Order Code RL30140
CRS Report for Congress
Received through the CRS Web
An Information Technology Labor Shortage?
Legislation in the 106th Congress
Updated June 14, 2001
Specialist in Labor Economics
Domestic Social Policy Division
Congressional Research Service ˜ The Library of Congress
An Information Technology Labor Shortage?
Legislation in the 106th Congress
Concern about an information technology (IT) labor shortage prompted
Congress, in Title IV of P.L. 105-277, to raise for 3 years the ceiling on H-1B visas
for skilled temporary alien workers. The law also called on the National Science
Foundation to assess employers’ future requirements for workers with computerrelated skills, the future education/training needs of U.S. students to ensure an
adequate supply of workers at requisite skill levels, and the potential costs/benefits to
the U.S. economy from admission of foreign workers with science and engineering
skills. The study requirement suggested that disagreement persisted about the
existence of an IT labor shortage. The imposition of a user fee on employers who file
H-1B visa petitions to largely go toward funding technical skills training and math,
engineering or science education similarly implied ambivalence over the preferred
solution to the perceived shortfall of IT workers.
While many (including the National Research Council in its congressionally
mandated report released in October 2000) would agree that the IT labor market
became tight in the late 1990s, the paucity of good data made it difficult to
unambiguously determine that an IT labor shortage existed. For example, although
employment increased rapidly for IT jobs overall during the 1990s and is projected
to continue to do so through 2008, job growth rates varied between individual IT
occupations. In addition, while the decline in bachelor’s degrees conferred in
computer/information sciences was used to demonstrate that firms faced a dwindling
supply of workers, bachelor’s degree holders in other disciplines commonly work in
IT jobs and other sources (e.g., community colleges) prepare persons for IT positions.
Although estimates of job vacancies were pointed to as proof of a labor shortage, they
are an insufficient indicator — especially those made without reference to the duration
of job openings or the wage levels associated with them. Moreover, an occupation
could have both a high unemployment rate (which suggests excess supply) and a high
vacancy rate (which suggests excess demand) if firms search in the labor market for
experienced workers who already possess the hottest IT skills and simultaneously
layoff, rather than retrain, their own IT employees. (A subtext of the IT worker
debate was whether demand exceeded supply for all workers; for workers already
trained in the latest skills; or for young workers to whom firms could pay relatively
low, entry-level wages.) Surveys also presented a mixed picture of IT pay which,
under shortage conditions, should increase much faster than wages for jobs with a
more abundant supply of labor.
Proposals were introduced in the 106th Congress to promote education and
training in math, science and engineering skills among U.S. residents. Other bills
returned to immigration policy as a way to lessen tightness in the IT labor market
while often also raising P.L. 105-277’s user fee and changing its allocation formula.
The legislation that the 106th Congress passed to address this issue continued the twopronged approach initiated by the 105th Congress, namely, further raising the limit on
H-1B visas as well as increasing the user fee for education/training purposes and
altering its allocation formula. With the IT sector experiencing reduced product
demand and employee layoffs in 2001, interest in this issue has waned.
The Demand for and Supply of IT Workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Demand Side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Supply Side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Measures of Labor Market Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
The Unemployment Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Vacancy Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Wage Increases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Selected Results from the ACWIA-Mandated Studies . . . . . . . . . . . . . . . 11
Legislation and Education/Training Programs . . . . . . . . . . . . . . . . . . . . . . . . .
The 105th Congress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DOL’s 56.3% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NSF’s 28.2% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
P.L. 105-220 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The 106th Congress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bills Amending ACWIA’s User Fee Level and Allocation Formula . .
Using Tax Incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Other Bills Unrelated to ACWIA’s User Fee and
Education/Training Provisions . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix: The Older Worker Issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
List of Tables
Table 1. Employment in Selected IT Occupations, 1989-1999 . . . . . . . . . . . . . . 2
Table 2. Employment, 1998 (Actual) and 2008 (Projected) . . . . . . . . . . . . . . . . 3
Table 3. Percent Distribution of Workers in Selected Occupations by Highest
Level of Education Completed, 1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Table 4. Unemployment Rates in Selected Occupations, 1989 and 1999 . . . . . . 7
Table 5. Median Weekly Earnings of Full-Time Wage and Salary Workers
in Selected IT Occupations, 1989-1999 . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Appendix Table 1. Average Wages in Selected ITOccupations, 1999 . . . . . . . 18
Appendix Table 2. Average Salary Offers to Bachelor’s Degree Candidates, by
Major Field of Study and by Job Function . . . . . . . . . . . . . . . . . . . . . . . . 19
An Information Technology Labor Shortage?
Legislation in the 106th Congress
Concern about an information technology (IT) labor shortage culminated during
the congressional debate over raising the ceiling on H-1B visas for skilled temporary
alien workers.1 The 105th Congress chose to raise the cap for 3 years on the
admission of nonimmigrant professionals who work in specialty occupations2 in Title
IV (the American Competitiveness and Workforce Investment Act, ACWIA) of P.L.
105-277 (the FY1999 Omnibus Consolidated and Emergency Supplemental
ACWIA also called on the National Science Foundation (NSF) to, among other
things, study and report on employers’ requirements over the next 10 years for
workers with computer-related skills, the future education/training needs of U.S.
students to ensure an adequate supply of workers at requisite skill levels and the
potential costs/benefits to the U.S. economy (i.e., to employers, workers and
consumers) from admission of foreign workers with science and engineering skills.
