Testing for a Unit Root against Nonlinear STAR Models

In this paper we propose a simple testing procedure to detect the presence of nonstationarity against nonlinear stationarity based on the Smooth Transition Autoregressive modelling approach. We provide an advance over the existing literature in three senses: first, we derive the limiting nonstandard distribution of the proposed NLADF tests; second, we establish the superior power performance of the NLADF test over the standard linear ADF test under the alternative of nonlinear stationarity via Monte Carlo simulation exercises; third we provide an application to ex post real interest rates from six major OECD countries, and find the NLADF test is able to reject a unit root in cases, whereas the linear ADF tests fail.

This paper is unavailable for download.