The Matthew effect in political science: head start and key reforms important for democratization

The study investigates whether countries are better positioned at the onset of liberalization to successfully democratize, testing for a Matthew effect—where initial advantages compound over time—in the context of political transitions. Building on decades of research on democratization that often assumes constant effects across cases, the authors examine whether initial conditions related to political institutions, economic development, and education significantly condition the likelihood of successful transitions. Using Dahl’s concept of polyarchy and recent work on identifying liberalization episodes, the paper sets out to compare initial institutional and socioeconomic conditions, as well as the pace of reform, between successful and failed democratization attempts from 1900 to 2018. The purpose is to identify whether prior advantages translate into higher success rates and faster reforms, thereby informing theories of democratization and policy approaches.

The paper situates its inquiry within literature on cumulative advantage (the Matthew effect) across domains such as science, markets, sports, and education (e.g., Merton 1968, 1988; Salganik et al. 2006; van de Rijt et al. 2014), and within democratization research emphasizing economic development and education (Lipset 1959; Teorell et al. 2019). Prior studies have noted roles for prior democratic experience (Bernhard & Karakoc 2007; Persson & Tabellini 2009; Gerring et al. 2012) and cultural foundations (Ruck et al. 2019), but not a general Matthew effect over a century-plus span. The paper also engages debates on electoral institutions in autocracies—whether elections stabilize authoritarianism through cooptation and elite management (Gandhi & Przeworski 2007; Lust-Okar 2009; Greene 2010; Magaloni 2008; Boix & Svolik 2013) or correlate with democratization upon regime failure (Wright & Escribá-Folch 2012; Wilson 2019; Bernhard et al. 2020). It revisits modernization theory debates on whether development increases democratization likelihood (Przeworski & Limongi 1997; Boix & Stokes 2003; Przeworski et al. 2000), proposing that development may not trigger transitions but may condition their success once underway.

Conceptualization: Democracy follows Dahl’s (1971) polyarchy, emphasizing institutional guarantees: free, fair, periodic elections; suffrage; and freedoms of association and expression. The primary indicator is V-Dem’s Electoral Democracy Index (EDI), a composite of 24 variables measuring elected officials, suffrage, election quality, and freedoms (Teorell et al., 2019; Coppedge et al., 2019). Liberalization episodes are defined as periods moving a country from non-democracy toward democracy by at least 10% of the EDI range. Sample restriction: episodes beginning in closed or electoral autocracies, excluding positive changes in already democratic regimes. Classification of episodes: Using Wilson et al. (2020), episodes are categorized as successful (transition achieves electoral democracy threshold and holds free and fair founding elections with winners assuming office), failed (liberalization begins but does not meet success criteria), or censored (ongoing by 2018). Dataset: 337 liberalization episodes across 155 countries (1900–2018): 146 successful (110 countries), 182 failed (91 countries), 9 censored. Analytical strategy: Compare initial conditions (the episode start year, as well as one year before and after) between successful and failed episodes for: (a) overall EDI; (b) the 24 EDI component indicators mapping to institutional guarantees; (c) GDP per capita; and (d) primary school enrollment (education). Assess the pace of reform during episodes by computing average annual changes in the 24 indicators, GDP per capita, and primary school enrollment, comparing successful vs. failed episodes. Statistical testing includes t-tests (e.g., overall pace difference p<0.001; specific indicators p<0.05) and checks for confounding by episode duration (differences not significant, p=0.618). Data sources: V-Dem Dataset (v9) for democracy indicators; macro indicators for GDP per capita and primary school enrollment as reported alongside V-Dem analyses.

- Initial institutional advantage: Successful and failed episodes differed significantly in EDI levels at episode start, as well as the years before and after (Fig. 1; 99.9% CIs). - Component indicators: Of 24 democracy indicators, 16 significantly differed at episode start; 14 differed the year before; 17 differed the year after (Supplementary Fig. S1). Consistent differences covered election quality, freedoms of discussion (men and women), opposition party autonomy, civil society activity, and media freedom (government censorship, harassment of journalists). By year one, additional differences appeared in party barriers and civil society repression; by year two, further differences in media criticism and media bias. - Socioeconomic conditions: Countries with successful episodes had significantly higher GDP per capita and primary school enrollment at episode start and in adjacent years (Fig. 3), echoing classic findings but newly framed as a Matthew effect. - Pace of reform: Successful episodes had a significantly higher average pace of change during liberalization (p<0.001). Indicators with higher pace for successful episodes (t-test p<0.05) included: eight freedom of expression measures (academic freedom; discussion for women; government censorship; media criticism; harassment of journalists; media self-censorship; range of views; media bias), two freedom of association measures (entry/exit of organizations; civil society repression), and election management body autonomy (Fig. 4). Episode durations did not differ significantly by outcome (p=0.618). No pace difference for GDP per capita; significant faster gains in primary school enrollment for successful episodes (p=0.01). - Overall: Evidence supports a Matthew effect in democratization: countries starting with stronger institutions, higher development, and better education not only succeed more often but also reform faster, including in areas without initial advantage.

The findings directly address the research question by demonstrating that initial advantages in political institutions, economic development, and education substantially increase the probability that liberalization results in democracy, and that these advantages accelerate reforms across multiple institutional domains. This constitutes a Matthew effect in democratization: advantage begets further advantage. The results help reconcile debates about elections under autocracy by suggesting that while elections can serve both autocratic and democratic ends, higher-quality electoral institutions make democratic transitions more likely. The demonstrated role of development and education refines modernization debates, indicating that although development may not initiate transitions, it strongly conditions their success once underway. The study underscores the importance of legacy effects and suggests that building civic, institutional, and human capital—especially election management capacity, media freedom, and civil society space—prior to or early in liberalization can be decisive for successful democratization. It calls for re-evaluating models that assume constant effects across countries and highlights potential thresholds and path dependencies that shape transition outcomes.

The paper provides the first systematic evidence of a Matthew effect in democratization over 1900–2018: countries with stronger initial democratic institutions, higher GDP per capita, and greater educational attainment are more likely to achieve successful transitions and reform faster during liberalization. It identifies specific domains—election quality, freedoms of expression and association, media freedom, and election management autonomy—where initial advantages and accelerated improvements are most consequential. These insights challenge assumptions of uniform effects in democratization research and offer a framework to reconcile competing views on authoritarian elections. Future research should identify threshold levels of institutional and socioeconomic ‘capital’ that tip transitions toward success, examine whether similar Matthew effects operate in democratic consolidation, and explore policy interventions that can feasibly build key capacities before or during early liberalization phases.