| Field | Before | After |
|---|---|---|
| Field Trial Status | Before on_going | After completed |
| Field Abstract | Before Regulation and enforcement around the use of public funds can reduce corruption, but does it also alter incentives to spend? In this paper, we investigate whether compliance uncertainty around spending rules can stifle valuable-to-the-public spending and distort policy choice. We begin by leveraging administrative data of local government accounts to establish a puzzling phenomenon: substantial shares of local budgets in Brazil are not spent despite clear needs for additional resources for public services. We also investigate how random audits, which might increase the salience of compliance uncertainty, affect public spending. Then, we leverage a collaboration with the Brazilian Council of Municipal Health Secretaries (CONASEMS) to conduct an experiment with municipal health secretaries in order to shed light on the impact of compliance uncertainty on secretaries’ choices. In the experiment, we offer valuable-to-the-public spending plans randomly varying the degree of compliance guarantees from the audit courts. | After Regulation and enforcement around the use of public funds can reduce corruption, but does it also alter incentives to spend? In this paper, we investigate whether compliance uncertainty around spending rules can stifle valuable-to-the-public spending and distort policy choice. We leverage administrative data and a collaboration with the Brazilian Council of Municipal Health Secretaries (CONASEMS, which organizes large conferences with municipal health heads responsible for managing, on average, a USD 1 million health budget. We first present three descriptive facts of the context consistent with the importance of compliance uncertainty for aggregate public spending. First, 20% of municipalities have experienced substantial incomplete spending for more than 10 years. Second, around one-fourth of the mayors are convicted with large penalties, even when there is no incidence of corruption. Lastly, regulation complexity is associated with incomplete spending (in the administrative data and a policymakers' perception in a survey). We then establish the causal link between compliance uncertainty and the spending decision, in a low-stakes lab-in-the-field experiment at CONASEMS conferences. We offer management support to implement a valuable public policy, namely a children's kit to reward parents who complete the vaccination scheme. We vary whether the children's kit includes "toys’’ (vs. "hygiene’’) items to be purchased with health earmarked funds. Preliminary evidence from the pilot shows that the toys policy bundle is perceived to be equally effective, but has a 20% reduced willingness to pay. Consistent with a risk-effectiveness trade-off, the toys policy bundle is perceived to have a higher risk. |
| Field Trial End Date | Before February 15, 2024 | After June 18, 2025 |
| Field Last Published | Before July 15, 2023 05:26 PM | After August 08, 2025 07:42 AM |
| Field Intervention (Public) | Before We administer a survey experiment to local bureaucrats in the health sector that seeks to understand their demand for spending plans that facilitate the execution of resources. We will introduce exogenous variation in the degree of compliance uncertainty around spending rules that each plan entails by randomizing the content of the provided information. We vary the source of the public funds used by the plans since rules and oversight vary depending on the source of funding in Brazil. We also randomly assign information on whether the spending plan will be interpreted as compliant with the rules by the control agency. With these random variations we will be able to understand whether reducing compliance uncertainty increase the implementation of effective health interventions. And whether this mechanism distorts spending of public funds. The survey experiment will be implemented in collaboration with CONASEMS, the Brazilian Council of Municipal Health Secretaries. We will administer it during CONASEMS conferences organized over time, starting the pilot at CONASEMS annual congress in July/2022. These events bring together local bureaucrats in the health sector from all over Brazil. | After We study the causal effect of compliance uncertainty on spending decisions in a lab-in-the-field experiment with municipal health heads, in partnership with the Brazilian Council of Municipal Health Secretaries (CONASEMS). We offer policymakers management support to implement a valuable public policy, namely a children's kit to reward parents who complete the vaccination scheme. Crucially, we introduce exogenous variation in the compliance uncertainty associated with the public policy while holding constant its perceived effectiveness. We vary whether the children's kit includes "toys’’ (vs. "hygiene’’) items to be purchased with health earmarked funds. With this random variation, we will be able to understand whether compliance uncertainty around the regulation of public funds alters incentives to spend and distorts policy choice. The experiment will be implemented in collaboration with CONASEMS, the Brazilian Council of Municipal Health Secretaries. CONASEMS organizes a variety of conferences throughout the year, in which all Brazilian municipal health heads participate. We will administer it during the 2025 National Congress of CONASEMS in June 2025. |
