Social Capital, Economic Conditions and Evaluative Well-Being: Comparing Life Satisfaction and the Cantril Ladder Across the West and the Global South

Social capital is widely recognised as a determinant of subjective well-being (SWB), yet most studies rely on a narrow set of indicators. Typically, the association between generalised trust and civic participation and a single evaluative well-being measure is explored, mainly in Western countries. Using Waves 1 and 2 data from the Global Flourishing Study, we test whether these conventional social capital indicators are the most important predictors for two different evaluative well-being measures: life satisfaction and the Cantril ladder. Additionally, we use economic measures as a benchmark, given that traditionally it has been among the principal indicators and most extensively discussed correlates of SWB, to compare the predictive importance of three social capital dimensions or blocks across measures and development contexts, i.e., Western countries and the Global South. We employ XGBoost models and SHAP values, normalised for cross-sample differences in the dispersion of the outcome variables, to rank predictors. A sequential Pseudo-R² decomposition is used as a scale-free robustness validation, and panel fixed-effects models with standardised coefficients are used to assess which relationships survive within-person identification. We find that the conventional social capital indicators, civic engagement and general trust, used predominantly in the literature, perform poorly. The two variables add little to both SWB measures across all samples (Western and Global South) and, in several specifications, worsen out-of-sample prediction. Among the social capital blocks (dimensions), belonging, driven by satisfaction with close relationships, is the only dimension that exceeds economic variables, and it does so only for life satisfaction, not for the Cantril ladder. In the Global South, material challenges have the largest standardised within-person association with life satisfaction (ß = -0.213) and exceed all social capital variables. In Western countries, satisfaction with relationships dominates (ß = 0.204). The Pseudo-R² decomposition shows the same broad block-level pattern, and the fixed-effects estimates further indicate that the main relationships are also evident within persons over time. Our findings indicate that conclusions about the importance of social capital depend on which dimension (and indicators) of social capital is measured, which SWB outcome is used, and in which region the analysis is conducted.