Inequality, Social Capital, and Crime: A Municipal-Level Mediation Analysis in Mexico

Abstract

This study analyzes the relationship between economic inequality, social capital, and crime across Mexican municipalities in 2020, adopting a mediation approach and an exploratory territorial analysis. Using data from the National Council for the Evaluation of Social Development Policy (CONEVAL), the National Institute of Statistics and Geography (INEGI), the Executive Secretariat of the National Public Security System (SESNSP), the National Statistical Directory of Economic Units (DENUE), and the National Electoral Institute (INE), a structural social capital index was constructed through Principal Component Analysis (PCA), a structural social capital index was constructed through Principal Component Analysis (PCA), incorporating indicators of civic participation, associational density, and local social infrastructure. The empirical strategy combines ordinary least squares (OLS) regression models, Exploratory Spatial Data Analysis (ESDA), bootstrap mediation models with 5000 resamples, and spatial regression robustness checks. The results show that inequality is positively and significantly associated with municipal crime rates; the Gini index is associated with vehicle theft (β = 0.115, p < 0.001) and homicide (β = 0.073, p < 0.001). After structural covariates are included, inequality loses statistical significance for property crime but remains associated with homicide (β = 0.098, p < 0.001). Social capital shows a significant negative association with both crime outcomes. The mediation analysis indicates a statistically significant indirect pathway linking inequality and crime through structural social capital; however, the form of mediation differs by crime type. For homicide, the evidence is consistent with complementary or partial mediation, whereas the covariate-adjusted vehicle theft model shows a competitive mediation or suppression pattern. The spatial analysis reveals territorial autocorrelation in all variables. Spatial regression robustness checks further show that structural social capital remains negatively associated with both crime outcomes after accounting for spatial dependence. The findings suggest that the inequality–crime relationship depends on contextual factors and local social structures and should be interpreted as statistical associations rather than causal effects.