THE STRATEGIC MANAGEMENT OF THE ENTITLEMENT...

Strategic Management Journal, Vol. 14, 75-93 (1993)

THE STRATEGIC MANAGEMENT OF THEENTITLEMENT PROCESS IN THE EMPLOYMENTRELATIONSHIPCHIP HEATH, MARC KNEZ, and COLIN CAMERERGraduate School of Busitiess, University of Chicago, Chicago, Illinois, U.S.A.

Over time, members of organizations develop entitlements—preferences about how theywish to be treated and beliefs about how they should be treated. The formation ofentitlements is an important subject for strategy researchers because employees resist changesthat violate their perceived entitlements; thus entitlements constrain an organization's abilityto adapt quickly in a changing environment. In this paper, we use psychological researchto propose a two-part model of entitlements formation: (i) Preference formation makespeople likely to resist change because preferences adapt to experience, and thus changeimposes painful losses; (ii) Belief formation leads to over-entitlement, and this producesresistance because employees perceive changes to be unfair or unjust. Over-entitlementhappens because (a) psychological limitations in judgement and (b) strategic distortions inthe character and content of information exchanged in relationships lead employees toperceive their entitlements as richer and more systematic than intended by the organization.Combined, the preference and belief formation processes can produce substantial resistanceto change.In stable environments firms may have an incentive to allow entitlements to develop sincethey enhance employee security and commitment. However, in changing environments,entitlements constrain firms' ability to mobilize resources to meet competitive challenges.After presenting our model of entitlement formation, we use the model to organize andanalyze a set of suggestions about how the employment relationship might be managed toavoid problems of entitlement formation, thus enabling firms to adapt more effectively ina dynamic environment.

INTRODUCTION

In the employment relationship, as in all relation-ships, we develop expectations about the relation-ship. We learn what the other party expects us tocontribute and we develop notions about what weshould receive in the relationship. These implicitor explicit expectations constitute the perceivedcontract in a relationship, and the benefits webelieve we deserve under the contract constituteour entitlements.

Key words: Entitlements, loss aversion, behavioraldecision theory, communication, incomplete contracts

The formation of entitlements is an importantsubject for strategy researchers becauseentitlements constrain the ability of organizationsto adapt to new competitive situations. Becauseemployees resist changes that violate their perceivedentitlements, if entitlements are not aligned withbusiness needs then organizations will experiencea substantial constraint on their ability to deploythe talent and attention of their workforce.

Conventional wisdom holds that the currentbusiness environment is becoming more dynamicand competitive, and that in response, organizationsmust be better able to learn and adapt to theirexperience in order to survive (Senge, 1990;Garvin, 1993; Guest, 1986; Beer, Eisenstat, and

CCC 0143-2095/93/100075-19© 1993 by John Wiley & Sons, Ltd.

76 C. Heath, M. Knez, and C. Camerer

Spector, 1990, Lawler, 1992). In dynamic environ-ments, it is more important to effectively managethe entitlements formation process because theability to move quickly and deploy resourcesfreely becomes an important source of strategicadvantage. For organizations to respond quickly,employees must be more flexible in their workstyles and tasks, must respond appropriately whenincentives change, and must be willing to absorbmore fluctuations in rewards as the environmentchanges.

In this paper, we present a model of entitlementsformation. By understanding how entitlements areformed, we can better understand resistance tochange, and also better understand how to preserveflexibility and adaptability. While the problemsof organizational change have been extensivelydiscussed, the emphasis has typically been onmanaging one-time transitions. These models takea reactive rather than a proactive stance towardresistance. When changes occur more frequently,it may be more effective to short-circuit theentitlements formation process and avoid resistancefrom the start. In this paper, we use the model ofentitlements formation to suggest not how resist-ance might be overcome, but how resistance mightbe avoided.

This paper is complementary to other strategyresearch on information processing and rent-seeking. We focus on the information processingstrategies that employees use to understand theirentitlements. In one way, our work concerns rent-seeking, or value-creation, in di factor market—themarket for labor—rather than rent-seeking inproduct markets. However, at another level, ourcomments on managing the entitlements processconcern the firm's ability to keep its workforcemotivated during times of change, which is essentialto carry out the search for product market rents.We think of our approach as strategic becausemany important modern strategy studies focus onrelations with customers or suppliers; we turnthese focuses inward, toward the firm's relationswith its workers as suppliers of labor.

AN OVERVIEW OF THE MODEL

Underlying any employment contract are basicnotions of employee motivation. Both expectedutility theory in the economic literature andexpectancy theory in the psychological literature

(Vroom, 1964) assume that employees are willingto take action when they expect that thoseactions will lead to a particular outcome thatthey find intrinsically or extrinsically desirable,Under these descriptions of employee motivation,the employment contract defines what actionsthe employer expects from the employee andwhat intrinsic and extrinsic benefits the employeeexpects from the employer (Rousseau and Parks,1992). In an ideal world such a contract wouldcompletely describe the expectations of bothparties in all relevant states of the world.Of course, bounded rationality prevents suchcontracts from being formed. Instead, employ-ment contracts are incomplete, and each sidefills in missing information as best they can.

In the next section, we present a model ofhow employees fill in gaps in the explicit contractto form perceptions of their entitlements. Wealso describe how the entitlements formationprocess can lead to resistance to change. Ourmodel has two parts, one describing how prefer-ences are formed, and one describing how beliefsare formed. First, in describing the preferenceformation process, we argue based on thepsychology of preference formation that peopleadapt to their experience (Kahneman, Knetsch,and Thaler, 1990; Zajonc, 1980), and that thepain of a loss from a status quo is much morepainful than an equivalent sized gain is pleasurable(Kahneman and Tversky, 1979). Since changetypically involves a loss, or a mixture of painfullosses and less-pleasurable gains, people will tendto resist change. Second, in describing the beliefformation process, we argue that because ofsystematic biases in the psychological process ofbelief formation by employees, and because ofstrategic distortions in the content and characterof information offered by the organization,employees will typically come to see theirentitlements as richer and more systematic thanwould a neutral observer. In turn, employeeswill resist changes, not only because they arepainful, but also because they will perceive thechange to violate the implicit contract, and thusto be unjust or unfair.

Our definition of entitlements as 'the benefitsthat people believe they deserve under the implicitcontract' highlights the fact that entitlements arebeliefs, and it may seem unnecessary to includepreference formation with belief formation aspart of the entitlements formation process. We

Entitlement Formation 11

do so because preference and belief formationare confounded in most situations. On average,people develop beliefs to justify the things theyprefer. It would provoke cognitive dissonance(Festinger, 1957) for people to prefer somethingthey do not deserve, and therefore at the sametime their preferences adapt to experience,people also tend to develop beliefs to justifytheir preferences. Secondly, people try not todevelop preferences for things they believe theydo not deserve or things that will not persist.For example, people may not allow their prefer-ences to adjust to unexpected or unusual benefits(e.g., Kahneman and Thaler, 1991). Thus, webelieve that preferences and beliefs will be highlyrelated in most situations, and therefore find ituseful to consider preference formation as a partof the process of entitlements formation.

In a stable environment, the entitlementformation process will not cause problems andwill perhaps add value to the relationship, givingpeople a sense of security and comfort. However,in a dynamic or changing environment,entitlements formation will make it difficult forrelationships to change. When change occurs,one party is likely to accuse the other ofmisrepresentation or unfairness, and resist thechange. After presenting the model of theentitlements formation process, we use the modelat the close of the paper to suggest waysorganizations might manage entitlements forma-tion to produce more fiexibility in the workrelationship. For example, if people develop apreference for the status quo (their current salary,responsibilities, etc.) and resist switching, it mayhelp to make the status quo less salient. Currentmoves toward 'broadbanding' in job titles seemto do precisely this by reducing the number ofjob titles available in the firm and thus makingit easier to move people from job to job withouta perceived 'loss' in status when titles change.

THE ENTITLEMENTS FORMATIONPROCESS

Preference formation

People often experience change as painful. Wesuggest that change is painful because preferencesadapt to experience and because losses from astatus quo are heavily penalized.

In search of simplicity, economic literature

typically assumes that preferences are exogenousand stable. While we assume that preferencesare endogenous and unstable, we assume theyadapt to experience in a simple and predictableway: people come to prefer what they experienceand they learn to like what they have (Zajonc,1980; Thaler, 1980). Preference adaptation mayoccur for almost any aspect of experience aslong as it is initially neutral or positive (tasks,perks, monetary benefits, status). This assump-tion is not new—it is implicit in much work oninertia in the face of change, and certain economicmodels (Constantinides, 1990; Cyert andDeGroot, 1987). However, over the last fewyears, researchers have made progress in under-standing the magnitude of the effects andin modeling the phenomenon (Tversky andKahneman, 1991).

