Getting Ahead of the Joneses: When Equality Increases Conspicuous Consumption among Bottom-Tier Consumers

Journal of Consumer Research, Inc.

Getting Ahead of the Joneses: When Equality Increases Conspicuous Consumption amongBottom-Tier ConsumersAuthor(s): Nailya Ordabayeva and Pierre ChandonSource: Journal of Consumer Research, Vol. 38, No. 1 (June 2011), pp. 27-41Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/10.1086/658165 .

Accessed: 15/05/2014 12:48

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR todigitize, preserve and extend access to Journal of Consumer Research.

http://www.jstor.org

This content downloaded from 194.29.185.52 on Thu, 15 May 2014 12:48:14 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/stable/10.1086/658165?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


27

� 2010 by JOURNAL OF CONSUMER RESEARCH, Inc. ● Vol. 38 ● June 2011All rights reserved. 0093-5301/2011/3801-0008$10.00. DOI: 10.1086/658165

Getting Ahead of the Joneses:When Equality IncreasesConspicuous Consumption amongBottom-Tier Consumers

NAILYA ORDABAYEVAPIERRE CHANDON

It is widely believed that increasing the equality of material possessions or incomein a social group should lead people at the bottom of the distribution to consumeless and save more. However, this prediction and its causal mechanism have neverbeen studied experimentally. Five studies show that greater equality increases thesatisfaction of those in the lowest tier of the distribution because it reduces thepossession gap between what they have and what others have. However, greaterequality also increases the position gains derived from status-enhancing consump-tion, since it allows low-tier consumers to get ahead of the higher proportion ofconsumers clustered in the middle tiers. As a result, greater equality reduces con-sumption when consumers focus on the narrower possession gap, but it increasesconsumption when they focus on the greater position gains (i.e., when consumptionis conspicuous, social competition goals are primed, and the environment is com-petitive).

Thorstein Veblen coined the term “conspicuous consump-tion” to describe the acquisition and display of possessions

with the intention of gaining social status (1899). It is wellestablished that poorer households save a smaller fraction oftheir income than richer households (1% for households in thelowest U.S. income quintile vs. 24% for those in the highestquintile) and that people at the bottom of the income or material

Nailya Ordabayeva ([email protected]) is assistant professor ofmarketing, Rotterdam School of Management, Erasmus University, PO Box1738, 3000 DR, Rotterdam, the Netherlands. Pierre Chandon ([email protected]) is professor of marketing, INSEAD, Boulevard deConstance, 77305 Fontainebleau, France. The authors gratefully acknowl-edge the financial support of R&D INSEAD and thank Robert H. Frank,Joseph Lajos, Olivier Toubia, Christophe Van den Bulte, and Homa Zar-ghamee for valuable comments. The authors also thank Liselott Pettersson,Cecile Adam, and Dimitri Vasiljevic at INSEAD Social Sciences ResearchCentre for help with the data collection. The helpful inputs of the editor,the associate editor, and the reviewers are gratefully acknowledged.

John Deighton served as editor and Rebecca Ratner served as associateeditor for this article.

Electronically published December 8, 2010

possession distribution spend more on conspicuous and incon-spicuous consumption (Bloch, Rao, and Desai 2004; Duesen-berry 1949; Dynan, Skinner, and Zeldes 2004; Moav and Nee-man 2008). Despite evidence that low savings and highconsumption, especially conspicuous consumption, are mostharmful to low-income households (Bagwell and Bernheim1996; Knell 1999), this pattern has intensified over the past 20years because of growing income and consumption inequality(Christen and Morgan 2005; Zhu, forthcoming).

The most common explanation for why inequality de-creases saving and increases consumption—especially con-spicuous consumption—among the least well-off consumersis that these consumers are trying to “keep up with theJoneses” (Christen and Morgan 2005; Dreze and Nunes 2009;Frank 1985a). In essence, the argument goes that consumersat the bottom of the distribution spend a larger proportion oftheir budget on status-conferring consumption in order toreduce the dissatisfaction they feel with their current level ofpossessions due to the widening gap between what they haveand what others have (Dupor and Liu 2003; Elster 1991;Solnick and Hemenway 1998). Although there is little doubtthat equality and consumption are linked, at least four im-



28 JOURNAL OF CONSUMER RESEARCH

portant issues remain unaddressed in the current literature.The first issue concerns the oft-repeated recommendation thatgovernments increase equality by imposing measures such asprogressive consumption taxes, which penalize high levels ofconsumption but not high levels of savings, or luxury taxes,which penalize specific status-conferring consumption suchas purchases of high-end cars or boats (Becker, Murphy, andWerning 2005; Frank 1985b). While a few theoretical modelshave examined these ideas (Duncan and Sabirianova Peter2008; Hopkins and Kornienko 2004), there is to date no directexperimental evidence that increasing equality does indeedincrease savings and reduce consumption. Second, existingstudies have looked at the effects of inequality on the overalllevel of consumption rather than on the consumption deci-sions of consumers at the bottom of the distribution, thosewho are most at risk if they overspend. Third, existing em-pirical analyses linking equality and consumption have notdistinguished between conspicuous (status-related) and in-conspicuous (status-neutral) consumption. Finally, prior re-search has relied on a simplified model of social comparisonsthat focuses on the role of the “possession gap” (the differencebetween what one person has and what other people have),and it has neglected the role of the “position gains” (theincrease in social rank) provided by consumption.

The main objective of this research is to test experimentallythe effects of increasing equality on the conspicuous andinconspicuous consumption of consumers at the bottom ofthe distribution. Consistent with prior research, which foundthat social comparisons are mostly made with a limited num-ber of individuals who are highly relevant and similar to theself (Festinger 1954), we focus on the effects of equality atthe level of people’s immediate social group (and not, say,at the national level) and study how spending decisions affectone’s status within this small group. Following Dreze andNunes (2009), we define “status” as one’s relative position(or rank) in a social group, where position can be broadlyconstrued and unobservable (e.g., in terms of income) or morenarrowly construed and observable (e.g., in terms of one’sendowment with specific status-granting possessions). Be-cause endowment is often only observed at an ordinal level(e.g., whether one owns a Coach or a Chanel bag), we measurethe equality of the distribution as the relative proportion ofpeople in the middle versus the extreme tiers. We considerthat the equality of a distribution increases when, ceteris par-ibus, the number of people in the middle tiers increases whilethe number of people in the top tiers decreases. Equalitytherefore decreases when people are more uniformly distrib-uted across all tiers because there are fewer people in themiddle and more people at the extremes. Clearly, there areother ways in which the equality of a distribution may beincreased. For example, taking from the rich to give to thepoor increases equality by narrowing the range of the distri-bution and increasing the endowment of the poor. However,by maintaining a constant range and endowment for the targetgroup, our operationalization of greater equality captures theeffects of policies such as luxury taxes and progressive con-sumption taxes and allows us to isolate the effects of two

important mediators—possession gap and position gain—while ruling out other mechanisms.

Using these definitions we build a framework of the effectof equality on the preference for conspicuous or inconspic-uous consumption over saving of people at the base of thedistribution. This framework clarifies the dual effects ofequality on the possession gap and the position gain pro-vided by consumption and the way these two factors influ-ence consumption decisions. In a series of five experimentswe test our key hypothesis that increasing equality actuallyincreases consumption when people at the bottom of thedistribution care about social position, that is, when pur-chasing status-enhancing (vs. status-neutral) products, whensocial competition (vs. social indifference) goals are acti-vated, and in a competitive (vs. cooperative) social envi-ronment. In the general discussion we review the implica-tions of these findings for consumer research on the effectsof status and for the debate on the value of redistributionpolicies such as consumption and luxury taxes as a meansof improving the welfare of those at the base of the pyramid.

THE EFFECTS OF EQUALITY ONCONSUMPTION

There is a large body of research in economics, social psy-chology, and consumer research on the effects of status andsocial comparisons on consumption (Amaldoss and Jain 2005;Berger and Heath 2007; Dreze and Nunes 2009; Frank 1985a;Griskevicius et al. 2007; Leclerc, Hsee, and Nunes 2005;Mandel, Petrova, and Cialdini 2006; Richins 1994; Ruckerand Galinsky 2008). In this section we introduce our researchquestion with a stylized example that allows us to explainhow we implement and measure increasing equality. We thenreview existing studies on the role of social envy and dis-satisfaction caused by possession gaps that suggest that in-creasing equality within a social group should reduce con-sumption among people at the bottom of the distribution.Finally, we present new hypotheses on the role of the socialposition gains provided by conspicuous consumption, whichsuggest the opposite.

