ebook img

DIRECTED ALTRUISM AND ENFORCED RECIPROCITY IN SOCIAL NETWORKS* Stephen PDF

37 Pages·2015·2.39 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview DIRECTED ALTRUISM AND ENFORCED RECIPROCITY IN SOCIAL NETWORKS* Stephen

DIRECTED ALTRUISM AND ENFORCED RECIPROCITY IN SOCIAL NETWORKS* Stephen Leider Markus M. Möbius TanyaR osenblat Quoc-AnhD o We conductedo nlinef ielde xperimentisn larger eal-worldso cial networkisn ordert o decomposep rosocialg ivingi ntot hreec omponents(1: ) baselinea ltruism towardr andomlyse lecteds trangers(,2 ) directedal truismt hatf avorsf riendos ver randoms trangersa,n d (3) givingm otivatedb yt hep rospecto ff uturein teraction. Directeda ltruismin creasesg ivingt o friendbs y5 2% relativet o randoms trangers, whereasf uturein teractioenf fectisn creaseg ivingb ya n additional2 4% wheng iv- ingi s sociallye fficienTt.h isf indingsu ggestst hatf uturein teractionaf fectgs iving througha repeatedg ame mechanismw herea gentsc an be rewardedf org rant- ing efficiency-enhancfinavgo rsW. e also findt hat subjectsw ithh igherb aseline altruismh ave friendws ithh igherb aselinea ltruism. I. Introduction Real worlds ocial networksp rovidea natural laboratoryt o studyp rosocialb ehaviorF. riendsh elp each otherf requentlyan d oftens ubstantiallyI.n the 1995 General Social Survey,5 5% of Americansr eportedt hat theyf irsta pproachedc lose friendsa nd familym emberws hent heyn eededt ob orrowa larges umo fm oney (Mobius and Szeidl 2007). In a 2007 UK YouGovs urvey4, 8% of respondentrs eportedle nding,o n average,$ 1,800 to friendsa nd relativesd uringt hep ast twelvem onths(Y ouGov2 007). Similarly, closef riendsa nd relativesa re thep redominansto urceo fi nformal insurancea gainst risk in developingc ountries( Townsend1 994; Udry1 994). *Thisp aperr eplacesa n earlierw orkingp aperv ersione ntitled" SocialC apital in Social Networks.W" e are gratefutlo Ed Glaeser,M urielN iederleA, damS zeidl, and seminarp articipantast the2 004 EconometriSco cietyS ummerm eetingst,h e 2004 SITE, the Max Planck InstituteR ingbergW orkshopo n Social Networks, the2 006 NorthA mericanE SA meetingsB, ostonU niversityN,e wY orkU niversity, theU niversityof C onnecticutT,e xas A&M, ColumbiaU niversityth, eI nstitutef or AdvancedS tudy,a nd the Universityo f Michiganf orh elpfulc ommentsW. e are particularliyn debtedt o RachelC rosona nd AI Rothf ore xtremelyh elpfucl onver- sationsd uringt he designs tage of our experimentsP.a ul Niehaus and Raphael Schoenlew ere excellentr esearcha ssistants.W e thankE lizabeth MurryT, yler Williamsa, nd Kate Amblerf orv eryc arefulp roofreadinagn d numerouss ugges- tions.R osenblatt hankst heF ederalR eserveB ank ofB ostonf ori ts hospitalityW. e are gratefutlo the Social SecurityA dministratioann d the Centerf orR etirement Researcha t BostonC ollegef orf inanciaslu pportt hrougha SandellG rant( Möbius, Rosenblata) nd to theN ationalS cienceF oundation(L eider). © 2009 byt heP residenta nd Fellowso fH arvardC ollegea nd theM assachusettsIn stituteo f Technology. TheQ uarterlJy ournaol fE conomicsN, ovember2 009 1815 1816 QUARTERLYJ OURNALO F ECONOMICS Economistsh ave exploredt wom aine xplanationso fp rosocial behavior- selflessa ltruisma nd expectationso fr eciprocafl avors in futuree xchanges.B ecause people tend to interactf requently withf riendst owardw homt heyh ave the strongestal truisticf eel- ings,t hese mechanismsa re difficultto distinguishe mpirically.1 In thisp aper,w e use twoo