Preference adaptation is troublesome whenchange is necesary because of another principleof psychological value: losses are more painfulthan gains are good (Kahneman and Tversky,1979). A variety of evidence indicates that losses(cost increases, benefit decreases) are about twotimes more painful than gains are pleasurable. Forexample, in laboratory settings which eliminateincentives to price strategically, the owners ofan item typically request about twice as much togive something up as a neutral observer is willingto pay to acquire it (Kahneman et al., 1990). It isremarkable that this 'endowment effect' (Thaler,1980), is easy to produce in laboratory exper-iments simply by giving people mundane con-sumer goods like coffee mugs or pens. Lossaversion is likely to be even more salient formore meaningful goods like job tasks, titles, orbenefits.'

Loss aversion makes change difficult at thefrontier—after preferences adapt to a currentstate, it takes a large increase in benefits on onedimension to compensate for a sacrifice onanother (Tversky and Kahneman 1991). Imagine,for example, an employee who is indifferentbetween two jobs at the time he or she is hired.One job (Job T) offers greater travel opportunity(a feature that this employee values) and anotherjob (Job P) offers slightly more pay. After

' Most researchers treat loss aversion as a relatively physiologi-cal phenomenon, but it is clear that other cognitive or socialfactors may exacerbate the effect (e.g., dissonance- Festinger,1957).


accepting either job, the combitiation of prefer-ence adaptation and loss aversion implies thatthe employee will resist a move to the other jobthat he or she previously found equally attractive.Before employment, the two jobs were equivalentbecause gaining an opportunity to travel was justas attractive as gaining some additional pay.After taking the job with more travel, giving upthe travel would be a loss, and that loss willoutweigh the gain of the extra salary associatedwith Job P. After taking the job with the highersalary, the loss of the extra salary could not beovercome by the possibility of gaining extratravel offered by Job T. When organizations areleft to offer people trade-offs at the frontier,employers may find their employees are surpris-ingly unsatisfied with the menu of options offeredto them.

Because preferences adapt to experience, andbecause losses impose relatively heavy hedonicpenalties, change will typically be unattractive.Change is more painful when people have moretime to adapt to and enjoy the status quo. Themore features of their experience that they like,the harder it is for any change to leave thembetter off and the greater the potential for changeto harm them.

Belief formation

The pain associated with change sets the stagefor resistance. However, people are more likelyto resist changes that are perceived as unfair.Researchers in procedural and distributive justicehave pointed out that people care a great dealabout the fairness of outcomes: they withdrawtheir enthusiasm and commitment from relation-ships they perceive as unfair (Greenberg, 1988),they actively resist unfair changes (Bies, 1987),and they are even willing to sacrifice personalgain to sanction those who are unfair (Kahneman,Knetsch, and Thaler, 1986; Lowenstein, Thomp-son, and Bazerman, 1989). In order to predictwhether people will treat change as an unfairviolation of an entitlement, we must understandwhat people believe they are entitled to underthe contract.

Helping people form accurate beliefs abouttheir entitlements would be easy if completecontracts could be written. However, as wementioned earlier, complete contracts are impos-sible to write because people are boundedly

rational: their foresight is incomplete and thetask of imagining and reconciling all potentialproblems exceeds their patience and mentalfacilities.^ Instead, contracts tend to be looselydefined and people develop beliefs about thecontract over time. We suggest that this processof belief formation is likely to lead to over-entitlement—that is, people will perceive theirentitlements as richer and more systematic thanwould a neutral observer. When change takesplace, over-entitlement will lead people to feelthat the implicit contract has been violated, andwill cause them to perceive changes as unfairand unjust.

In this section, we will consider two majorreasons for over-entitlement. First, over-entitlement occurs because people are cognitivelyand motivationally predisposed to form positiveviews of the content and stability of the employ-ment contract. Second, over-entitlement occursbecause the dynamics of relationships may leadorganizations to transmit only positive infor-mation about the status of the relationship. Eachof these biases—one driven by an incompleteprocess of inference on the part of individualemployees and one driven by incentives todistribute overly positive information on the partof organizations—work to ensure that employeeswill expect their outcomes to be more positivethan they are likely to be. People will seethemselves as more secure, more valuable, andentitled to more benefits than they should basedon an objective observers' view of their historyor of their current environment. Because of theseperceptions, change is likely to be surprising andnegative, and is likely to leave people feelingthat the relationship contract has been violated.

Over-entitlement due to the psychology of beliefformation

In this section, we examine how the psychologyof belief formation may lead to over-entitlement.We argue that the way that people form beliefswill lead employees to see their entitlements as

^ There is a significant amount of theoretical research studyingthe impact of incomplete contracts on the structure ofeconomic relationships. The works of Coase (1937), William-son (1985) and Klein, Crawford, and Alchian (1978) areamong those credited with motivating this research. Ourinterest is in the behavioral implications of incompletecontracts within the employment relationship.


richer and more systematic than intended bytheir employers.

RicherPerceived entitlements will be rich, for example,because people typically believe that they contributea great deal to their relationships, and in turnbelieve that they deserve equal contributions fromthe other party. However, people are likely tohave an exaggerated view of what would constitutean 'equal' contribution from the other party, sincethey are aware of and remember each of theirown contributions to a relationship, but are notaware of, nor do they remember, each of thecontributions of their partners. Empirical evidenceshows, for example, that if you add up individuals'perceived contributions to joint products, the totalgenerally exceeds 100 percent (Ross and Sicoly,1979); this is true for spouses' reports of theircontributions toward housework and for academics'contributions to scientific papers. In a world whereexchanges are largely premised on reciprocity orequity (Walster, Walster, and Berscheid, 1978),peoples' tendency to over-estimate their contri-butions may lead them to expect a great deal fromthe other party at exactly the same time that theother party is expecting a great deal from them.

Perceived entitlements will also be rich becausepeople tend to use their knowledge of themselvesto fill in gaps in their knowledge of others. In theabsence of concrete information to the contrary,people tend to assume that others see things likethey do and that others value the same things thatthey do (Ross, Greene, and House, 1977). Thesetendencies may lead people to overestimate theamount of motivation and knowledge they sharewith the other party in a relationship. On average,people will expect their companies and theirsuperiors to agree with them on what projectsshould be done, how they should be done, andwhat benefits they should receive for doing them.Their expectations are likely to be overly positivesince the actual desires of their partners will oftendiverge from their own.

SystematicEntitlements will be more systematic than theyshould be because people may not anticipateexceptional situations where entitlements mightchange, and because they may not recognizeexceptional situations when they occur. Due tolimited information processing ability, people

will generally not anticipate all the unusualsituations that may arise in the course of arelationship—people typically have limited timeto consider all possible alternatives and limitedability to imagine them (Simon, 1947). Forexample, people tend to treat very low probabilityevents as impossible and very high probabilityevents as being certain (Kahneman and Tversky,1979). Although any individual low probabilityevent will not occur very often, the combinedportfolio of unlikely events leads people toexperience a large number of surprises. Forexample, across a number of domains, whenpeople are certain that they are right about apiece of information or about their prediction ofan event, they are actually wrong about one timein five (Fischhoff, Slovic, and Lichtenstein, 1977).In addition, for strategic emotional reasons,people may be particularly unwilling to imaginenegative events. People typically assume thatbad events are relatively unlikely to happen tothem (Taylor and Brown 1988); for example,people say that they will be less likely than theirpeers to have an automobile accident, becomethe victim of a crime, have trouble finding a job,or become ill or depressed. Because they do notthink about all possible states of the world (andespecially because they do not think aboutnegative states), people may not have an adequatepicture of the conditional nature of theirentitlements.

Entitlements may also be overly systematicbecause people may assume too much consistencyin the actions of others or the environment. Inanalyzing the world, people typically assume thatthe world will act as it did before (Kahnemanand Tversky, 1973). In analyzing their relationshippartners, people tend not to focus on aspects ofthe situation that caused their partners to act ina particular way, and tend to attribute actionsto personality or traits rather than to the situation(Jones and Davis, 1965, Ross, 1977). This maylead people to overestimate the amount ofconsistency in other's behavior since they do notprocess the root causes of that behavior. Ifbehavior is driven by personality, it will bestable, if driven primarily by situations it willchange when the situation changes. For example,this principle says that employees who join amajor accounting firm in the fall are likely toattribute a relaxed work environment to theattitude of their boss or the culture of their


company. When tax season approaches, theymay be shocked and disappointed to find thattheir 'mellow' boss (or company) has suddenlybecome frenetic and demanding. Because theybelieve the world will be consistent, people mayalso not develop an adequate understanding ofthe conditional nature of their entitlements.

