IT companies heavily courting CS grads

By Louise Arnheim

As recently as three years ago, there was concern in academic circles that US universities might be producing more doctorates in computer science and computer engineering than the domestic economy could absorb. No more.

Today, university CS departments report difficulty in recruiting the new faculty they want, while industry labs are thirsty for Ph.Ds in nearly every CS and CE specialty area. What has happened? Where did the Ph.D. surplus go?

In fact, there is little evidence that there has ever been a surplus of doctorates in computing disciplines. What has occurred is a shift in the mix of academic and industrial positions for computer science Ph.D's. Data from the past four years of the CRA Taulbee Survey (see page 16) show that the jobs taken by new Ph.D's have shifted substantially towards industry. The number of academic positions has remained relatively constant from 1993 through 1996 while the number of Ph.D. graduates recruited by industry has grown steadily. Individuals whose expectation of employment after graduation is a tenure-track faculty position in a prestigious university may be disappointed, but universities have always been selective in their faculty appointments. Those who expect to find a challenging technical position in research or advanced development should be well pleased. They can choose among many opportunities.

Data from the annual Taulbee surveys are quite revealing. The number of Ph.D. graduates reporting themselves as not yet employed or unreported at the time of the survey has declined steadily since 1993. This number represents a pool of trained talent available for employment. It does not imply that these individuals remained unemployed or were forced to seek employment in lower-skilled jobs. It simply accounts for those who had not obtained definite employment at the time they were surveyed after graduation, for whatever reasons. If this trend continues, this pool of trained talent will be exhausted in 1997.

Another interesting statistic is the number of new Ph.D's who accepted employment outside of North America following graduation. This number is also in steep decline. Presumably, most of these persons are non-resident, non-nationals who have pursued graduate study in the US or Canada. Apparently, fewer are returning to their country of origin because they are being offered attractive employment opportunities here, following graduation. The employment statistics depict a growing demand for new Ph.D's in the face of a supply that remains nearly constant. They indicate that in 1997, the slack in this employment market will have been entirely used up.

Anecdotal evidence from all geographic sectors of the US supports these statistics. University research groups have reported serious difficulty in filling postdoctoral research positions; many are now recruiting for researchers abroad. Industrial laboratories have been unable to meet their hiring goals. Even university faculty positions are going unfilled because departments cannot find persons with the qualifications they want, although the number of faculty openings has not yet begun to grow significantly.

It is worth speculating about whether this trend is likely to continue. Since the trend is being driven by the demands of industry, it could end if the computer, communications and information technology industries in the US and Canada should cease growing. However, these industries are so tightly enmeshed in the overall economy that it is difficult to imagine a long-term reversal there, short of a collapse of the entire US economy. Rather, what appears to be happening is that companies in these industries have been systematically raising the education levels to which they tie their hiring goals, to accommodate the increasing technical complexity of their products and businesses. Companies that only occasionally hired at the Ph.D. level five years ago now have regular recruitment programs to hire new Ph.D's. Furthermore, early in the new decade, the children of the baby boomers (the echo) will begin to enroll in colleges and universities in large numbers. University enrollments can be expected to swell again for at least a decade. This is likely to bring a renewal in the growth of university faculties during the next decade.

Employment statistics, anecdotal evidence and plausible speculation all point to a severe shortage of new Ph.D's in computer, communications and information sciences and engineering. What of the supply? Unfortunately, the pipeline is long (the average time to a Ph.D. is closer to 6 years than to 5) and the number of new students enrolling in doctoral programs has not shown any significant upward trend over the past four years. The Ph.D. granting CS and CE departments have not rallied to the situation as they did a dozen years earlier, in part because the shift in the Ph.D. marketplace from academic dominance to industry dominance caught them by surprise.

National and state governments have not responded, either. Federal agencies have tended to look at science and engineering in the aggregate. They appear to have overlooked the fact that new doctorates have become scarce resource for the nation's most pervasive, most productive technology sector. The situation in the computing and communications disciplines is unique; there is no comparable shortage looming in physics, biology, chemistry, astronomy, oceanography or any of the traditional fields of engineering.

The current situation is not being corrected by market forces, either. New BS graduates, offered starting salaries that go into the $60K range, are tempted to delay indefinitely their entry into graduate school. Graduate students are being recruited out of the laboratory by industrial firms with exciting work to do and the money to get it done. University CS departments, faced with unprecedented competition in recruiting talented students, are offering "signing bonuses" to new students in competition with one another. This competition works to the benefit of the student who has already decided to enter graduate school but it does little or nothing to entice more students to apply.

If we are to head off a severe shortage of highly trained talent that will be needed to refresh our universities and infuse the computing and communications industries with new research throughout the next decade, there must be concerted action to stimulate graduate (Ph.D.) enrollments now. The CRA, through its member institutions can set targets and recommend streamlining of programs to increase their efficiency. Industry can sponsor graduate traineeships and offer internships to graduate students. Government agencies can fund fellowships and traineeships that provide multi-year support directly to students, targeting these to the areas of national need, which are the CS and CE degree programs. Doing so should stimulate enrollment by removing some of the financial uncertainty that makes many persons reluctant to embark on a five-year doctoral program, particularly those who face the repayment of student loans taken to finance an undergraduate education.

Richard Kieburtz is a Professor of Computer Science & Engineering at the Oregon Graduate Institute. He returned here after a two- year assignment as Division Director for Computer and Computation Research at the National Science Foundation. He can be reached at dick@cse.ogi.edu.