Miriam Golden’s presentation in CDDRL’s Research Seminar on April 23, 2026, addressed a central puzzle in democratic politics: why are incumbent reelection rates systematically higher in richer democracies? Drawing on cross-national data, she demonstrates a strong positive relationship between national income and reelection rates, a pattern that is both statistically robust and theoretically unexpected. This empirical finding motivates a reassessment of two dominant frameworks — accountability theory, associated with John Ferejohn, and selection theory, associated with James Fearon. Accountability models suggest that voters reward good performance and punish poor performance, but they do not explain cross-national variation in reelection rates. Selection models argue that elections filter out low-quality politicians, implying that poorer countries with lower reelection rates must have dishonest or incompetent politicians, yet empirical evidence does not align well with these inferences.

Golden proposes an alternative framework centered on “capacity gaps,” introducing the resources that politicians have available and voters' ability to discern political performance as key missing parameters. In poorer countries, both state capacity and voter interpretive capacity are constrained. Governments face fiscal and administrative limitations that restrict policy delivery, while voters struggle to distinguish whether poor outcomes result from incompetence, corruption, or structural constraints. As a result, the informational conditions necessary for effective accountability break down. Golden further argues that informational signals are asymmetric: markers of “bad” types, such as corruption scandals, criminal convictions, or dynastic ties, are visible and salient, whereas markers of “good” types, such as competence or honesty, are diffuse and easily mimicked. In these settings, even honest, competent, and well-intentioned politicians are likely to lose office because they are indistinguishable to voters from the malfeasant and incompetent. Even high-performing politicians may not be rewarded electorally, and good types gain no consistent advantage in reelection. 

To evaluate this framework, Golden presents multiple empirical investigations. First, she examines whether voters reward economic performance using within-country variation in GDP growth. The results show that higher growth increases reelection rates, but only in countries with high literacy levels. Since literacy roughly proxies voter discernment capacity, this suggests that performance matters electorally only when voters can interpret it. Second, she analyzes survey data from legislators in Italy and Pakistan to assess whether elections filter out low-quality politicians. She finds that politicians with “bad-type” markers, such as dynastic backgrounds or long tenure, exhibit higher tolerance for corruption yet continue to survive electorally, contradicting selection theory. Third, she tests whether poorer democracies have lower-quality politicians by examining education levels and relative salaries. She finds no meaningful differences in legislator quality across income levels and no relationship between salaries and reelection rates, further weakening selection-based explanations.

Overall, Golden’s approach reconciles several empirical anomalies: the income–reelection relationship, the conditional effect of economic performance, and the persistence of low-quality politicians. At the same time, important questions remain regarding causal identification and measurement, as proxies like literacy may capture broader development effects. Nonetheless, the framework offers a compelling shift in focus from politicians to voters, highlighting how limits in information processing can undermine both accountability and selection in democratic systems.