A Stylized Example: Effects of Equality onConsumption

Consider someone who is preparing for her 10-year highschool reunion. She needs to decide whether to take her oldunbranded bag or to buy a new handbag with a prestigiousbrand. (Note that this example could be readily applied toa masculine fashion accessory, such as a watch.) Figure 1shows the consumer’s expectation of the proportion of otherwomen attending the reunion who will have one of fivetypes of bags, from tier 5 (consisting of unbranded hand-bags) to tier 1 (consisting of limited-edition bags from pres-tige luxury brands). In the low-equality distribution, peopleare roughly uniformly spread across all tiers (10% of thepeople are in tier 5, 20% in tier 4, 20% in tier 3, 25% intier 2, and 25% in tier 1). In the high-equality distribution,



GETTING AHEAD OF THE JONESES 29

FIGURE 1

STYLIZED EXAMPLE: HIGH VERSUS LOW EQUALITY OF MATERIAL POSSESSIONS IN A SOCIAL GROUP

there are fewer people in the top tiers and more people inthe middle tiers (10% of the people are in tier 5, 40% intier 4, 20% in tier 3, 20% in tier 2, and 10% in tier 1). Thehigh-equality distribution is more concentrated (HerfindahlH index p .26 vs. .22 for the low-equality distribution),and the concentration occurs around the middle tiers. (Notethat we use the H index to measure the equality of possessiondistributions because, unlike the GINI index that we use forincome distributions, it does not require an interval scaleand can therefore be applied to ordered categorical data suchas the quality and prestige of handbag brands. The H indexis not a perfect measure of equality either, because it doesnot take into account the fact that the type of possessionscan be ordered. For example, a more equal distribution with40% of people in two adjacent middle tiers has the same Hindex as one with 40% of people in the top and bottomtiers. This is why our definition of greater equality requiresmore concentration in the middle tiers.)

The difference between low- and high-equality distribu-tions in figure 1 mirrors the growing inequality of incomedistribution in the United States over the past 30 years (Nunes,Johnson, and Breene 2004). More importantly, it captures theeffects of policies frequently used to increase equality, in-cluding luxury taxes (which would be applied to only themost expensive handbags) and progressive consumption taxes(which kick in at high spending levels). Both taxes reducethe number of people buying high-status bags and increasethe number of people buying cheaper bags. For example, theintroduction of a luxury tax on expensive boats in the UnitedStates in the 1990s reduced the demand for boats costingabove $100,000 by 93% (Hyder 1991). A similar shift wouldoccur in the case of restrictive quotas on imports of specifichigh-status products (e.g., Australia’s restrictive import quotas

on high-end wines and spirits), which, unlike taxes, do notdirectly affect product prices. Therefore, in figure 1 and inthe experiments we manipulate equality by changing theshape of the distribution but not the range of prices. Thisallows us to focus on the effects of changes in the distributionof people across tiers while controlling for the effects of price.

Note that the range (five tiers) and the proportion of peo-ple (10%) in the fifth (bottom) tier are identical in the twodistributions shown in figure 1 as well as those used in ourstudies. This is consistent with the effects of luxury andprogressive consumption taxes, which only influence whathappens in the middle and top tiers but not in the bottomtier. Keeping the range of the distribution and the percentileposition of the target consumer constant is also importantbecause it allows us to control for the well-known effectsof these two factors on people’s evaluation of their currentlevel of possessions (Parducci 1965). In the general dis-cussion we consider the effects of different redistributionpolicies that shrink the range of the distribution or increasethe endowment of the bottom tier instead of simply changingthe equality of the distribution (e.g., taking from the rich togive to the poor).

Keeping up with the Joneses: The Effects of thePossession Gap

What predictions would existing studies make about theeffects of such an equality increase on consumption? Socialcomparison theory argues that people have an inherent ten-dency to compare themselves to others in order to judge howwell they are doing (Festinger 1954). Although people engagein both upward (unfavorable) and downward (favorable) com-parisons, upward comparisons occur faster (Wood 1989) and




arise by default for self-evaluation purposes (Collins 1996).Unfavorable comparisons with what others have lead to envyand feelings of inferiority, which in turn motivate people withlower levels of possessions to “keep up with the Joneses” byattaining and publicly displaying the same level of possessionsas those who are better off (Christen and Morgan 2005; Clarkand Oswald 1998; Dupor and Liu 2003; Elster 1991; Fein-berg, Krishna, and Zhang 2002; McCormick 1983). However,although envy is thought to be a key motivation for “keepingup with the Joneses,” it is a complex construct, and its effectson consumption are not completely understood (Miceli andCastelfranchi 2007; Smith and Kim 2007).

A related stream of research (Frank 1989, 2007; Hsee,Rottenstreich, and Xiao 2005) focuses on the role of sat-isfaction with one’s current possessions (which is differentfrom envy) and of learning from the behavior of others (asopposed to competing with them). The key argument is that,when equality is low, people at the bottom of the distributioninfer from the higher number of people with high levels ofpossessions that acquiring possessions will lead to greaterincreases in their satisfaction with their possession even ifenvy or competition are irrelevant to the particular product(Frank 2007). Conversely, when equality is high, people atthe bottom of the distribution infer from the higher numberof people in the middle tiers who own products that are notthat different from theirs that upgrading would not increasetheir satisfaction as much. For example, owners of the basicversion of computer software should be more likely to up-grade to the full version if more people own the full versionbecause they would infer that it signals that the full versionis significantly better even if this particular product is con-sumed privately and is neither a source of pride nor envy.

To summarize, existing research builds on the generalpremise that people are subject to social comparisons andfocuses on the role of the possession gap between what thetarget consumer has and what others have. Importantly, thesestudies have not always specified who the others providingthe reference level are, and hence they have used differentmeasures of central tendency of the distribution (mean ormedian levels of possession) to measure possession gaps(Alpizar, Carlsson, and Johansson-Stenman 2005; Clark andOswald 1998; Duesenberry 1949; Festinger 1954; Knell1999). For our purposes, and since we are studying theimpact of small groups of relevant others (e.g., neighbors,former classmates at a high school reunion), we assume thatthe entire distribution is relevant when making social com-parisons and that it does not matter which particular measureof central tendency (mean or median) is used as long asincreasing equality reduces both of them, which is the casein our stylized example and all our experiments. Returningto our stylized handbag example, figure 1 shows that thepossession gap for people in the fifth (bottom) tier is nar-rower in the high-equality condition because the type of bagthat they own is closer to the mean or median handbag inthe high-equality distribution than in the low-equality dis-tribution. Existing research would therefore predict that in-creasing equality should reduce the spending of lowest-tier

consumers because it reduces their possession gap and hencemakes them more satisfied with their original bag and lesswilling to upgrade to a higher-status bag.

Getting Ahead of the Joneses: The Effects of thePosition Gains

We propose that conspicuous consumption decisions arenot only influenced by the size of the possession gap butalso by the status or position gain in the distribution thatconsumption would provide. This is consistent with Veb-len’s original argument that consumption can be driven bythe “desire of every one to excel every one else in theaccumulation of goods [where] the end sought by accu-mulation is to rank high in comparison with the rest of thecommunity in points of pecuniary strength” (1899, 39). Inother words, people are not merely satisfied to be on a parwith their peers by eliminating the possession gap; they takeinto consideration the improvement in their rank in the dis-tribution, or position gain, that they can obtain through con-sumption (i.e., the return on status consumption).

Taking into account improvement in status, or position gain,rather than just the possession gap leads to markedly differentpredictions regarding the effects of increasing equality for bot-tom-tier consumers. This is because while increasing equalityindeed narrows the possession gap for these consumers, it alsoincreases the percentage of people that can be surpassed andhence the potential position gains that can be obtained throughconspicuous consumption. Consider again the handbag ex-ample shown in figure 1. If a fifth- (bottom-) tier consumerdecides to buy a third- (middle-) tier bag, she will get aheadof the people in tiers 5 and 4. Since tier 4 comprises 40% ofthe people in the high-equality distribution but only 20% ofpeople in the low-equality distribution, the same conspicuousspending offers twice the position gain when equality is highthan when it is low. More generally, conspicuous consumptionprovides greater position gains for low-tier consumers whenequality is high (rather than low). Indeed, in the extreme caseof perfect equality, even the smallest conspicuous consump-tion will guarantee the top status position, and hence theposition gain hypothesis would predict maximum conspicu-ous consumption. In contrast, existing studies would predictthat the motivation to consume should disappear under perfectequality when the possession gap is eliminated (McCormick1983).