nlinef ielde xperimentisn a real world social networkt o solvet hisi dentificatiopnr oblemO. ur designa l- lowsu s to selectivelysw itcho fft her eciprocitmy echanismin some treatmentsa,n d therebys eparatelym easuret he strengtho ft he altruisma nd reciprocitmy echanisms. The abilityo fo ur methodologtyo distinguishb etweent hese twom echanismsh as severala pplicationsF. or example,m anyr e- searchersh ave documentedth ata s societiest ransitiont o market economies,m arkett ransactionsr eplace bilateral reciprocale x- changes,l eadingt o a declinei n social capital.2B y usingo ur diag- nosticg ames to identifycu lturesw heret he social ties are either heavilyr eciprocity-baseodr heavilya ltruism-basede,c onomists may be bettera ble to explain differenceisn the social disrup- tion caused by markets.A dditionallyo,u r finding(d escribedb e- low) that the reciprocalm echanismo nlya pplies to transactions thata re efficienstu ggestsa policyu se ofo urt echniquesD. evelop- mentp rogramssu cha s microfinancteh atu se socialt iest o encour- age lendinga nd insurancem ayb e moste ffectivien communities wheret he social networksh ave a strongr eciprocityco mponent, because relationshipsb ased on tradinge fficiency-enhancifnag- vors may help directr esourcest owarde fficienuts es. Thus, pol- icymakersc ould use experimentssu ch as ours to identifyar eas wherei nterventiomn ayb e moste ffective. In our experimentsw, e distinguisht hree componentso f prosocialg iving:( 1) baselinea ltruismt owardr andomlys elected strangers(,2 ) directeda ltruismt hat favorsf riendso verr andom strangers,a nd (3) givingm otivatedb y the prospecto f future interactionW. e beginb y directlym easuringt he social networks ofH arvardu ndergraduatets o identifyf,o re ach subject,s ocially close directf riends,l ess close friendso f friends,a nd socially 1. Economistsa nd sociologistus se bothf requencyof i nteraction(M armaros and Sacerdote2 006) and intensityo f altruismi nterchangeabltyo measuret he strengtho f social connection(sG ranovette1r9 74, 1985; Marsdena nd Campbell 1984;M itchell1 987;P erlmana nd Fehr1 987;M athewse t al. 1998). 2. See fore xampleC oleman( 1993) on the industrialr evolutionin Western societies,V ölker( 1995) on the changef romc ommunismto capitalismi n East Germanya,n d Yellen( 1990) on thei ncreasedp articipatioonf t he Kalahari !Kung in markets. DIRECTED ALTRUISMA ND ENFORCED RECIPROCITY 1817 distants trangersW. e thenc onducta serieso f onlineg ames wheres ubjectsm akeu nilateraal llocationd ecisionfs ors everal typeso fn amedp artnerasn d onen amelessp artner(a randomly selectedp articipanftr omth es ubject'ds ormitoryT)h. eg amesa re eitherm odifieddi ctatogra mesa, s inA ndreonain dM iller(2 002), or a new" helpingg ame."S ubjectsm akem ultiplde ecisionbs ut are paidf oro ned ecisionse lecteda t randomF. ors omed ecisions, neitherp articipanits toldw hicho ft hed ecisionm aker'sc hoices was implementeIdn. thesea nonymoudse cisionst,h ed ifference ina llocationbse tweenfr iendasn ds trangerasl lowsu s toq uantify them agnitudoef d irectedal truismI.n particularw, e findt hat subjectsse ndo na verage5 2%m orem oneyto c losef riendtsh ant o strangerWs. ea lso demonstratthe atg ivingt of riendiss strongly relatedt o the decisionm akers'u nderlyinbga selinea ltruism towardn amelessp artners. Foro therd ecisionsb,o tht hed ecisionm akera ndt hep artner arei nformewdh icho ft hed ecisionm aker'csh oicews ass electedf or paymentW. eu se thed ifferenbceet