Background Whether income inequality erodes citizens’ (VA) depends on institutional context. We examine how integrity—proxied by Control of Corruption (CC)—conditions the inequality–voice link across developed and developing democracies. Methods We assemble a harmonised, unbalanced 10-country panel (2003–2022) with VA as outcome and a parsimonious set of regressors: lagged Gini, lagged poverty, lagged GDP-per-capita growth, CC, Political Stability, and Gini×CC. We test cross-sectional dependence, serial correlation, and stationarity (IPS/CADF). Given mixed orders of integration and no panel cointegration in Westerlund tests with bootstrap p-values, our main specification is first-difference two-way fixed effects with country-clustered inference. A dynamic Augmented Anderson–Hsiao (AAH) estimator provides robustness. Multicollinearity among governance pillars is addressed via diagnostics (VIF) and a parsimonious design. Heterogeneity is assessed via separate models for developed and developing groups. Results Diagnostics indicate cross-sectional and AR(1) dependence; VA, Gini, poverty, and CC behave as I(1), while growth and stability are I(0); Westerlund fails to reject no cointegration (bootstrap p≈0.51 across Gt/Ga/Pt/Pa). In Δ two-way FE, CC is a robust positive correlate of VA, whereas the main Gini effect is small/unstable; the Gini×CC term shows that integrity conditions the inequality–voice link. In developed countries, higher integrity attenuates the (marginally) negative association of inequality with VA (interaction >0). In developing countries, integrity amplifies a negative inequality–VA association as CC improves (interaction
Pacheco-Jaramillo et al. (Fri,) studied this question.