2008216124054114拿起计量经济的武器

拿起计量经济的武器Econometrics:YourWeapon李春涛SchoolofEconomicsandFinance,theUniversityofHongKongEmail:[e-mail]计量经济学家林少宫1944年毕业于中央大学经济系1947年到美国，师从Herwicz和Modigliarni伊利诺伊大学获得经济学博士学位任教于俄亥俄州立大学1954年，他谢绝了当时统计学界、经济学界、数学界公认的大师霍特林(Hotelling)教授的工作邀请，放弃了国外优越的生活条件，毅然返回祖国，向教育部报到计量经济学家林少宫现代经济学一直被视为是“庸俗”的、为资本主义制度进行诡辩的资产阶级学科，经济学出身的林老师回国后不能教现代经济学，也就只好被分配到华中工学院数学教研室从事数理统计的教学和研究工作其间钱学森教授曾力邀林老师去北京工作，被爱才如命的教育家朱九思老校长强留下来田国强、艾春荣、时寿勇、宋敏、谭国富、徐殿庆、方振民、杨小凯、文梅、刘安平、黄晓东等SchoolofInformationWhatisinformation？Insiderinformation?李国宝的朋友梁启雄的女儿梁家安和丈夫王勤兢在4月13日至30日期间购买41.5万股道琼斯公司股份，价值1500万美元。新闻集团收购意图公开后，道琼斯股票价格急升58%，王勤兢夫妇所掌握的股票价值升至2300万美元，利润高达818万美元。Rumors？Fortuneorsin?April19,2007Stockmarketreactionbeforemacro-datareleaseThe(trilogy)filmMatrixInformationisdatacollectionanditsprocessingDataminingAttributesofInformationMathematicalmodelingDatacollectionComputerscienceDatamining(EconometricsMethods)Whattheschoolhasmissed?MorefocusonEconometrics?Aeconometricsdepartment?YourMeritsSolidbackgroundinMathematicsTrainingsinStatisticsandProbabilitytheoryIntensiveSkillsinComputerProgrammingSomebasicknowledgeofEconomicsWhatthejobmarketisrecruiting？QuantitativeanalysisforfinancialindustryEmpiricalresearchersforsocialscienceOpportunitiesforyouFinancialInstitutionsResearchAssistantGraduateProgram,domesticandabroadOsamabinLadenRecommendationsStatisticsEconometricsEconomicsTheorySoftwareDatasourcesOutlineWhatisEconometrics?Wheretouseit?ClassicalEconometricsCasesSoftwareIntroductionDatabaseswithinarm’slengthSuggestionsforyourthesisTheEmergingofEconometricsIdealTheoreticalModels:tooidealBasedonidealhypothesisToosimplecomparedwithrealworldMulti-factorinCausalityOwnershipmayaffectfirmperformanceFirmsize,legalenvironmentetcmayalsohaveeffectsModelscannotbeinclusiveenoughTwo-DirectionCausalityComplexityinDataContinuousmeasureddataDiscretedataCross-sectiondataTime-seriesdataPaneldataWhatisEconometrics?Econometricsisasetofquantitativetechniquesthatareusefulformaking“economicdecisions”economicmeasurementItisthebranchofeconomicsthatappliesstatisticalmethodsandmathematicalmodelstotheempiricalstudyofeconomictheoriesandrelationshipsInterdisciplinaryMathEconStatisticsMath.Econ.Math.Stat.Econ.Stat.EconometricsBacktoHerwiczEg1：Y—SavingX—IncomeuXYob)(PrTypicalEconometricsQuestionsEg2：Y—VoteforDemarcatesX—SocialEconomicsCharacteristicsuXYWidelyUsedinSocialScienceCliometricstherailwaysintheAmericaneconomyNorth,thecivilwarandtheendoftheNegroSlaveryeconomichistoryLawResearch中国“命案必破”是否制造了冤假错案？