weenth isn onanonymoaunsd thea nonymoutsr eatmentto m easuret hes eparatee ffecotf f uture interactioonn prosociabl ehaviorW. ef indt hatt hen atureo ft he futurien teractioenf fecdte pendsc ruciallyon the socialw elfare effectosf p rosociabl ehaviorW. heng ivingin creasejso ints urplus, subjectsin creaseg ivingt o friend(sr elativet o strangersb)y an additiona2l 4% in then onanonymoturse atmenTt.h us,d irected altruismis roughlytw icea s strongas futurien teractioenf fects in determininggiv ingb ehaviorW. heng ivingd ecreasejso ints ur- plus,s ubjectsd o notg ivem oret o friendisn then onanonymous treatmenItn. contrastt,h ed irecteadl truismef fecfta vorfs riends overs trangerbso thw henp rosociabl ehavioris sociallye fficient andw heni t is inefficient. Thesed ifferentieaflf ectosf f uturein teractioonn prosocial behaviora re well explainedb y the theoryo f repeatedg ames. Karlan et al. (2009) developa tractablet heoryf ora nalyzing repeatedg amesi n socialn etworkws, hichw e adaptt o our set- tinga nd call the enforcerde ciprocitmyo del.I n that model,a decisionm akerc an safelyg rantf avors(i n the formo f larger allocationst)o partnerws hent her elationshibpe tweent hemi s strongearn dm orev aluableb, ecauset hep artnewr ouldr atherr e- payt hef avort hand amaget hef riendshiGp.r antinfga vorsh, ow- ever,o nlyb enefitbs otht he decisionm akera nd partnerw hen givingin creasess ocials urplusT. hisc onflictwsi thB enaboua nd Tirole's(2 006)m odelw, herei ndividualgs ivei n ordert o signal 1818 QUARTERLYJ OURNALO F ECONOMICS beingo fa n altruistitcy peT. he signalingm odelp redicttsh ate x- cess givingt o friendssh ouldi ncreasef ora ll theg ames( and all exchangrea tes),b ecauseg ivingst illd istinguishaelst ruistfs rom selfisht ypesT. he enforcerde ciprocitmyo dela lso predicttsh at givingsh ouldin creaseif t hed ecisionm akera nd partnesrh area greatenr umbeorf c ommofnr iend(sc ontrollinfogr s ociald istance). Wec onfirmth isp redictioinn ourd ata.T hisp rovidefsu rtheerv i- dencef ort hee nforcerde ciprocimtyo deal ndd istinguishietsf rom relatedt heoriessu cha s preference-barseedci proci(tRya bin1 993; Dufwenberagn d Kirchsteige2r0 04).T hisr esulta lso highlights thei nfluencoef s econd-ordleirn ksa ndn etworskt ructuroen e co- nomicd ecisions. Last,w es howt hatd ecisionm akerws hoe xhibigt reatebr ase- linea ltruismar e treatedm oreg enerouslbyy t heirf riendsH. ow- everw, e showt hatf rienddso notr ewardin trinsikci ndnessb,u t rathert,h atk indp eoplet endt oh avef riendws hoe xhibigt reater baselinea ltruismth emselves. Ourp aperb uildso n a riche xperimentlaitl eraturoen other- regardinpgr eferencaensd c ooperatioAnl. truistbice haviotro ward strangerhsa s beeno bservedin a varietyo fl aboratorcyo ntexts (see Camerer[2 003]f ora n extensivseu rvey)H. offmanM,c Cabe, and Smith(1 996)s uggestth ata decreasein perceivesdo ciald is- tancei ncreasesd onationisn dictatogr ames.O ur paperr educes sociald istancein twow ays:( a) we distinguisbhe tweeng iving to friendasn d givingt o strangersa;n d (b) we eitherr evealo r do notr evealw hichd ecisionis implementeOdu. r onlinee xper- imentadl esigni s a significanmt ethodologicaadlv anceb ecause it providesa practicawl ayt o matchs ubjectsw itht heirr eal- worldf riendsL.3a b experimentths atr elaxt raditionaanl onymous matchinogf s trangertsy picallrye veald emographcihc aracteris- ticss ucha s gendero re thnicit(yfo re xamples,e e Fershtmaann d Gneezy[ 2001]).V eryf ewl aboratoreyx