call duty’’ or the responsibility to care for any of the practice’s patients on an urgent. and (4) group size. diagnostic images or specimens) that come through their respective departments. as-needed basis. or simply note the hours their colleagues maintain. physicians can collect and review common productivity measures (e. the highest levels of task interdependence occur in inpatient settings (Pauly 1996). physicians can free ride on their colleagues by seeing fewer patients or performing fewer procedures than their colleagues. or patient visits). substantially enhancing the value of physicians’ leisure time (Casalino et al. Empirical studies provide limited evidence on the performance effects associated with different combinations of task interdependence and reward system design. In medicine. Wageman (1995) ﬁnds that highly task-interdependent The Accounting Review American Accounting Association January 2010 . (1997) and Che and Yoo (2001) ﬁnd that in team settings. even in outpatient settings. When tasks are highly interdependent. discuss treatment approaches. agency models suggest that professionals can use group incentives to encourage cooperation and thereby improve output. (3) the ability of group members to monitor each other. 1976). HYPOTHESES Determinants of Group-Based Compensation In theory. To empirically test agency-based models of group incentives. Wageman and Baker (1997) show that when tasks are highly interdependent. Holmstrom and Milgrom (1991) and Itoh (1993) demonstrate that when it is Pareto-efﬁcient for agents to coordinate their efforts. principals should base compensation on group performance to induce members to monitor each other and punish free riders. and communication. monitor their colleagues. The management literature uses the term ‘‘task interdependence’’ to characterize the extent to which group members rely upon each other to perform their jobs (Wageman 1995.g. charges.g. group-based rewards are associated with increased output. Arya et al. 2003). physicians can. Compared with inpatient settings. another partner often sees his/her patients so that they can be treated in a timely manner. health care delivery in outpatient settings is much less task-interdependent. which in turn improve the group’s reputation and increase the value of the partnership. To prevent free-riding. at some cost. (2) income risk.. A team of physicians works together to perform a complex surgery or administer urgent care to a patient in an emergency room. the greater the gains from cooperation. coordination. In a ﬁeld study of 152 groups of service technicians. For example. Yet. cooperative gains include higher levels of output and quality. However. For example. under group incentive plans.. it is difﬁcult. to directly link output to individual members’ efforts and actions. Van de Ven et al. When tasks are interdependent.346 Pizzini III. In medicine. Wageman and Baker 1997. Physicians often consult with each other on cases to determine the best treatment plan. if not infeasible. ﬁrms choose group incentives if such incentives induce cooperative and insurance beneﬁts that exceed the sum of losses in efﬁciency from free-riding and the costs associated with monitoring group members. intra-group coordination has the potential to improve collective output and quality. Physicians in the hospital-based specialties of radiology and pathology coordinate with each other to process cases (e. collections. Task Interdependence Groups are typically formed to take advantage of efﬁciency gains that accrue from cooperation. The more taskinterdependent a job is. when a physician is unavailable. I examine four factors that are predicted to inﬂuence the beneﬁts and costs associated with team incentives: (1) task interdependence.
While virtually all physicians are fully ensured against the majority of direct lawsuit expenses. this loss will be spread among all partners when compensation is based on group performance. Interdependencies among multiple business units within the same ﬁrm reﬂect the extent to which a business unit manager’s actions affect the performance of other business units. other shocks include difﬁculty in fee collection. neither study examines the role of task interdependence in contract design. Evans et al.3 Physicians can also reduce ﬁnancial risk by entering into equal-sharing arrangements with other physicians. and the likelihood of increased malpractice insurance premiums. theoretical and limited empirical research suggest the follow hypothesis: H1a: The probability that a professional partnership bases compensation on group performance is increasing in the degree of task interdependence.’’ The Accounting Review January 2010 American Accounting Association ..Group-Based Compensation in Professional Service Firms 347 groups perform better when rewards are based upon group performance rather than individual performance. Prior research has shown that physicians who bear greater legal liability (due to state tort laws and court systems) seek to reduce overall ﬁnancial risk by practicing ‘‘defensive medicine’’ and accepting less cost risk (through capitation) in their contracts with Managed Care Organizations (Kessler and McClellan 1996. The risk-sharing feature of group-based compensation plans leads to the following hypothesis: H1b: Professional partnerships located in geographic areas with higher malpractice risk will be more likely to compensate members based upon group performance than those practicing in areas with lower malpractice risk. Although results of both studies are consistent with theory. If a lawsuit reduces the amount of revenues that one physician generates. the indirect costs of a lawsuit can be substantial (Kessler and McClellan 1996). as well as the uncertainty of the litigation process (Evans et al. Using an experimental study.g. Two related accounting studies examine the role of organizational interdependencies in the design of compensation contracts for business unit managers in hierarchical corporations (see Bushman et al. including damages and legal fees.2 Medical malpractice liability exposes physicians to the risk that a patient experiences a bad outcome. Kessler and McClellan (1996. invest heavily in their own human capital. Lawsuits require a physician to spend time preparing a defense. like many professionals. Lang and Gordon 1995). Wageman and Baker (1997) report that cooperation leads to superior performance when interdependent tasks are accompanied by group rewards. Income Risk Arising from a Malpractice Lawsuit Physicians. 2 3 For physicians. shifts in demand for particular specialties. 2006). Malpractice lawsuits are one example of a shock that can significantly reduce a professional’s earning capability. Taken together. variation in insurance coverage and claim denial. reputation effects stemming from litigation reduce future revenues. 1046). no-shows. Equalsharing arrangements enable risk-averse professionals to ensure themselves against idiosyncratic shocks that impair the value of their human capital (Gaynor and Gertler 1995. Empirical results suggest that the use of aggregate accounting performance measures (e. 254) deﬁne ‘‘defensive medicine’’ as the provision of ‘‘precautionary medical treatments that have minimal expected medical beneﬁt out of fear of legal liability. divisional operating income) is increasing in the level of organizational interdependencies.g.. I focus on malpractice risk as a source of income risk because the legal environment signiﬁcantly inﬂuences physician behavior. 2000. 2006. ﬁrm-wide operating income) relative to more direct performance measures (e.  and Keating ). 593). and simply luck (Gaynor and Gertler 1995.
Accordingly.g. Narayanan 1995. higher levels of income insurance in larger groups suggest that. and empathy is likely to be stronger among partners who are similar. If public signals of partners’ individual efforts exist.348 Pizzini Mutual Monitoring Theoretical research has shown that when it is advantageous to use team-based compensation. so that the income variance for any given sharing rule is smaller in larger groups (Lang and Gordon 1995. H1c: The probability that a professional partnership bases compensation on group performance is decreasing in the diversity of its members. Conversely. Lazear 1989. Encinosa et al. the insurance beneﬁts inherent in equal sharing plans predict a positive association between size and sharing.g. and envy. (1997) ﬁnd that use of equal-sharing decreases as group size increases. income risk is spread over more partners. ceteris paribus. Group Size Theoretical models provide competing perspectives regarding the effect of group size on sharing. the incentive for group members to supply effort and to monitor others decreases with size (e. 1997). As group size increases. job responsibility. The assumption that mutual monitoring and peer pressure are more effective in homogeneous groups leads to Hypothesis 1c.4 However. Alchian and Demsetz 1972.. but proﬁt-sharing does not increase as size increases beyond 30 members. Arya et al. shame. Quite simply.. sharing will be greater in larger groups. Encinosa et al. Institutional differences potentially explain the contrasting 4 In one-period models with budget-balancing. In a study of 583 medical groups. Most theoretical models rely on explicit contractual sanctions such as monetary penalties or the threat of expulsion. which deter free-riding (e.. Che and Yoo 2001. multi-period models demonstrate that peer pressure can create implicit sanctions. hence. Whether explicit or implicit. including guilt. Miller 1997. explicit sanctions require that a partner or group of partners outside the group being evaluated serve as the ‘‘sink’’ to administer punishments and rewards (Miller 1997). it is easier for a group member to monitor someone who is like himself in training. Homogeneity also increases the power of peer pressure. 1997). Huddart and Liang 2005). Narayanan (1995) and Che and Yoo (2001) use multiperiod games to relax the budget-balancing constraint and enable partnerships to correlate reward with effort through threat of expulsion or non-cooperation in future periods.g. Direct empirical evidence on the relation between group size and equal-sharing in partnerships is limited and mixed. homogeneity reduces monitoring costs and the incentive to shirk monitoring responsibilities. Milgrom 1988. Lang and Gordon (1995) ﬁnd that proﬁt-sharing in 329 law practices increases signiﬁcantly as ﬁrm size increases from fewer than 21 lawyers to 21–30 lawyers. Peer pressure relies on partners feeling empathy for each other. and work patterns than it is to monitor someone who differs on these factors (Kandel and Lazear 1992. the effectiveness of mutual monitoring and peer pressure is likely to increase in the homogeneity of team members. Encinosa et al. Pay-performance intensity decreases as group size increases. Accordingly. mutual monitoring and peer pressure can mitigate free-riding (e. Huddart and Liang 2004). group members can use these signals as grounds for imposing explicit or implicit sanctions upon fellow group members (Hansmann 1996). 1997). Holmstrom 1982. The Accounting Review American Accounting Association January 2010 . Huddart and Liang (2005) preserve budget-balancing with explicit sanctions by distributing the risk associated with the payment schedule across the other partners of the ﬁrm.
Still. 2002). The Accounting Review January 2010 American Accounting Association . this issue warrants further empirical investigation. prior evidence of freeriding in physician ﬁrms suggests the following hypothesis: H1d: The probability that a professional partnership compensates members based upon group performance is decreasing in group size.-born males practicing alone or in small. a more diverse set of physicians practiced in a wider variety of settings and organizations. Leone 2002). high-ability professionals select partnerships with members whose abilities match their own. professionals sort into partnerships of roughly equal ability. Partnerships composed of high-ability professionals are reluctant to offer partnership status to low-ability professionals to avoid subsidizing less-able partners (Farrell and Scotchmer 1988). neither study controls for several factors that are correlated with both group size and compensation plan. Group size increased and a greater percentage of physicians began practicing in groups. Leibowitz and Tollison (1980) posit that brand-name capital and intra-ﬁrm referrals make larger law ﬁrms economically efﬁcient. Aside from the selection effects created by team-based compensation plans. and geographic area. Intra-Group Variation in Productivity Theoretical research suggests that group incentives work best with individuals of similar ability and work ethic. loosely structured clusters (Mauney 2002). During the 20 years that have elapsed between Encinosa et al. physicians will have less need for income insurance than lawyers (Cooper et al. income risk. namely experience. Although Weiss (1987) provides much-needed empirical evidence on the effects of group incentives. Encinosa et al. it is not clear whether his results hold for professionals. while retaining medium-ability workers. Finally. He ﬁnds that output variability decreases as workers progress from a piece-rate plan in the ﬁrst month of employment to a group incentive plan in the fourth month. which may not be relevant to the current organization of medical practices. intra-group diversity. Professional corporations offering limited liability replaced partnerships and sole proprietorships as the dominant organizational form.’s (1997) survey and the 1998 survey used in this study. Income risk increased from 1978 to 1998 due to increases in malpractice risk. As a result. fundamental changes in the organization of medical practices and the level of income risk borne by physicians have signiﬁcantly affected contract design. (1997) use survey data from 1978.6 Accordingly. The following sections consider evidence on the relation between group incentives and performance. Empirical evidence on the performance effects of group incentives is primarily limited to Weiss’ (1987) case study of 752 newly hired manufacturing workers. Also. and greater use of capitation contracts (Mauney 2002. Health systems and managed care organizations began hiring physicians on a salary basis. Similarly. Kandel and Lazear’s (1992) model suggests that peer pressure creates incentives that further minimize deviations from group productivity norms. Analysis of worker ‘‘quit rates’’ suggests that group-based compensation drives out the low-ability and high-ability workers. legal work is generally more task interdependent than medicine. The selection and incentive effects of group-based compensation suggest the following hypothesis: H2: Intra-ﬁrm variation in productivity will be lower in professional partnerships that base compensation on group performance than it is in those that base compensation on individual performance. 5 6 To the extent that demand for physician services is stronger and less variable than that for law services.Group-Based Compensation in Professional Service Firms 349 results for lawyers and physicians.5 Also.S. By 1998. Effects of Team-Based Compensation on Performance Firms choose compensation systems to attract desirable members and encourage them to take appropriate actions. greater limitations on reimbursements from Medicare and other third-party payers. In 1978 most physicians were U.