Summary and Outline of Experiments

To summarize, we expect that increasing the equality ofthe endowment distribution will reduce envy and increasesatisfaction among people in the lowest tier of the distri-bution because it narrows the gap between their level ofpossessions and the possessions of other people in the socialgroup. We also expect that increasing equality will amplifythe social position gains provided by conspicuous con-sumption for people at the bottom of the distribution. There-fore, we expect that increasing equality will reduce con-




sumption when people do not seek status but will increaseconsumption when people seek status.

We test these hypotheses in five experiments. In study 1, wetest our key prediction about the effects of equality on envyand conspicuous spending decisions. In study 2, we directlymanipulate people’s focus on the possession gap or the positiongain and rule out the alternative explanation that people simplywant to join the majority. In study 3, we examine the effectsof equality on conspicuous and inconspicuous consumption andfurther examine the process underlying these effects by directlymeasuring satisfaction and perceived position gains. We testthe boundaries of the effects by looking at what happens whenpeople are not motivated to acquire status by priming themwith social competition or social indifference goals in study4A and by manipulating the competitive or the cooperativenature of the social environment in study 4B. Studies 4A and4B also allow us to test the effects of the equality of incomedistribution and not simply of the equality of the distributionof specific material possessions. All the participants were re-cruited near a large urban university, and so most of them,but not all, were university students. We collected demo-graphic data in three of the five experiments (studies 2, 4A,and 4B). Across these studies, 60% of participants werewomen, the average age was 22, and 84% lived on less thani1,000 per month, suggesting that they could imagine howit feels to have a lower level of income or fewer expensivematerial possessions than others. There were no effects ofgender in any of the studies in which this information wascollected.

STUDY 1: EFFECTS OF EQUALITY OFPOSSESSIONS ON SOCIAL ENVY AND

CONSPICUOUS CONSUMPTION

Method

We recruited 73 people outside a large urban universitycampus and asked them to participate in a study about ev-eryday decisions of home owners in exchange for a mealvoucher. The participants read a neighborhood newsletterdescribing the efforts of a local association to improve theappearance of the neighborhood, which included a pie chartshowing the number of houses in the neighborhood withflower gardens. We used this scenario because a pretest hadindicated that the appearance of one’s garden is a source ofsocial status in middle-class neighborhoods.

The distribution had five tiers: no flowers, one or twoflower bushes, three or four flower bushes, five or six flowerbushes, and seven or more flower bushes. To manipulateequality, the pie charts showed either the low-equality dis-tribution displayed in figure 1 (in which the proportion ofhouses in each tier was 10%, 20%, 20%, 25%, and 25%;H p .22) or the high-equality distribution (respectively,10%, 40%, 20%, 20%, and 10%; H p .26). Increasingequality thus reduced the number of households in the high-est tiers and increased the number of households in themiddle tiers of the distribution. As a result, the possession

gap for people in the bottom (fifth) tier was lower in thehigh-equality condition (where the median possession levelwas in tier 4) than in the low-equality condition (where itwas in tier 3). Conversely, the position gains provided bymoving from tier 5 to tier 3 were larger in the high-equalitycondition (where upgrading enabled tier 5 people to moveup in the distribution by 40 percentage points) than in thelow-equality condition (where it allowed them to move upby only 20 percentage points).

Participants were then asked to rate whether, after readingthe newsletter, a homeowner with no flowers in her garden(and hence in tier 5) would spend i45 on three flower bushes(and thus join tier 3) or whether she would choose to savethe money (on a 9-point scale, where 1 p “definitely save”and 9 p “definitely spend”). Participants were also askedto rate how envious this person would be of her neighbors’gardens (from 1 p “not at all” to 9 p “very much”). Theorder of the two questions was counterbalanced across par-ticipants.

Results and Discussion

We first conducted a pretest to verify that participantsunderstood the impact of the distribution manipulation onthe potential position gains. We asked 45 people similar tothose who participated in study 1 to rate the extent to whichbuying the flower bushes would allow the person in thelowest tier to improve her social position (on a 9-point scale,where 1 p “not at all” and 9 p “very much”). As expected,the gain in social position, and thus in status, was ratedhigher in the high-equality distribution condition (M p 6.7)than in the low-equality distribution condition (M p 3.7;F(1, 43) p 17.6, p ! .01), indicating that the equality ma-nipulation was successful.

To test our main predictions in the main study, we con-ducted two separate ANOVAs on the preference for spend-ing over saving and on social envy with distribution equalityas the between-subjects factor. The results showed that envywas significantly lower in the high-equality condition (M p4.7) than in the low-equality condition (M p 5.9; F(1, 71)p 4.0, p ! .05). In contrast, and consistent with our hy-pothesis, the preference for conspicuous consumption oversavings was significantly higher in the high-equality con-dition (M p 6.9) than in the low-equality condition (M p5.8; F(1, 71) p 4.1, p ! .05).

The results of study 1 support our prediction that increasingequality increases conspicuous consumption among consum-ers in the lowest tier of the distribution even though it de-creases their envy of other people’s possessions. This supportsour hypothesis that the distribution of possessions influencesboth envy and the potential position gains conferred by con-spicuous consumption and that conspicuous consumption canbe motivated by position gains independently of envy.

Study 1 had some limitations. First, it tested the posses-sion gap and the position gain hypotheses indirectly by mea-suring two separate dependent variables, social envy andconspicuous consumption decisions. It would therefore beimportant to directly examine the role of the two constructs




by manipulating people’s focus on either the possession gapor the position gain in each distribution condition. Second,bottom-tier consumers in study 1 were just below the largestgroup in the distribution (40% in tier 4), and this may havebeen responsible for some of the effects if, for example,people do not want to be just below the majority and ifinstead of getting ahead of others they simply want to jointhe majority. It remains to be seen whether the results wouldbe replicated if the modal group were somewhere else inthe distribution, for example, if bottom-tier consumers werealready part of the modal group instead of just below it. Weexamine these issues in study 2.

STUDY 2: EFFECTS OF EQUALITY,FOCUS ON POSSESSION GAP OR

POSITION GAIN, AND MAJORITY GROUPLOCATION ON CONSPICUOUS

CONSUMPTION

In study 2, we further tested the process behind the results ofstudy 1. First, we directed people’s attention to either the po-sition gain or the possession gap in the distribution, whichallowed us to reverse the effects of equality on conspicuousconsumption decisions. Remember that increasing equality re-duces the possession gap between the bottom tier and centraltendency measures of the distribution, while it increases theposition gains provided by conspicuous consumption. Prompt-ing people to focus on the possession gap should thereforeconfirm the prediction made in previous research that spendingwill decrease when equality is high and the possession gap isnarrow. However, prompting people to focus on the positiongain should lead to a replication of the results found (withoutprompting) in study 1—that spending will increase when equal-ity is high and there are greater position gains available fromconspicuous consumption.

The second goal of study 2 was to test the alternativeexplanation that people may simply want to join the majoritygroup. To achieve this goal, we manipulated the position ofthe majority group independently of the equality of the dis-tribution by placing it either in tier 5 (bottom) or in tier 2 ofthe distribution. If the results of study 1 were caused bypeople’s desire to join the majority and not by position gains,the equality manipulation in study 2 should have no effectbecause bottom-tier consumers are either already part of themajority group (in tier 5), and hence should not want to leaveit, or would not be able to reach the majority group (whenit is in tier 2) even after upgrading to a third-tier product. If,on the other hand, conspicuous spending is driven by positiongains or possession gaps created by equality (which is ma-nipulated independently of the location of the majority group),then the effects of equality should be independent of thelocation of the majority group.

Method

We recruited 214 people outside a large urban universitycampus and asked them to participate in a study about peo-

ple’s spending decisions for special occasions in exchangefor a meal voucher. We used a 2 # 2 # 2 between-subjectsdesign in which we manipulated distribution equality (lowvs. high), focused on the possession gap or the positiongain, and varied the location of the majority group (tier 5vs. tier 2). The participants read a scenario in which theyhad to choose a wedding gift for a classmate among productsspecified in an online wedding registry. The registry in-cluded five tiers (personal care products priced between i20and i59, tableware items at i60–i79, decorative items ati80–i99, home electronics at i100–i199, and furnitureitems at i200–i599) and allowed people to see how manyitems in the registry had already been bought by other wed-ding guests. We used different product categories, and notjust different price levels, in different tiers in order to makethe scenario more realistic for the participants. We also usedwider price ranges at upper-tier product categories than atlower-tier product categories in order to mirror real life (log-normal) price distributions of wedding gifts. The participantswere asked to imagine that they had initially decided to buya i40 bottle of perfume—a bottom-tier product for thecouple—but that, after seeing the distribution of other peo-ple’s gift choices, they were considering spending an addi-tional i50 to purchase a third-tier product (a i90 decorativevase). The scenario was therefore similar to the one used instudy 1 except that in study 2 status was derived from giftgiving rather than from spending money on oneself.