perimentesx plicitlrye ly on subjectso' ngoingr elationshipws itht heirf riendass we do in ourn onanonymoturse atmenItn. stead,r epeatedin teractions arei ncorporatdedir ectliyn tot hee xperimentdael signi,n creasing prosociabl ehavior(s ee Rotha nd Murningha[n1 978]a nd Murn- inghana nd Roth[ 1983]f ore arlye xamples)A. notablee xception is thes eminawl orko fG laesere ta l. (2000),w hom atchs ubjectast variouss ociald istanceisn a trustg ame.4In subsequenrte search, 3. Oure xperimentadle signf itst hed efinitioonf a n artef actualf ielde xperiment fromH arrisona nd List (2004). 4. See also Polzer,N eale, and Glenn( 1993) comparingal locationst o friends and strangersin an ultimatumg ame. DIRECTED ALTRUISMA ND ENFORCED RECIPROCITY 1819 Goereee t al. (2008)h ave adoptedt he anonymoutsr eatmenotf our experimentdale sign( usinga standardd ictatorg ame)a nd alsof inds tronegv idencfeo rd irecteadl truismin a schoonl etwork oft eenageg irls( alsos ee Brañas-Garzeat al. [2006]f ord ataf rom experimenwtsi thE uropeanu niversitsyt udents)T. o theb esto f ourk nowledgoeu, rd esigni s thef irstt o distinguisbhe tweend i- recteda ltruisman df uturien teractioenf fectins socialn etworks. The resto ft hep aperi s organizedas followsS.e ctionI I de- scribess everalr elevantth eorie-s enforcerde ciprocitsyig, naling, andp reference-barseedci proci-t yan dr eviewtsh eirt estableim - plicationsS.e ctionII I presenttsh ee xperimentdael signS. ection IV summarizetsh e main featureso f the data. Our empirical resultos nd irecteadl truismar ep resenteidn S ectionV . SectionV I analyzesd ecisionus ndern onanonymiItny S. ectionV II, we con- sidert wob enchmarkcso mparintgh ei mportancaen dm agnitude oft he directedal truisma nd futurein teractioenf fectsS.e ction VIII showst hatf riendtse ndt o have similarl evelso fb aseline altruismS.e ctionIX concludebs yd iscussintgh ei mplicationosf ourr esultsf ort het heoreticaanl d empiricaaln alysiso fp rosocial behavior. II. TheoreticalF ramework Web rieflrye viewt hep redictionosf t hreed ifferentht eories thate xplaind ecisionm akingw hent herei s a prospecotf f uture interactioWn. e refert hei nteresterde adert o thee arlierN BER workingp aperf ora detailedf ormatl reatmenotf each theory (Leidere t al. 2007). We assume throughoutth at therei s a decisionm akerM, , and a partnerP, , whoa re embeddedin a socialn etworkW. ec alculatet hes ociald istanceD mpb etweenth e decisionm akera nd thep artneras thes hortespta thc onnecting themf: ore xamplet,w od irectf riendhs avea sociald istanceo f1 , whereasa friendo fa friendis at distance2 (see FigureI ). The decisionm aker'asl locationde cisionisn t hea nonymoutsr eatment (neitherd ecisionm akern orp artnerle arnsa boutt he decision) and then onanonymoturse atmen(tb otha gentsl earna boutt he decisiona)r ed enotedb yx mpa ndx mpr, espective(lcyo deds o that a largerx impliesg reaterp rosociabl ehavior).5 We assume,a s a startingp ointt, hatd ecisionm akersh ave altruistipcr eferencwesh oses trengtvha riesw iths ociald istance. 