INFREQUENT PERMANENT SHOCKS AND THE UNIT TOOT IN QUARTERLY UK OUTPUT.

This article discusses whether quarterly post-Second World War Great Britain output is difference or trend stationary, i.e., whether output contains a unit root or not. This paper represents a further attempt to evaluate the robustness of the random walk hypothesis for Great Britain output. It could be argued, nonetheless, that any conclusions concerning the presence of a unit root in output should be tempered by the known low power of unit root tests to discriminate between random walk and deterministic trend models. An alternative approach to testing between the two models has recently been suggested by economist P. Kwiatkowski et al., in which they develop a Lagrange Multiplier test of the null hypothesis of trend stationarity against the alternative of a random walk. Computing their test statistic for the output series enables the authors to reject trend stationarity at almost the 1 per cent level. In the light of this result, the authors feel encouraged to maintain that detrended output does contain a unit root and, indeed, still appears to follow a random walk.