Working Papers

Abstract

This paper studies how the housing market in Colombia responds to policies aimed at increasing homeownership among low-income households. Private sector developers received tax incentives to build houses priced below a cutoff, and households received subsidies to buy houses below the same cutoff. The paper models an equilibrium between heterogeneous developers building differentiated housing and heterogeneous households purchasing homes. The model is estimated using bunching around the price cutoff induced by the policies. To benefit from the policy, households buy housing units up to 30 percent smaller and bunch at the policy cutoff. A counterfactual simulation shows that without supply-side incentives, developers may exit the market; their profits would be 14 percent lower. 

Abstract

In this paper, we examine infrastructure cost trends through the lens of the interstate highway system, using high-quality measures of the physical extent and quality of the interstate.

Contrary to popular belief, we find that the quality of interstate has improved considerably over the study period (1990-2008) and only limited evidence of excessive cost growth. Maintenance costs have remained stable and, while resurfacing costs have doubled, this price increase is entirely explained by inputs. Together, maintenance and resurfacing account for 2/3 of expenditures; new lane-miles however have exhibited unexplained excess cost growth. A calibrated model of highway investment suggests that the gas tax, despite being fixed since 1993, may cover the relevant user cost of capital for the US interstate.

Abstract

 This paper estimates the effect of a location-based price subsidy of utilities and property taxes on the housing market in Bogotá, Colombia. In Bogotá, neighborhood blocks are divided into 6 subsidy codes using and arbitrary cutoffs on a block quality score. I use the discontinuity introduced by the cutoffs in a Regression Discontinuity Design to study the effect of the subsidy on the housing market. I find that blocks receiving a higher subsidy have newer houses - implying more construction in such areas. I also find  a capitalization of the subsidy into housing prices: properties in areas receiving a small subsidy are cheaper than those receiving a bigger subsidy. These results suggest that a careful evaluation of location-based subsidies' incidence needs to consider potential capitalization into the housing market and other unintended effects such as new construction or renovations.

Abstract

We study the impact of a large public-policy push expanding internet access on secondary school test scores in Colombia. We use an instrumental variable approach exploiting the costliness of extending the existing internet infrastructure to connect new areas to identify causal impacts. We find that this internet expansion did have an effect on math test scores, which was concentrated in the bottom third of the test score distribution. Our estimates suggest that every 10% increase in the number of test-takers with access to the internet resulted in a 0.06 SD increase in math scores for this group. We find no significant effects on language test scores. We also show that this expansion did not result in increased rates of test-taking students participating in the labor force or in increased family incomes during this period, ruling out that these results may be driven by either substitution effects in the students time use or income effects resulting from increased family incomes. 

 Abstract

A rapid expansion in demand for post-secondary education triggered an unprecedented boom in higher education programs in Colombia, raising concerns about their relevance and quality. This paper shows that the penalty on student learning and labor market outcomes of attending a recently created program is large but, to a large extent, it is driven by student and program selection. Using rich administrative data that match higher education school admission information, socioeconomic characteristics of the young graduates, standardized test scores pre- and post-higher education, and formal labor market outcomes, we characterize this selection process by disentangling the relative roles of demand and supply forces. The main factor behind the learning penalty is student selection in baseline ability. In the case of labor market outcomes, the penalty is due to a combination of student and program characteristics.

Abstract

The literature that studies the determinants of inequality in Colombia finds that education is a key explanatory factor. However, studies in other countries shows that increases in education are not necessarily associated with reductions in inequality. This paper simulate different scenarios of increases in the educational level of the Colombian population and evaluate their effect on inequality. In particular it explores changes focused in a particular level of education (primary, secondary, tertiary), changes in more than one level, and changes focused in the regions with lower educational levels. The results suggest that inequality can be lower if the changes in the education are applied simultaneously in primary, secondary and higher education and focused on departments with low levels of coverage. Otherwise, the effects would be modest or even perverse: increasing the coverage of higher education in general, without focusing in departments with low coverage levels, may increase inequality.