Kent K. Curtis
National Science Foundation
Washington, D.C. 20550
Factors influencing the supply and demand of computer manpower are analyzed in the context of available data on scientific manpower including statistics on degrees awarded in various discipline at the Bachelor's, Master's, and Ph.D. levels, faculty mobility, job mobility among professionals, starting salary trends, comparative unemployment statistics and economic projections. It is found that there is a shortage of computer manpower which is expected to persist for the foreseeable future but that society is responding, perhaps as rapidly as possible, to provide the trained people required by business, industry and government. Only the educational institutions of the U.S. have what might be described as a crisis, a staffing problem which seems to have no solution within the context of normal supply/demand forces.
Although the subject of computer manpower supply and demand is of intense interest to many people, there are no experts. This is surprising because the gathering of statistics on all kinds of subjects is one of our national pastimes; yet, in this area the statistics we have in great abundance do not lead to any clear conclusions. Why?
For one thing, the question is ill-defined. There is no consensus on who should be included in the manpower pool being studied. The boundaries of every occupation are fuzzy, of course, but in the computer related occupations we don't even know where the center of gravity lies. The Bureau of Labor Statistics counts people according to occupational classifications which are approximately the same as the Federal government's job classifications. These include titles such as computer specialist, systems analyst, programmer, computer technician, etc., each including a wide range of educational levels, and the size of each group is so large (hundred of thousands) the academic research community cannot be identified in any of them. The BLS projections are interesting, provocative, and display useful trends but they are not helpful for analyzing the problems of computer science faculty in universities.
John Hamblen, who has made several surveys of computer science-related programs in educational institutions of the United Stated has provided a very valuable service, especially for the supply side of the question. However, his surveys do not include the many people who go into computer related jobs after graduating in physics, mathematics, economics, or many other disciplines. His summary statistics also need careful interpretation when they are applied to the problem of university faculties because he includes programs in Data Processing, Information Processing, and other titles, which expand his population of academic programs beyond what most computer scientists and computer engineers acknowledge as their domain of interest.
The National Science Foundation, which has a congressional mandate to monitor the supply and demand of engineers and scientists in the United States , only recently began to count computer scientists as a separate group. Until then they had only one entry called computing theory as a sub-category of mathematics. Even now, in classifying academic departments and gathering departmental statistics, they label all departments which have compound names, e.g. Electrical Engineering and Computer Science, by the first adjective. Thus, the departments at MIT or Berkeley make no contribution to NSF's statistics on computer science departments because they appear only in the numbers for Electrical Engineering.
On the other hand, the annual survey of departments done by Professors Samuel Conte of Purdue and Orin Taulbee of the University of Pittsburgh  restricts its domain to Ph.D.-granting departments which carry the label of computer science. As a result, although it includes some departments which are overlooked by the NSF survey, it omits all of those which are labeled computer engineering or are computer science options in departments of mathematics or business administration. In brief, all of the available data needs a great deal of interpretation, interpolation or extrapolation to make it useful for a study of manpower dynamics. This leaves every analysis open to question.
Another reason for confusion is that the dynamics of the computer manpower supply and demand are without precedent. We are experiencing an almost continuous change in the all-pervasive technology of information processing which creates a sudden, almost infinite demand for specialized talent. The industrial revolution which caused such tremendous social upheavals in the western world proceeded at a snail's pace by comparison. As a result, history offers very limited guidance in the analysis of the present situation and gives no help whatsoever in predicting the immense transients that may be induced by such a strong forcing function as the computer revolution.
Nevertheless, laying all of these caveats aside, some interesting observations can be made which offer food for thought. In this paper, some data will be shown which exhibit convincing trends about supply, although not reliable absolute magnitudes about either supply or demand, and some conclusions will be drawn which offer a plausible framework for understanding the data and explaining the phenomena we are observing. These observations and interpretations have been exposed to several industrial managers and academic administration who have found them to be consistent with their experience.
