The United States IT Workforce Shortage

Howard is a tenured full professor and chair of the Department of Computer Science at Hunter College of the City University of New York, CEO of Rubin Systems, a faculty affiliate of Cap Gemini America, and META Group research fellow. He is also editor and publisher of IT Metrics Strategies, in which a version of this article first appeared. Howard can be contacted at [email protected].

With all the attention being devoted to the Year 2000 computer crisis and the European Monetary Unit conversion, it is almost no surprise that a far larger crisis -- the information-technology (IT) workforce labor shortage -- is being overshadowed. In this article, I'll provide a multidimensional view of IT workforce labor shortage issues, drawing heavily on the findings of my own 1995, 1996, and first-half 1997 Worldwide Benchmark Project findings, and blending in the work and words of others, including the insights of Norm Brown and Frank McGrath of the Department of Defense's Software Program Manager's Network, the research and position paper produced by John LaFrance of the Department of Commerce, the investigation into this area performed by the Information Technology Association of America, and the data resources of the Stanford Computer Industry Project (SCIP). Among the many other groups that have begun to analyze these issues are the National Research Council, the Software Publisher's Association, the Computer Research Association, and various professional information technology/computer science-oriented societies.

If I had to provide a one-line summary of all the viewpoints, it would be: The shortage is real and its impact will dwarf such techno-issues as the Internet, e-commerce, and HDTV that are the focus of media attention today.

If I were also asked to provide a parallel one-line summary of what needs to be done, it would be: At an internal corporate IT level, focus on retention and recruiting strategies now; at a corporate business level, focus on the emerging new economics of IT; at a national policy level, enact an "endangered species act" -- the species is software professionals.

The Big Picture

There is no longer any separation between information technology and business. While the distinction between IT and business is definitely blurred, the linkage to productivity and business is clearly there. The U.S. Bureau of Labor Statistics (BLS) appears to support this view. According to the BLS,

Computer technology is altering the form, nature, and future course of the American economy, increasing the flow of products, creating entirely new products and services, altering the way firms respond to demand, and launching an information highway that is leading to the globalization of product and financial markets.

There is no question that there is a definitive linkage, albeit virtually uncharted, between technology and business performance. Scale this up, and the linkage moves up to economic performance and global competitiveness.

The Basic IT Issues

In U.S. corporations, the annual thirst for information systems is growing at 25 percent per year. While this number may seem high, consider the following change rate of information technology itself: For every two to three years that goes by in the 1990s, we have seen as much technology infusion in business as in all the 1980s. By 2005, each year will be the equivalent of the 1980s twice over in terms of information technology acquisition.

However, raw technology is not enough to provide business leverage and opportunity. It is the systems and software that activate the hardware, and they are a commodity that cannot be mined or manufactured today through automated means or mechanisms. Systems and software are developed by people -- and the right people are in short supply.

From tech journals to BusinessWeek, the IT labor shortage is finally attracting some degree of attention. Before discussing the numbers and labor force demographics of the shortage, here's a summary of the basic issues that the nation is facing in this regard:

• There is an acute shortage of software professionals in the U.S. -- the IT workforce.
• The size of the graduating classes of U.S. universities that would typically be expected to fill newly created positions for software professionals indicates that the demand for this labor category will not be met in the near term from these ranks.
• The pool of graduates from U.S. universities has the wrong skill set to meet the needs of U.S. industry.
• The pool of graduates from technical programs in the U.S. is also too small and under-skilled to meet demand.
• The global competitiveness of the U.S. software-development community is weakening.

Perhaps the biggest single issue has to do with the potential impact of the IT workforce shortage on the economy. Because of the fact that information technology and business are so interleaved/interwoven today, the workforce shortage may act as a braking force on the economy. Some simple numbers illustrate this point in two ways:

• The impact of the shortage on the ability of businesses to expand and remain competitive.
• The cost structure of doing business in our information-based economy.

One estimate of the shortage is that 200,000 jobs currently remain unfilled. Based on the results of the 1996 Worldwide Benchmark Project, the amount of software produced by the 200,000 professionals if the vacancies were filled could support an additional $500 billion in corporate revenue and $10 billion in associated income annually.

