Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 92
Building a Workforce for the Information Economy 3 Characterizing the Workforce Problem 3.1 FRAMING THE PROBLEM IN CONTEXT Issues related to the IT workforce are controversial. To better understand the issues at hand, it is helpful to start with what has become the most common formulation of the problem: IT employers of all kinds report having considerable difficulty in hiring workers for their open IT positions. A shortage of IT workers is inferred from these difficulties. Remedial actions are taken and/or proposed to increase the supply of IT workers. There is much about this common formulation that is useful and valid. But each element of the logical chain must be scrutinized carefully. At the end of careful analysis, a much more nuanced and qualified picture emerges. The third point concerning remedial actions is the subject of Chapter 6 and Chapter 7. Note, however, that the appropriateness of any given remedial action depends also on how one sees the origin of that problem, and the intent of this chapter is to focus on the first and second propositions. 3.2 REPORTS OF DIFFICULTY IN HIRING The reports of IT employers that have considerable difficulty in finding workers for their open IT positions are based on
OCR for page 93
Building a Workforce for the Information Economy Large numbers of vacancies,1 High turnover, Long times to fill positions, and Low unemployment rates for IT workers. While the committee believes that employers are accurately reporting their difficulties in hiring, these reports must be framed against a number of important contextual factors. In particular, high turnover rates and long times to fill positions requiring particular skills mean that significant vacancies will occur even if the supply of qualified workers is equal to the demand for them. Perhaps the most important consideration is the definition of a vacancy. The number of vacancies is sometimes used as a measure of the difference between supply and demand. But vacancy counts suffer from a number of problems. For example, data sources are not reliable. IT employers are increasingly turning to advertising job openings on the Internet. To the extent that vacancy counts are based on the amount of help-wanted advertising done in print (and if each help-wanted print ad represents one real open job), vacancy counts will understate the actual number of vacancies. Conversely, the lack of a posted vacancy may not indicate the lack of a position. Many IT employers have stated that they would be willing to hire an outstandingly talented individual even if no opening were immediately available, because s/he would be regarded as a strategic asset. Furthermore, advertised positions may not reflect immediate hiring needs, but may rather be posted in order to speed up hiring anticipated in the future (so-called “evergreen” positions for which IT employers are always recruiting).2 Or, advertising may reflect positions with particularly high turnover. Posted ads may reflect jobs that are contingent upon funding, for example, or on the presumed outcome of certain decisions that are out of the control of the immediate hiring manager. 1 Large vacancy rates are not surprising given the low levels of unemployment in the IT sector. As a rule, vacancy rates tend to move in inverse relation to the unemployment rate and positively with the employment and population rate. See, for example, Katz, Lawrence F., and Alan B. Krueger. 1999. “The High-Pressure U.S. Labor Market of the 1990s. ” Brookings Papers on Economic Activity, Vol. 0, No. 1). Available online at <http://www.irs.princeton.edu/pubs/working_papers.html>. 2 In some cases, vacancy counts may also reflect multiple countings of single openings. For example, consider N companies bidding on a large government contract. For planning purposes, it is prudent for each company to anticipate winning the contract —and it is a small step from that anticipation to anticipating the need to staff the project. But if all companies act in this way, and only one company can win the contract, then the number of anticipated openings is N times the number of openings that will actually be realized with the contract award.
OCR for page 94
Building a Workforce for the Information Economy A conceptual source of difficulty with the notion of “vacancy” is the fact that the number of jobs that “need” to be filled is not necessarily a well-defined quantity. By changing production modes, improving productivity, and so on, a firm may be able to do the same work with fewer workers. (This subject is addressed further at the end of this chapter but can include such measures as reorganizing workers into teams, improving management, organizing work and training workers to improve productivity and product quality, and using automated tools.3) A second major factor is the variation in hiring difficulty by type of worker sought. Demand is particularly high for some types of IT workers and less high for other types. In the aftermath of the Y2K programming problem, demand for COBOL programmers is declining, while demand for e-commerce specialists and Java programmers is increasing. In general, the time to fill a position varies with the position, but not surprisingly, positions requiring workers with “hot” skills take longer to fill. And, in the absence of data across the different types of position in the IT sector and in IT-intensive industries, advocates often make statements that imply uniform difficulty in finding workers. While an undifferentiated approach does not necessarily provide a misleading description of tightness in the market, policy remedies that are premised on the assumption of homogeneity may be misdirected. Finally, the variation in patterns and severity of difficulty experienced by individual employers is considerable, although the committee has not encountered an IT employer who has said that he or she has no difficulties in hiring. For example, the committee has heard reports of turnover rates—a contributor to vacancy rates—that vary from 5 percent to 50 percent per year, as against a national average rate of 11.7 percent per year for non-IT workers (Box 3.1). High turnover in a single job can result in an inflated number of vacancies. Consider, for example, a company of 100 employees experiencing no growth and with a turnover rate of 10 percent per year and an average time to fill a position of 6 months. This company will create five vacancies over the course of the year without an increase in the demand for IT work. With all that said, it is possible to make a rough quantitative estimate of the vacancy rate. If the vacancy rate V is defined as the number of jobs open, measured as a fraction of the workforce (i.e., open jobs divided by 3 As noted in Section 3.7.5, the extent to which tools, management, and organization can be used to reduce personnel needs is the subject of some controversy within the information technology community.
