IntroductiontoBehavioralEconomics復(fù)旦行為經(jīng)濟(jì)學(xué)課件

IntroductiontoBehavioralEconomics

行為經(jīng)濟(jì)學(xué)

方向博士OklahomaStateUniversityIntroductiontoBehavioralEco參考書(shū)作者:孫惟微：獨(dú)立經(jīng)濟(jì)學(xué)者，策略顧問(wèn)、熙代（北京）文化傳媒有限公司執(zhí)行董事。參考書(shū)作者:孫惟微：獨(dú)立經(jīng)濟(jì)學(xué)者，策略顧問(wèn)、熙代（北京）文董志勇

(作者)

董志勇(作者)

丹·艾瑞里

丹·艾瑞里

AShortSurveyFocusingfornowonindividualdecisions,whatdoesmainstream(standardneoclassical)economicsnormallyassume?AShortSurveyFocusingfornowAssumptionsRationality無(wú)限理性:Isperfectlyrational,makingchoicesthatconsistentlymaximizesomeexogenous,stablesetofpreferencesthatdependonabsolutelevelsofoutcomes(ratherthanchanges).經(jīng)濟(jì)學(xué)假設(shè)個(gè)人具有穩(wěn)定和連續(xù)的偏好，并用無(wú)限理性使這些偏好最大化Self-Interest無(wú)限自私自利:Isalmostalwaysalsoassumedtobeperfectlyself-interested,caringexclusivelyabouther/hisownconsumption,thoughthisassumptionisnotessentialtomainstreamtheory.

