Governance Initiative: Funded Projects

PDF version

The following is the list of projects funded by the Governance Initiative.

Jump to Seventeenth Round (Fall 2019)
Jump to Sixteenth Round (Spring 2019)
Jump to Fifteenth Round (Fall 2018)
Jump to Fourteenth Round (Spring 2018)
Jump to Thirteenth Round (Fall 2017)
Jump to Twelfth Round (Spring 2017)
Jump to Eleventh Round (Fall 2016)
Jump to Tenth Round (Spring 2016)
Jump to Ninth Round (Fall 2015)
Jump to Eighth Round (Spring 2015)
Jump to Seventh Round (Fall 2014)
Jump to Sixth Round (Spring 2014)
Jump to Fifth Round (Fall 2013)
Jump to Fourth Round (Spring 2013)
Jump to Third Round (Fall 2012)
Jump to Second Round (Spring 2012)
Jump to First Round (Fall 2011)

Projects Funded during the Seventeenth Round (Fall 2019):

Attacking Tax Evasion in Production Networks: Theory and Evidence from Paraguay
Researchers: Michael Best, Gaston Pierri, Evan Sadler
Location: Paraguay 
Type of project: Pilot study
Our project will combine new theory and an RCT in Paraguay to provide new in-sights into tax evasion by firms and how policymakers can respond to reduce evasion. We extend the canonical model of tax evasion in two ways. First, we model firms re-porting transactions with their trading partners, creating reporting spillovers through production networks. Second, we consider two types of tax reporting equilibria: Uni-lateral reporting and collusive reporting, and derive optimal enforcement targeting rules under both. The experiment combines two sets of treatments in a factorial de-sign. First, treated firms are assigned to one of three groups: I) An Electronic Billing Machine (EBM) linked to the tax administration; II) Desk audit; or III) Both EBM and audit. Second, clusters of firms are assigned to one of three targeting rules: A) Status quo: risk-rated targeting; B) Optimal targeting under unilateral evasion; and C)Optimal targeting under collusive evasion 

“Can you hear me now?”: Experimental evidence on improving public service delivery through non-electoral citizen participation" 
Researchers: Asim I. Khwaja, Tiffany M. Simon 
Location: Pakistan
Type of study: Pilot
We examine how citizens can hold policy actors (bureaucrats and elected politicians) accountable for public service delivery outside of election cycles. We do so by piloting a randomized controlled trial that tests the effectiveness of different methods of community-based mobilization of citizens in improving public school outcomes in Pakistan. Our initial research and fieldwork funded by a J-PAL GI Proposal Development Grant highlights two important dimensions to such participation: (i) policy actor type – whether citizens approach a political or bureaucratic actor and (ii) citizen gender–whether the citizens participating are (primarily) women or men. This pilot will implement a small rollout of the full study and inform its final conceptual and empirical design.  

The Selection and Making of Civil Servants: Evidence from China’s College Graduate Civil Services Program 
Researchers: David Y. Yang, Shaoda Wang 
Location: China
Type of project: Pilot 
The quality, incentives, and ideologies of local civil servants greatly affect policy implementation and governance efficacy. This project aims to understand what shape the background, preferences, attitudes, and behaviors of civil servants in rural China. We propose to randomly change selection algorithm in half of the job postings of the China’s College Graduate Civil Services Program in Guangdong. Combining the field experiment with surveys and official social security records, we will separately identify the selection margins (who apply to become local civil servants and which applicants the government selects) and the treatment margin (to what extent civil service experience changes the recruited candidates over time). We will provide the first empirical evidence on the treatment effect of civil service experience. It is important to incorporate the treatment margin in order to design the optimal selection mechanism, especially when there are tradeoffs between characteristics that the government deems desirable. 

Building Social Capital to Promote Political Engagement in Heterogeneous Communities 
Researchers: Leonardo Bursztyn, Ingar Haaland, Martin Kanz, Aakaash Rao, Christopher Roth, Anisha Sharma
Location: India 
Type of project: Pilot 
We study the extent to which heterogeneous groups are less capable of solving collective action problems and the mechanisms through which group heterogeneity operates. We hypothesize that social frictions induced by religious or socioeconomic diversity impede group coordination and cooperation, and thus interventions that reduce these frictions may enhance a group’s ability to solve collective-action problems and increase political participation. We test these hypotheses in a unique experimental setting: local WhatsApp groups in New Delhi, India formed for the purpose of organizing a community workshop with an NGO to address high levels of air pollution. We randomize the composition of these groups, and we further cross-randomize whether groups receive an intervention designed to reduce social frictions by emphasizing a common identity among group members. We examine effects on the success of the event, and we shed light on mechanisms with lab-in-the-field evidence. In addition to their relevance to the academic literature on public good provision and social capital, our results will inform policy for enhancing intergroup cooperation and effectively using social media as a tool to improve governance outcomes.

Projects Funded during the Sixteenth Round (Spring 2019):

Direct and Indirect Effects of Tax Enforcement: Experimental Evidence from the D.R.C.
Researchers: Augustin Bergeron, Nathan Nunn
Location: Democratic Republic of Congo
Type of Project: Full study
Tax compliance is a key challenge for low-capacity states seeking to expand their tax base. A first-order question that remains largely unanswered in the developing world context is: what are the direct and indirect /network effects of government’s enforcement efforts on tax compliance and bribe payments? This project will generate experimental evidence by randomly allocating connected firms to different tax enforcement treatments that vary in intensity or to a control condition in the context of a tax collection campaign in the city of Kananga in the Democratic Republic of the Congo.

Technology and Citizen Engagement in Local Tax Collection: Evidence from Ghana
Researchers: Karthik Muralidharan, Anders Jensen, David Lagakos
Location: Ghana
Type of Project: Pilot
Ghana, like most other developing countries, is characterized by low state-capacity for tax collection. The lack of tax revenues for public good provision is felt especially keenly at the local government level, where revenue collections and hence public good expenditures are minimal. In this study, we propose to conduct a pilot randomized controlled trial of a new technology that seeks to help local governments in Ghana increase property tax revenues. The main treatment group of households will receive visits by revenue collectors using the new technology, which replaces cash payment of taxes with electronic billing and payment. Two further, more exploratory, treatment groups will also receive presentations by national service personnel explaining the costs of noncompliance under the new technology or the benefits of compliance. The pilot will provide critical inputs to inform the design of a larger full-scale RCT including refining implementation details, measurement instruments, plausible effect sizes, and power calculations.