In terms of the cost structure of doing business, the labor shortage appears to be impacting and inflating the salaries of the existing IT workforce. While the Department of Labor claims that the average U.S. worker is getting a 4.1 percent increase each year, those in IT will soon average 20 percent or more. As a result, corporations will be anteing up a potential $10 billion in raises over what would normally be expected.

Other factors are important too. Add on the Year-2000 computer crisis impact on the IT workforce's workload, the need for Year 2000 conversion skills, and the cost of the nation's Year 2000 efforts, and the picture gets even more complex. Look at all this in the context of the total economic picture and consider that the IT workforce shortage may turn out to be the main cause of economic braking by preventing corporations from building information systems to support their business needs.

Different Views of the Same Phenomenon

A number of forward-thinking groups have devoted significant resources to examining whether or not an IT worker shortage problem exists. What follows is a snapshot of their findings:

The U.S. International Trade Commission View. In its June 1996 report, the U.S. International Trade Commission (ITC) sized the global market for computer software and services at about $277 billion. The ITC also predicted the software and services markets would grow at 12 percent, and 11 percent annually at least through 1998. In the projected $420 billion 1998 market profile, the U.S. would account for 56.6 percent of the software side and 46.8 percent of the services side.

The ITC also identified five market segments in which the U.S. competes:

• Applications software.
• Systems software.
• Systems integration.
• Custom programming services.
• Outsourcing services.

According to the ITC, the U.S. is the most competitive in the areas of applications software, systems software, and outsourcing services. Primary differentiators are price and the range of offerings. With price and the ability to stay "current" playing such an important role, the ITC concluded the labor shortage may have a profound impact on the competitiveness of U.S. software and services firms.

The National Research Council View. The Office of Scientific and Engineering Personnel of the National Research Council (NRC) has been involved in ongoing activities regarding the nation's needs for computer professionals for more than five years.

Its basic premise is that:

...the vitality of the computing enterprise and its components -- the academic disciplines, the industries, and the specialized computer professionals -- and its growth and contribution to the economy make supply and demand for computing professionals a matter of national interest....a technically skilled labor force will remain essential to securing competitive position in the global economy.

The observations and recommendations produced by the NRC are:

• There is an evolving demand for computer professionals but employment opportunities are shifting across technology platforms and environments. Though there is clearly overall growth in employment, some growth areas clearly lead others (for example, distributed computing and the Web) while others may lag (mainframe computing, with the exception of the Year 2000, for instance). Interestingly enough, the NRC finds a restructuring in the area of computing research with budget pressures on central industrial laboratories. At the same time the required level of skills of those in the workforce seems to be increasing with demand fueled by technology change and the need for specialization.
• The talent stream must be broadened in light of declining undergraduate computer science enrollments. Both a broader ethnic base and a broader discipline scope is needed to support the demand for workers.
• There is a need for ongoing training for those already in the field. In short, the NRC concluded that "changes in skill requirements are a way of life for computing professionals." In this context, frequent training must be planned for and corporations must be prepared to bear the financial burden of such training in their plans.
• The educational model needs to be changed. With the rapid changing of needed skill sets, educational institutions must focus on producing graduates that are positioned for lifelong learning. In addition, the NRC believes that education in other fields should have a stronger and fundamental educational component in the computer field to enable those in other disciplines to better embrace information technology. Such a policy may enable those from the other fields to migrate into the computer field.
• The gap between corporate needs and education must be filled by improving the linkage between the education system and the workplace. New models of cooperation are needed; new models of skills development are necessary.
• There is inadequate data about the field in general and the software labor market. In addition, skill portfolios must be understood so that a vehicle could be provided for tracking shifts in skill needs and supporting the national labor needs.

The Information Technology Association of America View. The Information Technology Association of America (ITAA) is a nonprofit trade association of more than 11,000 companies in the software and services industry. The ITAA surveyed more than 2000 large and midsize companies in the U.S. to provide a view of the U.S. IT workforce position.

The basic ITAA findings indicated that approximately 190,000 IT jobs were unfilled at the time of this 1996 study. However, the ITAA views this number as an understatement because the survey did not include smaller companies or government agencies. In addition, those surveyed indicated that demand for IT workers was higher than for other worker types.

