Liberalization of European migration and the immigration of skilled people to Sweden

Abstract

Migration policies can have a strong impact on the selection of immigrants, who in turn can affect the host country’s innovation development. This paper examines the effects of the liberalization of migration on the skill composition of immigrants from the EU-15 to Sweden after the inception of the European Economic Area (EEA) in 1994. We examine its effect on immigrants’ education levels and probability of becoming an inventor, comparing immigrants from the EU-15 with those from other developed regions in difference-in-differences regressions. The results show that the liberalization of migration had a negative effect on the educational profile of new EU-15 immigrants in the short run, but there is no such effect in the long run. Moreover, the liberalization of migration has no systematic effect on the EU-15 immigrants’ probability of becoming an inventor in neither the short nor the long run. These patterns are consistent with the theoretical implication that reduction in migration costs associated with the EEA mainly stimulated migration from the lower end of the education distribution.

Appendices

Appendix 1

The member countries of EEA; GDP, population, and GDP per capital in EU-15 countries in 1994

EEA countries for which the agreement had an impact on mobility to Sweden are the EU-15 countries: Austria, Belgium, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, and the UK. Normally, Sweden, Finland, and Denmark would be included, but as they are also Nordic countries, they have already had free mobility since 1954 and hence are excluded.

Table 6 Member countries of EEA at its inception and corresponding GDP, population, and GDP per capita in 1994

Table 7 Employment rates (%) for immigrants who migrated at ages 18–64 and were employed within 2 years after they migrated to Sweden, by year of migration and region of origin

Table 8 Linear probability models for migrants aged 18–64 being low, middle, or highly educated at the time of migration (no imputation on education data), 1990–1997 vs. 1990–2007

Table 9 Ordinary least square (OLS) regressions of the probability of a migrant being low, middle, or highly educated at the time of migration and becoming an inventor from the time of migration until 2007, 1990–2007 vs. 1985–2007

Fig. 4

Percentage of immigrants from the EU-15 as a share of all immigrants to Sweden, 2000–2015. Source: Statistics Sweden

Fig. 5

Missing data on education level as a share of all observations, with and without imputation. Source: Statistics Sweden

Fig. 6

Growing trend for the number of patent applications by Swedish born, EU-15 immigrants, and immigrants from “Other developed regions,” 1985–2007. Index 1985 = 100. Source: Statistics Sweden and CIRCLE data on inventors

Appendix 2

The Grogger-Hanson model with skill-differentiated fixed costs

The original model distinguishes between low-skilled (L) and highly skilled (H) workers, where wages (w) are specified as

$$ \ln {w}_k^L={\mu}_k, $$

(3)

$$ \ln {w}_k^H={\mu}_k+{\delta}_k, $$

(4)

where μ _k is a constant and δ _k is the wage premium associated with high skills in country k (“Sweden”). L means low-skilled and H means high-skilled. The fixed cost of moving from home country 0 (i.e., “the EU-15”) to k is C_0k. In the model, utility is linearly and positively dependent on net income, such that for skill level j = L, H:

$$ {U}_{i0k}^j=\alpha \left({w}_k^j-{C}_{0k}\right)+{\varepsilon}_{i0k}^j. $$

(5)

Assuming that the error terms ${\varepsilon}_{i0k}^j$ are independent and identically distributed (i.i.d.) with an extreme value distribution, the (natural) log odds ratio of choosing country k over 0, following McFadden (1974), is

$$ \ln \frac{p_{0k}^j}{1-{p}_{0k}^j}=\alpha \left({w}_k^j-{w}_0^j\right)-\alpha {C}_{0k}. $$

(6)

The selection pattern is obtained as the difference in log odds ratios of Eq. (6) among skill groups:

$$ \ln \frac{p_{0k}^H}{1-{p}_{0k}^H}-\ln \frac{p_{0k}^L}{1-{p}_{0k}^L}=\alpha \left[\left({w}_k^H-{w}_0^H\right)-\left({w}_k^L-{w}_0^L\right)\right], $$

(7)

In Eq. (7), fixed costs are eliminated. This means that positive selection takes place if the absolute wage premium difference is positive. If average wages are the same, there is positive selection if δ _k  > δ₀, but Eq. (7) can still imply positive selection even if δ _k  ≤ δ₀ if μ _k is much larger than μ₀.