The study requirement suggested that disagreement persisted about the existence of
an across-the-board IT labor shortage. And, the imposition of a filing fee on
employers of $500 per H-1B visa petition to be used largely to fund technical skills
training and math, engineering or science education suggested ambivalence about the
preferred solution to the perceived shortfall of workers with computer-related skills.
This report presents information on the demand for and supply of IT workers
during the 1990s. It analyzes the adequacy of measures that arguably demonstrated
the presence of an IT labor shortage around the turn of the century. The ACWIA
provisions that the 105th Congress intended to expand the supply of appropriately
educated or trained U.S. workers as a means of fulfilling employers’ demand for IT
personnel are next set forth, along with subsequent changes to ACWIA made by the
106th Congress. The report closes with a description of legislation introduced during
U.S. Congress. Senate. Committee on the Judiciary. Subcommittee on Immigration.
Hearings on the High Tech Worker Shortage and U.S. Immigration. 105th Cong., 2d Sess.,
February 25, 1998. (Hereafter cited as Senate Committee on the Judiciary, High Tech
Worker Shortage); and U.S. Congress. House. Committee on the Judiciary. Subcommittee
on Immigration and Claims. Hearings on Immigration and America’s Workforce. 105th
Cong., 2d Sess., April 21, 1998. For a summary of the hearings and of P.L. 105-277, Title
IV see CRS Report 98-531, Immigration: Nonimmigrant H-1B Specialty Worker Issues and
Legislation, by Ruth Ellen Wasem.
In addition to computer-related jobs, employers have obtained H-1B visas for such
occupations as physical, occupational, and speech therapists; electrical engineers; auditors and
accountants; university faculty; and physicians and surgeons.
the 106th Congress that was meant to further increase the number of domestically
available workers with computer-related skills.
The Demand for and Supply of IT Workers
The U.S. statistical system does not include a direct measure of occupational
imbalances in the supply of and demand for workers.
The Demand Side
The U.S. Bureau of Labor Statistics (BLS) collects data on current employment
and estimates future employment by occupation. These statistics represent only the
demand side of the labor market. They understate the real demand for labor to the
extent supply limitations constrain employment growth.
Table 1. Employment in Selected IT Occupations, 1989-1999
(numbers in thousands)
Total, computerrelated occupations
Source: U.S. Bureau of Labor Statistics. Employment and Earnings, January issues.
Note: The data are derived from a survey of households, the Current Population Survey, and use the
Census Bureau’s occupational classification system.
Over 2.2 million persons worked in IT jobs as computer systems analysts,
computer engineers, computer scientists and computer programmers in 1999 —
twice the number in 1989, the last peak in the business cycle. As shown above in
Table 1, the ranks of computer systems analysts, engineers and scientists expanded
to a much greater extent (174% or 983,000 workers) than those of computer
programmers (22% or 124,000 workers). In contrast, total U.S. employment rose by
slightly less than 20% between 1989 and 1999.
Employment gains among computer programmers, who are classified in the
technicians and related support occupational division, did not exceed the average
across all occupations until the latter half of the period. The much flatter growth
trajectory of programmers vis-a-vis professional computer-related occupations (e.g.,
computer systems analysts) may be partly related to technological considerations,
such as “consolidation of systems and applications, developments in packaged
software, advanced programming language and tools, and the growing number of
users who design, write, and implement more of their own programs.”3 In addition,
computer programmer is a narrowly defined group; BLS assigns new, emerging IT
occupations to the computer systems analysts, engineers and scientists group within
the professional specialty occupational division.
Table 2. Employment, 1998 (Actual) and 2008 (Projected)
(numbers in thousands)
Total, all occupations
Total, IT occupations
Computer support specialists
All other computer scientists
Source: Braddock, Douglas. Occupational Employment Projections to 2008. Monthly Labor
Review, November 1999.
Note: The base year data are derived from a survey of firms, the Occupational Employment
Statistics (OES) survey, and use the OES rather than the Census Bureau’s occupational classification
Veneri, Carolyn M. Here Today, Jobs of Tomorrow: Opportunities in Information
Technology Occupations. Occupational Outlook Quarterly, fall 1998. (Hereafter cited as
Veneri, Here Today, Jobs of Tomorrow.)
BLS projects that IT employment could increase to almost 3.9 million workers
by 2008. (See Table 2 above.) As has been the case, the professional IT occupations
are expected to grow much more than the technical IT occupation of computer
programmers. Nonetheless, employment of programmers is projected to expand by
much more than the average occupation, 30% versus 14%, respectively.
These demand-side data are somewhat wanting in detail, however. The figures
do not indicate the level — entry or experienced — at which firms want to hire IT
workers. A subtext of the IT worker debate was whether demand exceeded supply
for all workers; for workers already trained in the latest, sophisticated skills; or for
young workers to whom firms could pay relatively low, entry-level wages. (See
Appendix for data on average salaries for all workers by IT occupation and on
average starting salaries offered to bachelor’s degree recipients with IT majors or for
IT jobs.) ACWIA called upon the National Academy of Sciences to assess and report
on the status of older workers in the IT field. (See section below on Selected Results
from the ACWIA-Mandated Studies.)
The Supply Side
The supply-side data are even more problematic than the demand-side data with
regard to providing insights about conditions in the IT labor market. Because there
rarely is just one source of workers for a given job, it is very difficult to estimate the
supply of labor to a particular occupation and hence, whether an imbalance exists
between supply and demand.
A common assumption is that individuals with bachelor’s degrees in
computer/information sciences are the major source of IT workers. This supposition
underlies the notion that the decline in bachelor’s degrees conferred in
computer/information sciences through the early 1990s means that, in a period of
rising demand, employers have faced a dwindling supply of new qualified workers.