Self-Control無(wú)限意志力:Hasperfectwill-powerandtheabilitytomakeandfollowintertemporalplans(evencontingentones),withnoconflictbetweenthepreferencesofcurrentandfutureselves.AssumptionsRationality無(wú)限理性:I而行為經(jīng)濟(jì)學(xué)正是基于對(duì)以上三個(gè)“無(wú)限性”的反對(duì)提出了自己關(guān)于非理性人的新觀點(diǎn)1。正常生活中的經(jīng)濟(jì)個(gè)體應(yīng)該是非理性的，這樣一個(gè)非理性的人，并不具備穩(wěn)定和連續(xù)的偏好以及使這些偏好最大化的無(wú)限理性；2。即使知道效用最大化的最優(yōu)解也有可能因?yàn)樽晕铱刂埔庵玖Ψ矫娴脑蚨鵁o(wú)法做出相應(yīng)的最優(yōu)決策；3。其經(jīng)濟(jì)決策的過(guò)程中包含了相當(dāng)?shù)姆俏镔|(zhì)動(dòng)機(jī)和非經(jīng)濟(jì)動(dòng)機(jī)權(quán)重，人們會(huì)愿意犧牲自己的利益去幫助他人。而行為經(jīng)濟(jì)學(xué)正是基于對(duì)以上三個(gè)“無(wú)限性”的反對(duì)提出了自己關(guān)于行為經(jīng)濟(jì)學(xué)的興起主要有三方面的原因第一是基于對(duì)精神心理因素的研究?，F(xiàn)代經(jīng)濟(jì)學(xué)向心理學(xué)靠攏,同現(xiàn)代經(jīng)濟(jì)中的精神心理因素有關(guān)，與現(xiàn)代經(jīng)濟(jì)的日益“非物質(zhì)化”、“商品的個(gè)性化”有關(guān)。行為經(jīng)濟(jì)學(xué)的興起主要有三方面的原因第一是基于對(duì)精神心理因素第二，行為經(jīng)濟(jì)學(xué)拓寬了正統(tǒng)經(jīng)濟(jì)學(xué)的領(lǐng)域。經(jīng)濟(jì)學(xué)家一向認(rèn)為人是理性的動(dòng)物，即人總是傾向于追求最大收益或承擔(dān)最小成本。但人類(lèi)行為固然有其理性一面，也有其非理性的一面。所以，非理性行為可用科學(xué)的方法加以研究。(“不買(mǎi)最好的，只買(mǎi)最貴的?！?第二，行為經(jīng)濟(jì)學(xué)拓寬了正統(tǒng)經(jīng)濟(jì)學(xué)的領(lǐng)域。經(jīng)濟(jì)學(xué)家一向認(rèn)為人是第三，傳統(tǒng)經(jīng)濟(jì)學(xué)認(rèn)為，經(jīng)濟(jì)學(xué)是一門(mén)研究財(cái)富的學(xué)問(wèn)；而現(xiàn)代經(jīng)濟(jì)學(xué)提出，經(jīng)濟(jì)學(xué)也是一門(mén)研究人的學(xué)問(wèn)。人類(lèi)社會(huì)化的行為受復(fù)雜的社會(huì)關(guān)系制約，從而導(dǎo)致人的行為選擇并非都是建立在理性思考的基礎(chǔ)上。因此從純理性研究人的行為行不通。從經(jīng)濟(jì)主體講，經(jīng)濟(jì)過(guò)程最終體現(xiàn)為人的經(jīng)濟(jì)行為過(guò)程，因而對(duì)人的分析離不開(kāi)用人的心理對(duì)其行為的解釋。第三，傳統(tǒng)經(jīng)濟(jì)學(xué)認(rèn)為，經(jīng)濟(jì)學(xué)是一門(mén)研究財(cái)富的學(xué)問(wèn)；而現(xiàn)代經(jīng)濟(jì)亞當(dāng)斯密AdamSmith每個(gè)人都力求運(yùn)用自己的資本，生產(chǎn)出最大的價(jià)值。一般而來(lái)，他不會(huì)為了促進(jìn)公共利益，也不知道促進(jìn)多少。他只考慮自己的安全，自己的所得。正是這樣，他由一只看不見(jiàn)的手引導(dǎo)著，實(shí)現(xiàn)著他自己并不打算實(shí)現(xiàn)的目標(biāo)。與有意地促進(jìn)相比，在追求他自己得利益的過(guò)程中，往往能夠更加有效地促進(jìn)社會(huì)得公益。亞當(dāng)斯密AdamSmith每個(gè)人都力求運(yùn)用自己的資本，生直到最近，經(jīng)濟(jì)學(xué)尚被普遍看作一種必須依賴(lài)于對(duì)現(xiàn)實(shí)世界的觀察、而不能依靠在實(shí)驗(yàn)室里做受控制的實(shí)驗(yàn)來(lái)進(jìn)行研究的非實(shí)驗(yàn)性科學(xué)。許多評(píng)論者發(fā)現(xiàn)受利己主義和理性決策所支配的“經(jīng)濟(jì)人”(homooeconomicus)這一普遍假設(shè)存在局限性。然而，經(jīng)濟(jì)學(xué)研究已經(jīng)開(kāi)始朝著新的方向起飛。大量和日益增多的科學(xué)工作致力于對(duì)經(jīng)濟(jì)學(xué)傳統(tǒng)的基本假設(shè)的經(jīng)驗(yàn)檢驗(yàn)和修改，特別是有關(guān)無(wú)限理性、純粹利己主義和完全自治的假設(shè)。直到最近，經(jīng)濟(jì)學(xué)尚被普遍看作一種必須依賴(lài)于對(duì)現(xiàn)實(shí)世界的觀察、而且，目前的研究越來(lái)越依賴(lài)于來(lái)自實(shí)驗(yàn)室的新的實(shí)驗(yàn)數(shù)據(jù)，而不是從對(duì)實(shí)際經(jīng)濟(jì)的觀察中所得到的傳統(tǒng)的現(xiàn)場(chǎng)數(shù)據(jù)。最近的這些研究源于兩個(gè)區(qū)別顯著、但正趨于一致的傳統(tǒng)：一是認(rèn)知心理學(xué)中有關(guān)個(gè)人決策的理論和經(jīng)驗(yàn)研究；二是以實(shí)驗(yàn)的方法對(duì)經(jīng)濟(jì)理論預(yù)言的檢驗(yàn)。行為經(jīng)濟(jì)學(xué)和實(shí)驗(yàn)經(jīng)濟(jì)學(xué)已經(jīng)躋身于經(jīng)濟(jì)學(xué)最活躍的領(lǐng)域之列。2002的諾貝爾經(jīng)濟(jì)學(xué)獎(jiǎng)得主就是在這兩個(gè)領(lǐng)域從事研究的前輩。而且，目前的研究越來(lái)越依賴(lài)于來(lái)自實(shí)驗(yàn)室的新的實(shí)驗(yàn)數(shù)據(jù)，而不是2002:VernonSmithandDanielKahnemanreceivedtheNobelPrizeVernonSmith:“fortheuseoflaboratoryexperimentsasatoolinempiricaleconomicanalysis,inparticular,forthestudyofdifferentmarketmechanisms.”Founderofexperimentaleconomics.DanielKahneman:“fortheintroductionofinsightsfrompsychologicalresearchintoeconomics,inparticularwithregardtojudgmentsanddecisionsunderuncertainty.”Kahneman’sresearchisbasedonpsychologicalexperimentsandquestionnaires.Founderofbehavioraleconomics.2002:VernonSmithandDaniel實(shí)驗(yàn)經(jīng)濟(jì)學(xué)傳統(tǒng)上，經(jīng)濟(jì)學(xué)被看作一種完全依賴(lài)現(xiàn)場(chǎng)數(shù)據(jù)的非實(shí)驗(yàn)性科學(xué)。許多人認(rèn)為，這一點(diǎn)是經(jīng)濟(jì)學(xué)作為一門(mén)科學(xué)繼續(xù)發(fā)展的障礙。除非我們可以進(jìn)行受控制的實(shí)驗(yàn)，否則對(duì)經(jīng)濟(jì)理論的檢驗(yàn)將永遠(yuǎn)是有限的。僅僅根據(jù)現(xiàn)場(chǎng)數(shù)據(jù)，人們很難判斷一個(gè)理論是否失敗或何時(shí)失敗，并正確地指出失敗的原因。在理論與受控制的實(shí)驗(yàn)觀察兩者之間的反饋渠道在經(jīng)濟(jì)學(xué)中基本不存在。一種新興的、被稱(chēng)作“實(shí)驗(yàn)經(jīng)濟(jì)學(xué)”的研究領(lǐng)域的建立從根本上對(duì)這一觀點(diǎn)提出了挑戰(zhàn)。在受控制的實(shí)驗(yàn)室里，實(shí)驗(yàn)者以簡(jiǎn)單、抽象的形式模仿在市場(chǎng)和其他經(jīng)濟(jì)交往形式中出現(xiàn)的場(chǎng)景，以研究這些場(chǎng)景中的人類(lèi)行為。實(shí)驗(yàn)經(jīng)濟(jì)學(xué)傳統(tǒng)上，經(jīng)濟(jì)學(xué)被看作一種完全依賴(lài)現(xiàn)場(chǎng)數(shù)據(jù)的非實(shí)驗(yàn)性實(shí)驗(yàn)經(jīng)濟(jì)學(xué)的歷史早年已有前輩學(xué)者從事經(jīng)濟(jì)學(xué)的實(shí)驗(yàn)研究。50多年前，張伯倫(Chamberlin，1948)試圖通過(guò)實(shí)驗(yàn)來(lái)檢驗(yàn)新古典的完全競(jìng)爭(zhēng)理論。約翰·納什(JohnNash)——1994年諾貝爾經(jīng)濟(jì)學(xué)獎(jiǎng)得主——與他的同事通過(guò)設(shè)置一個(gè)實(shí)驗(yàn)對(duì)博弈論的預(yù)測(cè)能力進(jìn)行了早期研究。實(shí)驗(yàn)經(jīng)濟(jì)學(xué)的歷史早年已有前輩學(xué)者從事經(jīng)濟(jì)學(xué)的實(shí)驗(yàn)研究。50多弗農(nóng)·史密斯在亞利桑那大學(xué)十一個(gè)班級(jí)進(jìn)行了長(zhǎng)達(dá)六年的實(shí)驗(yàn)，驗(yàn)證了競(jìng)爭(zhēng)均衡理論。據(jù)此實(shí)驗(yàn)所撰寫(xiě)的論文《競(jìng)爭(zhēng)市場(chǎng)行為的實(shí)驗(yàn)研究》在1962年的《政治經(jīng)濟(jì)學(xué)雜志》發(fā)表，標(biāo)志著實(shí)驗(yàn)經(jīng)濟(jì)學(xué)的誕生。弗農(nóng)·史密斯在亞利桑那大學(xué)十一個(gè)班級(jí)進(jìn)行了長(zhǎng)達(dá)六年的實(shí)驗(yàn)，驗(yàn)弗農(nóng)·史密斯的意義最重大的工作涉及市場(chǎng)機(jī)制。他對(duì)競(jìng)爭(zhēng)性市場(chǎng)所做的創(chuàng)新性實(shí)驗(yàn)(1962)、對(duì)不同拍賣(mài)形式的檢驗(yàn)(1965，1976，1980)，以及對(duì)“誘導(dǎo)價(jià)值法”(induced-value，1976)的設(shè)計(jì)，都為這一領(lǐng)域的研究奠定了基礎(chǔ)。弗農(nóng)·史密斯的意義最重大的工作涉及市場(chǎng)機(jī)制。他對(duì)競(jìng)爭(zhēng)性市場(chǎng)所查爾斯·普洛特認(rèn)為：“實(shí)驗(yàn)室建立的經(jīng)濟(jì)與現(xiàn)實(shí)經(jīng)濟(jì)相比可能特別簡(jiǎn)單，但是卻一樣地真實(shí)。真實(shí)的人被真實(shí)的金錢(qián)所驅(qū)動(dòng)，因?yàn)檎鎸?shí)的天賦和真實(shí)的局限，作出真實(shí)的決策和真實(shí)的錯(cuò)誤，并為其行為后果而真實(shí)地悲喜”。查爾斯·普洛特認(rèn)為：“實(shí)驗(yàn)室建立的經(jīng)濟(jì)與現(xiàn)實(shí)經(jīng)濟(jì)相比可能特別HistoryofBehavioralEconomics

