Income Inequality, Governance Quality, and Political Engagement: A Cross-country Analysis of Disparities and Democratic Participation

Institutional governance as a mediator

Institutional quality, encompassing Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, and Political Stability, critically shapes civic participation (Kaufmann et al., 2010). Government Effectiveness enhances political engagement by increasing citizens’ confidence that participation yields meaningful outcomes (Evans & Rauch, 1999; Rothstein & Uslaner, 2005). Conversely, ineffective governance, marked by corruption and inefficiency, provokes political disenchantment (Hooghe & Marien, 2013).

The relationship between governance quality and corruption is inherently intertwined and pivotal to understanding civic engagement dynamics. High-quality governance institutionalises transparency, fairness, and responsiveness and establishes robust safeguards against corrupt practices. In contrast, poor governance environments often exhibit institutional weaknesses—such as lack of oversight, bureaucratic inefficiency, and limited civic Accountability—that create fertile ground for corruption to flourish (Kaufmann et al., 2010; Rothstein & Teorell, 2008). This erosion of trust in public institutions discourages citizen participation, particularly among marginalised groups who perceive political processes as inaccessible or rigged in favour of elites. Consequently, addressing corruption is not merely a legal or administrative priority but a fundamental requirement for enhancing institutional legitimacy and encouraging inclusive democratic participation (Johnston, 2005; Uslaner, 2008).

Control of corruption is particularly vital, as it maintains fairness and prevents elite capture of political processes (Rose-Ackerman, 1999; Johnston, 2005). High corruption erodes trust, discouraging formal political participation and occasionally triggering extralegal protests (Morris & Klesner, 2010; Rothstein & Uslaner, 2005). Effective corruption control fosters transparency and equitable representation, empowering marginalised populations to engage actively.

The Rule of Law promotes democratic participation by ensuring impartial legal frameworks, protecting political liberties, and enabling institutional redress for grievances (Tamanaha, 2004; Kaufmann et al., 2010). Weak rule-of-law environments suppress dissent and constrain civil society autonomy, significantly limiting civic engagement (Goldstone et al., 2010). Regulatory Quality intersects with civic freedoms, allowing diverse societal voices, independent media, and robust civil organisations to flourish (Djankov et al., 2003; Gleditsch et al., 2009). Political Stability and Absence of Violence underpin meaningful civic participation by reducing fears of repression and encouraging active political involvement (Dalton, 2004; Norris, 2011).

Voice and Accountability (VA) offers comprehensive insights into democratic health beyond voter turnout, encapsulating freedom of expression, association, and citizen participation in governance (Kaufmann et al., 2010; Cornell & Grimes, 2015). Countries combining high VA with equitable resource distribution sustain robust democratic participation despite economic fluctuations, whereas weak VA contexts often exhibit heightened corruption and institutional distrust (Wilkinson & Pickett, 2010; Morris & Klesner, 2010). Voice and Accountability (VA) is our dependent variable capturing democratic participation and accountability. Specifically, the analysis tests the relationship between Voice and Accountability and income inequality, seeking to uncover whether stronger democratic engagement mechanisms are associated with lower levels of income disparity across different governance contexts.

The choice of these variables rests on their well-established theoretical and empirical salience for explaining democratic participation, measured here through Voice and Accountability (VA), in ways that rival indicators cannot match. Income inequality, proxied by the Gini Index, systematically undermines political equality: higher inequality depresses citizen engagement and concentrates influence among the wealthy (Solt, 2008), reinforces a “one-dollar-one-vote” dynamic that distorts representation (Stiglitz, 2012), and erodes the interpersonal trust that sustains collective action (Rothstein & Uslaner, 2005). Extreme Poverty—captured by the headcount ratio at US $2.15 (2017 PPP)—adds a further constraint, as individuals struggling for subsistence lack the time, education, and resources required for civic involvement (Brady, Verba, & Schlozman, 2006). Complementarily, sustained GDP-per-capita growth signals broader economic opportunity: modernisation theory posits that an expanding middle class fosters political stability and participatory norms (Lipset, 1959). However, economic variables operate within an institutional matrix; thus, the inclusion of Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, and Political Stability acknowledges that robust, transparent institutions amplify the returns to participation and curb elite capture. In contrast, weak or corrupt frameworks breed cynicism and disengagement (Kaufmann et al., 2010). Together, inequality, extreme Poverty, economic growth, and institutional quality form an integrated explanatory set with demonstrable influence on VA, justifying their use over alternative metrics.

