Contextual Factors Behind Income Inequality: A Predictive Analysis
Abstract
Income inequality remains a central concern in economics and public policy, yet conclusions about its predictors often depend on both the factors considered and the measure of inequality used. This paper examines the predictive importance of a broad set of indicators covering structural factors, capital structure, financial-sector development, political institutions, regional classification, human capital, globalization, and technology. Using data for 88 countries from 1970 to 2017 and six inequality measures from the World Income Inequality Database (WIID), we apply a machine learning approach to assess the relative predictive importance of 35 candidate indicators across alternative income inequality measures. The analysis proceeds in three stages: a full specification including all candidate indicators, a category-representative specification with one common indicator from each conceptual category, and an inequality-specific specification allowing the most predictive indicator within each category to vary across inequality measures. Regional classification is the dominant predictor across all specifications and inequality measures, although its broad nature prevents identification of the specific regional characteristics underlying this result. Political and technology indicators provide the most consistent additional predictive information, whereas the importance of human capital, financial development, and globalization varies across inequality measures. Technology is particularly informative for top-income concentration, while returns to capital are more relevant for measures capturing gaps between the upper and lower parts of the income distribution.
Related
- Contextual Decision Making: Predictive or Generative
- Differential Privacy for Generalized Entropy Indices of Income Inequality
- Dynamic Spatial Bayesian Machine Learning Model: Applications to Intergenerational Economic Mobility and Geographic Income Inequality in the United States
- Conformal Predictive Distributions for Contextual Order Fulfillment Optimization
- Feeling Left Behind? Territorial Disadvantage, Well-Being and Cohesion across European Regions
BibTeX
@misc{aktas2026contextual,
title = {{Contextual Factors Behind Income Inequality: A Predictive Analysis}},
author = {Arda Aktas and Miguel Poblete-Cazenave and Narasimha Rao},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.36296},
eprint = {2610.36296},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.36296},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Contextual Factors Behind Income Inequality: A Predictive Analysis
AU - Arda Aktas
AU - Miguel Poblete-Cazenave
AU - Narasimha Rao
PY - 2026
DA - 2026/10/26
PB - arXiv
DO - 10.48550/arxiv.2610.36296
UR - https://arxiv.org/abs/2610.36296
AB - Income inequality remains a central concern in economics and public policy, yet conclusions about its predictors often depend on both the factors considered and the measure of inequality used. This paper examines the predictive importance of a broad set of indicators covering structural factors, capital structure, financial-sector development, political institutions, regional classification, human capital, globalization, and technology. Using data for 88 countries from 1970 to 2017 and six inequality measures from the World Income Inequality Database (WIID), we apply a machine learning approach to assess the relative predictive importance of 35 candidate indicators across alternative income inequality measures. The analysis proceeds in three stages: a full specification including all candidate indicators, a category-representative specification with one common indicator from each conceptual category, and an inequality-specific specification allowing the most predictive indicator within each category to vary across inequality measures. Regional classification is the dominant predictor across all specifications and inequality measures, although its broad nature prevents identification of the specific regional characteristics underlying this result. Political and technology indicators provide the most consistent additional predictive information, whereas the importance of human capital, financial development, and globalization varies across inequality measures. Technology is particularly informative for top-income concentration, while returns to capital are more relevant for measures capturing gaps between the upper and lower parts of the income distribution.
ER -