A GOLDEN RULE OVER TIME: RECIPROCITY IN … Classes/OA YA/articles/treating the...A GOLDEN RULE OVER TIME: RECIPROCITY IN INTERGENERATIONAL ALLOCATION DECISIONS KIMBERLY A. WADE-BENZONI

« Academy of Management Journal2002. Vol. 45. No. 5. 1011-1028.

A GOLDEN RULE OVER TIME: RECIPROCITY ININTERGENERATIONAL ALLOCATION DECISIONS

KIMBERLY A. WADE-BENZONINew York University

In this article, I offer theoretical and empirical groundwork for the study of intergen-erational issues and highlight the relevance of intergenerational hehavior to organi-zations. I focus on situations involving conflict hetween generations in which a presentgeneration must give up henefits or take on hurdens to act on the hehalf of futuregenerations. I introduce the notion of intergenerational reciprocity and, in four studies,found that the behavior of previous generations influences how a present generationtreats future generations.

As the son of a college professor and a professormyself, I hoped my sons would go to college. Butafter our oldest found that the courses he wanted totake at the university were full, he got discouraged.I went to the registrar, I'd heard he was a stickler forrules, so I was floored when the man asked, "Whatclasses would he like?" and took care ofthe problemon his computer.

Seeing my surprise, he explained that years before,he was in the same situation as my son. He'd met aprofessor who took him to the enrollment and gothim into every class he wanted. "That man was yourfather. By any chance, is he still alive?"

I nodded. He smiled, "Good," he said. "Tell himwe're even,"

Morgan, 2000

The above anecdote is a true story that capturesthe essence of reciprocity across generations. Whenthe registrar was a student, the author's fatherhelped him to get enrolled in the courses of hischoice. The registrar was not in a position to recip-rocate the act of kindness to the professor whohelped him. However, years later he was in a posi-tion to "repay the deht" by helping the grandson

I am grateful to Max Bazerman, Francisco Benzoni,Jeanne Brett, Doug Medin, Vicki Medvec, Dave Messick,and Ann Tenbrunsel for their insightful comments, gen-erosity with their time, and encouragement. I also thankMartin Evans, Elizabeth Mannix, and two anonymousreviewers for their helpful suggestions on earlier versionsof this work; Francisco Park for his help with the datacollection; and Violetta Averbakh for her help with thedata management. Finally, I gratefully acknowledge thegenerous funding for this research provided by the Dis-pute Resolution Research Center of Northwestern Uni-versity, the Kellogg Environmental Research Center ofNorthwestern University, National Science FoundationGrant #SBR-9511977, State Farm Companies Founda-tion, and the U.S. Environmental Protection Agency,

(who was now in a similar position) of the manwho had helped him.

The focus of this article is the role that reciproc-ity plays in the relationship among generations.The theoretical literatures on intergenerational is-sues can be grouped primarily into two generalapproaches: philosophical and economic. In thephilosophy and law literatures, discussions of whatshould be done on the behalf of future generationsare based on ethics, existing laws, moral reasoning,and societal norms (e.g., Barry, 1989; Richards,1982; Weiss, 1989), In the economics literature,arguments are made about what is the appropriatediscount rate^ in contexts involving trade-offs be-tween generations (e.g,, Kotlikoff, 1992; Portney &Weyant, 1999). In contrast to these normative ap-proaches that focus on making arguments aboutwhat a present generation should do on the behalfof future generations, my research examines factorsthat affect the extent to which the present genera-tion will act on the behalf of future generations,while highlighting the relevance of intergenera-tional behavior to organizations.

THEORETICAL BACKGROUND

Intergenerational Issues in Organizationsand Society

Some of the most important issues in societytoday involve trade-offs among the interests of dif-ferent generations. Societal intergenerational issuesare relevant to organizations when behaviors and

^ A discount rate refers to the extent to which futurebenefits are valued today (Brennan, 1995); the greater thediscount rate, the less future benefits count compared topresent costs.

1011

1012 Academy of Management Journal October

decisions made by organizational actors affect out-comes to society and when these issues affect or-ganizations' ability to perform well and surviveover time. One of the most prominent examples ofthis relationship between societal intergenerationalissues and organizational interests is global envi-ronmental change.

Organizations are the intermediaries that convertnatural resources into usable products and, thus,corporate activities are fundamentally linked toproblems of environmental sustainability acrossgenerations (Shrivastava, 1995). Business is argu-ably the most powerful institution on the planet(Harman, cited in Hawken [1993]), and as such maybe the only institution in the modern world pow-erful enough to foster the changes necessary forecological and social sustainability for future gen-erations (Hawken, 1993). Social responsibilityaside, since organizations act in an economic sys-tem that is inextricably intertwined with and de-pendent on ecological systems (Jennings & Zand-bergen, 1995), their activities are circumscribed bythe carrying capacity of the natural environment(Carroll & Hannan, 1995; Jennings & Zandbergen,1995). Indeed, organizational theorists have high-lighted that long-term ecological sustainability gen-erates important contingencies for organizationalsurvival over time and, thus, it is a desirable andnecessary consideration for organizations (Starik &Rands, 1995).

Although intergenerational terminology has tra-ditionally been applied to broad social issues, suchas the one described above, or to family issues,such as the transfer of wealth from parents to chil-dren, thinking intergenerationally is also a usefulway to conceptualize a set of issues within andamong organizations that share characteristics withthese more traditional applications. Past, present,and future sets of organizational actors can bethought of as different "generations" in organiza-tions. Managers often make decisions involvinglong-term consequences for their organizations.Since organizational membership changes overtime, a challenge business decision makers face isthat what may be in the best interest of the presentset of organizational actors is not necessarily whatis best for future organizational actors.

In this research, I developed a theory of intergen-erational behavior that is applicable to both (1)societal intergenerational issues relevant to organi-zations and (2) organizational contexts in whichactors are usefully conceptualized as generations.My empirical investigation included both of thesecategories.

Intergenerational Conflict

Intergenerational situations are problematicwhen there is a difference in preferences (or likelypreferences) between generations about how deci-sions should be made or resources should be allo-cated. Intergenerational conflict arises when onegeneration of actors faces a decision of whether ornot to incur costs themselves for the benefit ofanother generation. Situations involving intergen-erational conflict are typically characterized by acombination of features, regardless of the nature ofthe conflict, context, or level of analysis. One of themost prominent characteristics is that there is fre-quently power asymmetry between the actors, withthe present or older generation having all or most ofthe control over how issues will be handled orresources will be allocated, while the future, oryounger, generation has little or no voice in theseprocesses. This feature goes hand-in-hand with thefact that future or later generations often do nothave the opportunity to reciprocate the behavior ordeeds of previous generations. In addition, the con-sequences associated with a decision or action areoften decoupled, so that benefits are accrued im-mediately by the present generation, while burdensfall to future generations at a later point in time.Acting on the behalf of future generations may re-quire a reversal of that pattern in such a way thatthe present generation may have to endure or takeon a burden so that future generations can benefit.This trade-off among the interests of different gen-erations is at the heart of intergenerational conflict.

In my research, I define generation as follows: Ageneration occupies a role that may be an office,status, or set of responsibilities, and the time periodin that role is limited in the sense that one genera-tion does not occupy that role indefinitely. At somepoint, generational transition occurs, whereby onegeneration succeeds another and comes to occupythe roles formerly occupied by the generation thatit replaced. The amount of time that one generationoccupies a role can be either fixed (a class of MBAstudents is an example) or flexible (for instance,directors of departments in organizations).

Generations can be found in organizations andsociety at both the individual and group levels. Atthe individual level, a generation can be under-stood as a single individual while he or she occu-pies a particular role (CEO or dean of a college, forexample). The previous generation would be theperson who had occupied the role previously, andthe succeeding generation would be the person re-placing the present generation. Thus, the bound-aries of a single generation are demarcated by anindividual's tenure in that role. Note that not any-

2002 Wade-Benzoni 1013

one in any role would be usefully considered ageneration. The conditions of potential intergenera-tional conflict must first apply; these are powerasymmetry, decoupling of consequences, lack ofdirect reciprocity, and role transition between ac-tors. In organizations, a common intergenerationalcontext at the individual level is the situation inwhich there is frequent rotation through positions(internships, for instance). Intergenerational con-flict comes into play when an outgoing employee—the preceding generation—makes decisions involv-ing trade-offs between what is best for him- orherself and what is best for the incoming person, orsucceeding generation; one context would be theuse of discretionary funds associated with a position.

At the group level, generations are either "cohort-based" or "event-based" and can refer to smallerwork groups as well as to organization-, industry-and society-level groups. At the group level, a com-mon orientation, socialization process, and/or"psychological contract" with an organization cancause a group to be identified as a single genera-tion.^ In cohort-based generations, a group of indi-viduals demarcated as a generation has a commontemporal starting point in a common role, whichcauses them to hold a certain status contemporane-ously. For example, classes of Ph.D. or MBA stu-dents or groups of trainees in a management train-ing program may be considered cohort-basedgenerations.^

Event-based generations are created when theoccurrence of a significant event leads to the cre-ation of different generations in organizations, orcauses differentiation among generations to be-come meaningful, important, or apparent. A singlegeneration shares a common set of experiences thatshapes them in such a way that they share a set ofparadigms. For example, the appointment of a newCEO and/or a major revision of an organization'sstrategy can lead to downsizing the organizationand layoffs of many workers, followed by the hiringof newer workers at a later point in time. In thesesituations, different generations may be differenti-ated by the nature of employees' psychological con-tracts with the organization. The older generationmay have understood themselves to have a psycho-

^ Note that one person can be the decision maker act-ing or deciding on the behalf of a generational group;such is the case with organizational leaders such as CEOsand public policy makers.