In terms of IT salaries, IT worker compensation is two to six times higher than the typical manufacturing worker, and IT worker salaries in many skill areas rose an average of 20 percent between 1995 and 1996 while the U.S. average for all fields was 4.1 percent.

It was generally acknowledged that finding skilled IT professionals was the most difficult recruiting task that companies are facing; the inability to find skilled IT workers was the greatest barrier cited to companies' future growth.

The Computer Research Association View. The Computer Research Association (CRA) is a scientific society whose members are either academic departments or industrial laboratories that engage in fundamental research. In supporting a Department of Commerce study of the workforce shortage, the CRA provided the following views:

• The demand for software professionals is growing rapidly and demand currently outstrips supply.
• Foreign nationals are not taking jobs away from Americans; they are filling vacant slots that the American pool cannot fill and are additionally contributing to the nation's storehouse of knowledge and production capability.
• The nature of jobs in the field and the skills needed to hold them are changing rapidly.
• Educational institutions need to sustain a stable long-term response to the growth in demand for graduates in the computer field.
• Immigration reform should be utilized as a mechanism for providing access to labor.
• The nation and companies should invest in research that improves the effectiveness, quality, and ease of production of software.
• Federal policy is needed to encourage steady and consistent growth in the area. Federal programs should encourage minorities and women to enter the field.

The Software Publisher's Association View. The Software Publisher's Association (SPA), as its name implies, represents and lobbies for the Microsofts, Oracles, and Novells of the world -- companies that are keenly interested in policies that affect software onshore and offshore. The SPA is well aware of the information age economy and how its dynamics are fueled by information technology used by knowledge workers. In fact, the SPA predicts that 60 percent of U.S. jobs will require IT proficiency by 2000. The SPA clearly sees a national need to transform the industrial age workforce into an information age one. They advocate a high investment in the high-tech education and training of students and workers to boost U.S. economic productivity, increase wages, and create opportunities for minorities.

To put their suggestions into an economic perspective, the SPA has attempted to put the software industry into a financial perspective:

• In 1987 the software industry generated $55,000 in output per employee versus $87,000 in 1992 -- this represents a 66 percent greater increase than in other industries over the same period.
• Average salaries rose from $37,000 in 1987 to $55,000 in 1992.
• While the GDP grew at an annual rate of 5.9 percent, the software industry has been growing by more than 21 percent.

The SPA's view is that education plays a pivotal role in the workforce transformation and that basic education processes need to be reengineered and redesigned at all levels. In addition, technology itself must be fully integrated into the learning process. To make all this happen, the SPA advocates a new wave of tax incentives and financing options.

The DoD/Software Program Manager's Network View. Clearly, computer software is at the heart of the U.S. Department of Defense's (DoD) strategy to maintain military superiority into the 21st century, while reducing the size of annual defense expenditures. Although the U.S. Federal Government no longer dominates the expenditures on software systems in the United States, it remains a major player. Federal government agencies have spent more than $200 billion on information technology over the past decade.

Data collected by the Electronics Industries Association shows an enormous annual increase in software expenditures by the Department of Defense. Figure 1 (supplied by SPM) shows the enormous increase in the software part of DoD systems, while the increase in the hardware part is about even with the rate of inflation. This enormous increase in software expenditure represents an enormous increase in the demand for software professionals since most of the cost of software development and maintenance is the cost of labor. The number of computer scientists employed by the DoD, let alone by DoD contractors, is very large. According to statistics from the National Science Foundation, 26,932 computer scientists were employed by the DoD in 1991. The total employment of computer scientist by all federal agencies in 1991 was 53,097. More computer scientists were employees of U.S. Federal agencies in 1991 than any other field of science and engineering. The field with the second largest number of employees was electrical/electronic engineering, with 35,795.

There is no question that the workforce shortage will impede the DoD's ability to develop and maintain critical systems. While offshore resources remain available in the short term, the feasibility and advisability of using such resources in support of military needs is clearly an important issue of national security.

The Stanford Computer Industry Project View. The Stanford Computer Industry Project (SCIP) itself is a testimony to the importance of IT and the IT workforce in national and global economics. This project was launched in 1991 with a grant from the Alfred P. Sloan Foundation.

