Estimation and Empirical Testing of the putty-putty model

Abstract

In this chapter results will be presented of the empirical testing of the putty-putty model of chapter 3 on data from U.S. manufacturing and the Dutch manufacturing and total business sectors. The use of several samples to test a model increases the range of independent variation of the observations and thereby offers a partial compensation for the lack of experimental testing possibilities (see also chapter 1.2.3). The model presented in chapter 3 is formulated in continuous time and is therefore not directly applicable to discrete data. In section 5.1 of this chapter a discrete-time version of the model will be offered for this purpose. Section 5.2 discusses the assumptions about the firm’s information processing and expectation formation. Section 5.3 describes the (simple) stochastic structure imposed on the discrete-time model and presents the resulting parameter estimates for the different data sets. The performance of the model is further investigated in section 5.4 by means of an analysis of the final form. Data sources and measurement problems are discussed in appendices 5a and 5b.