^ Members of such cohorts may eventually be pro-moted at different rates, causing them to have differentlevels of status, at which point it may no longer be usefulto view them as a single generation in certain contexts.

logical contract according to which they were loyalto a specific organization and expected to buildtheir careers within it; in contrast, newer genera-tions understand themselves to have a psychologi-cal contract whereby they are loyal to their owncareers, do not expect job security, and expect tochange organizations frequently to gain promotionsand salary increases.

Intergenerational Discounting

Intergenerational decisions are further compli-cated when the consequences to future generations,whether they are positive or negative, increase overtime. For example, long-term investments are ex-pected to result in greater benefits for future mem-bers of organizations or society than were thoseforegone by present members by not using the cap-ital for present gain. Similarly, future generationscan experience more serious negative conse-quences as a result of the present generation leav-ing burdens for them than would be experienced bythe present generation had they handled the bur-dens themselves. For example, several decades ago,the W.R. Grace Company disposed of toxic wastesby dumping them on the ground. This course ofaction turned out to be an easy and inexpensivesolution only in the short term. After the chemicalscontaminated the well water in Woburn, Massa-chusetts, and the contaminated water was linked tonumerous cases of leukemia in children, the com-pany incurred greater costs from reputation damageand lawsuits than it would have had it disposed ofthe chemicals properly in the first place (Plater,Abrams, & Coldfarb, 1992). Not only were the coststo the organization greater in the long run, but sincethe costs emerged many years after the decision todump the chemicals on the ground, an essentiallynew set of organization members were required tocope with those consequences.

This article specifically examines intergenera-tional conflict situations in which consequences tofuture generations increase over time, such as thosedescribed above. That is, every unit of benefit con-sumed by the present generation costs future gen-erations more than one unit, and every unit ofburden left by the present generation translates intomore than one unit of burden for future genera-tions. In these cases, power asymmetry among gen-erations is exacerbated, since the parties who havecontrol over the decision process (the present gen-eration) are not the parties with the most at stake(future generations) and thus, the dependency offuture generations on the present generation is in-tensified. Although not every case of intergenera-


tional conflict involves increasing consequencesover time, I cbose tbis type of intergenerationalsituation as the focus of my investigation, not onlybecause it is a common feature of intergenerationalcontexts, but also because it is more problematicthan more straigbtforward allocation situations,given tbe escalation in the conflict between gener-ations when consequences increase over time.

In cases of intergenerational conflict, the depen-dent variable of interest is typically the extent towhicb the present generation acts on tbe bebalf offuture generations. When consequences escalateover time, self-interested intergenerational behav-ior can be "operationalized" in terms of intergen-erational discounting, wbich occurs when individ-uals prefer smaller, immediate benefits fortbemselves over larger benefits for otbers in tbefuture. Tbus, intergenerational discounting basbotb intertemporal and interpersonal dimensions.If individuals are choosing to benefit tbemselves attbe expense of greater benefit to future others, thenthey are discounting the value of that benefit tofuture generations. A central reason why individu-als tend to discount benefits to future generations istbat, although many of tbe consequences of theirintergenerational decisions are immediate to them,tbe associated consequences to future otbers arenot. Tbe long-term implications of intergenera-tional decisions are temporally and personally re-moved from decision makers, and thus tbey are lesslikely to understand, be concerned with, or even befully aware of future consequences. Tbis frame-work belps to explain wby individuals may noteven tbink in intergenerational terms when makingdecisions that affect future generations.

Intergenerational Reciprocity

Although philosophers and theorists have citedthe lack of immediacy of future consequences asone reason why people often do not act on tbebebalf of future generations, tbey bave pointedmore strongly to tbe absence of traditional bonds ofreciprocity as a factor tbwarting sacrifice for futuregenerations (Care, 1982). One generation itself typ-ically does not benefit from tbe sacrifices it makesfor future generations. Tbis feature of intergenera-tional decisions raises the following question: Whywould, sbould, or does one generation ever act onthe bebalf of future generations?

Pbilosophers and etbicists wbo advocate inter-generational justice try to answer this question byfocusing on tbe notion of "moral reciprocity," inwhicb the present generation treats future genera-tions as they would like to have been treated them-selves by the preceding generation (Rawls, 1971;

Richards, 1981). The presumption is that everyonewould prefer to be treated fairly by others, and sopeople should treat others fairly as well. Note, how-ever, that moral reciprocity, or how one would liketo be treated by others, may or may not correspondto how one is or was actually treated by others.When making decisions affecting outcomes to fu-ture generations, individuals may consider how theywould like to have been treated by previous genera-tions, but they may also be influenced by how theywere actually treated by previous generations.

Reciprocity typically refers to the mutual rein-forcement by two parties of each other's actions(Blau, 1964; Gouldner, 1959, 1960; Parsons, 1951).This most traditional form of reciprocity has beenlabeled "mutual reciprocity" or "restricted ex-change" (Ekeh, 1974) and is characterized by a"quid pro quo" mentality. The principle of reci-procity, however, has also been used more broadlyto refer to situations in which people feel obligatedto reciprocate others' actions, not by directly re-warding their benefactors, but by benefiting otheractors implicated in a social exchange situationthat includes their benefactors and themselves(Levi-Strauss, 1949). This type of reciprocity hasbeen labeled "univocal reciprocity," "generalizedexchange" (Ekeh, 1974; Levi-Strauss, 1949), and"circular exchange" (Malinowski, 1939). General-ized exchange involves three or more actors whoare linked in an integrated transaction in whichreciprocations are indirect, not mutual (Ekeh,1974). Generalized exchange is characterized bythe lack of one-to-one correspondence betweenwhat two parties directly give to and take from oneanother; this is the primary quality distinguishinggeneralized from restricted exchange.

In this article, I introduce and examine a relatedphenomenon called intergenerational reciprocity.in situations in which people cannot reciprocatethe good or evil left to them by previous genera-tions, I propose that they instead "reciprocate" bybehaving similarly to the next generation. In otherwords, people can pass on benefits (or burdens) tofuture generations as a matter of retrospective obli-gation (or retaliation) for the good (or bad) receivedfrom past generations. This notion is consistentwith the theoretical argument made by Becker(1986) that some obligations to future generationshave their source in the good received from pastgenerations.

Indirect reciprocation is a feature shared by gen-eralized exchange among contemporaries and in-tergenerational reciprocity. There are, however, im-portant differences between the two exchangesituations. In generalized exchange among contem-poraries, the actors typically exist in the social ex-


change structure simultaneously. Thus, at the timeof contribution individuals can still potentiallybenefit from the system at some point in the future.In contrast, in the intergenerational context, someactors (past generations] are removed from the so-cial exchange situation over time and therefore nolonger have the opportunity to benefit from theexchange structure, while new actors (later gener-ations) are added. Another important difference isthat, because of the temporal aspect of intergenera-tional relations, actors generally participate in theexchange in a clear sequence. Intergenerational ac-tors are at the receiving end first, with the oppor-tunity to contribute coming at a later point in time.With generalized exchange, the actors' opportuni-ties to contribute and benefit are often random anddetermined by fate.

These differences affect the respective incentivestructures of the exchange situations. The opportu-nity to benefit from the system is a primary motivefor contribution in generalized exchange. This mo-tive, however, is absent in the intergenerationalscenario at the time the actor decides how much tocontribute to future generations. Instead, actorswho benefit from actions of previous generationsmay be motivated by a sense of obligation to "repaythe debt" (as in the case of the opening anecdote] byproviding future generations with similar benefits.(Or, conversely, they may not feel obligated to do soif the previous generation did not help them.] Thissense of obligation can also be present in general-ized exchange among contemporaries if an actorhas already benefited from the system without yetcontributing. Theorists examining generalized ex-change among contemporaries, however, have notpointed to this factor as the primary motive forcontribution; instead, they have focused on the de-sire to contribute to the exchange system in order tokeep it functioning and to thus preserve the oppor-tunity for potential future self-benefit (Ekeh, 1974].

The lack of opportunity for future benefit in theintergenerational case causes a potential free riderproblem greater than the one that can arise withgeneralized exchange among contemporaries. Yam-agishi and Cook (1993] pointed out that since therewards that an actor receives usually are not di-rectly contingent on the resources provided by thatactor in generalized exchange, free riding can occur(that is, an actor can receive benefit without con-tributing]. Theorists have therefore highlighted theimportant role of trust in generalized exchange sys-tems. "Depersonalized" or "group-based" trust op-erates when individuals expect the probability ofreciprocity within a particular group to be high(Brewer, 1981]. It is not contingent on the beliefthat the specific recipient of a beneficent act will

return the favor, but rather, is based on confidencethat group members will reciprocate helpful behav-iors, even if the specific recipient of past favors ismissing or unable to fulfill his or her obligations(Kramer, 1993]. In the intergenerational case, how-ever, where actors benefit first and later have thechance to contribute with little or no probability ofbenefiting again, trust in the future reciprocation ofothers cannot be a motive for contribution.

