How Employers Cheat America’s Aging By Hiring Foreign Workers

By David North on June 24, 2012

Download a pdf of this Memorandum

David North is a CIS fellow who has studied the interaction of immigration and U.S. labor markets for more than 30 years. He thanks Jerry Kammer and Karen Jensenius, both CIS colleagues, and J.D. North for their comments on previous versions of this report.

One of the many reasons why the Social Security, Medicare, and Federal Unemployment Tax Act (FUTA) trust funds are in trouble is that American employers can avoid paying taxes to these funds by hiring certain classes of aliens, who, in turn, also do not pay these taxes.

The payroll tax loss for such a full-time worker at $30,000 a year was about $4,230 a year in 2010. That's a $2,535 bonus to the employer for hiring an alien rather than a U.S. citizen or a green card carrier and a $1,695 bonus to the alien worker, and thus a $4,230-a-year blow to the three trust funds collectively. At $60,000 a year, the damage to these funds is twice as great, or $8,460 a year.

There are over half a million of these pampered alien workers. A careless government, of course, does not keep track of these losses in payroll taxes. Why generate data on a subject you would rather ignore? The damage to the trust funds is remarkable, it is continuing and, sadly, the advocates of the elderly, such as AARP, are doing nothing about this problem.1

The difficulty is that those who benefit from these tax breaks, corporations like Disney, are much more powerful than the retirees who are hurt by them. As too often is the case in America, the narrow, intense lobbies routinely beat the generalized, but low-key interests of the rest of us. Table 1 shows our conservative, rounded, estimates of the trust fund losses.



A Word About the Trust Funds

These trust funds are used to finance the retirement of those too old or too disabled to work, to support healthcare for the aging and the disabled, and to fund unemployment benefits for those of all ages who are temporarily out of work. In another society these worthy activities would be supported, at least in part, by general tax funds, but because of the political power of the well-to-do in the United States the first two of these programs are funded by workers and their employers only. The third is funded by employers only.

If you live on your capital gains and your dividends, you will not pay a nickel into these funds, and if you manage the unlikely feat of doing that throughout your whole life you will not be eligible for any of these benefits, but then you will probably not need these programs anyway.

The surpluses in the larger two of these funds, Social Security and Medicare, have been slowly dropping as the population has aged, as the economy has sagged, and as more and more people become beneficiaries. The federal unemployment fund, the smallest of the three, has had similar problems. (There are also state unemployment insurance funds that are largely outside the scope of this paper.)

The Hill reported on April 23, 2012,2 for example, that the current surpluses in the Medicare and Social Security trust funds will run out by 2024 and 2033, respectively, that these funds will be dependent totally on newly arrived receipts thereafter to meet their various obligations, and that after those dates receipts are expected to be insufficient to cover the funds' payouts unless Congress acts, as it probably will have to, by either increasing the FICA and/or Medicare taxes, reducing benefits, postponing the retirement age, or some combination of those actions.

One of the admittedly lesser problems of these funds, and one rarely, if ever, discussed in public, is that there are small bands of alien workers — and their employers — who do not have to contribute to these funds.

To better understand the importance of these losses, and the mechanisms that produce them, the Center for Immigration Studies is providing what we believe to be the first published estimation of these revenue shortfalls, program by program.

Our calculations of the annual losses to the trust funds are based on the rates for the year 2010, when the whole package of payroll taxes was set at 14.1 percent of payroll, based on 10.4 percent FICA + 2.9 percent Medicare + 0.8 percent FUTA (Federal Unemployment Tax Act). This is a total reduction for the employer of at least 8.45 percent, and 5.65 percent for the alien worker.

The three flows of payroll taxes are these:

FICA. This is the Social Security tax; the initials stand for the Federal Insurance Contributions Act, and it dates back to the New Deal legislation of the 1930s. In normal times, but not at the present, the tax rate is set at 6.2 percent of wages paid by both workers and employers. Currently (this is being written in June 2012) the tax rate on workers has been reduced by two percentage points to help fight the recession. Thus the trust fund loss, at the moment, is 6.2 percent plus 4.2 percent, or 10.4 percent of wages when the selected groups of nonimmigrant workers are employed.3

FICA, however, is not levied on all wages, as Medicare taxes are. It is collected on the first $110,100 in income during 2012, so high-salaried people pay a lower percentage of their incomes to this tax than lower-income people do. (The upper limit is adjusted annually for inflation.) However, very few of the nonimmigrant workers we are discussing would be affected by this upper limit, were they to be covered by FICA.

Medicare. This tax is currently set at 1.45 percent of all wages for workers and 1.45 percent of all wages for their employers,4 for a total of 2.9 percent. The formula for this tax is much less complex than it is for the other two.

FUTA. The least, and most variable, of these tax breaks relates to FUTA, collected under the Federal Unemployment Tax Act.5 Employers, but not workers, pay both FUTA and state unemployment insurance taxes.6 Further, employers of selected nonimmigrant workers in some states are excused from paying state unemployment taxes on their wages.

The FUTA tax varied from 0.8 percent of payroll up in 2010 depending on a variety of factors, but for our purposes we are using 0.8 percent as the discount obtained by the employers of nonimmigrant workers and have ignored the state discounts. FUTA taxes apply to the first $7,000 of each worker's wages and the rate had fallen to 0.6 percent by 2012.

