## Publications

• Housing Discrimination and the Toxics Exposure Gap in the United States: Evidence from the Rental Market, The Review of Economics and Statistics (2022) 104 (4): 807–818. [REStat], [NBER WP No. 26805], Replication Files.

(with Peter Christensen and Christopher Timmins). Featured in Medium, Phys.org National Low Income Housing Coalition.

Abstract: This study uses a correspondence experiment in the online housing market to test whether racial discrimination constrains the housing choices of minority renters in markets that contain major sources of airborne chemical toxics. Within-property tests indicate that renters with African American or Hispanic/LatinX names are 41% less likely than renters with white names to gain access to properties in low exposure locations, though there is no evidence of constraints at high exposure locations in the same zip code. The same pattern appears in tests for constraints among highly comparable properties. Tests using data on the recent location decisions of renter households reveal evidence of a 16% race gap in the likelihood of residential exposures in the same markets.

• Can we stay one step ahead of cheaters? A field experiment in proctoring online open book exams, Journal of Behavioral and Experimental Economics, Journal of Behavioral and Experimental Economics 90, February 2021, 101653. [JBEE] (with Jose Vazquez and Eric Chiang)

Abstract: As more institutions of higher learning expand their offerings of online courses, the use of online assessments has become an important topic of discussion. Although the benefits of using online assessments can be substantial in terms of time savings and reduced material costs, the potential of greater instances of cheating, in the absence of a proctor, poses a substantial cost in protecting academic integrity. This has led to the development of many proctoring solutions to address this challenge. This paper uses a field experiment to analyze the effects of two proctoring methods on exam scores when students are informed in advance whether the exam is proctored or not. In both situations, best practices were used to minimize cheating. Our results show that students whose exams were not proctored scored over 11% higher on average than those whose exams were proctored, all else equal. But this effect is lower when web-based proctors were used instead of live in-person proctors. Although these results vary based on classroom format and student characteristics, our general finding is that when exams are not proctored, students are more successful at achieving substantially higher scores, presumably by engaging in collaborative activities that would be considered cheating.

• Sporting Events, Emotional Cues, and Crime: Spatial and Temporal Evidence from Brazilian Soccer Games, Economic Inquiry, Vol. 59, No. 1, January 2021, 375–395. [Economic Inquiry], [SSRN] (with Qi Ge and Rodrigo Schneider)

Abstract: This paper studies the effects of emotional cues and rational incentives on crime. Using police reports that provide granular information on vehicle robberies and thefts in the city of São Paulo, Brazil, we explore the impact of game outcomes of local soccer teams on crime and examine its spatial heterogeneity. Our estimates suggest that crime increase after home games in streets that are within a tight radius from the stadiums. The effect is driven by vehicle thefts with popular car models being more likely to be targeted and is particularly salient after upset losses.

• Unlocking Amenities: Complementarities in Public Good Consumption, Journal of Public Economics, Volume 182, February 2020, 104110. [pdf] [NBER WP No. 25107][JPubE] (with David Albouy and Peter Christensen). Featured in the New York Times, Psychology & Mental Health News, 7thSpace News.

Abstract: Public goods may exhibit complementarities that are essential for determining their individ- ual value. Our results indicate that improving safety near parks can turn them from public bads to goods. Ignoring complementarities may lead to i) undervaluing the potential value of public goods; ii) overestimating heterogeneity in preferences; and iii) understating the value of public goods to minority households. Recent reductions in crime have “unlocked” almost $7 billion in property value in Chicago, New York and Philadelphia. Still two-fifths of the potential value of park proximity,$10 billion, remains locked-in.

