The case for immigration restrictions is periodically debated in the political arena. The refugee crisis in southern Europe in recent months and the increased number of asylum seekers, who may turn into undocumented economic immigrants, has spurred discussion for stricter border enforcement and controls. In the UK, David Cameron promised few years ago to bring annual net migration down to ‘tens of thousands’ and put a cap on skilled non-European immigrants. With the British economy recovering, the cap has become binding for the first time in June of this year. Firms report problems in finding the right type of workers (Economist 2015a), yet relaxing immigration restrictions has little to no support among parties in the UK (Economist 2015b). In February 2014, Swiss voters narrowly approved a referendum to curb immigration from the EU after resentment to immigration grew, following the opening of labour markets to European workers (New York Times 2014).
The effects of immigration restrictions on the inflow of immigrants and, in turn, on native workers’ employment outcomes are prominently discussed among policymakers. In spite of high sounding statement about the need of ‘pulling up the draw-bridge’ to avoid a flood of immigrants who can take away jobs, there is little direct evidence in the economic literature on how more open immigration policies affect immigration flows and, in turn, native labour market outcomes. Existing studies have analysed the effects of immigration flows, comparing regions that receive more or fewer immigrants within a country, and interpreting the differences in outcomes as driven by immigration. However, examples in which different policies were adopted in otherwise similar regions – allowing a causal analysis of immigration policies on flows and on native labour markets – are rare.
The literature has leveraged the tendency of immigrants to settle in regions with a network of compatriots (e.g. Card 2001, Peri and Sparber 2009, Dustmann et al. 2013) or it has used ‘push episodes’ from sending countries such as the collapse of the Soviet Union (Borjas & Doran 2015) or the wave of refugees from Cuba (Card 1990) to learn about their labour market effects. These episodes, however, are not under the control of the receiving country and therefore from them we do not learn much about the effectiveness and labour market consequences of different immigration policies.
In a recent paper, we exploit the Swiss integration into the European labour market after 1999 and study the causal effect of removing restrictions on immigrant flows and on native employment and wages (Beerli and Peri 2015).
Access to the Swiss labour market: border vs. non-border regions
The Swiss case was unique in that two different parts of the country experienced different timing in the implementation of the free movement policy. Labour market access for cross-border workers (foreign workers commuting from Italy, France, Germany, and Austria), was gradually eased beginning in 1999 and fully liberalised in 2004.
- This type of workers could only work in the border region (BR), which encompasses municipalities close to the national border.
- They were not allowed in the non-border region (NBR).
Labour market access for other immigrants, who intended to work and settle in Switzerland, was also eased between 1999 and 2007 but symmetrically in all regions.
After 2007, the free movement policy was fully executed for cross-border workers and for all other EU immigrants in both regions.
Thus, the two different schedules created a time-window between 1999 and 2004, in which the border region became gradually more open to immigrants than the non-border due to cross-border workers. The difference in openness became most pronounced between 2004 and 2007 when cross-border workers had free access to border regions but no access to non-border regions.
We leverage this differential degree of openness of the border relative to the non-border to analyse the effect of removing immigration barriers on the inflow of new immigrants and on native labour market outcomes. As cross-border workers and other immigrants had similar demographic characteristics and skills, we look at the total share of foreign-born in the labour force and we adopt a flexible difference-in-difference framework. We analyse differential outcomes in the period 1999-2004, during which border region progressively eased the entry of cross-border workers, and in the period 2004-2007, in which cross-border liberalisation was fully executed in the border region.1
The effect on immigration
Figure 1 displays the evolution of the labour force share of new immigrants in border and non-border regions (top panel) and the difference between them (bottom panel) during the period 1994-2010. The figure shows that the share of new immigrants moved together prior to 1999 (pre-reform), with a difference around 7 percentage points. Between 1999 and 2010, however, the share of new immigrants increased from 12.6% to 18.2% in the border regions, and from 5.5% to 7.4% in the non-border regions. Hence, new immigrants as share of the workforce increased by 3.7 percentage points more in the border compared to the non-border region.
Figure 1. Share of new immigrants in BR and NBR (top panel) and their difference (bottom panel)
Notes: The left panel plots the evolution of the share of new immigrants on the total workforce in the border region (BR, left y-axis) and the same share in the non-border region (NBR, right y-axis). The right panel plots the difference in the share of new immigrants between both regions.
A more rigorous regression exercise yields a similar finding. The difference in share of new immigrants between border and non-border regions was stable prior to 1999 but it increased thereafter by about 4 percentage points up until 2010. Figure 2 shows that the cumulated gap in immigrant share, after controlling for several labour market characteristics, was significantly different from 0 for the first time in 2004 and remained significant after 2007.
