Why is it so hard to know the relationship between immigration and economic performance?

Here is the number of new permanent residents to Canada, as a percentage of the existing population, over time:

"New permanent residents" is not a perfect measure of immigration - it excludes temporary foreign workers, who have become much more important in recent years, and also ignores emigration, that is, the non-trivial number of newcomers who return to their home country. But it is as good a measure as we can get.

This chart, together with some basic information about Canadian economic history, demonstrates why it is hard to establish any kind of relationship between immigration and growth.

Sometimes - as when Canada accepted more than 37,000 refugees from the Hungarian uprising - immigration levels are determined by external events. But more often, immigration policy is a response to a domestic political agenda. The lowest immigration levels in the post-war period were experienced during the late 70s and early 80s. These were the years when last, and biggest, cohorts of the baby boom were hitting the labour market, and Canada's economic performance left something to be desired. Cutting immigration was a politically sensible move. But when the economy started to rebound in the 1990s, so did immigration targets.

Unfortunately the immigration data does not go back to the 1930s, but a similar pattern holds for that period. Severe economic depression was met with calls to restrict immigration, and the governments of the day were happy to respond.

The level of immigration is, to a significant extent, driven by economic conditions. Consequently, it is extremely difficult to know whether or not immigration promotes economic growth  If we see a positive relationship between immigration and growth in the data, we don't know if that relationship exists because immigration causes growth, or because growth causes immigration. (This paper argues growth drives immigration while this paper disagrees).

Some researchers have tried to solve this problem by looking at the impact of immigration at a sub-national level using city or regional data. This approach, however, is subject to the difficulty that immigrants tend to want to live there are jobs. If we see that cities with large immigrant populations grow more rapidly than those with few immigrants, is it because immigration drives growth and innovation, or because no sensible immigrant would want to move to a place with few job opportunities? Economists have used a variety of clever techniques to get around this "endogeneity" problem - an excellent though dated (1995) survey of the literature by Friedberg and Hunt and can be found here. Their conclusion:

The theoretical literature on immigration and economic growth suggests that the impact of' immigrants on natives' income growth depends crucially on the human capital levels of the immigrants. Empirical research on this question has yielded conflicting answers, and more work on this issue is needed. 

In the two decades since Friedberg and Hunt wrote that passage, more work on this issue has been carried out. Much of the research on this topic has been carried out using US data. However immigrants to the US, unlike immigrants to Canada, typically have lower human capital than the native born. There is a large group of scholars doing excellent work Canadian immigration policy. Here are links to fairly readable pieces by Ana Ferrer, David Green, Ted MacDonald, Mikal Skuterud, Chris Worswick, and Casey Warman.  Other notable names (at the risk of omitting people) are Arthur Sweetman, Garnett Picot, Charlie Beach and Canadian-in-exile David Card.

I am very doubtful that it is possible to predict, with any degree of certainty, how the current government's plan to "substantially increase" immigration levels will impact the Canadian economy - and the well-being of both native born and immigrant Canadians -  in the long run. In the past, we might have feared that immigrant engineers would end up driving taxis. Now that uber and self-driving cars are wiping out taxi driver jobs, it's hard to know what will happen.

 

Comments

[Avon Barksdale]

I find this discussion focusing on the wrong question. In effect you are asking what the optimal number of immigrants are. Economics teaches us not to think that way. The Soviets could get the price if coffee right. Why do we think we can do better?

Let people choose. Setting quotas is exactly the wrong way to proceed.

[Frances Woolley]

Avon: "In effect you are asking what the optimal number of immigrants are. Economics teaches us not to think that way."

Yes, you got it! "Optimal immigration theory" is precisely where I am going here.

Some economists are uncomfortable with that kind of thinking. I'm not. Your "let people choose" position is a normative statement about the most desirable kind of immigration policy (i.e. open borders). People who say "close the borders" are also making some kind of normative statement.

I would guess that the overwhelming majority of Canadians favour some kind of immigration policy that is neither completely closed nor completely open borders. So then the question is "what is the optimal number of immigrants?" If economists don't answer that question, other people will - and will do so without the analytical tools and empirical sophistication that economists bring.

