金融计量学习题及习题答案-上海财经大学

1诚实考试吾心不虚，公平竞争方显实力，考试失败尚有机会，考试舞弊前功尽弃。上海财经大学《FinancialEconometrics》课程考试卷一课程代码课程序号姓名学号班级题号一二三四五六七八九十总分得分Part1TermExplanation（20marks）1．WhiteNoise2．RandomWalk3．AkaikeInformationCriterion4．Jarque-BeraStatistic5．ChowTestImportantPoint：1．WhiteNoise：WhiteNoiseisthespecialcaseofstationarystochasticprocess.Wecallastochasticprocesspurelyrandomorwhitenoiseifithaszeromean,constantvarianceandisseriallyuncorrelated.2．RandomWalk:Randomwalkmeansthatthestochasticprocessisnonstationaryandvalueofthisperiodishighlyrelatedtothepastvalues.Forexample,thestockpricetodaymayequaltheyesterday’spriceplusarandomshock.Randomwalkwithoutdriftcanbeexpressedastttuyy13．AkaikeInformationCriterion:AICprovideawaytoselectthebetterregressionmodelamongseveralmodelsbycomparingtheirforecastperformance.ThelowertheAIC,thebettertheforecastperformancewillbe.AICwillalsobeusedtodeterminethelaglengthinARDLapproach.4．Jarque-BeraStatistic:TheJarque-Beratestisthetestofnormality.Wefirstcalculatetheskewnessandthekurtosis,anditisalsobasedontheresidualoftheregression.TheJarque-BeraStatistic=)24)3(6(22KSn,whereSistheskewnessandKisthekurtosis,2nissamplesize,andfornormaldistribution,S=0,K=3,ifJBstatisticisnotsignificantlydifferentfromzero,pvalueisquitelow,werejectthenullhypothesisthattheresidualisnormallydistributed.5．ChowTest:Thetestofstructuralchangeoftheregression.Theestimateoftheparameteroftheregressionmaynotretainthesamethroughtheentiretimeperiod;weusetheChowtesttotestwhethertherelationshipisstableandfindthebreakpoint.ItdeveloptheFstatistics=)/(/)(kNRSSmRSSRSSururr,thenullhypothesisistheregressionisstable.Part2Explainmainpurpose(s)ofconstructingfollowingtwomodelsandmakingcommentsontheempiricalresults.(25marks)1.GregoryChow(1966)whereM=naturallogarithmoftotalmoneystockYp=naturallogarithmofpermanentincomeY=naturallogarithmofcurrentincomeR=naturallogarithmofrateofinterest2.TaylorandNewhouse(1969)本题答题要点：1。模型一的建模理念是准备检验凯恩斯和弗里德曼德货币需求理论，根据t统计量的结果，模型并没有能明确说明支持或拒绝哪种理论；10分2。模型二也是准备检验凯恩斯和弗里德曼德货币需求理论，根据t统计量的结果，模型明确说明支持凯恩斯货币需求理论并拒绝弗里德曼德货币需求理论；10分3。模型二的特色之一是引入因变量M的前一期1tM做为自变量；2分4。两个模型都存在伪回归的嫌疑。2分5。专业英语词汇表达错误将被少量扣分。1分(.0540)(.13897)(.148)9965.7476.01321.069.11365.ˆ2RRYYMttptt(.0669)(.0597)(.0940)(.14284)9988.5878.3325.3274.06158.3067.ˆ21RMRYYMtttptt3Part3Explanationsoffourdiagnostictestsandmakingcommentsontheempiricalresultsoffollowingtwomodelsonthebasisofdiagnostictests.（25marks）4本题答题要点：1．分别对四个诊断检验予以说明12分2．分别对两图实证结果予以判读8分3．说明与图二相比，图一错误的根源在于将非线性的柯布－道格拉斯生产函数直接线性拟合4分4．