In 2017, the Trump administration announced that it intended to repeal the rule providing this work authorization. This February the administration followed through on that announcement with a notice of proposed rulemaking. The administration’s stated reason for repealing the rule is that it would create more jobs for U.S. citizens.

We believe a thorough benefit–cost analysis, as required under Executive Order 12866, would find this justification unfounded. Ending the ability of these workers—who are, by and large, well-educated and high-skilled—to hold jobs in the United States would at best have no net effect on Americans’ employment and likely would reduce Americans’ employment and wages. Further, ending EAD would hurt the U.S. economy and U.S. taxpayers.

The Economics of Skilled Immigration

The effect of any immigrant group on the U.S. economy depends on those immigrants’ skills and educational attainment. Highly skilled, well-educated workers, both foreign-born and domestic, have high employment levels, are less likely to avail themselves of public services such as food stamps and welfare, and are more likely to be in occupations that are hard to fill. As a result, they boost U.S. tax revenues while having little effect on government spending.

These skilled foreign workers benefit U.S. economic growth and employment, both for skilled and unskilled American workers. One reason for this is that skilled foreign labor has a relatively small substitution effect on skilled domestic workers because skilled foreign workers are relatively mobile and go where there are many available jobs. In contrast, the U.S. labor force is not so flexible: geographic mobility has gradually diminished in the United States since the 1950s and has fallen by 10% in just the last two years. The chief reason for this trend is the rise in two-income households, which increases the cost of moving for one spouse’s job.

Another reason that skilled foreign workers have a positive effect on domestic employment is that they create what economists call a “scale effect”: they boost overall economic activity, creating more opportunities and jobs for both skilled and unskilled domestic workers. This effect outweighs the small substitution effect for skilled domestic workers. For example, a 2014 study by Giovanni Peri, Kevin Shih, Chad Sparber, and Angie Marek-Zeitlin showed that reducing the number of skilled foreign workers coming to a community significantly reduced the wages of college-educated, U.S.-born workers in those communities who work with computers.

Skilled foreign-born workers have an unambiguously positive effect on unskilled U.S.-born workers. Skilled workers and unskilled workers are, in general, complementary, just as skilled workers and capital are complementary—that is, an increase in the quantity of one increases the demand and price for the other. Hence, an increase in the supply of skilled foreign workers increases the amount of capital in the economy and—along with it—the demand for unskilled workers. This results in higher wage and employment levels for unskilled workers, even without the scale effect.

Highly skilled foreign workers are also more likely to create new businesses than U.S. citizens with similar skills and education. Foreign-born workers in the United States are 30% more likely to start a new business than a native worker, and 25% of all startups in Silicon Valley have been founded by immigrants.

Workers currently holding an H‑4 visa are almost exclusively skilled workers. Thus, eliminating their work authorizations would have a small negative effect on skilled domestic workers, a large negative effect on unskilled domestic workers, and a significant negative effect on new business formation in the United States.

The Data

The American Immigration Lawyers Association distributed a questionnaire we prepared to its approximately 14,000 members, encouraging them to ask their H‑1B and H‑4 clients to complete the survey. At the same time, an H‑4 advocacy group called Save H‑4 EAD sent the survey to its members. The questionnaire had 25 questions pertaining to the immigrants’ level of education and area of study, the type and extent of work experience both in the United States and abroad, family status, and the visa status of the respondent.

We received responses from 4,708 individuals currently holding H‑4 visas, 90% of whom were female. Not surprisingly, our sample was highly educated. Recall that H‑4 visa holders are spouses of H‑1B visa holders, who are foreign workers in specialty occupations. H‑1B recipients overwhelmingly have college degrees in science, technology, engineering, and mathematical (STEM) disciplines and are employed in occupations like engineering, computer science, bio-sciences, and other high-tech areas.

Their spouses, H‑4 visa holders, tend to have similar educational backgrounds. Less than 1% of our sample had less than a college degree, and nearly 60% had a master’s degree, doctorate, or other professional or postgraduate degree. Some 83% of our respondents currently hold EADs; 75% of those EAD holders are currently employed in the United States, and almost 7% of them report being self-employed.

Employed EAD holders typically have held an EAD for two years and earn about $77,000 a year—an income that is above the U.S. median salary. Some 66% of them work in a STEM field, mostly in computer-related, engineering, or math or statistics jobs, earning on average about $83,000 annually. Some common self-reported job titles in our survey include systems engineers, software developers, automation engineers, quality assurance analysts, and data analysts—all jobs that U.S. employers have trouble filling.

Another 16% of respondents report working in the Business, Finance, or Management fields: this group typically reports holding occupations such as project managers or product managers and average about $73,000 annually.

