Research

Working Papers


  1. Procuring Firm Growth: The Effects of Government Purchases on Firm Dynamics, with Claudio Ferraz (PUC-Rio UBC) and Frederico Finan (UC Berkeley). [Submitted]


Firms in the developing world exhibit much flatter life-cycle dynamics compared to firms in developed countries. We examine the role of demand in limiting the growth of small and young firms in Brazil. We test whether firms that win government procurement contracts grow more compared to firms that compete for these contracts but do not win. We assemble a comprehensive data set combining matched employer-employee data for the universe of formal firms in Brazil with the universe of federal procurement contracts. Exploiting a quasi-experimental design, we find that winning at least one contract increases firm growth by 2.2 percentage points. These effects also persist well beyond the length of the contracts. We provide suggestive evidence that alleviating credit constraints is an important mechanism to explain these results.


  1. Agricultural Productivity and Deforestation: Evidence from Rural Electrification in Brazil, with Juliano Assunção (CPI and PUC-Rio), Molly Lipscomb (U. Virginia), and Mushfiq Mobarak (Yale) [Submitted]


We study the trade-off between economic development and environmental conservation in the context of agriculture. While increasing agricultural productivity is linked to economic development, it can have ambiguous effects on the environment. Improved productivity may provide farmers with incentives to expand the scope of farming and further encroach on forest lands, but it can also induce farmers to intensify their production and produce more output with less land. We examine these predictions using county-level data from five waves of the Brazilian Census of Agriculture combined with satellite data. We identify productivity shocks using the expansion of rural electrification in Brazil during 1960-2000. We show that electrification increased crop yields, and farmers subsequently shift away from land-intensive activities, sparring land. We find that without the increase in agriculture productivity brought about by the expansion of rural electrification in Brazil between 1970 and 2000, the rate of deforestation would have been almost twice as large as the actual rate. That effect is comparable to the most prominent and successful conservation packages ever implemented in the country. Our results suggest that synergies between growth and conservation exist and can create win-win situations.


  1. The Environmental Costs of Political Interference: Evidence from Hydropower Plants in the Amazon, with Juliano Assunção (CPI and PUC-Rio) and Francisco Costa (U. Delaware)


We estimate the impacts of ten recently built hydroelectric power plants in the Brazilian Amazon on forest loss. We find that the construction of the ten plants accounts for 21 percent of the observed forest loss in areas within 50 kilometers from the construction sites. However, all of this effect comes from only four plants. We interpret these results are evidence that existing regulations were effective in limiting environmental damages in six cases. We suggest that the dam's local economic impacts and political interference in the licensing processes explain the heterogeneous effects.



Selected Work in Progress


  1. Evaluation of Brazil's Agricultural Cisterns and Technical Assistance - P1+2 Program, with Marco Gonzalez-Navarro (UC Berkeley), Breno Sampaio (UFPE), and Vitor Pereira (ENAPE). [On the field] AEA RCT Registration ID: AEARCTR-0004109 .

Smallholders in developing countries are particularly exposed to climatic shocks. Furthermore, rural communities also display large gender imbalances. We evaluate a large-scale social program that helps poor smallholders adapt to climatic shock while trying to promote female empowerment. We cluster-randomize the delivery of large rainwater-fed cisterns for smallholder agriculture coupled with extension services and cash grants. Within the household, we also randomize the gender of the recipient.

  1. Evaluating Extension Services and Productive Cash Grants in Rural Brazil, with Marco Marco Gonzalez-Navarro (UC Berkeley), Breno Sampaio (UFPE), and Vitor Pereira (ENAPE). [On the field] AEA RCT Registration ID: AEARCTR-0003802

We evaluate a social program run by the Brazilian federal government that delivers agriculture extension services and cash grants to poor smallholders. The goal of the program is to increase farm production and improve wellbeing. We use a multi-site randomized controlled trial to assess its impacts and uncover the channels through which it alleviates rural poverty. The study takes place in three Brazilian states with very diverse socioeconomic and biophysical characteristics – Rio Grande do Sul, Minas Gerais, and Para.


  1. Heat and Care: Tale of a Tropical City, with Rudi Rocha (EAESP/FGV) and Vinicius Peçanha (UBC)

Temperatures display a U-shaped relationship with mortality, and that the link is primarily via human physiology. The bulk of this evidence comes from cross-city and epidemiological studies in developed countries. In this paper, we examine the heat-mortality relationship at a fine-grained level within Rio de Janeiro. We rely on novel satellite imagery sources on temperature and administrative health records at the individual level to build a neighborhood-by-month panel over 15 years. We find that hot days in a typical month in Rio account for 3.5% of all deaths due to non-external causes. Access to preventive health care can attenuate the marginal effect of temperature on cardiovascular deaths of the population 60+. We conclude that temperature shocks are localized within cities, implying that remedial policies should also be localized.