The empirical investigation of intergenerationalreciprocity that follows includes four studies thatexamined the allocation of resources across gener-ations in the presence of intergenerational conflict(that is, a trade-off between the interests of differentgenerations]. I propose that individuals' percep-tions of what the previous generation left for themwill affect what they leave for the next generationin such a way that the more the previous generationacted on the behalf of the present generation, themore the present generation will act on the behalfof future generations. In all four studies, resourcesincreased over time, so that self-interested behavioron the part of the present generation indicatedthe discounting of benefits to future generations. Thestudies were designed to capture and reflect thecombination of features (discussed previously] thatcharacterizes intergenerational situations, includingpower asymmetry, decoupling of consequences, lackof direct reciprocity, and role transition among actors.

STUDY 1

In Study 1, I took the first step in the empiricalexploration of intergenerational reciprocity by ex-amining a context in which limited benefits areallocated across generations. In this initial study, Ipredicted that the greater the amount of desirableresources a present generation perceives a previousgeneration as having left, the greater the amountthat the present generation will leave for futuregenerations.

The intergenerational context for Study 1 wasbased on the real-life crisis in fisheries around theworld, where stocks of many species of fish arenear collapse and there is much concern as to howa sustainable level of harvesting can be maintained.The collapse of the fisheries is a striking example ofan intergenerational situation that involves bothsociety-level concerns and generations of organiza-tional actors (illustrating that these are not mutu-ally exclusive]. From the perspective of a singlegeneration, it appears that it is in the economic bestinterest of fishing companies to catch as much fishas they can as fast as possible so that they canmaximize their immediate economic gains. Suchbehavior is detrimental to the long-term economic


interests of the industry because by following thiscourse of action over time, the fishing industry putsitself out of business by mismanaging its assets. Inother words, it eventually destroys the resource onwhich it depends for its sustenance and profits.The intergenerational conflict here is that by har-vesting at an unsustainable level, the earlier gener-ations experience benefits that pale in comparisonto the costs that later generations must bear (bank-ruptcy) as a result of the overfishing.

Methods

Participants. The Study 1 data were gatheredthrough a simulation involving a business decisionbased on the fisheries crisis described above (areal-life intergenerational issue at an industrylevel). Fifty-seven graduate students in an MBAprogram at a university participated in this study.This sample was used because the procedures in-volved making a major business decision. The in-dividuals in this sample were likely to have madebusiness decisions in the past and were likely to bemaking them in their future careers. Participantswere entered in a lottery to win $200 for theirparticipation in the study.

Procedures. Participants were put in the role ofan owner of a large commercial fishing business.They were told that this was the last year that theywould be in the business because they were retiringfor personal reasons. They had been informed bythe National Marine Fisheries Service (NMFS)about a problem of overexploitation of the fish re-source. The NMFS had asked them to reduce theirharvest by 50 percent in order to enable the speciesthey fished to sustain its existence into the futurefor the benefit of future generations of fishers. Theywere told that since they were the largest commer-cial fishing company in the industry, their harvestwould have the greatest impact on the resource.(The harvesting of the other, smaller companieswould be negligible, and so the behavior of otherfishing companies was not a relevant factor in theirdecisions. This instruction eliminated any poten-tial effect of an intragenerational social dilemma onthe decisions of participants.) At maximum capac-ity, they could harvest 1,000 metric tons. Thus, theNMFS was asking them to harvest only 500 metrictons this year. The request was not legally enforce-able; participants were asked to limit their harvestvoluntarily. Participants were also told that theirpersonal savings for retirement were modest.

The situation was such that they would not ben-efit from the sacrifice, but the cost of the collapsedresource to future fishers was much greater than thecost to the present fishers (the participants) of re-

ducing their harvest. Participants were shown witha graph that there was a positive correlation be-tween their profits and their harvest level. Partici-pants were also told that marine scientists esti-mated that the recovery of the resource wouldenable future fishing companies comparable totheir own to harvest 600 metric tons per year indef-initely beginning in a few years, after the fisheryhad had a chance to recover. This estimate wasbased on the assumption that future generations, inturn, would harvest at a sustainable level. Partici-pants were also told that, in contrast, if the fisherywas harvested unsustainably again this year, therecovery would not take place and future fisherswould experience tremendous losses.

The cost/benefit trade-off between present andfuture generations described in the experimentalmaterials mirrored the real-world situation. In ad-dition, the National Marine Fisheries Service is anactual organization whose responsibilities may in-clude attempts to persuade fishers to reduce theirharvests voluntarily.

Manipulation of intergenerational reciprocity.Participants were randomly assigned to one of twoinformation conditions. This manipulation repre-sented the behavior of the previous generation thatwould affect the present generation. Participantswere told that the NMFS had made a similar re-quest to the largest commercial fishing company afew decades ago, when the state of the fishery firstbegan to be an issue. In one variation of the infor-mation, constituting the first condition, the com-pany in the previous generation was said to havechosen not to respond to the NMFS's request; it hadcontinued to harvest at full capacity. Participantswere told that it is likely that this decision had hadan impact on the current condition of the fishery,which would be healthier today if the company hadchosen instead to reduce its harvest. However, thespecific levels of harvesting of the previous gener-ation were not specified. It was the general re-sponse of nonbeneficence toward the future gener-ation that was important in this manipulation.There were 27 participants in condition 1.

In the second information condition, the com-pany in the previous generation was said to havechosen to respond to the NMFS's request and hadcut its harvest in half Participants were told that itwas likely that this decision had had an impact onthe current condition of the fishery and that itwould be in worse condition today if the companyhad chosen instead not to reduce its harvest. Onceagain, the specific levels of harvesting (before andafter the reduction) of the previous generation werenot specified. It was the general response of benef-icence toward the future generation that was im-


portant in this manipulation. There were 30 partic-ipants in condition 2.

Measure of intergenerational behavior. Inter-generational behavior was measured implicitlyfrom participants' decisions. Participants decidedhow they would allocate resources between them-selves and future generations, knowing that futuregenerations would be affected by the allocationmore than they themselves would be affected. Par-ticipants were asked to indicate the amount of fishthey would harvest. This could be any amountbetween 1,000 and 500 metric tons, which repre-sented operating their business at 100 percent and50 percent capacity, respectively. This harvestingrange reflected real-world considerations by (1) notallowing harvests to exceed current levels, since1,000 metric tons was the company's maximumcapacity and (2) not allowing harvesting levels todrop to unrealistically low levels, since it is un-likely that companies would intentionally putthemselves out of business. In addition, the NMFSdid not say that it would help the fishery to gobelow the 500-metric-ton recommended level. Thecloser to 100 percent capacity (1,000 metric tons)they chose to operate, the less they were acting onthe behalf of future generations and the more theywere discounting the benefits to future generations.

Postdecision questionnaire. After making theirharvesting decisions, participants were asked toexplain in writing how they made their decisions.This question was intended for exploratory pur-poses and was intentionally very broad.

Manipulation check. In the same postdecisionquestionnaire, after they had answered the open-ended question above, participants were asked toindicate, using a seven-point scale (1, "not at all";7, "very much so"), the extent to which they agreedwith the following statement: "Previous genera-tions of fishers did their part to help preserve theresource."

Results

The manipulation check measure had a mean of1.70 (s.d. = 0.82) in condition 1 and 4.97 (s.d. =1.69) in condition 2 (p < .001), confirming theeffectiveness of the manipulation. The results sup-ported the hypothesis that the greater the amount ofdesirable resources the present generation per-ceives as having been left by the previous genera-tion, the greater the amount that the present gener-ation will leave for future generations. The overallmean harvest was 604 metric tons (s.d. = 160). Incondition 1 (previous generation did not reducetheir harvest], the mean harvest was 670 (s.d. =168) metric tons. In condition 2 (previous genera-

tion did reduce their harvest), the mean harvestwas 546 (s.d. = 131) metric tons. The difference inharvesting level by condition was significant [F =9.91, p < .01) in the predicted direction.

The results from the open-ended question offeredfurther insights into the decision process. First, allresponses were coded as to whether participantsexpressed a concern for future generations or not.Examples of statements indicating concern for fu-ture generations included "It is important to make asacrifice on the behalf of the industry and futurefishing companies" and "I want the next generationof fishermen to have the same opportunity I had."Their written responses were consistent with theirharvesting decisions; the majority of respondentsin condition 2 indicated a concern for future gen-erations (24 out of 31, or 77 percent), while only 9out of 27, or 33 percent, did so in condition 1. Thisdifference was significant (x^ = 11.80, p < .01).