It might seem that it would be in the public interest to fend off raids on these three funds, but from time to time, and using different techniques, various powerful employer groups have managed to be excused from paying payroll taxes — placing a larger burden on other employers, on other workers, and, of course on those using Social Security, Medicare, and unemployment insurance benefits.

Sadly both the Obama and the Bush II administrations have taken administrative actions that have actually expanded the damage done to the trust funds by these programs in recent years.

The Nonimmigrant Worker Programs

Initially it should be pointed out that many nonimmigrant workers, such as those in the H-1B, H-2B, and L-1 categories, do pay the full load of payroll taxes. So do aliens who have been in the country for five full years, but this is distinctly a minority population among the university-related alien student population.7

The six subclasses of nonimmigrants discussed in this report are exceptions to the general rule that workers in the United States and their employers all contribute to the three trust funds of interest.8

One thing is true about all of these programs: The rationale for excluding them from FICA, Medicare, and FUTA taxes is always ostensibly based on the interests of the workers, and rarely mentions the benefits to their employers. But both sides benefit, and usually benefit about equally.

The general notion for all six programs is that the members of those classes, all aliens, will not stay in America long enough to profit from any of the benefits linked to the three trust funds, and thus they should not have to pay into these funds.

I can see no rationale for the savings to the employers, nor could I find any argument for these employer savings beyond the superficial one that if the employee does not have to pay, why should the employer?

The better approach to the question of payroll taxes is that moneys are needed to fund these programs, that Congress has decided that work alone – and not other kinds of income – is to fund them, so we cannot afford to exclude any group of workers, or their employers, from contributing.

That reasonable approach, obviously, was not adopted.

One might also argue that if the rationale for non-payments to the trust funds relates to the workers, they should not pay into them, but employers should. That logical approach, too, has been ignored.

Further, and perhaps not recognized in the early days of these programs, these discounts create a perverse incentive; namely, employers looking at an F-1 worker for instance, and his or her American peers, will find that by hiring the nonimmigrant they can save at least 8.45 percent of the total payroll because of the inherent tax breaks. Thus the employer is encouraged — nay, subsidized — by the government not to hire the citizen or the green card worker.

This is an odd arrangement — largely unknown to the public — that hurts some young American workers and just about all older legal residents of the country; and it helps only a privileged segment of America's employers. My sense is this pattern persists largely because it is so obscure and because the concentrated, well-lobbied interest of the employers is much stronger than the small-scale, general interest of the rest of us.

Let's look at the six specific nonimmigrant worker programs where these discounts are available. We examine the general nature of each of the programs, the structural arrangements, the kinds of work performed by their participants, and a rough measure of the populations involved. We are usually using numbers for the year 2010 because this is the last year for which we can obtain even roughly comparable data or estimates on all six groups.

1. Summer Work Travel (SWT) Program (J-1)

This is nominally a cultural exchange program, designed to bring foreign youngsters, usually college students, to the United States, at their expense, to work and travel. It is supposed to encourage a rosy view of the United States in the minds of future foreign leaders.

The program is operated by the U.S. State Department, and it is one of the subsets of the J-1 Exchange program. It licenses certain middleman agencies to recruit the aliens and to place them in jobs in the United States. The admissions are for four-month periods. What little program supervision that exists is provided by the intermediary agencies. Neither the Departments of Homeland Security nor of Labor — both of which have extensive field operations in the United States — are involved in the program.

SWT, however, had a great deal of (thoroughly deserved in my view) negative attention during 2011 as a result of a strike by a group of alien summer workers at a factory controlled by the Hershey Corporation;9 the workers, all enrolled in the SWT program, complained of low wages, long hours, difficult working conditions, and other indications of exploitation.

The SWT program forces enrolled participants to pay for their own round-trip airfare, steep fees for the middleman agencies, and often-expensive housing and eating arrangements, but rarely pays much more than the minimum wage. The program also shoulders aside young American workers who might otherwise have these summertime jobs, often at resorts.

These are clearly jobs that U.S. residents can and should do, but the State Department's program gives employers a 8.45 percent discount to hire foreign workers, rather than American ones, through the peculiar provision excusing employers (and workers) from paying the normal payroll taxes.

The grim facts about the program were described in unsettling detail by my CIS colleague, Jerry Kammer, a skilled investigative reporter, in a highly useful CIS Backgrounder10 and in a video program entitled "The $100 Million Work Travel Industry".11

This is primarily a movement of young, single people and if there are any visas issued to their kin, they must be tiny in number. If there were any, they would be J-2s and permitted to work, but would be covered by FICA and the other payroll taxes.

Estimated Number of SWT Participants, 2010: 132,000. These were the admissions recorded by DHS that year; each was here for about four months.

2. Other Exchange Visas (J-1)

This is a mix of several sub-programs of varying lengths, ranging from the very useful, such as the exchanges of scholars,12 to marginal, such as the au pair program. In some cases all the participants are working, as with the au pairs and the teachers, while in others, such as the secondary school students, there is probably very little work for pay involved. To further complicate things, many are working for U.S. universities where FICA and the other trust funds do not apply to any of the students because of the higher education exception.

For more on the J-1 program, and its impacts on U.S. labor markets, see Daniel Costa's monograph "Guestworker Diplomacy: J Visas Receive Minimal Oversight Despite Significant Implications for the U.S. Labor Market", published last year by the Economic Policy Institute.13

These J-1 programs, including SWT, are run by the State Department alone, with minimal input from other departments.

