By Eric Crampton 10/09/2015


Richard Meadows at the Press asked me whether I’d seen any reports on the costs and benefits of accepting more refugees. I’ve not seen any real cost-benefit assessments, though I have seen indications that, in US data, huge numbers of refugees had zero effect on local labour markets – David Card’s work on the Marial boatlift.

I’ve copied below what I sent Richard; his final piece covers it very well though in shorter form.

“What it costs” depends so much on policy choices that it’s hard to give a straight answer. The government’s citing $80k per refugee, but I don’t know what that’s based on. If you’ve seen a source on it, I’d appreciate hearing about it. You could imagine that the total amount spent on somebody over several years, including, say, welfare payments at the start, any job training, education for the average number of kids, health services and the like all bundled up together over a few years could tally up to $80k, but I’d need to see the workings. I’d also be curious to know their estimates on the employment rates of those refugees over time and their earnings.

One somewhat pernicious feature of a progressive tax structure combined with WFF credits is that the average taxpayer pays very little on net to the Crown after you account for the value of WFF credits and in-kind benefits provided by the government. On the one hand that means that richer people that nobody much likes wind up paying for most things, but it also means that unless an incoming migrant is able to get quickly into the higher earning echelons, folks inclined to see others as a cost to the state would conclude that the migrants are a cost to the state. Do you blame the refugee for that or should you blame the tax structure?

Here are things I do know:

  • American work shows that even huge numbers of refugees don’t wind up depressing wages in the local labour market. See David Card’s work on the Marial boatlift, in which over a hundred thousand Cuban refugees showed up in Miami over a few months in 1980. It was a 7% increase in the local labour force; no negative effects on local wages.
  • Canada solves the problem of supporting refugees by encouraging community groups and private citizens to sponsor them. The sponsors agree to provide care, lodging, settlement assistance and support for 12 months, or until the refugee is self-sufficient, whichever is first. The sponsors provide a whole pile of additional support. See 2.6 here. The government gives financial support up to about $10k for refugees that aren’t privately sponsored; German support is a bit above that. But note that those are cash transfers and wouldn’t include things like the costs of education and health care that could be included in the NZ figure.

There’s a separate question of whether we do best by supporting refugees here or by sending cash to help them in Europe or elsewhere. Airplane tickets aren’t free, and it can be harder to get established in a place where there are fewer people from the same community. Do we do better by sending cash instead? Maybe. But then we get into the much broader question of where that kind of money can do most good. Suppose that the government’s right that settlement support is $80k.

For $80k, NZ would strongly improve someone’s life relative to what it was in Syria, or relative to what it might be like in some refugee camp on Europe’s borders. But $80k would also buy life-changing surgery for over 170 women in the third-world. How should we weigh such things? If we’re motivated by a desire to do good, is this the best good we can do? If there’s no way that the community could be rallied to support 170 women’s life-changing surgery in the same way that it could be to support one refugee, then the latter could be worth doing. None of this stuff is easy.

I’m not sure that’s particularly helpful. It’s not an easy problem, and I can’t point to well-established cost studies for New Zealand. I also worry about ‘costs to settle versus economic benefits to current New Zealanders’ calculations as they miss out on the benefits provided to the refugee who gets a shot at a new and better life. All of that should be in a mix in a proper net benefit assessment so it could be reasonably compared to other good and beneficial things that we can do, either privately or through the government.

Since then, I saw Peter Singer’s estimate that accommodating a refugee in Jordan costs 3000 Euros per year. Do we do better in supporting 15 refugees in Jordan or supporting one here?