Governments and organizations around the globe are seeking to expand children's access to computers and the internet as the United Nations calls for efforts to eliminate the digital divide. However, little is known about the effects this expansion may have on long-run human capital accumulation. This paper provides the first causal estimates of the effect of children's access to computers and the internet on adult educational outcomes such as schooling and choice of major. To establish a causal link, I exploit cross-cohort variation in access to computers and the internet among primary and middle school students in Uruguay, the first country to implement a nationwide one-laptop-per-child program. Despite a notable increase in computer access, educational attainment has not increased. However, college students who had been exposed to the program as children, were more likely to select majors with good employment prospects, and less likely to enroll in multiple majors at the same time.



The Political Coase Theorem (PCT) states that, in the absence of transaction costs, agents should agree to implement efficient policies regardless of the distribution of bargaining power among them. This paper uses a laboratory experiment to explore how commitment problems undermine the validity of the PCT. Overall, the results support theoretical predictions. In particular, commitment issues matter, and the existence of more commitment possibilities leads to better social outcomes. Moreover, we find that the link is valid when commitment possibilities are asymmetrically distributed between players and even when a redistribution of political power is required to take advantage of those possibilities. However, we also find that at low levels of commitment there is more cooperation than strictly predicted by our parameterized model while the opposite is true at high levels of commitment, and only large improvements in commitment opportunities have a significant effect on the social surplus, while small changes do not.

This paper was motivated by the structural changes that affected the Argentine financial system as a result of the 2001 economic crisis. The crisis lead some banks to close down and others to exit the Argentine market, while spurring changes in regulation. In this paper, I test whether the shock to the system affected the behavior of banks, and if so, whether this behavior became more compatible with market efficiency or market power: did the decrease in the total number of banks lead to increased productivity in the sector (by eliminating the least productive firms) or, on the contrary, lead banks to increase markups? To answer this question I use a structural approach and examine the Market Power Hypothesis (using Structure-conduct-Performance and RMP) and Efficiency Structure Hypothesis (using X-efficiency and scale efficiency) for all banking entities in the period 1994-2010. I use the Data Envelopment Analysis (DEA) technique to obtain reliable estimates of efficiency. Despite the large shocks to the system and the increase in concentration, I do not find significant differences before and after 2001. However, there are slight glimpses structural differences among certain groupings of banks after the crisis. In particular, I find evidence in support of the efficiency hypothesis for retail banks after 2001. This is the first study to develop a detailed evaluation of these hypotheses specifically for Argentina with a complete sample of banks. This research contributes to a more complete understanding of the market, illuminating discussions on banking policy and regulation.



Major natural disasters such as Hurricanes Katrina and Sandy cause numerous fatalities, and destroy property and infrastructure. In any year, the U.S experiences dozens of smaller natural disasters as well. We construct a 90 year panel data set that includes the universe of natural disasters in the United States from 1920 to 2010. By exploiting spatial and temporal variation, we study how these shocks affected migration rates, home prices and local poverty rates. The most severe disasters increase out migration rates and lower housing prices, especially in areas at particular risk of disaster activity, but milder disasters have little effect.

Protecting the environment is often plagued by collective action problems, and so it is important to understand what motivates politicians to act. This paper dwells on whether public information can influence demand in the population, and, if so, what are the relevant channels. I exploit the publishing in 1962 of the influential environmental science book Silent Spring - and the availability of U.S. congressional roll-call votes and census data, - to analyze how demand for environmental regulation changes in response to a radical informational shock. I define demand in terms of the total number of ‘green’ votes in Congress. My analysis has two steps. First, I evaluate the impact of my shock on average propensity for politicians to vote in favor of ‘green’ regulation, and find effects between 5 and 33 percentage points. Then, I look for heterogeneous effects of the shock, by including interactions with education, income, and exposure, and propose a framework to interpret my findings. My results suggest that public information, education and income interact in the demand for environmental regulation.

This paper provides causal estimates of the effect of secondary school’s access to computers and the internet on academic achievement. I exploit inter-temporal and cross-sectional variation in access to technology among schools in Argentina, the first country to implement a nationwide one-laptop-per-child program in secondary schools, and the largest program of this kind. Contrary to widespread evidence that one-laptop-per-child programs have had no effects on educational achievement for primary and middle-school students, I find improvements in promotion and graduation rates in public secondary schools around the time of the intervention. I also investigate the widely-cited hypothesis that schools that incorporate technology to their classrooms benefit the most from one-laptop-per-child programs, and find no evidence to support this claim.



The Conversation: New data-set explores 90 years of natural disasters in the US (2017)

VOX-LACEA: The Political Coase Theorem (2017)