What evidence do we have that there is a shortage of computer professionals? The most prominent is the anecdotal evidence which abounds. Everyone has heard many of the same kind of stories, and they need not be repeated here, but for the record let me show some data which support some of the most common.
It is interesting at this point to ask what are the factors influencing the career choices made by faculty and students when they are considering whether to enter industrial or academic careers? Fig. 7 shows the factors uncovered by the NSF survey of faculty mobility. It is noteworthy that salary was not the most compelling reason. Instead, "institutional disincentives" were. These include, of course, the uncertainty of tenure, a problem unique to academic institutions, which creates a feeling of job insecurity and frustration among young people who are starting academic careers. Other types of "institutional disincentives" cited most often were the difficulty and hassle associated with getting travel funds for professional conferences or other types of support for leading normal professional lives, heavy teaching loads because of the great influx of students into computer course and inadequate support of research, especially research in computer software or hardware systems. Few university administrators have experience with that kind of research which requires a high degree of support for facilities, time, and effort but is slow in yielding papers which can be published in the normal scientific literature. Universities have difficulty evaluating such work and this influences allocation of scarce university research funds or considerations of promotion. Many young computer researchers perceive industry as more hospitable and more rewarding for that kind of research.
Other anecdotal evidence of a shortage of computer manpower includes stories of intensive but often frustrating recruiting by industrial and academic organizations (just look at the ads in any journal or newspaper) and the fantastic salaries being offered. Many students, especially those with some relevant experience, have received amazing offers and some people have doubled their salaries by changing jobs. Statistical evidence to back up these stories is also available. Fig. 8 shows the increase in starting salaries for baccalaureate degree graduates during the period from 1974-1979. Starting salaries are used for this comparison because they are more flexible than salaries for established professionals and tend to be more sensitive as a barometer of hiring pressure. Fig. 9 shows that unemployment has been singularly low for computer professionals. Fig. 10 shows that computer science has had a higher growth rate in employment than any other field. Fig. 11 brings this data up to date by showing that only computer science had increased employment opportunities in 1982. Finally, a recent study by NSF, Fig. 12, shows that this imbalance in supply and demand is expected to persist for computer-related professions for the foreseeable future, even in the face of pessimistic economic assumptions. In fact, only the computer profession and aeronautical engineering show strongly encouraging prospects even under assumptions for the future of the economy which are very favorable to high technology.
Finally, let me note a rather general worry concerning education in the United States. Fig. 13 shows a suggestive correlation between trends in world trade and trends in technical education. One must be careful not to oversimplify, but this table is provocative. Something is out of balance which will redound to the long term disadvantage of our country.
Granted that there is an imbalance between supply and demand for computer professionals which is likely to persist for many years, how does society respond? We see part of the response quite clearly on our college campuses. Fig. 14 shows the total number of degrees awarded at the bachelor's, master's, and Ph.D. levels for the last several years. There has been almost no change. A comparison with the demographic data shown in Fig. 15 shows recent degree data to be consistent with demographic data. Assuming that consistency continues, we should expect a decrease of 20% to 25% over the next 15 years in the annual production of degrees. However, let us look at some of the fine structure. Figures 16 to 37 show data on degrees awarded in each of several fields of science, engineering, the arts, education, and the humanities for the last several years. Obviously there are large migrations in student interest toward engineering and away from education, letters and the humanities. Now look at Fig. 38 for computer science. That is true exponential growth, a more rapid increase in bachelor's degrees than in any other discipline. Undergraduates are swarming to major in computer science. Master's degrees are also up by almost 60%. These changes are imposing immense stresses on the institutions of higher education, causing large imbalances in teaching loads, a huge shortfall in capital funds for instructional and research equipment and a mad scramble for faculty slots and persons to fill them as they become available. Many institutions are being forced to limit enrollments in computer science, an action which is difficult to accept in either private or public institutions because of the traditional assumptions of our country concerning universal educational opportunity. Indeed, in this discipline we are being propelled toward a system of limited educational privilege, comparable to the "numerus clauses" of Germany, for example, which carries with it the potential for long-term problems, already evident in Germany, of unemployment or under-employment among educable youth due to lack of opportunity for appropriate education.
