The Impact of Human Capital Investment on Total Factor Productivity in the Age of Artificial Intelligence: An Empirical Study Based on Cross-Country Panel Data

Abstract

Human capital, as a fundamental driver of technological progress and economic efficiency, plays a pivotal role in determining cross-country differences in total factor productivity (TFP). Utilizing panel data from 142 countries spanning 1970 to 2019 drawn from the Penn World Tables version 11.0 (PWT 11.0), this study empirically examines the impact of human capital investment on national-level total factor productivity through a fixed-effects panel regression framework. The core findings are as follows: first, human capital exhibits a significant positive effect on TFP across all baseline specifications, with a benchmark coefficient of 0.2341 under country and year fixed effects, robust to the inclusion of multiple control variables and alternative estimation approaches. Second, trade openness and physical capital investment rate serve as partial mediating channels through which human capital promotes TFP growth, with mediation ratios of 9.96% and 7.52% respectively, confirming the existence of indirect productivity transmission pathways. Third, government expenditure significantly moderates the human capital-TFP relationship, with higher public spending amplifying the productivity returns to human capital investment and generating positive externalities consistent with the public-private complementarity hypothesis. Fourth, instrumental variable estimation using regional peer-country human capital and lagged human capital as instruments confirms the causal direction of the relationship, with 2SLS estimates exceeding OLS estimates in magnitude, suggesting that standard regressions produce conservative productivity return estimates. These findings carry important implications for developing economies seeking to accelerate productivity-driven growth through coordinated education investment, trade liberalization, and public expenditure policies.