Theoretical predictions of the effect of TFP growth on employment are ambiguous, and depend on the extent to which new technology is embodied in new jobs. We estimate a model for employment, wages and investment with an annual panel for the United States, Japan and Europe and find that TFP growth increases employment. For the United States TFP growth explains the trend change in unemployment. We evaluate the model and find that creative destruction plays no part in aggregate unemployment dynamics. The model can explain up to half of the estimated impact of growth on unemployment.