OCR for page 95
Building a Workforce for the Information Economy BOX 3.1 Two-Year Turnover Rates for IT and Non-IT Workers with Bachelor 's Degrees and Higher, from 1995 to 1997 High turnover characterizes many firms in the IT sector and IT-intensive firms, especially the former. From 1995 to 1997 the turnover in IT positions by individuals with a bachelor's degree or higher who either changed employers or changed jobs for the same employer was 38.3 percent (for a turnover rate of 19.2 percent per year), with the rate being almost 60 percent (30 percent per year) for those 30 years of age or younger (Table 3.1.1). The comparable rate for non-IT workers with an equivalent educational level was 23.4 percent (11.7 percent per year) with about a 50 percent (25 percent per year) turnover rate in the 30 and under age group. Comparison of the rates for the other age groups shows a significant increase for IT workers, at least to the age of 60. The primary reasons for the turnover are pay, at 25.4 percent for IT workers and 13.7 percent for non-IT workers, and working conditions, at 14.8 percent for IT workers and 9.0 percent for non-IT workers. Other reasons for turnover are similar for both groups, except that being laid off accounts for 6.7 percent of IT worker turnover and 4.7 percent of non-IT worker turnover. There is little difference in the layoff rate for non-IT workers across age groups, but for IT workers over the age of 40 years the rate rises to 7.7 percent. TABLE 3.1.1 Rates of Turnover by Age Group, 1995 to 1997 Age Group in Years IT Workers Non-IT Workers 30 or Under 58.3% 52.4% 31 to 40 41.3% 28.8% 41 to 50 34.1% 19.2% 51 to 60 29.9% 15.8% Over 60 18.7% 10.5% Total 38.3% 23.4% SOURCE: NRC staff analysis, based on NSF SESTAT data from 1995 and 1997. the number of jobs in the workforce), the following relationship can be used to estimate V: V = (T + G) × F, where T is the rate of turnover (measured as percent of workforce turned over per year), G is the rate of net of growth in employment in the industry (measured as percentage of growth per year), and F is the average time to fill an open position (measured in years).
OCR for page 96
Building a Workforce for the Information Economy As described in Section 7.2, the Bureau of Labor Statistics (BLS) estimates a growth rate in Category 1 IT professions of about 8 percent per year. Joint Venture: Silicon Valley Network indicates a value for F of 3.7 months for “high-tech jobs” and further estimates that the turnover rate in Silicon Valley is twice the national average.4 The national turnover rate for all workers is about 14 percent.5 All of these numbers give a vacancy rate of about 11 percent in these IT professions, if the conditions that characterize the Silicon Valley IT workforce obtain nationally. But, of course, they do not —the IT labor market is much tighter in Silicon Valley than elsewhere —and, as noted in Chapter 2, the majority of employers of Category 1 IT professions are not IT-sector firms (and not in Silicon Valley either). Note also two important characteristics of this model: Turnover and growth are portrayed as independent parameters. For the purpose of estimating vacancy rates as a function of turnover and growth, this is a reasonable portrayal. However, in fact turnover and growth are not likely to be independent of each other; that is, increased growth rates may well contribute to rising turnover rates as well. Such pressure arises from the fact that a worker may well leave a job in search of higher compensation—and in times of high growth in the IT sector and/or in IT-intensive industries, higher compensation may well be easier to find. In this model, vacancies due to turnover are three times the vacancies due to growth (i.e., 24 percent turnover per year divided by 8 percent growth per year). Workers who leave one position are available for another one. Put differently, if turnover could be reduced, say by a factor of 3 (from 24 percent to 8 percent), the vacancy rate would be half of what it currently is—a rate that would surely result in less concern. As noted above, high growth may also lead to higher vacancy rates, but because workers may leave a job for reasons other than higher compensation offered elsewhere, employers have a variety of nonfinancial as well as financial options for reducing turnover. A more refined model separates out the IT sector (employing about 25 percent of the Category 1 IT professionals) from the IT-intensive comp- 4 Joint Venture: Silicon Valley Network. 1999. Joint Venture's Workforce Study: An Analysis of the Workforce Gap in Silicon Valley. Palo Alto, Calif.: Joint Venture: Silicon Valley Network. 5 The Bureau of National Affairs estimates that the turnover rate for all employees in 1998 averaged 1.1 percent per month, or about 14 percent per year. See Bureau of National Affairs. 1999. Turnover Holds at Highest Levels of the 1990s, BNA Survey Finds. Washington, D.C., March 17. Available online at <http://www.bna.com/press/trn98r1.htm>.
OCR for page 97
Building a Workforce for the Information Economy anies (employing the remainder). For this latter group, estimating the rate of employee turnover T as 20 percent per year, employment growth G as 8 percent per year (per BLS), and position time-to-fill F as 2 months (1/6 year) yields a vacancy rate of 5 percent for the workforce of the IT-intensive companies. Assuming that the Silicon Valley calculation for vacancy rate above is valid for IT-sector companies in all geographic regions (which it is not), a composite vacancy rate of about 6 to 7 percent is indicated. 3.3 THE INFERENCE OF A WORKER SHORTAGE Accepting that most employers are indeed having some nontrivial difficulty in finding workers for their IT jobs, it is another conceptually distinct step to infer that there is a shortage of IT workers. An employer sees the workforce issue from the perspective of an individual firm. From this perspective, what matters is how many jobs are unfilled and how hard it is to attract workers to fill those jobs. But “shortage ” is a concept that applies more broadly to the entire universe of firms that use IT workers and compares overall supply with overall demand for these workers. Thus, the relevant question is the extent to which the latter can be inferred from many individual employer reports regarding their difficulties. Furthermore, part of the issue is terminological—the term “shortage” has many emotional connotations, and there is no uniformly accepted definition of the term among economists, to say nothing of common usage in discourse among employers. (Box 3.2 provides several definitions that have been presented to the committee in various forms.) BOX 3.2 Concepts of an Occupational Labor Shortage Perhaps surprisingly, there is no standard definition of an occupational labor shortage. In the “social demand model,” members of society believe that the labor market outcome results in too few workers in an occupation; in “ market disequilibrium models,” the amount of labor demanded exceeds what is supplied at the prevailing wage rate. The Social Demand Model Some analysts consider a shortage to be present if the number of workers in an occupation is less than what is considered the socially desired number. Under this
OCR for page 98
Building a Workforce for the Information Economy definition, a shortage of engineers exists if the analyst making the determination concludes that society would be better off if there were more engineers. This type of definition does not imply that the labor market is in disequilibrium; instead it describes a situation where the person who claims there is a shortage does not like the market's results. To quote Arrow and Capron, “Careful reading of such statements [of shortage] indicates that the speakers have in effect been saying: There are not as many engineers and scientists as this nation should have in order to do all the things that need doing such as maintaining our rapid rate of technological progress, raising our standard of living, keeping us militarily strong, etc. In other words, they are saying that (in the economic sense) demand for technically skilled manpower ought to be greater than it is—it is really a shortage of demand for scientists and engineers that concerns them.”1 Market Disequilibrium Models The committee uses the term “market disequilibrium models” to describe situations where the number of workers willing to work in an occupation is less than the number of workers employers would like to hire at the prevailing wage rate and where something prevents an adjustment that results in the market achieving an equilibrium. An equilibrium exists when the number of workers willing to work at the market wage rate is exactly equal to the number of workers employers would like to hire at that wage. Why might we have market disequilibrium occupational shortages? One possibility is that government or institutional forces restrict wage rates or entry into an occupation. For example, if the government wage structure limited the salary of IT workers in the federal government to half of what they could receive in the private sector, the number of workers demanded would exceed the number willing to work in the federal government, and there would be a shortage. Another reason that shortages might exist is that market changes occur faster than labor supply can adjust. In occupations such as IT, where it takes a number of years for people to gain the skills necessary to participate, sustained increases in the demand for labor can prevent supply from “catching up,” at least for a period of several years. A number of studies have developed formal models and applied them to the market for engineers, scientists, and IT workers.2 What these models have in common is that so long as demand grows more rapidly than supply grows, it is impossible for the market to reach equilibrium. 1 See Arrow, Kenneth J., and William M. Capron. 1959. “Dynamic Shortages and Price Rises: The Engineer-Scientist Case,” Quarterly Journal of Economics (73)2:292-308. 2 See, for example, Arrow and Capron, 1959, “Dynamic Shortages and Price Rises: The Engineer-Scientist Case”; Blank, David J., and George J. Stigler, 1957, The Demand and Supply of Scientific Personnel, New York, National Bureau of Economic Research; Radner, Roy, 2000, “A Simple Model of Aggregate IT Labor Market Dynamics,” Stern School, New York University, May 7; and Ryoo, J., and S. Rosen, 1996, “The Market for Engineers,” National Bureau of Economic Research, summarized in Forecasting Demand for and Supply of Doctoral Scientists and Engineers, Office of Scientific and Engineering Personnel, National Research Council, Washington, D.C.: National Academy Press (forthcoming).
OCR for page 99
Building a Workforce for the Information Economy 3.3.1 The Overall Labor Market It is important to frame any discussion of shortages in the IT labor market in the context of tightness in the overall labor market. The labor market is tight across the entire economy, not just in IT. Thus, individuals (incumbent workers, those entering the labor force, and those that may leave the labor force) have many more options available to them than when the overall labor market is more slack. Therefore, individuals who might be motivated to move into IT for economic reasons have many other fields in which they may work productively. (Of course, tightness in the overall labor market depends on the state of the economy, now in an unprecedented tenth year of expansion. It is anyone's guess when the economy will experience a downturn, but such a downturn would inevitably affect analysis of the workforce situation.) Note also that tightness in the labor market for IT workers is felt the world over (Box 3.3). 3.3.2 The Size of the Applicant Pool The size of the labor pool available for, and willing and qualified to do, IT work is a central issue in assessing the balance between supply and demand. Perhaps the most critical point is that the size of the relevant labor pool is not fixed—many factors affect it (e.g., compensation and wages, employer willingness to provide training for new hires6). But there is no direct measure of this pool, and much of the controversy about tightness in the labor market involves just this point. The number of applicants for an open IT job is often cited as an indication that there are large numbers of skilled workers currently available.7 However, most individuals seeking employment in IT apply for multiple jobs. From the job seeker's perspective, it makes sense to send out many applications, to increase the likelihood that a promising contact will be established. The job seeker may also apply for jobs for which s/he is only marginally qualified, or even unqualified, on the theory that a high probability of rejection is better than the certainty of rejection if the individual does not apply at all for a given job. And even individuals who are employed may be “testing the market, ” if nothing else to maintain 6 As discussed at greater length in Chapter 7, employers are often concerned that they will not reap the full benefits of investments they make in training employees because employees are generally free to leave for other employment. Thus, they may be less willing to provide training than they might otherwise be. 7 Matloff, Norman, “Debunking the Myth of a Desperate Software Labor Shortage,” testimony to the U.S. House Judiciary Committee Subcommittee on Immigration, April 1998.