Spotlight: A Field Experiment on Media Coverage and Local Governance
Researchers: Dylan Groves, Don Green
Location: Tanzania
Type of Project: Pilot
Does media attention improve local governance? In collaboration with two Tanzanian District Commissioners and two local radio stations, I propose to pilot a novel field experiment to assess the impact of investigative media reports on village water service delivery in rural Tanzania. The core of the experiment will randomize villages to receive visits from investigative radio journalists, who will record 30-minute radio segments about rural water service challenges featuring interviews with concerned community members. I propose a 2x2 factorial design to test two mechanisms of media influence: informing political authorities of service delivery failures and citizen needs, and incentivizing political responsiveness by widely publicizing local grievances. The pilot will implement the treatments in 60 villages, with support from government and NGO partners for a larger scale implementation based on the results. 

Teacher Rotation and Student Outcomes: Experimental Evidence from Uganda
Researchers: Frederico Finan, Ernesto Dal Bo, Kizito Omala, David Schonholzer, Isabelle Cohen
Location: Uganda
Type of Project: Full study
Can centralized teacher transfers improve job performance? One of the biggest challenges in supplying education services to poor populations is low personnel morale, low engagement and high absenteeism. This problem is particularly severe in Uganda, where teacher absenteeism rates are estimated at 27%. However, Uganda’s district- centralized teacher deployment system, where teachers are transferred every five years and can be placed in any school in a given district, offers a unique opportunity for low-cost motivation of teachers. In partnership with a rural district in Eastern Uganda, researchers are undertaking a randomized controlled trial to study whether rewarding top-performing teachers with a future posting of their choice incentivizes teachers to improve their attendance and the quality of their teaching. After a successful launch of field operations and a first year of incentivized transfers, we are now aiming to secure funding to complete the study.

Can Media Empower Women? Experimental Evidence on Training Journalists in Iraq
Researchers: Fotini Christia, Horacio Larregui
Location: Iraq
Type of Project: Pilot
Empowering women to participate equally as citizens and decision makers is a critical element of good governance. There are few contexts where this might be more important than post-ISIS Iraq, where strengthening and expanding women’s voices, while simultaneously shifting the attitudes of men, may help to put political and economic inclusion at the center of the reconstruction process. In partnership with Internews, we intend to pilot randomized interventions that target the beliefs and capacity of journalists on gender-related issues to estimate direct effects and spillovers through their professional networks, and leverage randomized variation in the dissemination of resulting media content to evaluate effects on citizen attitudes.

Improving Labor Courts: Fighting Corruption with Competition
Researchers: Enrique Seira Bejarano, Frederico Finan, Ernesto Dal Bo, Joyce Sadka 
Location: Mexico
Type of Project: Full study
Labor courts are essential for addressing grievances adequately and for well-functioning labor markets. Unfortunately, courts in developing countries function poorly. In Mexico, the law indicates that all lawsuits should be adjudicated within 3 months, but courts face backlogs of about 4 years. Delays are mostly due to the fact that a trial cannot start without the defendant first being notified in person by a court employee (notifier). But a notifier will often fail to notify a defendant because the defendant has paid him a bribe not to, or because delay helps him extract bribes from the plaintiff. In recent years, only in 30% of cases is the first hearing held on the date initially assigned when the case is filed. On average the first hearing is postponed 3 times, taking well over one year to be concluded. We propose an intervention to increase notification rates and reduce corruption in the labor courts. Notifiers are currently assigned to areas within which they act as the sole notifier, giving them monopoly control over notifications in their area. Thus, a notifier is well positioned to extract bribes from both the defendant and plaintiff. The notifier can guarantee the defendant that a notification will never take place and can simultaneously threaten the plaintiff with never making the notification. Our intervention will randomly rotate notifiers across areas. Under a rotation scheme, a notifier cannot guarantee to the defendant that, in exchange for a bribe today, a notification will continue to fail in the future. The defendant will likely need to bribe additional notifiers to keep the notification from happening. Similarly, under the rotation scheme the notifier cannot threaten the plaintiff with the notification failing in the future if a bribe is not paid today. Our main hypothesis is that the elimination of exclusive control by notifiers will decrease corruption and increase notification rates. We will measure the effect of our intervention on the probability that a case is notified (reflected in administrative data), on the speed of notification, and for which types of workers, firms, and notifiers the increase is largest. We will also measure effort, proxied by the number of firms visited per day and kilometers travelled by the notifier. This information will be captured by a GPS-enabled smartphone that the courts plan to introduce for all notifiers to use. We will gather survey-audit data on firms to determine if the address of the defendant exists, if firms were visited by the notifier, and if notifiers asked for a bribe. We will also collect worker surveys to measure if their lawyer asked them for a bribe for the notifier. The policy is scalable and of current relevance as the Mexican congress is in process of contemplating the adoption of a notifier rotation policy.

Taxation Toward Representation: Public Goods Provision, Tax Collection, and Democratic Accountability
Researchers: Jeremy Magruder, Benjamin Krause
Location: Haiti
Type: Full study
While theoretical and historical work has explored the relationship between public goods, taxation, and the formation of democratic institutions, there is a considerable gap in empirical evidence. Haiti is piloting a system for managing federal transfers and collecting municipal taxes which provides an opportunity to study the efforts of a weak state with extremely low tax compliance as it attempts to formalize society. We have partnered with the mayor of one of the largest cities in the country to implement a city-wide RCT of public goods, tax collection, and public exposure of tax compliance to identify effective interventions and disentangle channels for transitioning from a zero-tax equilibrium as well as the impact of such efforts on participation in formal channels of accountability including voting as well as participation in informal taxation and other non-governmental forms of civic engagement.