• Do More Eyes on the Street Reduce Crime? Evidence from Chicago’s Safe Passage Program, Journal of Urban Economics 110 (2019) 1–25. [pdf][JUE] (with Ruchi Singh and Dan McMillen). Featured in NPR, WBEZ, The 74, Policy Review

Abstract: Chicago’s Safe Passage program attempts to ensure the safety of student traveling to and from schools by placing civilian guards along specified routes. The program was launched during the 2009-2010 school year and was expanded to 140 schools by 2015-16. We use data from more than 10 years of geocoded Chicago police reports and school level data to analyze the Safe Passage program’s effects on crime rates and the rate of absenteeism from schools. Our findings suggest that the program is an efficient and cost effective alternative way of policing with direct effects on crime and student’s outcomes. Exploiting both spatial and temporal variation in the implementation of the program, we find that the presence of guards results in lower levels of crime, with violent crime declining by 14% on average. The rate of absenteeism is estimated to decline by 2.5 percentage points. We find no evidence of spillovers of crime to areas that are not along the Safe Passage routes.

• Prediction of SARS-CoV-2 infection with a Symptoms-Based model to aid public health decision making in Latin America and other low and middle income settings. Preventive Medicine Reports 27 (2022) 101798. [PMR] (with Andrea Ramírez Varela, Sergio Moreno López a, Sandra Contreras-Arrieta, Guillermo Tamayo-Cabeza, Silvia Restrepo-Restrepo, Ignacio Sarmiento-Barbieri, Yuldor Caballero-Díaz, Luis Jorge Hernandez-Florez, John Mario González, Leonardo Salas-Zapata, Rachid Laajaj, Giancarlo Buitrago-Gutierrez, Fernando de la Hoz-Restrepo, Martha Vives Florez, Elkin Osorio, Diana Sofía Ríos-Oliveros, Eduardo Behrentz)

Abstract: Symptoms-based models for predicting SARS-CoV-2 infection may improve clinical decision-making and be an alternative to resource allocation in under-resourced settings. In this study we aimed to test a model based on symptoms to predict a positive test result for SARS-CoV-2 infection during the COVID-19 pandemic using logistic regression and a machine-learning approach, in Bogota ́, Colombia. Participants from the CoVIDA project were included. A logistic regression using the model was chosen based on biological plausibility and the Akaike In- formation criterion. Also, we performed an analysis using machine learning with random forest, support vector machine, and extreme gradient boosting. The study included 58,577 participants with a positivity rate of 5.7%. The logistic regression showed that anosmia (aOR = 7.76, 95% CI [6.19, 9.73]), fever (aOR = 4.29, 95% CI [3.07, 6.02]), headache (aOR = 3.29, 95% CI [1.78, 6.07]), dry cough (aOR = 2.96, 95% CI [2.44, 3.58]), and fatigue (aOR = 1.93, 95% CI [1.57, 2.93]) were independently associated with SARS-CoV-2 infection. Our final model had an area under the curve of 0.73. The symptoms-based model correctly identified over 85% of par- ticipants. This model can be used to prioritize resource allocation related to COVID-19 diagnosis, to decide on early isolation, and contact-tracing strategies in individuals with a high probability of infection before receiving a confirmatory test result. This strategy has public health and clinical decision-making significance in low- and middle-income settings like Latin America.

• COVID-19 spread, detection, and dynamics in Bogotá, Colombia, Nature Communications volume 12,4726, 2021.[Nature], [CEDE WP], [RedNIE WP], [Nota de Política] Replication Files (with Rachid Laajaj, Camilo de los Rios, Danilo Aristizabal, Eduardo Behrentz, Raquel Bernal, Giancarlo Buitrago, Zulma Cucunubá, Fernando de la Hoz, Gabriela Delgado, Alejandro Gaviria, Luis Jorge Hernández, Leonardo León, Elkin Osorio, Andrea Ramírez Varela, Silvia Restrepo, Rodrigo Rodríguez, Martha Vives, Duncan Webb )

Abstract: Latin America has been severely affected by the COVID-19 pandemic but estimations of rates of infections are very limited and lack the level of detail required to guide policy decisions. We implemented a COVID-19 sentinel surveillance study with 59,770 RT-PCR tests on mostly asymptomatic individuals and combine this data with administrative records on all detected cases to capture the spread and dynamics of the COVID-19 pandemic in Bogotá from June 2020 to early March 2021. We describe various features of the pandemic that appear to be specific to a middle income country. We find that, by March 2021, slightly more than half of the population in Bogotá has been infected, despite only a small fraction of this population being detected. The initial buildup of immunity contributed to the containment of the pandemic in the first and second waves. We also show that the share of the population infected by March 2021 varies widely by occupation, socio-economic stratum, and location. This, in turn, has affected the dynamics of the spread with different groups being infected in the two waves.

