Robot Adoption, Organizational Capital and the Productivity Paradox, 2021, Job Market Paper (25th Razin Prize, best 2022 GU paper).
Major technological changes have come with an adjustment period of stagnant productivity before the economy operates at its full potential. The mechanism of this adoption process is still not well understood. Using event studies, I document that productivity increases with a five-year lag after the adoption of industrial robots in Brazil. I build a dynamic general equilibrium model with heterogeneous firms, endogenous robot adoption, and organizational capital accumulation. Imperfect transferability of organizational capital across technologies and labor reallocation of workers across tasks reduce productivity in the short run and the productivity increase occurs after some time. I combine employer-employee matched data with a novel measure of robot adoption and provide the first evidence of plant-level labor reorganization across occupations and organizational capital depreciation induced by the automation process. During five years after adoption, labor switching increases across occupations within firms, moving from production to support activities. I also show that firms’ organizational capital measured by workers’ firm-occupation specific experience depreciate and then slowly re-accumulate. When these processes stop, productivity gains appear. I use these results to estimate the model and analyze the transition path after the possibility of technology adoption opens. I use the estimated model to evaluate the dynamic effects of tax and subsidies targeted both on robots and organizational capital.
Conferences and Seminars: U Toronto, NBER SI, NASMES, SED, GWUSB, U Delaware, U Los Andes, U Queensland, ITAM Business, Banco de México, PUC Chile, Harvard, HBS, Boston U (TPRI), LACEA–LAMES, LACEA–RIDGE, World Bank, EEA, Georgetown (Macro, Trade), GU-McDonough (Strategy), AOM (TIM–Doctoral Consortium), Wharton Innovation Doctoral Symposium.
U.S. Robots and their Impacts in the Tropics: Evidence from Colombian Labor Markets, with A. Kugler, M. Kugler, and L. Ripani, NBER Working Paper N0. 28034 (R&R World Development, highlighted at The Future of Work - IDB Series).
Previous studies for developed countries show negative short-run impacts of automation on employment and earnings. In this paper, we instead examine whether automation by a key trading partner can hurt workers in a developing country. We specifically focus in Colombia's labor market, and how automation in the U.S. impacts Colombian workers by replacing exports from Colombia for cheaper robot-made U.S. products. We use employer-employee matched data from the Colombian social security records combined with data on U.S. exposure to robots in different sectors from 2011 to 2016 to examine if robots in the U.S. are displacing workers in Colombia. We find that U.S. robots decrease employment and earnings for Colombian workers in those sectors of local labor markets that have high levels of automation -measured as robots per thousand workers- in the U.S. labor market. In terms of turnover, as expected, there is an increase in dismissals and a decrease in hires for workers in sectors highly impacted by robots in the U.S. Moreover, the negative displacement effects of robots are greater for women; older workers; workers employed in small and medium sized enterprises, and workers employed in manufacturing. Importantly, local labor markets which exported the most to the U.S. in the past, are also the most affected by the increased adoption of U.S. robots, suggesting that Colombian workers may be losing employment to automated jobs reshored back to the U.S. Our estimates suggest that there may be sectors that benefit from automation due to productivity effects as the general equilibrium effects are nil at the local labor market level.
Cheating and Incentives: Learning from a Policy Experiment, with C. Martinelli, S. Parker, and A. Pérez-Gea. American Economic Journal: Economic Policy, 10: 298–325, February 2018 (highlighted by the AEA webpage).
We use a database generated by a policy intervention that incentivized learning as measured by standardized exams to investigate empirically the relationship between cheating by students and cash incentives to students and teachers. We adapt methods from the education measurement literature to calculate the extent of cheating and show that cheating is more prevalent under treatments that provide monetary incentives to students (versus no incentives or incentives only to teachers). We provide evidence suggesting that students may have learned to cheat, with the number of cheating students per classroom increasing over time under treatments that provide monetary incentives to students.
Closing the Achievement Gap on Mathematics: Evidence from a Remedial Program in Mexico City, with E. Gutierrez. Latin American Economic Review, 1–30, November 2014.
This paper evaluates the impact of an intervention targeted at marginalized low-performance students in public secondary schools in Mexico City, consisting in free additional math courses, taught by undergraduate students from the most prestigious Mexican universities. We exploit the information of all students’ (treated and not treated) transcripts enrolled in participating and non-participating schools. Before the implementation of the program, participating students lagged behind non-participating ones by more than a half base point in their GPA (over 10). Using a difference-in-differences approach, we find that students participating in the program observed a higher increase in their school grades, and that by the end of the school year, when the free extra courses had been offered for 10 weeks, participating students’ grades were not significantly lower than non-participating students’ grades. These results provide evidence that short and low-cost interventions can have important effects on student achievement.
WORK IN PROGRESS:
Sugar crops and the automotive industry in Brazil: the long-run consequences of industrial policy, with W. Kerr
Labor Market Power and Informality in Brazil, (RAIS & PME), with A. Pineda.
Cheating Networks, with C. Martinelli and S. Parker.