Strategic Management Journal Strat. Mgmt. J. (2008) Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.718 Received 22 March 2006; Final revision received 16 May 2008

ENVY, COMPARISON COSTS, AND THE ECONOMIC THEORY OF THE FIRM JACK A. NICKERSON* and TODD R. ZENGER Olin Business School, Washington University, St. Louis, Missouri, U.S.A.

An economic theory of the firm must explain both when firms supplant markets and when markets supplant firms. While theories of when markets fail are well developed, the extant literature provides a less than adequate explanation of why and when hierarchies fail and of actions managers take to mitigate such failure. In this article, we seek to develop a more complete theory of the firm by theorizing about the causes and consequences of organizational failure. Our theory focuses on the concept of social comparison costs that arise through social comparison processes and envy. While transaction costs in the market provide an impetus to move activities inside the boundaries of the firm, we argue that envy and resulting social comparison costs motivate moving activities outside the boundary of the firm. More specifically, our theory provides an explanation for ‘managerial’ diseconomies of both scale and scope—arguments that are independent from traditional measurement, rent seeking, and competency arguments—that provides new insights into the theory of the firm. In our theory, hierarchies fail as they expand in scale because social comparison costs imposed on firms escalate and hinder the capacity of managers to optimally structure incentives and production. Further, hierarchy fails as a firm expands in scope for the simple reason that the costs of differentially structuring compensation within the firm to match the increasing diversity of activities also rises with increasing scope. In addition, we explore how social comparison costs influence the design of the firm through selection of production technologies and compensation structures within the firm. Copyright  2008 John Wiley & Sons, Ltd.

INTRODUCTION externally managing activities. When market gov- ernance is costly or markets fail, integration is The theory of the firm has primarily focused on likely; when internal governance is costly or orga- the question of firm boundaries, articulating when nizations fail, market governance is likely. Cur- firms supplant markets and when markets sup- rent theory provides well-developed explanations plant firms. Coase (1937) originally articulated the for the failure of markets. Markets commonly fail logic that boundary choices depend on a com- when exchange demands highly specialized assets, parative assessment of the costs of internally and involves measurement difficulty, or requires the transfer of knowledge (Williamson, 1975; Klein, Crawford, and Alchian, 1978; Barzel, 1982). How- Keywords: theory of the firm; envy; comparison costs; ever, as both Coase (1937) and Williamson (1985) diseconomies of scale and scope; firm boundaries note, a theory of market failure alone is not a the- ∗ Correspondence to: Jack A. Nickerson, Olin Business School, ory of the firm, because such theory fails to resolve Washington University, Campus Box 1133, One Brookings Drive, St. Louis, MO 63130-4899, U.S.A. a basic puzzle: If firms are advantaged in manag- E-mail: [email protected] ing these complexities in exchange, ‘why is not all

Copyright  2008 John Wiley & Sons, Ltd. J. A. Nickerson and T. R. Zenger production carried on by one big firm?’ (Coase, highly autonomous internal units and then structur- 1937: 394). What precludes the firm from selec- ing market-like incentives within its boundaries. tively using the control features of hierarchy and Similarly, this argument does not explain why a otherwise replicating market incentives within its firm cannot integrate an independent firm or activ- boundaries? If firms can replicate market incen- ity and craft incentives that fully replicate the mar- tives and utilize authority to control exchange only ket incentives that previously existed. when needed, the firm should face no boundary Influence activities within organizations, as dis- limits. Indeed, absent impediments to replicating cussed by Milgrom and Roberts (1988; 1990a; market incentives within the firm, the boundaries 1990b), provide a partial explanation. When an of the firm become rather irrelevant and managers activity is integrated within the boundaries of the enjoy enormous flexibility in the design and struc- firm, it is placed in an arena in which incentives ture of organizations. now exist for both those managing the integrated Existing theory provides little in the way of activity and those positioned elsewhere in the firm explanation for the limits of the firm or the causes to politically influence the distribution of rewards of organizational failure. The prevalent view in allocated to those managing it. They argue that early writings on industrial organization was that such rent-seeking activity is wasteful and fully limits to firm size were due to ‘diminishing returns absent when the activity is autonomously con- to management’ (Sraffa, 1926; Kaldor, 1934). trolled by the market. Absent integration, there is Implicitly, Coase (1937) adopted a similar argu- simply no one to politic and therefore no influ- ment, assuming that at some point, integrating ence activities. While these influence activities additional transactions becomes more costly than undoubtedly play an important role in shaping the managing them externally. However, Coase pro- boundaries of firms, they provide at best a par- vided no clear explanation as to the origin of these tial explanation. The theory does not explain why costs or why so-called management costs should these organizational costs should increase with rise with firm scale and scope. Indeed, Coase has size or scope. Thus, while Milgrom and Roberts commented that the question of why the costs (1988, 1990a, 1990b, 1992) explain why selec- of internal organization increase with firm scale tive intervention is costly within firms, they do and scope remains unanswered (Coase, 1988). Our not explain why the marginal transaction or activ- goal is to develop a theory of organizational fail- ity becomes increasingly costly to integrate, as the ure, more precisely a theory of organizational dis- economies of scale and scope that is rooted in firm increases in scale and scope. Since Milgrom processes of social comparison as discussed in and Roberts’ (1988, 1990a, 1990b, 1992) argument the sociology and social psychology literatures. is invariant to size and scope, the theory fails to These processes of social comparison give rise to explain the limits to firm size and scope. what we label ‘social comparison costs’ (Zenger, In this study, we develop a theory of managerial 1992; 1994). Our theory argues that while transac- diseconomies of scale and scope, which are orga- tion costs in the market prompt access to author- nizational failures that constrain the boundaries of ity through integration, social comparison costs the firm. We begin by assuming that market fail- prompt managers to limit the degree of integration ures of the type commonly described in the trans- and otherwise take costly organizational actions to action cost economics literature provide the impe- restrict and efficiently manage these social com- tus to integrate activities within the boundaries parison costs. of the firm. Cognizant of these transaction costs We are, of course, not the first to explore the in the market that prompt the extension of firm question of organizational failure (see Hennart, boundaries, we develop a theory of countervail- 1994). Some have argued measurement difficul- ing social comparison costs that impel managers ties escalate with firm size and limit the capacity to restrict these boundaries or otherwise control of large firms to offer the high-powered incen- these costs. Thus, while market failures create a tives prevalent in markets (Barzel, 1982; Holm- type of centripetal force for moving activities out strom, 1989; Rasmusen and Zenger, 1990; Zenger, of the market and into a firm, our theory explains 1992; Williamson, 1985). But, this argument alone when and how organizational failures create a cen- does not explain why a large firm can’t replicate trifugal force for moving activities out of the firm the measurement precision of markets by creating andintothemarket.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

As a brief preview, our theory begins by examin- costs that arise as managers attempt to selec- ing the behaviors of both individual employees and tively infuse market incentives within the firm. managers in response to the infusion of market-like We then discuss the structure of these costs and incentives within the firm. Consistent with a wide identify and describe three levers, or organiza- range of social science literature, we assert that tional design choices, available to management individual employees invidiously compare their to attenuate them. We use two levels of analy- rewards with others they deem to be within their sis—employee decisions about how to attenuate referent group (for example, see Adams, 1963; envious emotions and managerial policy decisions Festinger, 1954). If perceived inequity arises, the about the structure of the firm—to develop an resulting negative feeling—what we refer to as argument for how social comparison generates dis- envious emotion—drives individuals to expend economies of scale and scope in firms. We then effort to ameliorate these perceptions of inequity. discuss the implications of our theory along with Such behaviors include reduced effort, influence a few caveats and conclude. activities, departure, noncooperativeness, or even outright sabotage. Such behaviors impose substan- tial costs on the firm—costs that we reference as Selective intervention and social comparison social comparison costs. costs Managers, of course, are not passive actors in responding to these comparison costs. Managers Our theory is designed to complement, not replace, anticipate social comparison costs or at least per- the logic of transaction cost economics. In trans- ceive them when they arise, and take active steps action cost economics, managers choose markets to attenuate them. Management, we argue, has over firms in order to access the superior incen- three structural levers at its disposal to attenu- tives that markets provide (Williamson, 1991). In ate social comparison costs: ‘compressing’ rewards other words, managers forego integration and thus by decoupling pay and performance, shifting ‘pro- limit the size and scope of the firm in order to duction technology’ to reshape social comparison access the high-powered incentives of the market and social referents, and redrawing the bound- (Williamson, 1985). Consequently, as Williamson aries of the firm. Our theory argues that each of argues, if managers could replicate the incentives these choices potentially imposes significant costs of the market within the firm, the firm would face on the firm. Therefore, the probability that man- no limits to its scale and scope, and hence no limits agers will select any one of them depends on to its boundaries. If firms could truly complement the magnitude of social comparison costs associ- the virtues of internal organization with the incen- ated with each. As social comparison costs rise, tives of the market, markets need never arise and managers are more likely to respond with damp- the boundaries of the firm face no limits. There- ened incentives, compromised production technol- fore, to understand the limits to the scale and scope ogy, or restricted boundaries to the firm. We argue of firms, it is useful to first explore why firms fail thatthescaleandscopeofthefirmareprimary as managers attempt to internally replicate market drivers of these social comparison costs. Hence, incentives. as scale and scope increase, the need to attenuate We are not the first to observe the difficulty social comparison costs through incentive dampen- that firms face in selectively replicating market ing, production efficiency compromises, or bound- incentives for subunits or activities within the firm ary restriction increases. The result is a theory (Williamson, 1985; Zenger and Hesterly, 1997). of organizational failure in which organizational Indeed, there is empirical literature that confirms costs rise with increasing scale and scope of the the greater difficulty that large firms confront firm. We believe our theory provides an expla- in structuring market incentives within the firm nation for managerial diseconomies of scale and (Garen, 1985; Zenger and Marshall, 2000). Fur- scope—an argument that is independent from tra- thermore, the literature is replete with examples ditional measurement, rent seeking, and compe- of failed attempts at such ‘selective intervention.’ tency arguments—that contributes to determining By examining a few notable illustrations of these firm boundaries. failed efforts, we hope to highlight the genesis of In the pages that follow, we use a series of social comparison costs and provide a backdrop illustrations to highlight the social comparison for our theory.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger

Illustration 1 30 to 50 percent above what was traditionally expected. These employees consequently earned Harvard’s efforts to manage internally their endow- a significant premium above what other more ment portfolio provide a fascinating example of skilled workers earned in other parts of the plant. social comparison costs. Until 2005, Harvard Man- Because of this differential, ‘[m]anagement was agement Company was a wholly owned subsidiary besieged by demands that the inequity be taken of Harvard University that managed Harvard’s care of’ (Strauss, 1955: 94). Ultimately, manage- $27 billion endowment portfolio. In an effort to ment returned the painting operation to its original mimic the rewards offered to fund managers of (and less productive) pay structure, which quickly privately managed hedge funds, the multiple fund led to the voluntary departure of workers on the managers within the Harvard subsidiary received team, as well as the departure of their supervisor. compensation based purely on a formula that linked bonus compensation to their fund’s perfor- mance relative to a benchmark fund of comparable Illustration 3 risk and classification (see Hall and Lim, 2003). In 1980 Tenneco Inc., acquired a relatively small These internally managed funds quite consistently company, Houston Oil and Minerals Corpora- outperformed the benchmark funds by a very wide tion (HOMC) (see Williamson, 1985: 158; and margin, which resulted in extremely high com- Milgrom and Roberts, 1992: 194). To encourage pensation for these Harvard employees. In fiscal the retention of HOMC’s exploration talent, Ten- year 2004, the top two fund managers received neco offered special salaries, bonuses, and ben- $25 million each. In fiscal year 2003, the top efits to HOMC employees—payments that were fund managers received $35 million each. In 2001 not offered to others at Tenneco. ‘Tenneco also and 2002 large sums were also paid out to fund agreed to keep [HOMC] intact and operate it managers due to exceptional relative performance, as an independent subsidiary rather than consol- even though the funds themselves declined in idate the acquisition.... Despite initial enthusiasm, value. In 2004, opposition from students, alumni, [HOMC’s] managers and its geologists, geophysi- and faculty to such pay reached a feverish pitch cists, engineers, and landsmen left in droves during (Harvard Magazine, 2004). Key alumni called the ensuing year’ (Williamson, 1985: 158). The for a university-wide forum to review the matter implementation of the customized compensation and threatened to withhold donations. Larry Sum- package had been delayed, because, as Tenneco’s mers, Harvard’s President, claimed that outsourc- vice president for administration maintained, ‘We ing the activity would lead to lower performance have to ensure internal equity and apply the and higher fees, while Ronald Daniel, the Uni- same standard of compensation to everyone...’ versity Treasurer, defended the practice by noting (Williamson, 1985: 158). that other universities which use ‘external invest- ment managers...do not face comparable scrutiny’ (Harvard Magazine, 2004: 73). In early 2005, Har- Illustration 4 vard Management Company significantly restricted Recently, MBA students at a North American busi- the maximum levels of compensation within the ness school, which will remain anonymous, com- subsidiary. By March of 2005, many of the key plained about the lack of access to the more ‘visi- fund managers had departed, taking with them ble and famous’ professors at the school. Seeking portions of Harvard’s endowment to manage exter- to satisfy student demand while recognizing that nally. these same professors were in high demand for consulting and speaking engagements, the dean Illustration 2 proposed individually negotiating overload teach- ing payments to faculty. This approach would Strauss’ (1955) describes a factory that manufac- enable the payment of ‘market rates’ for internal tured wooden toys that sought to enhance the teaching services. However, discussion with fac- productivity of a team of workers in the paint ulty quickly led to concerns over fairness, which department through a bonus pay system based could be resolved only if there was a set rate for on team output. As a result of the market-like all faculty to do overload teaching. It is interesting pay plan, productivity of this team increased to note that no such fairness concerns were voiced

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm about high speaking and consulting fees generated reactions (or anticipated reactions) of these other from outside the university. Negotiation with indi- employees prompted the plans’ cessations. Note vidual faculty on overload remuneration did not that in each case management was forced to make proceed. a trade-off between the incentive benefits of those affected by the plan, in terms of higher effort or the Illustration 5 attraction of superior talent, and comparison costs imposed by those not attached to the plan. In each At International Harvester in the 1960s, the union case the social comparison costs were determined requested that management increase the speed of to outweigh the incentive benefits. particular assembly lines where employees were These illustrations suggest that organizations rewarded in part based on output. Despite clear may fail relative to markets because social compar- benefits to the firm, management refused, con- ison costs limit the flexibility of managers in offer- cerned that higher pay for this group would pro- ing the ‘optimal’ market-like incentives for specific mote dissension elsewhere among those not ben- activities within the firm. Below, we illuminate efiting from the line speed increase (reported in more completely the mechanisms that underlie this McKersie, 1967: 222). argument. Moreover, we argue that as a firm’s scale and activity scope increase, comparison costs Illustration 6 also increase and with them the manager’s inflex- ibility in tailoring incentives for different inter- In the early 1970s, Syntex and Varian Corporation nal activities. These rising costs and inflexibility established a joint venture to commercialize jointly developed technology. The intent was to create an amount to organizational failures that constrain the entrepreneurial unit free from the bureaucracy of boundaries of the firm. the larger parent companies. Consistent with the desire to create the feel of a truly new start-up, the Social comparison and human behavior head of this new venture was given a compensation package characteristic of a small, entrepreneurial Our theory begins with the assumption that indi- firm. He was granted an equity stake and assigned viduals engage in a process of socially comparing a relatively small salary. As soon as the details their rewards to those received by salient referents. of this package were revealed to senior executives Thus, decisions about how rewards are allocated at Syntex, a major internal uproar ensued. Long define in part the nature of these social compar- before any value from the joint venture could be isons. Research on social comparison processes realized, the compensation package of the head and their effect on individuals are legion across the of the joint venture was realigned so as to be social sciences (Homans, 1961; Festinger, 1954; consistent with others at a similar level within Adams, 1963; Martin, 1981). Many theories indi- Syntex (former vice president of Human Resources cate that people engage in comparison activities, to second author, personal communication). focusing especially on relative income and the basis by which relative income is determined. Moreover, these comparisons tend to be invidi- Overview ous in the sense that individuals identify salient There are several common themes revealed in referents as those who are economically better these illustrations. In each, managers within the off. These theories assert that people care about organization sought to develop a high-powered, inequity; they react negatively to outcomes they market-like incentive mechanism, custom-tailored deem unfair; and they exert efforts to reduce neg- to a specific group or activity. In each case where ative feelings proportionate with the inequity they the incentive plan was in place long enough to perceive (Adams, 1963). This propensity to com- measure performance, the performance results for pare income to salient referents and the willingness those engaged in the activity were quite positive. to expend effort to reduce perceived inequity in rel- However, due to this success or anticipated suc- ative income, we assert, constitutes a fundamental cess, the plans produced wide variance in compen- behavioral assumption about human nature upon sation and, in particular, some very high compen- which our theory rests. sation levels that others in the organization per- Modern research on social comparison processes ceived as inequitable. The costs imposed by the begins with Festinger (1954), who assumed that

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger humans have an innate propensity to self-evaluate Social comparisons that lead to inequity percep- ‘based on comparison with other persons’ (1954: tions in individuals create a willingness to expend 138). Homans’s (1961) study constructed similar effort to attenuate these feelings of inequity. Such arguments, emphasizing that individuals particu- efforts increase the greater is the perceived larly compare income.1 Equity theory as developed inequity. For instance, Adams (1963: 427) argued by Adams (1963) posits that ‘[i]nequity exists for that ‘[t]he presence of inequity in [an individ- [an individual] whenever his perceived job inputs ual] creates tension in him ...proportional to and/or outcomes [typically measured by actual the magnitude of inequity present’. The pres- or expected income] stand psychologically in an ence of inequity ‘will drive him to reduce it.’ As obverse relation to what he perceives are the inputs with equity theory, feelings of relative deprivation and/or outcomes of [others]’ (Adams, 1963: 424). prompt behavioral responses to alleviate or avoid Relative deprivation theory similarly concludes such feelings (Martin, 1981). Generally speak- that feelings of deprivation arise from compar- ing, the literature argues that those who perceive isons of rewards to some referent (Martin, 1981). inequity relative to their social referents expend It is important to note that perceptions of inequity effort to eliminate this tension by restoring equity are not symmetric. That is, individuals commonly (see Walster, Berscheid, and Walster, 1973; Adams have inflated perceptions of personal contribu- and Freedman, 1976; Ma and Nickerson, 2007; for tion or performance (Meyer, 1975; Zenger, 1994). empirical reviews see Greenberg, 1982; Suls and The problem of establishing equity perceptions is Wills, 1991). further exacerbated by individuals’ propensity to An important issue underlying this research is compare pay only to those perceived as compa- identifying which referents are salient. In other rable in performance, but earning more (Martin, words, with whom do individuals socially compare 1981). This combination of upwardly biased self- themselves? The general conclusion in the litera- assessment and a propensity to compare income ture is that spatial proximity, degree of interaction, to those earning more nearly always ensures at and availability of information are primary deter- least some level of inequity perception. Thus, vary- minants of the choice of salient referents (e.g., Fes- ing perspectives on social comparison and inequity tinger, 1954; Kulik and Ambrose, 1992; Williams, perceptions are consistent with the notion that indi- 1975). Spatial proximity means not only propin- viduals are envious of salient referents perceived quity, but also a variety of other demographic to receive superior income relative to their contri- measures of social distance such as age, tenure, 2 butions. education, gender, and the like. As with spatial proximity, the degree of interaction and the avail- 1 Homans’s classic on social behavior (1961), which builds on ability of information may be endogenous to firm Blau’s (1955) social exchange theory, identifies an equivalent propensity. Homans claims that individuals engage in exchanges, boundaries and other organizational and produc- whether they are economic or social, because they receive some tion technology choices. These conclusions are benefit in excess of their costs (Homans, 1961: 72). Further, reminiscent of Aristotle’s observation (Rhetoric, individuals ‘perceive and appraise their rewards, costs, and investments in relations to the rewards, costs, and investments 1388, quoted in Goel and Thakor, 2003): ‘We envy of other men’ (Homans, 1961: 76). Homans’s key proposition in those who are near us in time, place, age, or rep- this regard is that ‘the more to a man’s disadvantage the rule of utation.’ We return to these issues below. distributive justice fails of realization, the more likely he is to display the emotional behavior we call anger’ (Homans, 1961: The literature suggests that individuals pur- 75). This proposition equates to individuals caring not only about sue several different behaviors or strategies to the absolute level of their income but also about relative income. reduce envious emotions that accompany per- Moreover, Homans acknowledges that such anger translates into action because individuals ‘learn to do something about it’ ceived inequity. First, individuals alter their own (Homans, 1961: 77). behavior. In particular, they reduce effort, bringing 2 At least since Aristotle, philosophers have recognized envy and contributions more in line with rewards (Adams, jealousy as fundamental propensities of human nature. Envy is the emotion that arises when one desires something currently possessed by another (Salovey, 1991). Envy is one of the seven deadly sins (Silver and Sabini, 1978) and perhaps the most per- did not delve into the emotional underpinnings of social compar- vasive but under-acknowledged one (Epstein, 2003). Arguably, ison theory, Salovey (1991) argues that envy and jealousy are it is envy and jealousy that generate feelings from social com- the specific emotions that accompany such appraisals. Indeed, parison that give rise to actions to reduce such feelings because invidious comparison, envy, jealousy, or some combination is people systematically care more about others ahead of them com- mentioned as the implicit or explicit motivation for behavior in pared to those behind them. For instance, while Festinger (1954) all three of the psychological perspectives mentioned above.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

