More New NACE Data

I’ve often referred to the data on starting salaries for new graduates, compiled by the National Association of Colleges and Employers. NACE gets its data from the schools (who are NACE members), and though there are possible issues with the data, they are very valuable for year-to-year comparisons. Over the years, they have consistently shown that salaries for Computer Science graduates have been flat, indicating that the industry’s constant drumbeat of a dire CS labor shortage is grossly inaccurate.

(I should mention that my last post on NACE data contained a couple of errors, which I have now fixed.)

The latest NACE report involves only a projection, but is still of interest. This one is on graduate degrees. In spite of all the industry lobbyist hoopla over the number of foreign students in U.S. doctoral programs, their focus in hiring foreign students is at the Master’s level. Recruiters from Intel have told me that a few times over the years, and HP said the same in a meeting with my department faculty, referring to the MS as the “sweet spot.” This is important to keep in mind, as the NACE data — taken at face value — would seem to suggest a very recent change to the MS-vs.-PhD situation.

At the Master’s level, NACE projects only a 1% increase for CS over last year, again consistent with previous years. At the doctoral level, though, the projected increase is 7.7%. Let’s take a closer look at that latter figure.

Of course, one thing that sticks out like a sore thumb is the fact that the 2015 figure is an even multiple of 1,000, $94,000. That suggests that the sample size was small, and that perhaps the median rather than mean was used. Indeed, past announcements by NACE have stated that they had insufficient data to publish figures at the PhD level.

Well, sure enough, the 2015 figure was based on just 13 responses — only a 1.2% response rate. If for instance, Stanford, whose graduates are very highly paid, did not respond in 2015 but did do so this year, that alone would likely raise the average quite substantially. No detail is given for the 2016 projections, but presumably the sample size was quite small there too.

I wouldn’t dismiss this PhD data out of hand, and indeed there may be something at work there related to the recent Big Data/Data Science craze. But as things stand, I wouldn’t put much stock in the PhD data. I do trust the BS/MS data, though, because as mentioned, those data have been consistently flat for many years.

But there is more to say, concerning this data. Specifically, note how small the salary premiums are in CS for pursuing graduate work. An American CS student with a BSCS degree who is considering obtaining a Master’s sees a wage benefit of about $10,000 accruing from that degree (confirmed for me by my campus placement office) — compared to a loss of, say, $100,000 arising from foregoing an industry-level salary during a 2-year MS course. Moreover, this student would see that he/she would likely get that $10,000 wage increment from raises received during 2 years in industry. And I haven’t even factored tuition into the picture here. The bottom line is U.S. students incur a heavy, permanent financial loss by pursuing a Master’s degree. The congressionally-commissioned NRC report reached a similar conclusion back in 2001 concerning the PhD.

All of this is highly relevant, because the industry’s favorite line is “We are forced to hire foreign students because not enough Americans go to graduate school.” For example, here is what I reported on a 2011 House hearing:

Yet, without fully realizing it, Texas Instruments V.P. for HR Darla Whitaker has now essentially admitted that all that “Johnnie Can’t Do Math” stuff was just slick PR. At the October 5 the House Subcommittee on Immigration Policy and Enforcement hearing titled, “STEM the Tide: Should America Try to Prevent an Exodus of Foreign Graduates of U.S. Universities with Advanced Science Degrees?”, Ms. Whitaker stated that TI has plenty of engineering applicants with Bachelor’s degrees, and thus does not hire foreign workers at that level. She stated TI does hire H-1Bs, and sponsors them for green cards, at the Master’s and PhD levels, where she says there is a shortage. This naturally led one of the congresspeople on the committee to ask Whitaker, why don’t the American engineering students go on to grad school? She replied that she supposed that the American students were anxious to get out and start making money.

To “start making money” may well be a draw, but as you can see from the data above, foregoing grad school is a good strategy to avoid LOSING money.

This in turn raises the question of why the wage premium for a graduate degree is so small. The answer, of course, is that the influx of foreign students into U.S. grad programs has held down wage growth at that level, as shown in work by UCB economist Clair Brown (1998 paper, 2009 book) and as forecast (and advocated) back in 1989 by an internal NSF report.

In other words, when the industry lobbyists say, “We need the H-1B program because Americans don’t go to grad school,” the reality is exactly the opposite: Americans don’t go to grad school because of the H-1B program.

Finally, a word about the generally high level of the wages, relative to, say, journalists. I say that because a journalist who interviewed me recently asked, as her very first question, “Even if the H-1Bs are bringing down wages for Americans, why is that a problem? They’re still making good money.” Apart from revealing likely bias in favor of the H-1B program, she is missing the point: Lots of American techies aren’t making that money,  because they can’t get tech jobs at all, as they are passed over by employers in favor the  foreign workers.

And as I so often point out, this is closely related to the age issue: Employers hire younger, thus cheaper, foreign workers in lieu of older (age 35+), thus more expensive, Americans. As a result, many qualified Americans either work sporadically as consultants, or leave the field altogether.

In my ASEE presentation earlier this week, I gave the example of “Dan” and “Ike,” two older Americans I know in the Bay Area, both of whom have been seeking positions in Data Science for over a year now. They both have Master’s degrees from top universities, and contrary to industry claims concerning older workers, they both have modern skill sets. Yet I see my foreign students getting jobs from the same employers that are rejecting Dan and Ike, in spite of not having any better skills or talent.

And clearly, these examples, and many others I’ve seen, show that the industry claim to hire foreign students due to having a graduate degree are false.