OCR for page 100
Building a Workforce for the Information Economy BOX 3.3 Tightness in the IT Labor Market Around the World “To get one man, Erik Masing bought a whole company. . . . Mr. Masing 's experience shows the extent to which lawmakers and bureaucrats [in Germany] have tried to slam the door on foreign high-tech professionals even as the nation's employers bemoan a shortage of engineers and programmers.” (Zachary, G. Pascal. 2000. “As High-Tech Jobs Go Begging, Germany Is Loath to Import Talent,” Wall Street Journal, January 17, pp. A1, A10) “The information technology industry faces a common problem of global proportions: Plenty of work to do and not enough people to do it. . . . [CIOs and IT executives around the world say that if the acute skills and labor shortages aren't addressed, national IT sectors will lose their competitive edge, economies will suffer, and innovation will slow.] The problem appears to be most threatening in Europe and Asia, where the scarcity of IT workers could stifle hopes for long-term economic growth. In Germany and the U.K., there's a 25 percent gap between jobs created and jobs filled. That's quite a disparity, considering that the countries together account for half of Europe's total IT production, according to the German Information Technology Association.” (Busse, Torsten, and Mary Brandel. 1998. “The Skills Struggle,” Computerworld Global Innovators, December 7, pp. 12-13) “. . . [T]he skills shortage remains one of the greatest bottlenecks to building a broad Asian base of new technology development. . . . “ (Guth, Rob, et al. 1998. “Despite Rampant Unemployment, Lack of IT Skills Threatens Asian Growth,” Computerworld Global Innovators, December 7, p. 15) “In Venezuela . . . the demand for skilled IT professionals has skyrocketed in the past two and a half years . . . says Jorge Mora, IS manager. ” (Perez, Juan Carlos. 1998. “In Most Regions, Labor Supply Can't Meet Skyrocketing Demand,” Computerworld Global Innovators, December 7, p. 17) “While we are again showing how not to have the right conversation about foreign-born high-tech workers . . . the rest of the developed world is waking up to the fact that America's cherry-picking of international tech talent amounts to an enormous competitive advantage . . . Our competitors are doing something about it. Germany, Canada, the United Kingdom and Australia, among others, have already entered or are preparing to enter the sweepstakes for high-tech workers.” (Papademetriou, Demetrios G. 2000. “The Global Fight for Talent,” Washington Post, March 21, p. A25) an understanding of trends in the sector in which the job seeker is interested and knowledge of his/her market worth. Furthermore, the trend toward multiple applications has been fueled by the ease of posting resumes on the Internet. The use of electronic resumes and Internet-based recruiting is increasingly common in many sectors of the economy, and is especially common in the IT sector. The number of job-related Web sites is growing rapidly, and the number of resumes posted already runs in the millions (4.9 million electronic resumes were posted in
OCR for page 101
Building a Workforce for the Information Economy 1999, and this figure is forecast to grow to over 16 million by 2002). 8 It is not unusual to receive hundreds or thousands of resumes in the process of filling desirable IT jobs.9 As a result, while an individual employer may in fact receive many applications from different individuals, a group of employers may be receiving applications from the same pool of applicants, and so the size of the overall applicant “pool” is not as large as it might appear extrapolating from the experiences of individual employers. Analysts also sometimes point to the number of individuals graduating yearly with computer science or related degrees and to the magnitude of layoffs in the IT sector. In the first case, the fact that many IT occupations—especially in Category 2—do not require significant amounts of formal IT training suggests that the number of computer science (CS) graduates produced yearly is not a good indicator for the labor supply as a whole taken across all IT occupations. On the other hand, it is difficult if not impossible to determine how many non-CS graduates have the skills to enter the IT workforce immediately. As for layoffs in the IT sector, they are considerable. For example, the outplacement firm of Challenger, Gray, and Christmas reported that, over the first 11 months of 1999, the computer industry had laid off 60,000 employees.10 This number is equal to approximately 3 percent of total employment in the “computer services” industry, as defined by the Bureau of Labor Statistics.11 Whether these individuals constitute a ready, immediate, and usable source of supply is unclear. On the one hand, they may not necessarily have IT skills, because, like all companies, IT firms employ many support staff, clerical staff, building maintenance staff, and others. And, even if those laid off are technical workers, they may or may not have the skills needed by employers that are growing and hiring. On the other hand, some organizations such as SAIC and the Communications Workers of America have demonstrated that to a certain extent, retraining of laid-off “employees” (i.e., ex-military personnel) to undertake IT work can be possible. The supply of “qualified” workers made available by layoffs may well depend on the efforts that are undertaken to enhance their technical skills. Finally, the size of the relevant applicant pool depends greatly on policy 8 Vaas, L., A. Chen, and M. Hicks. 2000. “Web Recruiting Takes Off.” PC Week, January 17, pp. 57-68. 9 Murphy, Kevin, and Zinta Byrne, “Applications of Structured Assessment in the IT Workforce,” commissioned paper prepared for the Committee on Workforce Needs in Information Technology, March 2000. 10 USA Today, December 10, 1999, p. A-1. 11 BLS uses the North American Industrial Classification System, which defines “computer-related business services” as companies that provide software services, data processing, and information services, and rental, maintenance, and other computer-related services.
OCR for page 102
Building a Workforce for the Information Economy decisions that regulate the number of foreign IT workers who may work in the United States. These include skilled IT workers on H1-B temporary visas (who must have at least a bachelor's degree), permanent immigrants, and foreign students graduating from U.S. colleges and universities. 3.3.3 Skills Shortages Versus Worker Shortages If there is a “shortage,” what is in short supply? In the common formulation described at the start of this chapter (and most often articulated by IT employers), the shortage is one of workers. But in the IT workplace, not any warm body can be an IT worker, nor are IT workers interchangeable with each other. Attention must also be paid to the ability of putative workers to do IT work. Thus, any shortage in fact refers not to IT workers but to qualified IT workers—with all of the discussion and nuance that relates to the definition of “qualified.” Specifically, critics of employer practices believe that the job definitions used by employers overstate or exaggerate the qualifications that are actually needed to perform the work required in job openings. Further, they argue, the employer demand for a “perfect” fit between an employee and a posted job opening creates artificial difficulties in finding human resources who would be able to do the work of a posted job with only a “little bit” of training. 3.3.4 Compensation Compensation must figure prominently in any discussion that relates the supply of labor to demand. The basic economic model is the following: Demand for labor relates the number of jobs that are available as a function of the compensation that employers are willing to offer for those jobs. Supply of labor relates the number of people willing to work as a function of the compensation that employers are willing to offer. When demand exceeds supply in a particular occupation, compensation tends to rise relative to compensation in other occupations that require similar education, effort, and working conditions. Rising compensation attracts into a field more people who are willing to work (increasing the current supply), decreases demand for those workers, and signals to those capable of being trained to begin studying for these jobs (increasing the future supply of new entrants).12 12 One caveat is needed. To the extent that high school students take courses of study that make it more difficult to study IT-related subjects in college, such students will be unable to respond effectively to market signals indicating the desirability of the IT field. While these students may not be precluded from undergraduate study in IT-related areas, they may well have to take remedial courses after high school —and thus the entry of these students into IT work is likely to be delayed even further.