[Jonathan Wiley]

Not everyone has the same intuitions as economists generally do. Complacently ignoring that fact can be dangerous. I'm writing this from a country where this very issue has been connected to the possibility of some sort of coup depending on what happens in the next month.

[Jonathan Wiley]

And if you want evidence of this intuition gap, you don't even have to leave this blog.

Blast from the past:
http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/05/how-not-to-evaluate-immigration-policy.html

[Jacques René Giguère]

Those who favor an open border suffer from fallacy of composition and a forgetting of ceteris paribus.
Letting a thousand people (or even one hundred thousand) is definitively differant than letting one million.
The few can adapt and use the preexisting level of physical and social infrastructure. The many overwhelm it. They are not bad persons, just too numerous.
5 millions caboclos moving into Rio de Janeiro doesn't make 5 millions more middle-class cariocas. Just a struggling favelas.
100K caboclos could make it in Canada. 10 millions would simply turn Canada's big cities into third-world urban nightmares.

[Lorenzo from Oz]

What Jacques said.

When we have robust models of what drives economic growth, the open borders folk may be worth listening to. Until then, erring on the side of caution is much more sensible. (Which, btw, I think Australia and Canada mostly due and the US is rather worse at, even though our rates of immigration are much higher than the US's.)

[Lorenzo from Oz]

That should be "mostly do".

[Avon Barksdale]

This post asks the wrong question. I think people like John Cochrane make a lot more sense on this issue.

What is the optimal number of imported tomatoes? Communist central planners tried to figure things out this way. We don’t. We decide on the optimal terms on which tomatoes can be imported, and then we let the market decide the number. So, let’s debate the optimal terms for immigration. How will we let people immigrate? What kind of people? Once we decide, let anyone who wants to immigrate come. On the right terms, the number will self-regulate, just like it does for tomatoes.

Econ 101: Figure out the price, set the rules of the game; don’t decide the quantity, or determine the outcome. When a society sets targets or quotas, the result is almost always a monstrous waste (see Venezuela). With a quota system, do you really think that our government bureaucrats could spot the family of a potential Sergey Brin, people who have little way of signaling just how great their contribution might be?

You might say that immigrants will overwhelm our welfare state. On moral grounds, it’s interesting that Canadians feel that their taxes should support an unfortunate stranger who happened to be born in New Brunswick and not one who happened to be born in Aleppo, but let’s put a pin in that. We can easily structure a deal that protects the finances of the government. Immigrants could pay a bond at the border, say $10,000 or whatever. If they run out of money, commit a crime, don’t get health insurance, etc., the bond pays for their ticket home. Or the government might establish an asset and income test - immigrants must show $10,000 in assets and a job or a working business within 6 months or you’re out. Overuse of social services is simply not a problem if we structure the deal.

Let’s think about the deal, not quotas. For every objection to open immigration, it’s easy enough to find terms of the deal that resolve the issues. The right terms will allow the optimal amount of immigration to settle itself, so that no bureaucrat in Ottawa (or professor at a university) has to come up with a quota. Get the terms right, and let every person who can benefit Canada come.

[Ralph Musgrave]

There's a glaring flaw in every study of the economics of immigration that I know of: they fail to take account of the fact that each net immigrant requires several tens of thousands of dollars of infrastructure. The cost of that is not charged specifically to immigrants: it's a cost born by the population at large.

[Brett Reynolds]

@Avon: I'm not an economist, so I'll probably get much of this wrong, but I'm sure that people aren't tomatoes.

You say, "figure out the price." That's sounds easy enough when when you can assume a homogeneous good and perfect information, as you can in Econ 101. With immigration, our "good" is quite varied and our information about demand is pitiful.

Of course, there's demand in terms of what companies will pay for people's labour, but demand for people's labour is a poor metric of demand for people. Then there's the question of what people are willing to pay out of their own pockets to help others because it makes them feel good. And what they're willing to pay when the people are in another country is much different from what they're willing to pay once they meet the people.

