Forecasting the Stage of the Business Cycle

Ullrich Heilemann and Heinz Josef Münch
Rheinisch-Westfälisches Institut für Wirtschaftsforschung and Universität Duisburg
Muench@rwi-essen.de

Abstract

Macroeconomic forecasts are traditionally stated as point estimates for one or several periods ahead. For many users, however, this approach may not correspond with their own uses and evaluations of macro forecasts. For them it may be more important to know when the economy is entering a recession or nearing the upper turning point (Webb 1994). To answer such questions requires, first, a sufficient precise and stable classification of the macroeconomic development and, second, adequate macroeconomic forecasts. The present paper examines the first requirement by overhauling a four stage classification scheme (upswing, upper turning point phase, downswing, lower turning point phase) of the West German economy based on a multivariate discriminant scheme developed by Meyer/Weinberg (1975) in the early 1970s for the U.S. and for West Germany, re-estimated and modified by Heilemann/Münch (1995). The stability of the classification scheme will be examined by extending the sample period of the classification scheme backwards from 1960 to 1950. The cyclical quality of present macroeconomic forecasts, including a macroeconometric model forecast (RWI-business cycle model) are tested by classifying them with the help of the classification functions derived from the discriminant analysis. As not all macroeconomic forecasts include all variables of the classification scheme, some modifications of the original functions will become necessary. While in a first round of tests the forecasts are deterministically interpreted, in a second round a stochastic view will be tried.

The results of the paper should, first, reveal constancy or change of the West German/German business cycle and its phases, lengths and its forming elements/variables (Gordon (ed.) 1986). The paper should, second, show whether German macroeconomic forecasts give a coherent picture of the various cycle stages and where improvement is needed or where the picture is spoiled by exceptional data or data constellations.