Testing for Non-Linear Dependence in Univariate Time Series An Empirical Investigation of the Austrian Unemployment Rate

In recent years interest has been growing in testing for stochastic non-linearity in macroeconomic time series. There are several inference procedures available. But not much is known about their behaviour on real world small-sized settings. This paper surveys some of these tests. Their performance is compared using monthly Austrian unemployment data that cover the period January 1960 to December 1997. It is found that the test procedures surveyed are complementary rather than competing. Several useful guidelines are provided for applying the increasingly complex test procedures in practice.