In February 1997, SCIP held a global software labor roundtable to discuss both the economic consequences of the shortage and evidence of the shortage itself. The roundtable also addressed the issue of the shortage's seeming acceleration. Its conclusions were:

• The demand for software may be growing nonlinearly and thereby fueling the acceleration of demand for labor.
• Growth in the past was masked by the rise of the offshore labor force, while at the same time onshore downsizing was taking place.
• Interest in careers in the field has declined thereby depleting the primary new labor source.
• There is a weakness in data available on the issue.

  In terms of recommended policy directions, the SCIP roundtable advises that:

• Certification of software professionals should be explored.
• Investment in training and attracting new sources of software labor should be made by companies.
• Immigration from selected sources should be encouraged.
• Investment should be made in the public sector in training and attracting new sources of software labor.
• Research in software-development methods and technology should be encouraged.
• Data must be improved; better censuses must be conducted.

It should also be noted that the SCIP has contributed research that indicates that the labor shortage does not exist. In any case, SCIP continues to maintain an aggressive research program in the area.

The Department of Commerce View. The U.S. Department of Commerce (DoC), guided by John LaFrance, has done the most extensive work to date in collating and synthesizing a large body of work addressing the IT workforce shortage.

In his draft report of April 1997, John LaFrance put forth this series of recommendations:

The role of government:

• Bring together all parties to the problem.
• Create a steering committee to create a plan of action.
• Establish a working group to set into motion pilot programs.
• Form a long-term-focused National Software Competitiveness Council.

The role of universities:

• Make long-term curricula changes to meet the changing demands of the IT industry.
• Focus on lifelong learning and skill upgrades.
• Add more computer science and engineering content to basic education.
• Increase industry interaction.

The role of community colleges and trade schools:

• Follow proposed University role.
• Strive for a higher degree of industry interaction.
• The role of K-12 education:
• Upgrade the technical content of basic education.

The role of training programs:

• Follow the training program models of Ireland, Canada, and India.

The role of immigration:

• Adjust immigration laws to encourage access to this labor pool.

Perhaps the greatest success of the DoC program is the IT Workforce Convocation which is being supported by President Clinton and organized by the ITAA scheduled for January 1998. It is likely that many of the DoC recommendations will be considered for implementation by the Convocation team.

Quantifying the Shortage: Computer-Science Program Metrics

Computer-science programs throughout the nation have traditionally been a source of entrants into the software engineering field. However, the number of computer-software-related majors graduating from U.S. colleges and universities has been declining since 1985 to the point where only 35,160 degrees were awarded in 1992-93. These 35,160 degrees consisted of 24,200 bachelor, 10,163 master, and 805 doctorate-level degrees. More recent data shows this figure has declined to about 25,000 in 1995-96. Details of this decline in computer-related degrees in the U.S. are shown in Figure 3.

Data collected by the National Science Foundation shows that roughly 25 percent of the computer-software-related majors graduating from U.S. colleges are foreign citizens, and only about half of these remain in the United States. This reduces the number of new computer software graduates that entered the U.S. labor force in 1992-93 to about 32,600, and creates an annual inability to staff 165,400 new computer-software-related job openings in the United States.

Projecting from past and current trends indicates an increasing cumulative shortage of software professionals in the U.S., beyond the current shortfall, even if the number of graduates with computer majors stops declining and remains steady at the 1992-1993 level through the year 2005 (see Figure 2). Clearly, steps must be taken to prevent this from happening.

Quantifying the Shortage: Job Market Metrics

According to the Bureau of Labor Statistics (BLS), in 1994 computer scientists, engineers, and systems analysts held about 828,000 jobs in the U.S., and computer programmers held about 537,000 jobs. These figures are somewhat lower than the Worldwide Benchmark Project approximation of 1.9 million to 2 million professionals employed in such positions currently.

