The Role of Schooling in Human Capital Production Functions: Estimation Challenges and Insights

• Authors: Inés P. Murillo Huertas and Josep-Lluís Raymond-Bara.
• BSE Working Paper: 1526 | October 25
• Keywords: HCPF , OLS , IV , omitted variables bias
• JEL codes: C18, I26, J24

This article estimates a human capital production function (HCPF) using both ordinary least squares (OLS) and instrumental variables (IV), finding that the returns to schooling in terms of competencies are higher when using the IV approach. To interpret this counterintuitive result, a proxy for individual ability is constructed using data from the Programme for the International Assessment of Adult Competencies (PIAAC). The higher IV-based returns align with OLS estimates when the sample is restricted to high-ability individuals. Moreover, when the ability proxy is included as a control in the HCPF, the OLS and IV estimates converge, suggesting that the discrepancy between the two methods can be explained by omitted variable bias related to ability. This extended HCPF is then used to compare the effectiveness of competency generation across a selection of OECD countries. The results indicate that countries with lower competency outcomes are primarily characterized by poorer effectiveness in transforming attained schooling into competencies.