We manipulated distribution equality by increasing theproportion of people in tier 4 from 15% in the low-equalitycondition to 30% in the high-equality condition. We inde-pendently manipulated the location of the majority group(40%) by placing it in tier 5 or in tier 2. When the majoritygroup was in tier 5, the distribution across the five tiers was40%, 30%, 15%, 10%, and 5% in the high-equality condition(H p .29); in the low-equality condition it was 40%, 15%,15%, 15%, and 15% (H p .26). When the majority groupwas in tier 2, the distribution was 10%, 30%, 15%, 40%,and 5% in the high-equality condition (H p .29) and 10%,15%, 15%, 40%, and 20% in the low-equality condition (Hp .26). The average gift (computed using the midpoint ofeach tier) was less expensive in the high-equality distributionthan in the corresponding low-equality distribution (respec-tively, M p i86 vs. M p i123 when the majority was intier 5 and M p i119 vs. M p i168 when the majoritygroup was in tier 2). As a result, the possession gap waslower in the high-equality conditions than in the low-equal-ity conditions.

We manipulated the participants’ focus on the possessiongap or the position gain by asking them to study the im-plications of their decision on either their rank in the dis-tribution or the type of product that they would offer as agift. In the position gain focus condition, we asked the par-ticipants to report what percentage of guests would have asuperior or an inferior gift to their own third-tier gift. In thepossession gap focus condition, we simply asked the par-ticipants to list the pros and cons of buying a fifth- and athird-tier gift. After this manipulation, the participants in-




dicated their preference between both gifts on a 9-point scaleanchored at 1 (i40 perfume) and 9 (i90 vase). After com-pleting the questionnaire, the participants were debriefed,handed a meal voucher, and dismissed.


We conducted an ANOVA on the gift preference withdistribution equality (low vs. high), focus manipulation (pos-session gap vs. position gain), majority location (tier 5 vs.tier 2), and all interactions as fixed factors. The results re-vealed a significant interaction between distribution equalityand the focus manipulation (F(1, 206) p 8.5, p ! .005). Aspredicted, and as shown in figure 2, when focusing on theposition gain people were more likely to spend an additionali50 on a third-tier gift in the high-equality condition thanin the low-equality condition (M p 5.4 vs. M p 4.4; F(1,206) p 3.9, p ! .05), but the opposite results were obtainedwhen participants focused on the possession gap (M p 4.3vs. M p 5.4; F(1, 206) p 4.5, p ! .05). All other factorswere insignificant, including the main effect of the locationof the majority group (F(1, 206) p .2, p p .67), the maineffect of equality (F(1, 206) ! .1, p p .90), the main effectof focus (F(1, 206) ! .1, p p .91), the equality by locationinteraction (F(1, 206) p .3, p p .58), the location by focusinteraction (F(1, 206) ! .1, p p .86), and the three-wayinteraction (F(1, 206) p .4, p p .54).

The results of study 2 support our hypotheses. First, wefound that the possession gap and the position gain bothinfluence conspicuous spending decisions but that unlesspeople are prompted to focus on the possession gap, higherequality leads to higher conspicuous consumption amongpeople at the bottom of the distribution. Conversely, whenpeople were primed to focus on the possession gap, reducingequality increased the possession gap and therefore moti-vated people at the bottom of the distribution to reduce thisgap by upgrading to a more expensive gift. Taken togetherwith the results of study 1, these results suggest that equalityincreases spending because bottom-tier consumers sponta-neously consider the position gains resulting from spending.Second, study 2 demonstrated that the effects of equalityare independent of the location of the majority group in thedistribution and that equality makes people want to spendmore and surpass—rather than join—a larger percentage ofpeople. When position gains were salient, for example,equality made bottom-tier consumers spend more even ifthat meant leaving the majority group. This suggests that itis people’s desire to be better than others, and not a desireto be just like others, that drives the effect of equality onconspicuous-spending decisions.

Studies 1 and 2 both examined decisions about whetherto save money or to engage in conspicuous consumption byacquiring products that conferred social status. In the fol-lowing three studies, we further test our framework by vary-ing the importance of status and hence the position gains.We do so by manipulating the degree to which the consumedproducts enhance status (study 3) by priming social com-petition or social indifference goals (study 4A) and by using

a social context that rewards either cooperation or compe-tition (study 4B). This allows us to further test the positiongain mechanism and to examine its boundary conditions.

STUDY 3: EFFECTS OF EQUALITY ONCONSPICUOUS AND INCONSPICUOUS

CONSUMPTION

The main objective of study 3 is to examine whether theeffects of equality on consumption are moderated by itsconspicuousness. Study 3 also allows us to test another por-tion of the framework—that increasing equality should in-crease the satisfaction of low-tier consumers with their cur-rent level of possessions (vs. the social envy in study 1)due to their smaller possession gap with others. We expectincreasing equality to improve the position gains that con-sumption provides to low-tier consumers as well as theirsatisfaction with their initial level of possessions. We alsoexpect the effects of equality enhancement on position gainsand satisfaction to be similar for status-neutral and status-enhancing products. In other words, people should realizethat equality makes them happier with their lot, yet it in-creases the position gains from consumption of any type ofproduct. When status matters (conspicuous consumption),position gains motivate consumption decisions, and there-fore increasing equality should increase spending amongbottom-tier consumers, as found in studies 1 and 2. Con-versely, when consumption does not enhance status (incon-spicuous consumption), position gains do not matter, andconsumption is driven by the possession gap and satisfactionwith what one already has. Due to the smaller possessiongap and greater satisfaction with what one has in the high-equality distribution, we expect that high equality shouldreduce inconspicuous consumption among bottom-tier con-sumers, consistent with the results of the possession gapfocus condition of study 2.

Method

In study 3, we manipulated the distribution equality (lowvs. high) and the conspicuousness of the consumption (lowvs. high) between subjects with three product replications(gardens, home decoration, and clothing). We used the sameprocedure as in study 1 but with three different scenariosto check the robustness of the findings and to manipulatethe conspicuousness of the consumption in two differentways: by using different products (flower bushes vs. pinetrees in the garden scenario, the size of a TV screen vs. thesize of a decorative mirror in the home decoration scenario)and by focusing on two different attributes of the sameproduct (the brand name vs. the type of fabric of a scarf inthe clothing scenario).

In the garden scenario, participants in the high-conspic-uousness condition read the same information as in study1, describing someone with no flower bushes in her garden,who, upon reading information about the number of housesin the neighborhood with flower bushes in a local newsletter,




FIGURE 2

STUDY 2: EFFECTS OF EQUALITY AND FOCUS ON POSSESSION GAP OR POSITION GAIN ON CONSPICUOUS CONSUMPTION

was considering spending i45 to buy three bushes to beplanted in the front garden. Participants in the low-con-spicuousness condition read a similar scenario except thatit was about pine trees to be planted in the back garden.The five tiers of the distribution were the same in bothconditions (zero, one or two, three or four, five or six, andseven or more flower bushes or pine trees). In a pretest, weasked 22 people to rate how much each type of plant in-fluenced one’s social status on a 9-point scale (anchored by1 p “not at all” and 9 p “very much”). As expected, flowerbushes were seen as more status enhancing than pine trees(M p 4.1 vs. M p 3.3; t(21) p 2.2, p ! .05).

The home decoration scenario described a person whowas considering improving the appearance of her livingroom. In the high-conspicuousness condition, she was con-sidering replacing her 19-inch flat-screen television with a32-inch flat-screen television. In the low-conspicuousnesscondition, she was considering replacing a 19-inch antiquemirror with a 32-inch mirror. In both conditions, the personhad been reading a magazine article that provided infor-mation about the popularity of five different sizes of tele-visions or mirrors (19 inches, 20–31 inches, 32–39 inches,40–45 inches, and 46 inches and beyond). The pretest con-firmed that the size of a television screen had a greaterimpact on social status than the size of a mirror (M p 5.4vs. M p 3.7; t(21) p 2.8, p ! .01).