OCR for page 122
Building a Workforce for the Information Economy clearly problematic, since very strong evidence exists that this relationship changes over time and will continue to change as the fixed costs of employment rise relative to variable costs and as the relative importance of overtime cost declines. Moreover, the assumption of such a fixed relationship amounts to an assertion of the interchangeability of persons and time worked, an assumption that is not valid in many sectors (including much of IT) in which team efforts are central. Finally, the BLS methodology neglects many dimensions in which adjustment may occur, including training and retraining, and especially response to changes in wages. None of the past changes in the relationships are assumed to have been affected by anything behavioral—everything is summarized in the time trend. For these reasons, it is likely that the underestimation of the growth in IT job categories will continue.32 The state of the labor market, of course, depends on more than demand —supply (by which is meant all sources of labor that could do useful IT work) matters as well. Elements of supply include individuals entering the IT workforce for the first time, individuals in the IT workforce who are inclined to leave the workforce unless given incentives to stay, individuals in the IT workforce who shift from areas of low demand to areas of high demand, individuals in other lines of work (or currently unemployed) who could move into the IT workforce, individuals currently working in a low-demand IT segment who could move into highdemand IT segments (perhaps with some retraining), and foreign IT workers who might be employed (either in the United States or abroad) to perform IT work on behalf of U.S. companies. Approaches to increasing supply are the focus of Chapter 7. 3.7.3 Skills for the Future Because it is—by definition—impossible to predict discontinuous changes in technology, assessments of the specific skill sets needed in the future can be based only on what is known today. With that caveat, it is likely that the types of IT worker in greatest demand over the long run will fall into three categories: Those who combine strong knowledge of a specific business with IT skills. As IT applications to support effective decision making become ever more pervasive throughout business, industry, and government, organizations in these sectors are likely to realize the benefits of process reengineering. 32 For more discussion of the limitations, see National Research Council, Office of Scientific and Engineering Personnel, 2000, Methods of Forecasting Demand and Supply of Doctoral Scientists and Engineers, the report from which this discussion is derived.
OCR for page 123
Building a Workforce for the Information Economy Effective reengineering of business processes requires good knowledge of what a firm is trying to do as well as good instincts for what IT can and cannot do. IT is becoming central to many fields, such as finance and health care, and thus a dual competency will be increasingly useful. Those with the skills to work with recent information technologies that have broad-ranging business application. Because information technologies change rapidly, those with the most recently acquired skills useful for technologies with broad application (or the ability to learn these skills quickly) are likely to be in very high demand. These individuals will be able to fit into a wide variety of organizational venues and businesses. And, because future applications will probably be more complex compared to the applications of today, those with the ability to manage complexity very well are especially likely to be in high demand. Those with extraordinary mastery of hard-core technology skills. There are always firms that need to squeeze the last bits of performance and functionality out of the information technologies that they use. While the number of such firms that are willing to pay large premiums for such efficiency is small (mostly because such efficiency is not necessary for most IT-sector or IT-intensive businesses), individuals with extraordinary mastery (“wizards”) will be in high demand for those that are so willing. 3.7.4 Project-based Employment For much of the latter half of the 20th century, many jobs were characterized by relatively high stability—long-term employment with one firm. Typically, these relationships involved assistance to employees for maintaining and upgrading skills to accommodate changing work assignments (because the firm perceives a stake in those skills) and salary structures that compensate junior or newer workers less than senior or older workers, in effect implementing a discount for lower productivity during the time new skills are being acquired early in a career. Of course, in practice, many workers did change employers, but such changes tended to be infrequent and measured in many years or decades. In another mode of employment that this report calls project-based, the firm uses a worker (and compensates him/her, either directly or indirectly) for a specific task, without obligation to continue employment beyond that task. (For ease of discussion, it is helpful to distinguish between the firm that needs work to be done and the employer that actually employs the worker. These may or may not be identical.) Project-based employment can take several forms: A “regular” employee of the firm, receiving benefits (e.g., health insurance), pension, or stock options or equity stakes, but without the
OCR for page 124
Building a Workforce for the Information Economy expectation that he or she will necessarily remain with the company after his or her stock options vest. Such an individual may well move from project to project within the firm, but must “job-hunt” within the firm once a given project has ended. An individual independent contractor (“self-employed”) with very well defined (and usually time-delimited) responsibilities to the firm, which has no responsibility to the contractor other than paying the agreed-upon fee. An employee of a third party such as a personnel firm, a temporary help service, or a consulting or contracting firm. In such instances, the firm contracts with the employer for either an individual to work on projects of the IT firm's choosing or a product or service that the employer will deliver to or on behalf of the IT firm. This practice, often referred to as outsourcing, started to become more common more than 10 to 15 years ago. In all of these instances, compensation for the worker is geared to the current worth of the worker to his or her employer for the duration of the task. The firm has no responsibility to the worker to ensure that his or her skills remain current or that he or she remains useful for another task for the same firm—skills are the worker's responsibility. Furthermore, in some instances, shorter job tenures are common, which may be due to the nature of the work in a sector in which workers can be displaced after a project ends. Both firms and individual workers may have some incentives to prefer project-based arrangements.33 For example, firms may obtain greater flexibility to address economic, strategic, and technological changes with project-based workers because they can more easily change the size and composition of their effective labor forces. Because such workers are not guaranteed jobs beyond their current project, the firm has opportunities to change workforce size and composition at the end of every project. Such arrangements can be useful in managing product cycles, during which personnel requirements are much larger before product release (i.e., during development) than afterwards, or in making conversions to new technology or products generally. 33 Much of this analysis is based on Kunda, Gideon, Stephen R. Barley, and James Evans, 1999, “Why Do Contractors Contract? The Theory and Reality of High End Contingent Labor,” draft working paper, June. While the discussion in this section addresses incentives, it is silent on the downsides of high-mobility employment, since it is intended only to explain why high-mobility employment is increasing. The Barley paper discusses many of the downsides, as does the following popular press article: Downey Grimsley, Kirstin. 2000. “Independent Contractors' Victory in Microsoft Case May Have Wide Impact,” Washington Post, January 16, p. H01.
