Work in Progress
"Do Rewards Work? Evidence from the Randomization of Public Works" (w/P. Carrillo, and E. Castro) Download draft (R&R)
Abstract: Positive inducements and rewards are becoming an instrument widely used by policymakers to generate good behavior. While certain incentive mechanisms, including moral suasion messages and rewards, have been shown to improve good behavior they tend to have a limited effect: they dissipate quickly and affect only the direct recipient of the treatment. Some inducements have backfired by crowding out intrinsic motivations. Under what conditions could positive inducements generate broad and persistent effects? This paper evaluates the effect of positive inducements on tax behavior by exploiting a natural experiment in which a municipality of Argentina randomly selected 400 individuals among more than 72,000 taxpayers who had complied with payment of their property tax. These individuals were publicly recognized and awarded the construction of a sidewalk. Results indicate that: i) being selected in the lottery and publicly recognized by the government has a positive but not persistent effect on future compliance; ii) receiving the sidewalk has a large positive and persistent effect; iii) high and persistent spillover effects exist: some neighbors of those who receive the reward comply more too, and these effects can be even larger than the direct effects; and iv) there is no financial motive effect; i.e., people do not pay their taxes just to participate in the lottery. Recognition serves only as a short-term incentive, but the provision of a durable and visible good has more persistent and broader effect by crowding in intrinsic motivations (through reciprocity and peer-effects). These findings provide evidence on features that make a positive inducement more successful, whether for tax compliance or other policy purposes.
"Don’t Blame the Messenger. A Field Experiment on Delivery Methods " (w/D. Ortega) Download draft (R&R)
Abstract: Does the method a government uses to communicate public policies to its citizens affect the effectiveness of those policies? Are impersonal methods such as a letter or an email as effective as those methods that include personal interactions, such as a phone call or a visit? The literature tends to avoid discussing the issue even though the chosen method may affect the effectiveness of an intervention and the interpretation of its results. This study reports the results of a field experiment in Colombia that varies the way the National Tax Agency contacts taxpayers who owe income, value added, and wealth tax payments. More than 20,000 taxpayers were randomly assigned to a control or one of three delivery mechanisms. Results indicate significant and siz- able differences across delivery methods. A personal visit is more effective than a letter or an email, conditional on delivery. The findings show that assessing the mechanisms through which policies are informed and publicized should be fully incorporated in the development literature.
"Compliance Spillovers Across Taxes. Results From a Field Experiment." (w/A. Lopez-Luzuriaga) Download draft (R&R)
Abstract: When the tax authority increases the enforcement for one tax, what happens to the level of compliance in other taxes (spillover effect)? In this paper, we present a simple analytical model that shows that the sign of the spillover depends on how taxpayers update their beliefs about penalties and detection probabilities for one tax after observing the deterrence actions the tax agency takes for another tax. The implications of the model are that penalties and increased detection are not necessarily interchangeable policy tools once we incorporate potential spillover effects. The results from a randomized field experiment that increases both penalties and detection probabilities are consistent with the expectations from the model. The evidence indicates that increasing the salience of fines and enforcement probabilities for those who do not comply with the payment of the property tax increases how much the same individual declares on the gross-sales tax. The result has ample implications for researchers bringing interventions to the field and for governments' enforcement strategies.