The effects of the EEA should be understood as increasing the possibility of going to Sweden without visa requirements, which thus lowers the cost of moving. Obtaining a visa was presumably easier (“low cost”) for highly skilled people than low-skilled ones who wanted to go to Sweden. This makes the Swedish case different from that of Austria, as Austria had “low costs” in its attraction of low-skilled workers from the EU-15 before the accession to the EEA (Huber and Bock-Schappelwein 2014). If we modify the Grogger and Hanson (2011) model to allow the fixed cost to vary, we can rewrite Eq. (5) to

$$ {\overline{U}}_{i0k}^j=\alpha \left({\overline{w}}_k^j-{\overline{C}}_{0k}^j\right)+{\overline{\varepsilon}}_{i0k}^j, $$

(8)

So that fixed costs differ by j, denoted ${\overline{C}}_{0k}^j$. We use a bar above variables to denote the pre-EEA situation and let the previously defined equations describe the post-EEA situation. We assume that ${\overline{C}}_{0k}^L>{\overline{C}}_{0k}^H$. Furthermore, Eq. (7) is modified as

$$ \mathrm{In}\frac{{\overline{p}}_{0k}^H}{1-{\overline{p}}_{0k}^H}-\mathrm{In}\frac{{\overline{p}}_{0k}^L}{1-{\overline{p}}_{0k}^L}=\alpha \left[\left({\overline{w}}_k^H-{\overline{w}}_0^H-{\overline{C}}_{0k}^H\right)-\left({\overline{w}}_k^l-{\overline{w}}_0^L-{\overline{C}}_{0k}^L\right)\right]. $$

(9)

Taking the difference between Eqs. (7) and (9) shows the change in selection patterns:

$$ \alpha \left[\left({w}_k^H-{w}_0^H\right)-\left({w}_k^L-{w}_0^L\right)\right]-\alpha \left[\left({\overline{w}}_k^H-{\overline{w}}_0^H-{\overline{C}}_{0k}^H\right)-\left({\overline{w}}_k^L-{\overline{w}}_0^L-{\overline{C}}_{0k}^L\right)\right]=\alpha \left[\varDelta {W}_{0k}^H-\varDelta {W}_{0k}^L+\left({\overline{C}}_{0k}^H-{\overline{C}}_{0k}^L\right)\right], $$

(10)

where $\Delta {W}_{0k}^j$ is the change in the wage differential for j-skilled workers following the reform and in comparison to the EU-15. The change in selection is (more) positive if the wage differential for highly skilled workers becomes more pronounced in the destination country after the EEA was introduced and if the returns to low-skilled workers decreased in the destination country relative to the home country. Although marginal taxes in Sweden fell in the early 1990s, they have risen somewhat since, such that Sweden has one of the highest marginal tax rates in the world (KPMG 2012). Moreover, World Bank data on Gini coefficients do not suggest that trends in Sweden with respect to income distribution changed substantially during the period 1990−2000 (Milanovic 2013). This means that it is hard to argue that net wages for highly skilled workers ($\Delta {W}_{0k}^H$) have generally increased for those who go to Sweden compared to other countries in the EEA. Therefore, $\Delta {W}_{0k}^H-\Delta {W}_{0k}^L$ has probably either gone down or is close to 0. That means that pre-EEA moving costs are critical to the results. Since we have argued that $\left({\overline{C}}_{0k}^H-{\overline{C}}_{0k}^L\right)<0$, this suggests a more negative selection change following the EEA.

Based on these arguments, we expect that the average skill level among immigrants from the EU-15 to Sweden declined after 1994, when Sweden became a member of the EEA.