行為經(jīng)濟(jì)學(xué)歷史AdamSmith,whoisbestknownfortheconceptofthe"invisiblehand"andTheWealthofNations,wrotealesswell-knownbookTheTheoryofMoralSentiments(《道德情操論》),whichlaidoutpsychologicalprinciplesofindividualbehaviorthatarearguablyasprofoundashiseconomicobservations.Forexample,AdamSmithcommented(1759/1892,311)that"wesuffermore...whenwefallfromabettertoaworsesituation,thanweeverenjoywhenwerisefromaworsetoabetter.”Lossaversion!損失厭惡HistoryofBehavioralEconomic認(rèn)知心理學(xué)的發(fā)展Beginningaround1960,cognitivepsychologybecamedominatedbythemetaphorofthebrainasaninformation-processingdevicereplacingthebehavioristconceptionofthebrainasastimulus-responsemachine.Theinformation-processingmetaphorpermittedafreshstudyofneglectedtopicslikememory,problemsolvinganddecisionmaking.Thesenewtopicsweremoreobviouslyrelevanttotheneoclassicalconceptionofutilitymaximizationthanbehaviorismhadappearedtobe.認(rèn)知心理學(xué)的發(fā)展Beginningaround1960,HerbertSimon西蒙現(xiàn)代決策理論的奠基人–卡內(nèi)基大學(xué)教授西蒙，心理學(xué)家和人工智能的創(chuàng)立者批判理性經(jīng)濟(jì)人的假設(shè)，提出“有限理性”(boundedrationality)概念,獲得1978年諾貝爾獎(jiǎng)。西蒙指出了新古典經(jīng)濟(jì)學(xué)理論的不現(xiàn)實(shí)之處，分析了它的兩個(gè)致命弱點(diǎn)：(1)假定目前狀況與未來(lái)變化具有必然的一致性；(2)假定全部可供選擇的“備選方案”和“策略”的可能結(jié)果都是已知的。HerbertSimon西蒙現(xiàn)代決策理論的奠基人–卡有限理性人的有限性包括兩個(gè)方面的含義，一是環(huán)境是復(fù)雜的，在非個(gè)人交換形式中，人們面臨的是一個(gè)復(fù)雜的、不確定的世界，而且交易越多，不確定性就越大，信息也就越不完全；二是人對(duì)環(huán)境的計(jì)算能力和認(rèn)識(shí)能力是有限的，人不可能無(wú)所不知。西蒙認(rèn)為,人們?cè)跊Q定過(guò)程中尋找的并非是“最優(yōu)”的標(biāo)準(zhǔn)，而只是“滿意”的標(biāo)準(zhǔn)。有限理性人的有限性包括兩個(gè)方面的含義，一是環(huán)境是復(fù)雜的，在非TverskyandKahnemanPerhapsthetwomostinfluentialcontributionswerepublishedbyTverskyandKahneman.Their1974Sciencearticlearguedthatheuristicshort-cutscreatedprobabilityjudgmentswhichdeviatedfromstatisticalprinciples.(AmosTverskyandDanielKahneman,“JudgmentunderUncertainty:HeuristicsandBiases,”Science185(1974),pp.1124–1131)Their1979paper"Prospecttheory:decisionmakingunderrisk"documentedviolationsofexpectedutilityandproposedanaxiomatictheory,groundedinpsychophysicalprinciples,toexplaintheviolations.ItwaspublishedinthetechnicaljournalEconometricaandisoneofthemostwidelycitedpaperseverpublishedinthatjournal.TverskyandKahnemanPerhapsthStepsthatearlyBEresearchtookFirst,identifynormativeassumptionsormodelsthatareubiquitouslyusedbyeconomists,suchasBayesianupdating,expectedutilityanddiscountedutility.Second,identifyanomalies—i.e.,demonstrateclearviolationsoftheassumptionormodel,andpainstakinglyruleoutalternativeexplanations(suchassubjects’confusionortransactionscosts).Third,usetheanomaliesasinspirationtocreatealternativetheoriesthatgeneralizeexistingmodels.Afourthstepistoconstructeconomicmodelsofbehaviorusingthebehavioralassumptionsfromthethirdstep,derivefreshimplications,andtestthem.StepsthatearlyBEresearcht例子1A.你一定能賺30000元。

