More Data on the Age Issue, Plus Clarification on H-1B Underpayment

As most of you know, I often point out that employers use the H-1B work visa program in order to hire young foreign workers in lieu of U.S. citizens and permanent residents over age 35.  Younger workers are cheaper, as a recent PayScale report illustrates.  (Note that older workers are more costly in terms of health benefits and the like as well.)  Though the analysis simply looks at college major, regardless of whether the person is still working in that field, the numbers (adjusted for inflation) are comparable to the report that I wrote at the invitation of the California Labor & Employment Law Review, a California State Bar publication.  Basically, the magnitude of  wage savings accruing from hiring the young is about 40%.

So you can see why the employers concentrate on the young H-1Bs, and for that matter new or recent grads in general.  Concerning the latter, see my recent blog post, “A Failed Recruitment,” particularly the material quoting Intel.

Which brings me to a point I’ve been meaning to make for a while.  I’ve always referred to what I call Types I and II wage savings accrued by hiring H-1Bs.  The former involves paying H-1Bs less than what comparable Americans make (the word comparable is key), while Type II arises from hiring young foreign workers instead of older Americans.  My Migration Letters paper estimated the Type I savings at something like 20%, and this is what I wish to clarify.

That 20% figure was based on a technique I like to use:  Take the industry’s claims at face value, and see where that leads.  I took that approach in my EPI paper on the technical quality of the H-1Bs, and used this technique in a different context in Migration Letters, concerning the wage savings issue.  In this latter case, I reasoned as follows:

  • The legally required wage for H-1Bs, called the prevailing wage, does not take into account special skills, say Android programming.  Yet the data show that most of the foreign tech workers are paid at or near the prevailing wage.
  • The employers claim that they hire H-1Bs because the foreign workers possess special skills that are rare among Americans.
  • The special skills typically command a premium of 20% or more on the open market.
  • Therefore, the Type I savings is about 20%.

What is often overlooked — including in my own statements, I must admit — is that the above analysis is predicated on the employers’ claims that they hire H-1Bs because qualified Americans are not available.  In reality, though, many employers are motivated not by Type I savings but by Type II.  And it’s no wonder, since the latter savings are much larger.  My point, then, is that even if an employer pays his young new grad H-1Bs the same as similar Americans, he is still saving something like 40% in the Type II sense.

In other words, if one drops the assumption that the employers are telling the truth 🙂 the overall Type I savings might be somewhat lower than 20%.  I say “somewhat,” because one must also note that most analyses, including mine above, are based on wage at the time of hire.  The NRC survey found that employers admitted to giving the H-1Bs smaller raises and so on, so there is a counter factor, making the 20% figure (or its adjusted version) actually too low.

To my knowledge, my papers are the only ones to take skill sets into account, even in the indirect manner described above.  Clearly, it is a key factor and cannot be ignored.

So, as is often the case with economic issues, we do not have a firm estimate of the magnitude of Type I savings.  But as I state in the Migration Letters article, the question of whether such savings exist should be considered settled, just from applying basic principles:

  • Immobile workers make less than mobile ones, all else being equal.
  • The H-1Bs who are sponsored for green cards (i.e. tech industry mainstream, e.g. Intel, Google etc.) are immobile.
  • Thus the mainstream H-1Bs are underpaid, compared to the market value they’d have if they were American.
  • In addition, the foreign workers tend to be willing to work for less, as a green card serves as large nonmonetary compensation for them.

So, again, the use of the H-1B program by employers to save labor costs ought to be considered a settled question.

In an upcoming post, I’ll present a “cookie theory” analysis of computer industry wage trends.  Stay tuned!


4 thoughts on “More Data on the Age Issue, Plus Clarification on H-1B Underpayment

  1. Norm, we’ve often discussed this, but let me put up a couple of thoughts here on your blog. First, I’d add a Type III factor, that any gross enlargement of the labor pool is going to tend to drop ALL wages, and that as much as anything was a key to the introduction of H-1B in the first place. Just walk in and look around any STEM workplace today, and *apparently* the H-1B impact on the gross population is HUGE, let’s say on the order of 50%. Independent of other factors, what should this do to any wage levels!?

    Second, is whether the age savings are real! The *wages* are less, but if the productivity is also less than the savings are nil – or negative! I’ve been doing this stuff a long time, and I’m pretty sure my productivity didn’t hit a high until I’d been at it for 20 years and more, and there are a lot of tasks the older me could do, that I would never want to give to the younger me. And that holds for others I’ve worked with as well, the best of the best don’t get that way without a lot of experience. So this is a negative Type II.

    And really, that should be a Type IV, the exclusion of the best as wages are depressed. At very least, the best don’t work for big companies that are structured to use the cheapest wages possible. They leave the field, or don’t enter in the first place, or if they also have business skills start their own companies. And even worse when the best *do* work for big companies simply to pay the rent, they find themselves in job slots engineered to fit only the cheap-average contributors. This is a shame, as it probably throws away 90% of the productivity of the best, no matter where they are from or what they are paid.

    The end result is that H-1B has depressed wages at least 50% since 1991, depending on your inflation model – the government numbers average less than 3% over that period, more realistic numbers would average about 5%. Within large swaths of STEM including where H-1B concentrates most, in IT, there are other factors as well over this time period that have reduced effective wages, one is just a “maturing” of the field that is often used as an umbrella in technology/market arguments. These other factors also apply previous to 1991, so that STEM wages overall since 1980 are probably reduced more like 75%.

    By 1980 numbers adjusted only for inflation and not impacted by either H-1B or “maturing” factors, a senior software developer position would average in round numbers somewhere around $400,000 instead of the area of $100,000. I would estimate the H-1B factors Type I, II, and III explain about half of this impact especially since they only occurred after 1991.


  2. I have been engaged in a long discussion involving H-1B hiring practices on Quora. I posted this post in a comment in response to a hiring manager at Twitter who argues that age is not an issue and H-1Bs are not used to save money.

    Quora seems to attract technical types. The question under discussion there is “Why can’t I get a job at some tech giants despite doing well on interviews?” I answered as best I could, but the manager, naturally, shot me down with his “facts”. Since you have more facts at your disposal than I do, I have added this blog post to the quora discussion.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s