OCR for page 125
Building a Workforce for the Information Economy In addition, the use of workers with varied job histories in the IT field enables firms to capture experience, intellectual sophistication, and knowledge that a new employee may have gained in working for a different firm (or on a variety of previous projects). While exploiting proprietary information belonging to competitors is illegal and unethical, a worker working at any given IT firm learns much that is not proprietary and that can benefit another IT firm. Moreover, the experience that a new worker can bring from a different IT firm means that the present IT firm does not have to subsidize the new worker's learning. This kind of human capital flow has been a feature of successful IT-focused regions, such as Silicon Valley. Finally, through the use of some types of project-based employees, such as independent contractors or service firms, firms can avoid paying fringe benefits and employment taxes. Similarly, the use of contractors or service firms allows a firm to avoid many of the costs of recruitment, training, and termination incurred to comply with laws protecting regular employees. Also, the use of contractors and service firms is sometimes “off the books” from a firm-wide perspective, which enables individual managers to obtain additional labor without having to clear it through the company hierarchy. Some workers also may find project-based employment advantageous, and their availability for such work can feed the growth of such employment.34 Project-based employment allows a worker to obtain experience at multiple firms and learn new skills more quickly by being a “regular ” employee of several different firms over a period of time. Another benefit of such varied experience may be financial: individuals who wish to build their own stock portfolio of pre-IPO shares may be able to diversify their holdings in order to improve their chances that at least one of the companies for which they have worked will in fact have a successful initial public offering. More generally, independent contractors may benefit from greater flexibility with respect to work schedules, the freedom to accept or reject assignments, or a greater range of insurance options (e.g., health plans). And, many “free agents” do well as high-end consultants or well-paid workers with skills in critical areas, because it is difficult for companies to find permanent employees with these abilities. At the same time, free agents with skills in less critical areas tend to make less than their permanently employed counterparts. Thus, the economic consequences of project-based employment are far from uniform. Finally, 34 A study conducted for the Kelly Services employment agency by EPIC/MRA, a marketing research consultancy, found that 64 percent of IT workers either want to work as a free agent or are open to the idea. According to the Kelly study, more than 36 percent of IT free agents earn more than $100,000 per year.
OCR for page 126
Building a Workforce for the Information Economy free agents must cope with occasional job instability and often receive fewer benefits than permanent workers. 35 Because both workers and firms may perceive benefits to these more flexible arrangements, project-based employment appears to be increasingly common, and Web-based businesses have begun to emerge to provide matchmaking between parties offering and seeking work. 3.7.5 Reducing Relative Needs for Personnel Through Tools and Techniques for Greater Productivity Historically, the personnel needs of many fields, e.g., agriculture and manufacturing, have been reduced by the use of machinery to displace labor. Indeed, many economic models of productivity36 are built around the assumption that, to varying degrees, capital can substitute for labor. Different management approaches have also helped to improve productivity in certain instances. It is thus reasonable to ask if such approaches might be promising in the field of IT. Tools For IT, the analog of mechanized agriculture is tools that enhance individual productivity, noting that productivity is, by definition, the ratio of useful output to human work input to produce that output. In agriculture, for example, the use of tractors, threshers, and the like enables the production of far larger amounts of food for a given number of farming hours than would be possible with the use of oxen-drawn plows and scythes. In information technology, widely used operating systems, language compilers, debugging tools, performance analysis systems, environments for program maintenance and integrated development, component frameworks, rapid prototyping tools, and problem-solving systems have all helped to increase the productivity of Category 1 workers. The development of more and better tools to enhance individual programmer productivity remains an active area of research and development. 35 Edwards, John. 2000. “Redefining IT Career Paths for the New Millenium,” IEEE Computer 33(1). 36 The committee notes that the term “productivity” is used in many different contexts. The basic definition of productivity, and the one used in this report, is the ratio of output per unit input. The term is also used in the context of describing an individual company's competitive edge (a company is productive if it can develop and sustain an edge over its competitors) as well as in the context of a workforce on the whole that can do more (a productive workforce is a “rising tide that lifts all boats”).
OCR for page 127
Building a Workforce for the Information Economy BOX 3.6 Illustrative Tools for Increased Productivity Software reuse is the systematic building of a family of products, drawing from a common core of elements including software code, design, domain expertise, and development processes. Reuse has the potential to reduce the amount of product development work that needs to be undertaken. However, software reuse is challenging because it entails deep understanding of the commonality of requirements that are at the core of the software development process. End-user programming (of which spreadsheets and Mathematica are good examples) shifts IT work away from IT workers toward end users, thereby reducing or eliminating the need for the end user to interact with a specialized IT worker to formulate and/or program an end-user problem, with the effective result that fewer programmers are needed. Industry and enterprise-wide software systems integrate functionality to support routine applications common to most firms in a given industry (e.g., systems to support common and generic hospital or banking operations). Such systems have the potential to reduce the need for large numbers of IT workers developing customized solutions and applications for specific firms. The primary challenges are in understanding the essential features common to most organizations in a particular industry or business without compromising the ability to specialize and customize that is the basis on which individual firms compete. Applications service providers provide IT services to large numbers of client firms on a centralized basis. Much of the software in use today is locally installed and maintained, a fact that requires client firms to employ their own Category 2 workers to assist with and be responsible for such tasks. By providing software that is installed and maintained on servers that are remote to the user firms, a large amount of effort that is currently duplicated among user firms can be replaced with a smaller amount of effort by the applications service provider. Implementation of the “ilities” has become increasingly important in today's complex software environment. Key properties of systems such as security, flexibility, reliability, manageability, quality of service, modifiability, and scalability—some of the “ilities”—have become more important as systems become bigger, more complex, and accessible (via the Internet) by more users. A better understanding of how to develop very large systems that encompass the “ilities” (as well as better implementation of what is already understood) is likely to reduce the number of people needed to develop and maintain systems. It is likely that such tools (as illustrated in Box 3.6) will indeed emerge as the result of continued research and development in the area, although the magnitude of the productivity improvements that can be expected from the use of tools is a matter of sharp dispute within the IT community.