However, as shown in Table 3, roughly one-half of computer programmers, systems
analysts, engineers and scientists in 1999 had completed bachelor’s degrees regardless
of field of study. Moreover, as already was shown in Table 1, the employment of IT
professionals doubled in the 1990s despite the downward trend in
computer/information sciences bachelor’s degree recipients early in the period.4 A
Commerce Department report that declared the existence of an IT labor shortage was
faulted for failing to adequately take into account sources of supply other than college
The National Center for Education Statistics collects data on degrees awarded in
computer/information sciences separately from data on degrees awarded in engineering (e.g.,
computer engineering). Degrees awarded in engineering exhibit the same declining trend as
degrees awarded in computer/information sciences. While a substantial share of bachelor’s
degree holders employed as computer engineers in 1995 had majored in engineering, large
proportions had majored in related or in unrelated fields. More specifically, among bachelor’s
degree recipients employed as computer software engineers, 35% had graduated with
engineering majors; 31% with computer/information sciences majors; 13% with mathematical
sciences majors; 11% with life, physical, social and related sciences majors; and 10% as
nonsciences/nonengineering majors. Veneri, Here Today, Jobs of Tomorrow.
graduates who had majored in computer/information sciences.5 And, the number of
bachelor’s degrees awarded in computer sciences reversed direction in the last few
years, increasing modestly between the early 1990s (24,200) and late 1990s (24,768).6
Table 3. Percent Distribution of Workers in Selected
Occupations by Highest Level of Education Completed, 1999
degree or PhD
Source: U.S. Bureau of Labor Statistics. Unpublished Current Population Survey data.
Includes a small number of workers without a diploma who completed less than 12 years of
Not only are persons with bachelor’s degrees in computer/information sciences
employed in non-IT jobs, but also persons with bachelor’s degrees in related and
unrelated fields of study work in IT occupations. For example, about 48% of
U.S. Department of Commerce. Office of Technology Policy. America’s New Deficit: The
Shortage of Information Technology Workers. Washington. September 29, 1997; and U.S.
General Accounting Office. Information Technology: Assessment of the Department of
Commerce’s Report on Workforce Demand and Supply. GAO/HEHS-98-106R.
Washington, March 20, 1998. (Hereafter cited as GAO, Information Technology:
Assessment of the Department of Commerce’s Report on Workforce Demand and Supply.)
National Center for Education Statistics. Digest of Education Statistics, 1999. NCES
2000-031. Washington, D.C., May 2000. (Hereafter cited as NCES, Digest of Education
bachelor’s degree recipients in 1992-1993 who majored in computer/information
sciences worked in non-computer occupations in April 1994. The same was true for
some 36% of bachelor’s degree recipients who majored in computer programming.
Alternatively, bachelor’s degree recipients who majored in the related fields of
engineering and mathematical sciences accounted for 19% of college graduates
employed as computer systems analysts in 1995, while bachelor’s degree recipients
who majored in nonsciences and nonengineering fields accounted for 28% of the total.
Similarly, bachelor’s degree recipients in related fields of study accounted for 26% of
college graduates employed as computer programmers, while bachelor’s degree
recipients in unrelated fields accounted for 9% of the total.7
Community colleges and private vocational institutions also prepare students for
IT jobs, as do company and military training programs. For example, associate’s
degrees accounted for 23.5% of the postsecondary degrees in computer/information
sciences conferred in 1996-1997.8 In addition, as shown in Table 3, 23.1% of
computer systems analysts, engineers and scientists as well as 31.8% of computer
programmers had completed postsecondary education below the bachelor’s degree
level in 1999. And, unlike the previously mentioned downward trend in bachelor’s
degrees awarded in computer science/information and computer engineering, the
number of associate (as well as masters and doctoral) degrees in IT and related
disciplines has been on the rise since the 1980s.9
Measures of Labor Market Conditions
Because neither demand nor supply can be measured directly, “the determination
of labor market imbalances must rely upon indicators ... The best-known example of
such an indicator is the unemployment rate ... Using only one indicator, however —
even one as relevant as the unemployment rate — can still lead to an incorrect
conclusion” about the existence of a labor shortage.10 Indicators in addition to those
previously discussed (i.e., the trend in employment and projected employment change)
include the occupational unemployment rate and the trend in wages. Vacancy
information also is an indicator, but such estimates are not currently part of the
government’s statistical system. BLS intends to reestablish a job openings/labor
turnover survey, but it will not yield measures of job vacancies by occupation.
BLS cautions that available labor market indicators should be combined with
information on the supply of labor to one or more related occupations (e.g.,
Veneri, Here Today, Jobs of Tomorrow. Note: The remaining major field of study (i.e., life,
physical, social and related sciences) accounted for 16% of systems analysts and 18% of
computer programmers with bachelor’s degrees.
NCES, Digest of Education Statistics.
U.S. Department of Commerce. Office of Technology Policy. The Digital Work Force:
Building Infotech Skills at the Speed of Innovation. Washington. June 1999. (Hereafter
cited as DOC, The Digital Work Force.)
Cohen, Malcolm S. Labor Shortages as America Approaches the Twenty-first Century.
