Job Market Paper
Abstract: We construct a novel longitudinal dataset from administrative records to examine the impact of selective college education on asset accumulation, consumer credit usage, as well as short and long-term earnings. Our empirical strategy is a fuzzy regression discontinuity design that employs the admission policies of a selective public university in Colombia, relying solely on the national high school exit examination scores. Scoring above the admission threshold has no short-term effect but raises access to consumer credit by 4 percent and earnings by 24 percent eight years after college entry. While the gains in consumer credit stabilize after 11 years after college entrance, earnings returns keep growing up to 32 percent 16 years after college entry. The impacts on asset acquisition take longer to emerge as admission raises the likelihood of homeownership by 12 percent when individuals are 30 to 35 years old. Students admitted to the selective university reported more days with formal jobs per year, hinting that better earnings and employment outcomes likely contribute to gains in credit market access. The results on financial indicators shed light on the college education impact on dimensions such as durable purchases and financial inclusion describing economic wellbeing in the long term.
Food assistance and mobility during COVID-19 lockdown 2021. with Juan Mogollon and Catalina Villamil. (Click here to download a copy)
UNDP LAC WORKING PAPER SERIES N. 27.
Abstract: This paper examines the effects of emergency food assistance on human mobility patterns between March and August of 20 20. We study a large public-private initiative in Colombia, created to deliver food aid to one million households at risk of falling into poverty and previously not included in other assistance programs. The impact is estimated using the quasi-exogenous roll out of the food distribution within municipalities. The high-frequency data set links detailed daily deliveries with georeferenced food recipients’ location and mobility indicators measuring out-of-home events. This information is then aggregated at the urban tract level. The specification includes a rich set of fixed effects using sector, time and department-week trends. The main findings indicate that receiving food assistance delivered to the home reduced outof-home mobility by 1.6 percent during a two-week window. The estimated coefficient after that window is imprecise, suggesting a short-term effect. The program reduced out-of-home activities by reducing visits to grocery stores. Delivering food assistance proved to be an effective intervention to reduce population movement, thus helping households in distress comply with national lockdown orders.