Recent empirical workin the study of political sophistication finds that citizens’ knowledge of politics is not only a function of their individual characteristics but also depends on the supply of information from their environment (the ‘information environment’). Yet this literature does not address the question of how the information environment may be shaped by institutional factors. This article aims to fill this void. It first argues that the relationship between a government and the media affects the information that is available to individual citizens. Using cross-national data, it then finds that less government interference with the media (1) positively affects political learning and (2) moderates the individual-level effect of education on learning.

Using China's Great Leap Famine as example, this article shows how political career incentives can produce disastrous outcomes under the well-intended policies of a dictator. By exploiting a regression discontinuity design, the study identifies the causal effect of membership status in the Chinese Communist Party's Central Committee—full (FM) Versus alternate members (AM)—on grain procurement. It finds that the difference in grain procurement between AMs and FMs who ranked near the discontinuity threshold is three times that between all AMs and all FMs on average. This may explain why Mao exceptionally promoted some lower-ranked but radical FMs shortly before the Leap: to create a demonstration effect in order to spur other weakly motivated FMs into action.

This article establishes a folk theorem for a model of repeated elections with adverse selection: when citizens (voters and politicians) are sufficiently patient, arbitrary policy paths through arbitrarily large regions of the policy space can be supported by a refinement of perfect Bayesian equilibrium. Politicians are policy motivated (so office benefits cannot be used to incentivize policy choices), the policy space is one-dimensional (limiting the dimensionality of the set of utility imputations), and politicians’ preferences are private information (so punishments cannot be targeted to a specific type). The equilibrium construction relies critically on differentiability and strict concavity of citizens’ utility functions. An extension of the arguments allows policy paths to depend on the office holder's type, subject to incentive compatibility constraints.

This article describes and investigates a model of strategic sequential decision making in networked policy-making environments with three agents. The primary interest is the effect of network structure on sequential policy making and information aggregation. The model and results illustrate how individual policy decisions of varying weight (in terms of a decision maker's unilateral effect on policy outcomes) can enable information aggregation in decentralized environments. In the studied environment, the incentive compatibility conditions for information aggregation are not invariant to network isomorphisms: individuals’ positions in the network matter. The study derives exact conditions for every acyclic network of three or fewer agents and illustrates the counterintuitive nature of comparative statics with respect to both network structure and individual agents’ policy preferences and discretionary authority.

This article puts forward an explanation of the resilience of corruption by arguing that this phenomenon offers net gains to a much larger share of the population than just corrupt government officials. Corruption is modeled as a solution to an allocation problem for a generic government good G. The defining features of this solution are the existence of a market for G when it is not supposed to exist, and the fact that, in that market, contracts cannot be enforced by an outside agent and transactions are usually secret. Corruption redistributes welfare toward ‘insiders’ who share some natural connection to the government and to other insiders. Corruption also redistributes welfare toward those who are skilled at imposing negative externalities.

A Hawthorne effect found in election studies is that pre-election survey participation increases voter turnout. Using the Swedish National Election Studies, Granberg and Holmberg (1992) showed evidence in support of this effect. However, their findings have been criticized and more recent studies have failed to find any treatment effect of pre-election survey participation (cf. Mann 2005). This study re-examines an updated version of Granberg and Holmberg's time-series cumulative data file covering eight additional election studies (for a total of 14 election studies from 1960 to 2010). These studies have an experimental component, since half of the sample was randomly assigned to be interviewed before the election and the other half after the election. By comparing validated turnout in the pre-election sample with the post-election sample, it is possible to estimate the causal effect of survey participation on voter turnout. The results show that participating in the pre-election survey indeed has a significant and positive effect on voter turnout. Moreover, this article evaluates whether the treatment effect is unevenly distributed in the population. Results show that citizens with a low propensity to vote are more affected by taking part in election studies than citizens with a high propensity to vote. The study also estimates the long-term effects of survey participation. Results show that participating in an election survey can have significant effects on voter turnout several years later.

Many citizens abstain from the political process, and the reasons for this abstention are of great interest and importance. Most scholars and pundits assume that greater electoral competition and the increased chance of pivotality will motivate citizens to participate. We test this hypothesis through a large-scale field experiment that exploits the rare opportunity of a tied election for major political office. Informing citizens that an upcoming election will be close has little mobilizing effect. Any effect that we do detect is concentrated among a small set of frequent voters. The evidence suggests that increased pivotality is not a solution to low turnout and the predominant models of turnout focusing on pivotality are of little practical use.