2020-09-26T15:03:54Zhttps://repository.dl.itc.u-tokyo.ac.jp/?action=repository_oaipmhoai:repository.dl.itc.u-tokyo.ac.jp:000420602020-09-08T01:26:09Z00062:07433:0743400009:07435:07436
Improving the Rank-Adjusted Anderson-Rubin Test with Many Instruments and Persistent HeteroscedasticityengStructural EquationLikelihood Ratio CriterionRAAR testModified RAAR test Many InstrumentsPersistent Heteroscedasticityhttp://hdl.handle.net/2261/18461Technical ReportKunitomo, NaotoYukitoshi, MatsushitaAnderson and Kunitomo (2007) have developed the likelihood ratio criterion, which is called the Rank-Adjusted Anderson-Rubin (RAAR) test, for testing the coefficients of a structural equation in a system of simultaneous equations in econometrics against the alternative hypothesis that the equation of interest is identified. It is related to the statistic originally proposed by Anderson and Rubin (1949, 1950), and also to the test procedures by Kleibergen (2002) and Moreira (2003). We propose a modified procedure of RAAR test, which is suitable for the cases when there are many instruments and the disturbances have persistent heteroscedasticities本文フィルはリンク先を参照のことDiscussion paper series. CIRJE-FCIRJE-F-5882008-09AA11450569application/pdf335日本経済国際共同センターhttp://www.cirje.e.u-tokyo.ac.jp/research/dp/2008/2008cf588ab.html2017-06-16