1963).3 Second, individuals seek to alter others’ design an organization that attenuates these social outcomes and rewards in several ways: directly comparison costs while simultaneously providing sabotaging others’ efforts, behaving noncoopera- effective incentives and efficient production tech- tively with them in collaborative settings, or sim- nology. ply lobbying those managers who assign their compensation (Cropanzano, Goldman, and Folger, Economizing on social comparison costs 2003; Pruitt and Kimmel, 1977; Smith and Walker, 2000). Finally, individuals may simply depart While transactions cost economics contends that from the referent group that prompts envy, for the manager’s job is to economize on transaction instance, by exiting the job, firm, or industry (Fes- costs, our theory suggests that the manager’s job tinger, 1954; Mui, 1995; Salovey, 1991; Sheppard, is to (also) economize on social comparison costs. Lewicki, and Minton, 1992; White and Mullen, We maintain that managers have at their disposal 1989). These behavioral strategies to reduce envi- three structural or organizational levers with which ous emotion impose costs on the firm, which we to attenuate social comparison costs: wage com- term social comparison costs (Zenger, 1992; 1994). pression, the choice of production technology, and Evidence of these social comparison costs is the choice of firm boundaries. All contribute to abundant in our motivating examples. For instance, explaining organizational diseconomies of scale. perceptions of the inequitable compensation led to Below, we not only highlight how these levers intense lobbying activities to reallocate rewards economize on social comparison costs, but also at Harvard, at the wooden toy factory, at the examine how each choice imposes its own costs North American business school, and at Syntex. on the organization. After discussing these levers, Tenneco’s insistence on ‘internal equity’ in com- we consider their relationship to managerial disec- pensation also implied a concern for these costs. onomies of scale and scope. Similarly, management at International Harvester was concerned with these social comparison costs Wage compression in contemplating selectively rewarding a subset of employees based on output. In all of our illus- Scholars from a range of disciplines suggest that trations social comparison costs overwhelmed any weakening the link between pay and performance gains that accrued or would have accrued from is a common response to social comparison pro- offering high power incentives to a select group cesses (Akerlof and Yellen, 1990; Frank, 1985; within the firm. Konrad and Pfeffer, 1990; Zenger, 1992). By tak- In summary then, humans have the propensity ing steps to reduce differences in income among to engage in social comparisons. These compar- workers, managers reduce the impetus for workers isons, in the presence of differences in relative to envy each other’s income. Indeed, by adopt- income among salient referents, trigger percep- ing uniform compensation, managers essentially tions of inequity and envious emotion that indi- avoid issues of distributive justice.4 When all pay viduals attempt to reduce. An overarching theme in the literature is that the greater the disparities 4 Even when compensation is uniform and distributive concerns among salient referents, the larger is the costly are rendered rather moot, procedural justice issues may still arise. effort individuals are willing to expend to reduce Distributive justice focuses on perceived fairness of the distri- bution of rewards (Homans, 1961) whereas procedural justice their perceptions of inequity. The specific actions focuses on perceived fairness of the process by which decisions undertaken by individuals to reduce their envious are reached (Greenberg, 1990). We do not incorporate procedu- emotion impose social comparison costs on firms. ral justice concerns into our argument for three reasons. First, empirical research indicates that distributive justice contributes Thus, the organizing challenge of the manager is to to variance of satisfaction with pay more than twice that of pro- cedural justice (Tyler, Rasinski, and McGraw, 1985; Folger and Konovsky, 1989), which suggests that distributive justice may 3 In addition to increasing or reducing effort to bring contri- be far more important than procedural justice. Second, Long, butions more in line with rewards, workers might also (over) Bendersky, and Morrill (2007) find that when tasks are separate invest in capabilities through training or education based on the and pay contingent, workers principally care about distribution of expectation of future rewards. To the extent that workers choose rewards—a perspective consistent with our view. They also find increasing effort and investing in capabilities to reduce envi- that when tasks are interdependent and wages are not contingent ous emotions, these behavioral responses can benefit the firm. (i.e., wages are compressed), workers may then care more about We set aside these behaviors that might benefit the firm while procedural justice. Their finding suggests that distributive justice developing our theory and return to them in the discussion. is a more important driver of structural responses by managers

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger is equal, perceptions of inequity must stem from It is useful to distinguish between two types of differing perceptions of performance rather than wage compression. Horizontal wage compression pay. A small but growing economics literature refers to the adoption of rather flattened or more has incorporated envy and social comparison into uniform compensation for individuals in equiva- formal models.5 These economists acknowledge lent or identical jobs. Thus, as the costs imposed that workers care about relative income, which on firms by social comparison processes increase, creates social forces that affect the wage pro- firms adopt more horizontally compressed or more files within and between firms. In an early com- uniform income for each position (e.g., Hicks, ment, Hicks (1955: 390) states that ‘[e]conomic 1955; Addison and Burton, 1981; Hamermesh, forces do affect wages, but only when they are 1975; Dunlop, 1947; Pencavel, 1977). Indeed, it strong enough to overcome these social forces.’ is well documented in labor economics that work- Economists acknowledge that disparities in com- ers undertaking the same job or occupying posi- pensation can ‘induce discontent among employ- tions within the same grade typically have minor ees, ... followed by uncooperative and unaccom- differences in relative income even though their modating behavior’ (Pencavel, 1977: 239). Thus, marginal products may vary greatly. This unifor- economists theoretically and empirically note that mity of income—income that is invariant with managers can and do attenuate social comparison performance—is a decision made by management. costs by flattening the wage structure compared A second type of wage compression—vertical to the marginal product schedule. In doing so, wage compression—arises among workers hold- the firm elevates pay beyond the marginal product ing different positions, or performing completely for some and constrains pay below the marginal different jobs. Vertical wage compression involves product for others, which may precipitate depar- reducing the variation in income across differing ture—a phenomenon called wage compression. positions or jobs, despite potentially widely vary- These economic arguments clearly resonate with ing marginal products associated with these jobs those of equity theory, relative deprivation, and and positions. Addressing the vertical wage com- social exchange theory discussed above. pression phenomenon, Frank (1984a; 1984b; 1985) offered a formal theory how invidious compar- isons and variation in utility for status lead to and that procedural justice may take on importance only after vertical wage compression in firms. Akerlof and wages are compressed. Third, scholars have found the central Yellen (1990) in their fair wage hypothesis argue issue of procedural justice revolves around worker commitment to the organization, which may be little affected by production that vertical wage compression is a best response technology and organizational boundaries, which are important by management to counter reduced effort, which levers in our theory. occurs as a response to envy among individu- 5 Although still rare, economists with increasing frequency have als with different skills. These and other scholars incorporated the concept of envy into a variety of analyses typically modeling preferences as interdependent. For instance, argue that preferences for perceived inequity may Foley (1967) introduced the notion of social comparison and cause managers to vertically compress wages by defined envy-free equilibria in which everyone likes his own narrowing income dispersion across levels and jobs allocation goods at least as well as that of anyone else. Such envy-free equilibria are equitable and have been generalized (e.g., Frank, 1984a; 1984b; Akerlof and Yellen, by Varian (1974) in welfare economics and used to investigate 1990; Hicks, 1955; Addison and Burton, 1981; such topics as tax policies to correct distortions created by envy Hamermesh, 1975; Dunlop, 1947; Pencavel, 1977; (Banerjee, 1990) and redistribution programs (Brennan, 1971). Others have shown that relative status may lead to risk-taking Zenger, 1992). (Brenner, 1987) and may cause people to abstain from becoming Our prediction is that as social comparison costs superior to others to avoid provoking envy (Elster, 1991). Mui increase, managers are more likely to engage in (1995) directly implements equity theory in a utility function and wage compression both vertically and horizontally. analyzes how varying legal institutions’ and agents’ propensities for envy jointly determine whether agents engage in innova- Such wage compression offers an economic lever tion, retaliation, sabotage, or sharing behavior. Other theoretical to ameliorate social comparison costs by reducing economics literature that is focused on the applications of envy- the stimulus of envy. Its application, however, nec- based preferences in certain economic situations includes Bolton and Ockenfels (2000), Charness and Rabin (2002) and Fehr and essarily implies efficiency losses as weak incen- Schmidt (1999). Experimental economic evidence on envy has tives discourage high effort and trigger a pattern been reported by Martin (1981), Cason and Mui (2002), and of adverse selection among workers in which the Zizzo and Oswald (2001). Additional empirical evidence can be found in Frank (1985), Pfeffer and Davis-Blake (1992) and highly productive but underpaid depart, while the Pfeffer and Langton (1993). less productive but overpaid remain.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