IV.. 1997. under group incentives plans (e. However.350 Pizzini Individual Productivity Agency theory contends that. Consistent with the AMA’s deﬁnition of group practice. I include controls for factors that are correlated with both productivity and contract design (clinical hours. Unlike prior studies. ceteris paribus. 1997).98 percent. Three particularly relevant empirical studies use data from a 1978 survey of 6.7 Gaynor and Pauly (1990) and Gaynor and Gertler (1995) ﬁnd that productivity is reduced when individual incentives are weakened either by introducing a ﬁxed salary component to individual compensation or by increasing the degree of revenue-sharing among group members. Surveys were mailed to 5. Limitations in the 1978 data used in these studies and signiﬁcant changes in the health care environment (see footnote 6) suggest that additional insight can be gained by revisiting equal-sharing arrangements with a larger. Additionally. and malpractice risk). Gaynor and Gertler 1995. Alchian and Demsetz 1972. respectively. Kandel and Lazear 1992. Encinosa et al. Consistent with these claims. Despite the large theoretical literature on group incentives plans.. Arya et al.772 practices. Limited empirical evidence on physicians suggests that weakening the pay-performance link lowers output and thus provides the basis for Hypothesis 3: H3: Individual productivity will be lower for professional partnerships that base compensation on group performance than those that base compensation on individual performance. recent theoretical studies suggest that in certain settings. more comprehensive.193 MGMA member practices in February of 1999. or late submission. duplication. I delete groups with fewer than three members and those not owned by physicians (AMA 1999. productivity per worker will be lower under group incentive plans than under individual incentive arrangements because group incentive plans weaken the link between individual performance and compensation (e. Holmstrom and Milgrom 1991). my productivity measures (charges and income) better reﬂect service complexity than the 1978 survey measure (ofﬁce visits per hour). direct empirical evidence is sparse (Prendergast 1999). (1997) conduct analyses at the group-level using 905 and 488 groups. Encinosa et al. Huddart and Liang 2005). I also delete not-for-proﬁt and medical 7 Gaynor and Pauly (1990) and Encinosa et al. Responses were received from 1. RESEARCH DESIGN Sample The data for this study come from a nationwide survey developed and conducted by the Medical Group Management Association (MGMA 1999a).g. yielding an adjusted response rate of 30. Gaynor and Gertler (1995) analyze a subset of 1. (1997) ﬁnd that productivity in equal-sharing groups containing more than ﬁve members is signiﬁcantly lower than productivity in groups using individual plans.g. subspecialty. Che and Yoo 2001. Holmstrom 1982. or even higher. I use the individual physician as the unit of analysis instead of the group. The Accounting Review American Accounting Association January 2010 . MGMA staff eliminated 163 surveys due to incomplete data reporting. In particular. and more recent data set. as previously noted.230 individual physicians. productivity will be no different. 1).353 physicians in 957 medical group practices (Gaynor and Pauly 1990.
100% Straight Salary: The compensation for the provider is 100 percent attributable to a ﬁxed salary. One-hundred ﬁfty practices (16 percent) use group-based compensation. 100% Productivity: The provider’s compensation is based wholly on individual productivity measures (gross charges.). 50% to 99% Guaranteed Salary: Between 50 percent and 99 percent of the provider’s compensation is paid by a base or guaranteed amount. The remaining 785 practices (84 percent). productivity. such as experience. which have unique objectives. ‘‘100% Equal Shares. GROUP COMP. etc. encounters. capitation). The MGMA sample is also very comparable to the AMA sample with respect to the number of physicians in family practice (13 percent for both) and the internal medicine specialties (34 percent and 32 percent. The remainder of the compensation is a set amount. The ﬁnal sample consists of 11. The Accounting Review January 2010 American Accounting Association ..971 physicians in 935 practices. respectively). 2. and Tables 2 and 3 report descriptive statistics for group-level variables and physician-level variables. 3. select option 2. 50% to 99% Productivity: Between 50 percent and 99 percent is based upon individual physician productivity measures. respectively). which fall 8 The AMA (1999) survey provides the most comprehensive data set on physician group practices available for this time period. but average group size for the MGMA sample is larger (12. The indicator variable for group-based compensation. RVUs. However.’’ identiﬁes groups that divide 100 percent of net proceeds equally among all members. 1. If a provider is compensated 50 percent from productivity and 50 percent is guaranteed salary. equals 1 for these practices.Group-Based Compensation in Professional Service Firms 351 school faculty practices. The ﬁrst category. The unit of analysis for tests of H3 is the physician. respectively. The unit of analysis for tests of H1a–H1d and H2 is the group. 4.8 versus 10.g. respectively. as follows: Please select the compensation method that most closely resembles how each provider’s compensation is computed. If 50 percent of the provider’s compensation is from a guarantee salary and 50 percent is from productivity measures. bonus. MGMA sample practices are larger than those in the AMA sample. respectively). select option 2. are aggregated to the group level for these tests. the MGMA sample has relatively more physicians that practice a surgical specialty (34 percent and 25 percent. 50% to 99% Productivity.8 Sample data vary by group and by individual physician. physician-level variables.0 physicians. 100% Equal Shares: All providers in the practice are paid equal shares based upon overall practice performance. Measurement of Compensation Method Compensation method is determined from a survey question that asks respondents to indicate the basis of compensation for each practice member. Table 1 reports variable deﬁnitions. 5. The proportion of single-specialty groups in the MGMA sample (77 percent) is highly similar to that of AMA groups (78 percent). respectively). the balance of compensation is determined by other factors (e. Uppercase and lowercase letters denote group-level and physician-level variables. but comparable with respect to specialty composition. Median MGMA group size is similar to AMA group size (7 versus 6 physicians.
task interdependence. and 0 otherwise. research stipends. physicians per 100. number of different specialties in group divided by group size. indicator variable that equals 1 if 50–99 percent of a physician’s compensation is guaranteed salary and the remainder is based on individual productivity. 21–31. and 0 otherwise. 11–20.000. RURAL.000 in population in county. incentive payments. indicator variable that equals 1 if 100 percent of a physician compensation physician is guaranteed salary. indicator variables set equal to 1 if number of physicians in group falls within speciﬁed size range. 7–10. (continued on next page) The Accounting Review American Accounting Association January 2010 . intra-group standard deviation of gross charges.352 Pizzini TABLE 1 Variable Deﬁnitionsa Variables Dependent Variables GROUP COMP group comp 100% productivity 50%–99% productivity 50%–99% salary 100% salary AVG CHARGES STD CHARGES CHARGE VARIATION charges income Independent Variables INTERDEPEND RISK SPECIALTY DIVERSITY FTE DIVERSITY GENDER DIVERSITY 3–6. within group standard deviation of gross charges (full retail prices billed) divided by average gross charges for the group. total compensation. indicator variables that equal 1 if the medical group is located in a rural. Medicare malpractice geographic practice cost index. number of physicians in group. and URBAN % 3 YEARS Deﬁnitions indicator variable that equals 1 if a predominant number of group members share equally in the net proceeds of the group. or urban area. full retail prices of services billed by an individual physician. average of gross charges (full retail prices billed) for the group. indicator variable that equals 1 if physician’s entire compensation is based upon individual productivity. indicator variable that equals 1 if physician receives an equal share of the group’s net proceeds. Ln(1 percentage of group revenues derived from capitation contracts). natural log of the number of physician in the group. suburban. standard deviation of the proportion of full-timeequivalent units each group member devotes to clinical activities. managed care penetration by county. bonuses. and 31 PHYSICIANS SIZE LN SIZE CAPITATION MCO PENETRATION COMPETITION SUBURB. indicator variable that equals 1 if 50–99 percent of a physician’s compensation is based upon individual productivity. and proﬁt distributions for an individual physician. respectively. and 0 otherwise. percentage of physicians in group with fewer than three years of experience. including salary. proportion of group members belonging to the nondominant gender.
For example. 100% Straight Salary). MIDATLANTIC. CALIFORNIA. The survey also allowed respondents to indicate ‘‘other’’ if none of the ﬁve responses listed was appropriate. empirical tests using the full sample will be biased against the hypothesized predictions. indicator variable that equals 1 if physician has a subspecialty. 25 proportion of full-time equivalent units (FTEs) a physician devotes to clinical activities. 2007. The compensation method used with the greatest number of members determines the compensation method assigned to the group. but do not consider equal-sharing arrangements. etc. SOUTHEAST. into one of the last four categories. I estimate the model’s predicting compensation method both with and without observations with compensation plans classiﬁed as ‘‘50% to 99% Productivity’’ and ‘‘50% to 99% Guaranteed Salary. 50% to 99% Guaranteed Salary. 309–310). As respondents are instructed to choose the method that ‘‘most closely resembles how each provider’s compensation is computed.10 Wording of the compensation question raises some ambiguity regarding classiﬁcation if compensation plans incorporate both individual and group incentives. (2007) identify factors associated with the choice of individual compensation methods (100% Productivity. years male subspecialty a 353 Deﬁnitions indicator variables that equal 1 if group is located in speciﬁed region. sample groups used the same compensation method with 90 percent of their members. NORTH ATLANTIC. or more than 25 years. NORTHWEST. 3–5 years.’’ To the extent that plans incorporating sharing are included in these categories. Their empirical tests of survey response data indicate that compensation plans were correctly classiﬁed (Ittner et al. CALIFORNIA. indicator variables that equal 1 if physician has been practicing medicine for 2 or fewer years.’’ Thus. clinical FTE years 2 years. less than 1 percent of physicians were placed in this category. UPPER MIDWEST. years experience practicing medicine. Group-Level variables are in UPPERCASE letters and Physician-Level variables are in lowercase letters. 50% to 99% Productivity. 9 10 Ittner et al. and 0 otherwise. 3 to 5 years. sharing may also be used with providers classiﬁed as receiving ‘‘50% to 99% Guaranteed Salary. Medical groups often use different compensation schemes with different members (Lee 1990). and 0 otherwise. base compensation on individual performance (GROUP COMP 0). Similarly. indicator variable that equals 1 if physician is male. LOWER MIDWEST.’’ respondents likely will classify such a plan as ‘‘50% to 99% Productivity. UPPER MIDWEST is the omitted region in tests of the hypotheses. SOUTHWEST. ROCKY MOUNTAIN. and 0 otherwise. suppose 60 percent of compensation is based on individual productivity. On average. The Accounting Review January 2010 American Accounting Association . with the remainder equally split among all partners. respectively.9. providers classiﬁed as receiving ‘‘50% to 99% Productivity’’ may have a substantial portion (up to 49 percent) of their compensation based upon group performance. N .Group-Based Compensation in Professional Service Firms TABLE 1 (continued) Variables NORTHEAST. EASTERN MIDWEST. SO.’’ To address this concern.