B.你有80%可能賺40000元，20%可能性什么也得不到。例子1A.你一定能賺30000元。例子2A.你一定會(huì)賠30000元。B.你有80%可能賠40000元，20%可能不賠錢(qián)。例子2A.你一定會(huì)賠30000元。例子3假設(shè)你面對(duì)這樣一個(gè)選擇：在商品和服務(wù)價(jià)格相同的情況下，你有兩種選擇：A.其他同事一年掙6萬(wàn)元的情況下，你的年收入7萬(wàn)元。B.其他同事年收入為9萬(wàn)元的情況下，你一年有8萬(wàn)元進(jìn)賬。例子3假設(shè)你面對(duì)這樣一個(gè)選擇：在商品和服務(wù)價(jià)格相同的情況下，認(rèn)知心理學(xué)家考慮的是一個(gè)交互作用的過(guò)程，幾個(gè)因素都會(huì)對(duì)決策產(chǎn)生重要影響，比如知覺(jué)、信仰或心智模式。諸如感情、態(tài)度等內(nèi)在動(dòng)機(jī)也會(huì)影響一項(xiàng)決策。此外，對(duì)以前決策及其后果的記憶是一個(gè)至關(guān)重要的認(rèn)知函數(shù)。在這種復(fù)雜觀點(diǎn)下，人類(lèi)行為被認(rèn)為是局部地適應(yīng)于一個(gè)既定的環(huán)境。行為具有適應(yīng)性，取決于環(huán)境和瞬間的感知狀態(tài)。

卡尼曼等人通過(guò)調(diào)查和實(shí)驗(yàn)收集到的事實(shí)，對(duì)經(jīng)濟(jì)學(xué)的理性假設(shè)，至少是對(duì)復(fù)雜決策情景下的理性假設(shè)提出了質(zhì)疑。例如，現(xiàn)實(shí)世界中的決策者不總是依據(jù)概率法則評(píng)價(jià)不確定前景，有時(shí)，制定決策會(huì)違背預(yù)期效用最大化原則?？崧呢暙I(xiàn)主要是關(guān)于不確定條件下的判斷和決策。認(rèn)知心理學(xué)家考慮的是一個(gè)交互作用的過(guò)程，幾個(gè)因素都會(huì)對(duì)決策產(chǎn)阿莫斯·特維爾斯基(AmosTversky)，丹尼爾·卡尼曼(DanielKahneman)，里查德·薩勒(RichardH.Thaler)，馬修·拉賓(MatthewRabin)，美籍華人奚愷元教授等是這一學(xué)科的開(kāi)創(chuàng)性代表。阿莫斯·特維爾斯基(AmosTversky)，丹尼爾·卡總結(jié)丹尼爾·卡尼曼運(yùn)用認(rèn)知心理學(xué)中關(guān)于心理過(guò)程的深刻見(jiàn)解，幫助我們更好地理解了人們制定經(jīng)濟(jì)決策的行為?？崧吞鼐S爾斯基對(duì)不確定條件下決策行為的研究最有影響力?？崧€對(duì)行為經(jīng)濟(jì)學(xué)等其他領(lǐng)域作出了開(kāi)創(chuàng)性的貢獻(xiàn)?？崧呀?jīng)成為近期在行為經(jīng)濟(jì)學(xué)和金融學(xué)研究領(lǐng)域中所出現(xiàn)的繁榮景象背后的一個(gè)主要靈感來(lái)源。他的研究對(duì)其他領(lǐng)域也產(chǎn)生了重大影響。