• Factors Associated with SARS-CoV-2 Infection in Bogotá, Colombia: Results from a Large Epidemiological Surveillance Study. The Lancet Regional Health - Americas Volume 2, October 2021, 100048. [The Lancet][SSRN] (with Andrea Ramirez Varela, Luis Jorge Hernandez Florez, Guillermo Tamayo-Cabeza, Sandra Contreras-Arrieta, Silvia Restrepo Restrepo, Rachid Laajaj, Giancarlo Buitrago Gutierrez, Yenny Paola Rueda Guevara, Yuldor Caballero-Díaz, Martha Vives Florez, Elkin Osorio, Daniela Rodriguez Sanchez, Leonardo Leon Nuñez, Raquel Bernal, Sofía Rios Oliveros, Leonardo Salas Zapata, Marcela Guevara-Suarez, Alejandro Gaviria Uribe, Eduardo Behrentz)

Abstract: Background: Epidemiologic surveillance of COVID-19 is essential to collect and analyze data to improve public health decision making during the COVID-19 pandemic. There are few initiatives led by public–private alliances in Colombia and Latin America. The CoVIDA study led by Universidad de los Andes contributed with RT-PCR tests for SARS-CoV-2 in population groups with mild or asymptomatic infections in Bogotá. The present study aimed to determine the factors associated with SARS-CoV-2 infection in working adults. Methods: COVID-19 sentinel epidemiological surveillance study, from April 18, 2020 to March 29, 2021. The study included people aged 18 years or older without a history of COVID-19. Priority for inclusion was given to two main occupational groups working during the pandemic: health care workers and essential services workers with high mobility in the city. Social, demographic, and health-related factors were collected via phone survey. Afterward, the molecular test was conducted to detect SARS-CoV-2 infection. Findings: From 58,638 participants included in the study, 3,310 (5·6%) had a positive result for SARS-CoV-2 infection. A positive result was associated with the age group (18-29 years), living with more than three cohabitants, living with a COVID-19 confirmed case, having no affiliation to the health system, reporting a very low socioeconomic status, and having essential occupations.

## Working Papers

• Can’t Stop the One-Armed Bandits: The Effects of Access to Gambling on Crime (pdf) (with Nicolas Bottan and Andrés Ham)

Abstract: We estimate the effects on criminal activity due to the approval of state legislation in Illinois that legalized video gambling. The bill gave municipalities discretion over whether to allow video gambling within their local boundaries. Many jurisdictions adjacent to Chicago opted in, while the City of Chicago opted out. These decisions create a natural experiment that allows studying the effects of increasing access to gambling on crime. Using detailed incident-level crime data and a difference-in-differences strategy, we find that (i) access to gambling increases violent and property crimes; (ii) these crimes represent “new” rather than displaced incidents; and (iii) effects are persistent over time. We further study downstream effects on property values, finding that properties adjacent to gambling establishments sell on average at a three percent discount.

• Racial Discrimination in the United States Rental Market [NBER WP No. 29516] (with Peter Christensen and Christopher Timmins)

Abstract: We report evidence on discriminatory behavior from the largest correspondence study conducted to date in the rental housing market. Using more than 25,000 interactions with rental property managers across the 50 largest U.S. cities, the study reveals that African American and Hispanic/LatinX renters continue to face discriminatory constraints in the majority of U.S. cities although there are important regional differences. Stronger discriminatory constraints on renters of color (particularly African Americans) are also associated with higher levels of residential segregation and larger gaps in intergenerational income mobility. Using matched evidence on the actual rental outcomes at the properties in our experiment, we show that correspondence study measurements of discrimination do indeed predict actual outcomes.

## Other Publications

• ECLAC South American Input-Output Matrix Associate Researcher for Argentina, Ecuador and Paraguay. Available here (In Spanish). Media