J-2s are spouses or children of J-1s, and while they may work with the assent of DHS, they are covered by the usual payroll tax programs.

Estimated Other J-1 Participants, 2010: 126,000. This estimate is based on a State Department webpage estimate, probably relating to the year 2010, that reports in an average month there are 170,000 J-1 participants in the nation.14 If you back out the 132,000 SWTs, who are here for four months each, the estimated average for the others becomes 126,000. Some of these benefit from the non-payment of payroll taxes, and some do not.

3. Foreign Students (F-1, M-1) Working Off Campus

Most, but not all, foreign students attending U.S. institutions of higher learning have one of these two visas. The F-1 visa is for academic students, and is much more widely used than the M-1 visa for vocational students. (Other non-citizen, foreign-born students include green card holders and J-1s.) The F-1 and M-1 visas are issued by the State Department after the applicants have secured a form issued by an American educational institution, the I-20, which indicates admission to that institution. There are no case-by-case screenings of the I-20 documents by any U.S. agency, but many would-be foreign students are, in fact, interviewed overseas prior to the issuance of the visa.

The heading above speaks of off-campus work for a reason: All university and college students, of whatever civil status, pay no FICA (or other trust fund) taxes for their on-campus work. So the lucky grad student whose department may be paying $20,000 a year for his or her work as, say, a research assistant, is treated the same by the payroll tax system whether native- or foreign-born.

Off-campus work is lightly regulated by the universities who have been given authority to manage it by the Department of Homeland Security. Generally, during the school year this is limited to 20 hours a week, and can be full-time during the summer. This is all spelled out through a DHS program called Curricular Practical Training (CPT), which is not to be confused with the somewhat similar Optional Practical Training (OPT), which only applies to F-1 and (a few) M-1 students after graduation.

Working off campus, the foreign students and their employers pay no payroll taxes,15 but their resident (U.S. citizen or green card) peers holding the same kinds of jobs and their employers do pay these taxes.

While I could find no data on the extent of off-campus employment, there is, if anything, too much data (of varying degrees of utility) on the larger question of the size of the alien student population. The process by which I estimated the sizes of the visa-bearing student population is explained in the Appendix.16 The portion working off campus is estimated in the "Trust Fund Losses" section below.

F-2 visa holders, i.e., relatives of F-1s are not allowed to work, and the same is true of M-2s.

Estimates of Combined F-1 and M-1 Populations in 2010: 583,000. Using the methodology spelled out in the Appendix, there appeared to be something like 568,000 F-1 students and 15,000 M-1 students in the 2010-2011 academic year, and this population has probably grown since then.

4. Recent Alien College Graduates (Largely F-1s) Working on OPT Permits

The OPT program allows F-1 (and a few M-1) students who have secured a degree and are no longer in college to work for a year in a job that relates to their education. During these OPT months the student is still in F-1 or M-1 status, and neither he nor his employer pay payroll taxes.

In 2008, the Bush administration decided that for F-1 graduates in a lengthy list of STEM (science, technology, engineering, and math) occupations this post-graduate period of OPT could be extended by an additional 17 months, for a total of 29 months.17

This decision was apparently made in deference to pressure from high-tech employers who wanted these workers at a discount.18

At the time USCIS estimated that there were 70,000 OPT workers. As of the end of FY 2010, DHS reported 94,465 OPTs, an increase of 35 percent over the 2008 figure, and said that 10,022 of the latter had signed up for the extra 17 months.19

This count of OPTs can only be regarded as an undercount, given the linked facts that universities might issue the 17-month document and not report it to DHS and DHS's casual supervision of the universities. Nevertheless we accept those numbers as a basis for estimating the number of OPTs on corporate payrolls at the end of FY 2012, which comes a few months from now.

Subsequently the Obama administration, in 2011 and again in 2012, decided to expand the list of occupations eligible for the OPT extension, but did not put a number on the size of this new population. Meanwhile more and more F-1 students, and more and more high-tech employers, learned about the program and its accompanying tax breaks. These employers started running ads saying that they were hiring OPTs. My sense is that the program, as a result of both of those developments, grew rapidly.

To project the number of current (end of FY 2012) people on extended OPT, I did a little back-of-the-envelope calculation. If in the first half of FY 2008 there was no one on the program, and if at the end of FY 2010, there were 10,022 of them, what was the rate of growth?

I decided to work that out, assuming that there were 1,000 OPTs hired on an extended basis in the second half of FY 2008, and that each OPT stayed in that status for 24 months. What I found was that in a straight-line projection, based on an increase of 40 percent every six months, by the end of FY 2010 there would be 10,100 of them, close to the 10,022 actual number.

Now, if you take the increase in the total number of OPTs from the 35 percent every two years noted above and project that for extended ones at 40 percent every six months, you get, in rounded numbers, a total of 128,000 OPTs and 39,100 in extended status. I decided that the last number was a bit high and reduced it to an estimate of 30,000 for those in extended status, leaving 98,000 in the usual one-year program.20

Typically there is a trial work period under OPT. Employers who are thinking about hiring recent alien graduates under the more expensive21 H-1B program have a chance to examine the potential H-1B workers and vice-versa. In my volunteer income tax assistance work with graduate students I have also heard, more than once, unhappy descriptions of some employers who "will use up your OPT eligibility and never offer you an H-1B".

The handful of M-1 (vocational workers) presumably using the program are included in these estimates.