Technology choice envy, the absence of team production limits the efficacy of behavioral strategies that impose com- A second economic lever at management’s dis- parison costs on the firm. For instance, in the posal in response to social comparison costs relates absence of team production, shirking imposes large to shaping a firm’s production technology, broadly costs on the worker in terms of lost productivity defined. From a classical microeconomic perspec- and income and imposes rather modest costs in tive, for any given desired output there is an opti- lost productivity for the firm. The verifiability of mal production technology that defines the con- individual performance causes influence activities figuration of assets and the structure of individual that appeal to management to raise pay or lower and collective tasks and activities within the firm. others’ pay to be credibly met with the response In addition to this ‘technical’ efficiency, the choice that the worker, not management, is responsi- of production technology also defines the degree to ble for earning less. Noncooperative behavior and which output is team produced, the spatial prox- sabotage are less feasible strategies because they imity of workers, and the degree to which workers lead to a reduction in the worker’s own produc- interact. Thus, managers shape both production tivity and income. Furthermore, independent jobs costs and social architecture when choosing a pro- reduce the opportunities for such behavior. By con- duction technology. trast, the envy promoted by pay variance with The social architecture defined by a produc- independent jobs may trigger behavioral strategies tion technology is important because it shapes that positively affect worker effort and benefit the social comparison processes, which can impose firm, since additional effort translates directly into costs on the organization in two ways. First, social higher income. Thus, social comparison processes architecture affects the cost of accessing informa- when jobs are independent may lead to social com- tion about peers’ performance and productivity, parison benefits instead of costs. which molds processes of social comparison. For By contrast, social comparison costs are likely instance, workers located in close proximity and quite high when managers in a team production performing similar tasks are likely to identify each environment attempt to link income to individ- other as salient referents and compare income, ual performance. As the scope of team produc- which, if different, can lead to social comparison tion increases, accurately observing and agree- costs. By increasing the physical distance among ing on relative contribution becomes more diffi- workers, management can restrict the scope of cult for workers and managers alike. In a team interaction and information sharing, thereby reduc- production setting, individual contributions are of ing the salience of these workers as referents.6 necessity subjectively determined. Consequently, Thus, management can reduce social comparison if managers assign pay based on these subjec- costs by increasing the physical and informational tive evaluations, this engenders influence activi- distance between jobs, which, however, departs ties because managerial subjectivity can be blamed from the technically efficient production technol- for differences in pay. Additionally, individuals ogy. without access to verifiable information rather eas- Second, the choice of production technology, ily develop inflated perceptions of their marginal specifically the extent of team production, shapes contributions (Meyer, 1975; Zenger 1994), which social comparison costs by defining the precision may amplify willingness to engage in influence with which individual performance is measured. In activities. Moreover, team production makes feasi- the absence of team production, workers and man- ble other behavioral strategies like noncooperative agers can clearly observe and verify relative indi- behavior and sabotage as well as shirking that are vidual contributions (Alchian and Demsetz, 1972). not effective in the absence of team production. While variance in worker income may stimulate Coincidentally, because of measurement difficul- ties, team production also reduces the effectiveness 6 Managers can and often do limit the disclosure of employee of behavioral strategies like increasing effort as a compensation (Lawler, 1990). Many companies require employ- ees to sign documents in which they agree not to discuss means of raising income. information about their personal compensation. However, often The underlying logic here is that in team pro- in such organizations and in many others, there is an infor- duction settings the social comparison costs that mal agreement among employees to anonymously share their salary information, which is then aggregated and shared as a accompany efforts to link pay and individual distribution. performance typically overwhelm the benefits.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger

Consequently, managers increasingly compress argue that the behavioral responses to social com- wages to reduce envy as the degree of team pro- parisons that occur within a firm are more costly duction rises. Thus, Leventhal (1976) and Deutsch than the behavioral responses to social compar- (1985) find that individuals prefer equal rewards isons that occur across firm boundaries. Second, rather than performance-based rewards when they the firm’s boundary may define an individual’s anticipate collaborating on tasks. Similarly, Frank salient reference group, which implies that altering (1985) and Konrad and Pfeffer (1990) argue that firm boundaries alters the scope of salient refer- production technologies that encourage collabora- ents, and thus the scope of social comparison and tion hinder pay based on individual contributions. its accompanying costs. We describe both mecha- Choosing an inefficient production technology nisms below. for the purpose of reducing social comparison Our first contention is that behavior stemming costs obviously imposes its own set of costs. For from invidious comparison manifests differently instance, in our illustration of International Har- within the firm than across a market interface. vester and the toy factory, management purpose- Within the firm, the presence of a central man- fully chose a technically inefficient assembly line ager shifts the costs and benefits of various behav- speed in order to avoid income dispersion that iors for reducing envious feelings. Consequently, would generate social comparison costs. Attenuat- the comparison costs imposed by these behaviors ing envy, while attempting to reward performance, accrue differently within firms than across mar- often calls for a departure from the technologically kets. When comparison occurs within the firm, the efficient choice of production technology; it often costs of both reduced effort and intense lobby- demands spatially separating workers and limit- ing to restrict others’ pay are felt very directly ing information flows to reduce social comparison. by the firm. By contrast, when a market inter- Thus, to reduce social comparison costs to a level face separates focal individuals from those with sufficient to allow aggressive pay for performance, whom they compare themselves, the comparison managers may need to alter production technology costs that these focal individuals can impose are to reduce collaboration, increase physical distance minimal. Across this market interface, there is no among employees, or actively contain information central authority who commonly assigns rewards. about productivity and income. Thus, our predic- Consequently, there is very little that those who tion is that as social comparison costs rise, man- perceive inequity across this market interface can agers are more likely to compromise the efficiency do in response. Certainly, the external manager is of production technology to mitigate these costs. unlikely to entertain pleas for cross boundary pay equity from those employed with an outside firm. To further illuminate this argument, consider two Constraining firm boundaries actors, Gen and Will. Assume they are engaged in similar tasks, that Will views Gen as a salient The final and most dramatic structural means to referent, and that Gen receives a higher income, constrain social comparison costs is to limit the which is known to Will. Because relative income boundary of the firm. While social comparison is transparent to both parties, Will envies Gen occurs both within and across firms, the scope and is motivated to take action to reduce this of such comparison and the corresponding costs perceived inequity. The critical question is how imposed play out very differently within firms than Will’s actions to reduce feelings of envy differ across firms. These differences allow management when Will and Gen are in the same versus different to use the boundary of the firm as a means of influ- firms. encing social comparison costs. The shift in social When Gen and Will are employed by the same comparison costs at the firm’s boundary occurs for firm, Will can engage in influence activities that two reasons. Either reason is theoretically suffi- affect Gen’s income or behavior. Will can lobby cient to generate our boundary prediction. First, the central manager to reduce Gen’s pay or to within the boundaries of the firm, the presence change the nature of her job and thereby indi- of the central manager who assigns compensation rectly reduce her pay. Moreover, he can affect her changes the cost-benefit analysis that individuals behavior by engaging in subtle and perhaps not-so- face in determining their various individual behav- subtle forms of retribution and sabotage. Within ioral strategies for reducing feelings of envy. We the firm, Will is likely to have close access to