It must be noted at this point that there have been rapid increases in demand for education in some other fields during the 1950's and 1960's. Note, for example, bachelor's degrees in some social science, biology and business. However, total college enrollment and federal support for education were increasing very rapidly during those periods, also. As a result, enrollments in all fields were increasing and the changes were primarily differential rates of increase into some fields. These were readily absorbed by the quite elastic, expanding university resources which existed at that time. Now, college and universities have static, soon to be declining, enrollments, and limited, inelastic resources, but are confronted with shifts of student interest of unprecedented proportions. It is no wonder that our educational institutions are feeling stress.
Setting these institutional problems aside for a moment, however, we see that one of the natural responses of our society to the shortage of computer professionals is to expand the number of students being educated for the computer professions as rapidly as our educational system can adjust. This is happening as a result of natural selection by students of preferred courses of study and is having a significant impact on the supply side of the balance.
Another indicator of social response is shown in Fig. 39. This shows, for 1978 bachelor's degree or master's degree recipients, the ratio of people employed in each field in 1980 to those who received a degree in that field in 1978. For example, chemistry shows a slight gain at the bachelor's level but none at the master's. Engineering comes out even at the bachelor's level but has a marked decrease at the master's level. The computer professions, on the other hand, are sweeping people up from all disciplines like a giant vacuum cleaner, with people being retrained on the job or through part-time education as necessary, but moving to fill the void.
These effects are just what one should expect in a free market. Demand is creating supply by utilizing the job mobility of young professional people. Such job mobility has surely been facilitated by the computer experience which many college students now get as part of their normal undergraduate education regardless of their major subject, and NSF's programs during the sixties to help colleges and universities acquire computing facilities for research and education did much to make that possible. It will be even further increased as personal computers become integrated into the educational environment of every student. These are natural economic and social processes which are progressing rapidly, and it is not clear what government can do, if anything, to facilitate them. It is true that capital funds for universities and colleges to acquire equipment are one bottleneck, but we hope that may be relieved by recent changes in tax laws which are intended to encourage donations of equipment to academic institutions. Obtaining qualified faculty is a greater and more imponderable problem, but faculty are being retrained and faculty slots are being reallocated. Indeed, the academic environment is changing, perhaps as fast as the inertial factors of tenure and the highly specialized nature of academic jobs allow, so that it is not clear that government action could have much leverage. Perhaps steps to increase computer literacy at the junior or senior high-school level would help relieve the service-course teaching burden in colleges and universities but that movement, too, is already in progress at a rapid rate through natural processes which are subject to their own social inertia.
In brief, the natural working of free-market conditions for computer manpower supply and demand is leading to an increase in supply which is suitable for most purposes. The supply is not increasing as rapidly as it could be absorbed but perhaps as fast as society can respond, and there is nothing properly termed a computer manpower crisis in business, industry, or government. Acceptable people are becoming available in steadily increasing numbers and the only real effect of the shortage is a somewhat inflated price for labor, a differential which provides the fuel that powers social adjustment.
The situation in the academic world is quite different. Academic jobs are so specialized and require such long periods of education and training that job mobility is very low  and each class of academic institution can draw upon only one source of supply for labor. Elementary and secondary schools can use only people with proper education credentials to satisfy state and local certification standards. As the statistics on degrees in education show, that is a rapidly diminishing pool. Institutions of higher education, on the other hand, are restricted to the pool of Ph.D.s for their labor supply and in computer science that is a small supply indeed, far out of balance with the expanding need. We must conclude that the educational institutions of the country cannot obtain the labor they need and have poor prospects of finding it in the near future. They face a real crisis. The migration of student interest is working against them, not in their favor, and the job mobility which allows so many people to enter computer professions in business, industry and government is not effective for educational institutions because of their highly specialized job requirements. Let us analyze their dilemma somewhat further in the next section.