Digital civic platforms: Opportunities and risks

Digital civic platforms offer significant potential as “equalisers” by lowering resource-based barriers to participation through simplified legislative processes, accessible voter guides, and enhanced perceived efficacy (Gil de Zúñiga et al., 2020; Simon et al., 2022; Brady et al., 2006). While this study does not empirically test digital participation as a variable, the broader relevance of these platforms lies in their ability to expand civic engagement, particularly in contexts of high inequality or limited institutional trust. UrVote is an illustrative case highlighting the promise—and the challenges—of digital tools in fostering more inclusive democratic processes. However, the effectiveness of such platforms depends heavily on stable regulatory environments, comprehensive internet access, robust cybersecurity, and adequate digital literacy to avoid replicating offline inequalities (Margetts, 2017; Chadwick & Dennis, 2019; Zuboff, 2019).

Concerns about the integrity and security of digital platforms are increasingly salient, given risks related to hacking, surveillance, data misuse, and manipulation by elite interests (Snowden, 2019; Obar & Oeldorf-Hirsch, 2020). Integrating robust cybersecurity measures—such as end-to-end encryption and transparent data governance—is essential to ensure legitimacy and citizen trust in digital platforms. Documentaries and investigative reports have highlighted how digital tools can inadvertently become part of broader surveillance and data exploitation systems, reinforcing the importance of safeguards protecting privacy and data sovereignty.

Data sources and variables

The empirical analysis draws on several reliable data sources, including the World Development Indicators (WDI), which provide measures for the Gini Index, Poverty Headcount Ratio at $2.15 per day (2017 PPP), and GDP per Capita Growth. Governance quality is evaluated through data from the Worldwide Governance Indicators (WGI), which includes Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, Political Stability and Absence of Violence/Terrorism, and Voice and Accountability. Where applicable, voter turnout data are obtained from the International Institute for Democracy and Electoral Assistance (International IDEA) or relevant national electoral authorities.

Measurement units and scaling

The Gini Index (WDI)¹ is reported on a 0–100 scale and interpreted in percentage points. The Poverty Headcount Ratio at $2.15/day (2017 PPP) is the percent of population below the threshold. GDP per capita growth is the annual percentage change. Worldwide Governance Indicators (WGI)—Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, Political Stability and Absence of Violence/Terrorism, and Voice and Accountability—use the standard WGI estimate metric ranging approximately from −2.5 (weak) to +2.5 (strong). For ease of interpretation, we report coefficients in original units, and provide a robustness table with standardized (z-score) regressors for comparability across scales.

The dependent variable, Voice and Accountability (VA), sourced from WGI, is rated approximately from ‒2.5 (indicating weak democratic practices) to +2.5 (indicating robust democratic practices). The main explanatory variables considered are the Gini Index, which captures overall income inequality; the Poverty Headcount Ratio, which highlights the proportion of the population living below $2.15/day; and GDP per Capita Growth, which reflects annual economic performance and its potential impact on civic participation.

The analysis includes governance indicators as potential moderating factors: Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, Political Stability, and Absence of Violence/Terrorism. These indicators explore the extent to which strong institutional contexts might buffer or exacerbate the influence of economic disparities on democratic engagement.