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

Gen’s work location and possess information use- output or shirk to attenuate envy, which is a well- ful to Gen, which could allow him to interfere with known and common response (Foster, 1972; Elster, Gen’s tasks. Such efforts in the anthropology lit- 1991; Mui, 1995). Such behavior introduces effi- erature (e.g., Knauft 1991) are sometimes referred ciency loss because it affects the envied worker’s to as leveling and are exemplified in all of our incentive for high effort. illustrations. While policies against retribution or The boundaries of the firm not only influence sabotage may limit such behavior, the monitoring the costs imposed when rewards are compared, but required to reduce these behaviors or to monitor these boundaries also define patterns of compari- effort are clearly costly themselves. son. By defining these comparison patterns, firm In the case that Gen and Will are employed by boundaries further shape the magnitude of social different firms, Will can take actions to affect his comparison costs. Individuals within a firm seldom own behavior such as working harder or invest- view those outside the firm as salient referents and ing in his ability if it would increase income, vice versa. Individuals outside the firm are unlikely engaging in influence activities to increase his own to impose comparison costs on the focal firm, in income, or reducing his effort. Alternatively, Will part because those inside the firm are less likely to can depart from the referent group. Will has lit- compare compensation with those outside. A vari- tle opportunity to affect Gen’s behavior or more ety of arguments may explain the greater intensity importantly Gen’s income because, as Milgrom of internal comparison. Closer physical proxim- and Roberts (1988, 1990a) argue, influence activi- ity or more intense social interaction within the ties are not effective across firms. Will might con- firm provides a partial explanation, though many sider noncooperation and sabotage outside of the who are not employees of the firm may still have firm; but, engaging in these activities may lead to close physical proximity and extensive social inter- intervention by civil authorities or social approba- action with those within the firm. Organization 7 tion. scholars argue that firm boundaries shape individ- While the costs and benefits of these alternative uals’ sense of identity (e.g., Kogut and Zander, responses depend on particular circumstances or 1992). Because individuals ‘identify’ with their context, it is nonetheless clear that the costs and employer, they compare pay to those within the benefits of the alternative actions Will can take organization. The hierarchical structure in which to reduce feelings of envy shift when salient ref- pay levels are essentially ratified or endorsed sup- erents are within the firm’s boundaries rather than ports this identification with the firm. The hier- across a market interface. The primary effect of the archical structure used to allocate rewards also boundary of the firm in this context is to define creates a de facto comparison process across gaps the ways in which Will can affect Gen’s behav- in the organization, either geographic or structural, ior and rewards. This effect may also come in a which would otherwise prevent direct comparison more indirect manner. Gen, knowing that Will can among employees. While distance or the absence successfully alter her output and pay even if she of social contact may preclude direct comparisons works hard, may simply shirk in anticipation of among employees, managers compare the compen- such behavioral responses by Will. Thus, envied sation of their subordinates with the compensation workers may find it in their best interest to share of other managers’ subordinates. Equity concerns and the capacity of managers to socially com- 7 Other arrangements, of course, could exist. For instance, con- pare promote implicit comparisons among all those sider the case where Gen is a well-paid contractor working for a within the borders of the organization. As a conse- government agency for which Will is an employee. If Gen and quence, there are strong pressures to standardize Will are colocated and he views her as a salient referent, then Will may envy Gen because of income differences. The number pay policies even across geographically distinct and type of behavioral strategies available to him to reduce his organizational units (Beer et al., 1984). Thus, our envious emotions expands because of colocation (i.e., sabotage prediction is that as the social comparison costs is now made feasible). Even so, not all behavioral strategies are available to him at low cost. For instance, political influence associated with maintaining high powered incen- activities are unlikely to be effective because Gen belongs to a tives within an activity rise, managers are increas- different organization. Influencing his manager to affect Gen’s ingly likely to outsource that activity. income likely would require revisiting the contract between two organizations, which, while not fully explored here, is likely to In sum, our arguments suggest that managers be far more costly to change than the income of coworkers. face a fundamental trade-off between enduring

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger the comparison costs that accompany high pow- Comparison costs and organizational ered rewards, and enduring the productivity losses diseconomies of scale and scope that accompany compressed compensation struc- tures, suboptimal production or task structures, Our arguments, thus far, highlight the role that and compromised boundary decisions. Managers the choice of incentives, production technology, must balance the social comparison costs that arise and boundary decisions play in shaping com- from envy with the costs imposed by horizontally parison costs. Our theory suggests that for any and vertically compressing wages, adopting inef- given activity a manager examines the comparative ficient production technology, and reconfiguring governance costs, both transaction and compari- organizational boundaries. We contend that man- son costs, of several alternatives: (1) governance agement’s best response to expected social com- through the market with high powered incen- parison costs is often to use the structural levers tives, (2) governance within the firm with high at its disposal to reduce the stimuli of envy, by powered incentives, though perhaps with a pro- narrowing the set of salient referents, or shifting duction technology that reduces social compari- the behavioral strategies individuals’ employ in son, and (3) governance within the firm, but with their effort to reduce envious emotions. Figure 1 rather weak, compressed incentives. Nothing to describes the basic relationships within our the- this point, however, highlights how the selection ory. Management selects the firm’s production among these choices is influenced by the scale or technology and social architecture, its compensa- scope of the existing firm. Yet, empirically there tion and incentive structure, and its boundaries. is evidence that scale and scope do matter in These choices not only create an incentive struc- selecting among these alternatives. For instance, ture and production technology that employees empirically, large firms are more inclined to com- operate within for producing output, but also cre- press wages than small firms (Garen, 1985; Zenger, ate a particular pattern of social comparison within 1994; Rasmusen and Zenger, 1990). We argue and without the firm. Individuals respond to this below that as scale or scope increase, social com- pattern of social comparison and impose social parison costs rise and consequently firms are more comparison costs. Management expecting these likely to respond with wage compression, produc- social comparison costs or updating its selection tion technology compromise, or a limit to the scope chooses production technology, firm boundaries, of the firm. Thus, decisions about how to govern and incentives structures in a way that balances the marginal activity depend on the scale and scope the costs imposed by compromising choices of of existing activities. To advance this argument, organizational design and boundaries with social we examine a simple organization that can vary comparison costs. along two dimensions: the scope of activities that

Management actions • compress wages • craft social architecture • select firm boundaries

Social Costs imposed by comparison costs Scope and pattern ‘suboptimal’ imposed by of social comparison organization design individual behavior

Individual actions • influence activities • effort • turnover • cooperation

Figure 1. Balancing social comparison costs and suboptimal organizational design

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm it performs and the scale of (i.e., the number of) the activity (or in this case, the firm) as social participants within these activities. referents.8 Envy arises when differences in income emerge, especially among workers engaged in an activity involving team production. In our simple, single Activity scale activity firm, differences in compensation arise We begin by discussing how the scale of an activ- because of differences in the manager’s assessment ity within the firm shapes the comparison costs that of each worker’s marginal product—assessments accompany efforts to link individual pay and indi- influenced both by ability differences and by mea- vidual performance within that activity. Of course, surement error. A difference in pay between two the sensitivity of comparison costs to activity scale workers gives rise to perceived inequity and envy depends in part on the nature of the production within the lower paid worker. If we assume that technology. When jobs are completely separate feelings of envy, and hence the effort the worker is and independent, comparison costs in linking pay willing to expend to reduce these feelings increase, to individual performance are quite limited and the greater is the pay differential between the two thus increasing the scale of the workforce intro- workers, then, at an aggregate level, perceptions of duces no real scale diseconomy. Individual perfor- inequity increase as the dispersion of pay increases mance is completely observable to all and employ- among workers engaged in identical activities. ees through their behavioral strategies can impose We contend that dispersion (the difference rather few costs on the firm. At the same time, between the maximum and all others) in the when managers simply forego efforts to link indi- marginal productivity of labor, and therefore the dispersion of resulting pay among those engaged vidual performance and pay, due to high compari- in an activity, increases with the scale (or number) son costs, and instead adopt horizontally and fully of individuals engaged in that activity. Under most, compressed (flat) wages in response to high levels if not all reasonable distributional assumptions, of team production, then again there is no real scale dispersion of ability and hence dispersion in pay diseconomy present. While there is a clear degra- increases with the number of workers, because the dation in employee effort, horizontally flat wages likelihood of drawing extremely high or low abil- are fully scalable within an activity, that is, com- ity workers from the distribution increases with the parison costs per worker do not rise when wages number of workers drawn (i.e., firm scale). Based show no dispersion. However, our interest is in on this model, the more employees that a firm has, showing that when attempting to link pay to indi- the greater will be the dispersion of worker abili- vidual performance within an intermediate level of ties. Because greater dispersion in ability generates team production, activity scale increases pressures focal individuals of very high ability and there- to compress wages, adopt a less efficient produc- fore very high pay, as scale increases so does the tion technology, or even restrict the scale of an magnitude of envy and resulting social compari- activity. son costs. Thus, unless management takes steps to To illustrate, consider a firm that performs a sin- alleviate these envious feelings, the firm can expect gle activity staffed by multiple employees engaged social comparison costs from reduced effort, non- in collaborative production. Assume that the firm cooperativeness, influence activities, departures, draws workers from a labor market pool in which sabotage, or departure to increase with scale.9 the workers are heterogeneous in their ability to contribute. As a starting point for our logic, assume that the workers are paid what the firm estimates to 8 While there are clearly limits to the number of comparisons in which an individual can engage, the hierarchical structure of be each worker’s marginal product of labor. How- organizations generates de facto comparisons; workers compare ever, because output is collaboratively produced to focal workers and managers compare pay of their subordinates and inputs are imprecisely measured, the assign- with other managers’ subordinates. Our assumption is that even in the absence of direct personal contact, word of inequity travels ment of individual contributions is both somewhat rapidly through the hierarchy. In particular, word of the highest arbitrary and imprecise, which provides no veri- paid worker, what we refer to as the focal worker, travels the fiable information to counter inflated perceptions fastest. 9 While not a focus of our study, note that firms will often of their marginal contribution. Also assume that seek to actively reduce the variance in ability by screening out each worker considers all other workers within those at either the low or high tail of the organization. Some