11 13.834) (0.299) (0.81 9.000) (0.000) (0.91 0.05 7.001) M-Wb (0.86 0.45 0.00 7. 0.29 0.000) (0.25 0.044) (0.10 2.57 0.08 1 Std.79 0.00 2.10 12.79 0.53 0.62 GROUP COMP (n 785) Mean 2.00 7. 0.10 0.82 Difference between Partitions ta (0.45 0.31 0.00 8.32 0.13 21.95 0 Std.19 0.62 0.819) (0.26 0.00 1.18 0.584) (0.00 1.43 0.95 Std.920) (0.79 0.006) (0.79 GROUP COMP (n 150) Mean 2.00 7.85 0.04 0.80 0.83 0.819) (continued on next page) Pizzini .19 0.98 0.07 0.20 0.87 0.34 0.09 Median 2.33 0.13 0.13 0.01 0.13 22.80 2.23 0.030) (0.22 0.83 0. 0.445) (0.25 0.12 7.17 0.85 0.46 2.000) (0.001) (0.11 0.07 7.01 Median 2.53 0.008) (0.10 3. Partitioned by Compensation Method All Groups (n 935) Mean Task Interdependence INTERDEPEND Income Risk RISK Intra-Group Diversity SPECIALTY DIVERSITY FTE DIVERSITY EXPERIENCE DIVERSITY GENDER DIVERSITY Group Size SIZE LN SIZE 3–6 PHYSICIANS 7–10 PHYSICIANS 11–20 PHYSICIANS 21–31 PHYSICIANS 31 PHYSICIANS 2.02 0.The Accounting Review American Accounting Association January 2010 354 TABLE 2 Descriptive Statistics for Variables Predicting Use of Group-Based Compensation and Variation in Productivity for 935 Physician Groups.92 0.060) (0.04 0.184) (0. Dev.08 Median 2.02 0.45 0.68 0.66 0.47 0. Dev.10 3.40 0.06 7.39 2.00 7.31 0.251) (0.007) (0. Dev.25 0.07 7.18 0.03 0.
14 0. Variables are deﬁned in Table 1. Dev.046) (0.049) (0.26 0. 355 .27 0.124) (0.45 0.25 1 Std. Dev.28 0.33 0. Dev.09 0.58 0.18 Mean 10.009) Median 8.90 0. 13.11 0.00 0.77 0.25 0.34 0.20 0.49 0.00 0.00 0.036) (0.084) (0.08 0.26 0.23 0.26 0.66 1.050) (0.26 0.01 0.31 0.16 0.226) (0.50 0.01 0.33 0. Signiﬁcance of Mann-Whitney U-statistic testing whether samples of groups are drawn from same population.35 Median 8.25 11.16 0.000) January 2010 American Accounting Association Signiﬁcance of independent samples t-test for equality of means for groups using group-based compensation and those using individual compensation.001) (0.32 0.80 0.00 0.000) M-Wb (0.The Accounting Review All Groups (n 935) Mean Physician Experience 3 YEARS % Market Conditions CAPITATIONc MCO PENETRATION COMPETITION SUBURB RURAL Intra-Group Variation in Productivity CHARGE VARIATIONc a b Group-Based Compensation in Professional Service Firms TABLE 2 (continued) GROUP COMP (n 150) Std.60 0.29 0.31 0.042) (0. 13.81 1.20 (0. 13.29 0.25 Median 4.18 Difference between Partitions ta (0.17 0.23 0.18 GROUP COMP (n 785) Mean 11. c n for CAPITATION is 924 (147 group. 734 individual). n for CHARGE VARIATION is 865 (131 group.79 1. 777 individual).28 0 Std.
Continuity of care refers to the longitudinal interpersonal physician-patient relationship that transcends multiple illness episodes and includes responsibility for preventative care and care coordination (Saultz 2003). family practice) and least important in hospitalbased specialties (e. the amount of individual inﬂuence a physician exerts over his patient panel. Continuity of care is most important in primary care specialties (e. Measurement of the Determinants of Group Incentives Task Interdependence A physician’s specialty largely determines task interdependence. one physician cannot be easily substituted for another without compromising patient satisfaction and quality.e. The Appendix summarizes results of a survey that asked physicians to evaluate different specialties with respect to continuity. inﬂuence on patient panel. is calculated using the average of the individual physician-level task interdependence scores. Schwartz et al. A group-wide measure of task interdependence. Saultz 2003. 100% Productivity. with an average of the physician ratings for continuity. 100% productivity. an emergency room physician treats whoever comes to the emergency room during his/her shift.234 physicians in the sample. for example. and the lower the degree of task interdependence (Atlantic Information Services. Physicians often distinguish their services from those of their colleagues by building strong reputations that lead to referrals. Conversely. Larger values for interdepend indicate higher levels of task interdependence.356 Pizzini Physician-level tests of H3 incorporate all ﬁve compensation categories (100% Equal Shares. Thus. task interdependence is low. When continuity is important to care delivery. neurology. and psychiatry. which include most surgical specialties. Inc. These include dermatology. ‘‘non-controllable lifestyle’’ specialties. and 100% salary. I measure task interdependence. 2003. The more inﬂuence any one individual physician can exert on his/her patient panel. and control over work schedule.. the less dependent that physician is on the assistance of his colleagues to be successful. 2003. A physician’s control over the hours worked each week varies dramatically by specialty (Taylor 2003). Taylor 2003). Ophthalmologists. require physicians to work long and unpredictable hours.. 100% Straight Salary). 50%–99% salary. 50% to 99% Guaranteed Salary. (1990) refer to specialties that allow the physician to control the number of hours they work as ‘‘controllable lifestyle’’ specialties. Table 3 (Panel A) contains the distribution of compensation methods across the 11. and the extent to which a physician controls the timing and amount of work performed (Schwartz et al. sharing ‘‘call’’). actively advertise vision correction procedures to attract patients.g. 50% to 90% Productivity.g. panel inﬂuence. Saultz 2003). The Accounting Review American Accounting Association January 2010 . radiology). Conversely. Physicians in non-controllable lifestyle specialties can make their work schedules more manageable by coordinating patient care and covering for one another (i. 2002). ophthalmology. the corresponding physician-level indicator variables are: group comp. The results are highly consistent with the practitioner literature on specialties. Nutting et al. denoted interdepend. The extant literature suggests that task interdependence is inﬂuenced by three factors: the importance of care continuity in the physician-patient relationship. Some physicians attract patients by marketing their services and offering extended hours. ‘‘non-controllable lifestyle’’ specialties are characterized by high levels of task interdependence because group members rely heavily upon each other to perform their jobs. and.. and schedule control. thus. denoted INTERDEPEND. in which physicians typically treat patients on a nonrecurring basis (Nutting et al. 50%–99% productivity. 1989.
The Accounting Review January 2010 American Accounting Association Group-Based Compensation in Professional Service Firms TABLE 3 Physician-Level Descriptive Statistics for Compensation Method. Productivity.3% $714.00 12.705 2. Partitioned by Compensation Method All Physicians n 11.106 $173.094 3.2% $923. $1.094 Mean Median Std.624 $213.278) (0.5% 11.970 14.1% 93.67 7.008) (0.81 1.234 Mean Median Std.451 0.821 0.8% 85.1 11.3% 13. and Specialty Panel A: Distribution of Compensation Methods n group comp Individual Compensation Plans 100% productivity 50–99% productivity 50–99% salary 100% salary Total 1.00 $622.3 24.000) (0.00 $754.921 1.000) (0.140 Mean Median Std.000) (0.78 7.794 0.04 Tests for Equality of Partitions ta M-Wb (0.8% 8.862 $174.7% 12.0% 10.632 0.14 8.14 9.929 1. Dev.8 100.1% 84.000) (0.888 $249. Productivity chargesc income Physician Characteristics clinical FTE years 2 years 3–5 years 25 years male subspecialty $899.0 Panel B: Productivity Data and Individual Physician Characteristics.959 13.4% 11. Dev.650 $345. Dev.1 19.91 group comp 1 n 1.000) (0.000) (0.7 35. Physician Characteristics.234 Percentage 9.330 11.00 $745.465 $175.158 0.000 $305.6% 12.00 14.10 7.4% 8.211) (0. $876.000) (0.144 1.003) (0.52 group comp 0 n 10.958 13.4% $684.389 0.359 $258.107.000 $204.832 1.00 12.000) (0.000) (continued on next page) 357 .961 2.9% 13.
03 6.66 0.25 1.135 125 139 1.90 1.04 1.The Accounting Review American Accounting Association January 2010 358 TABLE 3 (continued) Panel C: Distribution of Physician Specialties Specialty Allergy Anesthesiology Cardiology Dermatology Emergency Family practice Gastroenterology Hematology / oncology Internal medicine Nephrology Neurology Nuclear medicine Obstetrics / gynecology Ophthalmology Orthopedic surgery Otolaryngology Pathology Pediatrics Physiatry Psychiatry Pulmonary medicine Radiology General surgery Specialty Groupings6 Other Hospital-based Internal medicine subspecialty Other Hospital-based Primary care Internal medicine subspecialty Internal medicine subspecialty Primary care Internal medicine subspecialty Other Hospital-based Primary care/ surgical Surgical Surgical Surgical Hospital-based Primary care Other Other Internal medicine subspecialty Hospital-based Surgical Frequency 116 690 1.11 1.278 220 80 731 74 65 197 416 471 Percentage 1.24 12.83 0.14 10.240 140 206 13 717 361 1.38 3.58 1.21 11.60 2.75 3.415 326 200 1.96 0.19 (continued on next page) Pizzini .51 0.70 4.38 1.78 11.12 6.11 1.71 6.