Ann Arbor, Michigan, University of Michigan Press, 1995. p. 25. (Hereafter cited as Cohen,
educational attainment by field of study vis-a-vis skill requirements of employers) and
with “knowledge of the workings of the labor market” (e.g., whether there are lags
between rising demand and a supply response) in order to analyze existing or potential
labor shortages. “Conclusions about shortages should not be based on general labor
market statistics alone or anecdotal evidence alone.”11
The Unemployment Rate. While most observers would acknowledge that
the comparatively low jobless rates of IT occupations reflect a tight labor market,
fewer would agree that it proves a shortage. As shown in Table 4, the unemployment
rate in 1999 for computer programmers, at 2.3%, was identical to that of all
technicians. The jobless rate in 1999 for professional IT workers was 1.7%, just 0.2
percentage points below the rate for all professional workers. The unemployment
rates of IT workers in 1999 was virtually the same as in 1989 (see table note), the last
time there was concern about labor shortages generally.
Table 4. Unemployment Rates in Selected Occupations, 1989
All professional specialty occupations
Computer systems analysts, engineers & scientists
All technicians and related support occupations
Source: U.S. Bureau of Labor Statistics. Unpublished data from the Current Population Survey.
Note: Because of the fairly small number of workers in computer-related occupations, year-to-year
changes in their unemployment rates must be several tenths of a percentage point (0.6-0.9) to be
considered statistically significant.
An occupation’s unemployment rate is a tenuous measure of shortages because
it is based on only those in the labor force who report that their current or last job was
in that occupation.12 The jobless rate of computer programmers, for example, does
not pick up workers with programming degrees who either currently are employed
in other occupations or who are newly entering or reentering the labor force but have
not yet gotten a job, all of whom represent potential sources of supply. The jobless
rate of programmers also does not reflect individuals who may be qualified for such
work (e.g., math majors) but are now employed in another occupational group. In
Veneri, Carolyn M. Can Occupational Labor Market Shortages be Identified Using
Available Data? Monthly Labor Review, March 1999. p. 21.
Ibid. Note: The occupational unemployment rate is the number of workers without a job
in a given occupation as a percent of the labor force in that occupation. The labor force in a
given occupation is the number of workers currently employed in the occupation and the
number of unemployed workers whose last job was in the given occupation.
addition, it does not capture qualified workers who are out of the labor force (e.g.,
retirees with IT skills) who might reenter under the right circumstances. Thus, the
multiple paths to employment in many occupations not only make it difficult to
determine the supply of labor to particular jobs but also complicate the unemployment
rate’s usefulness as a conclusive indicator of an occupational labor shortage.
Vacancy Data. Two surveys conducted for the Information Technology
Association of America (ITAA) and from which IT vacancies were estimated were
heavily relied on to substantiate the existence of a labor shortage in the late 1990s.
ITAA surveyed 2,000 medium and large companies, of which 271 responded. (Firm
size was measured by number of employees or sales volume.) It was estimated from
this sample that roughly 190,000 IT positions were unfilled at all medium and large
firms in the United States, and that the vacancies represented about 10% of all IT
positions at these private sector firms.13 The Virginia Polytechnic Institute and State
University, in collaboration with the ITAA, surveyed 1,500 companies having at least
100 employees, and 532 responded to this survey. It was estimated that 346,000
positions for computer programmers, systems analysts and computer
engineers/scientists were unfilled at all U.S. companies with 100 or more employees
on their payrolls.14 The accuracy of either of these estimates is questionable,
however, because of the surveys’ low response rates (14% and 36%, respectively).15
A subsequent ITAA-commissioned survey used a much broader definition of IT
occupations.16 Firms (excluding non-profit organizations and government) reportedly
needed an additional 1.6 million IT workers in 2000, and a little over than one-half
(843,328) of the required positions would go unfilled due to a shortfall of qualified
IT workers. (The shortage measure was based on the proportion of applicants that
a nationally representative sample of 700 hiring managers at IT and non-IT for-profit
firms with at least 50 employees considered to be qualified for specific IT job
Information Technology Association of America. Help Wanted: The IT Workforce Gap
at the Dawn of a New Century, 1997.
ITAA and the Virginia Polytechnic Institute. Help Wanted: A Call for Collaborative for
the New Millennium, 1998.
GAO, Information Technology: Assessment of the Department of Commerce’s Report on
Workforce Demand and Supply.
In the two previous ITAA-sponsored studies, the following standard occupational categories
were used to define the IT workforce: computer programmers, systems analysts, engineers,
and scientists. In the latest report, a broader definition consisted of the following eight job
categories derived from skill standards of the NorthWest Center for Emerging Technologies
(NWCET) and an “other” category:
programmer/software developer; database
administrator/developer; web administrator/developer; network systems specialist; enterprise
information systems integrator; interactive digital media specialist; technical writer; and
computer systems (“tech”) support representative. (The NWCET was established in 1995
with an NSF grant and had as one of its goals the development of voluntary skill standards
that reflect the expectations of IT firms.)
categories.) Based on an estimated IT workforce of 10 million, a vacancy rate of
8.4% was calculated.17
Vacancies in general are an insufficient measure of labor shortage. An
occupation could have both a high unemployment rate — which suggests excess
supply — and a high vacancy rate — which suggests excess demand — if employers
and workers find it difficult to make matches due to imperfect information or due to
the different locations of job openings and qualified workers.18 The same situation
also could indicate a skills mismatch rather than a shortage per se if employers want
to fill vacancies with persons who already possess and have experience applying the
latest specialized skills (e.g., C++ expertise) rather than retraining their mid-career IT
employees who lack the currently hot skills. Evidence exists that companies, because
of short product-life and product-development cycles in IT-intensive industries, have
been pursuing a “buy” (from the external labor market) rather than a “make” (by
training employees within the firm) employment strategy.19
Moreover, data that cover one point in time do not indicate whether the current
level of IT vacancies is consistent with or differs from past conditions. A firm could
regularly have a high proportion of vacancies if it experiences high turnover or is
growing rapidly, or if vacancies take a long time to be filled.20 In addition to
information being absent on the duration of IT vacancies today and over time,
information also is lacking on the salaries associated with the vacancies. Such data
could help to demonstrate whether there is a shortage of qualified workers per se or
whether employers have not been raising wages sufficiently to attract enough workers
(e.g., from other occupations or from outside the labor force).
