Sign In

Communications of the ACM


Thoughts on 'The Frenzy About High-Tech Talent' and a Call For Corporate Action

View as: Print Mobile App Share:
Valerie Barr

Valerie Barr, Professor of Computer Science, Union College

The July 9, 2015, issue of The New York Review of Books carried a very thoughtful piece by Andrew Hacker.  In "The Frenzy About High-Tech Talent," Hacker discusses a number of books and reports that address whether or not there really is a need for more tech talent, the justification for the H-1B visa program, and issues in the American educational system.  Hacker is Professor Emeritus of political science at Queens College, part of the City University of New York, where he taught both political science and mathematics.

Throughout his piece, Hacker is basically questioning two things:

  1. Is there really an unfilled need for STEM graduates, or are we actually graduating too many so that many end up unemployed or employed in different areas?
  2. Are there flaws in the American education system, both at the K-12 level and in college, that lead us to be very dependent on foreign STEM graduates?

I recommend Hacker’s piece, and you might want to take a look at the books and reports on which he based it.  My interest here is not in debating the original texts, but I do think there are a number of key points that Hacker overlooks.  I raise these not so much to take issue with Hacker, but because his piece is but one instance of a set of oversights commonly made when people talk about job projections in STEM.

Liberal Arts and STEM

Several times in the piece, Hacker makes a distinction between STEM and liberal arts.  The first place this appears is in the statement "It's true that the U.S. has fewer people studying the subjects involved in STEM than many other countries.  The chief reason is that more of our students choose to major in business and liberal arts."  This comment implicitly elides "liberal arts" and "non-STEM," despite the fact that liberal arts colleges offer STEM majors, and that the original roots of the liberal arts included arithmetic, geometry, and astronomy (along with music, grammar, logic, and rhetoric).  At my own institution, Union College, 40% of our student body graduates with a STEM major (in addition to the traditional liberal arts, we also have engineering).    

This elision also arises in a table Hacker presents, drawn from Princeton Review.  It shows how students rate their professors, comparing technical institutions that focus on STEM (such as Georgia Tech and MIT) with liberal arts institutions.  Rather than looking at ratings of entire faculty groups, a more useful comparison would examine how students at each of those types of institutions rate faculty in just the STEM disciplines.  I suspect we would still see higher ratings for faculty at liberal arts schools because of factors such as small class size, one-on-one advising of students by faculty, number of office hours, little use of TAs, and heavy involvement of faculty in lab sessions.  Nonetheless, it would be more accurate to compare ratings of faculty who are in the same disciplinary groups, rather than comparing entire faculty bodies at the two types of institutions.

Where Are the Jobs Really?

The texts Hacker is reviewing, and his own information, seem to dwell predominately on overall job projections for the STEM fields.  Nowhere does there appear a breakout of the job forecast for computing related job categories.  With the increased ubiquity of computing across all industries and employment sectors, it seems unlikely that we will see the "deskilling" trend that may be occurring in engineering (whereby engineers create equipment that means they and others like them no longer have job opportunities).  We know that there are many jobs in the "tech sector" but there are also a lot of computing jobs in banking, finance, manufacturing, agriculture, healthcare, etc.  We can get an accurate picture of future job openings only if we can make a good determination of the computing jobs that exist outside of the "tech sector." 

Are There People to Fill the Jobs? (Hidden Pools of Talent)

There is much discussion and hand wringing about all of the looming job openings in computing and the dearth of people to fill them.  A lot of focus has been on the front end of the pipeline — how do we get more students exposed to CS early, and how do we recruit and retain a larger and more diverse pool at the college level?  These are very worthwhile questions.  But even if we snapped our fingers and woke up tomorrow with 10K trained CS teachers and fully implemented K-12 CS programs across the entire world, it would take quite a while for the students of today to take on the jobs that are open right now.  All is not lost however, if companies are willing to make some changes.  Consider the following:

1.  A 2008 report showed that 41% of technical women in tech companies leave within 10 years, or they transfer into positions on the business side (sales, marketing, etc.).  A more recent report shows that women working in science, engineering, and tech fields are 45% more likely than their male peers to leave the industry.  Let’s do some math.  There were 12K people at the recent Grace Hopper Conference.  Let’s guess that half of those were women students looking for jobs, and let’s say that they all were hired and will start new jobs next summer.  That’s 6K new tech employees, starting jobs in June 2016.  Unless something changes soon, somewhere between 2460 and 2700 of those new employees will leave.  That's not a winning strategy for filling a large number of open positions!

2.  Hacker points out, "employers see no need to encourage longer periods of employment, since each year cheaper graduates…arrive with their resumes." That may work in fields where the number of openings is reasonable and relatively stagnant. But that doesn’t work in the tech and computing related areas where there is huge need.

3.  In the past, in the early days of the computer age, when there were no academic CS programs, companies provided training for new employees.  In fact, many women who had been math majors and were planning on teaching careers instead were trained by major computing companies (like IBM) and went on to have long and productive careers (Turing Award recipient Fran Allen comes to mind immediately).

It is true that the industry changes quickly in some ways, with new tools, new approaches, and new languages. But there is a rich pool of potential employees who are being completely overlooked. The many women who have left tech positions could be brought back in and given training to bring them up to speed on the newest languages and development practices.  But this is a reasonable approach only if, at the same time, the tech industry makes a commitment to improving climate. There is no point in bringing back people who left tech if they are simply going to want to leave again in another 5 years. In fact, I imagine that bringing back a group of tech veterans who have greater maturity and experience could do wonders to improve climate in some of the tech companies.  But the companies have to commit. And they have to recognize that you can still be a cutting edge agile company even if the average age of your employees ticks up a bit.