The third scenario described a college student going ona class ski trip, who owned a hand-me-down scarf. Unde-cided about whether to buy a better scarf, she had observedthe scarves worn by other students and estimated that therewere five categories. In the low-conspicuousness condition,the four categories above hand-me-downs were ranked by

fabric quality (polyester, cotton, wool, and cashmere). In thehigh-conspicuousness condition, the four other categorieswere ranked according to the prestige of their brands (H&M,Zara, Ralph Lauren, and Chanel). In both conditions, thedecision involved whether to spend i40 on a scarf in thethird tier (either a cotton scarf or a Zara scarf). Althoughboth the fabric and the brand name of a scarf were observ-able, the pretest showed that the brand of a scarf had agreater impact on social status than the quality of the fabric(M p 5.6 vs. M p 4.8; t(21) p 2.9, p ! .01).

As in study 1, we manipulated the equality of the distri-bution across the five tiers using the distributions shown infigure 1: 10%, 40%, 20%, 20%, and 10% in the high-equalitycondition (H p .26) and 10%, 20%, 20%, 25%, and 25% inthe low-equality condition (H p .22). Just like in study 1increasing equality reduced the possession gap but increasedthe position gains obtained by upgrading to a tier 3 product.Participants were then asked to predict whether the bottom-tier person described in the scenario would choose to spendthe money or save it (anchored from 1 p “definitely save”to 9 p “definitely spend”). They also indicated how buyingthe product would increase this person’s position in the dis-tribution described in the scenario (from 1 p “not at all” to9 p “very much”). Finally, they rated how satisfied theybelieved this person was with her initial position in the fifthtier (from 1 p “very unsatisfied” to 9 p “very satisfied”).The order of the three questions was counterbalanced acrossparticipants.

Results and DiscussionWe begin with the analyses of the perceived position gain

and satisfaction data, which allow us to check the effec-




FIGURE 3

STUDY 3: EFFECTS OF EQUALITY ON CONSPICUOUS AND INCONSPICUOUS CONSUMPTION

tiveness of the manipulations. Because there were no dif-ferences across the three scenarios, we pooled the data acrossthe three replications and obtained a total of 153 responses.For perceived position gains, only the main effect of dis-tribution equality was statistically significant (F(1, 149) p29.1, p ! .01), while the effects of conspicuousness and itsinteraction with equality were not (F(1, 149) p .1, p p.71 and F(1, 149) p .2, p p .65, respectively). As expected,people realized that consumption (conspicuous or not) al-lowed them to get ahead of more people when equality washigh (M p 6.2) than when it was low (M p 3.8). Similarly,only the main effect of distribution equality was significantfor satisfaction (F(1, 149) p 10.7, p ! .01), and the effectsof conspicuousness and its interaction with equality werenot statistically significant (F(1, 149) p .04, p p .85 andF(1, 149) p .2, p p .66, respectively). People indicatedthat the fifth-tier person would be more satisfied with herinitial position when equality was high (M p 3.8) than whenit was low (M p 2.8), regardless of the conspicuousness ofthe consumption.

Turning now to the dependent variable, we found that,consistent with our predictions, the main effects of distributionequality and conspicuousness were not statistically significant(respectively, F(1, 149) p .1, p p .75 and F(1, 149) p 1.1,p p .29), but their interaction was significant (F(1, 149) p9.4, p ! .01). As shown in figure 3, increasing equality boostedspending for status-enhancing products but decreased spend-ing for status-neutral products. Contrast tests further showedthat conspicuous consumption was higher in the high-equalitycondition (M p 7.3) than in the low-equality condition (Mp 6.4; F(1, 149) p 3.8, p ! .05), which replicated the find-ings of study 1 and the result of study 2 in the position gainfocus condition. For inconspicuous consumption, however,

spending was lower in the high-equality condition (M p 6.0)than in the low-equality condition (M p 7.1; F(1, 149) p5.8, p ! .05), consistent with the results of the possession gapcondition of study 2.

Study 3 showed that the effect of equality on the pref-erence for spending over saving among low-tier consumersdiffers for conspicuous and inconspicuous consumption.First, it replicated in two new scenarios the main result—that increasing equality encourages bottom-tier consumersto spend on conspicuous consumption because it allowsthem to get ahead of more people. More importantly, study3 showed that increasing equality reduces spending on in-conspicuous consumption because it increases people’s sat-isfaction with their current level of possessions. Even thoughbuying status-neutral products provides the same improve-ment in position as buying status-enhancing products, po-sition gains matter less than the possession gap for productsthat do not contribute to status. Study 3 therefore shows aboundary condition for the effect of increasing equality onspending versus saving decisions.

An interesting question that arises from the results ofstudies 1–3 is whether these results would carry over to amore general setting in which endowment is more broadlyconstrued and less directly observable (e.g., income levelvs. the size of one’s TV screen). This is important becausein much of the research on status and on conspicuous con-sumption, status is determined more broadly by people’srelative wealth or income (Chao and Schor 1998; Duesen-berry 1949). In addition, it is important to further test thehypothesized moderating role of the importance of statusseeking found in study 3 by directly priming status-seekinggoals rather than by using different product categories. Fi-nally, studies 1–3 examined the trade-offs that people make




between consumption and savings. It remains to be seenwhether the effects found in these studies also apply to trade-offs between conspicuous and inconspicuous consumption.We examine these issues in the following two studies.

STUDIES 4A AND 4B: EFFECTS OFINCOME EQUALITY AND SOCIAL

COMPETITION GOALS ONPREFERENCES FOR CONSPICUOUS

VERSUS INCONSPICUOUSCONSUMPTION

In studies 4A and 4B, we manipulated the equality of incomedistribution to examine whether it influences the trade-offsbetween conspicuous and inconspicuous consumption. Wealso tested the moderating role of status importance in twodifferent ways—either by priming social competition or so-cial indifference goals (study 4A) or by varying the degreeof competitiveness of the social group (study 4B). Usingthese two different mechanisms allows us to test the ro-bustness of the moderating effect of status goal importancewith manipulations exhibiting either a higher degree of in-ternal validity (unconscious priming) or a higher degree ofexternal validity (scenarios about different social groups).We expect the results to be replicated regardless of themethod used to manipulate status goal importance.

Research has shown that activation of competition-relatedconcepts leads to more competitive behavior (Kawada et al.2004). For example, Griskevicius et al. (2007) showed thatactivating mating motives leads men to engage more inconspicuous consumption. Similarly, Rucker and Galinsky(2008) showed that people compensate for feelings of socialpowerlessness with a higher willingness to pay for statusgoods. Other studies have shown that people seek to over-take others and to maximize position gains more when thesocial group is competitive than when it is collaborative(Frank 1985b; Frank and Cook 1995). Therefore, activatingsocial competition goals or a competitive social environmentshould motivate people to maximize position gains. Sincechoosing a status-enhancing option can help one achievethis goal more successfully when equality is high (and po-sition gains are higher), we expect increasing equality toincrease the preference for conspicuous over inconspicuousconsumption when social competition goals have been ac-tivated (study 4A) or when the group is competitive (study4B). Conversely, activating social indifference and groupassimilation goals should encourage people to focus on per-sonal satisfaction and to avoid position gains that wouldexacerbate differences with the rest of the group when peo-ple seek to avoid them. Because lowest-tier individuals aremore satisfied with their lot and would gain larger unwantedposition gains through consumption when equality is high(as demonstrated in study 3), we expect increasing equalityto reduce the preference for conspicuous over inconspicuousconsumption when social indifference goals have been ac-tivated (study 4A) or in a collaborative group (study 4B).

Method

Both studies used a 2 # 2 between-subjects design withthe equality of the income distribution (low vs. high) andgoal prime (study 4A: social competition vs. social indif-ference; study 4B: competitive vs. collaborative group) asfixed factors. We recruited people near a large urban uni-versity campus to fill out a questionnaire in exchange for avoucher for a movie ticket. We obtained 69 usable responsesin study 4A and 70 in study 4B.