弗農(nóng)·史密斯是開(kāi)創(chuàng)以實(shí)驗(yàn)作為經(jīng)濟(jì)學(xué)經(jīng)驗(yàn)方法論的一個(gè)最有影響的人物。與卡尼曼不同，他沒(méi)有以挑戰(zhàn)傳統(tǒng)經(jīng)濟(jì)學(xué)理性決策理論而開(kāi)始，而是檢驗(yàn)了關(guān)于市場(chǎng)表現(xiàn)的假說(shuō)?？崧恼{(diào)查和實(shí)驗(yàn)主要關(guān)注個(gè)人的決策，而史密斯則將實(shí)驗(yàn)的重點(diǎn)放在特定市場(chǎng)環(huán)境下人與人之間的相互作用上。他還強(qiáng)調(diào)了方法論問(wèn)題，發(fā)展了具有實(shí)踐性的實(shí)驗(yàn)方法，建立了構(gòu)筑一個(gè)良好實(shí)驗(yàn)的標(biāo)準(zhǔn)。基于史密斯的成就，許多經(jīng)濟(jì)學(xué)家開(kāi)始把實(shí)驗(yàn)室實(shí)驗(yàn)作為一個(gè)基本工具。

總結(jié)丹尼爾·卡尼曼運(yùn)用認(rèn)知心理學(xué)中關(guān)于心理過(guò)程的深刻見(jiàn)解，幫MainTopics前景理論(ProspectTheory)啟發(fā)式與偏差(HeuristicsandBiases)錨定理論和現(xiàn)狀偏見(jiàn)(AnchoringandAdjustment,StatusQuoBias)稟賦效應(yīng)(EndowmentEffect)偏好理論(PreferenceReversal)時(shí)間貼現(xiàn)和跨期選擇(TimeDiscountingandTimePreference)心理帳戶(MentalAccounting)其他應(yīng)用MainTopics前景理論(ProspectTheoExample:StatusQuoBias

現(xiàn)狀偏見(jiàn)Decision-makershaveanoverwhelmingtendencytoadoptdefaults,tostickwiththestatusquoEvenwhenthedecisionisimportantandthestakesarelargeEvenwhenthedecision-makeristoldthatthedefaultissuboptimalExamplesfrom401(k)plans:participation,savingsrate,assetallocation(companystock).Otherexamples:insurancedeductibles,organdonationExample:StatusQuoBias

現(xiàn)狀偏見(jiàn)BrigitteMadrianandDennisShea(2001)

Design:AFortune500CompanySwitched401(k)defaultonApril1,1998.MadrianandSheaexaminebehaviorofnewhires.

OLDDefaultContribution:MustactivelysignupDefaultAllocation:None

NEWDefaultContribution:3percentofcompensationdeductedforplanDefaultAllocation:MoneyMarketFundBrigitteMadrianandDennisSh401(k)ParticipationIncreases

PercentatSpecifiedContributionRateSource:MadrianandShea(2001).ContributionRate401(k)ParticipationIncreases401(k)AssetAllocationAlsoChangedSource:MadrianandShea(2001).PercentofAssets401(k)AssetAllocationAlsoCImpactofAutomaticEnrollmentin401(k)Before,DuringandAfterSource:Choi,Laibson,Madrian,Metrick(2004)ImpactofAutomaticEnrollmentPolicyApplication:PensionReformBillof2006

The900PensionProtectionActof2006comesasthenumberofpeoplecoveredbyadefined-benefitpensionhassteadilydeclinedandawarenesshasgrownaboutthelackofadequatesavingsamongAmericans. Amajorityofworkers45andolderhavelessthan$50,000insavings,accordingtoasurveybytheEmployeeBenefitResearchInstitute(EBRI).What'smore,almost40percentofworkersover40don'tparticipateina401(k)whentheyareeligible. Thenewlegislationencouragescompaniestoautomaticallyenroll401(k)-eligibleemployeesandtoautomaticallyincreaseworkercontributionseveryyear.Italsoallowstheplanproviderchosenbytheemployertoofferinvestmentadvicetoworkers. Automaticenrollmentisexpectedtoboosttheparticipationratein401(k)plansbeyond90percent. ByJeanneSahadi,CNNMLibertarian-Paternalism:Setthedefaulttohelppeople,buttheycanoptout.PolicyApplication:PensionRe“不滿意七天可以退貨”的商品;汽車(chē)行也會(huì)借車(chē)給人試駕。為救美國(guó)經(jīng)濟(jì)于水深火熱，美聯(lián)儲(chǔ)主席伯南克有句名言：“如有必要，可用直升機(jī)撒錢(qián)?！币馔庵?cái)?shù)赝?/p>

