Long-run growth miracles and failures and human capital (PWT 9.1)

Abstract

The recent empirical growth literature has noted that most countries’ incomes do not grow uniformly over periods longer than a decade or so. We ignore the short-term, consider income variations over the long run (up to 65 years) and two growth regimes, and take growth relative to a numeraire country. The two regimes are: periods when a country’s income is catching-up (relative convergence) and when it is falling-behind (both relative and absolute divergence). To minimize parameter heterogeneity, we consider one geographical region, Sub-Saharan Africa (SSA). The average catching-up duration for this region is 19.4 years, average falling-behind is 28.1 years and the number of periods for the two experiences is almost the same. Performing growth accounting by panel estimation separately for the two sets of periods, we find when human capital is contributing to growth in a positive and statistically significant way, countries are catching-up; when it is not, they are falling behind. For the catching-up panels, we find total factor productivity (TFP) to be less important than human capital for both growth and income levels, and the excess-effect of human capital over TFP is about 2.5 times for growth than for income levels.

Keywords: Growth empirics; Catching-up versus falling-behind; Development accounting; Short-run versus long-run effects

JEL Classification: I25, J24, O47