Edited by John B. Davis and D. Wade Hands
Chapter 17: On the Role of Theory and Evidence in Macroeconomics
17 On the role of theory and evidence in macroeconomics Katarina Juselius1 17.1 INTRODUCTION Economists frequently formulate an economically well-specified model as the empirical model and apply statistical methods to estimate its parameters. In contrast, statisticians might formulate a statistically well-specified model for the data and analyze the statistical model to answer the economic questions of interest. In the first case, statistics are used passively as a tool to get some desired estimates, and in the second case, the statistical model is taken seriously and used actively as a means of analyzing the underlying generating process of the phenomenon in question. The general principle of analyzing statistical models instead of applying methods can be traced back to R.A. Fisher. It was introduced into econometrics by Trygve Haavelmo (1944) and operationalized and further developed by Hendry and Mizon (1993), Hendry (1987), Johansen (1995), Juselius (2006), and followers. Haavelmo’s influence on modern econometrics has been discussed, for example, in Hendry et al. (1989) and Andersen (1991). Because few observed macroeconomic variables can be assumed to be fixed or predetermined a priori, Haavelmo’s approach to econometrics requires a probability formulation of the full process that generated the data. Thus, the statistical model has to be based on a full system of equations. The computational complexities involved in the solution of such a system were clearly prohibitive at the time of Haavelmo’s monograph when even the estimation of a multiple regression was a non-trivial task. In today’s computerized world, it is certainly technically feasible...
You are not authenticated to view the full text of this chapter or article.