150 likes | 287 Vues
In this presentation, Rachel Glennerster, Executive Director of J-PAL at MIT, discusses the challenges of governance in democracies, including low voter knowledge and persistent corruption. The talk explores the role of citizens as direct monitors and voters, emphasizing mixed results from local monitoring initiatives. Glennerster highlights evidence showing that targeted campaigns can enhance voter behavior and participation. The presentation also raises critical questions about the effectiveness of community monitoring and the potential to integrate it with informed voter strategies to improve service delivery and governance quality.
E N D
Improving policies by improving governance Rachel Glennerster Executive Director, J-PAL Department of Economics, MIT PSI June 6, 2013
Overview • The problem • Citizens as direct monitors • Citizens as voters • All impact results based on randomized evaluations
The problem • Substantial increase in democracy but often very imperfect • Very low levels of voter knowledge • Persistence of corruption, vote buying • Little debate about policies • Increase in policies that benefit the majority • E.g free primary education • Still very poor services
High levels of provider absenteeism Absence Rates India Indonesia Uganda Bangladesh Uganda India Peru Indonesia Ecuador Bangladesh Peru 4
Direct monitoring by citizens • Encouraging local monitoring has mixed results • in Uganda for health, reduced absenteeism and infant mortality (Bjorkman and Svensson (2009) • In Indian for education, no impact on effort or service (Banerjee et al, 2010) • Existing beneficiary control systems were nonfunctional at project start • Top down audits more effective than increased community monitoring to reduce corruption in Indonesian road projects (Olken, 2007)
Improving representative democracy • Encouraging evidence that voting behavior is malleable to campaigns • Reduced ethnic voting 9% (Banerjee et al, 2010) • Increased female voting 12% (Gine and Mansuri, 2010) • Reduced violence by 11% (Collier and Vicente, 2008) • Mainly tested very specific information/messages • Can this be applied to a spectrum of aspects of quality? • Can it generate • increase competition on economic/quality issues? • Improvements in policies and services?
Info on corruption: Brazil • Federal audits of municipal governments • Randomized order of audits • Some results released prior to election, some afterwards • Allowed test of: • whether voters punish corruption, • how much do they punish corruption • and in what circumstances
Voting in urban India • Scorecards on municipal candidates disseminated in random slums • Legislative activity and committee attendance • Discretionary expenditure • Incumbent and challenger qualifications • High performing incumbents benefit from transparency, low performing are hurt • Respond to information that effects them • Spending on their slum Banerjee et al, 2011.
MPs in Uganda • Scorecards created for all MPs by AFLI • Attendance, participation, initiative in parliament and committees, peer reviews • Activity and accessibility in constituency • Randomly announced in advance where there would be intense dissemination of the scorecards • Change in short run self reported voting intentions • No change in actual voting • No change in politician behavior Humphreys and Weinstein, ongoing.
MP Debates in Sierra Leone Bidwell, Casey, and Glennerster, ongoing
Pre and post debate knowledge and stated preferences • 5% change stated party preference • 13% change to undecided
Outstanding questions • Can community monitoring/empowerment be made to work consistently and cost-effectively? • What are the key design elements to make it effective? • Is community monitoring only effective in certain environments? • Can providing information up the chain of command improve governance? • Can informed voters lead to better quality services? • Can we bring together the community monitoring and informed voter work? • For example, provide information on health worker absenteeism at district level in different districts prior to elections