Next, consider what we are willing to pay to avert a disaster in another country, which could have long-term costs for us. We'd like to buy insurance against the contingency that these conflagrations would affect our peace or our economy, but there's a market failure for this kind of insurance. Instead, we send money to the UN and put money in our armed forces.

When a grocery store overestimates the number of tomaotes needed, they rot. That's not a big deal when it comes to tomatoes, not for the store and not for the tomato. It is potentially a very big deal, both for countries and for individuals when they are left to rot.

In both the case of over and underestimating the need for tomatoes, the company's profit is not what it would have been. If a the grocery store is bad enough at getting this right, then it goes out of business, and the capital that was tied up there can be used for something profitable. Again, when it comes to people's labour, there's something like this going on at the corporate level, but when it comes to people themselves, this kind of "creative destruction" doesn't apply.

In short, when it comes to immigrants, there are a number of market failures, mostly related to lack of information, which means you can't leave it up to the market.

[Stephen Cryne]

One thing that I believe is being overlooked in this article is the need to focus on economic immigrants. Those are people with the skills and experiences Canadian businesses are clamouring for. today over 1/3 of employers can't fill positions. Hardest jobs to fill...Skilled trades, Engineers, sales professional and managers/executives. This is not just a "Canada problem" either. Developed countries around the globe (Australia, UK, US, Germany...even Hungary)cant find the people with skills today.

Canada should be developing immigration programs to bring the brightest and the best to our shores...not to keep them out.

The focus of the discussion should not be just on the numbers...but on the kinds of immigrants we select.

[Avon Barksdale]

Ralph and Brett,

Did we need to regulate the number of tourists? Tourist use public infrastructure. But we don't have a Japanese tourist quota. The issue here is about restricting people who might want to work. Determine the deal, not the numbers.

At the extreme suppose the deal was: you must have $6 million in the bank, have a job or business within 6 months of arrival, and have a graduate degree. What would it matter if 10,000, 100,000 or 1,000,000 of those types of people showed up?

No quotas. No "optimal numbers". Think about the deal.

[Frances Woolley]

Stephen Cryne: "today over 1/3 of employers can't fill positions."

Erik Lund gave a great description of a Vancouver supermarket's so-called labour shortage in the comments on the previous post. It's so good I'll reproduce it here:

Sure, the employment pitch is literally: "Wanted: frozen foods category manager; responsibilities include stock control, replenishment, restocking, and organising grocery store deep freezers and handling outside vendors. Work conditions will include deep freezer environment while maintaining world-class customer service and industry-leading personal appearance guidelines. Requirements: high level of physical fitness and upper body strength with implicit height requirement, unlimited availability, including night shifts. Compensation: minimum wage less union dues, no night shift or management premiums, no guarantee of minimum hours. Scheduling by seniority."

You can, as an economist, reasonably ask what the hell is going on when a labour shortage is met with a response like this.

Supply curves slope upwards. I'll start believing that there's a labour shortage when I start seeing upwards pressure on wages.

[Brett Reynolds]

@Avon:

1) Tourists (usually) leave after a short stay, immigrants usually don't.
2) Because of market failures, the question is how to come up with "the deal" or how to set the price. The price should include externalities. Research into the effect of migration on carbon output helps us understand that and put a price.
3) Look back at the post and subsequent discussion and you'll see that nobody here is recommending quotas.

[Joseph Savon]

Has there been any useful work on the Canada's capacity to support life at 1st World ( or any other) level?
My grandparents in Ireland raised 5 healthy children on less than a hectare, & Ireland still supports 6.4E6 people on ~1.2E6 arable hectare. Canada has ~5E7 arable hectare, which would suggest a population limit of ~230E6.

[Avon Barksdale]

Brett

What are all these externalities? You keep saying that but what are they? Where were all these externalities when millions showed up in the 19th and 20th centuries?

Yes, Frances is talking about an optimal number and that's poor economic reasoning in my opinion. Think about the optimal terms, then let the numbers sort themselves out. We know that Frances understands this line of thinking as she astutely points out that labour isn't in shortage if we don't see wages move up. It doesn't require some complicated model to see what the optimal numbers should be, watch the prices.

