欢迎来到天天文库
浏览记录
ID:59563379
大小:142.50 KB
页数:19页
时间:2020-11-11
《在计量经济学的Endogeneity : 同时等式模型(1)培训讲学.ppt》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、在计量经济学的Endogeneity:同时等式模型(1)THENATUREOFSIMULTANEOUSEQUATIONSMODELSConceptsEndogenousvariablesExogenousvariablesStructuralerrorsEachequationshouldhaveabehavioral,ceterisparibusinterpretationonitsown.SIMULTANEITYBIASINOLSsimultaneitybiasIDENTIFYINGANDESTIMATINGASTRUCTURALEQU
2、ATIONIdentificationinaTwo-EquationSystemThefirstissupplyequation.Thesecondisdemandequation.Ageneraltwo-equationmodelExclusionrestrictionsInotherwords,weassumethatcertainexogenousvariablesdonotappearinthefirstequationandothersareabsentfromthesecondequation.RANKCONDITIONFORID
3、ENTIFICATIONOFASTRUCTURALEQUATIONThefirstequationinatwo-equationsimultaneousequationsmodelisidentifiedifandonlyifthesecondequationcontainsatleastoneexogenousvariable(withanonzerocoefficient)thatisexcludedfromthefirstequation.Theorderconditionisnecessaryfortherankcondition.T
4、herankconditionrequiresmore:atleastoneoftheexogenousvariablesexcludedfromthefirstequationmusthaveanonzeropopulationcoefficientinthesecondequation.Estimationby2SLSThisissimilarinthepreviouschapter.SYSTEMSWITHMORETHANTWOEQUATIONSORDERCONDITIONFORIDENTIFICATIONAnequationinanyS
5、EMsatisfiestheorderconditionforidentificationifthenumberofexcludedexogenousvariablesfromtheequationisatleastaslargeasthenumberofright-handsideendogenousvariables.Overidentifiedequation,justidentifiedequation,andunidentifiedequation.Estimation2SLS3SLS–Moreefficientusingthere
6、sidualstodoGLSestimationatthethirdstage.Moresensitivetovariableselection.SIMULTANEOUSEQUATIONSMODELSWITHPANELDATA(1)eliminatetheunobservedeffectsfromtheequationsofinterestusingthefixedeffectstransformationorfirstdifferencing;(2)findinstrumentalvariablesfortheendogenousvaria
7、blesinthetransformedequation.Challenging,becauseforaconvincinganalysisweneedtofindinstrumentsthatchangeovertime.Howtodoifyouwanttoidentifythefixed-effects?Puttimedummiesandfixed-effectsdummiesinthestructuralmodel.Theend.此课件下载可自行编辑修改,仅供参考!感谢您的支持,我们努力做得更好!谢谢
此文档下载收益归作者所有