Another interesting trend was that, although anumber of respondents (28%) explicitly cited thebehavior of the previous generation as a factor thatinfluenced their decision in condition 2 (previousgeneration reduced harvest), none of the respon-dents referred to or mentioned the behavior of theprevious generation in condition 1 (previous gen-eration did not reduce harvest). The responses werecoded as to whether or not they referred to thebehavior of the previous generation as a factor inthe decision, and the difference between condi-tions was significant [y^ = 8.08, p < .01). An ex-ample of a reference to the previous generationwas, "My own profits have resulted because pastfishers considered and implemented reduced har-vest. I should do the same for others."

A potential concern in this study is that the sig-nificant difference by experimental condition inharvesting decisions was driven by demand effects.However, if that were the case, then one wouldexpect respondents in condition 1 to have tendedto maximize their harvests. However, only 3 did so(in the second condition, 2 did so, for a total of 5),and the majority of people, even in condition 1,reduced their harvests by 50 percent or less (52 outof 58 overall, or 90 percent; 23 out of 37, or 62percent in condition 1; and 29 out of 31, or 94percent, in condition 2). The results suggest thatthe majority of participants in both conditionsacted to some extent on the behalf of future gener-ations, but those in condition 2 did so to a greaterextent (that is, they were more generous). In condi-tion 2, 26 out of 31 participants (84%) harvestedthe minimum (500 metric tons), while in condition1, only 10 out of 27 (37%) did so. In addition, thequalitative data suggest that participants were en-gaging in the exercise in a way consistent with my


thesis. Nothing in any of the responses to the open-ended question suggested that participants wereguessing hypotheses or responding in a way thatreflected their perceptions of my expectations.

STUDY 2

In Study 1, I looked at an intergenerational allo-cation decision involving henefits. Since pastresearch suggests that allocations of henefits andhurdens represent different decision processes(Griffith & Sell, 1988; Lamm & Kayser, 1978; Man-nix, Neale, & Northcraft, 1995; Northcraft, Neale,Tenbrunsel, & Thomas, 1996; Sondak, Neale, &Pinkley, 1995; Tornhlom, 1988), one cannot as-sume that the findings from Study 1 will generalizeto hurden allocation. To examine this, in Study 2 Ilooked at whether intergenerational reciprocity ex-tended to situations in which people were allocat-ing hurdens hetween themselves and future gener-ations. My prediction was that the greater thehurden perceived to have heen left hy previousgenerations, the greater will he the hurden thepresent generation will leave for future generations.

A potential weakness of Study 1 was that, al-though the experimental materials were hased on areal-world situation, participants were playingroles rather than making decisions relevant to theirreal lives. Study 2, designed to address this con-cern, was hased on the topic of federal gasolinetaxation in the United States, an issue directly re-lated to greenhouse gas emissions and glohal warm-ing (the greater the tax increase, the greater thedecrease in annual greenhouse gas emissions, adelayed atmospheric effect resulting in significantmeasure from the consumption of fossil fuels). Thiscontext was chosen because global climate changeis one of the most important intergenerational is-sues worldwide, and it transcends national, gener-ational, and organizational boundaries and hasboth direct and indirect effects on organizations.Greenhouse gas emissions ahatement policies willtranslate fundamentally into encouraging techno-logical innovation and controlling energy genera-tion and consumption, which has direct economicimplications for the majority of industrial sectors.

One challenge in studying group-level intergen-erational issues is the presence of a social dilemmawithin a generation (an intragenerational social di-lemma) affecting actors' ability to act on the behalfof future generations. A decision maker in a presentgeneration might want to act on the hehalf of futuregenerations, but the potential impact of his or heractions might depend on others within the samegeneration also making decisions on the behalf offuture generations. The dilemma is that if the indi-

vidual sacrifices and no one else does, then theindividual gets hurt and the future generation stilldoes not benefit. In this kind of situation, the co-operation of many actors in the present generationis needed in order to benefit future generations.When studying group-level intergenerational proh-lems, researchers must find ways to separate theinfluence of social dilemmas on behavior from theintergenerational phenomena of interest. This oh-stacle was, admittedly, dealt with somewhat artifi-cially in Study 1. A strength of Study 2 is that thispotential confound was avoided naturally.

Methods

Participants. Sixty-one graduate students in anMBA program at a university participated in thisstudy. Participants were entered in a lottery to win$200 for their participation in the study. They weretold that comprehensive research was heing con-ducted to gather information on the issue of gaso-line taxation. The advantage to this method is thatparticipants were not imagining themselves in asituation. Instead, gasoline taxation was an issuewith which the participants were familiar and bywhich they were likely to he affected in their reallives.

Procedures. Participants were given some infor-mation on the federal gasoline tax, including thecurrent tax at the time of the study. Several benefitsassociated with the taxation of gasoline were out-lined for them, including reducing pollution(greenhouse gases), conserving natural resources,and improving national security.

Then participants were told that although a por-tion of the henefits descrihed would be realizedimmediately, most of the benefits resulting from anincrease in the gasoline tax would he enjoyed byfuture generations. Participants received further in-formation about fossil fuel and pollution levels thatillustrated the distribution of burdens and benefitsbetween present and future generations. Partici-pants were told that the relationship hetween theincrease in the gasoline tax and the benefits tofuture generations was roughly linear, such that thegreater the gasoline tax increase, the greater thebenefits. Finally, participants were reminded ofthe primary disadvantage to an increase in the tax:the price of fossil fuel would rise.

Manipulation of intergenerational reciprocity.Participants were randomly assigned to one of twoconditions. In both conditions, they received somehistorical information about the tax. In the firstcondition, participants were told that the federalgas tax had only been raised three times in the pastfew decades and that between 1932 and 1983, the


gas tax only increased from 1 to 4 cents per gallon.Thus, the emphasis was on what the previous gen-eration had not done for the present generation.There were 35 participants in condition 1.

In the second condition, participants were toldthat the federal gas tax had been in effect since theearly part of the century, had been steadily increas-ing since its inception, and had risen more dramat-ically (more than 400 percent) in recent decades.Thus, the emphasis was on what the previous gen-eration had done for the present generation. Therewere 26 participants in condition 2. Note that inboth conditions all information provided to theparticipants was real data on the actual tax.

Measure of intergenerational behavior. Partic-ipants were asked to indicate how much of a fed-eral gasoline tax increase they thought would beappropriate, given the relationship between thebenefits to future generations and an increase in thetax. They were told that the increase would be inaddition to the current federal and state tax and, forthe purposes of this survey, they should assumethat the burden of the tax would be distributedequitably over all sectors of society.

Intergenerational discounting was measured im-plicitly from participants' decisions. The greaterthe amount (in cents] they thought was appropri-ate, the more they were willing to take on a burdenfor the benefit of future generations, and the lessintergenerational discounting was occurring.

Results

The results supported the hypothesis that thegreater the perceived burden left by previous gen-erations, the greater is the burden that will be left tofuture generations. The overall mean gas tax thatparticipants suggested was 25.32 cents (s.d. =29.07). In condition 1 (previous generation has notdone its part), the mean gas tax was 18.84 cents(s.d. = 23.95). In condition 2 (previous generationhas done its part), the mean gas tax was 34.04 cents(s.d. = 33.32). The difference in gas tax by condi-tion was significant (F = 4.30, p < .05) and in thepredicted direction."*

STUDY 3

The data from Studies 1 and 2 demonstrated thatintergenerational reciprocity was possible in the

'' The high standard deviations were due primarily to afew participants in both conditions who could be called"zero-tax extremists." These individuals would not con-sider any increase in the federal gasoline tax appropriateregardless of any information they were given.

allocation of benefits and burdens. Study 3 was thenext step toward understanding this phenomenonbetter by considering what drives this effect. Howand why does the behavior of the previous genera-tion affect the intergenerational decisions of thepresent generation? The qualitative data fromStudy 1 suggested a combination of reciprocal ob-ligation (an illustrative comment is, "Someone haddone the same for me in the past") and following anorm or model provided by past generations ("Fol-lows tradition of another company in the past").

The results from Studies 1 and 2 are consistentwith research conducted within the framework ofsocial learning theory (Bandura, 1969; Bandura &Walters, 1963), which demonstrates that learningcan occur vicariously, through observation of otherpeople's behavior rather than through direct expe-riences. A study by Cowan, Langer, Heavenrich,and Nathanson (1969) demonstrated that modelinginfluences produced generalized and enduringchanges in moral judgments. Specifically, the au-thors found that modeling emerges as a significantdeterminant of moral judgments, regardless of thedirection in which judgmental behavior is beingmodified. Rutte, Wilke, and Messick (1987) sug-gested that in situations involving the sequentialuse of a resource pool, individuals can infer a normof either selfishness or generosity from the behaviorof the prior users of the resources. Conformance toinferred norms that can be implied by the behaviorsof others is a simple decision heuristic. Rutte andcolleagues noted that although understanding theuse of heuristics in simple situations must precedetheir analysis in more complex ones, more compli-cated situations (such as intergenerational alloca-tions) will naturally entail different heuristics deal-ing, for instance, with reciprocity.

I posit that when people face an intergenerationaldecision, they look for information on what mightbe appropriate intergenerational behavior to helpthem make their decision. The behavior of pastgenerations is one source of information people areespecially likely to use if it is easily accessible. Ipropose, however, that alternative models (definedbelow) of intergenerational behavior can influencebehavior toward future generations in much thesame way the behavior of the previous generationwas shown to do in Studies 1 and 2.

