UC Riverside

According to the U.S. Bureau of Labor Statistics, the share of foreign-born workers in the labor market increased from 14\% to 17\% between 1994 to 2008. At the same time, foreign-born union members increased, from 9\% to 11\%. Immigrants in the United States are an economically disadvantaged group in the labor market. Previous studies suggest that union members and other workers covered by collective agreements receive union wage premiums about 15 percent over nonunion members in the United States. Joining unions could be a good approach for foreign-born workers to receive higher wages. In this connection, the goal of this paper is twofold: one, to estimate the willingness of foreign-born workers to join unions, and two, to determine the union wage premium for foreign-born workers and whether there is a statistical difference in the union relative wage effect for foreign- and native-born workers. The results show that foreign-born workers have a lower probability of joining unions, ceteris paribus. The wage differential between union and nonunion workers for foreign-born workers is only 11.3\%, while that for native-born workers is 13.3\%. This 2-percent difference of the union impact on wages of native- and foreign-born workers is statistically significant. Among the foreign-born workers, Mexicans have the highest union relative wage effect (22.4\%). This study also finds that the union/nonunion wage differentials for both female foreign- and native-born workers are smaller than those for their male counterparts. Moreover, the union wage premium is greater for foreign-born workers in the private sector than for those in the public sector. By region, unions have higher wage impact in the West Coast than in the East Coast.

In light of the numerous criticisms leveled against estimating the wage differential between union and nonunion workers using the ordinary least squares (OLS) method, this study estimates the union impact on wages of foreign-born and native-born workers using the propensity score matching (PSM) methodologies (nearest neighbor and kernel matching methods) proposed by Rosenbaum and Rubin (1983), and compares the results with those obtained using the OLS approach. The data on the wages and salaries of male workers aged 16 years and above are obtained from the Current Population Survey and span the period 1994-2008. Both the propensity score matching and OLS estimates indicate that the union/nonunion wage differentials for male foreign-born workers lie between 12\% (OLS) and 27\% (PSM). In addition, our results suggest that there is little difference in the union/nonunion wage differential between native- and foreign-born workers. The estimates of the union impact on wages using the propensity score matching technique are higher than those derived using OLS for both native- and foreign-born workers. Furthermore, among the foreign-born workers, the union relative wage effect is found to be higher for Mexican-born workers (26-42\%), while for Asian-born workers it is lower and statistically insignificant based on the OLS. Further decomposition of the data into three different skill groups (high school dropouts, those with a high school degree, and those with a college degree or higher) reveals that, in general, the less skilled (high school dropout) foreign-born workers have the greatest union wage impact. However, the union wage premium is relatively larger for highly skilled workers (with a college degree or higher) among Mexican-born workers.

In chapter 3, we study the effect of job training on the US immigrant workers, using the 1996, 2001 and 2004, Survey of Income and Program Participation (SIPP) data. We improve upon prior studies by setting up our training evaluation model, studying the impact of training on both the average and the distributional earning of workers, and comparing the differences in the return to training for immigrant and native workers by applying the Quantile regression (QREQ) model, the DiNardo, Fortin and Lemieux (DFL) reweighting methods, and propensity score matching method (PSM). From our distribution study, we find that training has a positive effect on wages for immigrant workers for most parts of income distribution. The DFL reweighting technique shows that after removing all observable characteristics differences between trained and untrained workers, training still increases wage premium for both natives and immigrants throughout the income distribution. Our analysis provides strong evidence for the hypothesis that after corrected for observable characteristics differences between trained and untrained workers, the effect of training is relatively larger for rich natives, much larger for middle income natives and similar for the poor natives and immigrants. Furthermore, the PSM results show that the job training premium for foreign-born workers is between 0.063 and 0.184, whereas for native-born workers it is between 0.108 and 0.229. There is 4-percent difference in the job training premium between native and immigrant. All estimates are statistically significant. Our results suggest that OLS estimates underestimate the training premium.