Elizabeth Carlson, Pennsylvania State University
"Accountability among Traditional Chiefs: A Experiment on Transparency and Leakage in Malawi"
Traditional chiefs in African countries have the potential to serve as incredibly valuable partners in local development. Since chiefs are not elected, however, it is unclear whether and to whom they are accountable, and thus how resources provided to these leaders can be prevented from leaking. To identify mechanisms by which chiefs can be constrained, we distributed iron roofing sheets to 200 village-level chiefs with instructions to deliver the sheets to a needy household in their village. We then randomly varied whether the leader’s superiors, citizens, or neither were informed that the chief had been given sheets to distribute. We show that leaders who signed a letter to district officials acknowledging receipt of the sheets were 9 points less likely to divert the sheets than leaders whose superiors were not informed. Informing citizens about the sheets, on the other hand, did not have any impact on leakage. The largest effect was from direct monitoring by project personnel: a follow-up phone call reduced diversion by 15 points. We conclude that traditional leaders are not accountable to their citizenry, and bottom-up monitoring interventions will not be effective. However, chiefs are responsive to political superiors and donors, presumably because these actors control chiefs’ future access to resources. Leakage of resources provided to chiefs can be substantially reduced with simple and cost-effective, top-down monitoring protocols.
Gabriel Lenz, University of California, Berkeley
“The Importance of Knowing ‘What Goes With What’: Reinterpreting the Evidence on Attitude Stability”
Do citizens have meaningful views about public policy? Despite enormous scholarly attention to this question over five decades, researchers have not reached consensus on this question. While scholars such as Converse (1964) have argued that meaningful opinions are held only by a limited number of citizens, other scholars find that, after correcting for measurement error, meaningful attitudes are pervasive in the mass public, even among those with low education and low political knowledge. In this paper, we revisit this debate with a concept at the center of Converse’s theorizing but neglected by subsequent scholarship: knowledge of which issue positions go with which candidate and party—of "what goes with what." We find that much of the public lacks this knowledge and, consequently, much of the public has unstable views, even after correcting for measurement error. We conclude that less than half of US citizens hold meaningful views about even the most salient public policies.
Neil Malhotra, Stanford University
“Natural Disasters and Political Engagement: Evidence from the 2010-11 Pakistani Floods” (with C. Christine Fair, Patrick M. Kuhn, and Jacob N. Shapiro)
How natural disasters affect politics in developing countries is an important question. Research in sociology and psychology suggests traumatic events can inspire pro-social behavior and therefore might increase political engagement. Research in political science argues that economic resources are critical for political engagement and thus the economic dislocation from disasters may dampen participation. We argue that when the government and civil society response effectively blunts a disaster’s economic impacts, then political engagement should increase as citizens learn about government capacity. Using diverse data we show that the massive 2010-11 Pakistan floods had exactly this effect. Pakistanis in highly flood-affected areas became significantly more politically engaged than those less exposed. They acquired more political knowledge eighteen months after the floods and and turned out to vote at higher rates three years later. In a setting where an emerging democratic government was highly responsive in the wake of a natural disaster citizens subsequently invested in their own ability to shape the government’s activities and modestly rewarded the incumbent party. These results suggest that natural disasters may not necessarily undermine civil society in emerging developing democracies.
Gwyneth McClendon, Harvard University
“When Good News is Bad…and When it Isn't: Voter Coordination, Preferences, and Electoral Accountability” (with Claire Adida, Jessica Gottlieb, and Eric Kramon)
A large literature in political science argues that greater access to information about politician performance will help voters reward strong performers and punish the weak. We present results from a field experiment around the 2015 National Assembly elections in Benin that tested this logic. The field experiment, part of EGAP’s First Re-granting Initiative, had two objectives. First, the experiment tested whether greater access to information about legislator performance affected vote choice. Second, the experiment examined two possible moderators of the relationship between information and vote choice: the relevance of the information to voters’ wellbeing, and the ease of voter coordination both within-villages and across them. We leverage official behavioral data, a baseline survey, focus groups, and interviews to investigate how and under what conditions access to performance information influences vote choice.
Kristin Michelitch, Vanderbilt University
“Interventions to Improve Public Service Delivery via Politician Performance: A Field Experiment in Ugandan District Governments" (with Guy Grossman)
The persistence of poor quality public services in low and middle-income countries is commonly attributed to weak political accountability institutions, and thus, poor performance by elected officials. We argue that such poor performance stems from two key information constraints: (a) citizens' lack of politician performance information, and (b) politicians' lack of information about concrete public service deficiencies. In collaboration with a leading Ugandan NGO, we use a field experiment involving 406 politicians across 20 subnational governments over a legislative term to examine the effect of two interventions designed to address these informational barriers: the dissemination of politicians' performance `scorecard' to citizens and a text-messaging service enabling citizens to report public service deficiencies to politicians. We find that disseminating performance information increased politicians' performance as well as tangible development benefits for their constituents across a range of indicators, but only in competitive constituencies. The text-messaging service had no discernible effects. These findings imply that performance transparency and electoral pressure can jointly improve politician performance between elections, even in an electoral authoritarian regime, such as Uganda.
Jennifer Pan, Stanford University
“Party Institutions for Advancement, Participation, and Government Responsiveness in China"
Growing evidence of responsiveness in authoritarian regimes poses a puzzle for our understanding of what produces responsive government. Does responsiveness in the absence of electoral institutions corroborate participatory explanations of government responsiveness, or are non-electoral institutions generating incentives to respond? This paper disaggregates the possible effects of party institutions and participation in China by taking advantage of sharp age limits for retirement. Combining the results of an online field experiment among 2,103 Chinese counties with biographical data on the leaders of all of these localities, this paper estimates heterogenous effects in responsiveness among county-level executives who have the potential to advance in the party and those who are at the end of their careers. Results show that institutional and participatory explanations for responsiveness are complementary in China. Regardless of the advancement prospects of local leaders, the threat of collective action always generates large and significant increases in responsiveness, but localities where leaders face incentives for promotion are more responsive to threats of collective action than localities where leaders are at the end of their careers.
G. Bingham Powell, University of Rochester
“Congruence Astray: When and Why Elections Fail to Induce Ideological Congruence”
Congruence between the ideological position of the median citizen and his or her government is a widely examined measure of substantive representation and even democratic quality. The empirical analyses have focused on alternative election rules, observing average levels of congruence using various methods and in various time frames. But they have not attempted to describe or explain when congruence succeeds and when it fails—and why. This article elucidates the most widely accepted model connecting citizen ideological preferences to government commitments through elections. The role of elections in producing congruence is deconstructed into three stages—voter choices, vote-seat conversion, and formation of parliamentary government. Using citizen perceptions of self-placement and party placement in 64 parliamentary elections in twenty countries in the last twenty years, the degree of congruence failure is described as it cumulates across the stages. Drawing on theories and findings from the previous literature, it proposes and tests new hypotheses that explain why congruence failure accumulates and seldom recovers.
Mike Tomz, Stanford University
"Political Repositioning: A Conjoint Analysis" (with Robert P. Van Houweling)
You can find the conference program here.