Let us consider the conundrum facing the computer field in higher education first. It is experiencing an exponentially increasing demand for its product with an inelastic labor supply. How has it reacted? NSF has made a survey of the responses of engineering departments, including computer science departments in schools of engineering, to the increasing demand for undergraduate education in engineering. There is a consistent pattern in their responses and the results can be applied without exception to the computer field whether the departments are located in engineering schools or elsewhere. 80% of the universities are responding by increasing teaching loads, 50% by decreasing course offerings and concentrating their available faculty on larger but fewer courses, and 66% are using more graduate-student teaching assistants or part-time faculty. 35% report reduced research opportunities for faculty as a result. In brief, they are using a combination of rational management measures to adjust as well as they can to the severe manpower constraints under which they must operate. However, these measures make the universities' environments less attractive for employment and are exactly counterproductive to their need to maintain and expand their labor supply. They are also counterproductive to producing more new faculty since the image graduate students get of academic careers is one of harassment, frustration, and too few rewards. The universities are truly being choked by demand for their own product and have a formidable people-flow problem, analogous to but much more difficult to address than the cash-flow problem which often afflicts rapidly growing businesses. There are no manpower banks which can provide credit.
What flexibility does our society have to cope with this dilemma? Retraining existing faculty can be done only to a small degree. Faculty slots are being reallocated as they become available but that is a slow process in a period of inflexible or declining budget resources. Even so, universities cannot fill all of the slots they have available. The net result is that the lack of lateral job mobility among faculty from one discipline to another coupled with the small supply of the only type of labor universities and colleges can use forces them to become less efficient and effective producers when confronted with strong shifts in market demand for degrees. They become burdened with people whom skills are not in demand, which increases their nonproductive costs; at the same time they overwork people whose skills are in short supply, which threatens the quality of their educational product. Some people have begun to doubt whether universities can play, in computer science, their traditional role as the supplier of trained professional manpower. As a consequence, there are moves to establish specialized technical schools such as the Wang Institute or to rely more heavily on on-the-job training. If one could argue persuasively that this is a temporary or transient phenomenon, one might make a compelling case for federal government intervention to buffer the nonproductive extra costs of institutions while they adjust to new conditions. In the absence of a sufficient pool of qualified persons to hire as faculty, however, it is difficult to make that case.
On the other hand, if universities and colleges were made more attractive place of employment, they could compete more effectively for the available talent and motivate more students to prepare for and choose academic careers. It is true that this question of attractiveness is a relative one which is influenced by economic conditions. Poor economic conditions in business and industry make universities relatively more attractive but the problem in the computer field is not relative, it is one of absolute numbers and we need a long-term, stable solution. Our NSF survey has indicated the factors to be addressed in any attempt to accomplish this. The "institutional disincentives" are most important and can be modified only by the institutions themselves--either more manpower or else limitations on enrollments to reduce work loads; more institutional financial resources allocated for staff travel, secretarial support and other normal professional needs to reduce harassment and improve morale; more effective, or at least more apparent, consistency between the work required of faculty, namely, teaching, research and participation in committee and the criteria used for promotion to increase the sense among young faculty that they work in a rational system with reasonable job security. Of these, the availability of manpower is the factor which seems to be beyond institutional control the most. Yet if the institutions cannot control this factor, neither can anyone else, so at least we must examine it from the perspective of the university. Clearly, the objective should be to either expand the manpower pool which universities may draw upon or else reduce their manpower requirements. Many universities have already expanded their pool in one way by turning to temporary or adjunct staff to help with the teaching load. If the manpower shortage is a transient problem, this is surely the best way to deal with it. It is arduous--one university I visited recently is recruiting and managing 35 to 40 temporary teaching staff each year drawing upon local industry--and creates severe problems of quality control, but the educational enterprise can probably be continued in this mode for some time if there is hope of improvement. From the data at hand, however, it is not clear how equilibrium between supply and demand of permanent staff can be realized in less than a generation which is a long time.
