The Elasticity of Elasticity of Substitution Estimates: Is Anything Robust?

42 Pages Posted: 19 Sep 2007

See all articles by Andrew T. Young

Andrew T. Young

Texas Tech University - Rawls College of Business

Date Written: September 18, 2007

Abstract

We use a CES production function with no restrictions on technical bias to derive relationships between the growth in relative factor shares and (i) the capital to labor ratio and (ii) the ratio of marginal products. These relationships constitute a parsimonious specification used to identify the elasticity of substitution between labor and capital (σ). We estimate σ using (a) quarterly and annual data sets, (b) level and growth rate specifications, and (c) OLS as well 2SLS and GMM estimators utilizing (d) instruments that are tested for both relevance and exogeneity. Estimation in each of the above cases is carried out (e) both imposing and not imposing cross-equation restrictions. The most robust result is that estimation with imposed cross-equation restrictions and "demand-shifter" instrumentation yields σ estimates that are, in most cases, very close to unity.

Keywords: Elasticity of Substitution, Factor Shares, Labor's Share, Biased Technical Change

JEL Classification: E60, E22, E23, H25, O47

Suggested Citation

Young, Andrew T., The Elasticity of Elasticity of Substitution Estimates: Is Anything Robust? (September 18, 2007). Available at SSRN: https://ssrn.com/abstract=1015339 or http://dx.doi.org/10.2139/ssrn.1015339

Andrew T. Young (Contact Author)

Texas Tech University - Rawls College of Business ( email )

Lubbock, TX 79409
United States