Immigrants are increasingly important source of labour in Canada, the result of an aging Canadian population and low domestic birth rates. Immigrants, however, do not tend to perform well in the labour market in terms of wages (Hou and Lu 2017). Coulombe et al. (2014) find that the quality of human capital is considered by Canadian employers to be lower if credentials are obtained outside of Canada. Still, as Nadeau (2013) mentions, government employers may be in a better position to evaluate foreign credentials because they are large employers. Furthermore, the wage distribution in the public sector tends to be more narrow compared to that in the private sector. As a result of these factors, we may expect to see a smaller (if any) wage gap between immigrants and the Canadian-born in the public sector compared to the private sector.
The sparse evidence to date does suggest that the immigrant wage disadvantage is smaller in the public sector (Nadeau 2013; McIvor 2016; Zheng 2017). However, these studies generally only compare immigrant versus non-immigrant earnings within and not between sectors, nor do they address the wage gap at different levels of government. Current work in progress (Mueller, 2018) shows that there is a great deal of heterogeneity between these different groups of public employees within each province and also between provinces. We intend to document the immigrant wage differentials in the public and private sectors using two separate data sets: the Labour Force Survey (LFS) spanning 3 years on either side of 2015 (to get adequate sample sizes), and the 2016 Census (which contain data for 2015). These data both have strengths and weaknesses but used together are complementary. For example, the LFS, contains a variable for union status, an important determinant of wages, but lacks variables to disaggregate immigrants by admissions class. The Census, by contrast, contains a detailed class of admission variable but lacks a union status indicator. Using both datasets allows us to compare immigrant wages to those Canadian-born and both within and between the two sectors. Both the LFS and Census master files contain detailed public sector variables which allow this sector to be disaggregated into federal, provincial, and local levels of government administration as well as public sector jobs that are not public administration (e.g., health care, etc.).
Standard wage decomposition techniques (e.g., Blinder 1973; Oaxaca 1973) will be employed to see how changes in the gap can be explained by factors such as higher levels of education. Quantile regressions will be employed along with the decomposition techniques outlined in Fortin et al. (2010) to determine at which points (if any) in the wage distribution we still see wage gaps that cannot be explained by the usual factors that influence wages.