Estimating university-educated workers: Contribution to firms' efficiency using endogeneity-corrected stochastic frontier models

Estimating university-educated workers: Contribution to firms' efficiency using endogeneity-corrected stochastic frontier models


Seminar by Vincent Vandenberghe
On Friday, May 12, 2017 at 01:30 pm (B33 - Trifac 2)

...

In this paper, following Vandenbussche et al (2006) and Aghion et al. (2009),we posit that close to the technological frontier,efficiency growth rests on the presence of workers with advanced educational attainment. And we assess this assumption using a rich panel of Belgian firm -level data, covering the 2008-14 period. We first concentrate on estimating each firm’s technical efficiency (i.e. its distance to the efficiency frontier).The main originality of the paper is to combine proxy-variable methods à-la-Ackerberg et al. (2015) with stochastic frontier methods. The latter have been a standard for measuring firms' technical efficiency; but they ignore the risk of endogeneity. We remedy this problem by combining the Ackerberg et al. proxy-approach to endogeneity control, with a stochastic production frontier estimator We then regress each firm's efficiency growth rate on the share of university -educated workers, its (initial) distance to the efficiency frontier and (the main variable of interest here) the interaction between these two variables, whose sign provides a direct test of the Vandenbussche/Aghion assumption. Results are essentially twofold. First, controling for endogeneity matters when measuring firm-level (in)efficiency. Second, we verify that the closer firms are from the technological frontier, the more university-educated workers contribute to efficiency growth.

...