Estimates of Regular OPT and Extended OPT Populations: 98,000 and 30,000.

5. Cultural Workers (Q-1)

This is a strange little visa, largely designed for, and pushed through Congress by, the Walt Disney Co., to help staff its theme parks. Fortunately it has been studied in unusual depth by a law school professor, whose law review article includes these statements:

Disney has used a blend of chutzpah and ingenuity to forge new federal law establishing the Q visa … . An examination of Disney's immigration practices offers insight into the larger questions of who designs and benefits from immigration laws. These questions are particularly worthy of attention given the current call for federal immigration reform.22

The young, generally single people recruited by Disney overseas sometimes wear native costumes for their jobs and present an international aura, but they are primarily a low-wage, near-captive work force. Their stay, by law, is held to 15 months; immediate visa renewals are not available; and there are no provisions for visiting dependents, as there are in the F, J, and H programs.

While these provisions may be confining for the aliens involved, it is helpful for our estimating purposes. There were 1,589 visas issued to this group in 2010, and assuming some drop-outs and early returns by the visa holders, let's say that the average stay was a bit over 12 months. This translates to a rounded annual population of 1,600.

Incidentally, in several current IRS publications there are also references to workers with Q-2 and Q-3 visas being free of payroll tax obligations; this was once appropriate, but it is now out of date. Congress passed the Irish Peace Process and Cultural Training Program Act of 1998, creating the Q-2 and Q-3 visas, which turned out to be the ultimate example of a bespoke immigration category, i.e., a visa class tied to an extremely narrow population, in this case an even narrower group than those on the "Disney visas". To qualify as a Q-2 one had to be between 21 and 35 at the time of application, have been unemployed for more than 12 months, and not have graduated from college; further, one had to be a resident of either Northern Ireland (i.e., the British part) or to be from one of the six enumerated counties (Cavan, Donegal, Leitrim, Louth, Monaghan, and Sligo) in the Republic of Ireland that border on Northern Ireland. To qualify for a Q-3 visa one had to be married to, or the child of, a Q-2. The program, which was used sparingly, was repealed effective October 1, 2005.23

Too many such tiny categories clutter immigration law, showing the power of lobbyists, and most of them, unlike Q-2 and Q-3, remain on the books.

Estimate of Q-1 Population: 1,600.

6. Farmworkers (H-2A) Nation-Wide and Most H-2Bs on Guam

Most hand harvesting in American agriculture is done for large corporate operations, and most of that is done by foreign-born people. The overwhelming majority of these workers are illegal immigrants, mostly Hispanic. A minority of them are legal nonimmigrant (H-2A) workers, usually from Mexico, with the next largest group being from Jamaica. This nonimmigrant program is used more widely on the east coast than on the west coast.

Typically workers spend a number of months in the United States and then return to their homelands in the winter. Given this pattern, the population can be estimated rather easily by assuming that there was one worker, per year, for each of the 55,921 visas issued by the State Department in FY 2010, and thus the worker population was about 56,000. How many person-years of employment (or the average nonimmigrant class size times the proportion of the year employed in the United States,) however, is a little more difficult to calculate.

As to the H-2B workers on Guam, most of these are in construction, and are part of a small, but fast-growing population as a sizeable number of Marines are moving from Okinawa, Japan, to Guam and military bases on Guam will need to be expanded considerably. Given the cozy arrangements that the resident population (largely Chamorros) have created for themselves with the silent assent of Uncle Sam,24 few indigenous people do physical labor, so construction work is left to foreigners.

While Guam never had control of its immigration policies, as did the nearby Commonwealth of the Northern Mariana Islands, the mainland has given the Guam establishment some leeway in the operations of the H-2B program for temporary non-skilled workers. At the moment, for instance, the national ceiling on H-2B visas does not apply to Guam, and the Guam Governor's Office plays the role usually played by the U.S. Department of Labor in the issuance of H-2B permits; a role that, I was told years ago, leans heavily to the interests of the employers. Two other bits of background: Guam's total population is about 100,000; the H-2B workers from the Philippines, the main source nation, are free of payroll taxes because of a treaty with that nation.

My numerous efforts to get someone in the mainland or the Guam governments to give me an estimate of the H-2B workers on Guam were fruitless, so the estimate starts with the one number I was able to extract, the number of H-2B admissions to Guam in FY 2010, before the DOD build-up began. That number was 1,070.

Given that H-2B visas are good for a year, and can be extended for two additional years before the worker has to leave the United States for at least three months,25 we can assume that 1,070 admissions leads to an average population of H-2Bs on Guam of 3,000 under normal circumstances, and then assuming at least a doubling of the population for defense construction, I would estimate the population at about 6,000.

Estimate of H-2A Population: 56,000. Estimate of Guam's H-2B Population: 6,000.

The Grim Politics of Reform

Before considering the legislative obstacles to changing these peculiar arrangements, let's quickly dispose of the question of recapturing the trust fund losses incurred because of the Treaty with the Philippines. The stakes are relatively minuscule for the advocates of the elderly in this case, and large for the government of the Philippines (and thus for the State Department), for the governor of Guam (island governors are always treated very tenderly by the U.S. Congress), and for the Defense Department, which would have to pay most of the bills for the FICA and related payments.

To change the H-2B provision regarding work on Guam, both the United States and the Philippines governments would have to agree to the change. The United States alone is powerless to do anything about it.