For precision and clarity, here I take "previousgeneration" to mean the generation whose actionsdirectly affect the decision maker. I take "alterna-tive model" to mean an intergenerational decisionthat a decision maker can observe, but by which heor she is not affected. In this study, I predicted that,in the absence of information about the behavior ofthe previous generation, alternative models of in-

1020 Academy of Management fournal October

tergenerational behavior would influence the inter-generation allocation decisions of the presentgeneration.

A procedure involving the systematic replace-ment of individuals in a laboratory context to pro-duce an analogue of generational succession hasbeen viewed as an important methodological toolfor the study of social evolution (Gerard, Kluck-hohn, & Rapoport, 1956). Although I believe that itis important to be clear about the limitations of alaboratory study for explaining anything as com-plex as social evolution, the methodology de-scribed above has been considered legitimate inprior research on generational succession (Insko,Gilmore, Moeble, Lipsitz, Drenan, & Thibaut, 1980;Insko et al., 1982; Jacobs & Campbell, 1961; Rose &Felton, 1955; Weick & Gilfillan, 1971; Zucker,1977). Following this tradition, I used a similarmethodology in Studies 3 and 4. The studies citedabove generally examined the persistence of norms,traditions, and/or beliefs across generations; thisissue is relevant to my researcb, but to my knowl-edge none of these prior studies examined resourceallocation or contexts involving conflict of interestbetween generations.

Studies 1 and 2 required participants to makeintergenerational allocation decisions witbin twocontexts based on real-world intergenerational is-sues, fisheries and gas taxes, both of wbicb areamong the most important intergenerational issuesfor society and of bigb relevance to organizations.These issues provided ricb real-world contexts fortbe decisions; bowever, a limitation was that thedecisions were necessarily bypothetical and thusparticipants did not have to live witb tbe conse-quences of tbeir decisions. Studies 3 and 4 in-volved an allocation decision in which participantsexperienced actual consequences as a result of theirdecisions, tbus addressing tbis limitation. In addi-tion. Studies 3 and 4 examined tbe pbenomenon ofintergenerational reciprocity in a different popula-tion of participants.

Methods

Participants. Fifty-two individuals on tbe ad-ministrative staff (secretaries, area coordinators,area managers, and so fortb) at a university partic-ipated in this study. Tbey were each paid $5 toparticipate in the experiment.

Procedures. Participants were told that researchwas being done on bow people allocate resourcesbetween themselves and other people and that theywould be deciding how to distribute some moneybetween themselves and another person whowould be participating in tbe researcb after them.

Thus, participants were in the role of the presentgeneration and the "next person" represented thefuture generation.

Participants were each given an envelope con-taining $9 in single dollar bills and asked to taketbe amount that tbey wisbed to keep and leave theamount that they wished to give to the next personin the envelope. They were told that the amountthey left for the next person would be increased by50 percent. The money kept from the $9 was inaddition to the $5 that each was paid to participatein the research. Participants made their decisionsindividually and in private.

Manipulations. There were four experimentalconditions, with 13 participants in each one. Theinstructions in the first two conditions describedbow someone else who had participated in a simi-lar research project had bebaved; thus, they pro-vided alternative models of intergenerational be-bavior. One example was generous and the otherwas not. The instructions for the other two condi-tions indicated that the amount ($9) that the cur-rent participant was to allocate was the result of thedecision of the person who had preceded him orher in the experiment; the manipulation estab-lished the predecessor as generous in one case andnongenerous in the other (these conditions repre-sented the bebavior of tbe previous generation).

In tbe first condition, participants were given anongenerous example. Specifically, tbey were toldthat in a similar situation, someone had $24 andtook $18 (75%), leaving $6 (25%), which was in-creased by 50 percent ($3) and given to the nextperson. In the second condition, the example wasgenerous; they were told that someone else in asimilar situation had $8 and decided to keep $2(25%) and leave $6 (75%), which was increased by50 percent ($3) and given to the next person. Thus,it is the percentage kept or left of the originalamount that indicates the spirit of generosity orselfishness. In the third condition, participantswere told that the previous person had $24 anddecided to keep $18 (75%) and leave $6 (25%) forthe next person—who was the current participant.That $6 was increased by 50 percent ($3) to createtbeir $9. Tbus, the previous generation was rela-tively ungenerous. Finally, in the fourth condition,participants were told that the previous person had$8 and decided to keep $2 (25%) and leave $6(75%) for the next person (who was the currentparticipant). That $6 was increased by 50 percentand, once again, created tbeir $9. Thus, the previ-ous generation was relatively generous.

Measure, questionnaire, and manipulationcheck. Intergenerational behavior was measuredimplicitly by bow much of the resources (money)


participants kept for themselves. Since the money aparticipant left for the next person increased by 50percent, the more money the present generationkept for themselves, the more they were discountingthe value of the resources to the next generation.

After making their allocation decisions, partici-pants were asked to describe in writing how theyhad made their decisions.

After answering the open-ended question on howthey had decided, participants indicated the extentto which they agreed with a statement (1, "not atall"; 6, "very much so"). For people in conditions 1and 2, the statement was "The allocation decisionof the person who participated in a similar researchproject was generous." For those in conditions 3and 4, it was "The allocation decision ofthe personwho left me money was generous."

Results

Tahle 1 summarizes the means hy condition. Theresults support the hypothesis from Study 1 thatthe greater the amount of resources perceived ashaving heen left hy previous generations, thegreater the amount that will be left to ftiture gener-ations. A planned contrast analysis between condi-tions 3 (previous generation was not generous) and4 (previous generation was generous) indicated asignificant difference between the two [t = 2.97,p < .01), thus lending support, once again, for thepresence of intergenerational reciprocity in the al-location of benefits. The data also support the hy-pothesis that intergenerational behavior can be in-fluenced not only by the behavior of the priorgeneration, but also by alternative models of inter-generational behavior. Another planned contrastindicated a significant difference between condi-tions 1 (nongenerous example) and 2 (generous ex-ample) [t = 5.29, p < .001). Follow-up post hoc"pairwise" comparisons indicated (consistent withwhat is fairly obvious from the means] that condi-tions 1 and 3 were not statistically different fromeach other, nor were 2 and 4.

The actual allocations indicated that participants

TABLE 1Results, Study 3"

Experimental ConditionMean Amount

Kept

1. Alternative model, nongenerous 5.38 132. Alternative model, generous 2.92 133. Previous generation, nongenerous 5.31 134. Previous generation, generous 3.92 13

Overall = 1291, p < .001.

were responding to the spirit of generosity or non-generosity rather than mimicking the exact behav-ior of the previous generation or the alternativemodel of intergenerational behavior they received.That is, participants did not allocate the exact samedollar amounts or proportions of original resourcesto ftiture generations as the previous generationhad allocated to them. Rather, the responses re-flected a strong desire to share the resourcesequally, but responses tended to be either generousor not, depending on the behavior of the previousgeneration.

In their responses to the open-ended question, 21of the 52 (40%) referred to concern for equality.There were, however, different ways of thinkingabout equality that fell into three general catego-ries: (1) split what one has to begin with (that is,split $9 in half, although this meant deciding onone party getting an extra dollar, since they wereonly allowed to allocate in $1 increments), (2) splitthe ultimate amount, taking the increase into ac-count (that is, keep $6, which means the futureother also gets $6, after the increase), or (3) haveequal starting amounts (that is, keep $3, whichmeans the ftiture other gets $9—the same amountthe present generation started with—because thefuture other gets the $6 plus the 50 percent increaseof $3).

The manipulation check indicated that the ma-nipulations were effective. The mean response forcondition 1 (.x = 2.62) was significantly lower thanthat for condition 2 (x = 4.46, p < .01), and forcondition 3 (x = 2.69), it was significantly lowerthan for condition 4 (x = 5.31, p < .001), thusconfirming the effectiveness of the manipulations(overall F = 15.89, p < .001).

STUDY 4

The results ftom Study 3 indicate that an alter-native model of intergenerational behavior can in-fluence the behavior of a present generation in theabsence of information about the behavior of theprevious generation. This finding suggests that pro-viding a norm or model of appropriate intergenera-tional behavior is one reason why the behavior ofthe previous generation affects intergenerationaldecisions of the present generation. This is one ofthe mechanisms that I have suggested is likely tounderlie intergenerational reciprocity. The other,which is also evident in the qualitative data ftomStudy 1, is reciprocal obligation. It is likely thatthese mechanisms operate in conjunction to pro-mote intergenerational reciprocity. If these twopsychological processes (reciprocal obligation andmodeling) are intertwined in real-life intergenera-


tional decisions, it may not be possible, necessary,or even useful to isolate the role of each when thebehavior of prior generations is known because,despite any reciprocal obligation (or retaliation]that is created, the behavior of the prior generationcan always also serve as a model and thus, youcannot rule that out as a possible influence in thedecision process.

