close× Call Us
Title Can Access to Information Increase Community Monitoring & Service Provision? Evidence from a School Intervention in Mexico
Post date 07/19/2018
C1 Background and Explanation of Rationale Misappropriation and leakage of public funds remains a major impediment to improve service provision for poor people in developing countries (World Development Report, 2004). As a solution, policymakers and academics have championed information as a tool to make citizens more effective advocates for their rights. While there is some evidence that information can reduce leakage of resources (Reinnika and Svensson 2003), and increase citizens’ access to their entitlements (Banerjee et al. 2016), the success of these interventions depends on the willingness of a reformist government to disseminate relevant information. In contexts where government agencies do not disseminate information, or do so partially, Freedom of Information (FOI) laws could help citizens to learn about the “benefits and services they are entitled to and whether they are receiving the right amounts” (Banisar 2006, 7). Growing anecdotal evidence suggests that FOI laws have helped poor people gain access to government-subsidized food programs, and education. Yet there is little systematic evidence of the causal relation between the right to information and public service provision (Darch and Underwood 2009). A notable exception is Peisakhin and Pinto (2010)’s study, which draws from a field experiment in India where a group of confederates where randomly assigned to submit an information request under the Right to Information Act shortly after filling an application for a ration card. Compared to the control group, confederates in the treatment group were more successful in getting a card, and their waiting times were substantially shorter. Although these findings are encouraging, it is unclear if they can be extended to contexts where the benefit to be accessed is not a private, but a public good. We partnered with Mexico’s National Institute for Access to Information (INAI), and the local NGO Ciudadanos por la Transparencia (CxT) to evaluate the effects of two interventions. In a first treatment arm, INAI and CxT conduct training sessions to encourage the use of the access to information system among parents with school-aged children in Mexico. During the training session, trainers describe the workings of the system, explain how to submit requests of information, and illustrate how the system works using as an example a federal transfer program, Escuelas al Cien (E100), which transfers resources to public schools in impoverished areas of the country with the objective of improving schools’ infrastructure. After showing parents the publicly available information about the E100 program, and how to request additional information, trainers encouraged parents to demand accountability and to monitor the completion of the E100 works in their schools. In a second treatment arm, INAI and CxT disseminate information about E100, including the total amount of resources allocated to a school, the amount of resources disbursed, the progress of projects, and avenues through which parents could request more information and demand accountability. The effects of these two treatment arms will be compared to a control group. The study takes place in the state of Quintana Roo, Mexico. Our sample includes 100 preschool and primary schools, which have been allocated E100 resources, and are located in municipalities with low levels of criminal activity. Although federal resources have been allocated, there is a substantial delay in the implementation of the program. In March 2018, the program reported that in average 41% of the allocated resources had been disbursed to the schools in our sample, and 23% of the public works had been completed. From this group of 100 schools, 40 schools were randomly assigned to receive a training session, 20 schools were randomly assign to receive the information, and 40 schools were randomly assign to the control group. Out of the 40 schools receiving the training sessions, in 20 randomly selected schools the school principal was encouraged to attend the training session, and in the rest there was not such explicit encouragement. We used simple randomization to assign schools to the various experimental groups. In the schools assigned to receive a training session, CxT recruited 10 to 15 parents to participate in the session. In the schools assigned to receive the information, CxT distributed flyers to approximately 20 percent of the parents. At endline, we survey 15 parents per school in all our experimental groups. Hence, our sample will include 100 schools, and 1500 school parents.
C2 What are the hypotheses to be tested? The objectives of the study are to assess whether the interventions motivate citizens to demand accountability from their schools, as well as to increase their access to services and programs to which they have a right. Broadly, we expect that the training and information interventions will lead parents to try to monitor the public resources allocated to schools, both in general and specifically with respect to the E100 program. To measure these outcomes, we will submit requests of information to measure if the relevant government agencies have received requests of information pertaining to the implementation of the E100 program in our experimental schools. In addition to formal information requests, we will examine – through survey questionnaires – whether parents interacted with the schools through other methods such as by reaching out to school principals or the PTA. If monitoring of public resources by parents is successful, this may lead to an improvement of the school facilities and infrastructure. To measure these outcomes, we will collect administrative data of the E100 program, including the share of allocated resources that have been disbursed, and the progress of projects, if funded. We will collect this information 2 and 1 months prior to the intervention, as well as 2 and 4 months after treatment. Finally, with the endline survey, we will explore if the interventions influenced parents’ evaluations of their school’s infrastructure. Please see the attached pre-analysis plan for more detailed information about our hypotheses, measurement strategies, power calculations, estimations, and sub-group analyses.
C3 How will these hypotheses be tested? * To estimate intent-to-treat effects of the interventions, we will begin our analysis specifying a difference in means test, or differences in proportions, comparing: 1) schools assigned to the training sessions versus schools assigned to control; 2) schools assigned to the information-only treatment versus schools assigned to control; 3) schools assigned to the training sessions versus schools assigned to the information-only treatment; and 4) schools assigned to the training session without the school principal versus schools assigned to the training session with the school principal. Next, we will estimate treatment effects with difference-in-differences models when baseline measures of the dependent variables are available. At the individual level, for the means (proportions) tests, we will average outcomes at the school level, and conduct the tests with the grouped data. Then, using the individual level data, we will estimate effects regressing outcomes on treatment dummy variables, using school fixed effects. We will report standard errors clustered at the school level, which is the unit of randomization, for individual outcomes. Hypotheses tests will be conducted using two-sided tests, and p-values from corresponding t-tests will be reported. As mentioned, our plan is to estimate intent-to-treat estimates. However, we will also report compliance with assignment to treatment. In particular, in the training sessions we collected attendance data to record parents’ participation and the presence or absence of the school principal.
C4 Country Mexico
C5 Scale (# of Units) 100 schools, and 1500 individuals within experimental schools.
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number IRB Protocol ID: 2000021879
C9 Date of IRB Approval 10/20/2017
C10 Will the intervention be implemented by the researcher or a third party? 1. Instituto Nacional de Transparencia, Acceso a la Información, y Protección de Datos Personales; 2. Cuidadanos por la Transparencia
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? Yes
C13 JEL Classification(s) C93, D73, D83, H41, H52, H72, H75