This paper integrates two major strands of literature: the relationship between income inequality and political disengagement, and the role of institutional governance in enabling or constraining civic participation. These dimensions are brought together into a unified analytical framework built around nine core variables: (1) Income Inequality (Gini Index), (2) Poverty Headcount Ratio at $2.15/day, (3) GDP per Capita Growth, (4) Government Effectiveness, (5) Control of Corruption, (6) Rule of Law, (7) Voice and Accountability, (8) Regulatory Quality, and (9) Political Stability and Absence of Violence/Terrorism. By bridging economic and institutional determinants of political engagement, this study aims to provide a more comprehensive understanding of the conditions that shape democratic voices across countries and over time.

Limitations and gaps

The study acknowledges limitations inherent to its design and data sources. Firstly, data gaps exist, particularly for variables such as the Gini Index and Poverty Headcount Ratio, which may not be consistently updated annually for all countries included, potentially affecting the panel’s unbalance and completeness. Secondly, significant multicollinearity among the governance indicators presents analytical challenges; thus, strategies such as factor analysis or separate regressions may be necessary to mitigate inflated standard errors and interpretative ambiguity. Thirdly, despite employing fixed-effects models to reduce omitted variable bias, concerns around endogeneity persist, as reverse causation remains plausible—for example, improved governance potentially influencing lower inequality over time.

Measurement limitations also arise from the choice of indicators. Specifically, the Gini Index primarily captures income disparities and does not comprehensively account for wealth inequality or asset distribution. Similarly, the selected poverty line of $2.15/day may inadequately capture segments experiencing moderate Poverty. Finally, while the inclusion of both developed and developing countries enhances the generalizability of the findings, external validity may be limited when applied to non-democratic contexts or countries where civic spaces are significantly constrained. These caveats underscore the need for cautious interpretation of findings and highlight areas for further research.

While the proposed cross-country, longitudinal design clarifies many issues, further research might adopt mixed methods: in-depth qualitative interviews can capture how citizens perceive inequality and governance in diverse contexts. At the same time, experiments can isolate whether digital interventions truly boost engagement across income strata. Additionally, more advanced econometric strategies (e.g., dynamic panel models and instrumental variables) can tackle endogeneity concerns, including the possibility that higher engagement fosters lower inequality over time. Extending beyond formal democracies and including quasi-authoritarian regimes can reveal how repressive environments alter the relationship between resource disparities and civic life (Leigh, 2005).

Most cross-national studies that link economic conditions to political engagement still rely on income inequality—typically measured by the Gini Index, as in Solt’s seminal SWIID work (Solt, 2008). By centring our analysis on this updated poverty headcount, introduced to reflect higher global living costs, we address the resource-constraint logic of the Civic Voluntarism Model, which posits that material scarcity drains the time, money, and skills needed for participation (Verba et al., 1995). We also construct a unbalanced ten-country panel—five developed and five developing economies—to examine whether the Poverty–voice relationship varies across structural contexts, answering recent calls to avoid averaging away North-South heterogeneity (Haggard & Kaufman, 2021).

Our second innovation is conceptual. Instead of focusing on voter turnout or composite democracy scores, we model Voice and Accountability (VA) using the latest Worldwide Governance Indicators release and its well-documented methodology (Kaufmann et al., 2010). We then interact extreme Poverty with five institutional pillars—government effectiveness, regulatory quality, rule of law, control of corruption, and political stability—to test whether capable, rule-bound states can mitigate the participatory costs of deprivation, a mechanism theorised but rarely examined with full WGI granularity (Rothstein & Teorell, 2008). This approach extends the literature on turnout bias under inequality (Solt, 2010) and complements newer multidimensional poverty frameworks that emphasise institutional context (Alkire & Santos, 2014). Our focus on the US $2.15 line, the symmetric developed-versus-developing sample, and the institutional-interaction design carve out a novel empirical space that previous large-N panels have yet to explore in a single, unified model.

