Poverty Reduction in Burundi: Has Foreign Aid Helped? -A Machine Learning Perspective

Abstract

This study assesses the effectiveness of foreign aid in promoting development outcomes in Burundi using machine-learning techniques. Burundi remains highly dependent on external assistance, yet the extent to which aid translates into sustained improvements in economic growth, poverty reduction, health, education, and human development remains contested. Earlier Burundi-focused research has reported mixed results: aid has been associated with some improvements in education and child-mortality outcomes, while its effects on poverty reduction and growth have been less conclusive; political stability and governance conditions also appear important to aid performance. Using annual data for Burundi over the available study period, the research combines foreign-aid indicators such as official development assistance disbursements, sectoral aid allocations, and aid per capita with macroeconomic, institutional, demographic, and social indicators. Key explanatory variables include domestic investment, trade openness, inflation, exchange-rate movements, government expenditure, population growth, political stability, governance quality, and measures of human capital. The outcome variables include GDP per capita growth, poverty-related indicators, child mortality, school enrolment, and the Human Development Index, subject to data availability. The research aims to identify whether foreign aid is a meaningful predictor of development progress in Burundi and determine the institutional and macroeconomic conditions associated with stronger aid outcomes. The findings support more targeted, transparent and development-oriented policies in Burundi.

Keywords: Machine learning, foreign aid, official development assistance, poverty, growth, inflation, macroeconomics

DOI: 10.7176/JESD/17-6-05

Publication date: September 30^th 2026