“不滿意七天可以退貨”的商品;汽車(chē)行也會(huì)借車(chē)給人試駕。IntroductiontoLabExperimentsIntroductiontoLabExperimentLabExperiments“Onepossiblewayoffiguringouteconomiclaws...isbycontrolledexperiments....Economists(unfortunately)...cannotperformthecontrolledexperimentsofchemistsorbiologistsbecausetheycannoteasilycontrolotherimportantfactors.Likeastronomersormeteorologists,theygenerallymustbecontentlargelytoobserve.”(SamuelsonandNordhaus,1985,p.8)“EconomicTheory,throughaformaldeductivesystem,providesthebasisforexperimentalabstractionandtheexperimentaldesign,butsocietyinmostcasescarriesouttheexperiment,Therefore,theeconomicresearcherobservestheoutcomeofsociety’sexperimentorperformancebuthaslittleornoimpactontheexperimentaldesignandtheobservationsgenerated.Thus,bythepassivenatureofthedata,economicresearchersare,toalargeextent,restrictedintheirknowledgesearchtotheprocessofnonexperimentalmodelbuilding....theexperimentisoutsidetheresearcher’scontrol.”(“TheNonexperimentalModel-BuildingRestriction”inJudgeetal.(1988))LabExperiments“OnepossiblewPhilosophyofscience:thescientificmethodPhilosophyHowweunderstandtheworldScienceA“method”forachievingthisgoalDevelopmentoftheoryFormulationofhypotheses-predictionsTestsofpredictionsDesign&AnalysisHowwetestpredictions ofhypothesesStatisticsToolforquantitativetestsPhilosophyofscience:thesciScience—AssumptionsRealismWorldissubjecttoinvestigationRationality—logicRegularityResultscanbegeneralizedCausality—worldisordered—determinismDiscoverabilityorknowabilityofcausesScience—AssumptionsRealismDefinitionsInduction(orinductivereasoning):reasoningthatgenerallawsexistbecauseparticularcasesthatseemtobeexamplesofitexistDeduction(ordeductivereasoning):reasoningthatsomethingmustbetruebecauseitisaparticularcaseofageneral(universal)lawknowntobetrue Inductionspecificgeneral deduction5swansseen,allarewhiteallswansarewhiteDefinitionsInduction(orinducTypesofResearchDesignThreetraditionalcategories:ExploratoryDescriptiveCausalThechoiceofthemostappropriatedesigndependslargelyontheobjectivesoftheresearchandhowmuchisknownabouttheproblemandresearchobjectives.TypesofResearchDesignThreeBasicResearchObjectivesandResearchDesignResearchObjective AppropriateDesignTogainbackgroundinformation, Exploratorytodefineterms,toclarify problemsandhypotheses,toestablishresearchprioritiesTodescribeandmeasuremarketing DescriptivephenomenaatapointintimeTodeterminecausality, Causaltomake“if-then”statements BasicResearchObjectivesandExploratoryResearchExploratoryresearchismostcommonlyunstructured,informalresearchthatisundertakentogainbackgroundinformationaboutthegeneralnatureoftheresearchproblem.Byunstructured,wemeanthereisnoformalsetofobjectives,sampleplan,orquestionnaire.ExploratoryResearchExploratorExploratoryResearchItisusuallyconductedwhentheresearcherdoesnotknowmuchabouttheproblems.Exploratoryresearchisusuallyconductedattheoutsetofresearchprojects.ExploratoryResearchItisusuaExploratoryResearchUsesGainBackgroundInformationDefineTermsClarifyProblemsandHypothesis(refineresearchobjectives)EstablishResearchPrioritiesManyquestions;manysourcesDefiningtheproblem;gettinga“feel”ExploratoryResearchManyquesExploratoryResearchAvarietyofmethodsareavailabletoconductexploratoryresearch.SecondaryDataAnalysisExperienceSurveysCaseAnalysisExploratoryResearchAvarietyDescriptiveResearchDescriptiveresearchisundertakentodescribeanswerstoquestionsofwho,what,where,when,andhow.Descriptiveresearchisdesirablewhenwewishtoprojectastudy’sfindingstoalargerpopulation,ifthestudy’ssampleisrepresentative.DescriptiveResearchDescriptivCausalResearchCausalitymaybethoughtofasunderstandingaphenomenonintermsofconditionalstatementsoftheform“Ifx,theny.”Causalstudiesareconductedthroughtheuseofexperiments.CausalResearchCausalitymaybExperimentsAnexperiment

isdefinedasmanipulatinganindependentvariabletoseehowitaffectsadependentvariable,whilealsocontrollingtheeffectsofadditionalextraneousvariables.ExperimentsAnexperimentisdeExample:OurPropositionHypothesis:SouthernwhitemalesaremorepronetoaggressionthanareNorthernwhitemales.(conjecture)We’lllookatdifferentstrategiestoexaminehypotheses.(refutation)Example:OurPropositionHypothRefutationputtingtheoriestothetestWhatevidencesupports(and,moreimportantly,disconfirms)ourtheories?RefutationputtingtheoriestoDescriptiveResearchassessestheamountoraveragelevelofagivenvariableinapopulatione.g.,publicopinionsurveysnotatruetestofanhypothesisWhatisthewhitemalehomiciderateintheSouth?DescriptiveResearchassessestDescriptiveResearchWhatisthewhitemalehomiciderateintheSouth?adaptedfromNisbett(1993)DescriptiveResearchWhatisthDescriptiveResearchcriticalissuesrandomsamplingbasisofcomparisoninformative,butnotahypothesistestWhatelsecouldaccountforthefindings?DescriptiveResearchcriticaliCorrelationalResearchinvestigateswhetherchangesinonevariablearerelatedtochangesinanothervariableWhatistherelationshipbetweenbeingfromtheSouthandaggressivebehavior?CorrelationalResearchinvestigCorrelationalResearchcorrelationcoefficientsrangefrom+1.00to-1.00positive