Perceived entitlements may also be too system-atic because limited information processing mayprevent people from recognizing exceptionalsituations when they occur. Since limited infor-mation may lead their picture of the world tobe over-simplified, instead of recognizing thenuances of changing circumstances, people maysee a great deal of similarity between discrepantsituations. Since similarity judgments underliemany judgments of fairness—if one situation islike another then the same rewards ought to begiven, the same procedures followed, the sameoptions granted—people with an overly-simplisticpicture of the world may feel that they havebeen treated unfairly if they do not receive thesame treatment in two situations that theyperceive as similar (e.g., Levinthal, 1975). Thus,perceived entitlements will be overly systematic,not only because people may not anticipateexceptional situations, but also because they maynot recognize them when they happen.

Based on psychological research on learning andjudgment, we have suggested that belief formationwill lead to over-entitlement—^people will perceivetheir entitlements as richer and more systematicthan would a neutral observer. This process ofbelief formation is likely to cause problems indynamic relationships. For example, since peopleassume that others share their picture of the worldand have the same standards of correct andadequate behavior, when people are disappointedby their relationship partner, they may suspectthat their partner's motives are impure rather thanthe partner's picture of the world is different. If Ibelieve that my partners see the world like I do,then they could only disappoint me because theylack concern for my feelings, or even worse,because they are intentionally disdainful or dis-honest (Kramer, 1993).

Over-entitlement as a result of strategicinformation flows

The problems above arise from errors of inferenceand learning on the part of employees. In this

section we argue that employees may haveinflated expectations about their employmentcontracts as a reasonable response to biasedinformation they receive from their employer.Employers may strategically misrepresent infor-mation to their employees in an attempt tomaintain employee motivation. Broadly speaking,this misrepresentation may occur at two levels—at the upper levels of the firm where decisionsare made about the strategic direction of thefirm and how resources will be allocated, and atthe lower levels of the firm where employeesreceive feedback about their individual perform-ance.

At the upper levels of the firm, top managementhas private information about how they intendto allocate resources across the firm in the shortand long term. Depending on how and whetherthis information is conveyed, employees maydevelop a false sense of security from observingthe actions of the firm. Consider a generic firmwhich can find itself in three possible states: astatus-quo state which requires no change in thefirm's activities, a bad state which requirescutbacks to be made (e.g., a significant andpermanent fall in demand) and a danger state,which indicates that cutbacks might have to bemade if the state persists (e.g., a small drop indemand which may or may not be permanent).For simplicity, we assume that upper managementcan distinguish all the states of the world, whileemployees are be able to make only partialdistinctions. As discussed in the previous section,limited information processing ability is likely tocause employees' picture of the world to be over-simplified. For example, employees may be ableto distinguish major differences between states,like differences between good situations andsituations that are not good, but have difficultydistinguishing subtle differences, like thosebetween danger states and bad states (e.g.,between temporary and permanent falls indemand).

While the occurrence of a danger state makesa bad state more likely, upper management mayhave an incentive to avoid taking action during

' Anecdotal evidence and theoretical arguments (Zajac andBazerman, 1991) point out that upper managers are subjectto the same kinds of psychological distortions. In the currentargument, we only need to assume that upper managershave a slightly more complex picture of the world thanemployees.


a danger state. For example, the firm may avoidmaking temporary layoffs since the layoffs woulddecrease the remaining employees' sense ofsecurity and thus their level of effort (Brockner,1988)." By not responding to a danger state, thefirm keeps employees from becoming anxious ordiscouraged, and focuses their efforts on the taskat hand rather than on looking for alternativeemployment. However, by not taking negativeactions in the danger state, employees mayassume that the firm will not take negativeactions in 'similar' states in the future. Becausethey cannot distinguish danger states from badstates, the firm's lack of action in the danger stateleads employees to misperceive their stability. Ifthe firm is eventually forced to take action (whenthe danger state develops into a bad state),employees will view it as a violation of theemployment contract because from theemployee's view, the current situation is nodifferent than previous situations where the firmdid not take action.

In the Appendix we show how this kind ofinformation omission might be modeled in arational game-theoretic setting. In the model inthe Appendix, employees know that they maybe in a danger state, but do not reduce theireffort because they know that management hasan incentive to take the same action in both thestatus-quo state and the danger state. Thus, eventhough employees know their perceptions mightbe wrong, they retain their current beliefsbecause they don't know for sure whether theirperceptions are wrong. Although this model caneffectively capture the firm's incentives to distortcommunication, it is difficult to model over-entitlement using game-theoretic models. Thesemodels assume mutual rationality—workersunderstand the firm's incentives to distort infor-mation. In the behavioral model we present,employees react negatively to changes becausetheir picture of the world is overly-simplistic,and changes come as a surprise. Because therational model does not allow employees to be

•* Some evidence suggests that the effort effect can go bothways (Brockner et al., 1992). Companies tend to report thatmorale falls after cutbacks are made, but some companiesreport that employees work harder, hoping to keep theirjobs. In either case, managers express reluctance to take anyactions which might lead to a decline in morale (Bewley andBrainard, 1993), so managers' stated predispositions areconsistent with those we assume in the model.

surprised, in order to get over-entitlement, wehave to add the assumption that employeesbelieve that management is obligated to take thesame action in the future that they took in thepast.

Thus, strategic information withholding canlead employees to believe they are secure underthe current contract, when in fact they are not.The firm creates these beliefs by effectivelyignoring the existence of danger states throughtheir lack of action in those states. In the longrun, this practice causes workers to resist changebecause workers believe they are entitled to betreated today the same way they were treated inthe past. Employers may not only be temptedto withhold information, they may also activelybias their communication in a way that leads toincreased entitlements. To ensure the commit-ment of workers, employers may, for example,paint an overly rosy picture of the likelihoodthat a factory, work group, or company unit willexist in the future.

At lower levels of the firm, misrepresentationmay also lead employees to feel overly secure.Consider, for example, the relationship betweenan employee and the supervisor responsible forevaluating his or her performance. To preserveemployees' commitment to, and enthusiasm forthe job, supervisors may be reluctant to giveworkers negative feedback about their perform-ance,^ and in the extreme, may be prone to giveoverly positive feedback. Supervisors may alsogive positive feedback to preserve the socialharmony of the relationship—relationships aremore pleasant when the parties in the relationshipvalue each other. Social norms are designed topreserve signals of respect and liking even whenpeople do not respect or like each other. To theextent that supervisors do like the people theysupervise, it is natural for them to hedgecriticisms.

Combined, the desire to be nice and the desirenot to damage commitment or enthusiasm leadto striking biases in communication. This biasmanifests itself in the well-known tendency ofsupervisors to give compressed performancerankings. For example, Merck & Co. reported

' For example, Tesser and Rosen (1976) document thatpeople are reluctant even to pass along bad news to theperson affected by it. The situation is even more difficultfor supervisors who are responsible for the negative feedback.


that, on a 5-point scale where five indicatedsuperior performance, 93 percent of the work-force was given ratings from 3 to 4+ (Murphy,1991). These compressed ratings fail to adequatelyinform employees how management views theirperformance under the employment contract.Moreover, in a time of change when theorganization is forced to reallocate rents, it willbe difficult to use past performance as a faircriteria to redistribute rents.

Thus in this section we have argued thatcommunication biases may lead to over-entitlement. Upper management has an incentiveto distort feedback about the stability of theenvironment; supervisors have an incentive todistort feedback about individual performance.Because of these biased communications,employees may develop overly optimistic expec-tations about the stability of the world and abouttheir importance to the organization.

Why aren't the mistakes repaired?

Our description above paints an extreme pictureof the potential distortions in psychologicalcontracts. Our model argues that the psychologyof belief formation and the strategic nature ofcommunication lead employees to feel overlyentitled to benefits. It would be legitimate toask how this effect could persist over time. Whydo employees not realize that their pictures oftheir entitlements are exaggerated and overlysystematic?

The first answer is that the environment is adifficult one to learn in. As the last sectiondiscussed, the social environment often dedicatesitself to conveying misleading feedback. Also,the events that would require updating aretypically rare. Situationally, since change eventsare typically unusual, people may 'learn' fromtheir experience that they are relatively safe andsecure: 100-year-floods occur, on average, onlyonce every hundred years, and bankruptciesusually occur only once in the lifetime of acompany.