An additional 8% of employed respondents report working in the Healthcare Practitioner or Healthcare Support fields, in such occupations as physician, dentist, pharmacist, nurse, physical therapist, and healthcare business analyst. Once again, these are areas of high economic value, reflecting the high level of education and training among these H‑4 visa holders. The average earnings of this group is about $76,000 a year.

Self-employed EAD holders / Most of the 7% of employed EAD holders who report being self-employed appear to be independent contractors—technically self-employed, but doing work for one company. However, 2% of all the individuals we surveyed operate businesses that employ both themselves and others.

About 22% of this cohort were in the Business, Finance, and Management fields, another 22% were in a STEM field, and about 23% were in the Healthcare Practitioner and Healthcare Support fields. These self-employed have on average worked longer in the United States than those who work for others; 84% of the self-employed got their EADs in 2015 or 2016, as compared to only 70% of those employed by others. The self-employed report an average income of about $60,000 a year and employ five other people on average.

The Costs and Benefits of Repealing EAD for H‑4 Visas

The process for any administration to enact or rescind a rule is straightforward: it must formally publish its intent to do so, make the entirety of the proposed rule change available on its website, and allow at least 30 days for public comment. For a rule that has an estimated economic effect of at least $100 million—either to the government, the wider economy, or both—it must also pass a benefit–cost analysis pursuant to E.O. 12886. The order tasks the Office of Information and Regulatory Affairs (OIRA) with determining if the rule does indeed pass a benefit–cost threshold.

There are approximately 91,000 H‑4 workers with an EAD, and our estimates suggest that approximately 75% of this cohort are currently employed, with an average annual salary of $80,000. The sample we obtained is both representative and quite large—approximately 4% of the population, an almost unheard-of ratio—which implies that our earnings estimate is likely quite close to the true average. The arithmetic implies that this cohort’s annual income—and its annual contribution to U.S. gross domestic product—is approximately $5.5 billion, 55 times the $100 million threshold for a benefit–cost test.

Our estimates show that rescinding the EAD for H‑4 visa holders would reduce federal tax revenue as well as U.S. economic activity without creating any jobs on net for domestic workers.

Economic and tax revenue effects / The EAD program clearly affects the ability of American employers to hire and retain H‑4 visa holders. Our estimate of $5.5 billion in annual lost earning for this group provides a minimum estimate of their employers’ lost output. However, the H‑4 EAD program also affects the ability of American employers to hire and retain H‑1B visa holders. Some 28% of our employed H‑4 respondents reported that their EADs have been important in their families’ decision to remain in the United States; an identical 28% of the H‑4 respondents who are not currently employed but want to work identify the EAD as important to their remaining in the United States. This suggests that rescinding EADs could result in the loss of up to 25,000 H‑1B employees and the roughly $2 billion in U.S. production that they contribute to the U.S. economy. Thus, a better estimate is that the rule rescission could reduce U.S. GDP by around $7.5 billion per year.

Our estimates show that rescinding the EAD program for H‑4 visa holders would reduce federal and state tax revenue and U.S. economic activity without creating any jobs on net for domestic workers.

We can improve our estimate of the annual cost to the federal government from ending EAD in terms of forgone tax revenues. H‑4 visa holders earn about $80,000 a year and they all must have employed H‑1B spouses. After the $24,000 standard deduction and one $2,000 child tax credit, the H‑1B spouse would have about $54,000 in taxable income. At current tax rates, the H‑4 spouse’s income would result in an additional tax bill of $15,300. To that we must add both the employee’s and the employer’s shares of the payroll tax, about $12,200. Hence, each employed H‑4 visa holder would, on average, pay $27,500 in federal taxes annually. With roughly 68,000 H‑4 visa holders currently employed, that means that ending their employment would result in a federal revenue loss of around $1.9 billion annually. This estimate does not reflect the additional tax revenue that would be lost when American employers lose some of their H‑1B employees.

We also estimate the state and local taxes paid by this cohort by multiplying their total income of $5.46 billion by the average proportion of income that states and localities assess, 9.75%, via income, sales, property, and other taxes. This results in a state and local tax revenue loss of $530 million. This estimate is undoubtedly low because it assumes that the distribution of H‑4 visa holders resembles that of the overall population; in fact, they disproportionately congregate in high-tax states like California. Further, our calculation ignores the effect of the EAD rescission on employers’ ability to attract and retain H‑1B workers.

Employment effect / A question basic to the motivation for rescinding the work permit is whether the jobs currently held by H‑4 visa holders would subsequently be filled by U.S. citizens. This defies an easy or objective answer, but we can make a few relevant observations.

