Experience and growth

Abstract

This paper jointly estimates the social returns to physical and human capital with aggregate production functions using cross-country, first differenced panel data at frequencies of 10 years. Following cross-sectional wage regressions, the average experience of the workforce is added to average education as a proxy for the workforce. Growth in average experience is significantly associated with productivity growth. Moreover, controlling for experience increases the level and significance of returns to education.

Introduction

This paper estimates the social returns to human and physical capital from aggregate national production functions using first-differenced (at 10-year frequencies) cross-country panel data. Topel (1999) estimates a positive relationship between increases in aggregate education levels and productivity growth by applying functional forms drawn from the cross-sectional Mincer equations used by labor economists to estimate the relationship between education and wages. However, both Topel (1999) and Krueger and Lindahl (2001) emphasize the difficulty of finding large or statistically significant effects of aggregate changes in education when simultaneously estimating the returns to physical capital. The innovation here is the inclusion of the experience level of the workforce in the productivity regressions. Omitting experience may be an important source of downward bias on the gross returns to education. First, at the individual level, extra years in education will directly reduce experience. Second, at the aggregate level, periods in which many young, better educated workers are entering the workforce for demographic reasons are also periods in which the experience level of the workforce will be diluted for the same demographic reasons.