77 3.234 Percentage 1. b Signiﬁcance of Mann-Whitney U-statistic testing whether samples (group comp 0 and group comp 1) are drawn from same population. Variables are deﬁned in Table 1.00 Signiﬁcance of independent samples t-test for equality of means for physicians compensated based on group performance (group comp 1) and those compensated based on individual performance (group comp 0).94 0.12 1. n 10. c For charges.24 0.093 individual). January 2010 American Accounting Association 359 .074 (981 group.49 0. 9.26 100.The Accounting Review TABLE 3 (continued) Panel C: Distribution of Physician Specialties Specialty Cardiothoracic surgery Colon surgery Neurological surgery Plastic surgery Vascular surgery Urology Total a Group-Based Compensation in Professional Service Firms Specialty Groupings6 Surgical Surgical Surgical Surgical Surgical Surgical Frequency 218 14 139 55 87 366 11.
and the correlation between GENDER DIVERSITY and the percentage of females in the group will bias the results against H1c.12 All four measures of intra-group diversity are expected to be negatively associated with the use of group incentives. I measure intra-group diversity in clinical work (FTE DIVERSITY) using the within-group standard deviation of the amount of clinical fulltime-equivalent units (FTEs) reported for each group member (clinical FTE). I measure the risk of ﬁnancial loss arising from malpractice lawsuits (RISK) with Medicare’s 1999 Malpractice Geographic Practice Cost Index (MGPCI). 1999. Control Variables I control for a group’s propensity to use salary with new physicians by including a variable that measures the percentage of a group’s physicians that have been practicing 11 12 Specialty diversity is also computed using the sum of the squares of the proportion of group members in each specialty and the log of the number of different specialties in the group. clinical responsibility. Sunden and Surette 1998. Administrative responsibility. Table 3).11 Physicians in the same group sometimes devote different amounts of time to clinical activities. Levinson and Lurie 2004. including medical specialization. and gender. and continuing professional education reduce the time available to treat patients. If this is the case. I measure the relative number of males and females in a group (GENDER DIVERSITY) by subtracting the percentage of a group’s physicians that are of the same gender as the majority of physicians in the group from 1. Group size is also measured with the natural log of the number of physicians in a group (LN SIZE). When this variable equals its maximum of 1. These measures yield materially similar results in tests of the hypotheses. 11–15 PHYSICIANS. Bajtelsmit et al. 21–30 PHYSICIANS. 16– 31 PHYSICIANS. The Accounting Review American Accounting Association January 2010 . The omitted category represents practices with fewer than seven physicians. approach toward care delivery. Intra-group diversity in physician experience (EXPERIENCE DIVERSITY) is measured using the withingroup standard deviation of the number of years each group member has been practicing medicine (experience). The MGPCI reﬂects geographic variations in malpractice costs faced by physicians. 2002). teaching. These factors are likely to affect monitoring efﬁciency. The Health Care Financing Administration (HCFA) calculates the MGPCI from insurer premium and claim data (Task Order 0038. experience. and attitude toward compensation (Kletke et al. (2006). every member of the group is in a different specialty. Dwyer et al.g. HCFA 2000). Rizzo and Zeckhauser 2007). Powell and Ansic 1997. I include a measure of the relative mix of female and male physicians because research suggests that female physicians differ systematically from their male counterparts in their practice patterns. then women should prefer group incentives.360 Pizzini Malpractice Liability Following Evans et al.. 1990. equal 1 if the number of 20 PHYSICIANS. 7–10 PHYSICIANS. I measure intra-group diversity among specialties by dividing the number of different specialties in the group by the number of group members (SPECIALTY DIVERSITY). The following variables. relationships with patients. Group Size To accommodate potential changes in the direction of the relation between size and the use of team-based compensation. GENDER DIVERSITY is highly and positively correlated with the percentage of female group members. Physicians in the MGMA sample practice in one of 29 different specialties (Panel C. Intra-Group Diversity Physicians within the same group may differ on several factors that inﬂuence monitoring ability. Studies from the economics and psychology literatures ﬁnd that women are more risk-averse than men (e.. and 0 otherwise. group I measure size with ﬁve indicator variables (Kandel and Lazear 1992). and physicians in the group falls within the respective size category.
Evans et al. GROUP COMP 3 5 7 9 1 INTERDEPEND 2 RISK 4 SPECIALTY DIVERSITY EXPERIENCE DIVERSITY 7–10 PHYSICIANS 16–20 PHYSICIANS 31 PHYSICIANS CAPITATION COMPETITION i 14 8 FTE DIVERSITY 6 GENDER DIVERSITY 11–15 PHYSICIANS 10 21–30 PHYSICIANS 3 YEARS RURAL (1) 11 13 15 12 MCO PENETRATION 16 SUBURBAN 17 27 i 18 REGIONi. 2006): CAPITATION ln(1 percentage of revenues derived from capitation contracts) measures capitation. URBAN is the omitted variable.Group-Based Compensation in Professional Service Firms 361 medicine fewer than four years ( 3 YEARS). Leone.000 in population. and indicator variables representing 11 different regions of the country. physicians per 100. I measure intra-group variation in productivity (CHARGE VARIATION) by dividing the within-group standard deviation in gross charges by average charges. but the survey does speciﬁcally ask which physicians are partners. bad debts. contractual adjustments. SUBURBAN. the MGMA assumes that physicians with three or more years of experience are partners. are analyzed at the group level using the following cross-sectional logistic regression model. Physician Productivity To test H2. The indicator variable for the upper Midwest region is the omitted variable. total Health Maintenance Organization (HMO) enrollment per 10.13 I also control for the following marketrelated factors that potentially affect physician incentives (Hurley et al. All other market variables come from the MGMA survey. charitable adjustments. Hypotheses 1a–1d. which address determinants of group-based compensation. and other adjustments. A log transformation is used because the capitation percentage is skewed to the right. The Accounting Review January 2010 American Accounting Association . RURAL REGIONi Variables measuring managed care penetration and physician competition come from the Area Resource File. In reporting descriptive statistics. 1996. or rural metropolitan area. Control variables 13 The omitted category potentially controls for partnership status because medical groups typically do not admit new physicians into the partnership until after three years. Gross charges capture the time and intensity of physician services. 2002. indicator variables specifying whether the practice is located in an urban. I regress intra-group variation in productivity on the indicator variable for group-based compensation and a set of control variables. Gross charges are the full-dollar value of services provided at the practice’s established undiscounted rates before reduction by Medicare fee restrictions. suburban.000 in population. MCO PENETRATION COMPETITION URBAN.
size.929. and market conditions. The Accounting Review American Accounting Association January 2010 . The individual physician is the unit of analysis in tests of H3. Inclusion of net revenue reduced sample size by approximately 40 percent.14 Total compensation includes salary. Both variables are highly skewed to the right. malpractice risk. respectively) signiﬁcantly exceed the corresponding amounts for physicians with individual incentive plans ($684. but is eventually counteracted by increasing age (Gaynor and Pauly 1990. I included the natural log of net revenue as a dependent variable in robustness tests of H3. The results were highly consistent with the ﬁndings presented in the article. which range from 20 percent to 50 percent of total group revenues (MGMA 1999b). The data set does not provide information on adjustments at the physician level. 2 years. Prior research suggests that productivity improves with experience for most of a physician’s work life. net revenues (collections) are available at the group level for a subset of practices. and proﬁt distributions. which examines the performance effects of group-based incentives. The median value for this variable is 1. and thus reﬂect inter-group differences in income risk.362 Pizzini included in the OLS regression model (Equation (2)) that tests H2 are INTERDEPEND and the aforementioned measures of diversity. 3–5 years.000 and $305. To control for variation in income risk from adjustments. indicator variables that equal 1 if a physician has been practicing in her chosen specialty for the corresponding time period. respectively) (Table 3 Panel B). as most physicians in the sample spend all of their time on clinical activities. bonuses. contractual adjustments. and other adjustments. this control variable is not included in the ﬁnal models. Median charges and income for physicians with group incentive plans ($923. CHARGE VARIATION 3 5 7 9 1 GROUP COMP 4 2 INTERDEPEND RISK SPECIALTY DIVERSITY 6 FTE DIVERSITY CAPITATION COMPETITION RURAL 23 i 14 i EXPERIENCE DIVERSITY 8 GENDER DIVERSITY 10 LN SIZE MCO PENETRATION 12 11 13 SUBURBAN (2) REGIONi. The following variables control for the many factors outside the compensation contract that affect charges and income: clinical FTE the percentage of full-time-equivalent units (FTEs) that a physician devotes to clinical work. and 25 years 14 Gross charges do not include adjustments for Medicare fee restrictions. however.832. research stipends. Total adjustments vary widely among practices. therefore. I measure productivity with gross charges (charges) and total compensation per physician (total pay). I also include regressors for the natural logarithms of these variables (ln charges and ln total pay). therefore. reﬂecting effort in clinical and non-clinical tasks. Gross charges capture clinical productivity.000 and $204. incentive payments. Gaynor and Gertler 1995).
malpractice risk. Chi-squared statistics indicate that the model is a signiﬁcantly (p 0. or total pay. Correlations Tables 4 and 5 report Pearson correlations among group-level and physician-level variables. None of the correlations between group-level variables exceeds 0.26. The ﬁrst column is for the full sample of 935 groups. This variable controls for risk preferences if males are less risk-averse than females. RESULTS Determinants of Group-Based Compensation Table 6 results from logistic regressions examining the determinants of group-based compensation.30. indicator variable specifying which of 29 different specialties the physician practices (Table 3.’’ because these plans may include team incentives along with individual incentives. The Accounting Review January 2010 American Accounting Association . ln total pay. respectively) indicate reasonable explanatory power. and market conditions: productivity 3 6 9 39 41 1 100% productivity 4 2 50–99% productivity 5 50–99% salary 2 years subspecialty SUBURBAN 7 100% salary 8 clinical FTE 8 3–5 years 37 i 10 i 25 years 38 male specialtyi 40 52 i 43 CAPTITATION (3) MCO PENETRATION 42 COMPETITION i RURAL REGIONi. and 0 otherwise. and Nagelkerke R2 statistics (r 0.01) good ﬁt for both the full and reduced samples. twotailed) because capitation is chieﬂy used with primary care specialties. p 0. and compares groups that share 100 percent of their proceeds to those that use any of the four individual compensation plans. Only correlations among physician-level indicator variables for compensation method exceed 0. where productivity refers to ln charges. indicator variable that equals 1 if the physician has a sub-specialization.20 and 0.30.30. V.01. The omitted specialty is family practice. Panel C). INTERDEPEND is negatively correlated with CAPITATION (r 0. The second column pertains to a reduced sample that excludes groups with compensation plans classiﬁed as ‘‘50% to 99% Productivity’’ or ‘‘50% to 99% Guaranteed Salary. respectively. controls for potential economies or diseconomies of scope (Pauly 1996). and (deﬁned previously).Group-Based Compensation in Professional Service Firms 363 male specialtyi subspecialty SPECIALTY DIVERSITY indicator variable that equals 1 if the physician is male. charges. The OLS regression model used in tests of H3 (Equation (3)) also controls for size.
03) 0.00 0.30) 0.00) 0.09 (0.00) 0.10 (0.00) 0.07 (0.54) 0.07 (0.00) 1.00) 0.01 (0.11 (0.12 0.08 (0.28) 0.04) 1.06) 0.20 (0.52) 0.10 (0.11 (0.04 (0.11) 0.17 (0.02 (0.04 (0.00) (0.06) 0.02) 0. DIV.13 (0. DIV DIV.49) 1.00) 0.00 0.05 (0.00 (0.00 0.00) 1.94) 0.00) 0.03 (0.00 (0. GENDER 7–10 11–20 21–30 COMP INTERDEPEND DIV.00) 0.17) 0.00) 0.16 (0.26) 0.04) 0.00 (continued on next page) . PHYS PHYS 0.67) 0.03 (0.04 (0.04 (0.07 (0.00 (0.24 (0.01 (0.00) 31 PHYS LN SIZE RISK 3 MCO YRS CAPITATION PEN 364 January 2010 EXPERIENCE DIVERSITY GENDER DIVERSITY 7–10 PHYSICIANS 11–20 PHYSICIANS 21–30 PHYSICIANS 1.25) 1.00) 0.03 (0.04 (0.00 0.24) 1.02 (0.10 (0.20 (0.27 1.06 (0.31) 0.00 0.00) 0.83) 0. PHYS.01) 0.00) 0.00 0.10 (0.The Accounting Review American Accounting Association INTERDEPEND SPECIALTY DIVERSITY FTE DIVERSITY TABLE 4 Pearson Correlations among Group-Level Predictors of Compensation Method and Variation in Productivity GROUP SPECIALTY FTE EXPER.02 (0.
