An investigation into a non-linear stochastic trend model

The paper explores the empirical properties of a non-linear stochastic trend model which can be viewed as an intermediate case between a linear and a log-linear trend model. I assess the small sample distribution of the ML estimator by Monte Carlo simulations and use it to model some typical macroeconomic time series. The non-linear trend model turns out to be an important tool which warrants further analysis. I also compute impulse response functions and compare them with those obtained from a conventional linear model.