Wage Increases. If occupational demand is nearing or outstripping supply,
economic theory suggests that employers will bid up wages to attract workers.
“Thus, rapidly rising wages are consistent with a labor shortage.”21 The picture of pay
trends in the IT labor market is unclear, however. Private surveys show larger gains
than government surveys primarily due to differences in sample size and methodology.
The government sources are more likely to be accurate.22
Based on a source that compares hundreds of occupations over time, the relative
earnings trends of IT workers have varied by occupational group and by year. As
shown above in Table 5, the median weekly earnings of computer systems analysts,
engineers and scientists employed full-time in the private and public sectors rose by
42% between 1989 and 1999 — 4 percentage points more than the all occupations’
ITAA. Bridging the Gap: Information Technology Skills for a New Millennium. April
2000. (Hereafter cited as ITAA, Bridging the Gap.)
Cohen, Labor Shortages.
DOC, The Digital Work Force.
Testimony of Robert I. Lerman in Senate Committee on the Judiciary, High Tech Worker
Cohen, Labor Shortages, p. 33.
DOC, The Digital Work Force.
average. The differential is wider — at 10 percentage points — in the case of
computer programmers whose median earnings per week grew by 48% over the
period. After barely rising in 1998, the pay of computer programmers increased at
an above average rate in 1999 (6.5% and 5.0%, respectively). The increase in median
earnings of computer professionals (5.9%) in 1999 also surpassed the all occupations’
average gain (5.0%), but to a much smaller extent.
A compensation survey conducted for the ITAA by William M. Mercer found
that the average hourly pay of IT workers rose between 12% (for software
development architect) and 20% (for operating systems/software architect/consultant)
in one year alone, 1996. In contrast, a Computerworld survey found that in the same
year compensation changes ranged from -2% (for systems analyst, administration) to
0% (for senior systems programmer) to 5% (for chief information officer (CIO),
CIO/vice president of IT or information systems). Substantial salary increases did not
occur until 1997, according to this source, when CIOs averaged a 28% gain and at
the opposite end of the spectrum, senior systems programmers’ compensation grew
Table 5. Median Weekly Earnings of Full-Time Wage and Salary
Workers in Selected IT Occupations, 1989-1999
Computer systems analysts,
engineers and scientists
Source: U.S. Bureau of Labor Statistics. Employment and Earnings, January issues.
Melymuka, Kathleen. We’re in the Money. Computerworld, 11th Annual Salary Survey,
September 1, 1997. [http://www.computerworld.com].
According to data collected quarterly by the National Association of Colleges
and Employers (NACE), average starting salaries offered to graduates with bachelor’s
degrees in computer science rose by 54% between the September 1994 and 2000
surveys. Much of that increase occurred in one year alone, 1998, when the entry-level
wage offer for new graduates with bachelor’s degrees in computer science spiked by
12.7%.24 (For starting salary offers in other IT fields see Appendix.)
There was some evidence that pay raises in IT occupations may have slowed
somewhat in the waning years of the 1990s. Computerworld’s annual salary survey
found that the typical increase in IT salaries was between 4% and 5% in both 1998
and 1999. In 1997, it had been 11%. But, some IT jobs did continue to command
large pay raises (e.g., 20% for the average CIO and 11% for the average IT director
in 1999). The more subdued growth rate in IT pay partly may have been due to
having fewer workers tied up with Y2K-compliance efforts. In addition, firms
reportedly were expanding their (re)training of existing employees and thereby
internally growing the pool of IT workers.25 From a more long-term supply
perspective, enrollment in computer-related programs at 4-year colleges began to
Selected Results from the ACWIA-Mandated Studies 27
The National Research Council released a report in October 2000 which
complied with P.L. 105-277’s requirement of a study by the NSF concerning IT
shortages and by the National Academy of Sciences concerning older IT workers.
After analyzing the best available statistics, the Council’s Committee on Workforce
Needs in Information Technology found them wanting in terms of yielding a definitive
determination of whether an IT shortage existed. The committee chose to call the IT
labor market a tight one, and it expected the market — especially for IT occupations
in which there is a lengthy educational process — to remain tight for the foreseeable
future. With regard to older workers, the committee also found the data inadequate
to conclude whether employers of IT workers illegally engaged in age discrimination.
The report acknowledged that without H-1B workers to fill IT positions, the
sector would likely have experienced slower growth and that reliance on “foreign
Starting salary offers to new computer science graduates (bachelor’s degree level) averaged
$31,783 in September 1994, $33,712 in September 1995, $35,222 in September 1996,
$37,216 in September 1997, $41,949 in September 1998, $44,649 in September 1999 and
$49,055 in September 2000. NACE, Salary Survey, various issues.
Fryer, Bronwyn. Return to Sanity. Computerworld, 13th Annual Salary Survey.
September 6, 1999 and Goff, Leslie. The E-lusive Staff. Computerworld, 4th Annual Hiring
Forecast Survey. Note: For information on IT hiring managers’ views about the relative
effectiveness of different sources of skill development (e.g., on-the-job employer-provided
training) see ITAA, Bridging the Gap.
