Identifying poverty traps based on the network structure of economic output

TL;DR

This paper investigates how the structure of a country's production network influences poverty levels, introducing new metrics to predict and understand poverty traps based on product relatedness and connectivity.

Contribution

It develops two novel measures at the product level capturing short- and long-term poverty patterns using network analysis, enhancing understanding of poverty traps.

Findings

Poverty correlates with poorly connected products, especially natural resource-based ones.

The proposed measures are robust predictors of poverty even when controlling for other factors.

Examples demonstrate how network structure insights can inform poverty alleviation strategies.

Abstract

In this work, we explore the relationship between monetary poverty and production combining relatedness theory, graph theory, and regression analysis. We develop two measures at product level that capture short-run and long-run patterns of poverty, respectively. We use the network of related products (or product space) and both metrics to estimate the influence of the productive structure of a country in its current and future levels of poverty. We found that poverty is highly associated with poorly connected nodes in the PS, especially products based on natural resources. We perform a series of regressions with several controls (including human capital, institutions, income, and population) to show the robustness of our measures as predictors of poverty. Finally, by means of some illustrative examples, we show how our measures distinguishes between nuanced cases of countries with similar poverty and production and identify possibilities of improving their current poverty levels.