Another question raised by Study 3, however, iswhat will happen when a present generation issimultaneously exposed to an alternative model ofintergenerational behavior and information aboutthe behavior of the previous generation? In thiscase, the present generation may have two compet-ing models of how it should behave. If both modelssuggest a spirit of generosity, or both suggest aspirit of nongenerosity, they will not compete witheach other. Study 4 investigated what happens ifone model suggests generosity and the other modelsuggests nongenerosity. I predicted that the modeloffered by the previous generation would dominatein the presence of a competing alternative modelbecause the extra motive of reciprocal obligation(or retaliation) would be present. Thus, here I pre-dicted that in the presence of another, competingalternative model of intergenerational behavior, thebehavior of the previous generation will prevail asthe dominant influence on the behavior of the presentgeneration with respect to future generations.

Study 3 indicated that considerations of equalityand fairness came into play in participants' deci-sions. Prior research on conflict and resource allo-cation offers ample evidence that individuals makeegocentric or self-serving interpretations of situa-tions and judge as fair distributions of resourcesthat favor themselves (Babcock, Loewenstein, Issa-charoff, & Camerer, 1995; Bazerman & Neale, 1982;Diekmann, Samuels, Ross, & Bazerman, 1997; Mes-sick & Sentis, 1983; Neale & Bazerman, 1983;Thompson & Loewenstein, 1992; Wade-Benzoni,Tenbrunsel, & Bazerman, 1996; Walster, Walster, &Berscheid, 1978). In Wade-Benzoni et al. (1996),actors were in an asymmetric social dilemma inwhich a number of solutions could arguably bejustified as fair, depending on which informationparticipants chose to focus on. Results showed astrong tendency to (1) focus on solutions that favorself-interest and (2) justify self-interested behavioron the basis of arguments about fairness.

The design of Study 4 was based on the premisethat people need very little incentive to justify aself-interested decision or, in this case, one that istihed in a self-interested direction (since Study 3indicates that allocations also reflect the strong in-fluence of an equality norm). Because of the perva-sive tendency toward egocentric interpretations of

fairness, especially in resource allocation deci-sions, if there are two competing models—one sig-nifying generosity and one demonstrating self-interest—individuals might be expected to choosethe self-interested model, unless another force isalso at play. Thus, if they follow the behavior of theprevious generation in the presence of an alterna-tive model that allows them to justify a more self-interested choice, it would suggest that the influ-ence of the behavior of the previous generation hastrumped the alternative model of intergenerationalbehavior.

Methods

Participants. Thirty-seven individuals on the ad-ministrative staff (secretaries, area coordinators,area managers, and others) at a university were inthis study. Each was paid $5 to participate in theexperiment.

Procedures. As in Study 3, participants weretold that research was being done on how peopleallocate resources between themselves and otherpeople and that they would be deciding how toallocate some money between themselves and an-other person who would be participating in theresearch after them. They were each given a sum ofmoney ($9) and asked to decide how much of it tokeep for themselves and how much to leave for thenext person. They were told that the amount thatthey left for the next person would be increased by50 percent. Thus, again participants were in therole of the present generation and the next personrepresented the future generation.

Again, as in Study 3, participants received enve-lopes with the money (in dollar bills) and wereasked to take what they wished to keep for them-selves and leave what they wished to give to thenext person in the envelope. The money that theykept from the $9 was in addition to the $5 that theywere paid to participate in the research. Partici-pants made their decisions individually and inprivate.

Manipulations. There were four experimentalconditions. The first was a control condition andserved as a baseline in which participants were notgiven any additional information. In the secondcondition, they were given an example of very self-interested behavior on the part of someone in asimilar situation. Specifically, they were told theperson had $40 and took $38 (95%), leaving $2(5%), which was increased by 50 percent, therebyleaving $3 for the next person. This example pro-vided an alternative model of intergenerational be-havior and offered information justifying self-inter-ested tendencies in allocation decisions.


In the third and fourth conditions, participantsreceived all of the above information, including theself-interested example. In addition, they were toldthat the $9 they had to allocate was the result of thedecision of the person who had participated in theexperiment prior to themselves, and the $9 waswhat was left for them. In the third condition, par-ticipants were told that the previous person had$24, decided to keep $18 (75%), and left $6 (25%);that $6 was increased by the 50 percent to createthe $9 the participant then had. Thus, the previousgeneration was relatively ungenerous.

In the fourth condition, participants were toldthat the previous person had $8, decided to keep $2(25%), and left $6 (75%), which again was in-creased by 50 percent, thereby leaving $9 for thecurrent participant. Thus, the previous generationwas relatively generous. Note that conditions 3 and4 in Study 4 were identical to conditions 3 and 4 inStudy 3, except that in Study 4, the self-interestedexample was also provided.

Measure, questionnaire, and manipulationcheck. Intergenerational behavior was measuredimplicitly by how much of the resource (money)participants kept for themselves. Since the moneyleft for the next person increased by 50 percent, themore money the present generation kept for them-selves, the more they were discounting.

After making their allocation decisions, partici-pants were asked to describe how they made theirdecisions. They wrote down their responses on aform provided (with blank space for them to write).

In the same postdecision questionnaire, after an-swering the open-ended question above, partici-pants in conditions 3 and 4 were asked to indicatethe extent to which they agreed with the followingstatements using a six-point scale (1 = "not at all";6 = "very much so"): "The allocation decision ofthe person who left me money was appropriate,""The allocation decision of the person who left memoney was fair," "The allocation decision of theperson who left me money was generous."

Results

Table 2 summarizes the means by condition.Planned contrasts and follow-up post hoc pairwisecomparisons all indicated that the means in condi-tions 1, 2, and 3 were not significantly differentfrom each other, and the mean in condition 4 wassignificantly different from all other conditions.The behavior ofthe people in the control condition,who received no information as a guide for inter-generational behavior, was similar to that of theparticipants who received information on their pre-decessors' nongenerous intergenerational behavior.

TABLE 2Results, Study 4"

Condition

1. Control2. Alternative model only3. Alternative model plus

nongenerous previous generation4. Alternative model plus generous

previous generation

Mean AmountKept

5.435.144.75

3.55

n

77

12

11

" Overall F3,33 = 4.68, p < .01.

The data support the hypothesis from Study 1that the greater the amount of resources perceivedas left by previous generations, the greater theamount that will be left to future generations. Aplanned contrast analysis between conditions 3(previous generation was not generous) and 4 (pre-vious generation was generous) indicated a signif-icant difference between the two (t = 2.46, p = .01).Note that the presence of the self-interested exam-ple was constant across these two conditions. Thedata were also compatible with the hypothesis thatin the presence of another, competing model ofintergenerational behavior, the behavior of the pre-vious generation will prevail as the dominant in-fluence on the behavior of the present generationwith respect to future generations. Anotherplanned contrast indicated a significant differencebetween conditions 2 (only a self-interested exam-ple was provided) and 4 (both a self-interested ex-ample and information indicating that the previousgeneration was generous were provided) (t = 2.81,p < .01). In addition, the qualitative data are com-patible with this hypothesis. Specifically, althoughseveral people referred to the behavior of the pre-vious generation in conditions 3 and 4 as a factor intheir decisions, none of the participants in anycondition mentioned the self-interested example.

There was no significant difference between con-ditions 2 and 3, which was expected, because inboth cases all the information provided suggestednongenerosity. However, consistent with the re-sults of Study 3, the nature of the allocation deci-sions in Study 4, although they were relativelyself-interested among the participants in condi-tions 1, 2, and 3, also reflected concern for thewelfare of the future generation and considerationsof equality. The results from the follow-up open-ended question were also consistent with this ob-servation: 12 out of 37 (32%) of the participantsexplicitly mentioned fairness as a factor that influ-enced their allocation decisions, and 16 out of 37(43%) explicitly stated that concern for equality


was a factor in their decision process. The samethree equality-based decisions rules that were ob-served in Study 3 also dominated in this study.Thirty out of 37 participants (81%) used, and manyexplicitly stated, that their decisions were based onone of these three rules.

For the manipulation check, the three questionswere combined to form a scale. A reliability test in-dicated an alpha of .94. The mean response for con-dition 3 (x = 2.14) was significantly lower than themean response for condition 4 (x = 4.45; F = 21.91,p < .001), thus confirming the effectiveness of themanipulation of intergenerational reciprocity.

DISCUSSION

The pattern of results over these four studiesoffers consistent evidence that the behavior of aprevious generation influences the behavior of apresent generation toward future generations, inthe allocation of both benefits (Studies 1,3, and 4)and burdens (Study 2). The participants were fromtwo very different populations (business graduatestudents in Studies 1 and 2, and university admin-istrative staff members in Studies 3 and 4). Yet thefour studies yielded similar and compatible results,indicating greater generalizability of and confi-dence in the results than might otherwise be thecase. It is also notable that, contrary to what cynicsor advocates of game-theoretic and economic mod-els of cooperation might predict, in all the studiesthe majority of people acted on the behalf of futuregenerations to some extent, even though they expe-rienced no monetary benefit to themselves fromsuch contributions. This observation is especiallycompelling in relation to Studies 3 and 4, whereparticipants experienced real consequences of theirdecisions. In support of this trend, it was evidentacross studies that considerations of fairness,equality, and justice came into play.