Bachelor’s degree enrollments in the fields of computer science/computer engineering have
risen over 108% between 1996 and 1998. DOC, The Digital Work Force.
National Research Council. Building a Workforce for the Information Economy.
Executive Summary from a prepublication copy dated October 21, 2000.
workers will continue to be necessary for the immediate future.” The relatively large
number of H-1B visa holders in IT jobs put “nonnegligible” downward pressure on
wages, according to the committee. In other words, the availability of H-1B workers
kept the pay of IT workers from increasing as much as it otherwise would have in a
tight labor market.
Although the committee could not find an analytical basis upon which to
determine the “proper” level of H-1B visas, it offered several recommendations (e.g.,
the provision of training incentives to employers as well as the collection of more
timely, disaggregated data on the IT labor force in the United States). Other
recommendations dealt specifically with employment-based immigration policy and
with the federal government as an employer of IT workers.
Legislation and Education/Training Programs
The 105th Congress
In addition to temporarily importing more already qualified workers to alleviate
the tight IT job market, some Members of the 105th Congress wanted to promote a
home-grown remedy to the perceived shortfall of computer-related workers.
Consequently, P.L. 105-277 included a provision that requires firms filing petitions
to bring into the country, to extend the stay of, or to hire from another U.S. employer
nonimmigrant professionals on or after December 1, 1998, but before October 1,
2000, to pay a $500 user fee per petition.
Under P.L. 105-277, the fees deposited in the H-1B Nonimmigrant Petitioner
Account initially were allocated as follows:
The U.S. Department of Labor (DOL) is to issue grants amounting
to 56.3% of the fees for demonstration programs under the Job
Training Partnership Act (JTPA, Section 452(c)) or its successor, the
Workforce Investment Act of 1998 (WIA, Section 171(b)), to
establish demonstration programs and projects that provide training
in technical skills to employed and unemployed workers.The National
Science Foundation (NSF) is to use 28.2% of the fees for
scholarships to low-income students enrolled in programs that confer
associate, undergraduate, or graduate degrees in mathematics,
engineering or computer science.
The NSF is to expend 4% of the total fees to award merit-reviewed
grants under the National Science Foundation Act of 1950 (Section
3(a)(1)) for programs that provide opportunities for enrollment in
year-round K-12 academic enrichment courses in mathematics,
engineering or science; and 4% to carry out systemic reform
activities in K-12 under Section 3(a)(1) of the 1950 Act.The
remaining fees are to go to the Attorney General to reduce the
processing time of H-1B petitions and to improve the enumeration
of nonimmigrant workers, including the submission of periodic
reports to the House and Senate Committees on the Judiciary
(1.5%), and to the DOL to reduce the processing time of visa
applications and for enforcement (6%).
(See the section below on the 106th Congress for information about how it changed
the user fee’s level and allocation formula.)
DOL’s 56.3%. In the August 16, 1999 Federal Register (p. 44543-44554), the
Employment and Training Administration (ETA) announced the availability of grant
funds from ACWIA’s user fee to private industry councils (PIC) under the JTPA,
local workforce investment boards under the WIA, and regional consortia of PICs or
local boards for technical skills training programs in high-demand occupations (e.g.,
those for which employers have submitted H-1B applications)28 targeted at employed
and unemployed persons. There is a 50% non-federal matching requirement.
Although the federal grant must be used only for training services, applicants can
count non-federal resources for supportive services (e.g., transportation or child care)
toward the matching requirement.
In February 2000, the ETA announced the first-round awards of $12.4 million
in nine grants for the provision of training to some 3,000 workers. The typical grant
award was $1.5 million.29 The second-round grant solicitation appeared in the March
29, 2000 Federal Register (p. 16658-16670). In July, a total of $29.1 million was
awarded in 12 grants to serve approximately 5,000 workers. The awards ranged from
a little more than $900,000 to $3.0 million. The third-round grant solicitation was
issued August 1, 2000 (see Federal Register, p. 46958-46969). Some $54 million
was awarded in October 2000 to fund 22 grants that ranged from almost $1 million
to about $2.8 million.
NSF’s 28.2%. ACWIA authorized the Computer Science, Engineering, and
Mathematics Scholarships (CSEMS) program which is to award funds to accredited
institutions of higher education. They, in turn, are to select students for a 2-year
CSEM scholarship of up to $2,500 per year per student. Low-income,30 academically
talented students must be in associate, bachelors or graduate degree programs in
computer science, computer technology, engineering, engineering technology, or
mathematics to be eligible for scholarships. They must be pursuing their studies on
a full-time basis and be U.S. citizens, U.S. nationals, refugee aliens or permanent
About 280 proposals from public and private schools were submitted by the
August 30, 1999 deadline for first-round award applications. The NSF subsequently
announced the provision of $22.49 million to 114 institutions which will enable each
to offer about 40 scholarships annually over a 2-year period. The institutions’ total
Appendix A of the solicitation lists selected H-1B professional, technical and managerial
occupations (and fashion models) by number of job openings certified by the DOL.
