Fables of faubus?: Testing the sectoral shift hypothesis in the Netherlands using a simplified Kalman filter model

Abstract

The presence of structural breaks can seriously affect the outcome of standard regression methods like OLS. Although there are many methods available to deal with them, we focus here on a particular linear filtering method, namely the Kalman Filter. Its results vis a vis a regular OLS approach are illustrated by testing the sectoral shift hypothesis in the Netherlands. Although a rather simplified version of the Kalman Filter is used, it turns out to be a sufficient enough approximation. What we find, is that the variables capturing the sectoral shift hypothesis are the most important in explaining Dutch unemployment behaviour during the postwar period. Thus, the hypo-thesis is endorsed. On the other hand, our highly significant constant term indicates that the inclusion of other variables affecting unemploy-ment may alter the results. Our conclusion thus is a tentative one.