| Field Intervention End Date | Before February 15, 2024 | After June 18, 2025 |
| Field Primary Outcomes (End Points) | Before Our outcome variable will be participants’ demand for different strategies to execute resources. Following a Becker-DeGroot-Marschak (BDM) procedure, we will elicit participants’ maximum willingness to pay to receive these strategies. This will inform us about the local health officials’ demand for compliance guarantees. | After Our outcome variable will be participants’ demand for the proposed management support. Following a Becker-DeGroot-Marschak (BDM) procedure, we will elicit participants’ maximum willingness to pay to receive this support. This will inform us about the local health heads’ demand for compliance guarantees. |
| Field Experimental Design (Public) | Before The survey experiment will first describe some of the main problems that municipalities are dealing with in the health sector. We will present to health officials several initiatives that aim to overcome those challenges. These solutions will be presented along with a strategy to execute resources. We will use the Becker-DeGroot-Marschak (BDM) method to elicit participants’ demand for the proposed strategies. In the first round, the sample will be randomized into treatment and control group. The treatment group will receive information increasing the guarantee that the control agency will approve the expenses once the strategy to execute resources is implemented. This treatment reduces the compliance uncertainty associated with budget execution. This variation will let us test whether reducing uncertainty over compliance with spending regulation increase the implementation of effective health interventions by local bureaucrats in the health sector. In the second round, we will introduce variation in both the effectiveness of the proposed policy and the source of the public funds. We will test whether the compliance uncertainty raised by the execution of non-discretionary and more heavily regulated funds distorts public policy decisions as it makes local secretaries switch to safer but less effective public policies. We will also randomize in the second round the information on whether the spending plan will be interpreted as compliant with the rules by the control agency. With this variation, we will test whether the distortion can be explained by regulation risk. Finally, to be able to interpret the magnitude of the willingness to pay by local bureaucrats, we will introduce variation in another relevant dimension for policy implementation, as a benchmark. | After The survey experiment will first describe one of the main challenges that municipalities are dealing with in the health sector. We will present health officials with a valuable incentive public policy that aims to overcome this challenge. This initiative will be presented along with a proposal for management support to facilitate its implementation. We will use the Becker-DeGroot-Marschak (BDM) method to elicit participants’ demand for the proposed management support. Our experiment introduces exogenous variation in the compliance uncertainty associated with budget execution by randomizing the components of the incentive policy. Specifically, the treatment version includes non-health sector items to be procured with funds earmarked for health. This variation will let us test whether compliance uncertainty around the regulation of public funds alters incentives to spend and distorts policy choice. We also examine whether increasing the salience of the prosecution risk makes policymakers react more to compliance uncertainty. To do so, we randomize information about prosecution risk using data from the Federal Audit Court (TCU). Finally, we assess how much policymakers value a proposal that helps reduce this uncertainty. To this end, we randomize the content of the management support proposal, with some participants receiving a version that includes additional risk-reduction components. |
| Field Planned Number of Observations | Before 400 individuals per event | After 700 individuals |
| Field Sample size (or number of clusters) by treatment arms | Before We will randomize the sample into four groups. It will vary the information they receive associated with the spending plan. 1. Non-discretionary funds and low regulation risk 2. Non-discretionary funds and high regulation risk 3. Discretionary funds and high regulation risk 4. Non-discretionary funds and high regulation risk + benchmark constraint At the same time, each group will be randomized into an effective and less effective public policy. | After We do with three independent randomizations: Randomization A: Compliance Uncertainty Treatment A (High compliance uncertainty): 50% Control A (Low compliance uncertainty): 50% Randomization B: Salience of Prosecution Risk Treatment B (Increased salience): 50% Control B (No salience): 50% Randomization C: Management Support with Risk-Reduction Components Treatment C (With risk-reduction components): 30% Control C (Without risk-reduction components): 70% |
| Field | Before | After |
|---|---|---|
| Field Document | Before |
After
PAP_June 2025_AEA.pdf
MD5:
51f3aea6a97a7d874ba607762b3ef149
SHA1:
afd31d3884d7e9ce5fa05af509d2a64a7971b94a
|
| Field | Value |
|---|---|
| Field Document |
Value
PaP_July2023_AEA.pdf
MD5:
ac08dab21cd8968ff15aaa1334a84ef9
SHA1:
bc05fb8ae9fa8ea56e9b1cdb6a337a68545f2da9
|