The second answer is that people may not beaware of the limitations in their assumptions.The same information processing limitations thatproduce systematic positive errors in perceivedentitlements may also hinder people from realiz-ing the extent of their assumptions. Althoughthey are likely to have an over-simplified picture

of the contract, people may assume that theirpicture is accurate and complete. People arefrequently unaware of the assumptions they aremaking (Nisbett and Wilson, 1977), and as aresult they have difficulty diagnosing the sourceof problems when they take action based on amistaken assumption. For example, even experts,when asked to examine a fault tree which listsall the likely causes of a certain event (e.g.,asking auto mechanics to critique a list of all thepotential reasons a car might not start) tend tofocus on the information at hand, and fail tonotice what is missing (Fischhoff, Slovic, andLichtenstein, 1978). When people are not awareof their mistakes in belief, they may not takeadequate care to elaborate the contract or torespond to information in the environment.

The third, and perhaps most important, answerwhy errors will not be repaired is motivationalrather than cognitive; people may not be moti-vated to find out about limitations in theirperceptions of the contract. In stable environ-ments, over-entitlement is not dangerous, and infact may be helpful. Our lives are more pleasantwhen we can imagine that bad things are unlikelyto happen to us, that our relationship partnersshare our thinking about important events, andthat we contribute our fair share (and perhapsmore than our fair share) to our relationships.It is also pleasant to receive feedback fromothers that we are valued and that our work iscompetent and important. However, in changingsituations, these beliefs create unrealistic expec-tations and these unrealistic expectations maymake us less willing to adapt to changingcircumstances.

Resistance to change

In general, the process of entitlement formationwill cause difficulty for organizations because itleads to substantial resistance to change. Themodel of entitlement formation points out tworeasons why change will be resisted. First thepreference formation process points out thatchange will generally be painful. Preferencesadapt to experience, so we come to prefer thestatus quo. Since change typically imposes losseson some dimensions, change will typically be anegative experience because the psychologicalutility function penalizes losses more than itrewards gains.


Second, change will be resisted because thebelief formation process will lead people to seechange as unfair. Because people's view of theirentitlements will be overly rich and systematic,when a painful change causes people to reexaminetheir relationship, they are likely to concludethat changes are wrong or unfair. Since theirview of the psychological contract did notanticipate the changes, either the previousrelationship was 'misleading' or the changes bythe other party did in fact violate the contract.In reaction to the perceived violations, peoplewill withdraw their commitment and enthusiasmfrom the relationship, and may actively resist theefforts of the unfair partner (Bies, 1987). Indeed,when treatment is viewed as sufficiently unfair,the most committed people are likely to havethe strongest negative reactions since theiridentity is most invested in the relationship(Brockner, Tyler, and Cooper-Schneider, 1992).

In our model of entitlement formation, thepreference formation process leads people toexperience change as painful, and the beliefformation process leads people to perceivechanges as unfair. Thus, we assume that peoplewill resist change because changes will typicallypile insult on injury—over-entitlement will causepeople to see changes as insulting and unfair,loss aversion will cause people to experiencechanges as injurious and painful. Combined,the preference and belief formation processesproduce substantial resistance to change.

APPLICATION OF ENTITLEMENTFORMATION TO DYNAMICENVIRONMENTS

In this section we exercise our model ofentitlement formation by suggesting some waysto manage entitlement formation in dynamicenvironments. When researchers talk aboutorganizational change, they are often referringto a one-time change at a discrete point in time,in which the organization makes a transitionfrom one stable state to another. Kurt Lewin'smetaphor for the change process is often invoked:the change agent must 'unfreeze' a currentbehavior, 'change' the behavior to the preferredbehavior, and then 'refreeze' the new behavior(Lewin, 1947). In this section we explore theproblem of managing entitlements formation in

a dynamic environment where an organization isnot looking at a one-time change or transition,but faces a need to continually adapt in achanging environment. We thus explore ways toprevent behavior from refreezing in a state thatwill later have to be modified again.

Whether or not the reader chooses to believe(according to the conventional business wisdomcited in the introduction) that the current businessenvironment is more dynamic than before,the recommendations below will be useful formanaging dynamic environments wherever theyare found, for example, in industries or competi-tive contexts where change is more salient andfrequent, or within a firm for employees whoexperience frequent change: e.g., employees thatare on a development path that will involve anumber of short-term assignments.

In general, entitlements formation can beaddressed by addressing either component ofentitlement formation: organizations can attemptto manage preference formation and/or beliefformation. We will propose specific examples toillustrate how each aspect of the process couldbe managed. Most of the examples are notnovel—they have been proposed in either thepopular management literature or the academicliterature. However, the model of entitlementformation provides a useful psychological foun-dation to organize a number of suggestionsfrom diverse domains, and to understand whyindividual proposals may be psychologicallysound.

We also note here that while these followingexamples demonstrate ways that organizationsmight overcome problems of preference or beliefformation, each 'remedy' carries its own costs. Indeciding whether to implement these suggestions,organizations must weigh the benefits of managingpreference or belief formation against otherpsychological or administrative costs.

Managing preference formation

Preference adaptation causes people to experi-ence changes in benefits as a loss. If benefits aremanaged correctly, it may be possible to awardbenefits without having preferences adapt tothose benefits.

We will discuss two potential situations whereexplicit process management may prevent prefer-ences from adapting to benefits. First, people

84 C Heath, M. Knez, and C. Camerer

may not adapt to benefits that are segregated ina separate mental account. For example, it isprobably substantially less painful to lose a'bonus' than to lose an equivalent fraction ofbase salary (Kahneman et al., 1986). Second,people may not adapt to benefits that are notsalient. Since people may experience a loss whenany concrete feature of experience (e.g., a title)disappears, one way of avoiding losses is to avoidcreating or calling attention to such features(e.g., emphasizing the importance of titles).

Segregating benefits

One way of managing the preference adaptationprocess is to encourage a process that occursnaturally—the tendency to separate out resourcesinto separate mental accounts (Thaler, 1985).Even when resources are highly fungible, peoplereact differently to different resources, for exam-ple, people exhibit a high propensity to consumeregular income but a low propensity to consumeincome from investments or home equity (Shefrinand Thaler, 1988). In order to prevent preferenceformation, organizations may want to take careto encourage this process by helping people tosegregate idiosyncratic or infrequent benefits.

When people separate out a novel benefit intoa separate mental category, their preferencesmay be less likely to adapt to that benefit.Because preferences do not adapt, people avoida negative contrast or loss when the benefitdiminishes or disappears. For example, Kahne-man (1992, Kahneman and Thaler, 1991) suggeststhat cultural norms succeed in managing prefer-ence adaptation and contrast effects by creatinga class of 'special' occasions which allow extrava-gant consumption. By labeling an occasion as'special,' and thus mentally segregating thebenefit, people may be able to enjoy a birthdaymeal or anniversary celebration at a greatrestaurant without diminishing their enjoymentof their next dinner out at an average restaurant.By highlighting the unusual status of a 'specialtask force assignment,' employers may be ableto temporarily reward an employee with a moreinteresting task without decreasing the employee'senjoyment of the regular job once the task forceassignment ends.

By encouraging people to mentally segregatebenefits, an organization can help people resistdeveloping preferences for those benefits. One

organization with a substantial internationalbusiness has a special benefits package that isawarded to staff when they are away fromtheir own country. The package includes taxcompensation, ex patriot pay boosts, andadditional travel allotments for travel to andfrom their home country. The organization'sstrategy for minimizing endowment and contrasteffects is to take great care to label the extrabenefits as the 'International Staff BenefitsPackage,' and administer that package separatelyfrom the normal salary and benefits of employ-ment. Human resources personnel hold pointeddiscussions with new international staff dis-tinguishing between the International Staff Bene-fits Package which will disappear when theemployee returns home, and the regular benefitspackage which will continue. The discussions arenot entirely successful—the company complainsthat it remains difficult to get people to rotateback to their home country after experiencingthe rich benefits package. However, the exampledoes point to steps that might make it possibleto grant special benefits while mitigating thepossibility that those benefits will automaticallybe added to an endowment.

The current move to increase the amount ofpay awarded in the form of bonuses providesanother example of a management approach thatmay successfully manage preference formation.There are many reasons for this move other thanmanaging preferences. In a world of fiatterorganizations, promotion opportunities may nolonger be available to provide appropriate incen-tives to employees, and organizations mustreward performance in a more immediate waythan the old system which tied increased pay topromotion up the hierarchy to better-payingjobs (Baker, 1990). Bonus systems tie pay toperformance more effectively, providing immedi-ate rewards for appropriate behavior (which maychange rapidly as the environment shifts) withouttying the organization's resources to an outdatedincentive system.