Who will step up and be the first company to introduce quality in-house training, along with serious commitment to changing climate, as a multi-pronged strategy to recruit in the many women who left tech during the last decade or so?


Peter Bye

I have a couple of comments on Valerie Barr's interesting piece. FIrst, if there is a shortage of IT people, perhaps we should look beyond graduates in STEM subjects. I started in IT 50 years ago. Although I have a degree in physics, I have met excellent performers with degrees in languagaes, music and other subjects, as well as no degrees at all. We used to have aptitude tests and in-house training, as Ms Barr notes about the early days of the industry.

A second comment picks up on a remark Ms Barr quotes from the original NYRB essay: "employers see no need to encourage longer periods of employment, since each year cheaper graduatesarrive with their resumes." This represents a frightening view that experience does not matter. I have noticed, for example, that those replacing on-shore with off-shore people often expect nothing to change except the cost. In many cases, years of experience are being discarded.

Does experience really matter? I think it does. We encounter too many low quality systems, in particular regarding security, in our everyday lives. Perhaps lack of experience in those procuring, developing and deploying them is a significant reason for the poor performance.

Peter Bye

Mary Anne Wolf

I don't believe there is any shortage of people to work in STEM in the USA. There are at least 5 self-inflicted reasons why employers may believe that they experience a shortage.

1. In America, the supply of people trained in STEM is not uniformly distributed. Employers often try to locate jobs where the cost of living is low and the supply of STEM trained people is also low. If they can find one of few, or convince someone to relocate to a place where there are no other employers that use their skill set, then, yes, they might pay less. However, it is far more likely that they will simply not find someone with the right skills. Most people trained in STEM can do math
well enough to figure out that it is not worth moving to a location with no other job prospects. In software, hiring someone local to an area and having them also work from home in order to make them available 24x7 is common. Allowing someone not local to do a job should be more common than it is.

2. The demand for skills is ridiculously specific. Here is an example from a real ad:
Skill requirements:

Expert level skill set in WebSphere version 8 Administration running in Linux environment.
Expert level skill set in Apache/IBM HTTP Server Administration.
Expert level skill set in TOMCAT Administration.
Knowledge of Introscope Monitoring Tool will be a plus.
Expert in setup of Deployment manager, Nodes, WAS clusters, SSL configuration and application deployment.
Preferable experience in Sterling Commerce NDM Direct Connect.
Working knowledge on Databases - Oracle Performance Tuning and Optimization Shell/PERL scripting, Automation jobs Monitoring and Trouble shooting skills on UNIX.
Manage servers through F5 Load Balancer.

And that is not the whole ad.

The original reason for these impossible lists is to justify not hiring because someone is Black, old, disabled, female, gay, religious minority, etc. However, employers now seem to think that only the whole list will suffice, and from a single job during the last year. No wonder they don't find it. (This single job constraint is from recent personal experience.)

3. Employers do not invest in their talent seeking staff. Resume filtering is often done by inexperienced temporary contract recruiters who do not have STEM background and who cannot recognize synonyms. These contractors do not stay with one company for long enough to become familiar with the company culture. Companies also generally refuse to pay for experienced contingency recruiters. If they don't value ongoing expertise long term in finding and recognizing the candidates who will fit, no wonder they don't find any.

4. Age discrimination is rampant. Why would anyone in their right mind spend age 18 to 25 getting a BSCS and MSCS if they do not expect to have a career using their tech skills after the age of 30?

I believe that H1B visas should cost 5 million dollars per calendar year or fraction thereof, to be adjusted based on nationwide CEO total compensation increases. If corporations can afford to pay that much to executives, then they can afford to pay much less for the talent that business requires. It will mean, however, that H1B visa holders will be a small minority of employees. Smart employers might decide that being a little less specific and hiring someone who already lives there, or that locating the jobs where the desired talent is available, or allowing remote nonlocal workers, might make more business sense.

Education is simply not the problem. Whatever students learn, technology moves so fast that they will need to learn new things most of the time anyway.

Rudolf Olah

Michael O. Church has discussed the age discrimination and the culture of using up people's skills as fast as possible and tossing them out on his blog:

Age discrimination isn't that hard to notice; most of my coworkers have always been between 20 and 30 years old, and as I'm getting older there's some that are nearer to 35 but I expect there will be few 40+ year olds (I've only seen two 35+ year olds so far and that's over the last 8 years).

Investing in training and future potential is where big gains can be made. Heck, investing in research can be a huge payoff. We saw that with XEROX PARC where they invented much of modern computing in a 3-year time span. The payoff of training would be immediate quality improvements and much less technical debt.

Laura Farrell

My original degree was in Music. I've had to santize my CV of this fact due to the response of as much as 50% of employers, to whom it was akin to having announced that I had bubonic plague. In fact, it has been hugely advantageous, to the point where when I did take a second degree, I took it in Business rather than IT because I found IT much easier to learn on-the-job. It is a shame that there is such an accepted aura of outright discrimination against those without technology-specific backgrounds.

I agree with the problem where offshoring has eliminated the entry level tiers and with it much of the training without which I myself would not have enjoyed a 15 year career, however, in practice I have seen many of those outputed from in-house training to be remarkably poor later on as employees. I find a good amount of the "BS" comes from those, who erroneously inflate their abilities in order to get promotions and then create a total mess, both technically and as business decision makers. The biggest problem I have seen in 15 years is that of taking a very narrow view of the world, and basing decisions on that.

Displaying all 4 comments

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account