2004年2月开始，公安部向各地方公安局下达命令，要求公安部门必须破获一切命案这个命令没有科学依据警察不是超人出动大量的人力物力，造成破案成本上升制造冤假错案比如，警察找到一个死刑犯，这个人要死了，多顶一个命案的成本不大，于是就会顶替案件，让真正的罪犯逍遥法外研究方法：收集某一地区2004年2月前后各1年的涉命案例每个死刑犯背负的命案数量为y每个刑事犯罪分子的基本特征：年龄、性别、教育程度、职业、收入状况等时间变量T=1如果刑犯的审判日期在2004年2月以后，否则为0ModelTFarmerTAgeTEduFarmerAgeEduTY**********76543210如果beta1显著大于0，则说明命令下达以后，平均每个死刑犯的命案数量增加了，这是冤假错案的信号如果beta7显著大于0,说明命令下达以后,出身为农民的刑事犯罪分子，更可能被迫顶替命案FinanceOwnershipStructureandFirmPerformanceEfficientMarketHypothesisEventStudyCorporateR&DandCorporateGovernanceLawandFinancePrincipal-AgentProblemSeparationbetweenOwnershipandControlOwnersdonotmanagethefirmsdirectlyProfessionalManagersareagentforownersWithInformationAsymmetryandalackingofeffectivemonitoring,managerswillpursuitselfinterestattheexpenseofshareholdersLuxuryofficeandHotelBeautifulSectariesBusinessclassTunnelingSolution：InterestAlignment理论一：给经理人的股权越多，经理人的利益与股东的利益越一致(interestalignmenthypothesis)理论二：经理人的股权增加了，经理人的位置就稳定了，不怕失业了，不怕被人收购了，所以就不好好工作了（marketentrenchmenthypothesis）公司委托代理理论两个理论给出了相反的结论两个因素可能都存在，但是到底哪一个占优呢用实证研究来检验：收集几千家公司的绩效指标和股权结构指标，用股权结构指标对企业绩效作回归，看系数是否为正AccountingandAuditingDeterminantsofAuditOpinionEarningsInformativenessandMarketReactionEarningsmanagementandfirmAttributesEarningsinformativeness:EvidencefromChinaCorporatedisclosetheirearningsintheirannualreportsOnthedisclosingdate,marketreactdifferentlyfordifferentearningsanddifferentfirmsGenerallyspeaking,thehighertheearnings,thehigherthestockpricejumpsEarningsinformativeness:OtherfactorsmayalsoaffectmarketreactionsLargefirmsreactlessRiskyfirmsreactlessstrongEtcEarningsinformativeness:ThisresearchtesttherelationshipbetweenmarketreactionandauditfirmsizeOurtheoryisthat,largeauditfirmsareofhigherquality,thustheearningsaremorereliableMarketreactstrongerforreliableearningsThusmarketreactionispositivelycorrelatedwithauditfirmsizeMeasureofMarketreactionEventstudyFirmireleasedit’searningsondate0tdateonreturnsistockisRtdateonreturnindexmarketisRitmt'Agenda22titARCARRegressionResults(1)(2)(3)CAR[-2,2]CAR[-2,2]CAR[-2,2]UE3.3421***3.1627***2.4263*(2.953)(2.706)(1.931)UE*ClientSales0.0866**(2.289)UE*ClientAsset0.0980**(2.222)UE*Revenue0.1253***(2.730)UE*Joint0.3352**0.3274**0.3266**(2.394)(2.313)(2.342)UE*ClientSize-0.2069***-0.2079***-0.1975***(5.033)(5.045)(4.873)UE*Leverage0.01410.02230.0206(0.371)(0.592)(0.547)UE*Q-0.0461**-0.0489**-0.0499**(1.991)(2.107)(2.148)UE*SD-16.4964***-16.4612***-16.4708***(3.949)(3.940)(3.945)UE*Lo