The hubris of central planning is that government economists think they can write down some model and tell us things they can't possibly know. I'm much more worried about bullshit economics like the report that Stephen referenced in 2011 (and I mean the term in the technical Harry Frankfurt sense) than some loosey-goosey notion of an "information externality".

[Brett Reynolds]

Avon,

Carbon dioxide emissions could be one externality.

If you don't think the government economists can write down a model and tell us how many people we need, why do you think they can write one down to tell us what deal we need to offer?

[Avon Barksdale]

Brett

Carbon dioxide can be handled by carbon taxes. That issue has little to do with immigration. The externalities of carbon emissions, to the extent they are a significant enough effect, are mostly borne by people in the developing world, not people who live in Canada. Few argue against open immigration because it will make the great grandchildren of Africans worse off!

Economics is very clear - it's very hard if not impossible to manufacture an equilibrium. That's why the Soviet Union failed. We set the terms and let the market relax. We don't set employment quotas or provincal population quotas like the Soviets did. We let individuals sort it out for themselves. Economists who think they can set quotas are fooling themselves. Look at the bang up job they did with the cod fishery, etc.

[Avon Barksdale]

Brett

"1) Tourists (usually) leave after a short stay, immigrants usually don't."

Are you sure? At any given time how many tourists does Canada have inside her borders? When one leaves another comes in. If 25% of Japan decided to visit Canada next year, would we stop them from coming? Suppose that at any given time 25% of Japan was always inside Canada on vacation, would we stop that?

Immigration restrictions have nothing to do with people coming to Canada. It's about stopping them from working.

[C Trombley]

I think Avon is completely correct: the right way to think about immigration is in terms of price. Avon's analogy to a product is correct: immigration is difficult and expensive, illegal immigration is a bootleg product that comes with all the dangers. It's no different from me buying a guitar - that comes with the externalities of me making lots of noise (and worse, singing terrible old folk songs like I'm Mississippi John Hurt).

Anyway, Nick Rowe develops an immigration game where the population ends up "too concentrated" here:

http://worthwhile.typepad.com/worthwhile_canadian_initi/2016/02/anti-urban-economics.html

In this game _everyone_ is an immigrant so everyone bears the cost evenly. In real life, only some people are labeled "immigrants", so there is political pressure to push the costs of overconcentration on them. That's the essence of the immigration thing, I think.

[Brett Reynolds]

So how do you set the price?

[C Trombley]

Reynolds:

Ideally, an auction that can go negative if the Nash Equilibrium is below the symmetric optimum. In reality: pretty much arbitrarily and always positively. Also, most of the cost is in red tape, even then most of that is in time and uncertainty rather than cash. :/

[Avon Barksdale]

Brett

The terms are the "price". That part we can debate. That's the part the Parliament should be talking about. What are the terms? What consensus can we reach on the terms? On one extreme, you need to be a billionaire to immigrate; on the other we take everyone. But once we set the terms we don't pick numbers or quotas. Like I said, if the terms were everyone has to have a graduate degree and $6 million in the bank, why would we care how many of those people came.

[Stephen Cryne]

I am NOT talking minimum wage jobs here, consider...More than half of the high-skilled technology workers and entrepreneurs in Silicon Valley are foreign-born. Four Anglo-Saxon countries—the United States, the United Kingdom, Canada and Australia—constitute the destination for nearly 70 percent of high-skilled migrants (to the OECD) in 2010 (not minimum wage jobs). The United States alone has historically hosted close to half of all high-skilled migrants to the OECD and one-third of high-skilled migrants worldwide. In 2010, the United States hosted 11.4 million skilled migrants, 41 percent of the OECD total. 40% of the fortune 500 firms were founded by immigrants or children of immigrants. Canada has to compete for highly skilled talent, and our immigration system is one way of doing that.

[Jonathan Wiley]

Avon: "The hubris of central planning is that government economists think they can write down some model and tell us things they can't possibly know."

Do you think the hubris lies in the policy suggestions themselves or in the speed with which policy changes are made?

[Avon Barksdale]

Jonathan,

Only people can have hubris, policies can't.

[Erik Lund]

"I am NOT talking minimum wage jobs here, consider . . . [highly skilled immigrants are awesome sauce] . . . Canada has to compete for highly skilled talent, and our immigration system is one way of doing that."

Curiously, neither was I. I was talking about a position for which my employer has elected to offer a minimum wage. The position was vacated by an employee earning, not to parse the contract in its gory details. in excess of $26/hour. Now, it is perfectly reasonable to label it an "unskilled" job, in the sense that you can't take out a student loan to do it. But, just to launch into a non sequitur here, if you've thinking of a highly mobile, internationally-in-demand skilled trades employment, consider being a rigger sometime. Just arrange to be the son, daughter, nephew or neice of an existing rigger, and you're in! (That is, there are, or up to quite recently were, no trade college courses in "rigging.") Why? If there were, your son/daughter/nephew/niece might face competition on their next job, and their compensation might go down! Before you know it, there would be two trained riggers for every job, everyone of them desperate enough to start paying off their student loans that they'd accept any rate of pay the employer offered! And if wages started to tick up, you could just bring in some Bangladeshi riggers.

Oh, wait. That wasn't a non sequitur at all.

[Neil]

I am going to propose a baby quota.

Babies can't speak either of Canada's official languages, are very costly, and require tens of thousands worth of infrastructure throughout their lives.

First, we shell out an average of $9,700 just to get them here, while taking up space in billion-dollar hospitals. Then they consume another $142,000 of public money to give them a high school education. Not to mention the tax credits to parents, childcare subsidies, and general use of public infrastructure...the list goes on and on. Even after they graduate high school at long last most will still not become contributing members of society for several more years while they pursue yet more subsidized education.

We need to keep these numbers low so as not to overwhelm our capacity to handle them.

After all this it will take decades to pay back the public purse, if they ever do. In order to maximise our payback, we only want the best and the brightest, so we need a system to prescreen potential parents. We know that low income parents are likely have low income children, so that's the low hanging fruit. I'm sure we could analyse other demographic factors to screen out to maximise the odds of success. After that, we should require some sort of bond, which will be forfeit if your children commit a crime or turn out to have some sort of handicap preventing them from being net contributors.

----------------
I'm hoping the sarcasm is apparent in the policy suggestion above. If you subscribe to the view, as I do, that all humans are of equal worth, then our system for controlling immigration is just as amoral as my proposed system for restricting birth.

I'll agree that the problems of the world aren't something that Canada can tackle alone, but there's no reason we shouldn't be working with international partners to improve mobility rights for all people. In the meantime, Avon's suggestion of setting terms of immigration rather than numerical quotas seems like a step in the right direction.

[Jonathan Wiley]

Avon

I wasn't trying suggest that policies themselves can have hubris. I wanted to ask you whether you think the "social planners" would still demonstrate hubris if they advised the same basic policy goals based on the same models but advocated transitioning from the present set of policies to their preferred ones in a gradual process over a relatively long period of time, rather than abruptly over a short period of time.

[Avon Barksdale]

Neil,

Love it! I support open borders. If the compliant by the people who oppose me is that the welfare state can't handle it, fine, then let's talk about the deal. But let's stop talking about optimal numbers, quotas, or industry targeting.

Jonathan,

I'm not sure what you mean. I think that economists who think that they are social engineers drink a lot of the kool-aid that they make. It very quickly becomes toe of frog, eye of newt. As an example, I had a drinks with a group of colleagues this week and a government policy advisor on climate change came along as a friend of a colleague's. She told me how they are getting the subsidies on tidal energy just right. I nearly spit out my beer! I could tell that she was really sincere, that she really believed what she was saying. I asked her, "If you can do that, why can't you compute the optimal level of investment in just about everything?!" It puzzled her, but I could see that she thought that she might be able to do it! That's the hubris I'm talking about.