Rival Macroeconomic Models And Australian Stylised Facts

This paper reports the stylised facts resulting from the tests of rival macroeconomic models in explaining the Australian business cycle during the sample period 1966(3)-1995(3). The dominant rival paradigms such as the New Classical, Keynesian the Real Business Cycle theories have been tested using both Granger causality and non-nested testing techniques. The time-series data used for modelling the rival paradigms were processed using unit root and cointegration econometrics to guard against possible spurious regression inferences due to nonstationarity in the data. Parsimonious data congruent models for testing the rival paradigms were derived by the application of the general-to-specific methodology. The problem of non-spherical errors created by the use of generated regressors in the specification of business cycle models was tackled by replacing ordinary least squares by generalised least squares estimates. The empirical results supported the conclusion that hybrid macroeconomic paradigms encompassing both demand and supply side shocks provide more plausible explanations of the Australian business cycle than tests narrowly focussed only on demand side shocks. The study results challenges the narrow view that rival macroeconomic theories would have failed to provide meaningful guidelines to Australian policymakers to implement counter-cyclical policies during the study period.