This paper examines the sensitivity of Total Factor Productivity (TFP) estimates to alternative concepts of output and methods of estimation. On theoretical grounds, gross output is the superior concept to real output and econometric estimates of TFP growth are to be preferred to growth-accounting estimates under general assumptions. A panel regression on UK manufacturing industries reveals that controlling for market power and capacity utilisation adjustments, growth-accounting estimates of productivity growth are generally biased. The econometric results suggest that the recovery in productivity rates experienced in the 1980s by the UK manufacturing sectors and the slowdown in the 1990s was not as large as implied by the growth-accounting estimates.
- Capacity utilisation
- Gross output