Research

Implications of Zoning for Urban Gentrification in New York City

This paper estimates the effect of the sweeping rezoning program initiated by the Bloomberg administration in New York City from 2002 to 2013. Using an instrumental variable strategy incorporating regime change from the 1961 Zoning Ordinance, this paper finds a 10% increase in the floor area ratio (a physical change), defined as the building floor area to the lot size, decreases price per square foot by 1.7% and a 10% increase in the maximum floor area ratio (a policy change) results in a 2.4% decrease in price. The paper uses a spatial general equilibrium model to quantify the aggregate impact of the rezoning policy on low- and high-skilled workers. Counterfactual analysis finds that the price per square foot for residential floorspace drops as a result and aggregate welfare is increased for both groups, but residential segregation increases.

The Impact of Trade on Development: Evidence from Pastoralist Practices on the Ancient Silk Road joint with Mos Laoprassorn (Draft)

This paper studies the long-term effect of trade on development. We approach this question in the context of the ancient Silk Road, examining whether the locations along the highland Silk Road continue to be relatively more developed than other locations in the highland region along the Inner Asia Mountain Corridor that were not on the ancient Silk Road. We proxy for modern development using high-resolution satellite imagery. To provide a causal effect between proximity to the Silk Road and modern development, we adopt a novel instrumental variable, using a simulated seasonal mobility pattern of the nomadic pastoralists from Frachetti et al. (2017) as an instrument for the location of the Silk Road sites. We find a significant and robust positive relationship between proximity to Silk Road sites and modern development measures; an increase in the distance to the Silk Road by one standard deviation decreases the night lights intensity by 10.0%. Based on the elasticity of night lights with respect to GDP in the literature, this corresponds to a decrease in GDP of about 4.1%-9.7%.

Heterogeneous Effects of Trade Agreement Topics Over Time joint with Dyanne Vaught

Using unsupervised natural language processing techniques such as Latent Dirichlet Allocation and Structural Topic Model, we study the evolution of topics in trade agreements over time. In a standard gravity regression framework, a 10 percentage point increase in the Dispute/Arbitration Topic reduces trade flows by 3.5% and 10 percentage point increase in the Community/Membership Topic increases trade flows by 1.1%.