The National Academy of Sciences (NAS) has just published a major report, The Economic and Fiscal Consequences of Immigration. The ink was barely dry on the virtual page before wildly different interpretations of what the report actually said appeared in the media. The New York Times, for example, got an early peak at the study and quickly produced an article headlined “Immigrants Aren’t Taking Americans’ Jobs, New Study Finds.” The Washington Times, after looking at the same report, concluded that “immigration drains the government, sapping as much as $296 billion a year from federal, state and local taxpayers while depressing wages, at least in the short run, according to an authoritative study.”
The report, about 300,000 words on nearly 500 single-spaced pages, was prepared by a panel of about 20 social scientists, including economists, sociologists, and demographers. (Disclosure: I was a member of the NAS panel that prepared the report; this article reflects my take on what is in the report.) Much of is it written in technical-ese, comprehensible only to trained economists and likely to appeal only to immigration geeks fascinated by the minutiae of technical assumptions. It is inevitable that there will be disagreements on what it is we actually learn from all that dense verbiage, making it important not to rely on what a press release says but rather to highlight the crucial points by actually quoting the report and by linking to the specific tables that support that conclusion.
The report provides new and important information about the assimilation prospects of new immigrants. It is well known that immigrants have an economic disadvantage when they first enter the country. Many are not fluent in English; they are not familiar with how the U.S. labor market works; and on and on. Over time, the immigrants learn the language, acquire new skills, and begin to catch up, or assimilate, to the native norm.
It turns out, however, that the newer immigrants are not assimilating quite as rapidly as earlier waves. Here’s the table from the report that documents the slowdown of assimilation:
This table reports the percent difference in wages between specific immigrant waves and natives at different points in time. The data obviously show a lot of wage growth for immigrants who arrived in the 1960s and 1970s, but far less for immigrants who arrived in the 1980s and 1990s. For example, the 1965–69 arrivals had a 23.5 percent wage disadvantage at the time of arrival, and this narrowed to a 12.0 percent disadvantage after ten years. But the 1995–99 arrivals had a 27.3 percent wage disadvantage at the time of arrival, and it was still 26.9 percent after ten years. As the report modestly puts it, “the process of economic integration appears to have slowed somewhat in recent decades.”
The report’s documentation of a slowdown in assimilation brings to the immigration debate a significant new wrinkle. Even though we do not yet fully understand why this slowdown occurred, and the report makes little effort to determine the source of the slowdown, a crucial question lurks underneath: What does it mean for the future?
The report’s key conclusion on the labor market impact are stated clearly in the summary:
When measured over a period of ten years or more, the impact of immigration on the wages of natives overall is very small. However, estimates for subgroups span a comparatively wider range. . . . To the extent that negative wage effects are found, prior immigrants — who are often the closest substitutes for new immigrants — are most likely to experience them, followed by native-born high-school dropouts, who share job qualifications similar to the large share of low-skilled workers among immigrants to the United States.
Two distinct findings in that quote are worth emphasizing. The first is that low-skill workers who have qualifications similar to those of low-skill immigrants have lost out. Table 5-2, which summarizes the relevant evidence, is reproduced below so that readers can see (without contamination by any PR middlemen) that there are many studies detecting this negative wage impact, particularly on native high-school dropouts, many of whom are black or Hispanic.
The second finding — that “over a period of 10 years or more, the impact of immigration on the wages of natives overall is very small” — is based on conclusions from a separate strand of studies (reported in Table 5-1 of the report). These studies essentially “stream” the 42 million immigrants through a mathematical model of the American economy to predict what would happen in the long run, after the economy fully adjusted. And indeed those simulations indicate that the long-run impact of immigration on the average wage is exactly equal to 0.0 percent.
But there is a huge caveat that the report repeatedly emphasizes and that I know will be left out of much of the media discussion. The zero average wage effect in the long run is built in by the mathematics of the model that purportedly describes the American economy. As the report puts it: “In the case of structural studies, when capital is assumed to be perfectly flexible, [average] wage effects on natives are zero, although this result is built in by theoretical assumptions.” Put bluntly, the claim that the long-run effect of immigration on the average wage is “very small” has nothing to do with the data. That claim is instead a by-product of a mathematical assumption.
What the data clearly show is that wages drop for those natives who have skills that are most comparable to the skills of the immigrants.
The Fiscal Impact
The NAS panel calculated the short-run fiscal impact by comparing the cost of providing public services to immigrants with the taxes that those immigrants pay in a particular year. The report unambiguously concludes that, on a year-to-year basis, immigrants and their dependent children create a fiscal burden. This shortened version of Table 8-2 easily shows the magnitude of the burden:
There is a sizable fiscal burden regardless of which scenario is assumed. Scenario 1 assumes that the cost of providing public goods (such as national defense) is the same for an immigrant as it is for a native, while scenario 5 assumes that immigrants do not increase the cost of public goods at all. (The original table has eight scenarios, and all of them generate equally large fiscal burdens).
This is how the report describes the fiscal impact under scenario 1: “The total fiscal burden is $279 billion for the first generation group (average outlays of $15,908 minus average receipts of $10,887, multiplied by 55.5 million individuals).” It is equally easy to estimate the fiscal burden in scenario 5, where immigrants are assumed not to increase the cost of public goods at all. The average outlay is then $11,669, and tax receipts remain at $10,887, creating a fiscal burden of $43.4 billion.