correlation:increase/decreaseinthesamedirectionCorrelationalResearchcorrelatCorrelationalResearchCorrelationalResearchCorrelationalResearchcorrelationcoefficientsrangefrom+1.00to-1.00negativecorrelation:increase/decreaseinoppositedirectionsCorrelationalResearchcorrelatCorrelationalResearchCorrelationalResearchCorrelationalResearchcorrelationcoefficientsrangefrom+1.00to-1.00strengthoftherelationship:closenessto+1.00/-1.00,notbythevalence(+/-)Whichindicatesastrongercorrelation:-.74or+.21?CorrelationalResearchcorrelatCorrelationalResearchCorrelationalResearchCorrelationalResearchWhatistherelationshipbetweenbeingfromtheSouthandaggressivebehavior?Nisbett(1993)homiciderateand“southernness”: rs=.37inthesocialsciences(Cohen,1992)

r=.50(strong)r=.30(moderate)r=.10(small)CorrelationalResearchWhatisCorrelationalResearchstrengthsofthisapproachcanbearandomsampleactualbehaviorhasgoodgeneralizability(i.e.,externalvalidity)potentialfornumerousvariablesCorrelationalResearchstrengthCorrelationalResearchweaknessofthisapproachcannotinferacause-effectrelationshipfactorsforinferringcausalityassociation--yestemporalpriority--noruleoutaspuriousrelationship--noCorrelationalResearchweaknessCorrelationalResearchtemporalprioritysouthernnesshomicideratestimeCorrelationalResearchtemporalCorrelationalResearchtemporalprioritysouthernnesshomiciderateshomicideratessouthernnessortimetimeCorrelationalResearchtemporalCorrelationalResearchtemporalprioritydirectionofcausalityproblemsouthernnesshomiciderateshomicideratessouthernnessortimetimeCorrelationalResearchtemporalCorrelationalResearchruleoutaspuriousrelationshipsouthernnesshomicideratesanothervariablespuriousCorrelationalResearchruleoutCorrelationalResearchruleoutaspuriousrelationshipsouthernnesshomicideratespovertyrs=.38&.42(Nisbett,1993)CorrelationalResearchruleoutCorrelationalResearchruleoutaspuriousrelationship3rdvariableproblemsouthernnesshomicideratesanothervariablespuriousCorrelationalResearchruleoutCorrelationalResearchstrengthspotentialfornumerousvariablesgoodgeneralizabilityweaknessescannotmakecausalconclusionsdirectionofcausality3rdvariableproblemCorrelationalResearchstrengthAddressingProblemswithCorrelationalResearchdirectionofcausalityproblemDoes“southernness”leadtomoreaggressionordoesapropensityforaggressionleadtomore“southernness”?solution:wecause(i.e.,manipulate)oneofthevariablesinsultonegrouponSoutherners,butnotanotherAddressingProblemswithCorreAddressingProblemswithCorrelationalResearchdirectionofcausalityproblemifwecontrolwhoisinsulted,thenmeasureaggressiveness,weknowthedirectionofcausalitybut,westillhavethe3rdvariableproblempoorsocialskillsAddressingProblemswithCorreAddressingProblemswithCorrelationalResearchrandomassignmenttoconditionifsocialskillshaveaneffectonaggression,itshouldbeequalforbothgroupsWecanaddresstheproblemsofcorrelationalresearchbydoingexperiments.AddressingProblemswithCorreExperimentalDesigncharacteristicsmanipulationofavariablesolvesthedirectionofcausalityproblemrandomlyassigntoconditionssolvesthe3rdvariableproblemExperimentalDesigncharacterisExperimentalDesigntestingtheoryDoestheindependentvariablecausechangesinthedependentvariable?southernnessaggressioncauseExperimentalDesigntestingtheExperimentalDesigntheSoutherncultureofhonorhypothesisCohenetal.(1996)2(Southern/Northern)X2(insult/noinsult)ExperimentalDesigntheSoutherExperimentalDesignInsultNoInsultDistanceatwhichparticipantsgavewaytoconfederate(inches)ExperimentalDesignInsultNoInExperimentalDesignInsultNoInsultDistanceatwhichparticipantsgavewaytoconfederate(inches)ExperimentalDesignInsultNoInExperimentalDesignstrengthsallowsforcausalconclusionstobemade--besttestoftheoryweaknessesnotallquestionsareamenabletoexperimentsconcernsaboutgeneralizabilityExperimentalDesignstrengthsSummaryrefutation:processoftestingtheoriesdescriptiveresearchisinformative,butlimitedintheorytestingcorrelationalresearchismoreinformative,butdoesnotallowforcausalexplanationsexperimentsarethebesttestoftheoriesSummaryrefutation:processofIndependentVariableIndependentvariables

are

thosevariableswhichtheresearcherhascontroloverandwishestomanipulate.Forexample:price,framing,incentive,etc.IndependentVariableIndependenDependentVariablesDependentvariables

are

thosevariablesthatwehavelittleornodirectcontrolover,yetwehaveastronginterestin.Exampleswouldbeoutput,choice,etc.DependentVariablesDependentvCausalRelationshipIndependentvariableDependentvariable(e.g.,TobaccoTax)(Consumption)CausalRelationshipExtraneousVariablesExtraneousvariablesarethosevariablesthatmayhavesomeeffectonadependentvariableyetarenotindependentvariables.Extraneousvariablesmustbecontrolledthroughproperexperimentaldesign.ExtraneousVariablesExtraneousExperimentalDesignExperimentaldesign

isaprocedurefordevisinganexperimentalsettingsuchthatachangeinadependentvariablemaybeattributedsolelytothechangeinanindependentvariable.ExperimentalDesignExperimentaValidityInternalValidityExternalValidityValidityInternalValidityInternalValidityAretheobservedeffectsontheD.V.acauseofthetreatment?Orcouldtheyhavebeencausedbysomethingelse?InternalValidityAretheobserExternalValiditywhethertheresultfromaparticularstudycanbe:generalizedtothelargerpopulationofinterest.replicatedwithdifferentsubjects,researchsettings,andtimeintervals.generalizedtoamorenaturalenvironment.ExternalValiditywhethertherAdvantagesofLabExperimentsSubjectsarerandomlyassignedtothetreatmentconditions–rulesoutselectionbias.Itisknownwhichvariablesareexogenousandwhichareendogenous–allowstomakecausalinferences.Doesmoneycauseoutputordoesoutputcausemoney?Experimentercanmanipulatechangesintheexogenousvariables–allowsfortheisolationoftruecauses.Manyvariablesthatcannotbedirectlyobservedinthefieldcanbeobservedinthelab.