Studies 3 and 4 investigated the role of alterna-tive models of intergenerational behavior. Sincethey both involved participants from the same pop-ulation, and the allocation situation and instruc-tions were comparable, the two studies can be con-sidered together as a way to observe trends. In allcases in which only nongenerous models (eitherthe behavior of the previous generation or an alter-native model) were available to the present gener-ation, their behavior was strikingly similar to thatof participants in the control condition (and, infact, it was not statistically different from it inStudy 4). This observation suggests an alternativeexplanation for the data involving a multistage de-cision process, which I describe below.

When deciding how to treat future generations.

people first look for information about what mightbe appropriate behavior. One source of informationis how the previous generation treated them, al-though this need not be the only source, as demon-strated in Study 3. When the behavior of the previ-ous generation is accessible, people then judgewhether to accept it as appropriate or not. An ap-propriate model would reflect both fairness andfeasibility. If deemed acceptable, then people fol-low its example. If rejected as inappropriate, thentheir decisions are not influenced by it, and theymay search for alternative models. If an alternativemodel is found, then individuals once again judgeit as acceptable or not. If it is acceptable, then itmay influence their decision. This thesis wouldalso be compatible with the qualitative data fromStudy 1, in which participants cited the behavior ofthe previous generation as a factor influencingtheir decisions only when the prior generation wasgenerous.

Applying this thesis to the data in Studies 3 and4, in the cases (experimental conditions) in whichthe only available models were nongenerous (con-ditions 1 and 3 in Study 3, and conditions 2 and 3in Study 4), participants may have deemed themodels inappropriate and thus rejected them asinfluences on their decisions. This would explainwhy the decisions in these conditions looked muchlike those in the control condition, where no mod-els were available, in Study 4. In contrast, when theavailable model was generous (conditions 2 and 4in Study 3 and condition 4 in Study 4), then indi-viduals accepted it as legitimate intergenerationalbehavior and allowed it to influence their deci-sions. This thesis is compatible with the observedlower level of self-interested behavior when eitherthe previous generation or an alternative modelwas generous. Note that in these cases, individualsappear to be responding to the spirit of generosityrather than directly mimicking the behavior of pre-vious generations or alternative models of intergen-erational behavior.

A helpful extension to Study 4 would be to offerboth a generous alternative model and a nongener-ous previous generation to participants in a singleexperimental condition. If people follow^ed the be-havior of the prior generation (acted ungenerously)instead of the alternative model, then (in view ofthe data for Study 4), support for the hypothesis inStudy 4 would be stronger. If, instead, they actedgenerously, thus suggesting that they were follow-ing the behavior of the alternative model despiteaccess to information on the behavior of the priorgeneration, the data would lend support to the no-tion of the multiphase decision process I have justproposed. Implementing this experimental design


would be a useful direction for future researchsince it would, in effect, fully cross the conditionsfor Study 4, thus addressing this limitation of thecurrent research.

Future Research Directions

This series of studies on intergenerational reci-procity raises many questions. The results showthat intergenerational reciprocity can operate in al-locations of both benefits and burdens. An interest-ing direction for future research might be to di-rectly compare benefit and burden allocation in thesame study to explore whether intergenerationalallocation decisions differ depending on the va-lence of the resource that is involved.

Other questions researchers could ask include,When will a present generation look at the behaviorofthe previous generation and view it as a model ofwhat not to do? Or when will a generation do whatit thinks it should do, instead of reciprocating inkind? There are likely to be situations in whichindividuals follow the urging of moral theorists andtreat the next generation in the way they would liketo have been treated themselves instead of the waythey were actually treated. One hypothesis mightbe that trend-breaking behavior occurs at the ex-tremes—either positive or negative. Very bad treat-ment by previous generations may be perceived asunfair and inappropriate, and exceptionally goodtreatment may be viewed as unnecessary, beyondthe call of duty, or perhaps even impossible toreplicate. In these cases, alternative models of in-tergenerational behavior may be sought out that areconsidered more appropriate or reasonable.

Another issue raised by this research is that ofthe intent and capabilities of the previous genera-tions. For example, in some cases the previousgeneration might have had good intentions andtried its best to be generous toward the next gener-ation but had been prevented from doing so bycircumstances beyond its control. Is it, then, theactual outcome that influences the behavior of thenext generation toward subsequent generations? Ordoes the next generation take into account the ca-pabilities and intentions of the previous generationwhen it makes decisions that influence subsequentgenerations? This question is especially interestingwhen intent and outcome are not consistent (forexample, there is a good intention with a bad out-come) or when the intent is inferred from the out-come when the actual intent is not known.

A factor that may moderate intergenerational rec-iprocity is whether the resource being allocatedamong generations is continuing or finite. That is,will the next generation be the last one in the se-

quence, or will the sequence continue after the nextgeneration? This is an important issue for real-world applications since some resources have thepotential to continue indefinitely if managed prop-erly (for example, fisheries), and others (such asfossil fuels) are more finite, with allocation deci-sions realistically involve only a few generations.The effect of intergenerational reciprocity might beexpected to be weaker if a resource is finite. Thereason for this is that the behavior of the previousgeneration would be a more appropriate modelwithin the context of a continuing sequence, sincethe decision that the previous generation made wassimilar to the one that the present generation wouldnow be making. In contrast, if the resource is finite,the allocation decision is more like a two-partyallocation than one within the context of a continu-ing sequence.

Finally, the intergenerational problems exam-ined here are limited to situations in which actorsin differeiit generations do not simultaneously existin the focal social exchange context. Thus, therewas a complete decoupling of interests betweengenerations, in that the present generation did notsuffer any of the future costs or reap any of thefuture benefits associated with their decisions. Inmany real-world intergenerational contexts, espe-cially those in organizations, where time horizonsare shorter, things are not nearly as clean and neat.As some of the boundary conditions of the currentresearch are relaxed, additional complexities in in-tergenerational relations inevitably emerge. Futureresearch might explore situations in which differ-ent generations have opportunities to interact,some level of interdependence, and compatible in-terests as well as conflicting ones.

Implications for Organizations

Intergenerational behavior in organizations is im-portant when it is in the organizations' long-termbest interest to take future interests into account,yet the interests of the present generation are atodds with those of future generations. A provoca-tive aspect of Study 2 is that intergenerational be-havior could be changed by simply reframing theway the present generation viewed the same histor-ical situation, in one case emphasizing the gooddone by previous generations, and in the other,highlighting their negligence. One prescriptive im-plication for managers is to explicitly highlight thebeneficial intergenerational behaviors of previousgenerations and to downplay the nonbeneficialones in order to promote beneficent intergenera-tional behavior on the part of the present genera-tion of employees.


This research also has implications in the areasof productivity, training, learning, culture, andnorms in organizational contexts. For example,consider the case of incoming and outgoing em-ployees rotating through positions. These succes-sions constitute an individual-level intergenera-tional situation like that simulated in Studies 3 and4. The results of these studies offer some insightinto what might dictate the smoothness of suchtransitions. What will determine whether or not adeparting employee will give an incoming personkey pieces of information, such as files and notes,that will enable the incoming person to perform hisor her new job better and become acclimated moreeasily? This research suggests that the answer tothis question will partly depend on how the outgo-ing person was treated by his or her predecessorwhen he or she was first starting the job. In organi-zations, there is often no systematic monitoring orunderstanding of these transitions. This researchsuggests that it is important for organizations toprovide incentives for people to "transfer benefi-cently." Nongenerosity may perpetuate itself acrossgenerations and interfere with employee perfor-mance as well as with organizational learning overtime. Study 3 suggests that a possible interventionin this case would be to expose the present gener-ation to more appropriate or desirable alternativemodels of intergenerational behavior.

Managing intergenerational transitions well canhelp organizations to perform better in fast-chang-ing, unpredictable markets. Theorists and research-ers on pace have argued that managing transitionsin organizations is critical to successful time pacing(Eisenhardt & Brown, 1998). Good transitions cansave time and help companies perform better,while bad transitions can cause businesses to loseposition and fall behind. One type of transition thatcompanies face is the transition among differentgenerations of organizational actors. Intergenera-tional transitions, like other transitions in organi-zations, affect organizations' ability to be good timepacers, and good time pacing affects their ability tocompete. This research identifies intergenerationalreciprocity as a factor that may impact the effec-tiveness of these transitions.

Concluding Comments

Intergenerational considerations are relevant andimportant to individuals, organizations, and soci-eties alike. Intergenerational problems are typicallycomplicated and, consequently, they can be diffi-cult to study. In this article, I offer a framework forthinking about intergenerational situations system-atically and recognizing them in a variety of con-

texts. I also suggest one approach to the method-ological obstacles associated with intergenerationalresearch—conducting multiple studies to investi-gate an intergenerational phenomenon from differ-ent angles. It was my aim to bring attention to thetopic of intergenerational issues, offer suggestionsfor future research, and suggest initial insights thatmay help decision makers in various domains copemore effectively with these challenging problems.

REFERENCES

Babcock, L., Loewenstein, G., Issacharoff, S., & Camerer,C. 1995. Biased judgments of fairness in bargaining.American Economic Review, 85: 1337-1343.

Bandura, A. 1969. Social learning of moral judgments.Journal of Personality and Social Psychology, 11:275-279.

