The recent inflows of refugees and migrants to Europe have raised new questions about how migration policies should be designed. Migrant-recipient countries are concerned not just with the number of migrants arriving, but also with their ‘quality’. This column argues that policies that screen migrants based on observable characteristics can have a detrimental effect on the quality of migrants (measured by their income). Such policies might thus fail to improve immigrants’ labour market outcomes at their destination.
By Simone Bertoli, Vianney Dequiedt and Yves Zenou*
The recent inflows of refugees and migrants to Europe across the Mediterranean Sea have raised new questions about how migration policies should be designed. Migrant-recipient countries are not just concerned with the scale of incoming migration flows but also with migrants’ quality. This is usually measured through their wages, reflecting their productivity on the labour market at destination. Evidence of a fall in immigrants’ initial earnings in recent decades (e.g. Borjas and Friedberg 2009 and Borjas 2015 for the US and Aydemir and Skuterud 2005 for Canada) has prompted debates around the need to reform immigration policies in order to reverse this declining trend. Specifically, a growing number of countries are moving towards an immigration policy that screens potential immigrants on the basis of their observable characteristics, such as education and language proficiency, and grants better chances of admission at destination to applicants endowed with individual characteristics that are deemed desirable by the recipient countries. The EU Blue Card and the points-based system in Canada are good examples of such policies. Destination countries can be selective not just de jure, but also de facto, as potential migrants with different levels of education might have differential abilities to overcome identical policy-induced migration costs (Bianchi 2013).
A gap in our understanding
Quite intuitively, an attempt to increase the share of highly educated individuals among immigrants should be unambiguously beneficial for a destination country that aims at raising immigrant quality. But we have to keep in mind, as George J. Borjas observes, that actually “remarkably little is known about […] whether the chosen policy, in fact, has the desired outcomes in terms of the size and composition of the immigrant flow” (Borjas 2014, p. 215). Specifically, we should not overlook that the screening devices that are adopted by the countries of destination are not the only determinants of the size and composition of incoming migration flows. The scale and the quality of incoming migration flows result from the interplay of the out-selection mechanism determined by immigration policies with self-selection into the pool of potential migrants. The effect of adoption-selective immigration policies such as a points-based system is not only determined by the ensuing change in the way immigrants are selected from the pool of applicants, but also by how the immigration policy influences the incentives to self-select into this pool.
What remains unobserved for immigration officers
Self-selection into migration is not only driven by individual-specific characteristics such as education, which are observable. It is also affected by factors such as innate talent or propensity to take risks, which influence the individual’s productivity in the labour market both at origin and destination, and thus the incentives to move across a border. Since the seminal contribution of Borjas (1987), we know that migrants typically differ from stayers on unobservable characteristics, which can exert a first-order influence on the variability of earnings across individuals. Specifically, “education accounts for only a small portion of the variance in earnings across workers” (Borjas 2014, p. 29), with unobserved heterogeneity explaining most of the variability in earnings across workers who have the same level of education (Cunha and Heckman 2008) and the dispersion in earnings being proportionally larger with higher levels of education (Chen 2008). This implies that, while immigrants with a higher level of education earn, on average, a higher wage than immigrants with a lower level of education, there is a considerable overlap between the two wage distributions. Drawing on the waves of the American Community Survey conducted between 2001 and 2013, we computed the real wage income of educated (at least completed college) and uneducated (at most high school dropouts) foreign-born males aged 25 and above at the time of immigration. Educated immigrants earn an average real wage income of $67,000, which is more than three times the corresponding figure for uneducated immigrants $19,000). Still, if we randomly match one educated and one uneducated immigrant, the probability that the uneducated immigrant earns a higher real wage income stands at 18% (notice that this figure would stand at 50% if the two wage distributions were identical).
The influence of selective policies on self-selection on unobservables: New evidence
The relevance of individual characteristics that are related to immigrant quality and that are unobservable for immigration officers suggests that we should analyse how selective immigration policies, which screen potential migrants on the basis of their observable characteristics, also influence self-selection on unobservables.