Despite these other advantages, perhaps themost compelling arguments for bonuses arederived from psychological observation(Kahneman and Thaler, 1991). There is evidencethat bonuses have a different mental accountingstatus than regular income. For example, bonusmoney is more likely to be saved than regularincome, even when it is awarded consistently in


consistent amounts (Ishikawa and Ueda, 1984).Given that bonuses seem to be psychologicallyseparate, bonuses may allow the organizationmore flexibility than increases in base salary(merit increases, etc.) because they do notcommit the organization to a continued streamof pay. Increases in base salary are very likelyto be added to a person's endowment, andtherefore reductions in base salary are likely tobe treated as a loss. Many observers have notedthat nominal wages are sticky downward (Blinderand Choi, 1990). However, the magnitude of thephenomenon is surprising. In a recent study ofemployees at a large firm (Baker, Gibbs, andHolmstrom, 1993), 20 years of data involving49,276 salary changes recorded only 25 obser-vations where employees received a nominal cutin wages.

If bonuses are mentally segregated, it may beeasier to award and remove them than to awardand remove increments to base salary. Insteadof experiencing a loss (as would an employee ifsalary were cut), the employee merely experiencesthe reduction of a gain, an experience which ispsychologically much less painful (Kahnemanand Tversky, 1979).

Bonuses are not a cure-all. If the process isnot managed properly, people may adapt to theincome associated with bonuses, and treat theabsence of a bonus as a loss instead of a foregonegain, producing the same difficulty as changes inbase salary. Officials at one large financialinstitution say that employees reacted verynegatively to a decline in ESOP earnings duringa period of organizational distress—they hadbeen treating the ESOP earnings which wereprorated and placed in their paychecks as a partof their salary. By not effectively segregating theESOP earnings from the regular paycheck, thecompany made a strategic mistake in allowingthe employees to integrate the bonuses into theirendowment.

Perhaps the most difficult part of managingbonuses is that beliefs must be managed moreclosely. Because bonuses tie pay more explicitlyto performance, organizations must work veryhard to make sure that employees understandthe mapping between the reward and theiractivities. If, as we argue, people's model of therelationship is oversimplified, employees maynot discriminate between situations where theyshould receive the bonus and situations where

they should not receive it. People in organizationsare often quite unwilling to sacrifice their bonuseswhen their unit or organization is doing poorly—they reason that they have been doing 'their job'and are relatively unwilling to sacrifice becauseanother group has been doing poorly. In addition,because people are aware of their own contri-butions and not those of others, people mayfrequently believe their bonus is inadequate giventheir evaluation of their own performance vs.others'.

Making the status quo less salient

A second strategy to manage preferences is toavoid drawing employees' attention to aspects ofthe environment that may change and to avoidenriching the environment with features thatmight disappear at a later time. The existenceof a special title, special service, or specialprogram may draw people's attention to aspectsof the social environment that they wouldn'thave noticed otherwise. Once they notice, thenthey experience a loss when the title, service, orprogram disappears. These issues are particularlyacute with symbols of status, since organizationshave a strong ability to attach status to manyfeatures of the environment.

Many companies, for example, complain thatit is difficult to get people to make 'lateral'moves that might broaden their experience andincrease their long-run qualifications for seniorpositions (Jones, Braddick, and Shafer, 1991).They complain that people resist moves becausethey are too sensitive in the short run to subtledifferences in the status of job titles. There seemto be two main reasons for employee resistance,first, the 'career ladder' focuses people onmoving upwards, and accepting even a completelyequivalent position and title may be a lossrelative to aspirations; second, because titlesmake so many fine distinctions, it is difficult tofind another job title for a 'lateral' move that isin fact psychologically equivalent.

In response to this problem in mobility, somecompanies are reclassifying job titles into a fewbroad bands (Jones et al., 1991). Typically, thenumber of titles under 'broadbanding' are reducedby one half to two-thirds. For example, a firmwith 10 salary ranges, each with an associatedtitle, would reduce its number of salary ranges(and titles) to five. Each of these new ranges


would have a much greater spread between theminimum and maximum salary levels within theband (the goal is not to minimize salary spread,but to divorce salary from titles). One of theprimary goals of broadbanding is to increase thelevel of employee mobility and skills acquisition.By reducing visible job classifications, it ispossible for people to move from job to jobwithout worrying about a 'loss of status' becauseof the change in title.

While broadbanding is consistent with manag-ing preference formation, note that our modelof entitlement formation indicates that thisorganizational change will be difficult toaccomplish in companies that have historicallyhad more titles. In these companies, reducingjob classifications will deprive some employees—those who formerly had titles in the top half ofthe revised band—of part of their perceivedentitlement. One company that we know reportedbroad resistance to broadbanding by employeeswith higher than average job classifications.These employees perceived that their position inthe firm had been devalued, and hence that theyhad lost a benefit they had worked hard toobtain.

Managing belief formation

The section above gave two general kinds ofsuggestions about how to manage preferenceformation. The section below suggests two generalways to manage belief formation: (1) throughexplicit up-front analysis and negotiation, and(ii) through greater attention to ongoing com-munication.

Better up-front negotiation

The most frequently suggested remedy to theproblems of belief formation is to solve theproblem up front by making expectations, duties,and costs more explicit and by addressing andreconciling potential problems. Advice to soon-to-be-married couples includes recommendationsto explicitly discuss topics that are commonsources of disagreement (e.g., finances, child-rearing practices). Human resources guidebookstypically specify explicit policies for a class oftopic areas (absences, health days, paid andunpaid leave) which seem likely to cause prob-lems. Explicit discussions clarify otherwise-

ambiguous terms of the employment contract,and thus make it easier for employees to learnabout the work relationship and harder foremployers to misrepresent the relationship.

Evidence indicates, for example, that 'realisticjob previews,' which accurately describe thenegatives as well as the positives of jobs, actuallyreduce turnover (Premack and Wanous, 1985).Because they explicitly clarify the content of thejob for prospective employees, employees maybetter manage their own process of preferenceformation—learning not to treat unusual benefitsas expected or unusual negatives as habitual.Because negatives are anticipated, setbacks areless likely to cause employees to reevaluate therelationship. The fact that Realistic Job Previewsreduce turnover implicitly indicates that strategicinformation withholding does take place—employers must make specific plans to conveynegative information.

Explicit discussions are a reasonable solutionwhen it is possible to identify potential problemsand establish procedures to solve them. However,it is not possible to resolve all problems up frontbecause of the limitations in foresight discussedearlier. When people systematically ignore lowprobability events or events with negative pay-offs, contracts will remain incomplete. It is evenharder for people to write a complete contractin a dynamic environment, since the complexityof the environment makes it more difficult forpeople to imagine all contingencies and decideon appropriate actions.

Another way to make contracts more explicit,while avoiding the necessity of writing detailedagreements, is to make formal or informal pre-commitments about the scope and character ofthe employment relationship that make it moredifficult for the organization to strategicallymiscommunicate. For example, companies mayhire employees into a job or a status whichmakes certain benefits or rewards unattainable;some companies hire contingent workers to avoidthe trauma of laying off permanent employees.The strong informal norm against tenuringassistant professors at some universities clarifiesthe incentives for all parties: a senior professorcannot seduce an assistant professor into heavycommittee work by 'promising' that the workwill help their tenure case. Economic models haveargued that 'up or out' promotion mechanisms area good way to precontract to award workers


their marginal product (Kahn and Huberman,1988; Prendergast, 1993). These mechanisms mayalso serve to de-bias the information process.

Another solution that makes contracts moreexplicit, but that avoids problems of boundedrationality, is to use company norms or valuestatements to express general guiding principlesfor the relationship. Cultural norms may providea way of filling in gaps in the incomplete contract(Kreps, 1984; Camerer and Vepsalainen, 1990).Because they are general, cultural norms willnot handle every problem, but they mighteffectively guide decisions at difficult choicepoints, providing a coordinating mechanismfor firms and employees to renegotiate theirunderstanding of the agreement on an ongoingbasis. A company that cannot promise permanentemployment might offer 'permanentemployability'—making a general commitment todevelop people so that they are attractive tooutside employers if their original employercannot offer continuing employment. Companiesmight promise that they will offer employees 'achallenge, not a position.' An expressed concernwith 'continuous improvement' might also serveto signal to employees and managers that theyshould not expect the current job descriptionsand tasks to persist over time.

More ongoing information

When contracts cannot be effectively clarifiedup-front, then organizations must devote moreeffort to ongoing communication to managebeliefs as the relationship evolves. As discussedearlier, the combination of psychological infer-ence and strategic information flow may leademployees to feel entitled to work at their currentpace and enjoy the benefits associated with theircurrent job. Employees generally feel that theyare making an important contribution to theorganization—researchers have documented thatmost employees, most of the time, think thatthey are performing above average (Meyer,1975).