By the year 2005, the BLS projects 1,403,000 jobs for computer scientists, engineers, and systems analysts, and 723,000 jobs for computer programmers. This is an increase in the number of jobs over 11 years of 69 percent for computer scientists, engineers, and system analysts, and 35 percent for computer programmers. These projections are for new jobs. Tens of thousands of additional jobs will have to be filled each year to replace staff members who retire or leave the occupation for other reasons. The first-half 1997 Worldwide Benchmark Project results show the turnover in the U.S. to be in the 10-15 percent range, while the worldwide figure is in the 7-10 percent range.

The BLS also predicts that the annual average number of job openings in the U.S. due to new job creation and replacement needs from 1996 and 2005 will be 125,000 for computer scientists, engineers, and systems analysts, and 73,000 for computer programmers. This is a total annual demand for 198,000 new software professionals entering the U.S. labor force.

As an indication of the very serious shortage of software-skilled persons in the U.S. today, consider that on one day in February 1997, there were at least as many software jobs advertised as the entire output of software-degree graduates from U.S. colleges and universities in a full year. This large number of software job openings in one day is not an aberration. The number of job openings on any one day or week has been this high over the several months during which Frank McGrath of SPMN tracked such data.

To rapidly assess demand, McGrath performed a query on one of the largest listings of job openings on the Internet -- the Internet On-line Career Center. The query focused on C, C++, SQL, Java, Informix, Windows, Powerbuilder, and CICS skill sets. Table 1 shows the number of job ads found from this query from http://www.occ.com/ on three different dates.

The Census Bureau reported that computer and data-processing services in the U.S. added 14,000 jobs in the month of October 1996. If, based on the previous query on December 12, 1996, there were almost 27,000 software job ads listed in only one source of job ads, and many of these ads offered more than one job opening, and many of the software job ads in this single listing were not found (since not all software job skills were included in the query), then the 14,000 software jobs added in October probably fell far short of filling all the software job openings in that month.

Table 2 shows the number of jobs in the major segments of the U.S. economy in 1996 as determined by the Census Bureau. According to this same Department of Census publication, there were 1,220,000 persons employed in computer and data processing in early 1996. These jobs are part of the total number of services jobs shown in Table 2 in the same 1996 publication. The Census Bureau also reported the number of new jobs in each of these segments of the U.S. economy in October 1996. The rightmost column of Table 3 shows these numbers, and the center column combines this data with numbers from Table 2 to compute the number of new jobs as a percent of the total number of jobs in each of these segments of the economy. It is reasonable to make two conclusions from this analysis:

• The 13.8 percent growth in computer and data processing jobs means the market is trying to improve productivity and quality to compete in world markets.
• The 0.0 and 0.8 percent, 2.2 and 3.5 percent growth in manufacturing, government, construction, and retail shows that those segments, which offer most of the jobs to lower and middle income persons is not sufficient to provide new jobs and improve the standard of living into the future.

According to The Wall Street Journal, a 13.8 percent growth in software jobs represents a continuation of a steady increase in the rate of growth of software jobs since 1987. The Wall Street Journal reported the following rates of increases of software jobs in the U.S.:

• Between 1987 and 1994, 9.6 percent per year.
• In 1995, 11.7 percent.

During this time, the number of degrees awarded to software majors by U.S. colleges and universities dropped by 30 percent.

Clearly, the job market demographics support the existence of a worker shortage both empirically and statistically.

Quantifying the Shortage: Compensation Metrics

According to the March 1996 Supplement to the Annual Demographic Survey, jointly prepared by the Bureau of Labor Statistics and Census Bureau, the median household income in the U.S. is $34,076 and 70 percent earn less than $80,000 per year. In comparison, Table 4 shows annual salary information for software job openings advertised on the Internet. These figures show that a very high percentage of software job openings offer more than $50,000 per year and that many software jobs offer more than $80,000 per year. These yearly salaries for software jobs give individuals in the field incomes that exceed the yearly income of 70 percent of the households in the U.S., many of which have incomes from more than one person.

The salaries in Table 4 are for full-time employees with company benefits. However, a new mode of employment for software labor has materialized in the marketplace rapidly in the last few years -- the self-employed contractor who is hired at very high hourly rates to work on specific software development projects and then leaves to work on another project for another company. It can be hypothesized that these very high hourly rates result from market demand and the fact that most people believe that the company will summarily lay them off in the future if it is a short-term advantage for the company to do so.