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger

Consequently, as the scale of the activity in- from their subordinates, when they observe oth- creases, the pressure to attenuate the stimulus ers managers’ subordinates earning more than their to these rising comparison costs also increases. subordinates. Consequently, subdividing an activ- As the social comparison costs of rewarding the ity within the firm may do little to increase the marginal productivity of labor increase, the man- manager’s capacity to differentially reward these ager becomes more likely to either horizontally groups. To truly reduce comparison costs may compress pay, for instance, paying everyone the require changing the boundaries of the firm and average marginal productivity of labor, or to com- reducing activity scale. promise the technologically efficient production Thus, as the scale of an activity increases, the technology in a way that reduces interaction among social comparison costs of rewarding the marginal workers and limits their ability to view each other product of labor rise. As a consequence, the man- as salient referents. Either choice imposes its own ager must choose to either weaken incentives, set of costs on the firm. A decision to compress assigning a common level of pay to all work- pay imposes predictable costs of reduced worker ers, or simply compromise production technology. effort and adverse selection. For instance, high This compromise can take two forms: breaking ability workers are likely to exit the firm in search the firm into smaller internal units that are some- of firms that will pay them their marginal prod- what socially or geographically isolated, or more uct of labor, while low ability workers will be dramatically curtailing the boundary of the firm particularly attracted to this compensation policy itself. Thus, our prediction is that in attempting to from other firms paying the marginal product of link individual pay to performance, as the scale of labor (see Zenger, 1992). While the manager may an activity increases, social comparison costs rise elevate pay to retain these high ability workers and as a consequence compressing wages, compro- (Akerlof and Yellen, 1990), this choice has pre- mising production technology, or constraining firm dictable costs. The manager must either overpay boundaries all become more likely. for the low ability employees also attracted to the firm or heavily invest in screening procedures that reduce their number. Activity scope A decision to subdivide an activity into geo- We now turn our attention to how a firm’s activ- graphically or socially isolated groups may de- ity scope influences social comparison costs. Our crease social comparison and social comparison argument is that social comparison costs grow as costs, but at the expense of any scale economies the marginal productivity across activities becomes that arise from team production that is compro- increasingly dispersed. To explore this argument, mised by this technology choice. For instance, consider a firm that manages an array of het- subdividing workers into colocated groups may erogeneous activities that differ in their marginal decrease social comparisons across the entire orga- productivity. For convenience, assume that all nization by shaping social referents and inhibiting employees engaged in each activity receive the information flow and affect individual responses to average marginal product of labor for the activ- envy by decreasing opportunities to engage in non- ity. In essence, we offer employees a horizon- cooperative behavior. Yet, it is not clear that such tally compressed wage within activities in order to internal subdivisions are particularly effective in focus on the effects of variation across activities. constraining these social comparisons. Managers We say a firm displays scope when it integrates also are susceptible to invidious comparisons and, into two or more activities with varying average according to our view, invest in actions to reduce marginal products. The scope of the firm increases their own feelings of envy that arise from com- the more dispersed the average marginal product of parisons with other managers regarding the com- activities. Also assume that the production technol- pensation of subordinates. For instance, managers ogy of the firm requires at least some interaction experience envy or anticipate feelings of envy among the diverse activities, which implies that there exist efficiency benefits from colocation of organizations conclude that extremely high ability individuals activities and that the combination of this coloca- are simply ‘too costly’ to retain, where these costs are not merely tion and interaction of workers across the differ- the direct costs paid to the retained high ability individuals, but rather the additional comparison costs triggered throughout the ent activities ensures that the firm represents each remainder of the organization. employee’s salient reference group. Note again

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm that some of this social comparison may occur comparison costs and impose their own additional indirectly through the managerial hierarchy. costs in terms of foregone productivity. In deter- Under these assumptions, a firm with only one mining their best response, managers must con- activity incurs low social comparison costs because sider the fundamental trade-off between social pay is uniform within the activity and there is no comparison costs arising from envy and productiv- other activity with a different pay level with which ity losses coming from the adoption of inefficient to compare. When the firm integrates a second compensation and technology. Moreover, as activi- activity that differs in its average marginal product ties become more dispersed in terms of the average from the first, workers within these two activities marginal product of labor, both the comparison become the salient referent group and the poten- costs of rewarding average marginal product and tial for envy arises. We assume that the greater the the costs of vertically compressing wages rise. As a scope of activities, where scope is here defined as consequence, as activities become more dispersed the dispersion (the difference between the maxi- in their average marginal products, constraining the mum and all others) of average marginal products, vertical boundaries of the firm may become the the greater will be feelings of inequity among those effective response. workers receiving lower income. According to our Note that the increase in comparison costs logic, in response workers pursue behaviors to imposed by dispersion in the marginal productiv- reduce these envious emotions thereby imposing ity of activities is operative at both ends of the social comparison costs on the firm. Thus, feel- distribution. Managers who internalize activities ings of inequity and the corresponding comparison with low average marginal productivity activities costs increase directly with the scope of activities finditverycostlyanddifficulttomaintainlow within the firm. pay. The envy felt and the resulting comparison If disparate activities—activities with dispersed costs are reduced by placing the activity in an average marginal products of labor—are bundled organization composed of activities with more sim- within one firm, the manager faces a range of ilar average marginal productivity and more simi- options to address the potentially high comparison lar pay. Indeed, much outsourcing occurs because costs that would result from efforts to compensate managers are unable to adopt or maintain ‘mar- based on average marginal productivity. Managers ket’ wages for low productivity activities. Instead, can compress wages vertically and no longer pay social comparison costs in the form of influence the average marginal product for each activity. activities or reduced effort drive managers to ele- However, much like horizontal wage compression, vate these wages. On the other end of the spectrum, such vertical wage compression also imposes sig- managers who internalize activities with very high nificant costs on the firm. If those employed in marginal productivity also find it costly to adopt activities with high average marginal productivity market wages for these high productivity activ- of labor are paid well below that average, then ities because low marginal productivity workers those engaged in these high productivity activi- impose social comparison costs. Thus, our predic- ties are likely to exit the firm in search of firms tion is that social comparison costs associated with that do not vertically compress wages or that com- providing market-like incentives increase as the press wages less so. Our Harvard, Tenneco, and toy activity scope of the firm increases. Consequently, factory illustrations all provide evidence consistent the more divergent is an activity’s average individ- with this scenario. If the firm instead chooses to ual output from the average individual output of all simply elevate the wages of those engaged in activ- other activities within the firm, the more likely that ities with lower average marginal product of labor activity will be outsourced or the wages of those to be more similar to those with higher average within the activity vertically compressed to match marginal product of labor, this imposes its own the other activities of the firm. obvious costs of overpayment. Alternatively, the manager can constrain social Interactions between scale and scope comparison costs by departing from the technolog- ically efficient production technology, for instance, To this point in our discussion of scale and scope, isolating disparate activities and limiting what may we have examined two simplifications of the firm. be highly valuable interaction among them. Of In examining scale, we limited the firm to a sin- course, such efforts may only partially constrain gle activity. In examining scope, we limited the

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger

firm to rewarding average marginal compensation internal design of firms. We see our theory of social for each activity, that is, horizontally compress- comparison costs as playing a role in organiza- ing wages within activities. However, firm scale tional failure that is analogous to the role that and scope can interact in an important way to transaction costs play in market failure. While shape firm boundaries. In particular, we contend transaction cost economics views hierarchy as a that the social comparison costs that accompany an device to attenuate high transaction costs associ- increase in the firm’s activity scope can increase ated with market failures, social comparison cost with the aggregate scale of the firm. Thus, con- economics views the market as a device to atten- sider two single activity firms, one that is rather uate high social comparison costs associated with large with 10,000 workers and another that is quite organizational failures. Thus, while market failures small with 50 workers. When the small firm seeks create a type of centripetal force for moving activ- to add a new activity with a higher marginal prod- ities out of the market and into a firm, our theory uct of labor and to reward those engaged in that explains when and how organizational failures cre- activity, based either on average marginal prod- ate a centrifugal force for moving activities out of uct or individual marginal product of labor, the thefirmandintothemarket. 50 lower paid workers engage in comparison, feel Our theory of social comparison costs also has envy, and impose comparison costs. When the direct implications for the scope and scale of the larger firm attempts to add the same activity and firm. We argued that the larger the scale of an reward marginal product of labor, 10,000 employ- activity within a firm involving an intermediate ees engage in comparison, feel envy, and impose degree of team production, the more likely the firm comparison costs. Of course, admittedly some of is to horizontally compress wages with the accom- this comparison is indirect, because not all employ- panying reduction in incentives. Consequently, the ees will have direct contact with the newly added larger the activity, the more likely the firm is to employees. Indeed, it is likely that a greater per- partially outsource the marginal addition to that centage of employees in the small firm will have activity. Thus, a small firm is better able to offer direct contact with the newly added activity. How- workers their marginal product of labor. These ever, as we have previously discussed, the manage- incentives yield a more productive set of work- rial hierarchy, as well as informal networks within ers than larger firms. Our toy factory illustration the firm, will ensure that comparison costs, though may provide a case in point. It might have been perhaps at a diminished level, extend beyond the wiser for the toy manufacturer to outsource the toy boundaries of those who directly observe these painting activity and benefit from the high level individuals. Thus, as long as there are positive of worker productivity through a market interface comparison costs imposed by all within the bound- rather than keep the activity inside the firm, which aries of the firm, then the comparison costs of ultimately led to compressed wages, a loss of pro- adding this additional activity rise with the aggre- ductivity, and the departure of specially trained gate scale of the firm. Consequently, larger firms workers. in terms of the number of employees, whether A firm may also benefit from outsourcing when they are engaged in common or disparate activities, its necessary activities are diverse in scope. Our confront higher comparison costs in attempting to theory predicts that the firm may benefit by out- increase the activity scope of the firm than smaller sourcing those activities that are most distant in firms. As firm size increases, vertical pay compres- terms of their individual marginal product from the sion becomes more likely. Similarly, as firm size average activity. For instance, Tenneco’s acquisi- increases, management is more likely to either con- tion of HOMC internalized a set of activities with strain the boundary of the firm, and forego integra- a very different set of individual marginal prod- tion, or compromise on production technology by ucts compared to other activities within the firm. isolating geographically this newly added activity. The resulting horizontal and vertical wage com- pressions experienced by HOMC employees led to a high departure rate and, ultimately, an effi- DISCUSSION ciency loss for Tenneco. Structuring a relationship with HOMC through a contractual interface might Our theory suggests that social comparison costs have led to a superior outcome. Of course, moving play a pivotal role in shaping the boundaries and an activity outside the boundaries of the firm may

