Machine-learned patterns suggest that diversification drives economic development (1812.03534v1)

Abstract: We develop a machine-learning-based method, Principal Smooth-Dynamics Analysis (PriSDA), to identify patterns in economic development and to automate the development of new theory of economic dynamics. Traditionally, economic growth is modeled with a few aggregate quantities derived from simplified theoretical models. Here, PriSDA identifies important quantities. Applied to 55 years of data on countries' exports, PriSDA finds that what most distinguishes countries' export baskets is their diversity, with extra weight assigned to more sophisticated products. The weights are consistent with previous measures of product complexity in the literature. The second dimension of variation is a proficiency in machinery relative to agriculture. PriSDA then couples these quantities with per-capita income and infers the dynamics of the system over time. According to PriSDA, the pattern of economic development of countries is dominated by a tendency toward increased diversification. Moreover, economies appear to become richer after they diversify (i.e., diversity precedes growth). The model predicts that middle-income countries with diverse export baskets will grow the fastest in the coming decades, and that countries will converge onto intermediate levels of income and specialization. PriSDA is generalizable and may illuminate dynamics of elusive quantities such as diversity and complexity in other natural and social systems.