Making Government Purchases More Transparent and Effective
Researchers: Claudio Ferraz, Mauricio Romero, Enrique Seira (ITAM)
Location: Mexico
Type: Pilot

Projects Funded During the Fifteenth Round (Fall 2018):

Career Incentives in the Public Sector: Evidence from the Sierra Leonean CHW Program
Researchers: Erika Deserranno, Gianmarco Leon, Phillip Kastrau
Location: Sierra Leone
Type of Project: Full study
The proposed study will test the effect a steeper wage progression between low-tier employees (e.g., frontline service providers) and higher-tier employees (e.g. supervisors, bosses) has on organizational performance. Further, we will test whether if this effect differs in meritocratic vs. non-meritocratic promotion systems. While steep wage progression may incentivize lower-tier workers to perform better in a meritocratic regime (through career incentives), it may instead demotivate workers in a non- meritocratic regime (through fairness or inequality concerns). We collaborate with a national public organization - the National Community Health Worker Program in Sierra Leone - in which supervisors (high-tier workers) are in charge of monitoring and advising 9 health workers (lower-tier workers). When a supervisor decides to quit her position, one of the health workers is promoted to the supervisor position. Until now, promotions from health worker to supervisor were de facto made at the discretion of the head of each health unit and therefore mostly based on personal connections rather than merit. For this study, we collaborate with the Ministry of Health and Sanitation (MOHS) to shift – in a random half of 372 health units — to a meritocratic promotion system, which uses the number and quality of health services provided as the main input for promotion decisions. The experimental variation in meritocracy is cross randomized with variation in perceived wage steepness; i.e., in a random half of the 372 health units, health workers are informed about the large wage gap between their own salary and that of the supervisor (which at baseline is underestimated). The design will allow us to test how increasing the perceived vertical wage steepness affects low-tier workers’ effort in regimes that are more vs. less meritocratic, and whether this effect differs by worker ability and worker political connectedness. This study will leverage infrastructure from an ongoing study on pay-for-performance, which greatly contributes to its cost-effectiveness.1 Our results will contribute to our knowledge about the effects of career incentives in multi-tier organizations and help policy makers in Sierra Leone improve health care delivery in the country. 

Greener on the Other Side? Tax Collection and Horizontal Inequity
Researchers: Michael Best, François Gerard, Evan Kresch, Joana Naritomi, Laura Zoratto
Location: Brazil
Type of Project: Full study
This project works with the government of the city of Manaus, Brazil to improve compliance with the municipal property tax. Specifically, we aim to understand the role that horizontal inequity -- similar households being treated differently by tax policy -- plays in determining compliance with the tax. We do this by combining a novel survey experiment raising the salience of horizontal inequity with rich administrative data on tax liabilities and tax compliance. 

Building State Capacity and National Unity with Market Design: The Problem of Assignment in Kenya’s G-United Program
Researchers: Michael KremerWilla Friedman, Guthrie Gray-Lobe
Location: Kenya
Type of Project: Full study
This project examines a Kenyan government program, G-United, which places university graduates in primary schools around the country with the goals of boosting national unity and improving student learning. One part of the project examines the effect of the program on national unity objectives, including applicants’ and participating communities’ inter-tribal prejudice. Another examines the impact of market- design inspired strategies for assigning participants to positions on applicant retention, satisfaction, and performance. This part will examine whether an assignment system that incorporates applicants’ preferences will increase commitment and improve productivity, and will explore its impact on national unity objectives. Estimates of the effect of the assignment scheme on productivity and retention will, in turn, allow for an evaluation of the effect of the assignment scheme on overall program effectiveness, and of potential trade-offs related to the selection of different mechanisms. 

The Persistence of Inefficient Policy: The Case of Free Power in Punjab
Researchers: Nick RyanAnant Sudarshan
Location: India
Type of Project: Full study
Bad governance can be due not only to corruption but to deliberate, inefficient policy choices that pander to interest groups. The free provision of power to farmers in Punjab, India, is an example of such a policy. Farmers, given power for free, use too much, and have rapidly drained the groundwater resources of the state. The policy persists despite its high social costs due to the political power of the farmers’ unions. This project will conduct a large experiment, in partnership with the Government of Punjab, to offer lump-sum transfers in exchange for farmers’ voluntarily agreeing to face a marginal price for power. We use this experiment to study the persistence of inefficient policies. In particular, we will measure what farmers choose to enroll in the new tariff, and how this depends on both their expected financial gains and their political leanings (such as union affiliation). We will estimate adopters’ electricity demand to measure the efficiency gains from pricing power and use these estimates to learn about what the government’s redistributional objective must be in order to justify continuing the policy of free power. Improved state capacity, in the form of better monitoring and transfer technology, may now make targeted lump-sum transfers a viable alternative policy to the traditional and inefficient in-kind redistribution of electricity. This grant application is the result of over a year of preparatory work with the Government of Punjab culminating this June, 2018 in the launch of a pilot of the above transfer program. As a part of this pilot, farmers have voluntarily come forward to be metered for the first time in decades, and substantial lump-sum transfers have been made, showing the viability of the proposed intervention. The pilot is the flagship evaluation of J-PAL South Asia’s partnership with the Government of Punjab and has received the personal attention and approval of Punjab’s Chief Minister, for being of special importance to the state. 

The Effect of Information on Bargaining, Corruption and Trade: Evidence from Small Traders in Kenya and Uganda
Researchers: Eleanor Wiseman, Tavneet Suri, Jeremy Magruder
 Kenya and Uganda
Type of Project: Pilot
The research project focuses on a particular non-tariff trade barrier for small (and often informal) cross border traders. These traders (mostly women) face higher trade costs due to lack of information and corruption: they have to pay bribes to border agents, while crossing the border with their goods. An asymmetry of information between traders and border agents about what traders should be legally paying and what the trade procedures are may play a significant role in fueling corruption and informal trading. Moreover, little is known about how bargaining happens in the context of a bribe between traders and border agents and how the agreed upon bribe is set. Through creative data collection, this projects attempts to understand the role played by corruption/bribery in cross border trading and what the process of bargaining for a bribe looks like. In addition, through a RCT, I test whether information about the actual tariff to be paid at the border and trade procedures (based on the products/value they are importing/exporting) can lower the value of bribes paid at the border and/or affect small traders’ formal/informal route choices. Such reductions in trade costs may increase (and/or formalize) trade across the border and increase traders’ welfare. I also conduct an audit study (sending surveyors acting as traders across the border) to collect information about border agents’ behaviors and demand for bribes.