5. Thereforet,h ed ecisionm aker'su tilityis decreasingin x and thep artner's utilityis increasingin x. 1820 QUARTERLYJ OURNALO F ECONOMICS Figure I Examplest o IllustrateD ifferencbee tweenM aximumN etworkF low and Social Distance All linksa re assumedt o have unitc apacityT. he examplesi llustratet he dif- ferentf eatureso f the social networkt hat are capturedb y social distancea nd maximumn etworkf lowr, espectivelyt:h e additiono fc ommonf riendsw illn eces- sarilyi ncreasef lowb utc an leave sociald istanceu nchanged. We use a simplel inears pecificatiotno captured ecisionm akingi n thea nonymoutsr eatmentw, hichi s a naturale xtensiono fe xisting preferences-baseadlt ruismm odels:6 (1) xMp= ocZmp+ YiDmp+ YM+ tMP. We controfl oro bservabled emographicch aracteristicosf b othi n- dividuals{ Zmp).T he coefficienytm c apturest he decisionm aker's intrinsigc enerosityto warda ll partners(i ndependenotf s ociald is- tance),w hichw e refert o as his or her baseline altruism.T he co- efficienyt' determinehs ow the decisionm aker'sa ltruismv aries withs ocial distance,w hichw e call directeda ltruism. U.A. EnforcedR eciprocity Repeatedg ames providea naturalf ramewortko analyzet he decisionm aker'sa llocationx, mp,u ndert he prospecto ff uturein - teractionsH. oweverr, epeatedg ames typicallya dmitm anye qui- libria even in a two-personse tting- this multiplicitpyr oblemi s compoundedf ors ocial networksw hereh undredso fa gentso ften interact.K arlan et al. (2009) providea tractablef rameworkfo r 6. Andreoni( 1990) modelsa ltruisma s a "warmg low,"w hereas Fehr and Schmidt( 1999), Boltona nd Ockenfels(2 000), and Charnessa nd Rabin (2002) focuso n preferenceosv erp ayofdfi stributions. DIRECTED ALTRUISMA ND ENFORCED RECIPROCITY 1821 modelingr epeatedg ame effectisn social networksT. heya ssume thatd ecisionm akera nd partners hare a relationshipt hati s con- sumedi n thef uturea nd givesb otho ft hemu tilityV mpA- decision makerw ho is morep rosocialu nder nonanonymitcyo mparedt o anonymitysu, ch that xmp- xmp> 0, grantsa favort o the part- ner and can use the value of the relationshipt o the partnera s social collateralt o enforcer epaymento ft he favorI. f the partner refusest o returnt hef avort her elationshipb reaksd owna nd both agentsl ose thev alue oft he friendship. In our workingp aper,w e formallysh ow that in the unique subgamep erfecetq uilibriumt hed ecisionm akera llocatesa larger amountu ndern onanonymitcyo mparedt o anonymityon lyw hen givingi s efficienatn, d thus increasess ocial surplus.I n this case, botht hed ecisionm akera nd thep artnerk eep someo ft he surplus createdb yt hef avorI.n contrastw, heng ivingi s inefficienat f avor that has to be repaid makes both the decisionm aker and the partnerw orseo ff. We can also showt hat the differencien allocationsb etween the nonanonymouasn d anonymoust reatmentsx,m p- xmp,i s in- creasingi n thev alue oft her elationshipV mpb ecause the decision makeri s willingt og rant,a nd thep artneris willingt or epayl,a rger favorsT. his motivatest he followingem piricalm odelo fa llocation decisionsu ndern onanonymity: (2) xMp= r]ZMp+ 0xMp+ </>VM+p v M+ ¿mp- We includet he decisionm aker'sa nonymousd ecision,x mp,a s a covariate,b ecause our enforcedr eciprocitym odel predictsd evi- ationsf romt he anonymousb enchmarkT. he modela lso predicts thate nforcedre ciprocitayn d altruisma re substitutes(0 < 1): be- cause the decisionm akerw ill give the largeste nforceablef avor (wheng ivingi s efficienta)n d the partnerw ill retains ome oft he surplus,t he decisionm aker'sm arginalu tilityf romg ivinga n ad- ditionala ltruisticg ifti s smallert hani n the anonymousc ase (the partneri s nowr ichert hani n the anonymousc ase). We considert wo proxiesf ort he value of the relationship, Vmp-F irst,w e expectt hat relationshipv alue decreasesw iths o- cial distanceb ecause a decisionm akeri s less likelyt o interact witha sociallyd