Finally, this study is distinctive because it bridges the traditional divide between “hard-number” economics and more judgment-based institutional diagnostics. By modelling Poverty and GDP-growth figures alongside governance scores such as Voice and Accountability or Control of Corruption, we weave together quantitative resource constraints with qualitative assessments of how power is exercised. This multidimensional design allows us to test not only whether material scarcity suppresses civic voice but under what institutional conditions that effect is amplified or dampened—an interaction rarely captured when scholars treat economic and governance domains in separate silos.

Rising income inequality has become one of the most pressing socio-political challenges of the 21st century, not only for its economic consequences but also for its potential to undermine democratic participation. While economic disparities are often discussed in terms of Poverty and social exclusion, their political implications remain underexplored. This paper seeks to address that gap by examining how income inequality may erode civic voice and Accountability, particularly in contexts where institutional safeguards are weak. The central hypothesis is that higher income inequality reduces Voice and Accountability by concentrating political power, limiting representation, and weakening trust in democratic institutions—effects that may be amplified or moderated depending on the quality of governance.

Descriptive analysis

Before proceeding with the regression analysis, we explore the desciptive summay statistics. Table 2 reports means, standard deviations, and ranges for all variables over the panel. Table 3 splits the sample by development status and the bivariate relationships among the key variables through a correlation matrix (Table 4). This descriptive analysis allows us to identify general patterns of association between institutional indicators, economic variables, and political engagement. While these simple correlations do not imply causality, they provide a helpful starting point to assess the direction and strength of relationships—particularly between Voice and Accountability and factors such as income inequality, Poverty, corruption control, and government effectiveness.

The correlation analysis provides initial insights into the associations between institutional quality, economic conditions, and political participation. Voice and Accountability shows strong positive correlations with key governance indicators, particularly Control of Corruption (r = 0.98), Rule of Law (r = 0.97), and Government Effectiveness (r = 0.93), suggesting that stronger institutions are associated with higher democratic engagement. Conversely, it is negatively correlated with Poverty (r = –0.63) and the Gini Index (r = –0.37), indicating that higher inequality and poverty levels tend to be associated with a weaker democratic voice. These patterns highlight the potential mediating role of institutional quality in the relationship between economic conditions and political participation. However, it is important to note that these correlations do not account for country-specific or time-specific effects, nor do they control for confounding variables. Therefore, the econometric model—including fixed effects and interaction terms—will provide a more robust and accurate estimation of these relationships.

The relationship between income inequality and political engagement by exploring simple correlations and presenting formal regression results that account for institutional quality and economic factors. The results suggest that the simple direct link between economic inequality and political voice might not be linear or straightforward (See Figure 1).

Figure 1. Income inequality and voice and accountability.

By the authors, World Bank and WGI (2025).

In Figure 1, political participation tends to decrease as income inequality increases. However, the association appears weak, suggesting other factors may mediate or moderate this relationship.

In Figure 2, higher levels of Poverty seem weakly associated with lower Voice and Accountability. The trend is not as strong as in the corruption graph, but it is still slightly downward.

Figure 2. Poverty and voice and accountability.

By the authors, World Bank and WGI (2025).

In Figure 3, GDP growth shows no significant association with Voice and Accountability, suggesting that economic expansion alone does not necessarily translate into greater democratic engagement. Recognizing that inequality’s impact might depend on the quality of institutions, we introduced an interaction term between the Gini Index and the Control of Corruption.

Figure 3. GDP per capita growth and voice and accountability.

By the authors, World Bank and WGI (2025).

In Figure 4, higher control of corruption is associated with higher Voice and Accountability, separated by Developed vs Developing countries and an overall trend line.

Figure 4. Control of corruption and voice and accountability.

By the authors.