In study 4A and in the competitive scenario of study 4B,participants were asked to imagine that the human resourcesdepartment of the firm at which they had been working for 3years prepared a report on the salary distribution of people whohad been hired at the same time as they had. There were ninetiers of net after-tax monthly income (less than i1,500, i1,500–i1,999, i2,000–i2,499, i2,500–i2,999, i3,000–i3,499,i3,500–i3,999, i4,000–i4,499, i4,500–i4,999, and i5,000 ormore). The distribution was 5%, 15%, 35%, 15%, 10%, 5%,5%, 5%, and 5% in the high-equality condition and 5%, 15%,15%, 15%, 15%, 10%, 10%, 10%, and 5% in the low-equalitycondition. The GINI index (computed using the midpoint ofeach tier and i6,500 as the upper bound) was lower in thehigh-equality distribution (21.6%) than in the low-equality dis-tribution (22.1%). The equality manipulation therefore repro-duced the effects of a mild progressive income tax that reducedthe number of people in the top income tiers and increased thenumber in the middle tiers. The mean and the median salarieswere lower in the high-equality condition (respectively, M pi2,750 and median p i2,499) than in the low-equality con-dition (respectively, M p i3,100 and median p i2,999);hence, the possession gap was lower in the high-equality con-dition than in the low-equality condition.

Participants were asked to imagine that their net after-taxmonthly income was i1,900, which placed them in theeighth tier (20th percentile) of the distribution, just behind35% of people in tier 7 in the high-equality condition orbehind only 15% of people in tier 7 in the low-equalitycondition. Participants in study 4A and in the competitivescenario of study 4B were then asked to imagine that theywere planning to meet for dinner with a coworker and hadto indicate their preference between a trendy new Asianrestaurant with a i45 fixed-price menu (status-enhancingoption) and a traditional bistro with a i15 fixed-price menu(status-neutral option) on a 9-point scale item anchored at1 for “traditional bistro” and 9 for “trendy restaurant.” Apretest showed that the trendy Asian restaurant would havea stronger impact on social status than the bistro (M p 6.7vs. M p 4.2; t(21) p 5.4, p ! .01). After completing thequestionnaire, the participants were debriefed, handed avoucher, and dismissed.

Status importance was manipulated before participantsread the scenario with the income equality manipulation. Instudy 4A, participants completed a scrambled sentence task(Bargh and Chartrand 2000) presented as a verbal aptitudetest. In the social competition prime condition, the partici-pants created four five-word sentences that highlighted theimportance of social competition and comparisons (e.g.,




“success is a relative concept”). In the social indifferenceprime condition, the sentences highlighted the importanceof following one’s own preference and ignoring others’ (e.g.,“true happiness comes from within”). Both conditions alsoincluded three neutral sentences. In study 4B, there was noscrambled sentence. Instead, we relied on existing researchshowing that people are more likely to engage in cooperativebehaviors when primed with a “friend” concept than whenprimed with a “coworker” concept (Fitzsimons and Bargh2003). The competitive condition therefore used the samescenario as in study 4A, describing the income distributionof coworkers. In the cooperative condition, however, theincome distribution information was provided in a com-munity newsletter and applied to friends in the communityand participants were asked to imagine that they would meetfor dinner with community friends.


We first tested the effectiveness of the two status manip-ulations. In the first pretest, 47 people were primed with asocial competition or a social indifference goal using thesame scrambled sentence task as in study 4A. Afterwardthey engaged in a word-search task—a well-establishedmeasure of goal activation (Bargh and Chartrand 2000;Bargh et al. 2001). The word pool contained eight wordsrelated to social competition (e.g., “domination”), eightwords related to indifference (e.g., “independence”), andeight neutral words. In the social competition prime con-dition participants found more words related to competitionthan to indifference (M p 2.8 vs. M p 1.2; t(20) p 2.2,p ! .05), but the opposite was true in the social indifferenceprime condition, where people found more words related toindifference than to competition (M p 2.6 vs. M p 1.7;t(25) p 5.7, p ! .01). In the second pretest, we asked 53participants how important they felt it was to “do better thanothers” or to “assimilate to others” when these others arecoworkers versus community friends (1 p “not at all” and9 p “very much”). The results showed that it was moreimportant to do better than others among rivals (M p 5.5)than among friends (M p 3.6, t(52) p 6.2, p ! .001) butthat it was more important to assimilate to others amongfriends (M p 7.6) than among rivals (M p 3.7; t(52) p9.8, p ! .001).

To analyze the effects on the dependent variable in study4A, we conducted an ANOVA with income equality (lowvs. high), goal prime (social competition vs. social indif-ference), and their interaction as fixed factors. The two maineffects were not significant (F(1, 65) p .3, p p .62 forequality and F(1, 65) p 2.8, p p .10 for goal), but theirinteraction was statistically significant (F(1, 65) p 11.3, p! .01). As shown in figure 4A, in the social competitionprime condition people were more likely to choose thestatus-enhancing trendy restaurant when equality was high(M p 6.0) than when it was low (M p 4.4; F(1, 65) p4.2, p ! .05). The opposite was observed in the social in-difference prime condition: people were less likely to choose

the trendy restaurant when equality was high (M p 3.1)than when it was low (M p 5.3; F(1, 65) p 7.3, p ! .01).

We used the same procedure to analyze the data fromstudy 4B and conducted an ANOVA with income equality(low vs. high), social context (competitive vs. cooperative),and their interaction as fixed factors. As in study 4A, thetwo main effects were not significant (F(1, 66) ! .1, p p.98 for income equality and F(1, 66) p .9, p p .36 forgroup competitiveness), but their interaction was statisticallysignificant (F(1, 66) p 8.5, p ! .01). As predicted and shownin figure 4B, participants in the competitive condition weremore likely to choose conspicuous dining when incomeequality was high (M p 5.4) than when it was low (M p3.6; F(1, 66) p 4.3, p ! .05). Conversely, in the cooperativecondition participants were less likely to choose conspicuousdining when equality was high (M p 3.0) than when it waslow (M p 4.8; F(1, 66) p 4.2, p ! .05).

Overall, studies 4A and 4B show that the effect of equalityon the preferences of low-income consumers for conspicuousover inconspicuous consumption depends on how much theycare about status. When people care about status, either be-cause competition goals have been primed or because theyare among competitors, we replicate the findings of studies1–3. Again, we find that equality fuels conspicuous con-sumption among bottom-tier consumers even when equalityapplies to general income level rather than ownership of par-ticular possessions and when the decision involves choosingbetween different consumption options (vs. choosing whetherto spend at all or not). However, when people do not careabout status, either because they have been primed to beindifferent to others or because they are in a collaborativesocial group, equality reduces low-income consumers’ pref-erence for conspicuous consumption. Finally, the results ofstudies 4A and 4B reinforce the conclusion of the other studiesthat increasing equality does not necessarily reduce conspic-uous consumption by people at the bottom of the distributionwhen social rank matters. They also suggest that redistributionpolicies may be effective when the social environment down-plays competition for status and promotes cooperation.

GENERAL DISCUSSION

The objective of this research was to test in an experimentalsetting the oft-repeated prediction that increasing equality ofincome or possession would decrease conspicuous and in-conspicuous consumption among consumers at the bottom ofthe distribution. Our main conclusion is that increasing equal-ity does indeed reduce inconspicuous (status-neutral) con-sumption and conspicuous (status-relevant) consumption forpeople at the bottom of the distribution when they do notcare about status, for example, in a cooperative social context.However, we find that increasing equality actually fuels con-spicuous consumption when people at the bottom of the dis-tribution care about their social position. This is becausegreater equality increases the percentage of people in themiddle of the distribution and therefore increases the gain insocial position, and hence status, that a given spending on




FIGURE 4

STUDY 4A (A) AND STUDY 4B (B): EFFECTS OF INCOME EQUALITY ANDSTATUS IMPORTANCE ON PREFERENCE FOR CONSPICUOUS CONSUMPTION

conspicuous consumption offers to consumers at the base ofthe pyramid.

We find that these effects are robust regardless of whetherstatus is unobservable and broadly construed (as income)or observable and more narrowly construed (as endowmentwith specific status-conferring possessions), whether statusis derived from publicly spending on the self or acquiredthrough gift giving, and whether the decision is a trade-offbetween spending and saving or between spending on astatus-enhancing or a status-neutral option.

Implications for Policy Makers and Marketers

Our results suggest innovative ways for marketers and pol-icy makers to influence conspicuous consumption decisions.First, they show that we cannot simply assume that increasingequality will reduce consumption and that marketers and pol-icy makers should build a more holistic view of redistributionpolicies and their consequences. Specifically, our results sug-gest that the implications of redistribution policies need to bereconsidered for different social environments. For example,




we find that increasing income equality succeeds in reducingconspicuous consumption in cooperative environments andwhen people are indifferent to the social context. This sug-gests that redistribution policies may be particularly effectiveif supplemented with policies to promote resistance to socialpressure, which focus on relationships with friends and fam-ily. Echoing Putnam (2007), the promotion of a broad senseof “we” through popular culture, national symbols, education,and common experiences may not only increase trust butcould also reduce conspicuous arms races.