Another possibility would be to reduce manpower requirements by using technology both to make education more efficient and to move some aspects of computer training out of the universities into pre-college curricula. This approach has limited promise. It would clearly help reduce the demand for service courses in computer literacy and programming, but the hard-core dilemma concerns majors in computer science, not the brood community of users. We know very little, now, about applying technology to make education in advanced courses more efficient and it may take at least a generation to learn.
A third possibility would be to broaden the university community so it can draw upon a larger fraction of the existing manpower pool. From a labor management point of view, universities are the most homogeneous, restricted, single-purpose organizations imaginable. They will accept only one type of employee, namely a Ph.D., and have only one kind of job, i.e., a combination of teaching and research. That is traditional, but given the present dilemma one must ask is it necessary? Are there other ways of doing things which would utilize a broader range of talents, provide more diverse career opportunities and hence make universities attractive places of employment for a larger class of people? If so, a stable equilibrium between manpower supply and demand might be reached sooner.
In this connection, it is useful to consider the case of business and management curricula. Fig. 40 shows the number of degrees awarded in business during seven recent years. It is clear that bachelor's and master's degrees are increasing but note, especially, the very small number of Ph.D.s, approximately one doctorate for every 200 bachelor's degrees. This ratio can be sustained because professional curricula have been defined for business which draw upon the course offerings of many parts of the university and do not depend upon the Ph.D. faculty of one discipline to carry the teaching load. Defining a professional curriculum for computer science might be useful, too. Such a curriculum could recognize that many of the undergraduate technical courses in computer science do not need to be taught by Ph.D.s but could be handled equally well by high quality master's level people who do not, necessarily, have a research orientation. If such positions were given first-class status in universities, new career paths would be open and a large labor pool would be available. Many of the undergraduate and master's degree students might find a professional curriculum as much to their taste as the present core computer science curriculum with a corresponding reduction in number of majors in the latter. This would help bring supply and demand into better balance.
The second most important factor influencing the university environment was found to be research opportunities. In rapidly changing high technology fields it is extremely difficult for an institution's resources to encompass its requirements for capital equipment and operating costs for either education or research. This is a factor the Federal government can influence, however, and NSF has designed its Coordinated Experimental Research Program in computer science specifically for this purpose. As you know, our CER program reaches computer research no matter whether it is located in Schools of Engineering or Schools of Letters and Science and the ten awards made are split about half and half. Federal tax legislation can help contribute to the improvement of university environments, also, by encouraging the donation of equipment and money to educational institutions. Legislation passed in the last 18 months is intended to make tax incentives more effective for this purpose.
The third most important factor is salary and this, again, is an institution decision. Differential salary scales which provide preferred treatment for computer science and engineering faculty are becoming commonplace. As a result, it is apparent that the universities are adjusting in this dimension, albeit at the price of some internal stress and discontent.
The preceding analysis shows that the United States has an imbalance in the supply and demand for computer manpower which is expected to persist for the foreseeable future but that our society is adjusting rapidly to redress that imbalance in every sector except education. In education, the channels for adjusting supply and demand which are available to business, industry and government, principally mobility of personnel among job and shifts in the interests of students training for jobs, are ineffective and we cannot hope for short term solutions. This places our educational institutions under great stress and could lead industry to develop alternate methods or new institutions to provide the professional computer personnel it needs if the traditional educational institutions are unable to adjust. One can imagine fundamental changes in the structure and management of institutions, both of higher and pre-college education, which could help resolve the present dilemma but not without altering time-honored traditions. No other technological advance in history has been so rapid or so compelling in confronting education with the hard choice of either embracing fundamental change or accepting a reduction in its traditional role of training students for the future.