Saturday, August 26, 2017

Dispelling the migrant myths

The sudden arrival of over a million people to France when Algeria gained its independence in 1962 raised the unemployment rate for low-skill French workers by about one fifth of a percentage point.

Other such flows had either no effect or a positive effect on the local labour market.

The arrival of 125,000 Cubans into Miami in 1980 had no effect at all on unemployment and was followed by a small rise in average low-skill wages. 

The movement of Soviet refugees into Israel in the 1990s, enough to raise the country’s population 12 percent in just four years, saw a substantial rise in the wages of the occupations they crowded into, not cuts.

A recent study looked at the effect of waves of refugees like Iraqis and Afghans into Denmark and found that refugees did initially displace small numbers of native workers but most often into jobs requiring more complex tasks and native language skills, where they were more productive. The most affected natives typically ended up earning 3 percent more than they had before.

The unemployment effects of Balkan refugees in the 1990s were different in different parts of Europe. The choice to protect incumbent workers made it more difficult for unemployed natives to find new jobs where they could complement rather than compete with newcomers.

In Turkey there appears to be a small negative effect on native employment caused by the influx of Syrians, but not in Jordan. These studies are limited because they measure what happens to native workers who stay in the places and occupations where migrants cluster. They do not fully account for adaptation by natives who change or move to a new place, job or skill level.

Almost a million people arrived in Germany in 2015 because other countries did not accept substantial shares of them. The labour market effects on Germany appear to be minimal so far. But those effects would be negligible on all countries concerned if the burden had been broadly shared.

The clearest historical experience in this regard is the enormous flow of Hungarian refugees into Austria in 1956. They produced a spike of 3 percent in the population of Austria. Had Austria been left alone to assist them, there is little doubt that those 200,000 Hungarians, trapped in Traiskirchen and other camps, would have become a large economic and fiscal burden. What happened instead was that a coalition of 37 countries, on many continents, shared the responsibility. They moved so quickly that they had resettled over half of the refugees within 10 weeks. They were nothing but an economic boon to the countries they went to; one of them came to the U.S. and ended up founding the tech giant Intel. Coordination eliminated the burden.

Almost all refugees receive substantial public assistance when they arrive and years afterward. Countries vary in how much of this assistance refugees are asked to pay back later. But by far the most important determinant of the net fiscal effect is how quickly refugees integrate into the labor market and start generating tax revenue. Countries that actively deter asylum applicants from working are making the decision to increase their net fiscal burden. In the U.S., the average refugee becomes a net contributor to public coffers eight years after arrival. The assistance they received when they arrived was, in purely monetary terms, an investment with a positive return.

The International Monetary Fund projects that the new wave of asylum seekers and other migrants will raise economic growth over the long term by providing Europe with young and energetic workers. The labour market effects and fiscal effects are likely to be negative and small in the near future for some countries but more positive in countries with more flexible labour market regulations. The greatest determinant of these effects will be the extent to which continuing flows are shunted to just a few destinations or broadly shared. This depends less on migrants’ decisions than on the policy decisions of the recipient countries.

No comments: