Throwing Out the Baby with the Bath Water: A Comment on Green, Yoon and Kim

Green, Yoon and Kim (hereinafter GYK) contribute to the literature on estimatingpooled times-series{cross-section (hereinafter TSCS) models in International Relations(hereinafter IR). They argue that such models should be estimated with flxed efiectswhen such efiects are statistically necessary. While we obviously have no disagreementthatsometimesflxedefiectsareappropriate,weshowinthisresponsethatflxedefiectsareperniciousforIRTSCSmodelswithabinarydependentvariable(hereinafterBTSCSmodels)andthattheyareoftenproblematicforIRmodelswithacontinuousdependentvariable. Inthebinarycase,thisperniciousnessisduetomanypairsofnationsalwaysbeingscoredzero,andhencehavingnoimpactontheparameterestimates;forexample,manydyadsnevercomeintocon°ict. Inthecontinuouscase,flxedefiectsareproblematicinthepresenceofthetemporalstableregressorsthatarecommonIRapplications,suchasthedyadicdemocracymeasuresusedbyGYK.