The proverbial snowball in hell has a better chance of survival than getting justice for the trust funds on this small issue.

As to the balance of the payroll tax-avoiding arrangements, all rest on a vote by Congress at some point in history and in the case of the large losses due to OPT, also on an administrative decision made by Immigration and Customs Enforcement (ICE), an arm of the Department of Homeland Security.

Why should a law-enforcement agency, totally unrelated to the trust funds, have the power to make decisions about the flows of funds into the three trust funds? Well, Congress has given ICE the power to define who is an F-1 or an M-1 student and has decided that neither of these student classes, nor their off-campus employers, should have to bear payroll taxes. Then ICE simply re-defined "students" to include former alien students, working as much as two and a half years after graduation.

The extended OPT program, which does so much damage to the trust funds, was created by the Bush II administration and then further expanded by the Obama administration, showing the power that corporations and colleges collectively have on immigration decisions. A re-elected Obama administration or an incoming Romney administration could cut back, or eliminate the extended OPT program by administrative fiat, but neither action seems likely.

Changing the tax break for employers in both the H-2A and Q-1 programs will, of course, run into the skilled and vehement opposition of the Big Farming and Disney interests, and without a congressional champion for such a change — and there is none in sight — there is little chance of either change. The tax breaks for the J-1 program and for F-1s and M-1s would, similarly, run into a buzz-saw of opposition from the Big Education interests allied to the high-tech industries that make so much use of the OPT and H-1B programs.

The one hope for a little change in the right direction comes in the SWT program, and not because of a direct modification of the current set of tax breaks. My colleague Jerry Kammer, who has studied SWT in depth, tells me that the chances are excellent that the scandals of the summer of 2011 in this program, followed by a reexamination of the program by the State Department, will lead to tougher regulations. The new regulations, in turn, are likely to make the program more expensive and thus less attractive to some marginal users of it, so the numbers will probably drop, and thus the losses to the trust funds will decline as well.

Estimates of Trust Fund Losses

What follow are ballpark estimates that should be regarded as such until, one hopes, better data arrive. The margin for error is more like 10 percent or 20 percent one way or the other rather than 1 percent or 2 percent. But they do show that the various trust funds, mostly FICA, are losing well over $1.5 billion a year as direct result of these nonimmigrant programs. Separate estimates are offered for each of the six programs.

There are three variables in each estimate: 1) the payroll tax rates; 2) the average earnings per year for the aliens in question; and 3) the number of person-years of employment in each category. The number of person years, in turn is based on the average size of the nonimmigrant class being discussed times the proportion of the year in which it is employed in the United States.

Payroll Tax Rates. The easiest of these variables: for employers: 6.2 percent FICA (for Social Security), 1.45 percent for Medicare, and at least 0.8 percent for federal unemployment tax; some states, such as Florida, Maryland, and Pennsylvania, also exclude these workers from state unemployment taxes, but since this varies from place to place, we do not include it in our basic estimate of an 8.45 percent payroll tax rate reduction for the employers.

The workers, without the visa provisions, would be paying 4.2 percent (2010 through 2012) for FICA, and 1.45 percent for Medicare, for a total of 5.65 percent. (The two percentage points less of FICA levied on the workers is part of the federal economic stimulus program, and is supposed to be a temporary measure.) The total lost to the trust funds, at these rates, was thus at least 14.1 percent of payroll (8.45 percent + 5.65 percent = 14.1 percent).

Among the complications is that some visa classes exclude aliens from taxes for a full five years, as long as they stay in nonresident status; others, notably some of the J-1 subclasses, have shorter periods of tax avoidance, such as two years. The estimates that follow take these variables into account.

Average Earnings per Year. There does not appear to be much in the way of information provided by the government on this subject, so with the exception of the Summer Work and Travel J-1 program, where the employers provide some estimates, we must base our earnings estimates on general measures of wages for the kinds of work being done.

Person-Years of Employment. The most troublesome (and thus least reliable) of the estimates is this one. The government never counts or estimates this variable, but usually provides some clues as to the size of the population of interest through workload statistics on petition approvals, visas issued, or aliens admitted, three quite different measures. In addition, there are the estimates of university-related non-citizen populations by the Institute for International Education, an advocacy group for universities with foreign students.

The estimates of how much of the year the aliens are working are based primarily on the nature and the timing of their activities; graduate students, for example, can be counted on to stay in the country for the full 12 months, Summer Work Travel aliens for about four months.

Within the general framework described above, estimates were made of the trust fund losses caused by each of the visa categories listed on the first page of this report. The estimation technique used varied a bit by visa class. In every instance we used conservative techniques, which probably, in sum, underestimate the total damage to the trust funds. The details of the specific estimation techniques, class by class, follow:

1. Summer Work Travel (J-1). Since these participants are admitted on routinely non-renewable visas, and are admitted for periods of four months, the number of admissions and participants is the same.

To further simplify things, and since one of the principal attractions of this program for employers is the lack of payroll taxes, one of the private agencies promoting SWT workers provides a simple but effective online system for calculating these savings.

The title of this little program is called:

Payroll Taxes Savings Calculator
How much can you save?
By Promoting Cultural Exchange"26

We used the employers' calculator and their estimates, as follows: 17.14 weeks per admission and $8.00 per hour wages; then, on the advice of Kammer, who has talked extensively with the program's workers, I decided to use 50 hours a week, because so many of the workers have second jobs. All of this produced the following:

132,000 admissions (FY 2010) x 17.14 weeks = 2,262,480 weeks,

then 2,262,480 weeks x 50 hours per week x $8.00/hour = $904,992,000 payroll

The 14.1 percent discount creates a rounded loss of $128,000,000 for the trust funds.

