Who is helped and who is hurt by immigration? This question seems like it should be relatively straightforward.
In a previous essay in Public Discourse, I noted that immigration does not have any significant net aggregate economic effects. But the absence of an aggregate effect does not mean there are no distributional effects. After all, taking $10,000 from every person who reads this essay and giving it to the author of this essay has no aggregate effect on wealth, yet I still think that is an admirable idea. (You may have a different opinion.) So, maybe the economic effects of immigration are similar redistributions of income or wealth.
When we look at other policy issues, we know that distributional consequences can be large. Consider international trade and tariffs. It is well established that tariffs are harmful to the economy imposing them; levying tariffs makes your country poorer, because you are losing the benefits of international trade possibilities. However, not everyone is worse off. A tariff on automobiles does make domestic automobile manufacturers better off. While the gains to the automobile manufacturers are smaller than the losses to society, tariffs do benefit some at the expense of others.
The minimum wage also has distributional effects. When you raise the minimum wage, some people get a larger paycheck. Other people will lose their jobs when the employer decides not to hire as many people at the higher wage rate. Those people are worse off. The price of products will rise, so people who buy the product will be worse off. There are also longer run consequences. Faced with the higher wages, some employers will replace the now more expensive workers with machines, which makes the workers who lose their jobs worse off, but the workers who make the machines better off. And so on.
How does this work when we turn to the issue of immigration? As the number of immigrants in a geographical area or an industry rises, what happens to the indigenous population? Are they better or worse off? Does an increase in immigration cause wages to rise or fall? What are the effects on state or local government budgets? Do tax revenues rise faster or slower than government expenditures? Not surprisingly, there has been much research on these economic effects of immigration.
A Thought Experiment
Pretend you are an economist who wants to use data to answer these questions. You cannot answer them all at once, so start with a simple question. What is the effect of immigration on wages? Does an increase in immigration cause a rise or fall in the wages paid to the indigenous population? Obviously, this is a really important question. If, for example, a rise in immigration makes indigenous workers worse off, then it is perfectly understandable that they will resist further immigration.
This is a straightforward question that seems easy to answer. With a large dataset, we can compare changes in wages to changes in immigration levels. But first, we have an immediate problem to solve. How should we measure the volume of immigration into an area or an industry? There are two choices. You can either use:
1.) the change in the number of immigrants divided by the size of the original workforce, or
2.) the number of immigrants divided by the current workforce.
Those two numbers are obviously related, but they are slightly different. So, which one is the right way to measure this? Before reading on, decide which of those options you think is best.
Now that you have picked your answer, how certain are you that you picked the right way to measure it?
Ideally, it would not matter which of those two numbers you picked. Ideally, you would get the same answer either way. Unfortunately, that is not the case.
There is a very large debate in economics about the effect of immigration on wages, partly because measuring the size of immigration in those two ways generates completely different answers. One decision leads to results showing very large negative effects on wages; the other shows small positive effects on wages. Both of these results have been established by very capable economists who have been studying this matter for years. There is no consensus on the right way to measure immigration.
Next question. Be honest with yourself: if you knew which result went with which method of measuring immigration, would it be easier for you to pick the right way to measure immigration?
Another question: if you found out that the means of measuring immigration you chose above led to the result on immigration opposite to the one you currently believe to be true, would you change your mind about the desirability of immigration?
Why Is It So Hard to Determine the Distributional Effects of Immigration?
The fact that you almost certainly realized you would not change your mind about immigration based on the results of a study using a particular means of measuring immigration is not as problematic as it might seem. Should you change your views on immigration based on a single econometric study? Most certainly not.
The above thought experiment is a means to demonstrate that, when it comes to studying the distributional effects of immigration, the story is enormously complicated. Indeed, it is so complicated that it becomes incredibly difficult even to sort through all the economic implications.