First, the education and training of typical H‑4 visa holders mean that the jobs they occupy would be relatively difficult for employers to fill with American workers. Unemployment rates tend to fall steadily with educational attainment. In June 2018, the unemployment rate for people without a high school degree was 5.8%, two points above the national average, but 4.3% for those with a high school diploma and only 2.5% for college graduates. For those with professional degrees or doctoral degrees, the unemployment rates were 1.5% and 0.9% respectively. These numbers suggest that there is very little unused supply of U.S. labor with the ability to do the work of the typical H‑4 worker with a postgraduate degree and several years of professional experience.

A second, complementary point is that, as we approach the 11th year of an economic expansion, the availability of unemployed workers at any skill or educational level willing and able to do the jobs currently held by H‑4 visa holders is relatively slight. This is not to say that there is no pool of underemployed workers in the U.S. economy; the labor force participation rate, which measures the proportion of the adult non-institutionalized population that is active in the labor market, is 5–7 percentage points below where it was at the peak of the previous two business cycles. This suggests that there may be a pool of domestic workers willing to enter the labor market if the opportunity arose.

However, it is more likely that the labor force participation rate today does not have much room to increase: research published by Ike Brannon and Andrew Hanson finds that the gap between past rates and the current rate are due to demographics (a greater proportion of workers are above age 55), the crippling effects of opioid addiction, and the ongoing sluggishness of new home construction. None of these suggest that there is a sizable pool of unemployed college graduates with STEM degrees.

To reasonably approximate the number of H‑4 visa-held jobs that would be filled by U.S. citizens, we began with the Bureau of Labor Statistics’ unemployment rates by occupation for June 2018. We assumed that the frictional unemployment rate is roughly 2%, so any occupation with a 2% or lower unemployment rate would have no excess workforce whatsoever: all the currently unemployed workers in that occupation would be in the process of searching for and moving to a new employer. On the other hand, we assumed that if the unemployment rate in an occupation is 8% or higher, there is sufficient slack to allow all the H‑4 visa workers to be replaced. For occupational unemployment rates between 2% and 8%, we used linear interpolation: at a 4% unemployment rate, only one-third of the H‑4 workers would be replaced, but at a 6% unemployment rate two-thirds would be replaced.

In our sample, over two-thirds of the employed H‑4 EAD holders were in fields with a June 2018 unemployment rate below 2%: Management, Engineering, Legal, Healthcare Practice, and Computer and Mathematics. Interpolation suggests that only 8% of the H‑4 EAD workers would be replaced by American workers if the H‑4s lost their employment status—12% if we were to boost our slack estimates. We conclude that eliminating the employment status of 68,000 working H‑4 visa holders would result in the employment of only 5,500 to 8,200 U.S. citizens.

Remember that 2% of our employed H‑4 visa holders were self-employed, and they in turn employed an average of five workers. If the 68,000 employed H‑4 EAD holders all lose their employment status, 6,800 U.S. citizens would also become unemployed when the H‑4 business owners liquidate their businesses. This would almost exactly cancel out any employment gains accruing to U.S. citizens from H‑4 job replacement. And this ignores the other employment-creating effects from high-skilled H‑4 workers. Overall, ending the H‑4 program will likely reduce overall employment and wages for American workers.

No Discernable Economic Benefits — at Best

Rescinding employment authorization for H‑4 visa holders would result in substantial costs to the U.S. economy, to federal and state tax coffers, and to U.S. employers’ ability to attract H‑1B workers. There would be no employment or income gains by domestic workers to offset those losses because the relatively small gains to U.S. workers replacing H‑4 workers would be offset by the jobs lost when self-employed H‑4 workers were forced to close their businesses and dismiss their employees. Worse, the loss of these high-skilled foreign workers is likely to have a negative overall effect on Americans’ employment and wages.

In short, rescinding H‑4 visa holders’ ability to work fails to meet any credible benefit–cost analysis. This proposal should thus be rejected by OIRA.

Readings

• “Closing Economic Windows: How H‑1B Visa Denials Cost U.S.-born Tech Workers Jobs and Wages during the Great Recession,” by Giovanni Peri, Kevin Shih, Chad Sparber, and Angie Marek-Zeitlin. Partnership for a New American Economy, June 2014.
• “Foreign Scientists and Engineers and Economic Growth,” by Giovanni Peri, Kevin Shih, and Chad Sparber. Cato Papers on Public Policy 3: 107–184 (2014).
• “Wisconsin: A Blueprint for More Workers,” by Ike Brannon and Andrew Hanson. Badger Institute, August 2018.