Governance, Citizenship, and Accountability: Community-Centered Development in the Ugandan Health Sector
Researchers: Joshua Greenberg, Dean Yang
Location: Uganda
Type of Project: Pilot
Widespread accountability gaps in the public health sectors of low- and middle- income countries account for many government failures to deliver available, affordable health interventions. In Uganda, local political leaders have the potential to mitigate accountability problems by monitoring service provision at government health centers—but few do so. Limited citizen-politician information flow and inadequate politician training may explain this observation. I therefore propose a pilot study of two interventions: (a) quarterly citizen reporting meetings with local politicians on health service delivery and (b) politician skills training on monitoring local government health centers. The study will be one of a few to evaluate governance interventions targeting political economy inefficiencies.

Peer Effects and Conformity in Bureaucratic Inefficiency
Researchers: David Schonholzer, Erin Kelley, Gregory Lane, Elisabeth Sadoulet
Location: Kenya
Type of Project: Pilot
We propose studying the role of conformity and peer effects in corrupt behavior among bureaucrats of a large public service provider in Kenya. Peer pressure has been shown to have the potential to increase productivity and may have the potential to discipline peers, and this may be particularly true when supplemented with pressure to conform to a stacked peer group of role models. Exploiting GPS-tracked company vehicle data combined with administrative personnel and task records, we first aim to construct measures of corrupt behavior and productivity and study their relationship with bureaucrat characteristics. Second, we plan to run a series of randomized controlled trials re-assigning worst offenders and new hires to above median peers and exposing them to a stacked set of role models in job training seminars. Finally, we reveal the existence of the monitoring technology to all bureaucrats and observe subsequent outcomes.

Projects Funded During the Fourteenth Round (Spring 2018):

Performance Scorecards and Government Service Delivery
Researchers: Martin Mattsson 
Location: Bangladesh
Type of Project: Full study
Corruption and delays in government service delivery are positively correlated at the cross-country level, but it is not known if the relationship is causal in either direction. We propose a model of how the ability of bureaucrats to cause delays increases the amounts of bribes they can extract. Our experimental evaluation will measure the effect of a simple and cheap monitoring intervention aimed at reducing bureaucrats’ ability to delay service delivery. The experiment will test the hypothesis that the intervention not only increases the speed of service delivery, but also reduces bribe payments by reducing bureaucrats’ ability to delay services to citizens not paying bribes.

Improving State Response to Public Grievances
Researchers: Aprajit Mahajan, Prabhat Barnwal, Shekhar Mittal 
Location: India
Type of Project: Pilot
The Delhi Government’s Public Grievance Monitoring System (PGMS) seeks to redress citizen complaints about public goods and services. Currently, the redressal process is slow, resolutions are of poor quality, and citizen satisfaction is low. We seek to pilot interventions towards improving the quality of service delivery of the PGMS by examining the interplay between local politicians, bureaucrats, and citizens. We aim to answer three questions about the process. First, does providing local politicians with detailed current information on constituent grievances improve the speed and quality of resolution? Relatedly, does informing bureaucrats and/or citizens about this information sharing with local politicians have any incremental effects? Second, we examine whether providing detailed performance indicators to senior officials improves the speed and quality of resolutions. Finally, we seek to test whether improving the process by which complaints are matched (using machine learning methods) to relevant departments and officials within departments speeds up grievance resolution.

Building State Capacity and National Unity with Market Design: The Problem of Volunteer Assignment in Kenya's G-United Program
Researchers: Michael Kremer, Scott Kominers, Guthrie Gray-Lobe, Willa Friedman
Location: Kenya
Type of Project: Pilot
This project capitalizes on an opportunity to redesign a centralized labor market match to address two questions related to improving personnel management in bureaucracies and improving national unity. The program, G-United, places Kenyan university graduates in primary schools around the country to improve student learning and boost national cohesion, but the program suffers from high levels of volunteer attrition. In the first part of this project, we apply insights from market design to improve the assignment system to incorporate applicants’ idiosyncratic preferences to reduce attrition and improve applicant satisfaction and performance. Second, the project will evaluate the program’s influence on applicants’ attitudes regarding inter-tribal trust and national identity. An intermediate output of this project is the development and evaluation of a novel, incentive compatible salary adjustment process, allowing for financial bonuses to compensate for less preferred assignments. It is hoped that this mechanism may have broad applications in civil service assignment in both the developing world and beyond.

Social Spillovers and the Gender Gap in Political Participation in Pakistan
Researchers: Saad Gulzar, Luke Sonnet, Muhammed Yasir Khan
Location: Pakistan
Type of Project: Full study
We study the gender gap in politics in Pakistan. We construct social networks of women in 30 villages and design an experiment that creates a mechanism to encourage women who are most likely not to participate in politics – those with few social ties – to attend women’s meetings, register to vote, turn out, and run for office. We randomly select women in the social network and incentivize them to organize regular meetings where we provide important political information and encouragement. We study if these social interactions around politics can help reduce the gender gap in political participation by explicitly considering social spillovers in the design and analysis stages.

Representation through Information: Bringing Politician Actions Closer to Citizen Preferences
Researchers: Asad Liaqat
Location: Pakistan
Type of Project: Full study
Do politicians possess accurate information about citizen preferences and can better information bring politicians' actions closer to citizen preferences? In a sample of 360 local elected representatives in Pakistan, I test how accurate politicians are about citizen preferences and conduct interventions aimed at reducing the distance between politician actions and citizen preferences. The interventions involve (a) providing local politicians with accurate information on citizen preferences and (b) making them aware that their existing information comes from selected samples. Outcomes include (i) politicians' policy recommendations to party leadership, (ii) voting outcomes for the politician's party in the upcoming General Election, (iii) politicians' outreach to voters and voter attitudes towards them, (iv) politicians' demand for more information, and (v) attention paid by politicians to voter attributes.