Consider this scenario. A large number of immigrants who all happen to be carpenters move into an area. Suddenly, the supply of carpenters has gone up, so the wage rates for carpenters go down. This is bad for the indigenous workers. But, the price of hiring a carpenter also goes down, so more people hire carpenters to do work. As a result, there are more home improvement projects, so the demand for general contractors and electricians and plumbers also goes up, which drives up the wages of those other workers. Some of the carpenters who used to live in the area, seeing a fall in relative wages, either switch jobs or move to other parts of the country. A few of them expand their businesses by hiring the new immigrant carpenters. The youngest workers may decide that instead of being a carpenter, it is now worthwhile to invest their time and money in becoming an electrician or a plumber.
We could keep going like this for quite some time trying to measure all the ripple effects of an increase in immigration in one particular area in one particular industry. But it does not take long to see that it is difficult to know how exactly to measure the economic impact of the increase in the number of immigrant carpenters.
Did the story above show that the immigration helped or hurt the indigenous workforce? Even if we restrict the attention to the indigenous carpenters, are they on balance better or worse off? This is exactly what the econometric studies are trying to sort out, and it is incredibly difficult to get a robust set of results. And remember, the story above assumed the effect of the immigrant carpenters was a fall in the wages for carpenters, but it is not clear whether that is an accurate assumption to make.
Effect on Government Budgets
So maybe we should look at another way to think about the distributional effects of higher levels of immigration. Maybe we can look at state or local government budgets to see if the immigrants are imposing a net benefit or cost to the indigenous citizens of the place into which the immigrants are moving. After all, immigrants do put demands on government services. They also pay taxes. Surely we can figure out the net effect.
Unfortunately, once again, there is no good way to determine this effect. In fact, it is so hard to figure out how to study these effects that the number of studies even trying to do so is remarkably limited. The results from these studies are all over the place. Suffice it to say, if you know the conclusion you want to find, there is a study that shows that.
Why is this so complicated? Once again, there are endless numbers of spillover effects. Think about elementary education, which seems about as simple a case as could be imagined. Does an increase of immigration into a city cause the amount of spending on education to rise? Not necessarily. First, and most obviously, higher levels of immigration only affect elementary education expenses if the immigrants have children of elementary school age. Second, these immigrants must be moving into a home somewhere; if it is not a newly constructed home, who was living in the home before, and where did that household move? If an immigrant household without elementary schoolchildren moves into a home formerly occupied by a household with elementary school aged children, then the higher level of immigration reduced the number of children in the school.
This displacement effect is just the start of the story, however. Unless all the classes are at capacity, adding one more child to the school causes a very small increase in costs. But if enough new children show up, then the school will need to bear the extra cost of adding new classrooms. In other words, school expenditures are lumpy; not all immigrants cause the same increase in costs.
Now add in future effects on educational expenditures. Children grow up. Some go to public universities. Do we include the cost of educating the students in college? Eventually the immigrant children have children of their own. Do we include the costs of educating the next generation too? (This sort of long-run effect is particularly important when you start thinking about the effect on tax revenues.)
Any attempt to discover the effect of immigration on government budgets is thus highly susceptible to the decisions on how to measure all those benefits and costs and how to account for the effects on the indigenous population. The quick answer to “What is the effect of immigration on government budgets?” is “It depends on how you measure it.”
The Economics of Immigration
Is economics of any value in helping us understand the effects of immigration? Absolutely. It is only because of the careful work of economists that we are able to conclude that the economic impacts of immigration should not be the determining arguments for whether immigration levels should increase or decrease.
There is a problem with the attempt to lean too heavily on economic studies to bolster the case for raising or lowering immigration levels. As much as it pains me to admit this, economists are not immune to the temptations facing the rest of humanity. If you want to argue in favor of a particular policy, then it is possible to find an empirical argument to support your policy preference.
The advantage of economics as a discipline is that if you try to make a claim that cannot be supported by the data, then there is an army of economists who will happily show that your results are misleading or incorrect. But what happens when the data are not so clear? What happens when the problem is so intricate and complicated that perfectly defensible modeling choices generate diametrically opposite results? Within the guild of economics, this simply spurs on more research. Unfortunately, far too often economists succumb to the temptation to pontificate on policy issues even when there is no consensus within the profession.
There may indeed be very large distributional effects from immigration. There may be no distributional effects. But a fair look at the research on the matter should lead us to pause before we assert that these effects do or do not exist.