00) 0. PHYS PHYS 0.00 0.01 (0.09 (0.71) 0.23 (0.05 (0.06 (0.00 (0.00) 0. DIV.09 (0.11 (0.65) 1.00 0.00) 0.04 (0.38) 0.66) 1.31) 1.36) 0.01 (0.06 (0.04 (0.19 (0.10 (0.07) 0.81) 0.30 (0.70 (0.71) 0.54) 0.10 (0.25) 0.00 0.00) 0.00) 0.01 (0. GENDER 7–10 11–20 21–30 COMP INTERDEPEND DIV.00) 0.04 (0.19 (0.38) 0.04) 0.10 (0.00) 0.01 (0.00) 0.00) 0.00 0.03 (0.03 (0.The Accounting Review 31 PHYSICIANS LN SIZE RISK 3 YEARS CAPITATION MCO PENETRATION COMPETITION 0.11 (0.03 (0.55) 0.06 (0. correlations with CAPITATION are based on 924 groups due to missing data.07 (0.03 (0.30) 0.01 (0.29 (0.22 (0.00 0.07 (0.00) 0.04 (0.00) 0.06 (0.02 (0.00) 0.02 (0.07 (0.00 (0.02) 0.04 (0.22) 0.23) 0.30 (0.01 (0.04 (0.04 (0.00) 0.04 (0.09) 1. PHYS. .13 (0.85) 0.00) 0.01) 0. Variables are deﬁned in Table 1.13) 0.98) 0.37) 0.27) 0.04 (0.07 (0.56) 0.01 (0.00) 0.05 (0. DIV DIV.00) 0.11) 0.14 (0.29 (0.06) 0.00) 0.03 (0.04) 0.03 (0.20) 0.07) 0.00 (0.22 (0.21 (0.03 (0.04) 0.00) 0.08 (0.27 (0.00) 3 MCO YRS CAPITATION PEN 365 January 2010 American Accounting Association Sample consists of 935 groups.04 (0.10 (0.04) 0.00) 0.08) 0.00) 0.00 0.07 (0.02 (0.47) 0.12 (0.01) 0.94) 0.26) 0.00) 0.13 (0.18) 0.00) 0.57) 0.00) 0.15 (0.02 (0.00) 0.11 (0.08 (0.04 (0.20 (0.75) 0.03) 1.27) 0.29) 0. Two-tailed p-values are in parentheses.43) 0.17) 0.07 (0.05) 0.04 (0.00 (0.22) 0.16 (0.06 (0.00 (1.00) 0.04) 0.00) TABLE 4 (continued) GROUP SPECIALTY FTE EXPER.94) 0.21) 0.00) 0.04 (0.92) 31 PHYS LN SIZE RISK 1.05 (0.17 (0.06 (0.13 (0.01) 0.28) 0.12) 0.02 (0.00) 0.
27 (0.10 (0.18 (0.04) 1.00) 0.05 (0. Pizzini .06 (0.24 (0.02 (0.02 (0.00 0.00 0.00) 0.00) 0.00) 0.01 (0.05 (0.00) 0.02 (0.00) 0.03 (0. Two-tailed p-values are in parentheses.14 (0.00) 0.12 (0. Variables are deﬁned in Table 1.03) 0.20 (0.00) 0.00) 0.00) 0.00) 0.04) 0.00) 0.01 (0.00) 50–99% salary 50–99% productivity 100% productivity Clinical FTE 2 years 3–5 years 25 years male Sample consists of 11.00) 0.00) 0.00 0.01) 0.13 (0.00 0.02 (0.00) 0.09 (0.00 0.00) 0.00) 0.02 (0.36 (0.00 (0.00) 0.03 (0.00) 0.01) 0.00) 0.00) 0.00 (0.06 (0.01 (0.06 (0.04 (0.02 (0.19 (0.05 (0.00 0.12 (0.00) 0.00) 0.12) 0.21 (0.03 (0.03) 1.The Accounting Review American Accounting Association January 2010 366 TABLE 5 Pearson Correlations among Physician-Level Predictors of Physician Productivity group comp 100% salary 50–99% salary 50–99% productivity 100% productivity clinical FTE 2 years 3–5 years 25 years male subspecialty 0.02) 0.00) 0.00) 0.16 (0.00) 0.11 (0.71) 1.00 (0.28) 100% salary 1.01 (0.19) 0.00) 0.234 physicians.00 0.70) 1.04 (0.10 (0.02 (0.39) 1.76) 1.00) 0.02 (0.00) 0.06) 1.07 (0.07 (0.05) 0.00 0.10 (0.07 (0.00 0.42 (0.00) 0.00) 0.03 (0.00) 0.10 (0.12 (0.00) 0.00) 0.07 (0.05 (0.00) 0.10 (0.27 (0.00) 0.04 (0.01) 1.01) 0.
31 0.15 ( 0.47) 3.25) 0.90 0.12 (0.73 ( COMPETE SUBURB ( RURAL ( ROCKY MOUNTAIN ( NORTHWEST ( ( 1.92 1.31 1.17) 0.00 (0.67 0.03 1.09 0.51) 2.45) 1.49) 1. Litigation Risk.09*** ( 2.36*** ( 4.08*** ( 3.92** 2.89) 0.92 0.04) 0.67** ( 1.18) 0.19 1.35 0.23) 3.21 0. Intra-Group Diversity.00) 0.82 (1.12 1.76) 3.41 0.24** 2.10*** ( 2.35 (0.02** 2.49) 0.74 0.29) 0.94) 1.50 1.00 (0.29 Variables Constant Task Interdependence INTERDEPEND Risk RISK Heterogeneity SPECIALTY DIVERSITY FTE DIVERSITY EXPERIENCE DIVERSITY GENDER DIVERSITY Size 7–10 PHYSICIANS 11–20 PHYSICIANS 21–30 PHYSICIANS 31 PHYSICIANS Control Variables 3 YEARS % MCO PENETRATION Predicted Sign Full Sample Coefﬁcient Odds (z-statistic) Ratio 1.60** ( 1.31** 2.36) 0.48) 1.40* 1.69) 2.74 ( 0.55*** 2.30 (0.70) 1.12 (0.76) 0.63) 0.03) 0.81 0.29** (2.03** 2.48) 1.09 (0.01*** ( 2.27 ( 0.55** (1.00 2.11) 0.20 1.21) 0.12 (0.Group-Based Compensation in Professional Service Firms 367 TABLE 6 Group-Level Logistic Regression of Compensation Method (GROUP COMP) on Proxies for Task Interdependence.08 ( 0.36 1.20) 0.21 ( 0.17 ( 1.11*** ( 2.13 0.12 0.36 ( (continued on next page) The Accounting Review January 2010 American Accounting Association .13 0.00 0.94) 0.00 (0.31 1.68) 2.64 1.00 0.44) 0.44) 1.12 1.42*** (3.40*** (4.61) 0. and Size Exclude Groups Using 50%–99% Guaranteed Salary or 50%–99% Productivity Coefﬁcient Odds (z-statistic) Ratio 1.36 0.43) 0.02 1.53 2.73) 0.29) 0.13 1.27 0.05 0.53) 1.16) 0.86*** (2.
047). one-tailed) less likely to adopt equal-sharing plans 15 I tested the robustness of the results with alternative speciﬁcations of the model in Tables 6 (Equation (1)). Only statistically signiﬁcant regional control variables are reported in the table.01.59 ( 1. Including this control variable reduced the number of observations to 924. **. while the coefﬁcient on FTE DIVERSITY is not signiﬁcant. Finally. The untabulated results were materially consistent with those contained in Table 6.05) 416 150 566 114.01. the results are highly consistent with H1a–H1d.368 TABLE 6 (continued) Pizzini Variables SOUTHWEST EASTERN MIDWEST LOWER MIDWEST n (GROUP COMP n (GROUP COMP n (all groups) Chi-Square Statistic (signiﬁcance) Nagelkerke R2 Correct Classiﬁcation 0) 1) Predicted Sign Full Sample Coefﬁcient Odds (z-statistic) Ratio 1.44*** ( 3. three of the four proxies for intra-group diversity support the prediction that more diverse groups are less likely to adopt team-based compensation plans.842 0.96** ( 2.03) 1. INTERDEPEND is signiﬁcantly and positively associated with the use of team0. possibly due to the lack of variation in this variable.202 0.03) 0. Variables are deﬁned in Table 1.74) 0.1 (0. suggesting that physicians in more taskbased compensation (p interdependent specialties are more likely to use team incentives. In both samples.001) 0. The positive and signif0. therefore.53 (0. I also re-estimated the model using states as control variables in place of the regional control variables because tort reform regulation and liability payment limits vary by state.10) and the signiﬁcance dropped slightly (from p 0. it is not included in the ﬁnal model. and 1 percent levels. p 0.55 to 1.35 0. 5 percent. one-tailed).05 and p 0.04** ( 2. Overall.01.38 *. Specialty.039 to p 0.41) 785 150 935 117. The size of the coefﬁcient for RISK increased (from 0. controlling for capitation (CAPITATION) did not materially affect the results and capitation revenues were not signiﬁcantly associated with compensation method.73* ( 1. Inclusion of state control variables did not signiﬁcantly improve the model’s ﬁt.24 0. icant coefﬁcient on RISK in both the full and reduced samples (p respectively. *** Indicate that coefﬁcient is signiﬁcance at the 10 percent. Results regarding group size are consistent with H1d.28 0. one-tailed) associated with the use of (p group-based compensation. and p 0. respectively (one-tailed where coefﬁcient sign has prediction. The Accounting Review American Accounting Association January 2010 . in that groups with more than 30 members are signiﬁcantly (p 0.001) 0. experience and gender diversity are negatively and signiﬁcantly 0.266 0.05.01. I re-estimated the model after replacing INTERDEPEND with a set of 17 indicator variables representing a group’s specialty.769 0.20) 0.48 0.15 In both the full and reduced sample.01.27*** ( 3. respectively. two-tailed otherwise).55 Exclude Groups Using 50%–99% Guaranteed Salary or 50%–99% Productivity Coefﬁcient Odds (z-statistic) Ratio 1. one-tailed) indicates that groups facing greater malpractice risk are more likely to use team incentives.