High demand for software skills with corresponding high salaries is not something that happened abruptly in 1996. It has been building up for a number of years as evidenced by an article in the March 1995 issue of Money magazine. This article ranked jobs on the basis of a number of factors. The most heavily weighted factors were near-term demand, projected far-term demand, and compensation. This job ranking concluded that the first and second best jobs in the United States in 1994-95 were computer engineer and computer software-systems analyst, respectively. A computer engineer was defined as a person who designs and develops computer hardware and software. This March 1995 article also stated that computer engineer was the third highest paid occupation in the U.S. in 1994-95, trailing physician and airline pilot.

Quantifying the Shortage: Offshore/Immigration Metrics

Another interesting phenomenon is that the United States is to some extent the beneficiary of a computer science brain drain from foreign countries. Statistics from the National Science Foundation show that in April 1993, 13.6 percent of BA/BS, 29.0 percent of MA/MS, and 39.4 percent of Ph.D. computer-science graduates in the U.S. workforce were foreign born. Some of these figures are the result of the fact that roughly 2600 foreign citizen computer-science graduates from U.S. colleges and universities remain in the United States after graduation. Another factor is the H1B visa program that gives six-year nonimmigrant visas to persons in high-skill occupations in fields for which a major shortage has been certified to exist in the U.S. workforce. In 1994, it was certified that 64,426 openings in computer-related fields were eligible for H-1B visas. The number of openings for computer-related fields was larger than any other field except physical therapy. This, in itself, is evidence that there is currently a serious shortage of computer-software professionals in the U.S. labor market. (Although the H-1B program was recently expanded for IT workers, it remains under attack and will probably be reduced because some say that it takes jobs away from U.S. citizens and gives them to low-wage temporary immigrants.) There is also anecdotal evidence that foreign-citizen computer software workers in the United States are beginning to return home as computer software companies start to grow in their homeland. This is particularly true for Indian and South Korean citizens.

Global Competition

Figures 4 and 5 provide approximations of the size of the software professional workforce in a number of nations around the world and projections as to how the global levels of software production will be changing as we approach the new century in response to labor and performance factors. As can be discerned from industrial age lessons, the "balance of production power" may be changing in the information age. The key drivers are the availability of labor and worker productivity. Some insights into the shifting global dispersion of software production can be seen by examining the rise of India and Ireland.

India is perhaps the best example of a third-world emerging economy that has entered the information age. According to Forbes magazine, "India is poised to become a world leader in computer software." According to the Economist, "There are parts of the Indian economy that are skipping right over the "second" industrial age, and straight into the information era. India's software industry, which barely existed 10 years ago, notched up sales of more than $1.2 billion last year and is growing at over 40 percent a year. Unlike other Indian industries, it is highly competitive internationally. Around half the industry's revenues come from exports." India produces 20,000 computer-science graduates a year but demand is already so high that the industry's wages are rising by 20 percent a year. There currently are about 130,000 highly skilled software developers in the Indian labor market and starting salaries are less than $9600 a year or about one-fifth of what a comparably skilled U.S. programmer makes; the economy of Indian labor in this industry is also evidenced in associated unit costs of software production (see Figures 6 and 7).

Software developers in India are not only low-cost, they also have the leading-edge technology skills needed to develop software on computer networks. Ironically, this is due in large part to the fact that India could not afford mainframe computers in the 1970s and 1980s, and had to make-do by connecting low-cost personal computers and workstations into a network which has turned out to be the technology of the 1990s and the 21st century.

Ireland provides another example of foreign competition in software. Ireland has established four economic sectors to be at the foundation of the emerging economy of Ireland: software, microbiology, agriculture, and tourism. The government of Ireland has done a number of things that provide strong support to these industries.

In contrast to the United States where the number of software graduates has been declining since 1994, increasing numbers of college students in Ireland are majoring in software. The chairman of the software engineering department of Trinity College in Dublin, Ireland, stated last September that half of the engineering students in this most prestigious university in Ireland are majoring in software engineering.