Bandura, A., & Walters, R. H. 1963. Social learning andpersonality development. New York: Holt, Rinehart& Winston.

Barry, B. 1989. Theories of justice. Berkeley: Universityof California Press.

Bazerman, M. H., & Neale, M. A. 1982. Improving nego-tiation effectiveness under final offer arbitration:The role of selection and training. Journal of Ap-plied Psychology, 67: 543-548.

Becker, L. C. 1986. Reciprocity. New York: Routledge &Kegan Paul.

Blau, P. M. 1964. Exchange and power in social life.New York: Wiley.

Brennan, T. J. 1995. Discounting the future: Economicsand ethics. Resources, 120: 3-6.

Brewer, M. B. 1981. Ethnocentrism and its role in inter-personal trust. In M. B. Brewer & B. E. Collins (Eds.),Scientific inquiry and the social sciences: 345-360.New York: Jossey-Bass.

Care, N. S. 1982. Future generations, public policy, andthe motivation problem. Environmental Ethics, 4(3):195-213.

Carroll, C. R., & Hannan, M. T. 1995. Organizations inindustry: Strategy, structure, and selection. NewYork: Oxford University Press.

Cowen, P. A., Langer, ]., Heavenrich, J., & Nathanson, M.1969. Social learning and Piaget's cognitive theory ofmoral development. Journal of Personality and So-cial Psychology, 11: 261-274.

Diekmann, K. A., Samuels, S. M., Ross, L., & Bazerman,M. H. 1997. Self-interest and fairness in problems ofresource allocation: Allocators versus recipients.Journal of Personality and Social Psychology, 72:1061-1074.

Eisenhardt, K. M., & Brown, S. L. 1998. Time pacing:Competing in meu-kets that won't stand still. Har-vard Business Review, 76(2): 59-69.


Ekeh, P. P. 1974. Social exchange theory. Cambridge,MA: Harvard University Press.

Gerard, R. W., Kluckhohn, C, & Rapoport, A. 1956. Bio-logical and cultural evolution: Some analogies withexplorations. Behavioral Science, 1: 6-34.

Gouldner, A. W. 1959. Reciprocity and autonomy infunctional theory. In L. Goss (Ed.), Symposium onsociological theory: 240-270. New York: Harper &Row.

Gouldner, A. W. 1960. The norm of reciprocity. Ameri-can Sociological Review, 25: 161-167.

Griffith, W. I., & Sell, J. 1988. The effects of competitionon allocators' preferences for contxibutive and retrib-utive justice rules. European Journal of Social Psy-chology, 18: 443-455.

Hawken, P. 1993. The ecology of commerce: A declara-tion of sustainability. New York: Harper Business.

Insko, C. A., Gilmore, R., Moehle, D., Lipsitz, A., Drenan,S., & Thibaut, J. W. 1982. Seniority in the genera-tional transition of laboratory groups: The effects ofsocial familiarity and task experience. Journal ofExperimental Social Psychology, 18: 557-580.

Insko, G. A., Thibaut, J. W., Moehle, D., Wilson, M.,Diamond, W. D., Gilmore, R., Solomon, M. R. &Lipsitz, A. 1980. Social evolution and the emergenceof leadership. Journal of Personality and SocialPsychology, 39: 431-448.

Jacobs, R. G., & Gampbell, D. T. 1961. The perpetuation ofan arbitrary tradition through several generations oflaboratory microculture. Journal of Abnormal andSocial Psychology, 62: 649-658.

Jennings, P. D., & Zandbergen, P. A. 1995. Ecologicallysustainable organizations: An institutional ap-proach. Academy of Management Review, 20:1015-1052.

Kotlikoff, L. J. 1992. Generational accounting. NewYork: Free Press.

Kramer, R. M. 1993. Cooperation and organizationalidentification. In J. K. Murnighan (Ed.), Social psy-chology in organizations: Advances in theory andresearch: 244-268. Englewood Gliffs, NJ: Prentice-Hall.

Lamm, H., & Kayser, E. 1978. The allocation of monetarygain and loss following dyadic performance: Theweight given to effort and ability under conditions oflow and high intradyadic attraction. European Jour-nal of Social Psychology, 8: 275-278.

Levi-Strauss, G. 1949. Les structures elementaires de laParente [Parente's elementary structures]. Paris:Presses Universitaires de France.

Malinowski, B. 1939. The group and the individual infunctional analysis. American Journal of Sociology,44: 938-964.

Mannix, E. A., Neale, M. A., & Northcraft, G. B. 1995.Equity, equality, or need? The effects of orgeiniza-

tional culture on the allocation of benefits and bur-dens. Organizational Behavioral and Human De-cision Processes, 63: 276-286.

Messick, D. M., & Sentis, K. P. 1983. Fairness, preferenceand fairness biases. In D. M. Messick & K. S. Gook(Eds.), Equity theory: Psychological and sociologi-cal perspectives: 61-94. New York: Praeger.

Morgan, T. O. 2000. Life's like that. Reader's Digest,November: 105.

Neale, M. A., & Bazerman, M. H. 1983. The role of per-spective-taking ability in negotiating under differentforms of arbitration. Industrial & Labor RelationsReview, 36: 378-388.

Northcraft, G. B., Neale, M. A., Tenbrunsel, A., &Thomas, M. 1996. Benefits and burdens: Does it re-ally matter wliat we allocate? Social Justice Re-search, 9: 27-45.

Parsons, T. 1951. The social system. New York: FreePress.

Plater, Z. J. B., Abrams, R. H., Goldfarb, W. 1992. Envi-ronmental law and policy: Nature, law, and soci-ety. St. Paul: West.

Portney, P. R., & Weyant, J. P. 1999. Discounting andintergenerational equity. Washington, DG: Re-sources for the Future.

Rawls, J. 1971. A theory of justice. Gambridge, MA:Harvard University Press.

Richards, D. A. J. 1982. Gontractarian theory, intergen-erational justice, and energy policy. In D. MacLean &P. G. Brown (Eds.), Energy and the future: 131-150.Totowa, NJ: Rowman & Littlefield.

Rose, E., & Felton, W. 1955. Experimental histories ofculture. American Sociological Review, 20: 383-392.

Rutte, G. G., Wilke, H. A. M., & Messick, D. M. 1987.Scarcity or abundance caused by people or the envi-ronment as determinants of behavior in the resourcedilemma. Journal of Experimental Social Psychol-ogy, 23: 208-216.

Shrivastava, P. 1995. The role of corporations in achiev-ing ecological sustainability. Academy of Manage-ment Review, 4: 936.

Sondak, H., Neale, M. A., & Pinkley, R. L. 1995. Thenegotiated allocation of benefits and burdens: Theimpact of outcome valence, contribution, and rela-tionship. Organizational Behavioral and HumanDecision Processes, 64: 249-260.

Starik, M., & Rands, G. P. 1995. Weaving an integratedweb: Multilevel and multisystem perspectives ofecologically sustainable organizations. Academy ofManagement Review, 20: 908.

Thompson, L., & Loewenstein, G. 1992. Egocentric inter-pretations of fairness and interpersonal conflict. Or-ganizational Behavior and Human Decision Pro-cesses, 51: 176-197.


Tornblum, K. Y. 1988. Positive and negative allocation: Atypology and model for conflicting justice princi-ples. In E. Lawler & B. Markovsky (Eds.), Advancesin group processes, vol. 5:141-165. Greenwich, CT:JAI Press.

Wade-Benzoni, K. A., Tenbrunsel, A. E., & Bazerman,M. H. 1996. Egocentric interpretations of fairness inasymmetric, environmental social dilemmas: Ex-plaining harvesting behavior and the role of commu-nication. Organizational Behavior and Human De-cision Processes, 67: 111-126.

Walster, E., Walster, G. W., & Berscheid, E. 1978. Equity:Theory and research. Boston: Allyn & Bacon.

Weick, K. E., & Gilfillan, D. P. 1971. Fate of arbitrarytraditions in a laboratory microculture. Journal ofPersonality and Social Psychology, 17: 179-191.

Weiss, E. B. 1989. In fairness to future generations:International law, common patrimony, and inter-generational equity. New York: TransnationalPublishers.

Yamagishi, T., & Gook, K. S. 1993. Generalized exchange

and social dilemmas. Social Science Quarterly, 56:235-248.

Zucker, L. G. 1977. The role of institutionalization incultural persistence. American Sociological Re-view, 42: 726-743.

Kimberly A. Wade-Benzoni ([email protected])is an assistant professor of management and organiza-tional hehavior at New York University's Stern Schoolof Business. She received her Ph.D. in organizationbehavior from Northwestern University's KelloggSchool of Management. Her research interests includeintergenerational hehavior, the role of cognitive andmotivational factors in conflict management and deci-sion making, and the interrelationships hetween organ-izational and societal interests.

-AA-

A GOLDEN RULE OVER TIME: RECIPROCITY IN … Classes/OA YA/articles/treating the...A GOLDEN RULE OVER TIME: RECIPROCITY IN INTERGENERATIONAL ALLOCATION DECISIONS KIMBERLY A. WADE-BENZONI

Documents