in turn. 315) argue that groups generally choose location and specialty or specialties upon formation. Physicians who are compensated based on individual clinical productivity clearly have a very strong incentive to devote time purely to clinical 16 17 18 As income risk increases. who ﬁnd that medical groups are signiﬁcantly less likely to use equal-sharing arrangements as group size increases beyond ﬁve members.18 Physician experience is clearly predetermined. One regressor variable that is not clearly exogenous or predetermined is FTE DIVERSITY. 2003). The difference is likely attributable to substantial increases in income risk and signiﬁcant reductions in monitoring costs that have taken place during the 20 years that have elapsed between the time the two surveys were conducted (Mauney 2002. INTERDEPEND controls for decreasing marginal returns to size. and there is no evidence that physicians select into compensation arrangements based upon their own experience or the experience levels of other group members (Encinosa et al. Moreover. Gaynor and Pauly 1990). the signiﬁcant positive correlation between GENDER DIVERSITY and percentage of women in a group will bias tests against showing support for H1c. To the extent that decreasing marginal beneﬁts are associated with the absolute size of cooperative beneﬁts.16 Adoption of equal-sharing plans by groups with fewer than 30 members is consistent with agency-based predictions that free-riding and constraints on monitoring capabilities limit the use of group incentives in very large groups. 2003). These results differ from those of Encinosa et al. However. If marginal beneﬁts to cooperation drive the use of equal-sharing. they will seek out practices with equalsharing arrangements upon entering the workforce. and most operate for many years. However. group size. the point in this sample at which medical groups are signiﬁcantly less likely to use team incentives. The Accounting Review January 2010 American Accounting Association . An alternative explanation is that the marginal beneﬁts of cooperation decrease as group size increases. the substantial relocation costs. it is plausible that compensation method inﬂuences the amount of time each physician devotes to clinical activity. legal environment. (2007. However. It is unlikely that groups simultaneously choose compensation plans and locations upon formation because there are signiﬁcant barriers to the formation of a new group in a new location. Prior research suggests this is the case. and a physician’s ownership stake in a medical group. Although endogeneity is not likely to be a concern with FTE DIVERSITY. Ittner et al. it seems unlikely that these beneﬁts would decline at a size that is signiﬁcantly larger than the median group size. and management practices (Pauly 1996. the estimation results presented in Table 6 will potentially be confounded if the regressor variables are endogenous choices by the practice. 1997. Substantial improvements in medical billing and information technology have made it easier to monitor a greater number of physicians. Movement of physicians to new practices is limited by non-compete clauses. determine the group’s competitive market. physicians generally join established groups or take over practices of solo practitioners who are retiring in order to gain access to an existing patient base and existing administrative systems. It is plausible that if women are more risk-averse than men. Physicians cannot move their patients with them. then endogeneity should not affect the results. group size must increase correspondingly to provide the same level of insurance in 1998 that existed in a smaller group in 1978. Kessler and McClellan 2000. if the regressor variables are exogenous or predetermined at the time the practice selects a compensation plan. as previously discussed. thereby reducing the impetus for groups to use team incentives to induce cooperation. Location and specialty. Casalino et al.Group-Based Compensation in Professional Service Firms 369 than those with fewer than seven members. the group must incur the normal relocation costs of any business. (1997). Casalino et al. median group sizes for all specialties are well below 30 physicians. which measures the variation in the proportion of FTE units each group member devotes to clinical activity. At the same time. group size has no signiﬁcant effect on the choice of compensation plan in groups containing 30 or fewer members. Upon entering the workforce.17 As with most cross-sectional compensation studies. Leone 2002.
Finally. the ﬁxed-effects model fails to capture the effect of grouplevel variables. The Accounting Review American Accounting Association January 2010 . Total $26. Unfortunately. and other adjustments. which does not assume that observations within the same medical group are independent. The INTERDEPEND measure also controls for the adoption of standardized operating policies if specialties with greater task interdependence are more likely to adopt standardized procedures. R2 statistics are 0. The models using charges and total pay as the dependent variables predict that salaried physicians and those who receive 50 to 99 percent of their compensation as salary earn $66. I calculate standard errors using a robust variance estimator. the relation between total pay and compensation plan depends upon the size of the productivity component in the individual plan. as opposed to teaching residents or performing administrative tasks. many of which are signiﬁcantly associated with productivity and compensation method. Group-Based Compensation and Intra-Firm Variation in Productivity Table 7 reports ordinary least-squares estimates for tests of H2.208 less. There is no signiﬁcant difference between the amount of charges generated by physicians in group-based compensation plans and those in individual compensation plans of any type. two-tailed).19 This estimator assumes only that observations from different groups are independent of each other. intragroup variation in gross charges is signiﬁcantly lower in practices that use equal-sharing arrangements when compared with those that use individual compensation plans (p 0. in addition. practices with relatively poor patient panels should prefer equal-sharing plans to reduce each individual physician’s exposure to ﬁnancial risk. Together. and the random-effects models indicate that physicians in equal-sharing arrangements are at least as productive as those in individual plans. Observations from physicians in the same group are unlikely to be independent. FTE DIVERSITY is not signiﬁcantly associated with group-based compensation. practices with higher concentrations of uninsured and Medicaid patients are subject to greater ﬁnancial risk than those with wealthier patient panels. the data do not permit me to distinguish whether this result is attributable to selection and/or incentive effects. respectively. Including adjustments in the model (Table 8) reduces the sample size from 935 to 699 groups. bad debts. The random-effects model includes all group-level variables.370 Pizzini activities. In untabulated robustness tests.57 for ln charges and ln total pay. Group-Based Compensation and Productivity Table 8 reports regression estimates for tests of H3. two-tailed).01. but does not materially alter the results or conclusion.01. However. Measures of homogeneity and size are included in the model to help control for the relation between intra-group homogeneity and the use of standardized procedures. Adjustments equal the difference between gross charges (amounts billed) and net revenues (amounts collected) and include reductions for Medicare fee restrictions. However. Results indicated that physicians in equal-sharing arrangements generate signiﬁcantly more charges than those on individual plans with salary components that exceed 50 percent. the OLS. Moreover.983 and 0. I also estimated the relation between compensation and productivity with ﬁxed-effects and random-effects models. However. they earned signiﬁcantly more income than physicians in individual plans. respectively. if homogenous groups tend to choose equal-sharing rules and to adopt standardized operating procedures. intra-group homogeneity will lead to lower productivity variances. I control for clientele wealth effects with a group-level measure of total adjustments to gross charges. Therefore. the ﬁxed-effects. Accordingly. 254–258). however. a Hausman test of the coefﬁcients suggests that the model is misspeciﬁed. contractual adjustments.60 and 0. than those in equal-sharing plans (p 19 The robust variance estimator in Table 8 is an extension of the standard Huber (1967) / White (1980) / sandwich estimator of variance and corrects for clustering of observations by group (StataCorp 2001. As predicted. Fewer adjustments suggest less ﬁnancial risk.
the data do not permit me to distinguish between the effects of cooperation and mutual monitoring on productivity.63) 0. two-tailed otherwise). 5 percent. Taken together. The Accounting Review January 2010 American Accounting Association .05 (1.02*** ( 3.Group-Based Compensation in Professional Service Firms 371 TABLE 7 Group-Level OLS Regression of the Intra-Group Variation in Charges (CHARGE VARIATION) on Use of Group-Based Compensation and Controls Variables Variables Constant GROUP COMP Control Variables INTERDEPEND SPECIALTY DIVERSITY FTE DIVERSITY EXPERIENCE DIVERSITY GENDER DIVERSITY LN SIZE RISK CAPITATION MCO PENETRATION COMPETITION NO.414 *. respectively.15) 0. Second.27*** (8. CALIFORNIA n (groups) F (signiﬁcance) Adjusted R2 Predicted Sign Coefﬁcient (t-statistic) 0.72) 0.00 ( 0. (one-tailed where coefﬁcient sign has prediction.000) 0. thereby generating productive beneﬁts sufﬁcient to offset losses from free-riding. First.02 ( 0.02*** ( 3.06*** ( 3. ﬁrms use team incentives to induce cooperation. **.90) 856 27.14) 0. group-based compensation arrangements in homogenous groups encourage a level of low-cost mutual monitoring that is sufﬁcient to curb free-riding.47) 0.19) 0. pay does not differ between physicians in equal-sharing arrangements and those in individual plans with a productivity component of at least 50 percent.45) 0.05) 0.01*** (3.09* (1.05) 0.68) 0. Only signiﬁcant control variables for metropolitan area and geographic region are reported in the table. when group tasks are highly interdependent.00 (0. Unfortunately. Variables are deﬁned in Table 1.09) 0. *** Indicate that coefﬁcient is signiﬁcance at the 10 percent. the results on the determinants and performance effects of group incentives are consistent with two possible agency-based explanations.10*** (13. and 1 percent levels.05 ( 1.33) 0.70*** (12.23 (0.
136*** ( 2.04*** ( 3.62) 0.60) 0.41) 0.140*** (11.05 ( 0.15*** (4.45) 0.540 (0.27*** ( 7.15*** (10.39) 29.601* (1.016*** (9.40) 0.418*** ( 4.33) 67.81) 27.01 (0.14) 190.03) 0.941 ( 1.33*** ( 14.07) 0.633*** ( 3.28** ( 3.45) 0.08 ( 1.07** (2.01 ( 0.88) 37.074 853 charges Coefﬁcient (t-statistic) 502.60) 0.50) 0.21) 1.58) 232.46) 0.01 ( 0.42) 40.05*** ( 4.62) 78.822** ( 2.23*** ( 8.208* ( 1.19) 0.10*** (5.72) 0.11** (2.674 (0.70) 198.043*** ( 4.21) 15.08) 0. Group Characteristics.936*** (15.78) 26.019** ( 2.24) 0.21) 11.37) 51.07 ( 1.24) 0.866 ( 0.03 ( 0.10** ( 2.22) 0.222*** ( 10.41) 0.38) 0.549*** (5.90) 58.84) 11.60) 0.12*** (4.07) 100.762*** (11.672*** ( 5.49) 14.92) 0.30) 3.983*** ( 4.85) 65.00 ( 0.21*** (186.70) 29.074 (1.18*** ( 10.234 ( 1.32) 82.02) 86.31) 34.37*** ( 5.62) 894.074 853 ln total pay Coefﬁcient (t-statistic) 12.00 (0.98*** (25.03*** (168.81) 11.23*** ( 11.74) 0.247* ( 1.04) 0.556** ( 2.490 (0.64) 11.17) 58.23) 235.02 (0.27) 5.05 ( 1.89) 104.68) 30.05) 10.92) 5.234 919 Variables Constant Compensation Method 100% productivity 50–99% productivity 50–99% salary 100% salary Control Variablesb Physician Characteristics clinical FTE 2 years 3–5 years 25 years male subspecialty Group Characteristics SPECIALTY DIVERSITY LN SIZE Market Conditions RISK CAPITATION MCO PENETRATION COMPETITION RURAL SUBURB n (physicians) n (groups) (continued on next page) The Accounting Review American Accounting Association January 2010 .04 ( 1.28) 0.12) 0.23) 0.260*** ( 9.011*** (5.92) 194.60*** (19.11** ( 2.295 ( 0.91) 0.708 (0.73) 15.234 919 total pay Coefﬁcient (t-statistic) 367.01 (0.41) 0.05) 46.05) 0.71) 0.426*** ( 9.675 ( 1.59) 0.08** ( 2.06** ( 3.037 (0.02** ( 2.372 Pizzini TABLE 8 Physician-Level Regression of Gross Charges and Total Pay on Compensation Method and Control Variables for Physician Characteristics.66) 0.18) 39.965 ( 0. and Market Conditions with Robust Standard Errors and Clustering on Physician Groupa ln charges Coefﬁcient (t-statistic) 13.442** (2.159 (1.58) 1.944*** (4.15*** ( 10.49) 10.12) 38.17) 0.06) 0.12) 53.208 ( 0.983*** ( 12.