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm imply additional costs. Market failures that pro- of governance costs in considering further addi- vide the impetus for placing the activity within the tions to the firm. Because the firm’s best response firm are confronted once again should an activity policies depend on managerial diseconomies of be outsourced. For instance, Larry Summers, Har- scale and scope from activities already internal- vard University’s president, was keenly aware that ized, the boundaries of the firm are path depen- outsourcing Harvard’s investment portfolio would dent. Thus, the marginal transaction, as Coase potentially lead to lower performance, but certainly (1937) puts it, depends not only on the attributes higher fees. In the short term, Harvard decided of the exchange but also on characteristics of that outsourcing the activity was too costly and the organizations on both sides of the transaction. instead compressed wages by restricting the maxi- Those marginal activities that are more distant in mum level of compensation. When this caused the terms of their marginal products from the acquir- departure of key fund managers, increased out- ing firm’s activities will be more costly to inte- sourcing became the more efficient option. Note grate. For instance, in our illustration of Syntex and that while our illustrations have focused on dif- Varian’s joint venture, the differentiated incentive ficulties in internally organizing and rewarding structure offered to the head of the new venture activities with high individual marginal products, might have survived if Syntex had been a smaller, similar issues arise when attempting to internally entrepreneurial firm. Since the marginal productiv- organize and reward activities with low individ- ity of the joint venture appears to have been distant ual marginal products. In such instances, inter- from the rest of Syntex’s activities, the failure of nally organizing low marginal product activities selective intervention appears to be consistent with and rewarding based on marginal product gener- our theory. ates envious employees who are less productive Several limitations pervade our framework. Our than they would be if they were bundled with theory admittedly focuses on the ‘negative’ effects employees earning more similar wages. Moreover, of envy. Social comparisons lead to workers adopt- such envy and social comparison costs may cause ing behavioral strategies that impose costs on the the firm to elevate wages for the low marginal organization. Yet, envy may also stimulate a com- product activity. While this action may attenu- petitive response. Invidious comparison may cause ate social comparison costs, simply increasing pay workers to expend greater effort to increase their imposes its own obvious costs on the firm. Thus, output if they expect doing so will lead to higher in determining the optimal boundaries of the firm, income. Or, workers might (over) invest in capa- management balances the costs that accrue to mar- bilities if they have the expectation that doing so ket failures when an activity is outsourced with the leads to higher income. We anticipate that within costs that accrue to organizational failures, which limits management may be able to shape expecta- we argue are managerial diseconomies of scale and tions so that workers pursue these organizationally scope that derive from social comparison costs. ‘beneficial’ behavioral strategies to reduce envious Our discussion assumes the decision to internal- emotions. Incorporating these possibilities into our ize an activity depends on a comparison of gov- framework might lead to second order effects. For ernance costs, including social comparison costs, instance, workers might be less willing on the mar- between internal and external sourcing. In mak- gin to impose social comparison costs on the firm ing this comparison for the marginal activity, we if they believe they can reduce envy by increas- must consider the magnitude of diseconomies of ing investment in their capabilities, which would scale and scope imposed by social comparison. impact the degree to which managers could offer The magnitude of these costs depends on the set differentiated incentives. While managing worker of activities already organized within the bound- expectation might affect management’s best struc- aries of the firm and their relation to the activ- tural response to invidious comparisons, we main- ity considered for addition. Therefore, the costs tain that the basic set of relationships illuminated and benefits of integrating a marginal activity will in our framework should still hold. depend not only on the attributes of the exchange, Another concern with our framework is that as Williamson has maintained (Williamson, 1985), it relies on the conjunction of two distinct per- but also on the attributes of the firm. Firms spectives of human nature—two distinct mod- engaged in different sets of activities or a different els of man. Scholars comfortable with an emo- scale of activities will experience different levels tional model of man where individuals engage in

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger social comparison and experience negative affect indicates that social comparison and the resulting in response to inequity may be uncomfortable with costs are likely to be the greatest among those adjoining to this model a rational calculus in which workers who are the ‘closest’ salient referents. behavioral strategies are chosen to reduce envious Envy should be greatest among peers suggesting emotions. In contrast, scholars who work with a that the best response for management may be to rational model of man my feel uncomfortable with define or frame an employee’s set of peers and adjoining an emotional model of man. While fully then within this set dramatically compress wages. exploring this combination is beyond the scope of This approach leads to discontinuities in income our current study, we find that conjoining these that are linked to promotion or changes in level. models of man is both realistic and fruitful for Such discontinuities distance the promoted worker theory development. Obviously, our assumption as a salient referent thereby attenuating envy and will evoke debate and we hope stimulate further social comparison costs. discussions about various elements that should be These applications suggest that theories built on incorporated into the organization theorist’s model social comparison costs and envy may lead to a of man. theoretical synthesis that connects several frag- Our predictions focus on the use of three struc- mented and disconnected theories of human behav- tural levers for mitigating social comparison costs. ior inside of organizations. While such connec- At present, our theory does not offer specific pre- tions are presently only suppositions, the connec- dictions on when each lever is used or in what tions established by our theory suggest potential order they might be used. We also do not theorize inquiries. about the extent to which these levers are substi- tutes or complements. Such questions await further theoretical and empirical development. Our framework also offers several implications CONCLUSION that we mention below but that are beyond the scope of this study to fully develop. Our frame- work potentially opens up what may be a new Our goal has been to develop a more complete area of research about the location (i.e., produc- theory of the firm by highlighting an important tion technology choices) of workers. For instance, source of organizational failure. To do so, we why is it that some firms locate R&D distantly have explored how social comparison and envy from manufacturing or locate manufacturing dis- impose social comparison costs within firms. Man- tantly from marketing and sales? Or, why do some agers must make governance decisions cognizant firms locate workers distantly from other firms? of both social comparison costs and transaction Our theory suggests that these location choices are costs. Management’s best response to social com- strategic choices in the manager’s tool kit for shap- parison costs may involve a range of options: ver- ing salient referents. Knowledge spillover exter- tically or horizontally compressing compensation, nalities and agglomeration economies provide two adopting increasingly inefficient production tech- approaches for making activity location choices. nology, or shifting entirely the boundaries of the The consideration of envy provides another expla- firm. nation for determining these choices. An economic theory of the firm must explain Mainstream labor economics focuses on asym- both when firms arise and when markets arise. metric information and tournaments to explain Most economics theories of firm boundaries, like why promotions are accompanied by a substan- transaction cost economics, predict when hierar- tial jump in income (Baker, Gibbons, and Murphy, chies supplant markets. Like a centrifugal force, 1994; Medoff and Abraham, 1980). While asym- market failures increase the cost of using mar- metric information and tournament models have kets and push activities within firm boundaries. been used to explain this phenomena (e.g., Lazear Until now, the countervailing centripetal force that and Rosen, 1981; Gibbons and Waldman, 1999), pushed activities outside of the firm has largely we believe that wage compression within the same been assumed. Acknowledgment of social compar- level and job classification, that is, horizontal ison costs and envy, we argue, provide at least one wage compression, primarily results from social theoretical underpinning to explain these disec- comparison costs. The extant literature on envy onomies that shape a firm’s boundaries and design.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

ACKNOWLEDGEMENTS Charness G, Rabin M. 2002. Understanding social preferences with simple tests. Quarterly Journal of Economics 117(3): 817–869. We would like to thank seminar and conference Coase RH. 1937. The nature of the firm. Economica participants at the Academy of Management Meet- 4(16): 386–405. ings, Dartmouth, INSEAD, University of Southern Coase RH. 1988. The Firm, the Market and the Law. California, University of Pittsburgh, University of University of Chicago Press: Chicago, IL. Washington, the BYU/University of Utah Strategy Cropanzano R, Goldman B, Folger R. 2003. Deontic Organization Science justice: the role of moral principles in workplace Conference, and the Winter fairness. Journal of Organizational Behavior 24(8): meeting. We owe special thanks to Stuart Bun- 1019–1024. derson, Kurt Dirks, Li Ma, Hideo Owan, Brian Deutsch M. 1985. Distributive Justice: A Social Psy- Silverman, and two anonymous referees. chological Perpsective. Yale University Press: New Haven, CT. Dunlop JT. 1947. A review of wage-price policy. The Review of Economic Statistics 29(3): 154–160. REFERENCES Elster J. 1991. Envy in social life. In Strategy and Choice, Zeckhauser RJ (ed). The MIT Press: Cambridge, MA; Adams JS. 1963. Toward an understanding of inequity. 49–82. Journal of Abnormal and Social Psychology 67(5): Epstein J. 2003. Envy (The Seven Deadly Sins series). 422–436. Oxford University Press: New York. Adams JS, Freedman S. 1976. Equity theory revis- Fehr E, Schmidt KM. 1999. A theory of fairness, ited: comments and annotated bibliography. In competition, and cooperation. Quarterly Journal of Advances in Experimental Social Psychology (Vol. 9), Economics 114(3): 817–868. Berkowitz L, Walster E (eds). Academic Press: New Festinger L. 1954. A theory of social comparison York; 43–90. processes. Human Relations 7(2): 117–140. Addison JT, Burton J. 1981. On institutions and wage Foley DK. 1967. Resource allocation and the public determination. Journal of Labor Research 2(1): sector. Yale Economic Essays 7(1): 45–98. 99–109. Folger R, Konovsky MA. 1989. Effects of procedural and Akerlof GA, Yellen JL. 1990. The fair wage-effort distributive justice on reactions to pay raise decisions. hypothesis and unemployment. Quarterly Journal of Academy of Management Journal 32(1): 115–130. Economics 105(2): 255–283. Foster M. 1972. An introduction to the theory and Alchian AA, Demsetz H. 1972. Production, information practice of action research in work organizations. costs, and economic organization. American Economic Human Relations 25(6): 529–556. Review 62(5): 777–795. Frank RH. 1984a. Are workers paid their marginal prod- Baker GP, Gibbons R, Murphy KJ. 1994. Subjective ucts? American Economic Review 74(4): 549–571. performance measures in optimal incentive contracts. Frank RH. 1984b. Interdependent preferences and Quarterly Journal of Economics 109: 1125–1156. the competitive wage structure. Rand Journal of Banerjee AV. 1990. Envy. In Economic Theory and Economics 15(4): 510–520. Policy: Essays in Honour of Dipak Banerjee, Dutta B, Frank RH. 1985. Choosing the Right Pond: Human Gangopadhyay S, Mukherjee D, Ray D (eds). Oxford Behavior and the Quest for Status. Oxford University University Press: Bombay; 91–111. Press: New York. Barzel Y. 1982. Measurement costs and the organization Garen JE. 1985. Worker heterogeneity, job screening, of markets. Journal of Law and Economics 25: 27–48. and firm size. Journal of Political Economy 93(4): Beer M, Spector B, Lawrence PR, Mills DQ, Walton RE. 1984. Managing Human Assets. Free Press: New York. 715–739. Blau PM. 1955. The Dynamics of Bureaucracy: A Study of Gibbons R, Waldman M. 1999. The theory of wage and Interpersonal Relations in Two Government Agencies, promotion dynamics inside firms. Quarterly Journal University of Chicago Press: Chicago, IL. of Economics 114(4): 1321–1358. Bolton GE, Ockenfels A. 2000. ERC: A theory of equity, Goel AM, Thakor AV. 2007. Optimal contracts when reciprocity, and cooperation. American Economic agents envy each other. Paper presented at The Review 90(1): 166–193. American Economic Association Meeting (January), Brennan G. 1971. Horizontal equity: an extension of an Chicago, IL. Available at SSRN: http://ssrn.com/ extension. Public Finance 26(3): 437–456. abstract=779384. Brenner R. 1987. Rivalry: In Business, Science, Among Greenberg J. 1982. Approaching equity and avoiding Nations. Cambridge University Press: Cambridge, inequity in groups and organizations. In Equity and UK. Justice in Social Behavior, Greenberg J, Cohen RL Cason TN, Mui VL. 2002. Fairness and sharing in (eds). Academic Press: New York; 389–435. innovation games: a laboratory investigation. Journal Greenberg J. 1990. Organizational justice: yesterday, of Economic Behavior and Organization 48(3): today, and tomorrow. Journal of Management 16(2): 243–264. 399–432.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj J. A. Nickerson and T. R. Zenger