2. Other J-1. This estimate is softer than the previous one because of the sharply differing characteristics of the various populations. Of the 126,000 participants let's say that half of them, or 63,000, worked in labor markets that would routinely be subject to FICA and the other payroll taxes, that their average wages were something like $30,000 a year, and that those who did work, did so all year. This would produce:

63,000 workers x $30,000 per year = $1,890,000,000 payroll

Applying the 14.1 percent discount produces a loss to the trust funds of $266,000,000 a year.

3. F-1 and M-1 Students Working Off-Campus. Many foreign students do not work off campus. The more prosperous ones often spend summers in their homeland; able graduate students in well-funded universities, particularly those seeking PhDs, have year-around employment as research assistants on campus. Others get lesser summertime jobs on campus. My sense is that master's-level grad students probably are more likely to work off campus than bachelor's students or PhD candidates, and that M-1 students, the vocational ones, probably are more likely to have non-campus jobs than F-1 students. (I say that because my sense is that the F-1 students are more likely to be in well-funded institutions than the M-1 students.)

As to the total population of F-1 and M-1 students, we are using the 568,000 and 15,000 estimates outlined in the appendix.

The challenge is to estimate the proportion of those populations that work off campus. Let's say three-quarters of the M-1 students and one-half of the F-1 students work off campus in the summer, and that half this proportion, three-eighths of the M-1 students and a quarter of the F-1 students, work off campus during the school year. These are probably conservative estimates.

Usually off-campus work is limited, by ICE, to 20 hours a week during the school year. For convenience sake, let's say that summer work runs 10 weeks. Further, we are guesstimating the average work week for off-campus work as 12 hours, as many do not appear to get the maximum of 20 hours. These (admittedly heroic) assumptions produce the following estimates:

F-1 Summer: 284,000 students x 40 hours x 10 weeks x $10/hour = $1,136,000,000 payroll

F-1 School Year:142,000 students x 12 hours x 40 weeks x $10/hour = $682,000,000 payroll

M-1 Summer: 11,250 students x 40 hours x 10 weeks x $10/hour= $45,000,000 payroll

M-1 School Year: 5,625 students x 12 hours x 40 weeks x $10/hour $27,000,000 payroll

Total Payroll: $1,890,000,000

Using the 14.1 percent discount, this would indicate a loss to the funds of $266,000,000 a year, rounded.

4. F-1 and M-1 Graduates Working under the OPT Program. The two OPT populations previously estimated (at 30,000 for the extended program and 98,000 for the regular one) are all full-time workers, so what we need to do is to estimate their likely wages. Let's assume that the 30,000, with more technical skills and hired for longer-term periods, will be paid more than the 98,000.

As a measure of what recent alien college graduates (some with advanced degrees) earn, we turn to an independent website that uses the Department of Labor's extensive data system on new H-1B hires, It shows that the average annual wage for jobs certified, nationally, was $69,751 in FY 2010. Let's scale that back a bit to $60,000 a year for the 30,000 in extended OPT (they are about to be H-1Bs) and to $35,000 in regular OPT.27 Then we have:

30,000 extended OPT workers (FY 2010) x $60,000 = $1,800,000,000 payroll plus

98,000 regular OPT workers (FY 2010) x $35,000 = $3,430,000,000

Total Payroll: $5,230,000,000

With a 14.1 percent discount, this creates a loss to the fund of $737,430,000 or, rounded, $737,000,000.

5. Cultural Exchange (Q-1). The law review article cited earlier says that Disney, the principal employer, routinely pays a little over the minimum wage,28 so we will say $7.50 an hour and given a 2,000-hour work year we have:

1,600 Q-1 workers (FY 2010) x 2,000 hours x $7.50/hour = $24,000,000 payroll

When the 14.1 percent discount is applied it produces a loss to the three trust funds of $3,384,000 a year, rounded to $3.4 million.

This is, however, not the only savings that Disney gets out of the program, since it has excluded these workers from union contracts, and thus pays them less than comparable American workers. Further, Disney benefits from the rental and meal charges it lays on this essentially captive work force. (There are also some non-Disney users of this program, and the losses due to their participation are included within the estimates shown above.)

6. Foreign Farm (H-2A) Workers on the Mainland and H-2B Workers on Guam. There are two different sources for these discounts: federal legislation for the farmworkers (Congress loves growers) and for the H-2B (non-ag) workers on Guam, a treaty with the Philippines that eliminates the payroll taxes on this particular group of primarily construction workers. Most H-2B workers on Guam are from the Philippines; since the employers can choose where their alien workers come from, few are likely to choose more expensive workers from other places.29

This treaty provision, by the way, reinforces a broader policy point made in an earlier CIS Backgrounder, "Beware of Indirect Immigration Policy Making",30 that it is usually a disaster for restrictionists when someone other than Congress makes immigration policy decisions, in this case the diplomats at State. Further, it is much more difficult to change a treaty than it is to change a statute, so treaty provisions tend to go on forever.

Of the estimated total trust fund loss of $127,000,000 in these two categories, fully $102,450,000 relates to mainland growers' employment of alien farmworkers, with the balance coming from Guam.