Encinosa et al.600 charges Coefﬁcient (t-statistic) 44. **. However. and (2) individuals’ The Accounting Review January 2010 American Accounting Association . the level of task interdependence in health care delivery has increased and improvements in information technology have lowered monitoring costs (Casalino et al.000) 0. G1. 1997). coefﬁcients for region and specialty are not reported in the table. (1997) found that equal-sharing groups with more than ﬁve members averaged signiﬁcantly fewer patient visits than groups that did not use equal-sharing. and size in a ﬁrm’s choice of compensation contract. Differences in the specialty composition of the samples may also contribute to the contrasting results. M v ˆ ˆ V k 1 (G) (G) ˆ uk uk V where: ˆ V ( 2 lnL / lnB2) uj. both task interdependence and homogeneity are positively associated with the use of team incentives.000) 0. and.88 (0. Encinosa et al. VI.25 (0.53 (0. empirical evidence on (1) the role of task interdependence. CONCLUSION This study makes several important contributions to the extant literature by providing large-sample. Gaynor and Gertler 1995. *** Indicate that coefﬁcient is signiﬁcance at the 10 percent. Levinson and Lurie 2004). that are independent. Encinosa et al.429 *. extant research shows mixed support for the performance effects of intra-group homogeneity (see Van Knippenberg and Schippers  for a review). 2003.71 (0.... Substantial changes in the organization of medical group practices and the health care environment likely account for the difference in results. The results are also consistent with the social categorization perspective of diversity. 1 and: u(G) k j Gk However. which holds that groups function more smoothly when they are homogeneous (Williams and O’Reilly 1998). My results differ from prior studies. In particular.. which used a 1978 data set of physician group practices (Gaynor and Pauly 1990. Panel C). respectively. mutual monitoring. and 1 percent levels. a Standard errors are calculated using a robust variance estimator that does not treat all observations as independent. it controls individual physician specialty by including 28 different indicator variables (Table 3.GM. controlling for diversity in the robustness tests of the productivity regressions does not materially affect the results.Group-Based Compensation in Professional Service Firms TABLE 8 (continued) ln charges Coefﬁcient (t-statistic) 81. 5 percent.361 ln total pay Coefﬁcient (t-statistic) 75. (two-tailed). The formula for this calculation is below: b The regression model controls for region by including 11 indicator variables.’s (1997) sample does not include hospital-based specialties. income risk.000) 0. while I ﬁnd no productivity differences in groups that have as many as 30 members. However. G2. Furthermore. which tend to practice in relatively large groups and use group incentives. suggesting that both cooperation and monitoring affect productivity.571 373 Variables F (signiﬁcance) Adjusted R2 total pay Coefﬁcient (t-statistic) 35. but does assume that observations can be divided into M groups.000) 0.
First. In particular.. Holmstrom and Milgrom 1991. I ﬁnd that physicians whose compensation is based upon group performance are not less productive than those whose compensation is based upon individual performance. Third. as opposed to group. Accordingly. but not causality. who ﬁnd that legal risk inﬂuences physician contracts with MCOs. and more comprehensive data set than that used in prior research. and the quality and quantity of inputs (e.g. the ﬁndings indicate that physician ﬁrms choose group-based compensation when team incentives are effective in inducing levels of cooperation and mutual monitoring that are sufﬁcient to offset and/or prevent freeriding. as opposed to external monitoring costs. this study measures mutual monitoring costs. and future capital gains from the sale of partnership stakes. Keating 1997). Wageman 1995. professionals are typically organized into labor-managed ﬁrms.. and accounting. This study is subject to a number of limitations. I have included control variables to address competing explanations. Finally. Wageman and Baker 1997). but none has found empirical support for the premise that ﬁrms design incentive systems to match task interdependency. more recent. Prior studies have established technological differences across work groups (e. and group size in determining the extent to which they use group incentives to compensate employees. In contrast to prior research. Furthermore. The current results also complement Evans et al. age.. but I The Accounting Review American Accounting Association January 2010 . Although corporations differ fundamentally from labor-managed ﬁrms in governance and organization. insurance. Second. Findings are consistent with the premise that ﬁrms choose equal-sharing arrangements when homogeneity in training. perquisites. (2006). While my results are consistent with agency theory. which share key organizational characteristics with medical practices. Arya et al. Consistent with recent theoretical research (e.g. These results are also of interest to managers and owners of corporations who increasingly rely upon employees to work in teams to accomplish goals. the study provides important evidence on the performance effects of group incentives by examining individual. ﬁrms that use group incentives exhibit signiﬁcantly less variation in productivity than those that use individual plans. As with prior compensation research. they are also consistent with alternative explanations. Crosssectional studies such as this can establish associations.374 Pizzini responses to group incentives. Lang and Gordon 1995). Consistent with prior research. this study provides empirical support for research that models partnerships as insurance devices and thus helps explain the existence of limited liability partnerships (Gaynor and Gertler 1995. I develop measures of task interdependence for different physician specialties and ﬁnd that differences in task interdependence are associated with differences in compensation plans in a manner consistent with agency theory. consulting. mutual monitoring effectiveness. output. These results are also related to accounting research that examines the inﬂuence of organizational interdependencies on compensation contract design (Bushman et al. 1997. hours) are easily monitored in many professional service ﬁrms (Hansmann 1996). this study provides a new perspective on the performance effects of group incentives in labormanaged ﬁrms by using a larger. the data exclude a number of direct and indirect compensation sources. such as pensions.g. The results are potentially useful for other professional ﬁrms. and most perform work that varies in the level of task interdependence. and relates these costs to contract design. Che and Yoo 2001). such as law. all professionals generally make signiﬁcant investments in their own human capital. Where possible. compensation committees and managers should still evaluate task interdependence. and gender facilitates mutual monitoring by group members. 1995.
internal medicine subspecialty. consistent with Nutting et al. ophthalmologists. Physicians rated each specialty (using a scale of 1 to 5) on the following three dimensions: (1) the importance of maintaining continuity in care (denoted continuity). the physician survey respondents rated continuity of care to be most important in the primary care specialties (means range from 1. dermatologists. (1990). and (3) the extent to which a physician controls the timing and amount of service provided (denoted schedule control). (2) the amount of individual inﬂuence a physician exerts over his patient panel (denoted panel inﬂuence). (2003) and Saultz (2003). surgical.Group-Based Compensation in Professional Service Firms 375 cannot completely rule them out. APPENDIX DETERMINATION OF TASK INTERDEPENDENCE MEASURE The task interdependence measure (denoted interdepend) is based on a survey of 18 physicians in executive education programs at The University of Texas at Dallas. For example. Moreover. Physicians rated continuity to be least important for hospital-based specialists (means range from 4. analysis of professionals in the single ﬁeld of medicine limits the ability to generalize to other professions and work settings. and other.73). The variable interdepend is an average of the three reverse-coded responses for continuity.17 to 1. and psychiatrists exert the greatest inﬂuence on their patient panels and have the greatest control over their work schedules. higher ratings indicate greater beneﬁt from intra-group cooperation and teamwork. panel inﬂuence. The results are highly consistent with the practitioner literature on medical specialties.67 to 5. Finally. As predicted by Schwartz et al. Responses to each survey question have been reverse-coded so that lower numbers indicate that continuity is of greater importance and that physicians have greater patient inﬂuence and greater schedule control. the study expands the literature on group incentives by providing large-sample empirical tests of agency-based predictions. survey respondents indicated that plastic surgeons. who typically treat patients on a nonrecurring and urgent basis (Taylor 2003). Despite these limitations.00). hospital-based. Table A-1 contains the survey results. such as mutual monitoring and freeriding. Hence. and schedule control. nor can I distinguish between the effects of incentives and selection on performance. I am unable to distinguish between the performance effects of different agency-based constructs. Specialties in Table A-1 have been grouped into one of ﬁve major medical categories: primary care. The Accounting Review January 2010 American Accounting Association . Future studies using panel data would enable researchers to investigate the performance effects of introducing and discontinuing group-based compensation plans.
58*** 3.64** 3.43 3 3 3.17*** 3.43 2.76*** 3.07*** 2.79*** 1.34*** 3.40 1.93*** 2.86 3.06*** 2.86*** 3.00 3.5 4.00 2.57*** 3.5*** 4* 3 4* 4 3 1.67*** 3.40 2.71*** 2.07 2.98 2.5 3** 2.12 2.The Accounting Review American Accounting Association January 2010 376 TABLE A-1 Mean and Median Responses to Physician Survey Measuring Task Interdependencea.67*** 3.67*** 3.48** 2.21*** 3.50* 2.93*** 2.5*** 3*** 3*** 3*** 3*** 3** 2 2.00 2.67*** 4.29 2.51*** 3.56*** 3.00*** 2.80*** 5.5** 2 1** 1*** 1.93*** 3.43 2 2 2 3** 4*** 5*** 4.17 1.57*** 3.00*** 2.71*** 1.86** 1.90*** 2.b Continuity Mean Median 1.24*** 3.67*** 3.78*** 3.14 4.30*** 3.33*** 3.5*** 4.5*** 4*** 3*** 3*** 3*** Panel Inﬂuence Mean Median 2.67*** n-groups Primary Care Family Practice Internal Medicine Pediatrics Obstetrics-Gynecology Hospital-Based Anesthesiology Emergency Medicine Pathology Radiology Surgical General Surgery Vascular Surgery Cardiothoracic Surgery Neurological Surgery Orthopedic Surgery Plastic Surgery Ophthalmology Otolaryngology Urology 129 89 34 66 43 5 7 27 32 7 21 22 151 6 37 22 39 (continued on next page) Pizzini .79*** 2.43*** 3.57*** 3.93*** 2.67 2.37*** 1 1 1 2* 5*** 5*** 5*** 5*** 3*** 4*** 4*** 4*** 3.64*** 2.73* 4.5 Schedule Control Mean Median 2.86 2.47*** 3.93 1.20 1.11*** 3.14 1.57* 3.04 2.47* 3.33*** 3.33*** 3.57*** 4.33* 2.93*** 2.30*** 3.86** 1.00*** 4.5 Interdependc Mean Median 1.36*** 1.37*** 3.33*** 3.33*** 3.17*** 3.33*** 3.50 2.20 2.5** 2*** 2 2.21 3.