It is not surprising that employees resistchanged or increased job requirements whenthey think that their performance is already morethan adequate. Employees may find it confusingor unjust when they are asked to change whatthey thought was satisfactory performance. Indynamic environments where required behavior

is changing, people may need more explicit andfrequent communication to help them modifytheir picture of the world (McCann and Gilkey,1988). Because learning new behaviors is difficult,without sufficient warning that their skills orperformance are lagging, people may not devotesufficient time and attention to learn new skillsor modify their performance.

Our earlier discussion pointed out two areaswhere strategic information biases exacerbatedover-entitlement—at the bottom level of thecompany where people get feedback on theirperformance, and at the top level of the companywhere people get feedback about the stability ofthe environment. More accurate feedback aboutperformance helps people anticipate what actionsthey need to take in order to perform well, andallows them to learn what benefits they shouldexpect. More accurate feedback about theenvironment also helps people think about whatkinds of rewards they can expect in a changingenvironment. Managing belief formation in adynamic environment requires more carefulcommunication at both levels.

At lower levelsOrganizations can, for example, commit toevaluate performance in a veridical or timelymanner. For example, at the lower level of theorganization, organizations can prevent managersfrom avoiding performance evaluations byimposing a certain time schedule on evaluations.Many companies, for example, commit to haveyearly performance reviews for all employeesand more frequent reviews for new employees.Organizations can also take steps to makethe content of performance evaluations moreveridical. For example, organizations can forcemanagers to use a forced ranking system whichmakes it impossible for a manager to giveeveryone positive feedback. Alternatively, firmscan make evaluations more veridical by tyingcertain benefits to evaluations. For example,promotion reviews in academics are typically tiedto tenure. The practice of tying benefits toevaluations minimizes the incentives for super-visors to distort feedback because giving anemployee an adequate review requires that thesupervisor award large benefits contingent onthat verdict. Some have suggested that this isone reason that wages may be tied to jobs(Prendergast and Topel, 1993).


Given the magnitude of positive illusionsabout performance, veridical performanceevaluations are likely to contain bad news formost employees. Employees are likely to bedemoralized after hearing that their supervisorand their organization feel that they arecontributing less than they thought. On average,since employees receive news that is not aspositive as they would have expected, feedbackforces a negative reevaluation of the relation-ship, and raises concerns that the evaluator isbeing unfair. Aware of these issues, supervisorsare likely to resist delivering bad news: reducingthe positive bias requires evaluators to devotetime and effort to make the evaluations moreobviously fair (and thus justifiable to theemployee) and to manage the consequences ofdisabused expectations.

Because veridical performance evaluations aredifficult to administer, organizations should thinkseriously about how often and with what level ofdetail to evaluate their employees. In a stableenvironment, employees' illusions and manage-ment's reluctance to communicate bad newsprobably do not cause problems—employees feelsecure and valued and the firm may be able tocarry few marginal performers. In a more dynamicenvironment, the advantages of giving people moreimmediate feedback about their behavior mayoutweigh the disadvantages of disappointment andthe high costs of creating an evaluation systemthat wUl be perceived as fair.

At upper levelsIn a dynamic environment, it is also moreimportant to carefully communicate how thefirm stands within its environment. Accurateinformation about the environment may helpemployees manage their preferences andbeliefs; if, for example, employees know thatthe company is undergoing difficult times, theymay be less likely to develop beliefs that theconipany owes them promotions and less likelyto allow themselves to become comfortablewith perks or benefits that may disappear ifhard times continue. A variety of evidenceindicates that people can manage their prefer-ences and their coping resources more effec-tively when they have prior warning that anevent may happen (Krantz, Grunberg, andBaum, 1985). When bad states occur, employeeswith an overly positive sense of security will

be worse off than employees that had a morerealistic notion of their security.

Employers could precommit to share certainkinds of information. For example, openingcompany books may serve as an early warningsystem to insure against overly positive infor-mation flows. By sharing information on anongoing basis, companies can insure that theyare not tempted to hide information when badstates occur. The continual openness makesinformation more credible, and provides areputational incentive for continued opennesseven in bad states. If the organization commitsto convey information, then the absence of theinformation becomes a clear signal of badintentions.

The need for more communication creates adilemma, since there will be many moresituations where an organization may experi-ence a loss than situations where the organiza-tion actually experiences a loss. Employeeswho are warned of potential changes may bemore willing and able to change when the timecomes, but they may also be more stressed onan ongoing basis as they worry about potentialchanges.

Thus organizations must consider how muchinformation to communicate to employees in'danger' states. Note that our argument impliesthat full communication is not always good. Inrelatively stable, unchanging environments,withholding bad news is probably not a badstrategy—parents infrequently tell their childrenabout all possible dangers in the world, andspouses do not divulge every occasion wherethey were upset with each other. In a stablebusiness environment, it may be reasonable toallow employees to develop a strong sense ofsecurity because the organization will seldombe forced to change in a way that imposeslosses on the employees.

In general we can think about two differentstrategies of communication, paternalistic andcommunicative. Paternalistic managers shelteremployees from bad news and have employeeswho are overly optimistic about the world.When losses happen that force entitlements tobe cut, the employees of these managers willbe particularly upset and unadaptive. On theother hand, communicative managers whoalways inform employees about where thecompany stands will have less devastated


employees in the loss states because they havehad a chance to anticipate the bad events, butat the cost of having stressed and anxiousemployees in many states of the world wherebad events might happen.

In this section we have suggested two waysto manage belief formation—through betterup-front negotiation and through better ongoingcommunication. In general, belief formationmay be harder to manage than preferenceformation. Belief formation is hard to manageup-front because we have limited abilities toimagine the contingencies necessary to writecomplete contracts, and it is hard to manage onan ongoing basis because of the disappointment,stress and anxiety that often results from thetransmission of veridical information. However,firms in dynamic environments may not have achoice about whether or not to manage beliefs;active management may be necessary to avoidnegative reactions from employees who willtend to see changes as violations of the implicitcontract.

CONCLUSION: THE DILEMMA OFMANAGING IN DYNAMICENVIRONMENTS

Strategy researchers have recognized that thenewly dynamic, competitive environmentrequires organizations and their employees tobe more fiexible and adaptive. In order toadapt, employees must be more fiexible andadaptive. Some have argued that organizationscan achieve competitive advantages by changingtheir management approach to take moreadvantage of the skills and problem solvingabilities of workers (Kanter 1983; Beer et al.,1990).

The management literature over the last fewyears has noted that companies need to offera stable work environment in order to encourageemployees to respond fiexibly and creatively(e.g., Lawler, 1992). The demands onemployees in dynamic, competitive environ-ments are high. As Lawler (1992) states, 'Highinvolvement management techniques place sub-stantial demands on employees in terms oftheir ability to solve problems, contribute togroup discussions, and, of course, perform awide array of technical work-related activities

that contribute to the organization's basiceffectiveness.' It seems reasonable and fair thatin order for employees to be willing to adaptto new roles and to participate actively in theimprovement process, companies should offeremployees the security of not being fired onthe basis of their help in improving a process.

However, it is important to recognize thatthe demands for adaptability and stability arecontradictory—tbe same dynamic environmentthat makes it necessary for employees to bemore fiexible may also make it impossiblefor companies to offer a stable employmentrelationship. Dynamic environments requirefiexible workers, fiexible workers require astable employment relationship, but stableemployment relationships may not be possiblein dynamic environments.

We have no suggestions for resolving thisdilemma, indeed it is probably not resolvable.As indicated by our suggestions in the previoussection, we do note, however, that firms willfind it more costly to manage the employmentrelationship in a dynamic environment. Whileover-entitlement is costly for firms in a dynamicenvironment because it hinders their abilityto change, it benefits companies in stableenvironments by increasing employees' security.Employees who feel assured about their futurewith a firm do not have to be paid as much,are more willing to make sacrifices for thefirm, and are willing to allow paternalisticmanagers to withhold information from them.As security decreases in a dynamic environment,companies will have to invest more resourcesinto managing employment relationship. Com-panies will have to provide more explicit socialand monetary insurance against bad changes inthe environment. This may involve payinghigher wages, bearing costs of general trainingin addition to company-specific training, orcommunicating more. In each case, costs willincrease.

We have proposed a model of the entitlementsformation process that assumes that people willresist change because of preference and beliefformation. We also used the model to organizea set of suggestions that organizations mightuse to manage these processes in a dynamicenvironment. By better understanding theentitlement formation process, managers inorganizations will be in a better position to


ensure that their organizations remain able toadapt and change.

ACKNOWLEDGEMENTS

We thank Joel Brockner, Richard Cyert, RobGertner, Canice Prendergast, Denise Rousseau,and Sally White for comments on earlier draftsof this paper.