In this article we have examined the effects of redistributionpolicies that reduce the number of people in the upper tiersand increase the number of people in the middle tiers of thedistribution while keeping the number in the lowest tier con-stant. This was done to rule out the alternative explanationthat equality effects may be driven by changes in the endow-ment of low-tier consumers, by changes in the range of thedistribution, or by changes in the price and availability ofproducts in each tier. It would be interesting to examine whatwould happen if equality were increased through redistribu-tion policies that take from the rich to give to the poor andhence increase the endowment of people in the bottom tiers.It would also be interesting to examine the net effects ofdifferent types of redistribution policies, because they oftendo not just change the distribution of the population acrosstiers but also change product prices and availability as wellas consumers’ wealth and access to credit (e.g., interest rateregulations). Since redistribution policies may affect consum-ers’ spending and welfare in many ways—directly by chang-ing possession gaps and position gains and indirectly by fund-ing new programs using the new tax revenues (e.g., education,health care)—it would be important to study the net effectsof redistribution policies on the decisions and welfare of bot-tom-tier consumers.

Our research also has important implications for market-ers. First, marketers could enhance the perceived status ben-efits of their products or services by highlighting not justtheir exclusivity but their status improvement benefits. Forexample, companies could provide consumers with infor-mation about their rank or percentile and the proportion ofconsumers at each status level (e.g., “You are among the30% of our customers with gold status and with three extraflights you would join the top 10% of our customers in theelite platinum tier”). Marketers could also take into accountthe degree of equality in their customer base when makingpricing decisions. For example, they could charge more fordeluxe product variants in markets with a more homoge-neous consumer base and a competitive social environment.Instead of using a traditional pyramidal structure with alinear reduction in the number of people in each consecutivestatus tier, they could structure their loyalty programs tomatch the most profitable upgrades with the highest positiongains.

Our results also advance understanding of the intensityof competition among people and organizations with similarperformance levels. Lehmann (2001) argued that one reasonwhy market shares are exponentially distributed may be that

managers care about market share ranks and compete moreintensely when the gap between their market share and thatof their closest competitor is small; hence, ranks are lesssusceptible to change when the gap is large and more dif-ficult to bridge. Our results suggest that Lehmann’s resultscould be generalized by looking at the equality of the marketshare distribution rather than simply the proximity of theclosest competitor. Following the results of Leclerc et al.(2005), who found that the importance of status diminisheswhen true product quality is easy to evaluate, these effectsshould be particularly strong when only ordinal informationis readily available.

Implications for Consumer Research on Status

The existing literature on status in consumer research hastended to focus on social envy and dissatisfaction with one’scurrent endowment, two negative and backward-looking fac-tors. Our research aims to contribute to this literature by high-lighting the importance of social position changes as a positiveand forward-looking antecedent of conspicuous consumptiondecisions. Our findings also have implications beyond statusresearch. For example, they suggest that range-frequency the-ory (Parducci 1965) should incorporate people’s expectationsabout changes in percentile position, and not just the initialpercentile position, when evaluating their current position andthe actions that would shift them from the initial to the finalposition. They also suggest that optimal distinctiveness theory(Brewer 1991) may take into account the fact that differen-tiation is influenced not just by the size of the group to whichone belongs but also by the size of the group of people inthe social hierarchy that could be surpassed. In addition, ourfinding about the moderating role of status gain may help inunderstanding why people sometimes engage in conspicuousconsumption to differentiate themselves from their peers—the snob effect—while at other times they do so to affiliatewith their peers—the bandwagon effect (Amaldoss and Jain2005; Berger and Heath 2007; Leibenstein 1950). Our resultsmay suggest that the desire to maximize position gains maybe linked to downward social comparisons and dissociativemotives (Han, Nunes, and Dreze 2010; White and Dahl 2006,2007), just like the desire to minimize possession gaps islinked to upward social comparisons and associative motives(Escalas and Bettman 2003). It would be interesting to ex-amine these links in the future.

Our findings also have direct implications for understand-ing status perceptions. Dreze and Nunes (2009) showed thatadding a lower tier in a customer loyalty program increasesthe perceived status of people in top tiers. Our study extendstheir work by examining the effects of the distribution ofpeople in each tier (vs. the number of tiers), by looking atpeople in the bottom tiers (vs. just the top tiers), and byexamining effects on consumption (vs. status perceptions).An important difference between our work and existingwork on status (including the work by Dreze and Nunes) isthat the position gain hypothesis is independent of the statuslevel and the magnitude of the upgrade (i.e., the number oftiers surpassed). It makes the same prediction regardless of




status level and magnitude of the upgrade, as long as thepercentage of people that can be surpassed is identical. Forexample, in figure 1 our theory would predict that peoplein the middle tier (tier 3) would be less motivated to spendto reach the top tier in the high-equality distribution becauseit provides smaller position gains than the low-equality dis-tribution. Still, it would be interesting to test whether beingpositioned at the extreme ends of the distribution leads tospecific behaviors, either diminishing or increasing sensi-tivity to status change. More generally, it would be inter-esting to extend our work to examine the effects of statuschange and not just the effects of change in the distributionof status. For example, positive and negative changes instatus may have asymmetric effects and their effects maybe different if they are driven by changes in one’s income(other people’s income remaining constant) or by changesin other people’s income (one’s income remaining constant).These two issues are particularly relevant given the recenttransition from a long period of economic expansion to aperiod of economic recession and overall income stagnation.

Finally, it would be useful for future research to test thecentral assumption of inequality research that people havean accurate perception of the status distribution within theirsocial group. This is important because the generalizabilityof our findings rests on the assumption that people have, ifnot an accurate representation of the endowment distribu-tion, at least the ability to notice changes in the equality ofthis distribution as we define them (i.e., more people in themiddle and fewer people at the extremes). Although pre-vious studies have shown that people have a sophisticatedand fine-grained understanding of social categories withintheir social groups (Kraus et al. 1993), they may be subjectto systematic biases depending on their position in the dis-tribution, and we do not know how accurately they perceivethe equality of a distribution nor what measure of centraltendency they use most frequently. More generally, it wouldbe interesting to examine whether it is the actual or perceivedposition or the actual and perceived level of endowment thatbest predicts saving and spending decisions.

REFERENCES

Alpizar, Francisco, Fredrik Carlsson, and Olof Johansson-Stenman(2005), “How Much Do We Care about Absolute versus Rel-ative Income and Consumption?” Journal of Economic Be-havior and Organization, 56 (3), 405–21.

Amaldoss, Wilfred and Sanjay Jain (2005), “Conspicuous Con-sumption and Sophisticated Thinking,” Management Science,51 (October), 1449–66.

Bagwell, Laurie Simon and B. Douglas Bernheim (1996), “VeblenEffects in a Theory of Conspicuous Consumption,” AmericanEconomic Review, 86 (June), 349–73.

Bargh, John A. and Tanya L. Chartrand (2000), “The Mind in theMiddle: A Practical Guide to Priming and Automaticity Re-search,” in Handbook of Research Methods in Social andPersonality Psychology, ed. Harry T. Reis and Charles M.Judd, New York: Cambridge University Press, 253–85.

Bargh, John A., Peter M. Gollwitzer, Annette Lee-Chai, KimberlyBarndollar, and Roman Trotschel (2001), “The Automated

Will: Nonconscious Activation and Pursuit of BehavioralGoals,” Journal of Personality and Social Psychology, 81 (6),1014–27.

Becker, Gary S., Kevin M. Murphy, and Ivan Werning (2005),“The Equilibrium Distribution of Income and the Market forStatus,” Journal of Political Economy, 113 (April), 282–310.

Berger, Jonah and Chip Heath (2007), “Where Consumers Divergefrom Others: Identity Signaling and Product Domains,” Jour-nal of Consumer Research, 34 (August), 121–34.

Bloch, Francis, Vijayendra Rao, and Sonalde Desai (2004), “Wed-ding Celebrations as Conspicuous Consumption,” Journal ofHuman Resources, 39 (3), 675–95.

Brewer, Marilynn B. (1991), “The Social Self: On Being the Sameand Different at the Same Time,” Personality and Social Psy-chology Bulletin, 17 (5), 475–82.

Chao, Angela and Juliet B. Schor (1998), “Empirical Tests of StatusConsumption: Evidence from Women’s Cosmetics,” Journalof Economic Psychology, 19 (February), 107–31.

