Know more

Tuesday, April 12th 2022

Frederic Ang (Wageningen University) will present "Robust nonparametric analysis of dynamic profits, prices and productivity: an application to French meat-processing firms", joint with Pieter Jan Kerstens.

Abstract:

This article develops a statistically robust nonparametric framework to analyze profits, prices and productivity in a dynamic context that appropriately considers adjustment costs. The change in dynamic profit (current value of all future profits) is decomposed into dynamic price change and dynamic productivity change. These novel components respectively correspond to a dynamic Bennet price indicator and a dynamic Bennet quantity indicator. The latter is further decomposed into dynamic technical change, dynamic technical efficiency change and dynamic mix efficiency change. It is shown to be a superlative indicator for the dynamic Luenberger indicator for appropriately normalized prices: if the dynamic directional distance function can be represented by a quadratic functional form with time-invariant second-order coefficients and there is dynamic profit-maximizing behavior, then an appropriately price-normalized dynamic Bennet quantity indicator coincides with the dynamic Luenberger indicator. The application focuses on 1,638 observations of French meat-processing firms for the years 2012 ́2019. The decomposition framework uses data envelopment analysis, to which we apply a recently developed m-out-of-n bootstrap to obtain robust estimates and confidence intervals. Overall, this framework provides a powerful tool to analyze economic performance and guide the resource reallocation required for increasing profits and productivity in the long run. The dynamic profit increases by on average 0.320% p.a. in the studied period. It is driven by a modest dynamic productivity growth of on average 0.496% p.a., which is partly offset by dynamic price decline of on average 0.176% p.a. The components of dynamic productivity growth fluctuate substantially. However, they are often statistically insignificant, which highlights the importance of using the m-out-of-n bootstrap.