REFERENCES

Baker, G. (1990).'Pay for performance for middlemanagers: Causes and consequences'. Journal ofApplied Corporate Finance, 3 (3), pp. 50-61.

Baker, G., M. Gibbs and B. Holmstrom (1993). 'Thewage policy of a firm'. Working paper. HarvardUniversity and University of Chicago.

Beer, M., R. A. Eisenstat and B. Spector (1990).The Critical Path to Corporate Renewal, HarvardBusiness School Press, Boston, MA.

Bewley, T. and W. Brainard (1993). 'A depressedlabor market, as explained by participants'. Workingpaper. Yale University.

Bies, R. J. (1987). 'The predicament of injustice: Themanagement of moral outrage'. In L. L. Cummingsand B. Staw (eds.) Research in OrganizationalBehavior, Vol. 9. JAI Press, Greenwich, CT,pp. 289-319.

Blinder, A. S. and D. H. Choi (1990). 'A shred ofevidence on theories of wage stickiness'. QuarterlyJournal of Economics, 105, pp. 1002-1015.

Brockner, J. (1988). 'The effects of work layoff onsurvivors: Research, theory, and practice'. In L.L. Cummings and B. Staw (eds.). Research inOrganizational Behavior, Vol. 10. JAI Press,Greenwich, CT, pp. 213-256.

Brockner, J., T. R. Tyler and R. Cooper-Schneider(1992). 'The influence of prior commitment to aninstitution on reactions to perceived unfairness: Thehigher they are, the harder they fall'. AdministrativeScience Quarterly, 37, pp. 241-261.

Brockner, J., S. Grover, T. F. Reed and R. L. Dewitt(1992). 'Layoffs, job insecurity, and survivors' workeffort: Evidence of an inverted-U relationship'.Academy of Management Journal, 35, pp. 413-425.

Coase, R. H. (1937). 'The nature of the firm',Economica, 4, pp. 386-405.

Constantinides, G. M. (1990). 'Habit formation: Aresolution of the equity-premium puzzle'. Journalof Political Economy, 98, pp. 519-543.

Cyert, R. M. and M. H. DeGroot (1987). BayesianAnalysis in Uncertainty in Economic Theory. Roman& Littlefield, Totowa, NJ.

Festinger, L. (1957). A Theory of Cognitive Dissonance.Stanford University Press, Stanford, CA.

Fischhoff, B., P. Slovic and S. Lichtenstein (1977).

'Knowing with certainty: The appropriateness ofextreme confidence', Journal of Experimental Psy-chology: Human Perception and Performance, 3,pp. 552-564.

Camerer, C. and A. Vepsalainen (1988). 'The economicefficiency of corporate culture'. Strategic Manage-ment Journal, 9, pp. 115-126.

Fischhoff, B., P. Slovic and S. Lichtenstein (1978).'Fault trees: Sensitivity of estimated failure prob-abilities to problem representation'. Journal ofExperimental Psychology: Human Perception andPerformance, 4, pp. 330-334.

Friedman, A. and Z. Shapira (1988). 'Employeeinformation and beliefs about promotion decisions'.Unpublished working paper. Hebrew University.

Garvin, D. A. (July-August 1993). 'Building a learningorganization'. Harvard Business Review, pp. 78-91.

Greenberg, J. (1988). 'Cultivating an image of justice:Looking fair on the job'. Academy of ManagementExecutive, 11, pp. 155-157.

Guest, R. H. (1986). 'Management imperatives forthe year 2000', California Management Review, 28,pp. 66-70.

Ishikawa, T. and K. Udea (1984). 'The borius paymentsystem and Japanese personal savings'. In M. Aoki(ed.). The Economic Analysis of the Japanese Firm.North-Holland, Amsterdam.

Jones, E. E. and K. E. Davis (1965). 'From acts todispositions: The attribution process in personperception'. In L. Berkowitz (ed.). Advances inExperimental Social Psychology, Vol. 2. AcademicPress, New York, pp. 220-266.

Jones, M. B., C. A. Braddick and P. M. Shafer(1991). 'Will broadbands replace traditional salarystructures?', Journal of Compensation and Benefits,7, pp. 30-35.

Kahn, C. and G. Huberman (1988). 'Two-sideduncertainty and up-or-out contracts'. Journal ofLabor Economics, 6, pp. 423-444.

Kahneman, D. (1992). 'Reference points, anchors,norms, and mixed feelings'. Organizational Behaviorand Human Decision Processes, 51, pp. 1-17.

Kahneman, D., J. L. Knetsch and R. Thaler (1990).'Experimental tests of the endowment effect andthe Coase Theorem', Journal of Political Economy,97, pp. 1325-1348.

Kahneman, D., J. L. Knetsch and R. Thaler (1986).'Fairness and the assumptions of economies'.Journal of Business, 59, pp. S285-S300.

Kahneman, D. and R. Thaler (1991). 'Economicanalysis and the psychology of utility: Applicationsto compensation policy'. The American EconomicReview Papers and Proceedings, 81, pp. 341-346.

Kahneman, D. and A. Tversky (1973). 'On thepsychology of prediction'. Psychological Review,80, pp. 237-251.

Kahneman, D. and A. Tversky (1979). 'Prospecttheory: An analysis of decisions under risk',Econometrica, 47, pp. 263-291.

Kanter, R. M. (1983). The Change Masters. Simon &Schuster, New York.

Klein, B., R. G. Crawford and A. A. Alchian (1978).


'Vertical integration, appropriable rents, and thecompetitive contracting process'. Journal of Lawand Economics, 21, pp. 297-336.

Knez, M. (1993). 'Delayed downsizing and the efforterosion problem'. Working paper. University ofChicago.

Kramer, R. (1993). 'The sinister attribution error'.Working paper, Stanford University.

Kreps, D. M. (1984). 'Corporate culture and economictheory'. Working paper, Stanford University.

Krantz, D. S., N. E. Grunberg and A. Baum (1985).'Health psychology', Annual Review of Psychology,36, pp. 349-383.

Lawler, Edward E. (1992). The Ultimate Advantage:Creating the High-Involvement Organization, Joss-ey-Bass, San Francisco, CA.

Levinthal, G. S. (1975). 'The distribution of rewardsand resources in groups and organizations'. In L.Berkowitz and E. Walster (eds.). Advances inExperimental and Social Psychology, Vol. 8. Aca-demic Press, New York, pp. 91-131.

Lewin, K. (1947). 'Group decision and social change'.In E. E. Maccoby, T. Newcomb and E. Hartley(eds.). Readings in Social Psychology. Holt, Rine-hart, and Winston, New York, pp. 459-473.

Lowenstein, G. F., L. Thompson and M. H. Bazerman(1989). 'Social utility and decision making ininterpersonal contexts'. Journal of Personality andSocial Psychology, 57, pp. 426-^1.

MacLeod, B. and J. Malcomson (1989). 'Implicitcontracts, incentive compatibility, and involuntaryunemployment', Econometrica, 57, pp. 447-480.

McCann, J. E. and R. Gilkey (1988). Joining Forces:Creating and Managing Successful Mergers andAcquistions. Prentice Hall, Englewood Cliffs, NJ.

Meyer, H. H. (1975). 'The pay-for-performancedilemma'. Organizational Dynamics, 3, pp. 39-50.

Murphy, K. J. (1991). 'Merck & Co., Inc. (A)',Harvard Business School case 9-491-005.

Nisbett, R. E. and T. D. Wilson (1977). 'Telling morethan we can know: Verbal reports on mentalprocesses'. Psychological Review, 84, pp. 231-259.

Premack, S. L. and J. P. Wanous (1985). 'A meta-analysis of realistic job preview experiments'.Journal of Applied Psychology, 70, pp. 706-719.

Prendergast, C. (1993). 'The role of promotionin inducing specific human capital acquisition'.Quarterly Journal of Economics, 108, pp. 523-534.

Prendergast, C. and B. Topel (1993). 'Favoritismin organizations'. Working paper. University ofChicago.

Ross, L. (1977). 'The intuitive psychologist and hisshortcomings: Distortions in the attribution process'.In L. Berkowitz (ed.). Advances in ExperimentalSocial Psychology, Vol. 10. Academic Press, NewYork, pp. 174-220.

Ross, L., D. Greene and P. House (1977). 'The falseconsensus phenomenon: An attributional bias inself-perception and social perception processes'.Journal of Experimental Social Psychology, 13,pp. 279-301.

Ross, M. and F. Sicoly (1979). 'Egocentric biases inavailability and attribution'. Journal of Personalityand Social Psychology, 37, pp. 322-336.

Rousseau, D. M. and J. M. Parks (1992). 'Thecontracts of individuals and organizations'. In L.L. Cummings and B. Staw (eds.). Research inOrganizational Behavior, Vol. 15. JAI Press,Greenwich, CT, pp. 1-43.

