While all foreign worker programs hurt American workers, some are more harmful than others. Now it is becoming clear that the Trump administration is threatening to close down the least harmful of these programs, while expanding the most harmful.
Before we get into the specifics of what the administration is doing with these programs, let me suggest a way to evaluate them.
There are five major foreign worker programs and at least as many minor ones. Let's confine our analysis to these five:
- H-1B, for alien college graduates (and a handful of fashion models);
- H-2A, for foreign farmworkers;
- H-2B, for unskilled non-farm workers;
- OPT, for foreign graduates of U.S universities in the form of one- to three-year, non-renewable visas; and
- H-4, for the spouses of (employer-favored) H-1Bs; usually college graduates.
The government, perhaps deliberately, does not count or even estimate the total number of workers in each of these programs; the best estimates are that the five range in size (workers active at some point in the year) from around 100,000 each in the H-2B and H-4 programs up to about 750,000 in H-1B.
All of these programs negatively impact American workers directly by taking jobs that citizens and green card holders would otherwise have, and all impact U.S. workers indirectly by expanding the labor force, thus loosening the labor market and discouraging upward pressure on wages. Big Business, of course, and its multitudinous lobbyists, dispute these impacts, and talk (often without much empirical support) about labor shortages. They never mention that employers would prefer the luxury of picking and choosing among a surplus labor force rather than scrambling to hire the workers they need.
So all foreign worker programs starts with these two black marks (displacement and labor force expansion) against them, but that is not the end of the story.
Three of the foreign worker programs also create an atmosphere of indenture; if you work for an employer through one of the first three H programs, your employer not only can fire you, it can threaten you with deportation. This is a hidden benefit to employers that is not generally discussed (not all of them use it, of course). No American worker is going to be forced to leave this country if she displeases her employer — and many employers prefer a "stationary" workforce, one in which the worker cannot leave the employer, or cannot do so easily, simply to get a better job elsewhere.
Two of the the five programs (H-2A and OPT) carry government subsidies, something that is not widely known. An employer of an OPT college grad, or an H-2A farmworker does not have to pay the usual payroll taxes, but if the same employer hired an American to do the same job at the same salary it would have to pay into the Social Security, Medicare, and Federal Unemployment trust funds, for total of a little over 8 percent of payroll. Understandably, some employers prefer to hire alien workers at a discount, rather than U.S. workers at the usual rates.
One of these five programs, H-4, neither indentures the worker, nor is subsidized.
This suggests that we can divide all five of the foreign worker programs into three groupings:
- The one that both allows indenture and is subsidized (H-2A);
- Those that are either subsidized (OPT) or allow the indenture atmosphere (H-1B and H-2B), but not both; and
- The one where neither a subsidy nor an indenture is involved, H-4.
A rational public policy would seem to suggest that the H-2A program has four black marks against it (bearing in mind the replacement and loosening of the labor market impacts); that the next three (H-1B, H-2B, and OPT) have three black marks each; and H-4 has only two. Such a policy might indicate that the program that most needs reduction and reform is the one for farmworkers, and the one that needs the least attention is H-4.
So how has the Trump administration reacted to these five programs? Just about exactly opposite relative to the amount of harm done as sketched above. Here's a summary of the Trump administration treatment of these programs, from the most harmful to the least harmful.
H-2A. There is no statutory ceiling on the number of these farmworkers admitted each year, and the Department of Labor certified 172,654 positions in 2016 and 242,762 in 2018 (the 2017 figure is not available), thus a major expansion of the program.
OPT. Here are the recent numbers for this population of alien graduates of U.S. universities: 2015, 160,623; 2016, 192,841; and 2017, 219,635. The administration, and I have been listening carefully, has not said a word about this program, which primarily benefits alien alumni and their employers. It is often used by major H-1B employers as a useful, middle step between alien college student and H-1B worker.
H-2B. The administration has twice allowed one-time additions of 15,000 unskilled, non-ag workers, adding to the 66,000 legislative ceiling. A government sleight-of-hand, allowing some previously admitted workers to come in beyond the ceilings, adds to the total numbers. This year it will allow an additional 30,000 such workers.
H-1B. These are mostly college graduates, mostly working in high-tech fields. The administration has permitted DHS adjudicators to ask more questions about these jobs, and there has been a rise in denial rates, so perhaps the growth of this program has been slowed. Employers, of course, act as if anything less than, say, a 98 percent approval rate, is an outrage. (For more on the harm done by this program, see this recent Backgrounder by my colleague Matt Sussis.)
H-4. These are typically college graduate spouses (primarily women) of senior, employer-favored H-1B workers. The latter usually have been in the United States for several years and have been selected by their employers for green card sponsorship, which will be delivered only many years hence. The administration is threatening to close down the program, but has not yet done so. We estimate, through a process described below, that there were about 103,000 H-4s working in the United States on September 30, 2018, and assume that this number has risen substantially since. (A handful of H-4s are married to H-2A and H-2B workers, but they are rare.)