Performance Base Incentives in Multi-Layered Organizations: Evidence from Sierra Leonean Community Health Workers
Researchers: Erika Deserranno, Philipp Kastrau, Gianmarco Leon
Location: Sierra Leone
Type of Project: Full study
Pay-for-performance (P4P) has shown to be effective in improving job performance in both private and public organizations. Existing literature has studied the effect of P4P within a single tier of the organizational hierarchy, mostly focusing on frontline workers and rarely paying attention to their supervisors and/or managers. However, incentivizing one layer of the organization (either the lower- or upper-tier) might affect effort across both layers of the hierarchy, either through effort complementarities or fairness concerns. Further, given a certain budget it is also unclear how to structure incentive schemes between these different layers of the organization. Is it optimal to adopt flat incentive structures that reward different layers of the organization equally? Or instead adopt steeper incentives structures that reward one of the two layers more than the other? Our research project aims to provide the first evidence on financial incentives across different layers of the organization. We plan to test this in the context of a large national public organization that is structured around two key figures: community health workers (CHW) and their peer supervisors (PS).

Reducing Free-Riding in Public Goods: An Experimental Approach to Improving Payment Compliance for Trash Collection in the Dominican Republic
Researchers: Christopher Neilson, Jose Antonio Pellerano
Location: Dominican Republic
Type of Project: Pilot
This is a collaborative project with the City government of Santo Domingo to design and evaluate different interventions to improve payment compliance for public garbage collection. In a first stage, we will analyze data on payment histories, as well as undertake targeted interviews and focus groups to gain a deeper understanding of the determinants of payment compliance. In the second stage, we will design and implement several randomized controlled trials to evaluate different behavioral and non-behavioral approaches to reduce free-riding. The interventions considered are: 1) Simplification of communication and payment instructions; 2) A lottery for households with timely payments; 3) Informing households about the consequences on their credit history of not paying the bill; 4) Informing households about debt relief options, and 5) Informing citizens about actual compliance rates in high compliance neighborhoods. This will shed light on how to improve a state’s capacity to effectively collect payments for government services.

Projects Funded During the Thirteenth Round (Fall 2017):

Digitalization of Local Tax Collection in Urban Cote d'Ivoire
Researchers: Pascaline Dupas, Deivy Houeix
Location: Cote d'Ivoire
Type of Project: Pilot
The great majority of individuals in lower income countries are self-employed, which considerably limits the ability of governments to tax income. “Tax collectors” who walk from stalls to stalls on a daily basis to collect taxes in cash from vendors are ubiquitous in Cote d’Ivoire as in many other countries, but there is widespread evidence of collusion between tax collectors and tax payers. As many countries move towards greater autonomy of local governments, the ability of local governments to raise revenue is becoming increasingly crucial. This proposal aims to estimate the impact of moving from a cash to an electronic tax collection system in urban municipalities of Cote d’Ivoire. With the help of an Ivorian consulting firm called BMI, a number of municipalities are switching to a fully electronic payment system that will allow considerably greater monitoring of tax collectors performance. The digitalization is expected to increase not only compliance but also the tax base, leading to higher total revenue for local authorities. It could also increase accountability of local governments if compliant taxpayers are more likely to demand public goods than non-compliant ones. This offers a unique opportunity to: (1) understand the extent of collusion that was going on under the current (cash and paper-based) system; and (2) experiment with incentive schemes to limit new forms of collusion from emerging; and (3) study how local spending decisions are influenced by accountability to local taxpayers.

Tax Compliance and the Timing of Taxation: Experimental Evidence from Kenya
Researchers: Lorenzo Casaburi, Dina Pomeranz
Location: Kenya
Type of Project: Pilot
Increasing tax collection has become a key priority of developing countries, where tax evasion is a major concern. The project will investigate to what extent improving the match between the cash flow and the frequency of tax payments increases compliance among small firms that are subject to the turnover tax or value added tax (VAT). Models and analysis on the impact of taxation on firms in developed countries have often focused exclusively on how taxation affects firms’ profitability, rather than their cash flow. This assumes that firms can use credit to smooth cash flow over time. However, this is often not possible for most small firms, particularly in developing countries. Working in partnership with the Kenya Revenue Authority (KRA), the research team plans to design and pilot three interventions focused on the timing of taxation.

Improving Labor Courts: The Effects of Publicly Provided Lawyers and Statistical Information
Researchers: Enrique Seira, Joyce Sadka, Christopher Woodruff
Location: Mexico
Type of Project: Full study
Labor courts are essential for well-functioning labor markets and the provision of justice. In Mexico, labor courts are characterized by long delays in trials, low settlement rates, misinformation and overconfidence among plaintiffs, and lawsuit inflation. We aim to understand the extent to which these problems are caused by principal-agent issues between plaintiffs and private lawyers, and if they can be mitigated by 2 interventions: (1) encouragement to consult with (free) public lawyers, who have lower incentives to inflate expectations, and, (2) the provision of personalized statistical predictions of expected case outcomes as well as legally-mandated entitlements. We work with a sample of plaintiffs coming to the court before retaining a lawyer or filing a case. We measure the effect of treatment on pre-trial settlement rates, case duration and the amounts won by plaintiffs. We also measure lawsuit quality, expectations of parties about winnings, post-case satisfaction, and measures of consumption and employment.

Community Engagement and Lottery Efficiency: Learning from Nepal
Researchers: Abraham Holland
Location: Nepal
Type of Project: Pilot
What is the role of community preferences in driving the viability of village relocation programs following natural disasters? Nepal’s post-2015 earthquake village resettlement program, an effort to shift high-risk communities to safer, previously unsettled, locations provides a setting in which to examine this question. A key challenge in this program is how to fairly and transparently allocate new private residential space to households, while maintaining individual and overall community satisfaction with the relocation effort. In response, the government plans to administer a land lottery, which is selected via local plebiscite, in each affected village. This paper focuses on a potential driver behind the popularity of different lottery types, social preferences, and seeks to understand its implications for community engagement efforts supporting the plot allocation process.