Senge, P. M. (1990). The Fifth Discipline. Doubleday,New York.

Shefrin, H. M. and R. H. Thaler (1988). 'Thebehavioral life-cycle hypothesis'. Economic Inquiry,26, pp. 609-643.

Simon, H. A. (1947). Administrative Behavior. Macmil-lan, New York.

Taylor, S. E. and J. D. Brown. (1988). 'Illusion andwell being: A social psychological perspectiveon mental health', Psychological Bulletin, 103,pp. 193-210.

Tesser, A. and S. Rosen (1976). 'The reluctance totransmit bad news'. In L. Berkowitz and E. Walster(eds.). Advances in Experimental Social Psychology,Vol. 9. Academic Press, New York, pp. 193-232.

Thaler, R. H. (1980). 'Toward a positive theory ofconsumer choice'. Journal of Economic Behaviorand Organization, 1, pp. 39-60.

Thaler, R. H. (1985). 'Mental accounting and consumerchoice'. Marketing Science, 4, pp. 119-214.

Tversky, A. and D. Kahneman (1971). 'Belief in thelaw of small numbers'. Psychological Bulletin, 2,pp. 105-110.

Tversky, A. and D. Kahneman (November 1991).'Loss aversion in riskless choice: A referencedependent model'. Quarterly Journal of Economics,pp. 1039-1061.

Vroom, V. H. (1964). Work and Motivation. Wiley,New York.

Walster, E., G. W. Walster and E. Berscheid (1978).Equity. Allyn & Bacon, Boston, MA.

Williamson, Oliver E. (1985). The Economic Insti-tutions of Capitalism. Free Press, New York.

Zajac, E. J. and M. H. Bazerman (1991). 'Blind spotsin industry and competitor analysis: Implicationsof interfirm (mis)perceptions for strategic decisions'.Academy of Management Review, 16, pp. 37-56.

Zajonc, R. B. (1980). 'Feeling and thinking: Prefer-ences need no inferences', American Psychologist,35, pp. 151-175.


APPENDIX

In this Appendix we show how the strategicomission of information by upper level manage-ment discussed in the section on belief formationcan occur when employees are rational but notfully informed. We show that there exists aperfect Bayesian Nash equilibrium in which thefirm does not reveal its true state, and theemployees understand that such a possibilityexists. Intuitively, this happens because the firmdoes not want to reveal when times are 'almostbad' (because employee motivation suffers) andwhile employees know this they cannot forcemanagement to reveal all of their information.

Consider a firm which operates in one ofthree states—low, medium, and high denotedSL, SM, and SH respectively, for a three periodhorizon. Let Ns G {NL,NM,NH} be the optimalnumber of identical workers to employ in eachstate. The firm's state follows a three-stateMarkov chain P^ with positive transition prob-abilities Pss, where s,s' E {SL,SM,SH} and P^is common knowledge. Furthermore, assumethat the firm must pass through the mediumstate, that is, PHL = PLH = 0- For the purposesof this illustration assume that it is commonknowledge that the firm is in the high state inperiod one, but at the beginning of period twothe firm has private information about its truestate. In each period each of the identical, riskneutral workers selects one of two possibleeffort levels: a, = high effort, where the workerincurs effort cost c > 0, and aa = low effort,where he or she incurs zero effort cost. Weassume that while the workers' effort levels areobservable, they are not verifiable, and hence,noncontractible. This requires that the employ-ment relationship be based on self-enforcingimplicit contracts.* Let W' and R equal aworker's wage given effort level &i and a2in the previous period, respectively, whereW' > R. And for simplicity assume that thewage a worker receives if he leaves the firm(their reservation wage) is equal to R. Finally,assume that W' - C > R.

A worker's actual wage depends on whetherthe firm reduces its employment level given achange in state. Since workers are identicalassume that the probability of not being laid offfrom a current employment level of Ns to areduced employment level Nj is:

Since the firm is in state H in period one, aworker's expected utility in period two given ahigh effort level in period one is

EW(a||H) - c = PHHW' + pHM[qHMW'

+ (1 - qHM)* ] - c.

His expected utility from a low effort level issimply R, since his low effort level wage and hisreservation wage are assumed to be equal.^Hence, a worker will select aj in the high stateif:

PHHW' + pHM[qHMW' + (1 - qHM)R] - O R .

(1)

Equation (1) reduces to:

(pHH + pHMqHM)(W' - R) > C, (2)

That is, the expected benefit to a worker fromtaking the high effort action is greater than thecost. Assume relation (2) holds. Hence, allworkers find it optimal to select the high effortlevel in the first period.

Each worker has beliefs \i. = {(i,L,(JiM>M-H} aboutthe firm's true state in period two, where (J-H -t-fjLM + (JLL = L Prior to the firm's employmentdecision in period two, the beliefs of eachworker equal the transition probabilities, |Xo ={PHL.PHM.PHH}- Each worker uses Bayes' rule toupdate their beliefs after the firm makes its employ-ment decision, and before they select their effortlevels. Finally, in period three assume there is noeffort aversion problem and workers select aj.

'' See MacLeod and Malcomson (1989) for an elaboration ofthis approach.' A key aspect of this analysis is that a worker's currentwage depends entirely on his or her effort level in the

previous period. For simplicity we ignore discounting, whichhas no effect on the results." Note that by equating the workers" low effort wage withtheir outside wage the individual rationality constraints andthe incentive compatibility constraints are equal.


Our goal is to show the existence of a perfectBayesian pooling equilibrium in which the firmselects the high employment level in the secondperiod in both the high state and the mediumstate.^ In such an equilibrium the workers' beliefsafter the firm announces NH equal their priorbeliefs, that is, |i.(NH) = M-Q. Such an equihbriumrequires that the firm's profits from over-employment in the medium state with high effortlevels are greater than its profits from employingthe 'optimal' number of employees under loweffort levels. That is, assume that'TrM(NHlai) > TTM(NM|a2).''' This requires that theworkers find it optimal to select the low effortlevel if the medium state obtains, and in turnrequires that the following relation holds.

EW(ai|M) - c = pMH^i + pMM^i

+ pML(qML^' + (1 - qML)'*] - c < R. (3)

=>(pMH + pMM -I- pMLqML)(W' - R) < C. (4)

So the expected benefit to a worker from takingthe high effort action is less than the cost.Comparing equations (2) to (4) we see that if qMLis significantly smaller than qHM then it is feasiblefor both of these inequalities to hold. The intuitionis that workers face a much higher chance of beinglaid-off in the low state relative to the mediumstate, and hence, they do not have sufficientincentive to select the high effort action.

Let EW(H) denote a worker's expected wagefrom selecting a, in the high state (left side ofequation [1]), and let EW(M) denote a worker'sexpected wage from selecting ai in the medium

state (left side of equation [3]). A poolingequilibrium requires that the workers find itoptimal to select the ai under beliefs(jLo = {pHL,pHM,pHH}. Since pHL = 0, this con-dition requires equation (5) below to hold.

< (5)

That is, if the worker's prior beliefs, given bythe transition probabilities, place sufficient weighton the high state then it is optimal for them toselect the high effort action, despite the possibilitythat the firm is actually in the medium state.

Hence, given the above set of assumptions, aperfect Bayesian equilibrium exists where it isoptimal for the firm not to reveal its true state.Workers understand that the firm has an incentivenot to reveal such information, but since thefirm selects the same action in either state, it issufficiently likely that they are in the high stateto justify selecting the high effort level.

This brief model ignores several importantissues. For example, what if the firm simplyraises the wage in the medium state to make theincentive compatibility constraint binding? Thisoption is clearly possible in the frameworkdeveloped above, and this and similar issues areaddressed in Knez (1993). But note that raisingwages in the medium state (the danger state)implies the firm will want to lower wages if itreturns to the high state. Empirical evidence,along with the arguments developed in thispaper, suggest that this is very difficult for firmsto do.

' A perfect Bayesian equilibrium is a set of strategies andbeliefs such that strategies are optimal given beliefs and thebeliefs are obtained from equilibrium strategies and observedactions using Bayes' rule.'" Note that we are assuming that the firm's individualrationality and incentive compatibility constraints hold. Theimportant aspect left out here is a punishment strategy thatprevents the firm from reneging on its promise to pay ahigher wage under high effort in the previous period. Thiscan be dealt with by simply assuming that workers quit ifthe firm reneges and the firm incurs employee replacementcosts. See Knez (1993) for the details.

THE STRATEGIC MANAGEMENT OF THE ENTITLEMENT...

Documents

Transcript of THE STRATEGIC MANAGEMENT OF THE ENTITLEMENT...