Estimating Productivity When Primal and Dual TFP Accounting Fail: An Illustration Using Singapore's Industries

For both primal and dual TFP growth accounting to properly account for productivity growth, assumptions of constant returns to scale and perfect competition are necessary. This paper shows that without these assumptions, while both TFP growth accounting measures remain equal if factor shares are constant, they are also equally bad at measuring productivity growth. This paper proposes a structural regression to estimate productivity growth based on more general production and cost functions. Using Singapore's industries as illustrations, this paper finds that the assumptions are widely rejected, and the estimated productivity growth exceeds both the accounting measures. When the same methodology is applied to the aggregate Singapore data, the estimated productivity growth is 4.4 percent per year, significantly higher than that of Young's (1992) and Hsieh's (2002).