Econometric model

After exploring the initial associations through a correlation matrix, we proceed with the econometric analysis. While the descriptive results offer preliminary insights, a more rigorous model is needed to assess causality properly—the following regression controls for institutional and economic factors to explain variations in Voice and Accountability better. To examine the association between income inequality and political engagement, we used a unbalanced panel dataset covering ten countries (five developed and five developing) from 2002 to 2022. The variables included the Gini Index (income inequality), poverty headcount ratio at $2.15/day (2017 PPP), GDP per capita growth, and key governance indicators from the World Bank’s Worldwide Governance Indicators (Government Effectiveness, Control of Corruption, Rule of Law, Regulatory Quality, Political Stability and Absence of Violence/Terrorism, and Voice and Accountability).

An Ordinary Least Squares (OLS) regression was conducted, incorporating country and year-fixed effects to control for unobserved heterogeneity and time shocks. Cluster-robust standard errors were employed to account for intra-country correlation.

Before estimating the model, we performed a Variance Inflation Factor (VIF) test to diagnose multicollinearity among the explanatory variables. Before the main estimations, a multicollinearity diagnostic was conducted, motivated by the perception that several qualitative governance indicators might be highly correlated. Variance Inflation Factors (VIFs) were calculated, revealing substantial collinearity among variables such as Government Effectiveness, Rule of Law, and Regulatory Quality. As a result, these indicators were carefully selected or combined in the final model to avoid estimation biases. The results showed high VIF values (above 45) among governance indicators, especially Government Effectiveness, Regulatory Quality, Rule of Law, and Control of Corruption.

Given the theoretical and empirical overlap among these measures, we retained Control of Corruption as the primary governance indicator to avoid distortion of standard errors. This decision was based on its statistical relevance and central role in explaining Voice and Accountability in previous research. Also, given the potential for endogeneity between income inequality and political engagement (e.g., higher Voice and Accountability could reduce inequality), we addressed this issue by lagging key independent variables in one period (Gini Index, Poverty, GDP Growth). This strategy assumes that current levels of Voice and Accountability are affected by past inequality and economic conditions rather than simultaneous interactions.

Lagged independent variables are included in the model to mitigate potential endogeneity concerns, particularly reverse causality between economic conditions and political engagement. By lagging variables such as income inequality, Poverty, and GDP growth by one year, we ensure that these factors temporally precede changes in Voice and Accountability. This approach strengthens the causal interpretation of the results by reducing simultaneity bias (Wooldridge, 2010) and aligning the analysis with theoretical expectations that socioeconomic and institutional structures shape political participation over time (Acemoglu & Robinson, 2012; Persson & Tabellini, 2000).

Model specification

VoiceAccountability_it = Voice and Accountability score for country i in year t

Heterogeneity by institutional context

We split the sample into developed (Australia, Germany, Japan, UK, US) and developing (Brazil, India, Indonesia, Mexico, South Africa) groups and estimate separate two-way fixed-effects models (country and year), with country-clustered standard errors and the harmonized window (Table 6).

Developed countries. Lagged inequality is marginally negative for Voice and Accountability at the reference integrity level (Gini_{t−1} = −0.0387, p = 0.065), and the Gini × Control of Corruption interaction is positive and significant (+0.0280, p = 0.0036). Thus, the marginal effect of inequality becomes less negative as integrity strengthens, crossing zero around Control of Corruption ≈ 1.38 on the WGI scale (−2.5 to +2.5). The negative main coefficient on Control of Corruption (−0.6657, p = 0.098) is imprecise and should be read cautiously given small-N and collinearity within group.

Developing countries. The standalone inequality slope is near zero (−0.0032, p = 0.814), while Control of Corruption is positive and significant (+1.1406, p = 0.016). The Gini × Control of Corruption interaction is negative, borderline (−0.0197, p = 0.0738): as integrity improves, the marginal effect of inequality turns more negative (e.g., at CC = +1, ∂VA/∂Gini ≈ −0.023), whereas under weak integrity (CC < 0) the inequality slope can be slightly positive. Overall, institutional quality modulates the inequality–voice link in opposite ways across groups: it attenuates the adverse inequality effect in developed contexts but amplifies it in developing contexts as integrity rises. We interpret these as within-country associations, not causal effects.

