NSF Proposal 0721068 Eric Verhoogen, Columbia University
Exporting plays a central role in the development strategies of many countries. A crucial issue in policy debates around globalization is how broadly the gains from integration, and exporting in particular, are being shared. It is well established that exporting plants tend to pay higher wages than non-exporting plants within the same industry, in developing as well as developed countries. But it is not clear that this reflects a causal relationship between exporting and wage premia that is, wages above what workers would be paid elsewhere on the labor market. The task of estimating the effect of exporting on wage premia is complicated by two forms of selection bias. On one hand, plants that enter the export market tend to be strong performers on many dimensions, and it is not clear that exporting per se is responsible for high wages in such plants. On the other hand, the workers hired by exporting plants may have valuable skills that, while unobservable to researchers, would be equally rewarded elsewhere on the labor market. Thus while exporting may raise average wages within a plant, this rise may simply reflect changes in the skill composition of the workforce, rather than wage premia.
This project develops a rich new micro-dataset on Mexican manufacturing plants and workers in order to estimate the effect of exporting on wage premia separate from the confounding effects of plants' selection into the export market and workers' selection into exporting plants. The project links a longitudinal plant-level dataset from the Mexican statistical agency, a dataset that contains information on exports and other plant characteristics over time, with an employer-employee dataset from the Mexican social security agency which follows individual workers as they move from plant to plant. The project uses the peso crisis of late 1994 as a natural experiment that was generating changes in the propensity of Mexican manufacturing plants to export as a result of the steep real depreciation of the peso. It develops a new econometric methodology that takes into account changes in the economy-wide value of skill in estimating the effect of exporting on wage premia. The project is one of the first to use a natural-experiment methodology to examine the determinants of wage premia at the plant level.
The project makes broader contributions on a number of levels. It informs public debates about how the benefits to international integration are being shared. It also generates an important public good , a combined employer-employee and longitudinal plant dataset for Mexico , that is expected to be of broad service to research communities both in the United States and Mexico. Research collaborations between the United States and Mexico are likely to benefit.