Hall BJ, Lim J. 2003. Incentive pay for portfolio (Vol. 3), Cummings LL, Staw BM (eds). JAI Press: managers at Harvard Management Company. Case Greenwich, CT; 53–107. 902–130, Harvard Business School, Boston, MA. McKersie RB. 1967. Payment by results systems in the Hamermesh DS. 1975. Interdependence in the labour United States. In Forms of Wage and Salary Payment market. Economica 42(168): 420–429. for High Productivity, Supplement to Final Report, Harvard Magazine. 2004. John Harvard’s Journal: International Management Seminar, Versailles, France ‘Extraordinary’ bonuses. 106(4): 69–73. (26–29 September); 215–236. Hennart JF. 1994. The ‘comparative institutional’ theory Medoff J, Abraham K. 1980. Experience, performance, of the firm: some implications for corporate strategy. and earnings. Quarterly Journal of Economics 95: Journal of Management Studies 31(2): 193–207. 703–736. Hicks JR. 1955. Economic foundations of wage policy. Meyer HH. 1975. The pay-for-performance dilemma. Economic Journal 65: (September): 390–404. Organizational Dynamics 3: 39–49. Holmstrom B. 1989. Moral hazard and observability. Milgrom P, Roberts J. 1988. An economic approach to In Uncertainty in Economics: Readings and Exer- influence activities in organizations. American Journal cises (Economic Theory, Econometrics, and Mathe- of Sociology 94(Supplement): S154–S179. matical Economics) and Information (revised edn). Milgrom P, Roberts J. 1990a. Bargaining costs, influence Diamond P, Rothschild M (eds). Academic Press: costs, and the organization of economic activity. In New York; 327–344. Perspectives on Positive Political Economy,AltJ, Homans GC. 1961. Social Behavior: Its Elementary Shepsle K (eds). Cambridge University Press: New Forms. Harcourt, Brace & World: New York. York; 57–89. Kaldor N. 1934. The equilibrium of the firm. The Milgrom P, Roberts J. 1990b. The efficiency of equity Economic Journal 44(173): 60–76. in organizational decision processes. American Klein B, Crawford RG, Alchian AA. 1978. Vertical Economics Review 80: 154–159. integration, appropriable rents, and the competitive Milgrom P, Roberts J. 1992. Economics, Organization, contracting process. Journal of Law and Economics and Management. Prentice-Hall: Englewood Cliffs, 21(2): 297–326. NJ. Knauft BM. 1991. Violence and sociality in human Mui VL. 1995. The economics of envy. Journal of evolution. Current Anthropology 32(4): 391–428. Economic Behavior & Organization 26: 311–336. Kogut BM, Zander U. 1992. Knowledge of the firm, Pencavel JH. 1977. Work effort, on-the-job screening, competent to capabilities, and the replication of and alternative methods of remuneration. In Research technology. Organization Science 3: 383–397. in Labor Economics: An Annual Compilation of Konrad AM, Pfeffer J. 1990. Do you get what Research, Ehrenberg RG (ed). JAI Press: Greenwich, you deserve? Factors affecting the relationship CT; 225–258. between productivity and pay. Administrative Science Pfeffer J, Davis-Blake A. 1992. Salary dispersion, Quarterly 35: 258–285. location in the salary distribution, and turnover among Kulik CT, Ambrose ML. 1992. Personal and situational college administrators. Industrial and Labor Relations determinants of referent choice. Academy of Manage- Review 45(4): 753–763. ment Review 17(2): 212–237. Pfeffer J, Langton N. 1993. The effect of wage Lawler EE III. 1990. Strategic Pay: Aligning Organiza- dispersion on satisfaction, productivity, and working tional Strategies and Pay Systems (Jossey-Bass Busi- collaboratively: evidence from college and university ness and Management Series). Jossey-Bass: San Fran- faculty. Administrative Science Quarterly 38(3): cisco, CA. 382–407. Lazear EP, Rosen S. 1981. Rank-order tournaments as Pruitt DG, Kimmel MJ. 1977. Twenty years of experi- optimum labor contracts. Journal of Political Economy mental gaming: critique, synthesis, and suggestions for 89(5): 841–864. the future. Annual Review of Psychology 28: 363–392. Leventhal GS. 1976. The distribution of rewards and Rasmusen E, Zenger TR. 1990. Diseconomies of scale resources in groups and organizations. In Advances in in employment contracts. Journal of Law, Economics, Experimental Social Psychology (Vol. 9), Berkowitz L and Organization 6(1): 65–92. (ed). Academic Press: New York; 91–131. Salovey P. 1991. Social comparison processes in Long CP, Bendersky C, Morrill C. 2007. Reaping what envy and jealousy. In Social Comparison: Con- is sown: the impact of managerial controls on temporary Theory and Research, Suls J, Wills TA subordinates’ fairness evaluations. Working paper, (eds). Lawrence Erlbaum Associates: Hillsdale, NJ; John M. Olin School of Business, Washington 261–285. University, St. Louis, MO. Sheppard BH, Lewicki RJ, Minton JW. 1992. Organiza- Ma L, Nickerson JA. 2007. Worker envy and strategic tional Justice: The Search for Fairness in the Work- organization. Working paper, John M. Olin School of place. Lexington Books: New York. Business,Washington University, St. Louis, MO. Silver M, Sabini J. 1978. The perception of envy. Social Martin J. 1981. Relative deprivation: a theory of Psychology 41(2): 105–117. distributive injustice for an era of shrinking resources. Smith PC, Walker JW. 2000. Layoff policies as a In Research in Organizational Behavior: an Annual competitive edge. Competitiveness Review 10(2): Series of Analytical Essays and Critical Reviews 132–145.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj Envy, Comparison Costs, and the Economic Theory of the Firm

Sraffa P. 1926. The laws of returns under competitive Williamson OE. 1975. Markets and Hierarchies.The conditions. The Economic Journal 36(144): 536–550. Free Press: New York. Strauss G. 1955. Group dynamics and intergroup Williamson OE. 1985. The Economic Institutions of relations. In Money and Motivation: An Analysis of Capitalism. The Free Press: New York. Incentives in Industry, Whyte WF, (ed). Harper & Williamson OE. 1991. Comparative Economic Organiza- Row: New York; 90–96. tion: The Analysis of Discrete Structural Alternatives. Suls JE, Wills TA (eds). 1991. Social Comparison: Administrative Science Quarterly 36: 269–296. Contemporary Theory and Research.Lawrence Zenger TR. 1992. Why do employers only reward Erlbaum Associates, Inc.: Hillsdale, NJ. extreme performance? Examining the relationships Tyler TR, Rasinski K, McGraw K. 1985. The influence among performance, pay, and turnover. Administrative of perceived injustice on the endorsement of political Science Quarterly 37(2): 198–219. leaders. Journal of Applied Social Psychology 15: Zenger TR. 1994. Explaining organizational disec- 700–725. onomies of scale in R&D: agency problems and the Varian HR. 1974. Equity, envy, and efficiency. Journal allocation of engineering talent, ideas, and effort by of Economic Theory 9: 63–91. firm size. Management Science 40(6): 708–729. Walster E, Berscheid E, Walster GW. 1973. New direc- Zenger TR, Hesterly WS. 1997. The disaggregation of tions in equity research. Journal of Personality and corporations: selective intervention, high-powered Social Psychology 25: 151–176. incentives, and molecular units. Organization Science White GL, Mullen PE. 1989. Jealousy: Theory, 8(3): 209–222. Research, and Clinical Strategies. Guilford Press: Zenger TR, Marshall CR. 2000. Determinants of incen- New York. tive intensity in group-based rewards. Academy of Williams RM. 1975. Relative deprivation. In The Idea Management Journal 43(5): 141–162. of Social Structure: Papers in Honor of Robert K. Zizzo DJ, Oswald A. 2001. Are people willing to pay Merton, Coser LA (ed). Harcourt, Brace, Jovanovich: to reduce others’ incomes? Annales d’Economie et de New York; 355–378. Statistique 63–64: 39–62.

Copyright  2008 John Wiley & Sons, Ltd. Strat. Mgmt. J. (2008) DOI: 10.1002/smj