Modelling structural change using the Kalman Filter

Abstract

Structural change is endemic in the Eastern European economies and the newly emerging Commonwealth of Independent States, yet conventional econometric modelling techniques proceed under the assumption that there is a structurally stable ‘true’ economy to be discovered. This paper explores the consequences of endemic structural change for econometric modelling by considering the model reduction problem when the data generation process is itself undergoing structural change. The resultant econometric model, it is argued will generally exhibit time varying parameters where much of the structural change is reflected in the changing parameters. The use of Kalman Filters to estimate such changing parameters is then discussed and a range of specifications which allow the inclusion of different forms of identifying information is given. The paper then illustrates these ideas by modelling the determination of the black market exchange rate in Poland over the period from the mid 1970s to the early 1990s.