More information on grant awards can be found at the following internet website:
Determination of financial eligibility relies on U.S. Department of Education criteria for Pell
grants and for Graduate Assistance in Areas of National Need.
awards include funds for administrative and other expenses.31 Applications for
second-round grants were due to the NSF by August 3, 2000 and awards were made
during Spring 2000. About $24 million went to 110 institutions.32
P.L. 105-220. Subtitle C of WIA created The Twenty-First Century
Workforce Commission to study the skills needed to obtain IT jobs, explore ways to
increase the supply of IT workers, and compare the success of U.S. and foreign
programs at training individuals for IT employment. The Commission’s report was
released in June 2000. Among other things, it endorsed not only preparing more U.S.
students and upgrading the skills of incumbent workers to fill IT jobs, but also raising
the H-1B visa cap as a response to skill shortages in conjunction with increasing the
fee for employers who file H-1B petitions. As it turned out, this was the path
ultimately chosen by the 106th Congress.
The 106th Congress
Bills introduced during the 106th Congress to alleviate the tightness in the IT
labor market took several tacks, which are described below. The one ultimately
settled on saw the Congress returning to immigration policy as a remedy. This
approach was prompted, in large part, by an announcement from the Immigration and
Naturalization Service (INS) that the 115,000 limit on H-1B visas for all of FY2000
effectively had been reached in March 2000. In FY1999, the 115,000 cap had not
been hit until June 1999.
Bills Amending ACWIA’s User Fee Level and Allocation Formula.
Several bills were offered that included raising the user fee and altering its allocation
formula. Both the House and Senate ultimately passed S. 2045 (the American
Competitiveness in the Twenty-First Century Act of 2000) and the President signed
it in October 2000. Among other things, P.L. 106-313 raised the cap on H-1B visas
to 195,000 annually between FY2001 and FY2003 while making additional visas
available for FY1999 and FY2000. It exempted from the cap aliens employed by
institutions of higher education, nonprofit research organizations or governmental
research organizations. P.L. 106-311, enacted on October 17, 2000, raised the user
fee from $500 to $1,000 for those petitions filed 2 months on or after the date of
enactment. (For more information on other changes specific to immigration policy
The total grant award includes up to 5% of the total scholarship amount for student-support
infrastructure (e.g., recruitment of students from groups underrepresented in CSEM fields
including women, racial/ethnic minorities, and persons with disabilities; retention of CSEMS
recipients to degree completion; and support in employment placement) and up to 5% for
project management and administration (e.g., confirmation of scholarship applicant’s
eligibility and evaluation of program outcomes). Note: For information on the
underrepresentation of certain groups in the IT labor force, see Freeman, Peter and William
Aspray. The Supply of Information Technology Workers in the United States. Washington,
D.C., Computing Research Association, 1999; National Science and Technology Council.
Ensuring a Strong U.S. Scientific, Technical, and Engineering Workforce in the 21st
Century. April 2000; and National Science Foundation. Women, Minorities, and Persons
with Disabilities in Science and Engineering: 1998. NSF 99-87, February 1999.
For more information see: [http://www.ehr.nsf.gov/ehr/due/programs/csems/].
see CRS Report RL30498, Immigration: Legislative Issues on Nonimmigrant
Professional Specialty (H-1B) Workers.)
P.L. 106-313 amended the allocation of funds in the Nonimmigrant Petitioner
Account in the following ways:
The DOL’s share for training fell from 56.3% to 55.0%.
The NSF’s share for the previously described CSEMS program fell
from 28.2% to 23.5%. The level of scholarship awards rose from
$2,500 to $3,125 (the Pell Grant amount), with scholarships
renewable for up to 4 years.
An additional 15% went to the NSF to carry out a direct or matching
grant program to support private-public partnerships in K-12
education in lieu of the 8% for K-12 activities enumerated in
ACWIA. (Funds for the systemic K-12 reform activities referenced
in ACWIA were subsumed in the new, broader NSF program.)
The Attorney General continued to get 1.5% to reduce the
processing time of H-1B petitions and to improve the enumeration
of nonimmigrant workers, while the DOL’s share for reducing the
processing time of H-1B applications dropped from 6% to 5%.
As in ACWIA, the legislation directed the DOL to use its 55.0% share to create
demonstration programs to provide technical skills training for both employed and
unemployed workers. P.L. 106-313 specified that the training need not develop skills
commensurate with a 4-year college degree and that it should prepare workers for a
range of occupations along a career ladder. Further,
Consideration shall be given to the use of grant funds to demonstrate a significant
ability to expand a training program or project through such means as training
more workers or offering more courses, and training programs or projects resulting
from collaborations, especially with more than one small business or with a labormanagement training program or project.
P.L. 106-313 directed the Secretary of Labor, in consultation with the Secretary
of Commerce, to award
75% of the grants to WIA’s local workforce investment boards or
consortia of such boards in a region. These grants require a 50%
non-federal match. (Under ACWIA, 100% of the DOL grants had
to go to workforce investment boards or consortia.)
25% of the grants to partnerships consisting of at least two
businesses or a business-related nonprofit organization that
represents more than one business. These grants require a 100%
Further, at least 80% of all DOL grants from the user fee must be awarded for
skills training in high technology, information technology and biotechnology (e.g.,
skills required for “software and communications services, telecommunications,
systems installation and integration, computers and communications hardware,
advanced manufacturing, health care technology, biotechnology and biomedical
research and manufacturing, and innovation services”). And, no more than 20% of
the grants shall go toward training persons for skills in a single specialty occupation
(as defined in Section 214(i) of the Immigration Nationality Act).
P.L. 106-313 also required the DOL and NSF to track and monitor the
performance of programs funded from the Nonimmigrant Petitioner Account. The
two agencies are to submit a report on the programs’ performance to the House and
Senate 1 year after the date of enactment.
