Affordable Access

Parameter identification in Dynamic Economic models

Authors
Disciplines
  • Economics
  • Mathematics

Abstract

Parameter identification in dynamic economic models Articles | Autumn 2010 Economic Bulletin | Banco de Portugal 169 PARAMETER IDENTIFICATION IN DYNAMIC ECONOMIC MODELS* Nikolay Iskrev** 1. INTRODUCTION Parameter identifi cation is a concept which every student of economics learns in their introductory econometrics class. The usual textbook treatment of identifi cation leads one to think of identifi cation as a technical issue relevant only to empirical work, and to regard identifi cation problems as caused by either defi ciencies of the available data, or of the statistical methodology used to estimate the models. In this note I will argue that the analysis of identifi cation has an important economic modeling aspect, and that it may be very useful to researchers who are not interested in estimation. I will focus the discussion on the class of dynamic stochastic general equilibrium (DSGE) models which have become one of the main analytical tools of modern macroeconomics. The essence of my argument is that when the economic model supplies a complete characterization of the data generating process, parameter identifi cation may be treated as a property of the underlying theoretical model. Parameters will be unidentifi able or weakly identifi ed if the economic features they represent have no empirical relevance at all, or very little of it. This may occur either because those features are unimportant on their own, or because they are redundant given the other features represented in the model. These issues are particularly relevant to DSGE models, which are sometimes criticized of being too rich in features, and possibly overparameterized (Chari, Kehoe, and McGrattan, 2009). A second reason why it is important to study identifi cation is its econometric implications. The reliable estimation of a model is impossible unless its parameters are well identifi ed. Again, this is crucial for DSGE models as their use for quantitative

There are no comments yet on this publication. Be the first to share your thoughts.