Imperfect Price Deflation in Production Systems

Abstract

Numeraire prices that are measured with error create challenges for econometric estimation. A straightforward approach for a model with linear input demands, such as generated from a quadratic normalized profit function, is proposed where the numeraire price is measured with error. Numeraire measurement error is likely because expected output price is measured imperfectly by actual output price. An approach using generalized method of moments is developed to estimate such errors-in-variables (EIV) systems that avoids use of extra-sample data or additional structural specifications. Monte Carlo examination of small sample properties shows promise. Measurement error is statistically significant using aggregate U.S. agricultural data.