Demand-Driven Enforcement of Labor Law in Bangladesh
Researchers: Laura Boudreau
Location: Bangladesh
Type of Project: Full study
Weak states with poor institutions often lack the capacity to implement and/or to enforce labor regulations aimed at improving working conditions. Increasingly, private actors are enforcing labor standards in these countries, but the effects of their interventions on local firms and workers is unknown. This research partners with a set of multinational retail and apparel firms to enforce local labor laws on their suppliers in Bangladesh. Specifically, I design and implement a randomized controlled trial with Bangladeshi garment factories, randomly introducing worker-manager safety committees in supplier establishments. I measure the impacts of this intervention on factories’ compliance levels and productivity as well as on workers’ welfare. I also explore the intervention’s broader effect on equilibrium compliance levels in the sector.

Institutional Reform and De Facto Women's Rights in Punjab, Pakistan
Researchers: Erica Field, Kate Vyborny
Location: Pakistan
Type of Project: Pilot
Women’s rights to make life choices such as whether and when to marry and divorce are intrinsically important. Yet in many contexts, the law on the books is substantially more progressive than the de facto practice of the law. Pakistan is a good example: the government officials who register marriages often follow their own judgment rather than comply strictly with the law. Many are not even aware of the laws on the books. To address this, the government of Punjab is conducting the first ever mandatory training of marriage registrars in the province to inform them on women’s rights in marriage, including child marriage and women’s right to initiate divorce, and of the legal consequences to registrars for non-compliance. If the study identifies an effect of the training on legal practices, it will prepare the ground for study of the impacts on women and their families through a household survey.

Teacher Rotation and Student Outcomes: Experimental Evidence from Uganda
Researchers: Frederico Finan, Ernesto Dal Bo, Kizito Omala, David Schonholzer, Isabelle Cohen
Location: Uganda
Type of Project: Full study
One of the biggest challenges in supplying education services to poor populations is low personnel morale, low engagement, and high absenteeism. This problem is particularly severe in Uganda, where teacher absenteeism rates are estimated at 27%. However, Uganda’s district-centralized teacher deployment system, where teachers are transferred every five years and can be placed in any school in a given district, offers a unique opportunity for low-cost motivation of teachers. In partnership with a rural district in Eastern Uganda, researchers will undertake a randomized controlled trial to study whether rewarding top-performing teachers with a future posting of their choice incentivizes teachers to improve their attendance and the quality of their teaching.

Mobilizing Local Leaders to Rebuild the Social Compact
Researchers: Asim Khwaja, Benjamin Olken, Adnan Khan 
Location: Pakistan
Type of Project: Full study
A significant challenge to the provision of local public services–water, sanitation, waste removal, etc.–in developing economies is the inability to raise adequate resources, especially through local taxation. In many countries, the social compact, whereby citizens agree to pay taxes to fund government services that are then credibly and transparently delivered, is broken. In this project, we investigate whether mobilizing local political leaders to strengthen the link between local collections and urban services can increase citizens’ willingness to pay for services, improve service delivery, and ultimately revitalize the social compact. We test this in major urban centers in Punjab, Pakistan selected for an ongoing randomized evaluation in which local governments are required to allocate tax revenue according to citizen preferences for local services.

Evaluations Funded During the Twelfth Round (Spring 2017):

Designing Performance Indicators for Career Incentives
Researchers: Elisabeth Sadoulet, Alain de Janvry, Qiong Zhang
Location: China
Type of Project: Full study
We consider a case of career incentives given to local civil servants hired by the Chinese Government to serve the broad needs of the population. In collaboration with the regional officials and a provincial research center focusing on local governance, we propose to test alternative performance evaluation schemes, ranging from the status quo of delegation to local officials with total discretion, to a transparent performance score based on easily observable high signal-to-noise ratio indicators. We will also test two intermediate schemes, a transparent scoring formula including a larger set but of lower signal-to-noise ratio indicators, and giving discretion to the local officials to decide on the weighting scheme but not the components of the performance score. The objective is to test whether the levels of precision, transparency, and delegation in assigning performance scores improve the power of incentives, and to detect potential trade-offs in performance along non-incentivized dimensions due to multi-tasking.

When Castes Collide: Does Competition Mediate Contact?
Researchers: Matt Lowe
Location: India
Type of Project: Full study
52% of upper castes in India continue to practice “untouchability.” These group relations persist despite castes living in close proximity (albeit in segregated villages). Further integration may be a natural policy response, but little is known about whether this would work, and about how the type of cross-caste interaction might matter. This field experiment tests both the effects of integration, and whether effects are weakened (strengthened) by competitive (collaborative) incentives. I use cricket tournaments to integrate men from all castes, with men randomly assigned to teams (giving variation in cross-caste exposure). The teams are then randomly assigned to Individual vs. Team Pay (competitive vs. collaborative incentives). Mid-experiment results show that (a) incentives affect group cohesion, with players on Individual Pay teams wanting to interact less with team-mates; and (b) incentives affect post-tournament behavior, with Team Pay stimulating greater cross-caste trade and trust in others. These early results suggest promise that incentives can be used to mould social relations. This proposal would complete the experiment with a larger sample.