OCR for page 128
Building a Workforce for the Information Economy Management and Organization Management strategies and organizational approaches may result in higher productivity. Software development does not have a consistently good track record for quality, and from the earliest days of programming, there have been concerns about cost and time overruns and program reliability. As systems became larger, the opportunity for error became greater. Indeed, because the development of large systems requires teams, and all teams are organized in some manner, the role of organization and management is manifest. Over the past few years, the Standish Group has surveyed a wide range of organizations on the outcome of their IT projects.37 Overall the results are poor. In 1994, 31 percent of projects were canceled before completion, and a further 53 percent were completed over budget. Two years later, the corresponding results were that 40 percent of projects were canceled before completion and 33 percent were completed over budget, and in 1998, the results were that 28 percent of projects were canceled before completion and 46 percent were completed over budget. Although these results do show some improvement over time, the results for 1998 are still very unsatisfactory. Common sources of project cancellations and overruns include ill-defined or changing requirements, poor project planning or management, uncontrolled quality problems, unrealistic expectations or inaccurate estimates, and naive adoption of new technology.38 In addition, individual projects can be organized in ways that minimize personnel needs over the life cycle of a product. For example, software inspection for quality at the front end of the development process has been shown to reduce dramatically the amount of rework and debugging that otherwise needs to be done at the end of the process.39 Thus, even though initial costs are higher, overall personnel needs can be lower.40 37 Johnson, Jim. 1990. “Turning Chaos into Success,” Softwaremag.com. December. Available online at <www.softwaremag.com/archive/1999dec/Success.html>. 38 McConnell, Steve. 1998. Software Project Survival Guide. Redmond, Wash.: Microsoft Press, pp. 20-33. 39 See Wheeler, David A., Bill Brykczynski, and Reginald N. Meeson. 1999. “Software Peer Reviews,” pp. 454-469 in Software Engineering Project Management, Richard H. Thayer, ed. Los Alamitos, Calif.: IEEE Computer Society Press. 40 For various reasons, certain development environments militate against such an approach. For example, a small start-up company seeking to bring to market a single product against severe time-to-market pressures may not have the funding to make up-front investments in architectural design and quality assurance. Such pressure simply increases the importance of management discipline (e.g., in building such investments into the original business plan). (In some cases, pressures to be first to market can lead to excessive neglect for the test and debug phases, making the end user the beta tester!) Some interesting commentary on these pressures can be found in Minasi, Mark. 1999. The Software Conspiracy. McGraw-Hill Companies. See also Computer Science and Telecommunications Board, National Research Council. 1999. Trust in Cyberspace. Washington, D.C.: National Academy Press.
OCR for page 129
Building a Workforce for the Information Economy Finally, firms that use IT workers may outsource non-core competencies that involve IT to specialized service firms, thus reducing the need for in-house IT staff. Today, many non-IT companies, such as financial institutions, are reducing their need for whole categories of IT personnel because they cannot compete internally with what they can buy relatively cheaply on the outside; as a result, they have less need to build IT systems in house. With outsourcing, the same amount of work needs to be done through the economy as a whole, but a specialized service firm may well be able to undertake work more efficiently because of efficiencies of scale or greater experience and knowledge in providing the service. Arguments about the importance of management practices in the development of IT software have been made for at least three decades. There is disagreement in the IT community about how much more can be gained by better management practices. Some believe that much can be gained if every project used the best management practices, for example. (See Box 3.7 for a discussion of illustrations.) Others feel, however, that while there is undoubtedly some productivity to be gained, there is no silver bullet in such an approach, largely because software development is an extremely complex process.41 The Likely Impact of Improvements in Productivity For a fixed amount of work, the impact of productivity improvements, whether from tools or management techniques, is to reduce the number of personnel needed to perform that work. If the work that needs to be done (otherwise known as demand for labor) increases, then the impact of productivity improvements is to reduce the number of personnel needed below the number that would be needed in the absence of such improvements. Furthermore, to the extent that productivity improvements make it easier for people with less expertise and training to accomplish what previously took greater skill, such improvements may well serve as one stimulant of additional demand for IT labor. Historical experience, compiled by Boehm,42 suggests that the use of improved tools and improved management has resulted in an annual 41 See, for example, Brooks, Frederick P., Jr. 1995. “No Silver Bullet,” in The Mythical Man-Month, Anniversary Edition, Reading, Mass.: Addison-Wesley. 42 Boehm, Barry W. 1999. “Managing Software Productivity and Reuse,” IEEE Computer 32(9):111-113.
