This paper examines the cyclical relationship between exports and outp
ut in the U.K. using structural time-series modelling. The time series
on exports and output over the period 1885-1993 are decomposed, and t
he cyclical components are extracted. Causality testing as applied to
the cyclical components reveals the absence of causality from exports
to output except in the post-1945 period. The results of causality tes
ting are confirmed by the results of estimating a structural time-seri
es model in which exports appear as an explanatory state variable. Thr
ee different explanations are suggested for the change in this relatio
nship during the most recent period.