Research

We estimate the causal impact of communication infrastructure quality on growth and structural transformation. We use variation across US counties' Internet speeds in 2018 and build an instrument using ARPANET, a military network that preceded the modern Internet, with its location documented in historical government reports. We find that doubling Internet speeds increases the 4-year employment growth by 3.3 to 6.1 percentage points. Faster Internet shifts economic activity towards high-skilled services and away from non-tradeable services while increasing inequality. Industry linkages, capital-skill complementarity, and ICT workers' sorting rationalize our results. Medium and small cities and rural areas drive our results.

Immigrant entrepreneurs are a major driver of economic growth, and their decisions about where to locate can greatly affect the entrepreneurial ecosystem of a country. Meanwhile, increasingly uncertain immigration environments might discourage immigrants from establishing new ventures in host countries. This paper exploits the unexpected result of the Brexit referendum to investigate the relationship between immigration uncertainty and the entry of immigrant-founded ventures of different quality. We propose a model of immigrant entrepreneurial entry and introduce a new measure of venture quality at founding. We find that a surge in uncertainty decreases the growth rate of new immigrant-founded firms by 3.2%. The reliance on other immigrant actors exacerbates the negative effect that uncertainty has on entry. Moreover, low- and high-quality firms are the most affected, while the effect for medium-quality firms is negligible. Back-of-the-envelope calculations suggest that Brexit discouraged the entry of around 620 low-quality and 250 high-quality firms. Our model suggests that while founders of low-quality ventures might decide to take up employment, founders of high-quality ventures might be better off establishing their companies in another country.

We study the location and occupational composition of establishments within firms between 1981 and 2016. Using Danish administrative employer-employee matched data, we document four interesting results regarding the internal spatial organization of firms. First, we show that the average number of establishments per firm increased by 21%. Second, the average distance of establishments and workers to their headquarters increased by around 100%. Third, these changes are mainly driven by the decentralization of production and business service workers and higher use of the latter. Fourth, we show that the ratio of managers to production and clerical workers within firms increased by 80%, driven particularly by headquarters and establishments located in the largest cities. These facts imply that firms are not simply becoming more spatially dispersed; instead, some activities, such as managerial, are increasingly being concentrated near firms' central offices.

This paper studies the spatial organization of firms, both theoretically and empirically. Two new facts in Danish register data motivate the analysis. First, firms have become more fragmented over time. Second, headquarters (HQ) establishments have become more manager-intensive, despite a significant increase in managerial wages at HQ locations. We study the roles of spatial wage disparities and communication costs in explaining these two trends. Immigration shocks are the source of identifying variation for changes in relative labor supply. We estimate elasticities of substitution across establishments within firms of 3.5 for workers and 0 for managers. Estimates indicate that increases in the wage of managers at the HQ relative to non-HQ locations, explain 66\% of the increase in HQ managerial intensity. This can be explained by the associated increasing demand for headquarter services and managerial knowledge as satellite establishments become larger. Simulations suggest that wider wage gaps across locations can also lead to more establishments per firm, and this effect strengthens as communication costs fall.

In this paper we exploit the arguably exogenous staggered implementation of an extensive criminal procedural reform in Colombia between 2005 and 2008 to assess its intended and unintended consequences. The reform had explicit objectives, such as guaranteeing due process protection of the accused, reducing the use of pretrial detention, making the processing of criminal cases more efficient, reducing procedural times, and improving the mechanisms for early termination of criminal processes. Our results show that the reform achieved most of its goals. Namely, a significant reduction in the use of pretrial detention of about 17%-34%; a large and significant reduction in procedural times (18%); an increase in the use of mechanisms for early termination of the criminal process through settlements (43%-66%); and a large and significant increase in the percentage of cases that reach adjudication. Nevertheless, the reform also had negative unintended consequences on arrest, clearance, and crime rates. Our results indicate that arrest rates decreased by about 33% and clearance rates by 16%-27%. The reform also directly affected the incentives for criminal behavior and led to an increase in both property crimes (19%) and violent crimes (17%) as a result of the implementation of the reform. Our paper shows that well-intended reforms aimed at increasing due process protection can create unintended consequences in the administration of justice that led to increases in crime and raises the question of how to balance constitutional protections with public safety by creating special provisions and guidelines directed to mitigate potential adverse effects on crime rates.

This paper analyzes the incidence of free trade agreements (FTAs) on the use of certificates of origin (CO) of exports in Colombia. Using firm-product-destination level data, together with Difference-in-Differences models with a staggered treatment, we find that the use of CO slightly decreases after the signing of an FTA; that is, the use of preferential tariffs does not increase after the entry into force of the agreement. Furthermore, we find evidence of an initial illusion effect: a positive impact in the initial two years, followed by a subsequent decrease. We conclude that eliminating non-tariff measures is key to achieving better use of trade agreements and unlocking their economic potential. 

I study the welfare consequences of land use regulations for low- and high-skilled workers within a city. I use detailed geographic data for Cook County and Chicago in 2015-2016, together with a spatial quantitative model with two types of workers and real estate developers who face regulations. For identification, I use the 1923 Zoning Ordinance, which was the first comprehensive ordinance in Chicago. Results suggest that more mixed-use zoning and looser floor-to-area limits lead to cheaper housing and higher wages. Counterfactuals suggest significant welfare gains of relaxing regulations in areas where it is more stringent, relative to areas where restrictions are less tight, especially for high-skilled residents due to gains in residential market access.

We study the relationship between microbusinesses’ access to formal credit and their transit toward formalization in Colombia. We use the country’s Microbusiness Survey—a novel database including both formal and informal microbusinesses—together with two definitions of formality: a binary and a multidimensional definition that exploits the array of rules faced by businesses. We document that those microbusinesses that request formal credit have a 12.5% higher probability of not being completely informal. Using the ownership of collateralizable assets as instruments, we confirm that the relationship between both variables is more than just a simple correlation: formal credit can help a microbusiness escape total informality. This positive effect of formal credit is larger for the first stages in the firms’ path toward formalization. On the contrary, informal credit can end up reinforcing firm informality.