Competition and Coordination in Infrastructure: Port Authorities' Response to the Panama Canal Expansion (JMP)
Who should build infrastructure and how much should they build? In the United States, transportation infrastructure decisions are mostly made at the state or local level. I study seaports during a period when massive investments were being made on the East Coast to prepare for the larger vessels that could navigate an expanded Panama Canal. Competition between the ports spurs more investment and lower prices, but if ports do not internalize their business-stealing effect on others, there may also be overinvestment--especially with subsidization. With data on all container imports and capital costs of the major US ports, I estimate a model of the investment game that port authorities play with one another. I find the spatial distribution of investment on the East Coast would be very different if ports coordinated their activity.
Works in Progress
Skill Premium, Trade, and Human Capital (with Camilo Alvarez and Fernando Arce)
Our paper seeks to explain why the college wage premium first rises and then falls after emerging countries liberalize trade. We study a two-sector trade model with heterogeneous firms and human capital accumulation. From the firms’ side, our model has two sectors: the first, a perfectly competitive sector, the second, monopolistically competitive sector selling differentiated goods. The monopolistically competitive sector is more skill intensive. In the model, households are grouped in a family structure where they are endowed with two types of labor, low-skilled and high-skilled.Households can choose to invest in human capital, in which case they pay an endogenous tuition cost and forgo wages earned from working in a low-skilled occupation for that period. When a country liberalizes trade the homogeneous sector is perfectly traded and therefore maintains its autarky size. At the same time, trade expands the market for the monopolistically competitive sector, which has transitory effects on the skill premium. Using Mexican data starting in the late 1980s, we show that the model can quantitatively account for the evolution of the skill premium.