H-4 migrants can work anywhere they choose, can change jobs without endangering their status, and get no subsidies. The employer class thus has no special interest in these workers, other than that they are additions to the labor force, something employers always like. The H-1B employers of their spouses (usually husbands) also support the program because it lowers the potential pressure from their H-1B worker-spouses for raises. But there is little lobbying by employers for this program. For more on this program, and how it reflects gender bias in the H-1B program, see this post by my colleague Dan Cadman.
The H-4 program is thus much less harmful than any of the other four, and it is the only one threatened by the administration for complete termination. It has one non-economic aspect that makes it more vulnerable than the others: It was created in the Obama years and, as we know from other governmental sectors, the current White House loves to undo Obama initiatives. The program is for a favored subset of H-1B spouses, for those whose spouses have been nominated for green card status by their employers; because of numerical ceilings, however, these cards are often delayed for 10 years or more.
Meanwhile, in another policy-related irony, the government of India (91 percent of the H-4s are from India) is quietly pressing for the preservation of the program — but is doing so despite the fact that spouses of foreign workers in India (there are a few) are not allowed to work. But who ever said governments always operate rationally?
The CIS Estimate of the Size of the H-4 Population
Since I could not find an estimate, much less a count, of the number of H-4 visa holders with work permits (known as EADs, or employment authorization documents) I decided that to try to devise one using government workload data. (Only the hardiest of readers will want to wade through what follows.)
Some of the important variables can be measured with accuracy, but other, smaller ones, such as the rate of return to the home country and the conversion to other nonimmigrant status, cannot. For example, an H-1B spouse could get an H-1B of her (or his) own or go to work for the World Bank, or the UN, or an embassy, or switch to a J-1 (exchange visitor) slot. Or she could go back to school on an F-1 visa, and work as an OPT. All of these options would seem to account for a few H-4s, but not for many of them.
While the input (the issuance of the EADs) is known, we must also make allowances for the subtractions, such as for the migration variations mentioned above and also for a few deaths or divorces, as well as the happy ending of H-4 status, a green card. EADs are issued for the lifetime of the associated visa, and since H-1Bs are routinely initially granted that status for three years and are renewed usually for another three years; one variable we must take into account is the ratio between new and renewed EADs, and that distinction is not recorded in routine USCIS statistics, but happily is available for one time period.
Additions. With that background, let's turn to numbers. There were, according to one USCIS tabulation, the following H-4/EAD documents, of all kinds, issued in fiscal years 2015 through 2018:
- FY 2015: 26,858 (part year)
- FY 2016: 41,576
- FY 2017: 46,671
- FY 2018: 53,891
- Total: 168,996
Note the steady rise, year after year, a pattern we often see when a new benefit is offered. Some of the 168,996 were, however, extensions of previous grants rather than new grants; or, in a few cases, the replacements of lost EAD documents. This dataset provides no information on this point.
Fortuitously, the Congressional Research Service's Jill H. Wilson secured this information from USCIS:
As of December 25, 2017, USCIS had approved 126,853 applications for employment authorization for H-4 visa holders. These count all approvals since May 2015 when the rule was implemented. This number includes 90,946 initial approvals, 35,219 renewals, and 688 replacements for lost cards.
This means that 28.3 percent of the issuances before Christmas Day 2017 were not initial ones. When we apply that percentage to the 168,996 above, it produces 121,140 (or rounded to 121,000) first-time approvals as of the end of FY 2018.
Subtractions. However, as noted earlier, some H-4s must have changed status through the various routes mentioned earlier, from a handful of deaths and conversions to other migration statuses to at least some movements to countries other than the United States. (Canada, from time to time, makes passes at U.S. H-1Bs, promising quicker permanent alien status.) Let's assume that this attrition comes to something like 2 percent a year, or 8 percent over the four years, to total of 9,680 (or a rounded 10,000) over the four years, leaving 111,000 H-4s in that status at the end of FY 2018.
In addition to those departures, there must be some H-4s who secured green cards in those four years, but how many? DHS presumably either has this number, or could obtain it, but I have never seen anything written on the subject.
H-1Bs and family members typically secure their green cards through segments of the second and third employment preference classes, each of which has a numerical ceiling of about 40,000 aliens. When country of origin ceilings also come into play, it means that no more than 7 percent of a class may be from a single country, so the maximum number for H-1B (together with H-4) conversions, for a single nation, would seem to be about 2,800 per preference. This is important because 91 percent of the H-4s are from a single country, India.
Let's now focus just on the H-4s from India; the ratio of workers (mostly once H-1Bs) in each of these categories to spouses and children (H-4s) is about even in both the second and third employment preferences, so it would seem that 1,400 H-4s would be eligible for second preference slots each year, as would 1,400 H-4s in the third preference. Let's assume that there are about 400 non-workers among them, leaving about 1,000 EAD-bearing H-4s to secure green card status in each of the second and third preferences.
This would mean, at most, 2,000 spouses departing from H-4 status each year, bringing the population of EAD-carrying H-4s down from 111,000 to 103,000, on September 30, 2018, give or take 10 percent.
With about 100,000 additional H-1B workers being added to that population every year, many of them married, and with only a few thousand H-4s leaving that status annually, the total number of working H-4s is bound to expand in the years to come unless the administration does, in fact, move to terminate the program.