In developing countries, inequality alone does not significantly affect Voice and Accountability. However, when corruption is poorly controlled, inequality significantly exacerbates political disengagement. Thus, institutional quality—specifically, the control of corruption—modulates the relationship between income inequality and political voice.

We assess long-run comovement using Westerlund (2007) tests on the parsimonious model. Using bootstrap p-values robust to cross-sectional dependence, the tests do not reject the null of no cointegration (Gt = 3.08, p_boot = 0.51; Ga = 3.08, p_boot = 0.51; Pt = 11.91, p_boot = 0.51; Pa = 0.28, p_boot = 0.51). Given mixed integration orders and the lack of cointegration, we focus estimation on first-difference two-way fixed effects and provide a dynamic AAH robustness; level specifications are reported as descriptive associations.

Westerlund tests with bootstrap p-values do not indicate cointegration in the parsimonious system (Gt/Ga/Pt/Pa: p_boot = 0.51 throughout); we therefore centre inference on first-difference FE and a dynamic robustness.

We also estimated a dynamic specification via the Augmented Anderson–Hsiao (AAH) estimator. Results show strong persistence in VA (lagged VA = 0.701, SE = 0.072, p < 0.001), a positive and significant association for Control of Corruption (0.326, SE = 0.141, p = 0.023), and a negative inequality×integrity interaction (−0.0053, SE = 0.0030, p = 0.083). Other covariates are not statistically significant. The model explains a large share of within-country variation (R² = 0.788). Under stability (φ ≈ 0.70), the implied steady-state integrity effect is ≈ +1.09 (= 0.326/(1−0.701)), while the interaction implies that the marginal effect of inequality turns negative when Control of Corruption ≳0.67. All four statistics + p-values and a one-line takeaway (“CSD present → clustered SEs; robustness with Driscoll–Kraay/CCE where reported).

Interpretation

Until this stage, income inequality alone did not show a statistically significant effect on Voice and Accountability without considering its interaction with corruption control. However, additional diagnostic tests and model refinements were conducted to ensure the robustness of these initial findings. These adjustments aimed to better account for potential cross-sectional dependence, heteroskedasticity, and long-run relationships among the variables.

A comprehensive series of diagnostic tests will be conducted to ensure the robustness of the panel data estimations. First, cross-sectional dependence among countries will be assessed using the Pesaran (2004) scaled LM test and the correction proposed by Pesaran, Ullah, and Yamagata (2008). This step is critical to verify whether the standard panel model assumptions hold. The Wu–Hausman test will be applied to formally decide between fixed effects, random effects, or pooled OLS estimation, ensuring the appropriate control of unobserved heterogeneity.

Second, the stationarity properties of the panel data will be examined, given the relatively large number of units and periods (10 countries × 21 years). The Im–Pesaran–Shin (IPS) test (2003) and the cross-sectionally augmented Dickey-Fuller (CADF) test (Pesaran, 2007) will be used to detect unit roots while accounting for cross-sectional dependence. Orders are mixed; we therefore test for panel cointegration and center inference on Δ-FE and a dynamic robustness. Suppose variables are found to be integrated of order one (I(1)). In that case, panel cointegration will be tested using the Westerlund (2007) methodology, which is robust to serial correlation and cross-sectional dependence. Cointegration is not confirmed” and state that the main model is first-difference two-way FE with clustered/Driscoll–Kraay SEs; levels kept as descriptive robustness, estimation strategies will be adapted by specifying error correction models or differencing non-stationary series.