istantp artneri n thef utureS. econd,t he enforced reciprocitmy odelo fK arlan et al. (2009) suggestsm aximumn et- workf lowa s an alternativem easure.T he maximumf lowc ounts then umbero fd istinctp athsb etweend ecisionm akera nd partner (see Figure I). Networkf lowc apturess tructurafl eatureso ft he 1822 QUARTERLYJ OURNALO F ECONOMICS socialn etwortkh ata ren otc apturebdy s ociald istanceal one.F or exampleh, avingm oref riendisn commoinn creasens etworfkl ow, whereass ociald istanceis unchangedN. etworfkl owf ormalizeas commoinn tuitioinn thes ociologlyit eraturteh atd ensen etworks strengthetnru stb yf acilitatiningf ormaarl rangemen(tCso leman 1988,1 990).I n Karlane t al.'s (2009)e nforcerde ciprocitmyo del, each commonfr iendin creasest hes ocialc ollaterabl etweent he decisionm akera nd the partnerT. he decisionm akeri s able to extracrt epaymentfso rl argerf avorsb,e causei ft hep artnerd e- faultsh e ors hew illl oset her elationshiwpi tha ll oft hec ommon friendass wella s thed ecisionm akerT. hroughouwt,e calculate networfkl owb yo nlyi ncludinlgin kst hata re at mosta distance K = 2 awayf romth ed ecisionm akerK. arlane t al. (2009)j ustify thisc hoiceo fa circleo ft rusto fK = 2 byr eviewinsgo meo ft he existingem piricalli teraturien economicasn d sociologysu, cha s Granovetter(1974). TableI summarizetsh em ainp redictionofst hee nforcerde ci- procitmy ode(la s wella s thes ignalinga-n dp reference-barseedci - procitmy odelsd iscussedb elow). II.B. Signaling In recentw ork,B enaboua nd Tiróle( 2006) proposea sig- nalingm odelt hatp rovideasn alternativteh eoryf ore xplaining greategr enerosittyof riendusn dern onanonymiItny t.h eirf rame- worka,g entsc area boutb eingp erceiveads altruisti(cr atherth an greedyt)y pess,o theya ctm oreg enerouslwyh ent heira ctionsc an be observedM. oreoverit, is reasonablet o extendt heirm odelt o assumet hati ndividualcsa rem orea bouts ignalingg enerosittyo friendtsh ant os trangerbs,e causet heya rem oreli kelyt oi nteract withf riendisn thef uture. A keyd istinguishipnrge dictioonft hes ignalinmg odeils that decisionm akerss houldi ncreaset heira llocationd ecisionu nder nonanonymictoym paredto anonymitbyo thw henp rosociabl e- haviori s efficienant d wheni t is inefficienLta.r gera llocations are just as good( ifn otb etters) ignalso fg enerositwyh eng iv- ingi s inefficieanst w heng ivingis efficienAtd. ditionalleyx,c ess givingu ndern onanonymiitsy in dependenotf t hel evelo fa ltru- ismi n Benaboua ndT irole'(s2 006)m odelb, ecauses ignalinugt il- itya nd distributionuatli lityar e additivelsye parableT. hise ffect contrastws itht hee nforcerde ciprocimtyo delw, herea ltruisman d favorasr e substitutes. DIRECTED ALTRUISMA ND ENFORCED RECIPROCITY 1823 TABLE I Testable Predictionsa bout Decision Makers'A llocation Decisions under nonanonymitvye rsusa nonymity Enforced Preference-based reciprocity Signaling reciprocity Greaterg enerosity Yes Yes Yes towardf riend(s0 > 0) wheng ivingis efficient Greaterg enerosity No Yes No towardf riend(s0 > 0) wheng ivingi s inefficient Altruistidc ecision Yes No Yes makersa re relatively less generoust oward friendcs omparedt o strangers Maximumn etworkf low Yes No No is a separatep redictor ofg enerositbye yond sociald istance Note.W e estimatet he empiricaml odelx ^jp = r¡Z+ Oxj^p+ ^MP + VM+ €MP>w herex ^p and x'jp are the decisionm aker'sa ctionsu ndern onanonymiatnyd anonymitrye,s pectivelayn, d V^p describest he valueo ft her elationshibpe tweend