Christen, Markus and Ruskin M. Morgan (2005), “Keeping upwith the Joneses: Analyzing the Effect of Income Inequalityon Consumer Borrowing,” Quantitative Marketing and Eco-nomics, 3 ( June), 145–73.

Clark, Andrew E. and Andrew J. Oswald (1998), “Comparison-Concave Utility and Following Behavior in Social and Eco-nomic Settings,” Journal of Public Economics, 70 (October),133–55.

Collins, Rebecca L. (1996), “For Better or Worse: The Impact ofUpward Social Comparison on Self-Evaluations,” Psycholog-ical Bulletin, 119 (January), 51–69.

Dreze, Xavier and Joseph C. Nunes (2009), “Feeling Superior: TheImpact of Loyalty Program Structure on Consumers’ Percep-tions of Status,” Journal of Consumer Research, 35 (April),890–905.

Duesenberry, James S. (1949), Income, Saving and the Theory ofConsumer Behavior, Cambridge, MA: Harvard UniversityPress.

Duncan, Denvil R. and Klara Sabirianova Peter (2008), “Tax Pro-gressivity and Income Inequality,” Working Paper No. 08–26,Andrew Young School of Policy Studies Research, GeorgiaState University, Atlanta, GA 30303.

Dupor, Bill and Wen-Fang Liu (2003), “Jealousy and EquilibriumOverconsumption,” American Economic Review, 93 (March),423–28.

Dynan, Karen E., Jonathan Skinner, and Stephen P. Zeldes (2004),“Do the Rich Save More?” Journal of Political Economy, 112(April), 397–444.

Elster, Jon (1991), “Envy in Social Life,” in Strategy and Choice,ed. Richard J. Zeckhauser, Cambridge, MA: MIT Press,49–82.

Escalas, Jennifer Edson and James R. Bettman (2003), “You AreWhat You Eat: The Influence of Reference Groups on Con-sumers’ Connections to Brands,” Journal of Consumer Psy-chology, 13 (3), 339–48.

Feinberg, Fred M., Aradhna Krishna, and Z. John Zhang (2002),“Do We Care What Others Get? A Behaviorist Approach toTargeted Promotions,” Journal of Marketing Research, 39(August), 277–91.

Festinger, Leon (1954), “A Theory of Social Comparison Pro-cesses,” Human Relations, 7 (May), 117–40.

Fitzsimons, Grainne M. and John A. Bargh (2003), “Thinking ofYou: Nonconscious Pursuit of Interpersonal Goals Associatedwith Relationship Partners,” Journal of Personality and SocialPsychology, 84 (January), 148–64.




Frank, Robert H. (1985a), Choosing the Right Pond: Human Be-havior and the Quest for Status, New York: Oxford UniversityPress.

——— (1985b), “The Demand for Unobservable and Other Non-positional Goods,” American Economic Review, 75 (March),101–16.

——— (1989), “Frames of Reference and the Quality of Life,”American Economic Review, 79 (May), 80–85.

——— (2007), Falling Behind: How Rising Inequality Harms theMiddle Class, Berkeley: University of California Press.

Frank, Robert H. and Philip J. Cook (1995), The Winner-Take-AllSociety: Why the Few at the Top Get So Much More than theRest of Us, New York: Penguin Books.

Griskevicius, Vladas, Joshua M. Tybur, Jill M. Sundie, Robert B.Cialdini, Geoffrey F. Miller, and Douglas T. Kenrick (2007),“Blatant Benevolence and Conspicuous Consumption: WhenRomantic Motives Elicit Strategic Costly Signals,” Journalof Personality and Social Psychology, 93 ( July), 85–102.

Han, Young Jee, Joseph C. Nunes, and Xavier Dreze (2010), “Sig-naling Status with Luxury Goods: The Role of Brand Prom-inence,” Journal of Marketing, 74 ( July), 15–30.

Hopkins, Ed and Tatiana Kornienko (2004), “Running to Keep inthe Same Place: Consumer Choice as a Game of Status,”American Economic Review, 94 (September), 1085–1107.

Hsee, Christopher K., Yuval Rottenstreich, and Zhixing Xiao(2005), “When Is More Better? On the Relationship betweenMagnitude and Subjective Value,” Current Trends in Psy-chological Science, 14 (5), 234–37.

Hyder, Zakia (1991), “Boating Titles Swamped by Luxury Tax,”Folio: The Magazine for Magazine Management, September1.

Kawada, Christie L. K., Gabriele Oettingen, Peter M. Gollwitzer,and John A. Bargh (2004), “The Projection of Implicit andExplicit Goals,” Journal of Personality and Social Psychol-ogy, 86 (April), 545–59.

Knell, Markus (1999), “Social Comparisons, Inequality, andGrowth,” Journal of Institutional and Theoretical Economics,155 (December), 664–95.

Kraus, Susan, Carey S. Ryan, Charles M. Judd, and Reid Hastie(1993), “Use of Mental Frequency Distributions to RepresentVariability among Members of Social Categories,” SocialCognition, 11 (1), 22–43.

Leclerc, France, Christopher K. Hsee, and Joseph C. Nunes (2005),“Narrow Focusing: Why the Relative Position of a Good inIts Category Matters More than It Should,” Marketing Sci-ence, 24 (2), 194–205.

Lehmann, Donald R. (2001), “The Impact of Altruism and Envyon Competitive Behavior and Satisfaction,” InternationalJournal of Research in Marketing, 18 ( June), 5–17.

Leibenstein, Harvey (1950), “Bandwagon, Snob, and Veblen Ef-fects in the Theory of Consumers’ Demand,” Quarterly Jour-nal of Economics, 64 (May), 183–207.

Mandel, Naomi, Petia K. Petrova, and Robert B. Cialdini (2006),“Images of Success and the Preference for Luxury Brands,”Journal of Consumer Psychology, 16 (1), 57–69.

McCormick, Ken (1983), “Duesenberry and Veblen: The Dem-onstration Effect Revisited,” Journal of Economic Issues, 17(December), 1125–29.

Miceli, Maria and Cristiano Castelfranchi (2007), “The EnviousMind,” Cognition and Emotion, 21 (3), 449–79.

Moav, Omer and Zvika Neeman (2008), “Conspicuous Consump-tion, Human Capital, and Poverty,” Discussion Paper No.6864, Centre for Economic Policy Research, London EC1V7RR, UK.

Nunes, Paul F., Brian A. Johnson, and R. Timothy S. Breene(2004), “Selling to the Moneyed Masses,” Harvard BusinessReview, 82 (7–8), 94–104.

Parducci, Allen (1965), “Category Judgment: A Range-FrequencyModel,” Psychological Review, 72 (November), 407–18.

Putnam, Robert D. (2007), “E Pluribus Unum: Diversity and Com-munity in the Twenty-First Century: The 2006 Johan SkyttePrize Lecture,” Scandinavian Political Studies, 30 ( June),137–74.

Richins, Marsha L. (1994), “Valuing Things: The Public and Pri-vate Meanings of Possessions,” Journal of Consumer Re-search, 21 (December), 504–21.

Rucker, Derek D. and Adam D. Galinsky (2008), “Desire to Ac-quire: Powerlessness and Compensatory Consumption,” Jour-nal of Consumer Research, 35 (August), 257–67.

Smith, Richard H. and Sung Hee Kim (2007), “ComprehendingEnvy,” Psychological Bulletin, 133 (1), 46–64.

Solnick, Sara J. and David Hemenway (1998), “Is More AlwaysBetter? A Survey on Positional Concerns,” Journal of Eco-nomic Behavior and Organization, 37 (November), 373–83.

Veblen, Thorstein (1899), The Theory of the Leisure Class, NewYork: Penguin.

White, Katherine and Darren W. Dahl (2006), “To Be or Not Be?The Influence of Dissociative Reference Groups on ConsumerPreferences,” Journal of Consumer Psychology, 16 (4), 404–14.

——— (2007), “Are All Out-Groups Created Equal? ConsumerIdentity and Dissociative Influence,” Journal of ConsumerResearch, 34 (December), 525–36.

Wood, Joanne V. (1989), “Theory and Research Concerning SocialComparisons of Personal Attributes,” Psychological Bulletin,106 (September), 231–48.

Zhu, Ning (forthcoming), “Household Consumption and PersonalBankruptcy,” Journal of Legal Studies.



Documents

Getting Ahead of the Joneses: When Equality Increases Conspicuous Consumption among Bottom-Tier Consumers