Structural change and income inequality: a meta-analysis

This paper performs a meta-analysis of the literature on the relation between structural change and within-country income inequality. Structure is understood here as the sectoral composition of an economy. The meta-analysis is performed on 686 individual regressions coming from 44 papers. Results indicate no evidence of publication bias but also no evidence for an overall effect of structural change on inequality. However, results also indicate that significant changes in the effect size and sign come from different decisions taken in the empirical setup. Particularly, the decision of measuring structure as the size of agriculture or as the size of industry drives results to opposite directions in similar magnitudes. It is possible that these cancel each other out in the overall picture, leading to the observation of the lack of an overall effect. Other decisions that cause significant changes in the effect size include the data source for inequality, the functional form, the use of an econometric technique robust to endogeneity, the use of heteroskedasticity-robust standard errors, and the inclusion of covariates related to structure, inequality, demography, development level, and labour markets.