Innovation, structural changes, and income inequality: New insights from machine learning approach

Innovation and structural changes dynamics in developed and developing countries require measuring their potential heterogeneous effects on income inequality, which has become a major global concern guiding recent empirical research. In this context, the current study aims to estimate the distributional consequences of innovation and structural change using parametric and nonparametric estimation techniques for 29 developed and 14 developing countries from 1996 to 2020. The results of Kernel Regularized Least Squares estimations (KRLS) show nonlinear and time-varying paths of distributional effects of innovation and structural changes. Although innovation increases income inequality in both developed and developing countries, results support heterogeneous distributional effects of industrialization and deindustrialization as proxies for structural changes. Industrialization increases income inequality in developed countries and decreases it in developing countries, while deindustrialization reduces inequality in developed countries and increases it in developing countries. The study recommends policies to reduce income inequality associated with innovation processes in developed and developing countries. It argues that developing countries following the traditional path of structural change through deepening industrialization is beneficial for reducing income inequality, while implementing policies aimed at deepening deindustrialization represents a positive path towards reducing income inequality in developed countries.