The Myth of America's Tech-Talent Shortage

So it turns out the United States is not, in fact, the educational wasteland tech industry lobbyists would have you think.

Companies like Microsoft often claim that America is suffering from an economically hobbling shortage of science, math, and computer talent. The solution, they argue, is to let employers fill their hiring gaps by importing tens of thousands of educated guest workers beyond what the law currently allows. Much as farmers want to bring in field workers from Mexico on short-term visas, software developers desperately want to bring in more coders from India.

The Senate's current immigration bill would grant their wish. As written, it vastly increases the annual limit on H1-B visas, which allow corporations to bring employees with a bachelor's degree to the U.S. from overseas for up to six years. Roughly half the guest workers who currently arrive through the program come for computer-related jobs. When Facebook CEO Mark Zuckerberg announced earlier this month that he was forming a political action group to back the reform effort, it was in part seen as a move to ensure that the H1-B provision would make it to President Obama's desk intact.

There's just one problem. That whole skills shortage? It's a myth, as was amply illustrated (yet again) by a report this week from the Economic Policy Institute. It still might be the case that tech companies are having trouble finding specific skill sets in certain niches (think cloud software development, or Android programming), but there simply aren't any signs pointing to a broad dearth of talent.

Our Programmer Surplus

Colleges, for instance, are already minting far more programmers and engineers than the job market is absorbing. Roughly twice as many American undergraduates earn degrees in science, technology, engineering, and math disciplines than go on to work in those fields. As shown in the EPI graph below, in 2009 less than two thirds of employed computer science grads were working in the IT sector a year after graduation.

Meanwhile, the unemployment rate for programmers is still stuck well above its pre-recession average.

Could it be that schools aren't teaching their students the right stuff, that despite their fancy credentials, today's grads lack the programming chops or logical prowess needed to succeed at a Google or Microsoft? Not so much.

In industries where talent is scarce, economists generally expect wages to rise, as desperate companies go chasing after what few qualified souls they think can do the job. That's exactly what's happened to oil and gas engineers over the last decade during the energy boom, for instance. But while there have certainly been anecdotal accounts of Silicon Valley firms tossing outrageous sums at elite college students, in the big picture, programmer salaries have been stagnant ever since the dotcom bubble went bust more than a decade ago. The pattern holds whether you look at the national data, or just at traditional tech centers such as Silicon Valley, the Route 128 corridor outside Boston, Dallas, or Austin, where you'd expect competition for talent to be hottest.

The "Indentured" Workers Problem

The H1-B program's fiercest critics, such as University of California, Davis computer science professor Norman Matloff, have long derided it as little more than a pipeline for cheap "indentured" labor. Companies are technically supposed to hire H1-B immigrants only if there are no Americans available to do the job, and then are required to pay them on par with U.S.-born professionals. Thanks to an array of legal loopholes in the way appropriate wages are calculated, though, it doesn't necessarily work out that way.

Often, it comes down to a matter of age: Companies frequently save money by hiring a young, less experienced immigrant instead of an older American who would command a higher salary. And because the bureaucratic hurdles make it difficult for H1-B holders to switch jobs, guest workers have notorious difficulty bargaining for promotions or raises. They also can't go off and start their own businesses, as they'd lose their visa. Unlike green card holders, they're professionally chained in place.

The program has also fed the pernicious growth of IT outsourcing firms. These companies use H1-B visas to import low cost tech workers by the thousands, who they hire out to American corporations as substitutes for better-paid, in-house staff. The Boston Globe reports that just 4 of these companies -- New Jersey-based Cognizant Technology Solutions along with India-based Tata Consultancy Services, Wipro, and Infosys -- claimed 20 percent of the 134,780 H1-B visas that were approved in 2012.

In light of these concerns and what we now know about the "skills shortage," is there any way to justify expanding H1-B?

Possibly, but only with caution.

The Benefits of High-Skill Immigration

Hollow rhetoric about our unskilled workforce aside, there are a few compelling arguments in favor of H1-B. First, and most simply, it's reasonable to think that companies should be able to quickly and easily fish for talent abroad in the cases where it's truly necessary.

Moreover, there may be evidence that the H1-B program has been a net plus for American workers. Earlier this year, U.C. Davis's Giovanni Peri and Colgate's Chad Sparber released a draft paper suggesting that that from 1990 to 2010, the influx of H1-B workers actually accounted for an astounding 10 to 20 percent of yearly U.S. productivity growth, and added $615 billion to the economy. That, they say, boosted both wages and employment for U.S. born scientists and engineers. There were undoubtedly winners and losers; while average salaries rose, some older Americans almost certainly gave up their jobs to younger immigrants with fresher skills. But if the paper is right, that price seems worthwhile in the grand scheme of things.

Finally, over the long term we really do want more immigration. Economies grow essentially one of two ways: productivity growth, or population growth. H1-B helps with the latter. It obviously is not a perfect tool for the job (after all, it's technically supposed to be a "non-immigrant" visa). It would be far better if we simply handed skilled workers more green cards -- and to be fair, the immigration bill is designed to do that as well -- since that gives them the freedom to shop around for jobs, bargain for higher pay, and set down roots in a community. But inviting in a group of educated, decently paid, law-abiding professionals who might one day become permanent residents through employer sponsorship or through marriage isn't a terrible second option.

A Policy Compromise

So perhaps a compromise is in order. Today, the the number of generic H1-B visas available to for-profit companies is capped at 65,000, with another 20,000 set aside for foreign students who earn an advanced degree at U.S. universities.* The Senate plan would give those numbers an MLB-sized shot in the arm. The cap would jump to 110,000 visas minimum, but could go as high 180,000, based on a formula that would account for both demand from employers and the unemployment rate for skilled professionals. Another 25,000 visas would be around for advanced degree earners.

This is too much, too fast for program that's such a mixed bag in its present state. Instead of rushing to bulk it up, Congress should just make the system more flexible by leaving the cap at 65,000 and carefully indexing it to health of the job market (the current formula could probably use some tweaking). That would give it room to grow without creating a sudden surge of new guest workers into an economy that may not be primed to handle them.

It would also create time to judge Congress's efforts to fix the program's many faults. The immigration bill both increases the minimum salaries employers will have to pay H1-B workers and comes loaded with financial penalties aimed at cracking down on companies, like the outsourcing firms, that over-rely on the visas. But will that be enough to head off future abuse? We don't know, and it seems unwise to super-size H1-B until we do.

Our immigration system should be designed to adjust to the needs of the market, but for the time being, we seem to have as many brains as the market needs. Let's not act like we don't.

__________________

*Academic institutions and nonprofits aren't subject to the cap at all.

A note to regular readers: A few Atlantic fans might recall that I've written much more blithely in support of increasing the H1-B program cap in the past. As I've learned more about the program, my thoughts about it have obviously changed. While I would love it if all of my opinions were unimpeachably well thought out, I think I have an obligation to be transparent in cases where they were not.