OCR for page 130
Building a Workforce for the Information Economy BOX 3.7 Elements of Software Engineering Management Experience in software engineering suggests that many of the items described below characterize software development projects that have come in on time, on budget, and with the “expected” quality. Following these best practices—especially if they are “followed” but not implemented appropriately—is not a guarantee of success but may well increase the likelihood of project success, and many project failures can easily be traced to the absence of such practices. Have a clear vision, in which project teams work toward a limited and clear set of goals. Use prototyping tools that allow users to interact with system mockups. The use of such tools helps the user to understand what he or she wants the system to do, and thus addresses one of the most common risks—changing requirements after the system has been largely developed. Demand accurate estimates that reflect the true state of affairs. Under pressure of deadlines, it is often tempting to give time and effort estimates that are overly optimistic. To the extent that such estimates are believed by management, they set up a project for failure. Focus on quality from the start. Uncorrected product defects that occur early in the development process cost much more to fix than product defects that occur later. Under time pressures, quality control often falls by the wayside, but it is the role of management to ensure that this does not occur. When appropriate, implement projects in a known and familiar technology. New technologies can offer important and powerful new capabilities, but their use often reflects a high-risk strategy—because the expertise to exploit these capabilities properly is often not available. Avoid micromanagement, which is demoralizing to project teams. Management should be responsible for staffing, training, motivation, morale, and the work environment but should leave the technical people to do what they do best. Manage risk actively. Management should plan the project so that the largest areas of risk are worked on early in the project. After each increment of work, the project risks should be reassessed and a plan developed to address each remaining risk. These items are not intended to be complete or exhaustive—they merely illustrate some of the things that are entailed in good management. SOURCES: Construx Software (Bellvue, Washington): <http://www.construx.com>; Scott Ambler's Web Site for Object-Oriented Developers: <http://www.ambysoft.com/processPatternsPage.html>; and Software Program Managers Network (Arlington, Virginia): <http://www.spmn.com/best_practices.html>.
OCR for page 131
Building a Workforce for the Information Economy growth rate in software productivity of about 7 percent for the past 30 years. If this trend can be extrapolated to the future, it will approximately compensate for the 8 percent annual increase over all of the “core” IT occupations over the next 10 years projected above. If this is true, the impact of productivity improvements will be to maintain the approximate status quo with respect to the current tightness in the IT labor market. However, as noted above, recent comparisons of actual to projected demand indicate that U.S. government projections have underestimated the rate of job growth by almost a factor of 2. And, the rate of job growth in certain specialties is likely to be even higher. Boehm estimates that annual improvements in productivity of 10 percent are possible if new techniques are used more widely, though his analysis is qualified by both the fact that it uses equivalent machine language instructions as the output measure (a controversial measure in the IT community) and the fact that it is based on large software systems usually built on contract for the federal government (a point that may limit its applicability to many important software development efforts in the private sector). These comments indicate that while productivity tools and techniques have a role to play in reducing personnel requirements below what they would otherwise be, they are not likely to play a decisive role in reducing the current tightness in the IT labor market. 3.8 RECAP A primary manifestation of a tight labor market is the fact that many—perhaps most—employers of IT workers report large numbers of vacancies for IT positions. However, turnover and company growth contribute equally to the vacancy rate. Overall, today's IT labor market is tight, although the nature and extent of such tightness vary by employer, by type of IT work involved, and by geographical locale. The fundamental driver of this tightness is growth in the use of IT throughout a strong economy, a tightness that is significantly exacerbated by the currently low unemployment rate in the overall labor market. The behavior of wages, a common indicator of labor shortages, presents a mixed picture. Base wages for Category 1 workers (apart from those in hardware, about which little is known) have been rising at a rate of about 4 percent per year in constant dollars. However, this overall behavior masks much more rapid wage growth in certain subspecialties and less rapid growth in others. Furthermore, the presence of unexercised and/or unvested stock options and equity stakes in the compensation of workers in a relatively new and growing industry may help to explain the fact that mean wages in the IT sector have risen only somewhat more rapidly than wages in other sectors of the economy. Because
OCR for page 132
Building a Workforce for the Information Economy stock options and/or equity stakes that represent deferred compensation are an increasing part of worker compensation packages in IT, as suggested above, wages alone become a poorer measure of total compensation as time goes on. Thus, the omission of stock options and equity stakes is problematic in comparing wage trends in IT versus those in other sectors. The federal IT workforce presents special problems. Many government functions depend on IT, but government's ability to respond as the market would respond is limited, especially with respect to the compensation packages it can offer. The time horizons of the current tightness in the labor market are hard to predict. In the long term (measured in decades), continued growth in the IT sector and in the use of IT by IT-intensive firms is highly likely. On the other hand, all sectors—as well as the economy at large—experience periods of greater and lesser growth (or even contraction). When they occur, downturns in the IT sector and in the IT-intensive industries may reduce the amount of IT work that they can supply, increase the number of workers who are available to do (or are trainable for) IT work, and reduce the value of capital available (e.g., stock options) for compensation of IT workers. Such downturns are also likely to result in reduced demand for IT workers, with a consequent decline in the need for foreign temporary nonimmigrant IT workers and a slack labor market. Current projections for job growth for Category 1 IT workers, which do not take into account the possibility of such downturns, indicate strong growth for the next decade, about 7 percent per year. Historically, such projections have understated actual growth rates by as much as a factor of 2. Analytically, there are only a few ways of dealing with tightness in a labor market—to increase the productivity of individual workers so that a smaller number of workers can do the same work that a larger number of workers can do in the absence of productivity measures, and to increase the number of qualified workers. (Reducing demand for IT products and services, and hence the need for IT workers, is a third logical possibility, but one that contradicts the premise of continued growth in the use of IT.) Productivity can in principle be increased through the use of tools (e.g., integrated programming environments) and/or the use of different organizational or managerial strategies (e.g., structuring projects with more “up-front” design effort to reduce “downstream” personnel needs). Over the past 40 years productivity gains have been achieved through new management paradigms and technology. Similar gains are likely in the future, although they are not likely to play a decisive role in reducing the current tightness in the IT labor market. Approaches to increasing the supply of workers and making more effective use of the existing workforce are discussed in Chapter 6 and 7.
Representative terms from entire chapter: