We’ve seen a lot interesting and deservedly favourable commentary about Alex Tabarrok’s argument that liberalising high-skill immigration policy in the US would be an economic boon.
Hey, we’re all for it – and it’s more politically feasible than wider substantive reform on illegal immigration, in part because its entrepreneurial and societal benefits are intuitively easy to grasp.
But it also reminded us that we’ve been meaning to write something about low-skill immigration policy, which isn’t as glamorous but, we think, still has under-appreciated economic importance.
The jobs we take
In a 2005 paper, immigration economist Giovanni Peri explained the concept of imperfect substitutability, which is mostly a jargon-y way of saying that the jobs taken by new migrants tend to be different from those taken by native-born workers (emphasis ours):
One reason for this imperfect substitutability is that, for given skills, U.S. and foreign born workers often choose different occupations (see Card 2001 for more detail).
This is particularly true for workers with high and low levels of education (rather than with intermediate levels).
For instance among the lowly-skilled (HSD), foreign born workers are highly represented in occupations like tailoring (where 54% were foreign born in 2000) and plaster-stucco masoning (where 44% were foreign-born in 2000), while U.S.-born workers are highly represented among, say, crane operators (where less than 1% was foreign-born in 2000) and sewer-pipe cleaners (where less than 1% foreign-born). …
Foreign-born workers are highly represented in scientific and technological fields (45% of medical scientists and 33% of computer engineers are foreign-born) while U.S.-born workers are highly represented among lawyers (less than 4% are foreign-born) or museum curators and archivists (less than 3% are foreign-born).
Moreover, even within the same profession, often the U.S. and foreign-born provide different services, and hence benefit from complementing each other, regardless of education level.
For instance, Chinese and American cooks do not produce similar meals, nor do Italian and American tailors provide identical types of clothes.
Similarly, a European-trained physicist (more inclined towards a theoretical approach) is not perfectly substitutable with a U.S.-trained one (more inclined towards an experimental approach), and a French architect will likely create a starkly different building than an American one.
The point that Peri goes on to make is that the productivity gains from specialisation and economic benefits generated by immigration are especially amplified at both the high-skill and low-skill levels.
And once this process of specialisation has begun, its reversal can be painful for the local economies that try to undo it.
The ripple effects of labour shortages
We’ve been watching the story of how crackdowns on illegal immigrants in various US states have led to local labour shortages this year.
Not all of these states have passed laws quite as aggressive as, say, Arizona’s, which requires police to stop and possibly arrest anyone they even suspect of having entered the country illegally. (The legitimacy of this law will soon be considered by the Supreme Court.)
Other states have passed less controversial laws that force employers to use the federal E-verify system to confirm that their employees are in the country legally. Some have passed both.
Click to enlarge the image below to see a graphical representation of the states that have embraced these measures in the last couple of years, or go straight to the source for a more detail on the individual states:
Yes, it’s probably too early to pass certain judgment on the long-term effects of these laws. Most of the evidence is anecdotal.
Yet the anecdotal evidence that we do have seems to cut in the same direction – which is that employers have had immense difficulty filling these jobs with Americans or legal immigrants, and the effects are far more likely to be damaging than beneficial for local economies.
For one example of non-anectodal evidence we’ve come across, we read through a report released this summer by the Georgia Center for Agribusiness and Economic Development – and we suspect its conclusions apply more broadly to other states that are similarly punishing the employers of illegal immigrants and chasing away immigrants through other means.
The report estimates the economic losses to the state from the labour shortages reported by employers after Georgia passed its E-verify law.
Here’s the relevant excerpt, emphasis ours:
Lost fruit and vegetable production in the state in 2011 resulted in losses not only to the producers involved but to input suppliers to the producers as well as those who benefit indirectly from the expenditures that would have been created in the state and local communities. …
The so-called multiplier impacts occur as the money normally generated from the fruit and vegetable production was not spent on purchasing other goods and services.
For instance, not only would the input suppliers suffer lost sales, but so would all the retail businesses that would have had customers from those employed by the fruit and vegetable producers and suppliers.
The impacts reach further into the economy as the retailers then ultimately purchase less from others.
On and on the impacts spin through the state and region’s economy, resulting in fewer goods and services or state output produced in a multitude of industries and business not directly associated with fruit and vegetable production.
So there’s an effect well on down the supply chain, and because immigrants allow these employers to compete on cost, they prevent industries from shifting to other states or, in some cases, abroad. This is before getting into how the immigrants themselves can be a source of local demand.
“Dirty jobs”
The notion that there might exist labour shortages in a country with such a dramatic unemployment problem might seem a bit strange.
And the question of just why jobless American workers don’t fill the gap left by immigrants who get chased out is complicated, and one that also trespasses into the realms of psychology and sociology.
Bloomberg Businessweek’s Elizabeth Dwoskin gave it a thoughtful treatment in her big cover story last month (where the”dirty jobs” of the title comes from). And it’s a question we asked in a conversation with Peri, about which more below.
Thus far this is a local story, and the impact of a minority of states pursuing stricter laws might well be negligible to the national economy.
In many of these states, the immigrant population is both small and new. These immigrants will have little trouble moving to more hospitable neighboring states, who will benefit from their arrival to partially offset the loss in other states.
Yet we think it’s worth contemplating what the effect would be if an increasingly bigger part of the country began moving in this direction.
And since we also wanted to frame all of this in a larger context, we called Peri. His recent work seemed especially relevant to this discussion, as his papers have focused on differences in migration patterns between states, and how these differences affected economic outcomes.
We’ve already covered some of his conclusions, but we found the conversation illuminating and the lengthy transcript is posted below, with only light edits.
—————-
FT Alphaville: How much does your work on imperfect substitutability apply to what’s happening in states where immigrant crackdowns have led to labour shortages? I take it you’re not surprised.
Peri: I’m not surprised, because in some specific occupations – agricultural workers being one, but also in some construction workers and some personal services – the vast majority of those occupations have been filled by immigrants for the last ten years.
These are jobs at a wage level, and for a type of skill they require, are typical jobs for people with a very low level of education, of which there are very few US-born citizens in the US. And for the type of physical work they require, perhaps because of the stigma, perhaps because of other opportunities, very few American workers will take.
In states where you’ve had the fleeing of immigrants, either because of the the toughening of the laws, as in Alabama, you’ve seen that it’s incredibly hard to fill these jobs. So no, I’m not surprised.
I think a lot of people would say that this is all very strange given that the US has such a terrible unemployment problem, especially so many long-term unemployed. The idea that there aren’t people to take these jobs that are opening up would seem to defy common sense. Can you explain that?
In many states where immigrants have come, and they’ve tended to be those with low levels of education and bad English skills, and they’ve tended to take overwhelmingly these types of manual jobs.
And as they’ve kept coming, the companies have filled their manual needs with immigrants, and natives have moved towards jobs that are much more interactive.
These manual jobs are all over the spectrum – agricultural, construction, in personal services, household services, food preparation – in a lot of sectors, not just agricultural. Companies have been able to staff their manual positions with immigrants and expand in the direction of jobs.
So if you think of farms or construction companies, they certainly do need workers, but they also need supervisors, managers, people who take care of their sales, interact with their customers, engineers, headmasters.
So typically, companies that have hired immigrants have expanded their total employment more than just their immigrant jobs, and they have created jobs for Americans.
So what you’re seeing in the agricultural sector, for example, which is a small sector, what’s happened is that the few jobs which are American jobs in that sector are jobs linked to clerical and supervisory and sales roles.
Now, what happens is that if you don’t have the manual workers because the immigrants leave, these companies pack up and move, also eliminating the other jobs at the local level, or in many cases they outsource their line of production.
For agriculture there are only some goods you can do that for. So let’s say you grow tomatoes in California. If you can’t have a competitive price for workers there, you’ll have tomatoes imported from Mexico, for example.
And what’s happened is that you’ve destroyed jobs in the tomato industry, not just the immigrant jobs, but also the jobs of Americans.
So the complementarity, this is a topic I have researched a lot, is that by having the workers specialise, it’s hard to immediately substitute workers in a certain task.
And when you make it hard for those workers, you’re also endangering the companies and all the other jobs that are linked to them through these connections.
Given that some of these sectors have such a big group of immigrants working in them, I can see how these companies would get in trouble. So moving them away isn’t going to generate a lot of opportunities. On the contrary, it will generate a lot of losses.
So your last paper studied the impact of different migration patterns between states…
Well, what I studied is essentially the variation and differences in the presence of immigrants in different US states. So what I’ve tried to isolate is which states have received big inflows of immigrants, either because of their geographical position close to the border, or because there was a big presence of older communities of immigrants to attract new immigrants.
So this research is based on comparing states with many immigrants and states with very few immigrants.
In states with very few immigrants, you see fewer jobs for natives created as well, and also that the natives in these states tend to stay in more manual types of jobs. So Mississippi is a state with few immigrants, and Arizona has a lot. And in Arizona you see natives in more interactive types of jobs.
You see the opposite in Mississippi, where natives still do more manual types of jobs because companies have not expanded as much thanks to inflows of new workers.
Ironically and interestingly, Alabama is one of the states which has become toughest, and yet it’s one of the states with the fewest immigrants. There’s a very small group of them, and they’ll move easily out of the state because they’re also mostly recent migrants.
In a sense, Alabama should actually be trying to attract them given its economy, and yet they’re doing the opposite.
If you were thinking these laws were being pushed by a negative economic effect, you’d think they would be pushed by California and Texas, which have received huge inflows.
But to the contrary, these states have largely benefited from these inflows and are more pro-immigration.
So would you say that we can extrapolate from these local effects on states like Alabama and Georgia and assume that if they were applied at large to the rest of the country, something similar would happen to the whole national economy?
Yes, and I think you would especially see two effects.
One, you’d see a lot of companies leave the country altogether and you’d have a lot more offshoring. If you can’t produce competitively in this country, companies will simply use the labour force of other countries.
Both in agriculture and especially in manufacturing, they could easily relocate. If these jobs are lost, all the jobs connected to them would be lost.
Second is the question of what would happen to the unemployed immigrants. This is a harder question. We’d have to pay a massive amount for deportation, and there would be incredible legal issues because many immigrants have children here now and are very connected to the system.
One thing I’ve learned over time is that the boundary between documented and undocumented is less clear than we think. Many of the undocumented every year get a green card because they have relatives or spouses who are American. They are just undocumented for a while because the process is so slow.
I was looking at a statistic recently that of the 1 million green cards issued each year, about 200,000 of them go to people who previously were undocumented. [FTAV note: we haven't been able to very this.]
So people have a legitimate claim to being legal residents but the slowness of the process and the delays puts them in this position. So the issue is very complicated. But the idea that enforcement and moving people out is a good idea is a fallacy. It has big economic costs.
Your work seems to be a confirmation of the straightforward concept of gains specialization. But in the stories of what’s happening in Alabama, where native-born Americans won’t take these jobs, it’s not because they lack technical skill or lack training; these are manual jobs. So how does it apply here?
Yes, so the unemployed are people who are looking for a job but can’t find it. And many people would simply rather keep looking for a job that’s better paid or a more clerical type of job than take these manual jobs.
Keep in mind also that unemployment benefits have become more generous after the recession, so many of these people that’s another reason not to take jobs that pay barely more than unemployment benefits, and that are much harder than their previous jobs.
But one of the key points I think is this. Fifty years ago half the American labour force didn’t have a high school degree. Now only about 9 per cent don’t have a high school degree. So these manual, simple, low-paying jobs are jobs that only people with a very low level of education will accept. Others will keep looking.
The foreign-born don’t have this option to the same degree. For them it’s a very big improvement on their wages from what they would get in Mexico, for example, and so they’re willing to take these jobs.
This dynamic isn’t going to be changed. This isn’t entirely a specialization dynamic issue, you’re right, it’s more about outside options. Americans are more likely to conclude that they have outside options, so they don’t take these jobs.
And if you study the issue a little more carefully, and you look at the Americans who don’t have a high school degree, you’ll observe that many of them have some other issue. If you live in the US and don’t finish high school, you’re simply more likely to be the kind of person who has a hard time keeping a steady job. So even at the same wages, they’ll be less productive.
The complaint you hear from some of these employers is that some of these Americans will come for a couple of days, but they won’t work hard and then they’ll leave.
So there’s the issue of having outside options and there’s an issue that the selection of people who would be doing these jobs are relatively low-production people. Others, like those with a high school degree or more, will keep looking until they land a job with a better wage. So the solution to unemployment is actually to create more of those jobs.
In the stories about what’s happened in Alabama, one of the objections you’ll hear from the native-born workers is that they won’t accept the work because the wages and conditions are exploitative. They say the farmers have been getting away with it for years because they’ve had a pool of people that’s been willing to work for them for exploitative wages.
Well, but if you look at it first from the employer’s point of view, either you can hire a worker at a higher wage who will be much more productive, or you need to pay a wage that keeps you in the market.
When you are competing, in a sense, with companies that can pay much lower wages abroad, or if it’s a product that can be imported instead, if your cost of production is too high then you’re out of business.
Now, in many industries the way that Americans have kept their jobs is by increasing the productivity of the workers. So you can pay more workers in the US if they’re more productive, and you can afford to produce here if the workers are more productive.
But the economic tendency has been that productivity has increased more in jobs that are cognitive- and communication-based, involve the use of computers, and those kinds of jobs. For manual jobs, productivity hasn’t improved.
So yes, I’d say that the only way of keeping those jobs here is to keep the cost lower.
So you say, “But what if there were stronger unions or laws that forced farmers to pay more for the work?”, my prediction is that a lot of this work simply wouldn’t be here. It would be outsourced.
So I see from the partial point of view, you can save some jobs for a little while and have them pay well, but eventually the competition will drive the jobs away.
It’s beyond the scope of our conversation to get into how to reform all of US immigration policy, but specifically with respect to the situation in places like Alabama, what would be a good way to deal with this?
The best policy idea I’ve seen and would myself propose is to have a temporary visa program with the possibility after four or five years either for the immigrant to go back or to stay.
So my idea would be to have a guest worker program, but one that’s flexible to accommodate the demand that is here for these jobs. So right now it’s very rigid, there’s a cap. I’d like the employers to have more influence.
My idea is for everyone to come in on a temporary visa at first, I would say for four or five years. And my idea is to write the policy so that there are incentives for some of the immigrants to go back after the four or five year term is over.
One way to do this is to force part of the wages to be put away into savings that you’d give back to the immigrants if they agree to leave after five years. But that would also allow those who really wish to stay to build a long-term relationship with their employer.
Because generally I like the idea of a guest worker program, but the caveat is that then people are not willing to build long-term relationships with the employers, and employers will then be constantly rotating their employees.
So there should be a temporary visa at the beginning, but if a worker has been working hard and consistently and is willing to be sponsored in the long run, they should have access to permanent residency.
But there should also be a monetary incentive to go back so that out of, say, two hundred thousand people you’d get as temporary workers, maybe sixty or seventy thousand will end up staying, and the rest will go back.
Now, anything like this would have to implemented at the same time as fixing the problem of all the people who have already been here undocumented for a long time. So just solving this piece of the temporary problem won’t work if you don’t also deal with the people who have been here for a while, and deciding whether we should also admit them into the program. My feeling is that yes, they should be, maybe after paying taxes and fees. But they should be admitted.