Monitoring Publicly Funded Health Service Delivery through Private Agents: A Pilot Study with the Government of Rajasthan, India
ResearchersPascaline Dupas, Radhika Jain
Location: India
Type of Project: Pilot
Low quality care and leakages are known to plague the Indian public health system. Although the public sector is largely free (representing a large subsidy), poor households either forego care or pay high costs to go to the private sector, exposing themselves to considerable risk. In response, the Government of Rajasthan (GoR) has launched a program to cover the costs of secondary and tertiary care at empaneled private hospitals for low-income households, effectively outsourcing health care delivery to private agents. Private sector participation has expanded rapidly, comprising 70% of tertiary care 14 months into the program. However, the GoR is facing a new set of governance challenges in monitoring and appropriately incentivizing these private agents. Exploratory audits we conducted in January-February 2017 suggest partial capture of the subsidy by private hospitals, as well as over-reporting and over-provision. Our proposal for a pilot GI study is to conduct systematic audits that will enable us to estimate the rates, types, and incidence of fraud and leakage in the program. The findings will be used to design and test interventions that better align principal and agent objectives in three ways: top down monitoring of facilities, bottom up measures to improve patient information, and changes in reimbursement rates to meet private sector participation constraints. Identifying ways to effectively deliver health care through the private sector has the potential to increase the returns to scarce public health rupees.

The Elasticity of Tax Compliance: Evidence from a Low-Capacity State
Researchers: Augustin Bergeron, Jonathan Weigel, Gabriel Tourek
Location: Democratic Republic of Congo
Type of Project: Full study
Compliance is a key challenge for low-capacity states seeking to expand their tax base. A first order question that remains largely unanswered in the developing world context is: What is the elasticity of tax compliance with respect to the marginal tax rate? We will generate experimental evidence on this question by randomly distributing coupons that lower certain households’ property tax liability in the context of a tax collection campaign in the city of Kananga, Democratic Republic of Congo. The effects of the program on bribery will also be examined.

Establishing Responsive Linkages between Politicians and Voters
Researchers: Miriam Golden, Saad Gulzar, Luke Sonnet
Location: Pakistan
Type of Project: Full study
We partner with provincial legislators in Khyber Pakhtunkhwa, Pakistan to test whether integrated voice response (IVR) technology can springboard communication between politicians and voters and thereby improve accountability, responsiveness, and development. IVR allows politicians to record messages in their own voice and deliver them via robocalls; citizens can then respond to questions posed by pressing keys on their phones. The first stage of the experiment randomizes (at the household level) politician robocalls and elicitation of voter preferences. Stage two provides aggregated citizen preferences to the politicians, and randomizes (at the polling station level) a responsive call from the politician that is either generic or specific to the feedback received. We examine if either or both of these stages improve citizen trust in and engagement with government, political efficacy, support for the politician, and ability to sanction poorly performing representatives. The experiment offers legislators a direct channel to voters, bypassing traditional local elites and reducing the need for corruption and vote-buying. 

How Income Growth Shapes the State: A Village-Level Randomized Controlled Trial on Unconditional Cash Transfers, Community Engagement, and Local Politician Responses in Kenya
Researchers: Michael Walker, Kate Orkin
Location: Kenya
Type of Project: Pilot
Unconditional cash transfers to poor households are one of the most important innovations in poverty reduction in the last quarter-century. We build on two studies which randomly allocate the rollout of an unconditional cash transfer program, run by GiveDirectly, a foreign NGO, across 1,041 villages in Western Kenya, with variation in the proximity of rollout to the August 2017 election. We propose collection of community-level data on public goods provision, engagement with citizens by politicians, and citizen civic participation, combined with consolidation of administrative data on these issues. We study how political processes and politicians respond to changes in village wealth, and, because transfers are not state provided, we can conduct a clean test of whether, as modernization theory predicts, increases in income increase civic participation and levels of social capital. In addition, the large scale of the cash transfer program offers a unique opportunity to study politician responses.  

Politician Entry, Selection, and Performance in Sierra Leone
ResearchersKatherine Casey, Rachel Glennerster, Anna Schrimpf
Location: Sierra Leone
Type of Project: Full study
This project explores the relationship between regional party strongholds, where competition in the general election is weak and generates little accountability pressure, and the poor performance of elected officials. In such areas, the competition of import occurs at the primary stage, and internal party selection processes thus directly determine the characteristics and performance of elected representatives. In much of the developing world, these processes are opaque, controlled by elites, and likely hampered by information constraints. We propose to evaluate whether direct vote primaries, which broaden participation in the selection process to all card-carrying party members, and party primary conventions, which feature informative debates between aspirants, facilitate the election of more competent and accountable leaders. 

Projects Funded During the Eleventh Round (Fall 2016):

Age, Organization, and Accountability: Evidence from the DRC
Researchers: Nathan Nunn, James Robinson, Sara Lowes, Eduardo Montero
Location: Democratic Republic of Congo
Type of Project: Full study
This project explores the extent to which age sets, a traditional social structure in many African societies, can be exploited to improve local oversight, governance, and the provision of public goods. Age sets are initiation rituals that create cohesive groups of men of approximately the same age. A consequence is that the younger age sets in a village, because of their cohesion and allegiance to each other, provide a check on the balance of the village elite, including the chief, who belongs to a different and older age set. We are interested in whether programs that provide public goods are better managed by the village elite when age sets are present. Researchers are also interested in better understanding whether public goods provision can be improved if programs are adapted to account for the presence of age sets in a village.

Plugging the Leaky Bucket: Generating Experimental Evidence to Support Indonesia's Social Protection Reforms
Researchers: Rema Hanna, Ben Olken, Abhijit Banerjee, Joe Amick, Elan Satriawan, Sudarno Sumarto
Location: Indonesia
Type of Project: Full study
The Government of Indonesia (GoI) is embarking on a radical shift in Raskin, Indonesia’s largest targeted social assistance program, moving from in-kind transfers run by the government to a novel delivery system: restricted electronic vouchers that can be redeemed at any registered private sector retailer. The key reason for this reform is to reduce the extensive leakages and corruption that occur during the current pubic distribution process. Our previous research has shown that poor households only receive approximately one-third of their subsidy, and the GoI believes this problem can be alleviated by changing the structure of the bureaucracy’s delivery mechanism so that private sector agents, rather than government officials, are the ones handling the last-mile of service delivery. Yet, outsourcing creates its own new set of governance challenges in terms of how the government structures and then monitors its program. We are working with the GoI to design three experiments to test interventions intended to strengthen the transition to vouchers and private sector agents: a) strengthening the supply side, to ensure the elimination of private monopolies and ensuring high quality service, b) providing a tool for monitoring fraud, and c) ensuring that beneficiaries can navigate the technological and bureaucratic hurdles of the new e-voucher system. In this proposal, we request funds for the first experiment, to determine how the government can recruit a network of retailers to ensure sufficient high quality service, and eliminate the monopoly power that could lead to leakages in the new privatized system.