In new work (Bertoli et al. 2015), we provide a theoretical analysis of the effect of a scale-preserving increase in selectivity of immigration policies on migrants’ quality. The latter is measured by the logarithm of the wage at destination while the former is defined as a reduction in the policy-induced migration costs that educated potential migrants face matched by a corresponding increase in the migration costs for uneducated individuals that leaves the scale of migration unchanged. The analysis reveals that the migrant quality is a non-monotonic function of the share of educated immigrants when migrants are positively selected on unobservables, i.e. the average of their log wages stands above the expected value of the entire distribution of log wages. In such a case, migrant quality is maximised when migration costs are such that the quality of educated and uneducated individuals is equalised at the margin; the average log wage of an educated individual who is indifferent between migrating and staying coincides with the average log wage of the corresponding set of uneducated individuals. A further (scale-preserving) increase in selectivity would reduce migrants’ quality, notwithstanding the ensuing increase in the share of educated immigrants, whose quality that is, on average (but not at the margin), higher than the quality of uneducated immigrants. The analysis of the theoretical model also reveals that the quality-maximising share of educated immigrants is a function of the scale of migration. Specifically, an increase in the scale of migration is associated with a reduction in the quality-maximising share of educated immigrants.
Our theoretical model shows that a scale-preserving increase in the share of educated migrants can produce a detrimental effect on migrant quality when migrants have, on average, a higher level of ability than stayers. Thus, increasingly selective immigration policies might not just be “unfriendly to development” (Pritchett 2006), but they might also fail to attain their main goal of raising migrant quality, notwithstanding the fact that they increase the share of educated immigrants.
Aydemir (2011) provides evidence that immigrants to Canada admitted through the points-based system have a significantly higher level of education than immigrants admitted through non-selective channels, such as family reunification provisions, but they do not perform better on the Canadian labour market. Our analysis helps us understand why selection on observable characteristics might fail to improve the immigrants’ labour market outcomes at their destination.
*About the authors:
Simone Bertoli, Assistant Professor at the CERDI, University of Auvergne
Vianney Dequiedt, Professor of economics, University of Auvergne; director, CERDI
Yves Zenou, Professor of Economics at Stockholm University and CEPR Research Fellow
Aydemir, A (2011), “Immigrant Selection and Short-Term Labor Market Outcomes by Visa Category,” Journal of Population Economics, 24(2), 451-475.
Aydemir, A and M Skuterud (2005), “Explaining the Deteriorating Entry Earnings of Canada’s Immigrant Cohorts, 1966-2000,” Canadian Journal of Economics, 38(2), 641-671.
Bertoli, S, V Dequiedt, and Y Zenou (2015), “Can selective immigration policies reduce migrants’ quality?,” Journal of Development Economics, doi:10.1016/j.jdeveco.2015.11.002.
Bianchi, M. (2013): “Immigration Policy and Self-Selecting Migrants,” Journal of Public Economic Theory, 15(1), 1-23.
Borjas, G J (1987), “Self-Selection and the Earnings of Immigrants,” American Economic Review, 77(4), 531–553.
Borjas, G J (2014), Immigration Economics. Cambridge, MA: Harvard University Press.
Borjas, G J (2015), “The Slowdown in the Economic Assimilation of Immigrants: Aging and Cohort Effects Revisited Again,” Journal of Human Capital, forthcoming.
Borjas, G J, and R M Friedberg (2009): “Recent Trends in the Earnings of New Immigrants to the United States,” NBER Working Paper No. 15406.
Chen, S (2008), “Estimating the Variance of Wages in the Presence of Selection and Unobserved Heterogeneity,” Review of Economics and Statistics, 90(2), 275-289.
Cunha, F, and J J Heckman (2007), “Identifying and Estimating the Distributions of Ex Post and Ex Ante Returns to Schooling,” Labour Economics, 14, 870-893.
Pritchett, L (2006), Let Their People Come, Washington: Center for Global Development.