Figure 2. Estimated difference in immigrant share, between BR and NBR, coefficients and 5%-confidence intervals (base year = 1998)
Notes: The figure plots coefficients (straight line) and the 5%-confidence interval (dashed lines) of a difference-in-difference estimate with immigrants as share of labour force and including municipality and year fixed effects as controls for local, industry-driven demand shocks.
The effect on labour market outcomes
Having established that relaxing the restrictions on EU immigrants induced a significant growth, although certainly not a flood of immigrants, the next question we tackle is whether and how this influx affected natives’ labour markets. We exploit the same differential policy treatment between border and the non-border regions and find that average wages and hours worked by natives were not affected by it.
- However, when analysing the effects separately by education group, we find that natives with a college degree benefited from the liberalisation policy in terms of higher wages;
- Middle-educated workers (with upper secondary education but no college degree) suffered some decline in employment.
- Low-educated native workers were not affected.
This finding is puzzling at first, as the largest group of newly arriving immigrants was college educated and a standard model with skill complementarity would suggest that highly educated natives compete most directly with immigrants and should endure more negative effects.
To explore this puzzle, we investigate whether immigration had an effect on the job specialisation of natives. Previous research (Peri and Sparber, 2009, 2011, Foged and Peri 2013) suggests that natives’ job specialisation is likely to respond to the inflow of immigrants. As more immigrants take jobs, natives move to occupations where competition from immigrants is lower and complementarity effects are stronger. We find that highly educated natives were able to climb into higher management positions in response to the inflow of similarly educated immigrants, explaining some of their wage gains. On the other hand, middle-educated natives did not upgrade their positions but, instead, were reallocated to less-challenging job tasks and were replaced by immigrants in the ‘intermediate task’ range. So while in general natives responded to immigrant inflow, highly educated experienced an ‘upgrade’ of their jobs while middle-educated did not.
Discussion and conclusions
Our findings add to the existing evidence on the effects of immigration on native workers.
- First, by directly tackling immigration policies, we see that fully opening the border to neighbour countries increased immigrants only by 4 percentage points of the labour force over eight years.
- Second, we find that such an increased inflow did not have significant aggregate effects. Some groups of workers, however, experienced wage benefits while other experienced employment losses.
The recent research in this area has recognised the importance of looking at policy changes within countries to learn about their causal impact on foreign workers and labour markets. Dustmann et al. (2015), using an episode that allowed cross-border Czech workers into Germany, showed that in the short-run a substantial influx of unskilled workers reduced wages of unskilled young natives while unskilled older workers suffer employment declines. They explain this finding with the fact that old workers might either have larger labour supply elasticity (due to attractive outside options) or face larger wage rigidity than young workers. Our analysis shows that employment responses might also depend on the degree of occupation and task adjustment by native workers with different education levels. Such adjustment affects wage and employment effects.
Beerli, A and G Peri (2015), “The Labour Market Effect of Opening the Border: New Evidence from Switzerland”, NBER Working Paper 21319.
Borjas, G J and K Doran (2015), “Cognitive Mobility: Native Responses to Supply Shocks in the Space of Ideas”, Journal of Labour Economics, 33 (1), 109—145.
Card, D (1990), “The Impact of the Mariel Boatlift on the Miami Labour Market”, Industrial and Labour Relations Review, 43 (2), 245—257.
Card, D (2001), “Immigrant Inflows, Native Outflows, and the Local Labour Market Impacts of Higher Immigration”, Journal of Labour Economics, 19 (1), 22—64.
Dustmann, C, U Schönberg and J Stuhler (2015), “Labour Supply Shocks and the Dynamics of Local Wages and Employment”, Manuscript, University College London, March 2015.
The Economist (2015a), “How to kneecap the recovery”, 18 June.
The Economist (2015b), “Raise the Drawbridge”, 9 April.
Glitz, A (2012), “The Labour Market Impact of Immigration: A Quasi-Experiment Exploiting Immigrant Location Rules in Germany”, Journal of Labour Economics, 30 (1), 175—213.
The New York Times (2014), “Swiss Voters Narrowly Approve Curbs on Immigration”, 9 February.
Peri, G and C Sparber (2009), “Task Specialization, Immigration, and Wages”, American Economic Journal: Applied Economics, 1 (3), 135—169.
1 An important concern is that both regions could have experienced different pre- or post-1999 economic trends after the burst of the dot-com bubble. In the paper we establish that both regions are similar in terms of their industrial structure, and we control for local, industry-specific demand shocks and show similar pre-1999 economic trends.