Political Capital and Performance of Politicians
Researchers: Yasir Khan, Saad Gulzar
Location: Pakistan
Type of Project: Full study
We propose to study if politicians improve performance in office through investment in political capital. We randomly vary three channels of investment; training, peer learning and mentorship. We also propose to study if peer learning takes place through social utility when peers are competitors.

Two Approaches to Community Development
Researchers: Katherine Casey, Rachel Glennerster, Edward Miguel, Maarten Voors, Erwin Bulte
Location: Sierra Leone
Type of Project: Full study
Decades of unaccountable leadership, conflict and underdevelopment have limited the reach of the state into rural Sierra Leone and left communities with a dearth of public services. This project explores two distinct mechanisms to bridge the gap between communities and the government. The first is an evaluation of the long run effects of a participatory community driven development (CDD) program, which devolved financial and implementation control over public services to communities themselves, accompanied by intensive social facilitation. The second assesses a low cost technocratic alternative that identifies and supports high competence community members to take better advantage of development programs offered by the government. It leverages local talent, addresses information barriers, and augments existing managerial capital with basic training in project management. A third component elicits expert beliefs about the efficacy of these two approaches, assesses their forecast accuracy, and evaluates whether beliefs change in response to new information.

Participatory Development and Social Preferences: Experimental Evidence from Bangaldesh
Researchers: Selene Ghisolfi, Serena Cocciolo, Jakob Svensson
Location: Bangladesh
Type of Project: Full study
Since the 1990s, Community Driven Development (CDD) has become an increasingly common approach for development assistance. This has led the academic literature to assess the extent to which CDD programs can affect social norms and behaviors in benefiting communities. One fundamental challenge in this evaluation consists in the difficulties of observing and measuring social preferences. An increasing number of studies rely on lab-in-the-field experiments in order to test whether participatory development enhances participation, trust and social values in receiving communities. The existing evidence is mixed, calling for further research on this topic, which has relevant policy implications on the potential long run impact of participatory governance. We will contribute to this debate by providing novel evidence on the impact of a CDD program on two dimensions which, to the best of our knowledge, have not been previously addressed by the existing studies: (i) fairness norms on equity and equality of redistribution; (ii) citizens' preferences over participatory decision-making.

Projects Funded During the Tenth Round (Spring 2016):

Agricultural Electricty Subsidy Reform and Power Leakage in Rajasthan
Researchers: Nick Ryan, Anant Sudarshan
Location: India
Type of Project: Pilot
In many Indian states, a quarter or a half of the power pumped into the grid disappears, and the other part is sold at prices far below cost. These implicit and explicit electricity subsidies are meant as a lifeline to farmers, allowing the spread of irrigation using electric pumps to extract groundwater. Instead, they have locked rural India into a destructive cycle of groundwater depletion and bad power supply. Researchers will collect data on the nature of power leakages and design an alternate subsidy program, which will transfer electricity subsidies as a lump-sum amount, instead of through low, distortionary electricity prices. A planned experiment with the Rajasthan power distribution companies will measure the effects of this subsidy reform on leakages, energy consumption, and revenue.

Citizen Demand for Corruption: Evidence from Roadway Tolls in the D.R. Congo
Researchers: Otis Reid, Jonathan Weigel
Location: Democratic Republic of the Congo
Type of Project: Full study
This study examines the determinants of petty corruption in the DRC. In pilot data, 42 percent of motorcycle taxi drivers reported paying less than the official rate at tolls. To explain this high rate of non-compliance, we randomly offer motorcycle taxi drivers one of three incentives to obtain an official receipt at the toll, or a control condition . Relative to an official toll of 2000FC (US$2.20), the incentives are: a 1000FC-2500FC payment to the driver, a 2000FC donation to a local health NGO, or a 2000FC donation to the government, earmarked for health. The first treatment induces drivers to either pay the official amount or demand a lower bribe price. The latter two treatments compare the responsiveness of tax-payer behavior to differences in the perceived effectiveness of the recipient of tax payments. We will cross-randomize with an audit treatment to compare the effectiveness of citizen-side interventions with bureaucrat interventions.

Can Smart Technology Systems Improve Direct Benefit Transfer Performance and Increase Participation? Evidence from MGNREGA in India
Researchers: Rohini Pande, Charity Troyer-Moore, Eric Dodge, Yusuf Neggers
Location: India
Type of Project: Full study
The implementation of social protection programs remains a challenge in developing countries, often to the particular detriment of the most vulnerable intended beneficiaries. We will investigate the potential of a new internet- and mobile-based management and monitoring platform, developed in direct collaboration with the Indian Ministry of Rural Development, to improve the administration of a large government welfare scheme. A randomized controlled trial across multiple states will be conducted in which we provide the platform to different levels of the bureaucratic hierarchy responsible for program administration. The study will determine the extent to which heightening officials' performance incentives versus lowering their costs of information acquisition is effective, and whether complementarities between the two exist, in reducing payment delays and subsequently improving program uptake. We will additionally examine how impacts are mediated by bureaucrats’ own personality traits.

Government Mobile Salary Payments in Afghanistan
Researchers: Michael Callen, Tarek Ghani, Josh Blumenstock
Location: Afghanistan 
Type of Project: Full study 
Governments must pay their employees for states to function. Frequent delays and leakage of salary payments can undermine government effectiveness. These problems are severe in Afghanistan. Mobile Salary Payments (MSPs)—a system enabling employees to receive their salaries directly via mobile money—present a potential solution. At scale, MSPs can also grow the mobile money ecosystem and promote financial inclusion. We propose a randomized evaluation of MSPs with the Ministry of Education for a sample of approximately 200,000 teachers spanning 340 of the 396 districts in Afghanistan. We also propose to evaluate an additional Call Detail Record (CDR)-based attendance monitorin