  1. Land Rights and Land Use, with Juliano Assunção (CPI and PUC-Rio) and Marco Gonzalez-Navarro (UC Berkeley).

One common feature of the tropics is ill-defined and weakly enforced property rights. A large literature investigates the economic consequences of these weak institutional settings. However, the environmental effects are less well understood. Twelve percent of the Brazilian Amazon territory (an area equivalent to France) is unclaimed land i.e., public land that is not oversight by any public agency and is often privately occupied. Deforestation in these areas is higher than in the rest of the Amazon. Using high-resolution spatial data, we study the effects of a large-scale land-titling program that secured property rights to private occupants. We use variation in the timing of titling conditional on enrollment in the program.


  1. Keep Cool and Carry On: The Effects of Air Conditioning in Rio de Janeiro’s Public Schools

A growing body of literature finds that high outdoor temperature has detrimental effects on learning. However, little is known about the effects of lowering classroom temperatures when outdoor temperatures are high. To fill this gap, I use rich data on test scores and socioeconomic data on pupils and explore the staggered adoption of classroom air-conditioning (AC) in Rio de Janeiro’s primary public schools.


  1. The Economics of Lobbying in European Climate Policy, with Ulrich Wagner (U. Mannheim)

Regulated firms exert considerable effort to lobby against regulation. If successful, lobbying can distort the intended welfare effects of regulation. When such attempts to influence regulation fail, firms may choose to relocate to unregulated jurisdictions. In the case of climate policy, the relocation of polluting firms can be very detrimental to welfare as it induces carbon leakage. The European Union’s Emissions Trading Scheme (EU ETS) seeks to prevent carbon leakage by distributing free emission allowances to regulated firms. Given the evidence of lobbying in this context, we estimate the welfare costs of lobbying by organized interest groups in the context of permit allocation in the EU ETS. To do so, we use novel data on lobbying specific to the free permit allocation n the EU ETS and use it to estimate a model of optimal permit allocation that includes lobbying efforts by regulated firms. Specifically, we estimate the objective function of a regulator who trades off the loss of permit revenue versus keeping jobs, carbon emissions, and political support within the EU. We use the fitted model to perform counterfactual simulations that tell us how much lobbying distorts the allocation of emission allowances.


  1. Who Watches When the Watchmen Won’t Watch?

A series of command-and-control policies introduced by the Federal Government in the middle 2000s successfully brought deforestation rates in the Brazilian Amazon to an all-time low. As of 2018, many of these policies were weakened. An initiative by the Brazilian Federal Prosecution Service, which is independent from the government’s Executive branch, started to criminally charge deforesters. The program only prosecutes deforesters responsible for areas above 60 hectares. Leveraging on this discontinuity, I combine property boundaries with high-resolution satellite data to assess the deterrence and incapacitation effects of this initiative.


  1. Public Procurement in São Paulo: Leveraging Big Data to Improve Performance, with Joana Naritomi (LSE), Michael Best (Columbia), and Augustin Chaintreau (Columbia/Computer Science)

Public procurement is one of the core functions of the state, providing key inputs into the delivery of critical public services. We apply machine-learning techniques to transaction-level data on the universe of business-to-business (B2B) and business-to-government (B2G) transactions to benchmark B2G transactions with B2B transactions, to answer a simple question: Do governments pay “too much” for the goods and services they purchase? If so, what are the determinants of this price inflation? Do governments select higher-cost firms? Or do the same firms practice different prices when selling to the government?


  1. Decentralization of Tax Collection, with Arthur Bragança (CPI), Diogo Britto (Bocconi), Alexandre Fonseca (RFB), Breno Sampaio (UFPE), and André Sant’Anna (BNDES)

Property taxes in developing countries can be hard to enforce and monitor, particullarly in rural areas. Oftentimes, tax collection relies on self-assessment In this paper, we use tax-payer level data combined with satellite data to characterize tax compliance and calculate the amount of tax evasion on a property tax levied on rural properties in Brazil. We then study the roll out of a policy that transferred to municipalities the collection and auditing of this tax. The rich taxpayer-level data allows to study taxpayers’ margins of adjustment.

  1. The Bahia Restoration Project, with Leah VanWey (U. Brown/Sociology), Stephen Porder (U. Brown/Ecology), Daniel Piotto (U. Federal do Sul da Bahia), Rui Rocha (U. Estadual de Santa Cruz), Jorge Chiapetti (U. Estadual de Santa Cruz).

This is an interdisciplinary collaboration with local NGOs with the goal of studying ecological and socioeconomic aspects of forest restoration in private properties. The project is set in the South portion of the Brazilian state of Bahia, an important preservation area of the Mata Atlântica (Atlantic Rainforest). With an infrastructure that collects longitudinal farm-level data for nearly 3,000 farms, this project is a platform for experimentation and observation.