Our study uses data from a carefully-designed urban poverty survey undertaken in late 2009 by the General Statistics Office that first created a purpose-build sampling frame, and then interviewed 3,349 randomly-selected households covering 8,208 persons, half of whom were migrants. We estimate Mincerian-style earnings equations, after correcting for selection bias, and use the estimates to decompose the earnings differential following the method proposed by Blinder and Oaxaca. Using this method, at least 60%, and possibly all, of the gap can be explained by differences in endowments (such as education and experience).
A more elaborate decomposition, following the method proposed by Brown, Moon, and Zoloth, first seeks to explain how workers sort into different sectors (state, private, individual), or occupations (blue or white collar, in industry or services), and then examines wage differentials within each sector or occupation. Using this approach, about half of the wage gap may be explained by endowments, suggesting that migrants face headwinds when they try to find jobs in favored sectors, and when they seek comparable wages to those received by residents. The system of residential permits (ho khau) may contribute to the difficulties faced by migrants, particularly when seeking jobs in the public sector.
Our results are broadly similar to, although more stable and plausible than, those found for the major cities in China, which also operates a (stricter) system of residential permits. We attribute the solidity of our results in large part to the high quality of the data that we were able to use. An interesting question, which the data did not allow us to address, is who can shift from migrant to resident status, and under what circumstances.