Systems Dynamics of the Workforce Shortage

There is growing evidence that the IT workforce shortage is a result of what is known in the engineering community as "degenerative feedback." Degenerative feedback occurs when the response of a system to an output of the process drives the system to deteriorate. The workforce shortage is a driver of such a feedback loop in terms of its influence on staff turnover and job retention. This two-edged sword will impact both the organizations that the professional is leaving and the one that the professional is joining.

Fundamental "peopleware" statistics show that when a person -- even one highly educated, experienced, and competent in software development -- is added to the development team for large software systems some time after the beginning of the development, this person functions at a level of much lower productivity for months while learning the details of this particular software project. During this time, not only do these people function at substantially lower productivity than they will six months later, but they are a nonproductive drain on the time of the persons who have been on the project for some time who must spend time teaching the newcomer. Thus, when the forces that drive software development in the U.S. cause software developers to stay with one company and software development project a steadily increasingly short amount of time before they voluntarily quit and take a new job with a different company on a different software development project, there is degenerative feedback. What are the mechanisms that motivate skilled software staff to quit at a rapidly increasing rate? A severe software labor shortage combined with a supply/demand market economy, results in rapid increases in the compensation companies are willing to pay for skilled software developers.

This phenomenon has strong implications for compensation strategies in the field. Well-intentioned labor regulations in the U.S. make corporate legal and human resource executives very reluctant to allow the percentage of the annual raises given to workers to be significantly different for different employees for fear of a law suit from the employee getting the lower raise. The result is that the only way that a typical highly productive software developer with a high-demand specialization can get a substantial raise is to job hop -- that is, to quit and go to a different company that can offer a big increase in salary with much less danger of being taken to court by a disgruntled employee.

The degenerative feedback in the context of the workforce shortage is that anticipated decreased time between changing jobs driven by opportunities for greater compensation results in lower productivity, which means more persons are needed to develop a given amount of software. The market then responds to this by further increasing the compensation for software-skilled persons, which then feeds back to further shorten the time between changing jobs. This degenerative feedback resulting in the need for more people to develop a fixed amount of software is in addition to the increased demand for software-skilled persons due to the rapid increase in the amount of software that our society needs to remain competitive in world markets.

Recommendations: A National Agenda

A joint government, industry, and university alliance must be organized to address the IT workforce shortage. A National Software Competitiveness Council (NSCC) should be formed with the following charter:

• Identifying underlying factors that affect U.S. software competitiveness.
• Developing a national strategic plan.
• Maintaining information for competitive calibration and benchmarking of U.S. software competitiveness.
• Defining and coordinating potential changes to policy, regulation, and law.
• Fostering government, industry, and university collaboration in the development and implementation of academic programs supporting U.S. national needs.
• Communicating and enhancing awareness of the issues challenging U.S. competitiveness in the information age.
• Helping develop measures of global competitiveness for the information age.

Recommendations: What Organizations Must Do Now

With all the evidence presented herein, it should be clear that the IT workforce shortage is real. The increasing demand for business systems, the coupling of technology and business, the high rate of technology change, and demands of the Year 2000 crisis will continue to be the major forces driving the need for additional labor in a market where there may be none to be found through traditional means. Therefore, to manage the risk of the labor shortage while keeping an enterprise moving forward in the information age, a few key critical actions should be to be taken now:

• Develop and nurture new sources of labor through university alliances, creation of cross discipline training/retraining programs, and global labor sourcing.
• Reformulate internal compensation and reward programs to keep high-value staff, attract new staff at rates which make economic sense, and binding employees to the company; reward needed skills and behaviors and use core staff to attract new staff.
• Manage the skill inventory as an active portfolio; use forecasting and projections; and invest in redevelopment and retraining.
• Focus on increasing the leverage of the existing staff; higher productivity and quality through reuse; capturing and disseminating knowledge; reducing rework.
• Leverage external knowledge in the form of packages, systems outsourcing and business process outsourcing.

And most importantly in this fluid age of technology and people, maintain an aggressive program of benchmarking, performance calibration, and performance management. Watch turnover, retention, and "peopleware" metrics -- these may be the leading indicators of your future performance.

Acknowledgment

My thanks to Frank McGrath for his enormous contribution to this paper. Much of the data and views relating to the job market and skill-sets were drawn from his work on the software labor force in the U.S. Frank can be reached at [email protected].

DDJ