Thus, with a rigorous econometric strategy to address potential statistical issues inherent in panel data. Diagnostic tests were conducted to assess cross-sectional dependence (Pesaran, 2004; Pesaran, Ullah, and Yamagata, 2008), stationarity (Im, Pesaran, and Shin, 2003; Pesaran, 2007), cointegration (Westerlund, 2007), heteroskedasticity (Breusch & Pagan, 1979; White, 1980), and autocorrelation (Durbin & Watson, 1950; Breusch, 1978; Godfrey, 1978). A fixed-effects panel model with robust clustered standard errors was estimated based on these diagnostics ( Table 4). The findings offer strong empirical support for the proposed relationships, particularly regarding the adverse effect of income inequality on Voice and Accountability once institutional factors are correctly accounted for.

While initial estimations suggested a limited direct association between income inequality and political engagement, the refined model—correcting for heteroskedasticity, cross-sectional dependence, and non-stationarity—revealed a statistically significant negative relationship. This evolution highlights the importance of applying robust econometric procedures in cross-country analyses (Wooldridge, 2010) to uncover underlying structural effects that simpler models might obscure. Thus, the study’s conclusions are considerably strengthened by these methodological corrections.

Nevertheless, caution is warranted. Some diagnostic tests, such as the cross-sectional dependence and unit root tests, were conducted using approximated methods due to technical constraints. Although these approaches align with standard practices in empirical political economy (Baltagi, 2005), they might not fully capture complex dynamic interactions across countries. Therefore, while the results are consistent with theoretical expectations and supported by corrected inference, they should be interpreted as robust associations rather than definitive causal proof.

Robustness and limitations

To ensure the robustness of the findings, several methodological safeguards were implemented. Fixed effects (Hausman, 1978) were used to control for unobserved country-specific heterogeneity, and lagged independent variables helped mitigate potential reverse causality (Wooldridge, 2010). Residuals were tested and corrected for heteroskedasticity and cross-sectional dependence, and panel cointegration was confirmed, supporting a long-term equilibrium relationship. However, the study is limited by the relatively small number of cross-sectional units (countries) and the inability to fully execute second-generation panel techniques such as Driscoll–Kraay standard errors (Driscoll & Kraay, 1998) or Common Correlated Effects estimators (Pesaran, 2006). Future research could extend this analysis by applying more sophisticated models explicitly designed to handle cross-sectional dependence and dynamic panel structures. Despite these limitations, the results presented here provide a robust empirical foundation for understanding the conditional effects of income inequality on political voice.

We estimated a dynamic panel for Voice and Accountability using the Augmented Anderson–Hsiao (AAH) estimator (Chudik & Pesaran, 2022) to address persistence, endogeneity, and common shocks. The differenced equation includes the lagged dependent variable ΔVA_{t−1} and uses VA_{t−2} as an internal instrument; we augment the specification with cross-section averages (levels and differences, with lags) of VA and the regressors to absorb unobserved common factors. The covariate set mirrors the parsimonious model: Gini_{t−1}, Poverty_{t−1}, GDP per-capita growth_{t−1}, Control of Corruption, Political Stability, and the Gini×Corruption interaction. Estimation is performed on the harmonised sample window, with standard errors clustered by country; we verify the absence of second-order serial correlation in differences and report robustness to cross-sectional dependence. This design delivers short-run within-country associations that are consistent with the diagnostics (no panel cointegration), while allowing a transparent interpretation of persistence and integrity effects.

Policy implications and future research directions

Addressing structural barriers

Suppose inequality and Poverty are key drivers of civic disengagement. In that case, policy efforts must prioritise redistributive mechanisms (e.g., progressive taxation, social safety nets) and educational opportunities that can level the playing field (Stiglitz, 2012; Hacker & Pierson, 2010). These structural remedies, combined with strong governance—particularly in corruption control—can help restore trust among underprivileged groups, encouraging them to participate in elections or local assemblies.

