In Discussion Paper No. 1336, Research Fellow Dan Ben-David, Robin Lumsdaine and David Papell focus on the question of whether or not gross domestic products are characterised by a unit root. As a number of earlier studies have pointed out, the existence of a unit root implies that output shocks are persistent – an apparent contradiction to many current macroeconomic models. For example, any negative shock to the economy in the traditional, exogenous growth model will be followed by a transitional period of faster growth until the economy once again returns to its original steady-state path. A unit root in output contradicts the mean-reverting, or stationarity, predictions of this neoclassical growth model. While increasing the span of the data increases the power of the unit root test, it also increases the likelihood that the data is characterized by a structural break.

This paper shows that the inclusion of a second break makes it even less likely that output is characterised by a unit root, with a rejection of the unit root null in three quarters of the cases. Most of the trend breaks are associated with a change in output levels. As the neoclassical growth model predicts, the magnitude of these level changes is shown here to be related to changes in growth rates during the period following the break.