67** 2.20** 1.5 3** 3** 1.00 2.89 2.5 2 Schedule Control Mean Median 3. respectively for tests of equal means (medians).5 2 3 3 1*** 2** 1*** 1*** 1*** Interdependc Mean Median 2.80*** 2.60*** 1.80*** 2.51*** 1. **.60** 2. Hence. a Mean and median physician responses to questions in Task Interdependence Survey.00** 2.86 2.00 2. c task interdependence is the mean of each physician’s responses to question on continuity.01.36** 2.43 2.86 1. b Means (medians) for each variable for each specialty (except family practice) were compared to the corresponding mean (medians) for family practice using a paired t-test (signed rank test) for the 14 respondents who ranked each pair. higher ratings indicate greater beneﬁt from intra-group cooperation and teamwork. Mean and median values computed from 14–18 responses per variable.67*** 2.13 2.57*** 2.25*** 3*** 3*** 2* Panel Inﬂuence Mean Median 2.42** 2.47*** 2.67** 1.50 2.67* 2.05. 377 .70*** 1.67* n-groups 94 24 18 13 11 12 13 8 4 1 935 *.50 1.36*** 2.87*** 1.73* 2*** 3*** 2** 2*** 2*** 3*** 2.50** 2.43*** 1.29*** 2. Continuity—How important is continuity of care in this specialty? Inﬂuence patient—To what extent can a physician in this specialty individually inﬂuence the size and / or nature of his / her patient panel or case load? Control schedule—To what extent can a physician in this specialty control when and how many hours are worked in a given week? Responses are on a scale of 1–5 and have been coded so that lower numbers indicate that continuity is of greater importance and that physicians have greater patient inﬂuence and greater schedule control.27 1. patient inﬂuence.14 2.84*** 2.79** 1.50 2.58*** 1.21 2.97 2.87*** 1. and schedule control.5 3 2.50 2. and 0. *** Indicate p-values of 0.86 2.27*** 2.10.50*** 3 2.33 1.67 2.14*** 1. Higher numbers indicate greater beneﬁt from intra-group cooperation and teamwork.5 3*** 1** 2.The Accounting Review Internal Medicine Subspecialty Cardiology Gastroenterology Hematology-oncology Nephrology Pulmonary medicine Other Allergy Neurology Dermatology Physiatry Psychiatry Total January 2010 American Accounting Association Group-Based Compensation in Professional Service Firms TABLE A-1 (continued) Continuity Mean Median 1.63*** 2. 0.67 2.
Inc. Encinosa. Production. V. CA: University of California Press. M. A. J. Chicago. and P. B. Springﬁeld.. 1997.. Glover. J.C. P. Moral hazard in teams. and job design. P. Stoddard. Hansmann. J. Task Order 0038—Final Report. asset ownership. M. F. and P. Third Update to the Geographic Practice Cost Index. 1999. Bushman. III. NJ: Mathematica Policy Research.. J. Gold. E. and J. Devers. A. 2001. L. The sociology of groups and the economics of incentives: Theory and evidence on compensation systems. The Accounting Review 81 (5): 1045–1071. 2002. Scotchmer. D. and risk sharing. McKee. Jianakoplos. Princeton.gov / NationalHealthExpendData. 1991. A. repeated tasks. Journal of Accounting and Economics 23 (1): 7–30. MA: Harvard University Press. Rebitzer. Partnerships. H. Bernasek.. The American Economic Review 62: 777–795. Gaynor. Arrangements between managed care plans and physicians II. Lake. List. Almond. J. and J. Holmstrom. 2003. Archives of Internal Medicine 163: 1958– 1964. 1997.hhs. 2006. 1995. A. K. Proﬁt sharing and monitoring in partnerships. and economic organization. and N. 2000. 1993. 1999. R. D. R. Milgrom. incentives. Coalitions. The RAND Journal of Economics 26 (4): 591–613. The American Economic Review 91 (3): 525–541.: AIS Inc. http: / / www. Getzen. and P. and P. and H. and A. 1972. 2002. Casalino. prepared by KPMG. and D. Optimal incentives for teams. 2008. Indjejikian. Laud. Huddart. Economic and demographic trends signal an impending physician shortage. Reed. Farrell. In Proceedings of the Fifth Berkeley Symposium in Mathematical Statistics and Probability. 1996. Atlantic Information Services. Gertler. 1967. 2005. III. Huber. Centers for Medicare & Medicaid Services. and M. 1988. NBER Working Paper 5933. 1996. Bajtelsmit. In Report Physicians Payment Review Commission. National health expenditure accounts 2006 highlights. Y. M. T. American Medical Association (AMA).cms. H. 1982. Beneﬁts of and barriers to large medical group practice in the United States. Liang.. 1995. Inc. Journal of Law Economics and Organization 7: 24–52. Getzen.. M. Health Affairs 21 (1): 140–154. Berkeley. H. The Journal of Industrial Economics 33: 199–215. K. Che.. S. Kyonghee. Gilkeson. The Ownership of Enterprise. Yoo. Medical Group Practices in the US. Fellingham. information costs. The behavior of maximum likelihood estimates under non-standard conditions. Gender differences in deﬁned contribution plans. and S.. T. Nagarajan. Pauly. and J. ———... H. and S. Economics Letters 76 (2): 151–158. legal liability and incentive contracts. J. The Accounting Review American Accounting Association January 2010 . Arya. Lake. Washington. and N.. R. IL: AMA.. Hurley. 1990. A.. T. Evans. K. 1984. L. A. W. Cooper. 2002. J. J. The Bell Journal of Economics 13 (2): 324–340. Compensation and productive efﬁciency in partnerships: Evidence from medical group practice. The Journal of Political Economy 98 (31): 544–573. Journal of Economic Theory 60: 410–427. Multitask principal-agent analyses: Incentive contracts. P. and J. Demsetz. Dwyer. Cambridge. R. Financial Services Review 8: 1–10. H. Moral hazard and risk spreading in partnerships. The Quarterly Journal of Economics 103 (2): 279– 297. Aggregate performance measures in business unit manager compensation: The role of intraﬁrm interdependencies. Journal of Accounting and Economics 40 (1–3): 153–187. T. E. Health Care Financing Administration (HCFA). Gender differences in revealed risk taking: Evidence from mutual fund investors. Physician Compensation Plans Used by 57 Medical Groups. A brand name theory of medical group practice. Teams. Journal of Accounting Research 33: 101–128. M. Itoh. 1999..378 REFERENCES Pizzini Alchian. and implicit incentives. VA: National Technical Information Service. Gaynor. Uncertainty. Smith. ———.
American Journal of Public Health 80 (3): 300–304. 2004. Miller. Pay equality and industrial politics. and S. Tadelis. IL: American Medical Association. A. Medical liability. supply. 1992. Journal of Accounting and Economics 44 (3): 300–327. 2003. In Report of the Council on Medical Services. Working Paper No. D. and ———. C. Journal of Accounting and Economics 24 (3): 243–273. Performance-based compensation in member-owned ﬁrms: An examination of medical group practices. 1995. Kessler. Physician Compensation and Production Survey. Economics of multispecialty group practice. 2003. Zeckhauser. The Journal of Political Economy 97 (3): 561– 580. D. and defensive medicine. W. Nutting. Flocke.. Lazear. 2002. Proﬁt sharing and the role of professional partnerships. Journal of Economic Psychology 97 (18): 605–628. 1999 report based on 1998 data. F. and M. Lang. Keating. Journal of Economic Theory 77: 285–299.. C. and D. Levinson. 2007. Journal of Economic Literature XXXVII: 7–63. and R. Kandel. Determinants of divisional performance evaluation practices. 1996. P. E. 1997. R. S. Economic Inquiry XVIII: 380–394. ———.. and E. 1999b. The growing proportion of female physicians: Implications for U.. N. The Quarterly Journal of Economics 120 (1): 131–171. Stange. Olin Program in Law and Economics. 2005. Gender differences in risk behavior in ﬁnancial decision-making: An experimental analysis. and M. J. Kletke.. S. When most doctors are women: What lies ahead? Annals of Internal Medicine 141: 471–474. An economic approach to inﬂuence in organizations. P. Deﬁning and measuring interpersonal continuity of care. 2007. M.. R. D. Lee. A. Annals of Family Medicine 1 (3): 134–143. A. Chicago. Goodwin. J. 1988. 1997. M. 1980. 1989. C. A. K. The rise in professional liability insurance premiums. and R. S. Prendergast. Gordon. CO: MGMA. Silberger. E. The Journal of Ambulatory Care Management 19 (3): 26–33. Englewood. Leone. Stanford Law School.. Marder. Medical Group Management Association. McClellan. Efﬁciency in partnerships with joint monitoring. John N. 191. Powell. 1990. Mauney. Lurie. Saultz. P..Group-Based Compensation in Professional Service Firms 379 Ittner. Milgrom. The Journal of Political Economy 100 (4): 802–817. Monitoring physicians: A bargaining model of medical group practice. Zyzanski. B. A. Partnerships as insurance devices: Theory and evidence. F. Levin. CMS Report 12-A-02. A. Peer pressure and partnerships. J. The provision of incentives in ﬁrms. 1996. Moral hazard in repeated professional partnerships. Do doctors practice defensive medicine? The Quarterly Journal of Economics (May): 353–390.. Free riding. The Journal of Human Resources 8 (1): 37– 56. 2000. Englewood. W. Tollison. J. and K. Ansic. M. and A. ———. J. M. American Journal of Sociology 94 (supplement): 154–179. Lazear.. CO: MGMA. G. Journal of Health Economics 9: 463–481. The relation between efﬁcient risk-sharing arrangements and ﬁrm characteristics: Evidence from the managed care industry. Newhouse. The RAND Journal of Economics 26 (4): 614–629. Contemporary Accounting Research 11 (2): 895–917. V. 1999. 1997. 1995. 1973. Leibowitz. Rizzo. and team production in legal partnerships. D. Larcker. A. Cost Survey. Continuity of primary care: To whom does it matter and when? Annals of Family Medicine 1 (3): 149–155. Narayanan. managed care. and N. Pushing incomes to reference points: Why do male doctors earn more? Journal of Economic Behavior & Organization 63: 514–536. Pizzini. The Accounting Review January 2010 American Accounting Association . W.. and P. 2002. shirking. 1990.S. 1999 report based on 1998 data. J. The economics of group practice. 1999a. Journal of Management Accounting Research 14: 99–117. Pauly.
Interdependences and group effectiveness. Work group diversity. K. Stata Statistical Software: Release 7. Van Knippenberg. Gender differences in the allocation of assets in retirement savings plans. Research in Organizational Behavior 30: 77–140. College Station. 1995. 1998. E. Schippers. D. Philadelphia. R. R. Wageman.0 User’s Guide. Haley. A. StataCorp. Incentives and worker behavior. W.. edited by Nalbantian. E. Taylor. 1980. How to Choose a Medical Specialty.. TX: Stata Corporation.. and G. NJ: Rowan and Littleﬁeld. A. Sunden. J. A. and C. Incentives and cooperation: The joint effects of task and reward interdependence on group performance. Controllable lifestyle: A new factor in career choice by medical students.. White. 1997. 2003. H. 4th edition. Weiss. Econometrica 50: 1–25. Griffen. 1987. O. Incentives. Academic Medicine 64: 606–609. and G. Young. 2001. 2007. In Information. Determinants of coordination modes within organizations. C.. Demography and diversity in organizations: A review of 40 years of research. A. 137–149. and M. R. J. American Sociological Review 41: 322–338. B. Jr. Annual Review of Psychology 58: 515–541. Koenig. Surette. and W. The American Economic Review 88 (2): 207–211. Williams. ———. and Risk Sharing. Towota. Strodel. V. Y. Jarecky. A heteroscedasticity-consistent covariance matrix estimator and a direct test of heteroscedasticity. Van de Ven. 1990.. A. K. Baker. 1976. A. H. H. W. 1998. PA: Elsevier Science. Delbecq. The Accounting Review American Accounting Association January 2010 . O’Reilly. Journal of Organizational Behavior 18: 139–158. Administrative Science Quarterly 40: 145–180.380 Pizzini Schwartz. A. and R.
. or email articles for individual use. download. users may print.Copyright of Accounting Review is the property of American Accounting Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.