Courant Research Centre: Poverty, Equity and Growth - Discussion Papers 207
In what the authors name 'a first pass through the data', McMillan et al. (2014) have recently addressed the question: what determines the magnitude of growth-enhancing structural change - defined as gains to average labor productivity resulting from a reallocation of labor across sectors? This paper extends their cross-section work to a panel data set of 5- and 10-year intervals from 1970 to 2010 for 29 (mostly developing) countries. Controlling for a wide range of control variables and time-invariant unobserved heterogeneity, the results present support for growth-enhancing structural change to be the outcome of a conditional domestic convergence process towards, what I term, a country's idiosyncratic state of efficient allocation. The regressions further indicate that the removal of labor market rigidities and improvements in gender equality in education correlate with larger gains from structural change in a statistical and economical meaningful way. However, these relationships are not found in countries with large (gender) inequality in education or strong labor market rigidities, respectively. The study also shines some light on the channels through which the variables potentially affect gains from structural change.