专业英语词汇表达错误将被少量扣分1分Part4Prof.MiltonFriedmanarguedthattherewasapositiveassociationbetweeninflationandmoneysupply.PleaseexaminethisargumentusingECMandGrangerCausalitytest.（30marks）1．对取过对数的变量进行平稳性检验（说明在何种显著性水平条件下的判断）5分2．对做过差分的变量进行平稳性检验（同阶单整）5分3．协整检验（包括协整的意义）5分4．ECM模型的构造和解释7分5．Granger因果检验（要求说明检验阶数的选择）7分6．专业英语词汇表达错误将被少量扣分1分5诚实考试吾心不虚，公平竞争方显实力，考试失败尚有机会，考试舞弊前功尽弃。上海财经大学《FinancialEconometrics》课程考试卷二课程代码课程序号姓名学号班级题号一二三四五六七八九十总分得分Part1Termexplanation（20marks）1．SpuriousregressionRegressionofonetimeseriesvariableononeormoretimeseriesvariablesoftencangivenonsensicalorspuriousresults.Spuriousregressionoftenshowsasignificantrelationshipbetweenvariables,butinfact,thiskindofrelationshipdoesnotexist.2．QStatisticQstatisticisoneofthetestsofnon-stationary.testingthejointhypothesisthatalltheuptocertainlagsaresimultaneouslyequaltozero.3．Durbin-WatsonStatisticTheDurbin-Watsonstatisticisatestforfirst-orderserialcorrelation.Moreformally,theDWstatisticmeasuresthelinearassociationbetweenadjacentresidualsfromaregressionmodel.TheDurbin-Watsonisatestofthehypothesis0Ifthereisnoserialcorrelation,theDWstatisticwillbearound2.4．SchwarzCriterionItisdefinedas:nRSSnSICnk,imposingapenaltyforaddingregressorstothemodel..Part2Explainthemainpurpose(s)ofconstructingthefollowingtwomodelsandmakingcommentsontheempiricalresults(25marks)0:0,0kHk221ˆ~()mkmkQnkk61．GregoryChow(1966)whereM=naturallogarithmoftotalmoneystockYp=naturallogarithmofpermanentincomeY=naturallogarithmofcurrentincomeR=naturallogarithmofrateofinterest2．TaylorandNewhouse(1969)本题答题要点：1。模型一的建模理念是准备检验凯恩斯和弗里德曼德货币需求理论，根据t统计量的结果，模型并没有能明确说明支持或拒绝哪种理论；10分2。模型二也是准备检验凯恩斯和弗里德曼德货币需求理论，根据t统计量的结果，模型明确说明支持凯恩斯货币需求理论并拒绝弗里德曼德货币需求理论；10分3。模型二的特色之一是引入因变量M的前一期1tM做为自变量；2分4。两个模型都存在伪回归的嫌疑。2分5。专业英语词汇表达错误将被少量扣分。1分Part3.1MakecommentsonthefollowingtwosharepriceindexesusingdescriptivestatisticsincludingJaque-Berastatistic.SHAstandsforShanghaistockmarketsharepriceindexandSZAstandsforShenzhenstockmarketsharepriceindex（10marks）05010015020040060080010001200140016001800Series:SHASample2/01/199231/12/1999Observations2087Mean988.9936Median958.9130Maximum1842.610Minimum293.7500Std.Dev.351.0635Skewness0.103000Kurtosis2.103348Jarque-Bera73.60335Probability0.000000(.0540)(.13897)(.148)9965.7476.01321.069.11365.ˆ2RRYYMttptt(.0669)(.0597)(.0940)(.14284)9988.5878.3325.3274.06158.3067.ˆ21RMRYYMtttptt7050100150200250300100150200250300350400450500550Series:SZASample5/10/199231/12/1999Observations1890Mean295.8065Median296.3310Maximum561.5640Minimum95.26100Std.Dev.125.4013Skewness-0.022203Kurtosis1.645256