Measurement Error in Education and Growth Regressions*
*The first author gratefully acknowledges financial support by the Portuguese Foundation of Science and Technology (FCT; ref. SFRH/BD/5114/2001). For comments on previous versions of the paper, we thank Thijs van Rens, Ana Rute Cardoso, and the participants at the Tinbergen Institute Seminar, the 2003 Microeconometrics Conference at CEMFI, the IXth Spring Meeting of Young Economists, the 2004 European Economic Association and Econometric Society Conference. We also thank anonymous referees for their helpful comments.
Abstract
The use of the perpetual inventory method for the construction of education data per country leads to systematic measurement error. This paper analyzes its effect on growth regressions. We suggest a methodology for correcting this error. The standard attenuation bias suggests that using these corrected data would lead to a higher coefficient. Our regressions reveal the opposite. We discuss why this is the case.
