IFI-Linked Fiscal Pressure and Human Development Outcomes: A Machine-Learning Analysis

Abstract Book of the 9th International Conference on Future of Social Sciences and Humanities

Year: 2026

[PDF]

IFI-Linked Fiscal Pressure and Human Development Outcomes: A Machine-Learning Analysis

Pretty Karibo

ABSTRACT:

International financial institutions (IFI) often attach fiscal conditions to lending, yet research on IFI conditions and Human Development Index still tends to treat health, education, income separately. It also pays less attention to whether the relationship between fiscal pressure and human development differs across developing regions. This study looks at how IFI-linked fiscal pressure is associated with human-development outcomes across developing economies. The analysis uses a country-year panel of 76 developing economies from 1990 to 2023. The four indicators of HDI and the HDI itself are modelled. Government primary expenditure and multilateral debt service are used as the main fiscal predictors, while the United Nations M49 classification is used to represent regional context and a set of machine-learning models is then compared. The modelling results show that Random Forest performs best across all five outcomes. They also show that lower government primary expenditure is associated with weaker predicted human-development outcomes, while higher expenditure is associated with stronger predicted outcomes. However, the size of these relationships is not the same across health, education, income and region. The findings are predictive and associational rather than causal, as the data are observational. The modelling strategy is not designed to identify treatment effects. The study contributes by bringing together the dimensions of human development, regional variation and nonlinear modelling within one empirical framework.

Keywords: Human Development; Government Primary Expenditure; Multilateral Debt Service; Machine Learning; Developing Economies; Regional Differences