A New Test of Asymmetric Stationarity in the Presence of Deterministic Trends: Simulation and Empirical Evidence

Using local-to-unity detrending, a modified momentum-threshold autoregressive test is derived to allow the unit root hypothesis to be tested against an alternative of asymmetric stationarity about a deterministic trend. Monte Carlo evidence is presented to show the increased power of the proposed test in the presence of asymmetric adjustment relative to the familiar Dickey-Fuller (1979) test and the momentum-threshold autoregressive test of Enders and Granger (1998). The empirical relevance of the test is illustrated via an application to New Zealand national output over the period 1870-2001, where in contrast to findings obtained using alternative unit root tests, the unit root hypothesis is conclusively rejected.