Long-run implications of neoclassical growth models: empirical evidence from Australia, New Zealand, South Korea and Taiwan

Long-run implications of the one-factor neoclassical growth models are tested by using data from Australia, New Zealand, South Korea and Taiwan. Testable propositions about the long-run relations between log of real output, consumption, and investment are investigated using unit root tests and cointegration analysis. Also, common trends analysis is used to identify the sources of the permanent shocks driving the common long-run trend. Further, variance decomposition is used to examine the extent to which innovations to the common trend component are able to explain short-run fluctuations in the data. Empirical results provide limited support for the neoclassical models. In addition to productivity shocks, permanent innovations to investment are found to be important components of the long-run stochastic trend governing the dynamic behaviour of real output, consumption and investment. Variance decomposition results suggest that although innovations to common trend components are important, this is unable to explain a substantial fraction of short-run fluctuations in the data.