Replicability–providesthebasisforstatisticaltests.Criticscanruntheirownexperiments.AdvantagesofLabExperimentsSObjectionstoLabExperimentsExternalvalidity:Canwegeneralizeourinferencesfromthelabtothefield?ObjectionstoLabExperimentsEFormillenniathesunriseseverymorning.Yet,thisdoesnotallowyoutomaketheinferencethattomorrowmorningthesunwillriseagain.Nevertheless,almostallpeoplebelievethis.Thisconfidenceistheessenceofinduction.Ifmanyexperimentshaveshownthat–givenacertainsetofconditions–robustandreplicableregularitiesemerge,wecanhavefaiththatthesameregularitieswilloccurinrealitygiventhattheconditionsaremet.Therefore,anhonestscepticwhodoubtstheexternalvalidityofanexperiment,hastoarguethattheexperimentdoesnotcaptureimportantconditionsthatprevailinreality.FormillenniathesunrisesevObjections–lackofrealismLabexperimentsareunrealisticandartificial.Mosteconomicmodelsareunrealisticinthesensethattheyleaveoutmanyaspectsofreality.However,thesimplicityofamodeloranexperimentisoftenavirtuebecauseitenhancesourunderstandingoftheinteractionofrelevantvariables.Thisisparticularlytrueatthebeginningofaresearchprocess.Whetherrealismisimportantdependsonthepurposeoftheexperiment.Oftenthepurposeistotestatheoryorunderstandingthefailureofatheory.Thentheevidenceisimportantfortheorybuildingbutnotforadirectunderstandingofreality.Objections–lackofrealismLaOtherObjectionsParticipantsarejuststudents–lackofrepresentativityThestakesaresmallThenumberofparticipantsissmallParticipantsareinexperiencedResponsesTakeothersubjectpools(workers,soldiers,CEOs)Conductrepresentativeexperiments(Fehretal.2003)Increasethestakes(CameronEI1999,Slonim&RothEctra1997,Fehretal.2002).Effectsarestronger.Increasethenumberofparticipants(IsaacandWalker,J.Pub.E1994)Inviteexperiencedparticipants(Kagel&Levin,AER1986).Effectsstillholdforexperiencedparticipants.OtherObjectionsParticipantsaInvestmentBehaviorandEmotionMyopicLossAversion(BenartziandThaler1995).Peoplepreferbondsoverstocks.Shivetal.(2005),PsychologicalScienceAllparticipantsweregiven$20ofplaymoneyandtheywouldreceiveagiftcertificatefortheamounttheywereleftwithattheendofthestudy.InvestmentBehaviorandEmotioInvestmentBehaviorandEmotion20roundsofinvestmentdecision.Oneachround,theyhadtodecidetoinvest$1.00ornotinvest.Iftheydecidednottoinvest,theywouldkeepthedollar.Iftheydecidedtoinvest,theywouldhandoveradollarbilltotheexperimenter.Theexperimenterwouldtossacoin.Iftheoutcomewereheads(50%),theparticipantlose$1;ifthetossweretails,thentheywin$2.50.Threegroupsofsubjects:Normalpeople,patientswithbraindamagesrelatedtoprocessingofemotions,patientswithbraindamagesnotrelatedtoemotions.InvestmentBehaviorandEmotioResultsResultsTargetPatients–average$25.70Normalpeople-$22.80ControlPatients-$20.07TargetPatients–average$25.前景理論P(yáng)rospectTheory前景理論P(yáng)rospectTheory產(chǎn)生背景傳統(tǒng)經(jīng)濟(jì)學(xué)行為選擇理論–馮。諾伊曼和摩根斯頓的最大期望效用理論(ExpectedUtilityTheory–VNM)EU=Σpiu(xi)u(xi)為效用函數(shù)，pi為事件發(fā)生的客觀概率

RiskAversion(Concave)產(chǎn)生背景傳統(tǒng)經(jīng)濟(jì)學(xué)行為選擇理論–馮。諾伊曼和摩根斯頓的偏好公理占優(yōu)性:決策者永遠(yuǎn)都會(huì)選擇那些具有優(yōu)勢(shì)的策略。傳遞性:A>B，B>C，所以A>C獨(dú)立性:決策者對(duì)兩個(gè)策略的選擇不受第三個(gè)不相干的策略影響。完全性:X優(yōu)于Y或劣于Y或一樣。偏好公理占優(yōu)性:決策者永遠(yuǎn)都會(huì)選擇那些具有優(yōu)勢(shì)的策略。阿萊斯悖論1988年諾貝爾經(jīng)濟(jì)學(xué)獎(jiǎng)得主ALLAIS。1。下面兩個(gè)選擇，您更喜歡哪個(gè)呀？選擇A：100%機(jī)會(huì)贏