The mainland estimate came about this way:

Let's assume: that there were 56,000 H-2A workers, as suggested above, that each was in the country for 10 months, that three-quarters of these months included 40-hour weeks of employment, and that wages averaged $10 an hour,31 or

56,000 x 10 months (visa duration) x 75 percent employment = 420,000 months of paid work, then

420,000 months x 173 hours/month x $10/hour = $726,600,000 payroll.

Applying the 14.1 percent discount, the trust funds' loss comes to $102,450,000.

As for Guam, and its 6,000 or so H-2B workers, I assume that virtually all will continue to be from Philippines,32 and thus free of FICA and other trust fund contributions. These are full-time, year-round jobs, so the important variable is the wage level. Because both the Governor of Guam and the Defense Department have other priorities, construction wages on Guam are much lower than in the States.

USCIS routinely goes through a wage-rate determination for 17 different construction occupations on Guam and the one for 2011 was printed in the Federal Register.33

The proposed hourly wages ranged from $11.35 to $17.62 and the raw average of those numbers was $14.55. Thus:

$14.55 x 2,000 hours a year = $29,100 a year x 6,000 = $174,600,000 payroll

Given the 14.1 percent discount, this produces a $24,619,000 loss in Guam, and when that is added to the H-2A loss, the total is about $127,000,000 in losses to the trust funds from these two H-visa sources.

The total of these estimated annual losses to the trust funds thus comes to $1,527,400,000 a year; and given the past growth of these programs, it will probably rise every year. These moneys need to be restored to our Social Security, Medicare, and FUTA funds.

Appendix: Estimating the Size of the F-1 Student Population

It is necessary to estimate the size of the nonimmigrant student populations in order to understand the impact of the various tax breaks given to them and their off-campus employers. Some of the raw materials for such an estimate are shown here. The sources are noted in Table 2. Numbers are rounded to the nearest thousand.



These are measurements of different things, so should not be compared directly. Visas are issued once in a period of years, and often do not need to be renewed for a long time. Admissions often reflect more than one trip to the United States in the course of a year by a single alien and sometimes only a single trip over many years.

The Open Doors census data gathered by the Institute of International Education (IIE) on the foreign student population in the United States in a given year suffers from partial to complete non-responses from as much as one-third of the institutions of higher learning; this response rate varies from year to year, reaching a low of 65.2 percent in 2006-2007, with the responses presumably being less frequent from institutions that have smaller numbers of foreign students. (That census, by its own definition, does not cover M-1 alien students.)

SEVIS stands for the Student and Exchange Visitor Information System, a database maintained by the Department of Homeland Security and designed to count the current F-1 population, rather than flows into that population. I do not know enough about that system to evaluate its accuracy, but note that its F-1 numbers appear to be the lowest shown above. It secures its data from institutions educating F-1 students, institutions whose willingness to provide data to DHS may be similar to their willingness to provide responses to the annual IIE survey.

In short, none of these numbers are, in themselves, totally acceptable measures of the F-1 population at any given time.

A reason for the seeming discrepancy between the visa issuance and admissions numbers for the F-1s, on one hand, and the Open Doors numbers, on the other, is the large group of aliens who are no longer students, but who continue to have F-1 status legitimately under current rules. They are in the OPT class mentioned in the text. SEVIS covers both F-1 students and OPT workers, so the low level of its total is all the more puzzling.

Bear in mind that many alien students are not on either F-1 or M-1 visas, and the IIE census notes significant groups of green card and J-1 visa holders. In earlier surveys they routinely found about 86 percent of the international students to be on F-1 visas, hence the 622,000 estimated number above (which is about 86 percent of the 723,000 total number of students reported in Open Doors for the 2010-2011 academic year).

Further, F-1 visas are routinely issued for the duration of study, so the three-year total of F-1 visas above is significant. Many PhD students are here for five years or more, bachelor's students for four years, and master's students (along with some dropouts) for two years or less, but I would expect that three years of F-1 visas would be close to the total number of F-1s in the country at any one time unless the F-1 dropout rate is much higher than I suspect.

As to the small population of M-1 students, the data above show the 35- or 40-to-1 ratio of F-1 visas to M-1 visas. Since M-1 educations generally are shorter than those experienced by F-1s, say 18 months long on average, I estimate the M-1 population at 15,000; that is a little over 150 percent of a single year's visa issuances.

Weighing all these F-1 numbers and systems against each other I estimate that the number of F-1s in college is about halfway between the estimate based on IIE data of 622,000 and the SEVIS number (which also includes some OPT people) or 513,000. That estimate is rounded to 568,000, and is more likely to be too low than too high.

Summary: In the academic year 2010-11 there were an estimated 568,000 F-1 students and 15,000 M-1 students in the United States. Both populations have probably grown by several percentage points since that time.

End Notes

1 The author had a pleasant and thorough meeting with three AARP staff members on one aspect of this topic a year or so ago. At that time the losses to the trust funds appeared to be a little under a billion dollars a year, rather than the $1.5 billion-plus estimated currently. Given the other major fiscal problems of the funds, the loss because of the alien workers' provisions was regarded as regrettable, but not large enough to warrant a major effort.