Governance reforms to build trust

Corruption undermines fairness, so improving Control of Corruption is a top priority. Transparent procurement, open budget initiatives, and independent anti-corruption agencies have shown promise in raising institutional credibility (Rose-Ackerman, 1999). Additionally, bolstering the Rule of Law ensures that no individual or group is above the law, reducing intimidation or marginalisation of civil society. Implementing e-government platforms for public services can enhance Government Effectiveness, while coherent regulatory frameworks (i.e., strong Regulatory Quality) facilitate open media and robust civic organisations (Djankov et al., 2003).

Addressing structural inequalities through redistributive policies, progressive taxation, and educational investments is crucial for mitigating disengagement (Stiglitz, 2012; Hacker & Pierson, 2010). Strengthening governance through improved Control of Corruption, Rule of Law, and Government Effectiveness is equally essential in rebuilding institutional trust and fostering civic participation (Rose-Ackerman, 1999; Djankov et al., 2003).

From a policy standpoint, our integrative approach underscores the necessity of multi-pronged strategies: tackling structural inequalities through tax reforms or social programs, curbing corruption, guaranteeing the rule of law, and leveraging technological innovations to promote Accountability. Also in a previous study, we concluded that a strategic combination of social programs as progressive taxes, subsidies and vouchers can lead to more equitable outcomes (Pacheco-Jaramillo & Malliaros, 2025). Empirically, future cross-national analyses employing panel data will help clarify the interplay between these governance variables and economic disparities, potentially identifying tipping points at which inequality evolves into large-scale disaffection or triggers reforms. Ultimately, bridging the resource gap and reinforcing democratic institutions is paramount to ensuring that all citizens—regardless of socioeconomic status—feel invested in and capable of participating in public life.

Expanding on theoretical and empirical directions, this merged study offers a foundation for deeper investigations into how economic disparities shape civic engagement across diverse national settings. In doing so, it contributes to an emerging consensus that robust institutions, equitable resource distribution, and inclusive technological solutions are indispensable in nurturing a genuine, stable democracy capable of withstanding the pressures of globalisation, polarisation, and persistent inequality.

Leveraging digital civic platforms

Drawing on resource-based and psychological disengagement theories (Brady et al., 2006; Dalton, 2004; Solt, 2008), we argued that rising inequality and extreme Poverty weaken public trust and reduce participation, while robust governance metrics often mitigate these adverse dynamics (Kaufmann et al., 2010; Rothstein & Uslaner, 2005). The synergy between strong institutional performance and redistributive policies can sustain inclusive political systems in which even marginalised citizens exercise voice (Uslaner & Brown, 2005). At the same time, the advent of digital civic platforms offers promising avenues for surmounting resource constraints and engaging the disaffected (Gil de Zúñiga et al., 2020; Simon et al., 2022), provided that basic conditions of stable governance and adequate digital infrastructure are met.

Beyond institutional reforms, digital civic platforms could help offset resource deficits that hamper engagement among low-income communities. By providing simplified legislative summaries, automated “How to Vote” guides, and public forums, such initiatives can reduce the informational barriers that perpetuate political inequality (Simon et al., 2022). However, technology alone is insufficient if underlying governance conditions like political stability remain weak or digital access is uneven (Gil de Zúñiga et al., 2020). Governments and NGOs could collaborate to expand broadband infrastructure, promote digital literacy, and ensure platform neutrality. Doing so might significantly augment the capacity of poor or isolated communities to hold elected officials accountable.

Future research should incorporate mixed-method approaches, combining qualitative insights into citizen perceptions of inequality and governance with quantitative analyses employing advanced econometric techniques like dynamic panel models to address endogeneity. Expanding research to authoritarian or quasi-democratic contexts could provide valuable insights into civic participation under repressive conditions (Leigh, 2005). Also, future research should explore broader dimensions of participation, such as protest movements or online engagement, and apply mixed methods to deepen understanding of citizen perceptions. Enhancing Voice and Accountability in unequal societies requires more than economic growth; it demands inclusive, trustworthy, and responsive governance.