"A Heavy Hand or a Helping Hand? Information Provision and Citizen Preferences for Anti-Crime Policies" (w/D. Gingerich) Download draft (submitted)
Abstract: Welfarist justifications of democracy and the growing push for greater citizen participation in policy decisions presume the existence of an electorate with policy preferences that are responsive to pertinent information. Is this accurate, or are citizens' preferences for public policy dominated by pandering, lies, and exaggerations? We address this question by providing a model and empirical test of how citizens' policy preferences respond to information in the arena of anti-crime policy. Our model shows that preferences for anti-crime policy hinge on expectations about the levels of crime. If the crime rate is expected to be high, then citizens favor punitive measures. By contrast, if the crime rate is expected to be low, then citizens prefer policies that favor prevention and redistribution. Thus, information about the crime rate|by guiding citizens' expectations|plays a central role in shaping the policies citizens demand. We test our claim by using an information experiment embedded in the 2017 LAPOP survey conducted in Panama. Respondents were randomly assigned to three treatment groups: one shown a trend of homicides increasing, another shown a decreasing trend, and a third receiving no information. Consistent with the model, respondents in the first group were more inclined towards punishment and less towards social policy than respondents in the other groups. Yet because the treatment provided inaccurate information about crime trends in the country, their response was not guided by recent events but by the misleading framing of the message. This result helps explain the prevalence of suboptimal policies and the pervasive role of misinformation in policy decisions.
"Transparency and Trust in Government. Evidence from a Survey Experiment" (w/Bruno Cardinale L). Download draft (submitted)
Abstract: Does providing information improve citizens' perception about government transparency? Does all information matter the same for shaping perceptions about the government? This article addresses these questions in the context of an online randomized survey experiment conducted in Argentina. Results show that providing information to citizens matters for shaping perceptions about transparency, and the content of the information matters for affecting the evaluation people make about the government. Those who received a ``positive'' treatment (showing that the government was over-performing on its promises) increased their trust in the government more than those who received a ``negative'' treatment (showing that the government was underperforming). The evidence highlights that the channel between transparency and trust may be mediated by the performance of the government.
"Who Decides on Social Policy? Social Networks and the Political Economy of Social Policy in Latin America and the Caribbean" BOOK (w/A. Bonvecchi)
Download book prospect here
"Who’s calling? The effect of phone calls as a deterrence mechanism" (w/D. Ortega)
Abstract: There is an ample literature that focuses on the determinants of tax compliance. Several field experiments have evaluated the effect and comparative relevance of sending deterrence and moral suasion messages to taxpayers. Absent from the discussion so far has been evaluating the effect of different delivery mechanisms of those messages. We conduct a field experiment in Colombia in which we vary the way the National Tax Agency contacts taxpayers with due payments for income, value added, and wealth taxes. More than 34,000 taxpayers were randomly assigned to a control, or they were contacted by phone by the agency. Results indicate large and highly significant effects for increasing the payment of due taxes. These effects are higher than those usually obtained through the use of letters. The results in this paper have ample implications for the literature and for policy design. The experiment allows having the first estimates of the efficiency and effectiveness of using phone calls instead of letters, to argue about optimal enforcement strategies, and it sets the stage for further theoretical and empirical work. Moreover, it also provides evidence on the effectiveness of phone calls that could be compared to those in the get out the vote literature.
"Property taxation and noncompliance: Evidence from high-frequency panel data" (w/C. Traxler)
Abstract: This project studies property tax noncompliance in a large city in Argentina. We exploit a complex tax reform. The reform implied a series of quarterly or bi- annually nominal tax increases (or decreases) over a period of more than three years. We measure the behavioral responses to within-variation in property taxes using high-frequency panel data on the timing of the payment (and non- payment) of the monthly tax bills for all households and property owners in the city. Preliminary results from two-way fixed effects estimates indicate that higher (real) taxes lead to more delayed tax-payments and an increase in the rate of non-payment. Quantitatively, however, these effects are small on average. Studying heterogeneity of responses according to property values, we find much stronger responses for smaller and less valuable properties. In a second step, we analyze social-interaction effects. Using within-variation in a district’s minimum-tax level, we find that “unaffected” property owners respond to (minimum-tax induced) increases in non-compliance rates in their block. Using an updated wave of data we will substantiate this observation and translate our estimates into a social multiplier of non-compliance.