2 See

3 See (Circular E), Employer's Tax Guide for use in 2012, IRS,, page 1.

4 Ibid.

5 See Employer's Tax Guide, IRS Publication No. 15, section 14,

6 There is an exception in New Jersey, where the workers pay less than 0.4 percent of the first $30,300 of their earnings into the state unemployment insurance fund. This is the case because New Jersey unions long ago decided that they would have more influence over unemployment compensation payments if their members were nominal contributors to the system. (The author is an alumnus of the New Jersey State Department of Labor and Industry.)

7 I also have some personal experience with this variable. For the last dozen years I have been running a volunteer income tax assistance program for graduate students at a Washington, D.C.-area university. My colleagues and I see many alien students, only a handful whom have been in the United States for five years or more. They file the usual 1040 form, while the more numerous newer arrivals use the 1040NR, the non-resident form. The federal tax law is routinely more generous to 1040 filers than to 1040NR filers.

8 The six classes of nonimmigrants discussed in this paper all relate to the U.S. economy and thus logically could be covered by the payroll taxes of interest. There are other groups of nonimmigrant workers, however, such as other nations' diplomats and NATO officials, who are physically in the United States, but work in the economies of their home nations or for multinational organizations and thus could not logically be regarded as subject to U.S. payroll taxes, and are not. They are in the visa classes A, G, and NATO.

9 See, for instance, this August 13, 2011, article in the New York Times,

10 The CIS Backgrounder can be seen at

11 For the video presentation on the Summer Work Travel program, see

12 The by-now famed blind Chinese dissident Chen Guancheng was undoubtedly admitted as a J-1 Exchange Scholar.

13 See

14 See

15 During our income tax work with the graduate students we sometimes find that neither the F-1 student nor the student's off-campus employer knows that the payroll taxes are not to be collected and a process has to be set in motion subsequently to get the money back to the student. Under those circumstances the employer must have taken on the student without knowing about the payroll tax breaks that accompany such hirings.

16 That there are three visa categories covering foreign students, F, J, and M, is consistent with the general multiplicity of visa categories that have grown over the years.

17 For more on this decision, see this USCIS document issued at the time,, and another issued in 2011,

18 In addition to the 29-month extended OPT program, there can be two other additional extensions of OPT/F-1 status: a two-month grace period for all F-1s before they actually have to leave the nation, and an indefinite period in which the OPT worker is waiting for an approved H-1B visa to become active, with OPT payroll rules prevailing during this "Cap-Gap" period. For more on those arrangements, see my blog "USCIS Leans Over Backwards to Facilitate the H-1B program", at, and the references therein.

19 See the 2008 USCIS document cited in endnote 17,, and a blog I wrote at the time,

20 See For the source of the USCIS data on those in OPT status, see the reference to a Congressional Research Service report in Table 2 in the Appendix.

21 Workers hired under H-1B are more expensive than those hired under OPT for four reasons: 1) the lack of payroll taxes under OPT; 2) H-1B-specific fees, several thousand dollars per worker, that are absent in OPT; 3) the generally higher wages paid to H-1Bs than to OPT workers who, by definition, are fresh out of college; and 4) the regulation in the H-1B program that does not pertain to the OPT program that an employed H-1B must be paid whether there is paid work or not; there is no paid "benching" to use in the industry term, in the OPT program.

22 See Kit Johnson, "The Wonderful World of Disney Visas", 63 Florida Law Review 915 (2011), at She is, or was when she wrote this splendid article, a visiting assistant professor at the North Dakota School of Law.

23 See 2003 Yearbook of Immigration Statistics, p. 188, Office of Immigration Statistics, DHS, Washington, D.C., 2004.

24 The author worked for a couple of years in the 1990s with the Office of Insular Affairs, U.S. Department of the Interior, and made several working visits to Guam and to the nearby Commonwealth of the Northern Mariana Islands.

25 See this USCIS document for more on the H-2B rules:

26 The cheerfully titled calculator is a product of Contract Seasonal Staffing Solutions, LLC, of Spring Lake, N.J. It uses a discount of at least 0.8 percent for the non-payment of FUTA, which applied to earlier years. Our understanding is that the 2012 discount is lower, at least 0.6 percent. The calculator can be seen at

27 See The $35,000-a-year salary for all recent college graduates is based on a recent report by the Economic Policy Institute, Washington, D.C., at

28 See the Florida Law Review article cited in end note 22. The author must have had a wonderful time writing this article; it is full of (inevitable) chapter headings such as "Whistle While You Work". The federal minimum wage, incidentally, is $7.25 an hour.

29 There also is a local tradition, dating back to the days of when both entities were parts of the Spanish Empire, of recruiting people from the Philippines to do manual labor on Guam.

30 See

31 Professor Philip Martin, University of California, Davis, America's leading farm labor expert, suggested some of the parameters used in the estimation formula. Farmwork tends to fluctuate, both seasonally and week by week, more so than construction or office work. Further, H-2A workers are routinely guaranteed three-quarters-time employment as part of their contracts.

32 Ironically, DHS went to some trouble last year to expand the list of nations whose workers were eligible to be H-2Bs, so that workers from a number of small Pacific island nations (think Nauru, Tuvalu, and Tonga) could participate the expected DOD construction boom on Guam, as I noted in a blog ( at the time. The problem is that no Guam contractor is likely to hire from those islands when their workers will cost more than those from the Philippines because of the treaty. Did DHS go ahead with the announcement not knowing of the treaty implications? Or was it aware of the problem and made the announcement anyway? Neither answer speaks well for the agency.

33 See Federal Register, March 17, 2011 at