Title of article :
Estimation ofDSGEmodelswhenthedataarepersistent
Author/Authors :
Yuriy Gorodnichenko، نويسنده ,
Issue Information :
روزنامه با شماره پیاپی سال 2010
Abstract :
Dynamicstochasticgeneralequilibrium(DSGE)modelsareoftensolvedandestimated
under specificassumptionsastowhethertheexogenousvariablesaredifferenceor
trendstationary.However,evenmilddeparturesofthedatageneratingprocessfrom
these assumptionscanseverelybiastheestimatesofthemodelparameters.Thispaper
proposesnewestimatorsthatdonotrequireresearcherstotakeastandonwhether
shocks havepermanentortransitoryeffects.Theseprocedureshavetwokeyfeatures.
First, thesamefilterisappliedtoboththedataandthemodelvariables.Second,the
filteredvariablesarestationarywhenevaluatedatthetrueparametervector.The
estimatorsareapproximatelynormallydistributednotonlywhentheshocksaremildly
persistent,butalsowhentheyhavenearorexactunitroots.Simulationsshowthat
these robustestimatorsperformwellespeciallywhentheshocksarehighlypersistent
yet stationary.Insuchcases,lineardetrendingandfirstdifferencingareshowntoyield
biasedorimpreciseestimates.
Keywords :
Persistent dataFiltersTrendsUnit rootSpurious estimatesBusiness cycles
Journal title :
Journal of Monetary Economics
Journal title :
Journal of Monetary Economics