The NAS panel tries to put these statistics in context by noting that we are now running a fiscal deficit exceeding over a trillion dollars a year, so that everyone is a fiscal burden. Left unsaid is an equally important point: Some burdens are avoidable, and some are not. We may not be able to do much about the fiscal burden of the native-born population. But there are many obvious policy options available to ensure that the already-large burden is not further increased by immigration.
The NAS report also provides additional estimates of the short-run fiscal impact, focusing on expenditures and taxes at the state and local-government level. Here is a short version of Table 9-6, reporting the data for the largest states and for the country as a whole:
Again, the NAS find that immigrants generate a fiscal burden in the short run. The report states: “First generation independent person units (which include first and second generation children assigned to independent first generation persons) cost the states on net about $1,600 each. . . . These estimates of the fiscal impact imply that the total annual aggregate impact of the first generation and their dependents, averaged across 2011–13, is a cost of $57.4 billion.”
The data are so unambiguous that it is very easy to summarize what we learn. On a year-to-year basis, the taxes that immigrants pay do not cover the public expenditures they trigger. And the shortfall seems to be at least $50 billion annually.
The NAS also tried to calculate the long-run fiscal impact, taking into account the taxes and expenditures of immigrants and their descendants over a 75-year period. This long-run calculation allows for the possibility that immigrants might help fiscally, as the native population is aging and there is not enough money to fund the liabilities in Social Security and Medicare unless we drastically raise taxes or cut benefits. Immigration brings in new taxpayers who can help fund some of those liabilities in the future.
As the report repeatedly notes, however, the bottom line of the long-run calculation can be manipulated to be either a very positive number or a very negative number by making different assumptions. The long-run fiscal impact of the average immigrant is positive only if immigrants do not affect the cost of public goods and we assume that future tax rates and benefit payments will follow the projections made by the (obviously infallible) Congressional Budget Office. If you get rid of either of those assumptions, Table 8-12, which I reproduce in a simplified form to make it more easily readable, shows that the positive long-term impact of an immigrant (and descendants) contributing a net of +$58,000 over the next 75 years becomes a loss as large as -$119,000.
The report repeatedly emphasizes the fragility of the long-run evidence: “Forward-looking projections of the net fiscal impact of an additional immigrant and descendants generate a relatively wide range of possible results.” And: “The future path of fiscal policy is important for assessing the fiscal impacts of immigrants.” And: “The treatment of spending on public goods is important for assessing the fiscal impact of immigrants.”
Let me translate. Assumptions matter, and different assumptions lead to wildly different answers. I think there is an elephant in the room; the NAS report alludes to it but cannot bring itself to say it out loud. So I will: All estimates of the long-run fiscal impact are useless!
The bottom line
The NAS report does not conduct the final calculation that adds up the economic gains and compares that number with the fiscal burden. But anyone with a pencil and the proverbial back of an envelope can do so using the numbers in the report. Immigrants have both a labor-market impact and a fiscal impact. Do the economic gains generated by working immigrants outweigh the fiscal burden that immigrants impose?
The report concludes that there is a short-run fiscal burden. Across all levels of governments, the annual burden ranges from a minimum of $43 billion to nearly $300 billion, depending on what is assumed.
But what does the report say about the “immigration surplus,” the increase in wealth accruing to the native population as a result of immigration? As immigrants enter the labor market and reduce the wage of natives, they increase profits for the employers. Plus the immigrants themselves produce additional output, generating even more profits. In the end, the aggregate wealth of natives — both workers and firms — rises, and there is a redistribution of wealth from workers to firms. The NAS report their estimate of the immigration surplus in chapter 4:
Immigrant labor accounts for 16.5 percent of the total number of hours worked in the United States, which . . . implies that the current stock of immigrants lowered wages by 5.2 percent and generated an immigration surplus of $54.2 billion, representing a 0.31 percent overall increase in income that accrues to the native population.
That short paragraph contains a lot of important information. First, the immigration surplus is relatively small, about $54 billion. Unfortunately, the report avoids giving a transparent estimate of the size of the wealth transfer from workers to firms, reporting instead that, on average, wages went down by 5.2 percent. It would be better if they had reported the actual number of dollars involved in that transfer. That number, it turns out, would be about $500 billion. So, yes, immigrants created an additional $54 billion worth of new wealth, but a byproduct of that creation was a wealth transfer of half a trillion dollars.
I would add a huge caveat to the $54 billion estimate of the surplus. It ignores all the externalities that immigrants create along the way. The externalities are both good and bad. The good: The entry of extremely high-skill immigrants surely accelerates innovation, makes us more productive, and has a beneficial impact on economic growth. The bad: The entry of some high-skill immigrants, such as those who enrolled in flight schools and learned to fly planes and then flew them on September 11, 2001, can make us all much worse off. The NAS did not even try to quantify the value of all the many positive and negative externalities (and, in fact, neither has anybody else). So, in the end, all we really have to go on is an estimated surplus of $54 billion in the short run.
If we then take the report’s estimates of the surplus and the fiscal burden at face value, it is self-evident that the impact of immigration on the aggregate wealth of natives is, at best, a wash. Instead, the impact of immigration is distributional. Those who compete with immigrants are effectively sending billions and billions of dollars annually to those who use immigrants.
There is a lot of temptation, particularly in the middle of a presidential campaign in which immigration is one of the core issues that completely differentiate the two candidates, to spin aspects of the NAS report in ways that will further a particular narrative. There are enough scenarios in the NAS report that would enable the construction of practically any factoid. As always, it is crucial to open that hood and look inside before you buy into it.