ecisionm akera nd partner(p roxiedb ys ociald istance). U.C. Preference-BaseRde ciprocity A thirdp ossiblem echanismfo rf uturien teractioenf fectiss preference-barseedci procitDyu.f wenberang dK irchsteig(e2r0 04) developa psychologicgala met heorym odelo f sequentialr eci- procityw,h erea n individuatlr eatsk indly(u nkindlyt)h osew ho have treated/wtilrle ath imo r herk indly(u nkindlyA). s in en- forcerde ciprocittyh,e d ecisionm akerin creasehs is orh era lloca- tionb etweenn onanonymiatnyd anonymitoyn lyw heng ivingis efficienatn,t icipatintgh atw iths omep robabilitays,s umedt o be decreasinign sociald istancet,h ep artnerm aya ct to benefitth e decisionm aker. Underp reference-basreedc iprocithyo,w evert,h e partner's desiret or eturnth ed ecisionm aker'fs avoris intrinsirca, therth an designedto p reservteh er elationshiwpi tht hed ecisionm akero r commofnr iendTs.h ereforuen, liket hee nforcerde ciprocimtyo del, preference-basreedc iprocitwyo uldn otp redictth att hen etwork flowm easurein dependentcloyr relatews ithi ncreasedg enerosity 1824 QUARTERLYJ OURNALO F ECONOMICS | Networekli citatioin ! Allocatiogna me j Wavei I ^"^" Dinctaamtoegrlea smsp easr( atnnoenr ) I "^^" Dictnaatmogrea dmp aerst( anneorn ) ÏÏect Coordinatiotnas k -land -land ^design J I 1 I ^ Dictatorg ames (nonanon) I ^ Dictatorg ames (nonanon) nameless partner named partner game (anon) named partner i 1 Helping game (nonanon) named partner Figure II Overviewo fE xperimentaDl esign( Waves1 and 2) In Wave 1, we randomizedw hethers ubjectsf irstm ade all the namelessd eci- sionso rw hetherth eyf irstm adea ll then amedd ecisionsF. ore ach namedp artner inW ave1 ,w e randomizewd hethersu bjectsf irstsu bmittetdh eira nonymoucsh oice fort hatp artnero rt heirn onanonymoucsh oiceW. ea lso randomizedw hethera sub- ject firsts ubmittedal l anonymou(sn onanonymoucs)h oicesf ora ll partnerso n a singles creeno, rw hethera subjects ubmittebdo tha nonymouasn d nonanonymous choiceso n a separates creenf ore ach partnerF. inallyw, e randomizedth eo rderi n whicht het hreed ifferenetx changer atesw erel istedo n subjectss' creens. afterc ontrollinfgo rt hef requencyof f uturein teraction(v ia social distance). III. ExperimentalD esign FigureI I presentst hem ajorf eatureso fo ur experimentadl e- sign.I n ordert o recruitm ores ubjectsa nd map a largers ocialn et- work,a ll communicatiown iths ubjectsw as conductedb y e-mail and all choicesw ere submittedo n a website( rathert han in a laboratoryt) hat subjectsc ould access witha passwordt hrough theiro wn web browsersW. e conductedt wo waves oft he experi- ment.I n each wave,w e firstu sed a novelt ask to elicitt he social networkt ruthfullWy.e thenh ad subjectsp laya n allocationg ame witho thers ubjectsi n the networkI.n Wave 1, we used modified dictatorg ames withv aryinge xchanger ates,a nd in Wave 2, we used a new helpingg ame. In bothw aves, decisionm akersf irst decidedo n allocationsb etweent hemselvesa nd otheru nnamed ("nameless")p artners,a nd then,a fewd ays later,m ade several allocationsb etweent hemselvesa nd named partners( identified by real firsta nd last name) at variouss ocial distances.F or both waves,a singled ecisionw as randomlys electedf orp aymenta nd all playersw erei nformeodf t heire arningsb y e-mail.

Description:
Quoc-Anh Do. We conducted online to GARP: An Experiment on Leider, Stephen, Markus M. Mobius, Tanya Rosenblat, and Quoc-Anh Do, Di-.
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.