"Making People Pay their Debts: The role of salience, information, and computational costs in a tax amnesty (Pre-analysis plan)" (w/E. Castro) Download
Abstract: This document describes the pre-analysis plan of a set of interventions that will take place in the city of Santa Fe, Argentina. In these interventions we will evaluate the role that messages play on the probability that taxpayers will cancel their debt with the Tax Administration by subscribing to a payment plan in the context of a tax amnesty. The interventions are the following. First, a group of recent debtors -debt less than five years old- will be divided into 3 groups. The control group will receive the old notification. Treatment one will receive a new notification that is more colorful and presents easy to follow descriptions of the payment plans. Treatment 2 will receive the new notification but including the computation of interests saved under each one of the payment plans also. The second intervention has been designed for those who have older debt that is close to prescribing. These taxpayers will be divided into two groups, one receiving the old notification and one receiving the new notification, including the analysis of savings. With these interventions we plan to evaluate the role of messages, salience, and information on debt payment, on the selection of a payment plan, and on compliance with current tax obligations (spillover). While there are several recent field experiments looking at the role of messages on tax payment, there is no evidence for this in the context of a tax amnesty. Moreover, even though it has long been considered that reducing transaction and informational costs should be a relevant policy instrument, there is little to no evidence about the role that reducing the costs of understanding the specifics of a payment plan could have on tax payment and compliance behavior.
"Public Goods, Information, Trust and Tax Compliance (Pre-analysis plan)" (w/B. Cardinale Lagomarsino) Pre-analysis plan
Abstract: Tax morale has been shown to affect individuals' willingness to pay taxes. Among others, people pay taxes because doing so is part of a voluntary exchange with the government (reciprocity). There is evidence that people comply more when they see the government in action and public monies being used for the good of the community. There is also partial evidence that information about what the government does with the money matters too. This result has been more elusive as some papers have shown evidence but others have not. One possibility for the high variance in results could be that interventions have differed in the intensity of the treatments (i.e., some researchers have worked in communities where they could show higher efficiency/efficacy of the government than others and that has driven the results). Other possibility is that average results mask high heterogeneity across individuals based on their priors about the efficiency and efficacy of the government. These priors may be affected by their previous experience with the government. In this project we aim to disentangle among these hypotheses by evaluating the role of messages in the context of a large infrastructure campaign (i.e., high intensity treatment). We aim to evaluate the marginal effect of informing taxpayers about the use of public monies and check for heterogeneity according to the services each taxpayer receives (e.g., pave roads or dirt roads), and according to changes to their stock of public services change (e.g., those who have changed from dirt to pavement recently). In terms of policy relevance, we can evaluate the marginal effect of information on top of the effect that public works should have had by themselves (people see the works and change their payment behavior). In this project, we also evaluate the effect of promises about future public works in the context of a local government with relatively low levels of trust but recently engaged in the expansion of public works. This is important in terms of policy implications: if people believe in promises, then governments could use them to finance future works in advance.
"Government Behavioral-Contingent Responses and Tax Compliance. Evidence from a Field Experiment"
Abstract: Tax evasion is a common feature in developing countries, due in part to low enforcement capacity and to a lower effect from moral determinate such as reciprocity with the State. Previous research has concentrated on one-shot interactions between the tax authority and taxpayers. In that setting, deterrence and moral messages affected tax compliance, and results tended to fade over time. One reason why these effects fade over time is that no changes occur in the long-term relationship between those who pay and those who collect taxes because of this sporadic interaction. In this project, a large scale field experiment in a Municipality in Argentina, we introduced a design in which the government's response changes over time according to the taxpayers behavior. As such, the taxpayer has a continuous feedback from the results of her actions. Those who pay are acknowledged and those who don't are reminded of their obligation. Consequently, our project tests the impact of an innovative experimental design that alternates the provision of positive and deterrence messages contingent on taxpayers’ compliance behavior.
Field Experiment with Aguas de Quito (water service provider) in Quito, Ecuador: Transparency, Payments, and Water Consumption (w/P. Carrillo)
Field Experiment in Mexico City: Pollution, Information, and the Demand of Public Policies (w/B. Hoffman)
Survey Experiment on the Politics of Crime (w/D. Gingerich)
Age Eligibility, Elections, and Information Gathering (w/R. Vlaicu)