Using Tax Incentives. Companion bills H.R. 838 and S. 456 would have
amended the Internal Revenue Code (IRC) to allow an employer an income tax credit
for expenses that he paid or incurred in connection with the provision of technology
training. The IT training credit would have been equal to 20% of the employer’s
annual IT training program expenses. If the IT program were operated in certain
areas (e.g., an urban or rural empowerment zone/enterprise community, a school
district where at least one-half of the students are eligible for subsidized lunches, or
a disaster area) or by a small employer (i.e., 200 or fewer employees on the payroll),
the rate would have been 5 percentage points higher. The maximum amount of IT
program expenses employers could have taken into account when calculating the
credit was $6,000 per individual for the taxable year. IT program expenses were
defined as a program that trains computer programmers, systems analysts, and
computer scientists or engineers; involves a partnership between employers and state
training programs, school districts, universities or certified commercial IT training
providers; and that entails the employer paying or incurring at least 50% of the costs.
The IT training credit would have been part of the general business credit under the
H.R. 5004 (Technology Education and Training Act of 2000) also would have
amended the IRC to provide a credit for 100% of the expenses paid or incurred by
taxpayers in connection with IT training programs. Although the maximum amount
of IT program expenses per individual generally would have been $1,500, it could
have gone up to $2,000 if the training programs were operated in empowerment
zones/enterprise communities, school districts in which at least one-half of the
students are eligible for subsidized lunches, disaster areas, rural enterprise
communities, rural economic area partnership zones, or if the programs were operated
by small employers (i.e., those who had 200 or fewer employees daily in each of 20
or more calendar weeks in the current or preceding calendar year). The amount of
the credit would have been reduced for employees whose employers paid or incurred
qualified training expenses. IT training program expenses were those that lead to an
industry-accepted IT training certification for the participant. The NSF, through a
newly created advisory group, would have developed a list annually of the latest
industry certifications that could qualify for the credit.
Other Bills Unrelated to ACWIA’s User Fee and Education/Training
Provisions. H.R. 709 (The Technology Education Capital Investment Act of 1999)
would have authorized the appropriation of funds to involve the NSF, Secretary of
Education, Secretary of Commerce, and a Technology Workforce Commission in an
effort to alleviate the perceived shortage of IT workers. Under the legislation, the
NSF would have expanded its informal science and math education programs (i.e.,
science education provided outside a formal school setting) as well as its national
advanced scientific and technical education program under Section 3(a) of the
Scientific and Advanced Technology Act of 1992. The Secretary of Education would
have made grants to states for the purpose of awarding scholarships to students in
programs leading to a postsecondary degree in science, math, engineering or a related
field. The Secretary of Commerce would have made grants to institutions of higher
education so that they could develop industry-sponsored internship programs which
give undergraduate engineering students the opportunity for hands-on training at local
businesses. The Technology Workforce Commission would have examined the
causes of and possible solutions for the IT labor shortage as well as the comparative
efficacy of programs in the United States and other countries to increase the supply
of IT workers, especially those programs that offer secondary or postsecondary
education other than a 4-year bachelor’s degree. It would have issued a report to the
President and the Congress.
H.R. 1265 (The Mathematics and Science Proficiency Act of 1999) would have
authorized a demonstration project through the NSF to encourage interest in the fields
of math, science and IT. Under the bill, the Director of NSF would have awarded
grants to local educational agencies in five urban and five rural areas so that they
could develop an IT program which builds upon or expands math, science and IT
curricula; purchase necessary equipment for such a program; and provide teacher
training in the IT, math and science fields. Maximum grant awards could not have
exceeded $300,000. Grant applications would have had to include assurances of
agreements with private sector representatives for such things as donation of
computer hardware and software, establishment of internship and mentoring
opportunities for students who participate in the IT program, and donation of
scholarship funds for students who have participated in the IT program. The Director
would have been required to assess the effectiveness of the activities carried out under
this legislation and to conduct a longitudinal study of students who received
scholarships including the number of students who graduated from institutions of
higher education with degrees in math, science or IT and the number of graduates
who took jobs in these fields. The bill authorized to be appropriated to the NSF
$3,000,000 to carry out the Act.
Appendix: The Older Worker Issue
As previously mentioned, the debate over whether there was an IT labor
shortage around the turn of the century is related to the kind of workers employers
wanted versus those who were available. Some asserted that firms wanted younger
workers who were well-schooled in the latest IT skills, who were more willing to put
in very long hours allegedly because they have not yet taken on non-work
responsibilities (e.g., raising a family) and who could be paid less than more senior IT
workers. IT-intensive firms countered that for reasons of competitiveness they could
not take the time, nor did they have the staff available, to train mid-career IT
employees in currently hot IT skills. Moreover, given the tightness of the labor
market, companies argued that the employees to whom they provided training might
well then take jobs with other firms.
The first table below shows the earnings of all workers, regardless of age or
experience, in selected IT occupations. The second table relates to starting salaries
offered (1) to students graduating from bachelor’s degree programs in IT-related
disciplines regardless of the occupation they are entering, and (2) to students
graduating from bachelor’s degree programs, regardless of the field of study, who are
being hired into IT-related jobs. A comparison of the data in the two tables not
unexpectedly reveals that the starting salaries of presumably inexperienced, young
workers typically are below those of all workers on average.
Appendix Table 1. Average Wages in Selected IT
Computer hardware engineers
Computer & information scientists, research
Computer software engineers, applications
Computer software engineers, systems
Computer support specialists
Computer systems analysts
Network & computer systems administrators
Network systems & data communications
Source: BLS’ Occupational Employment Statistics survey database.
Note: Data for 1999 reflect the redesigned Standard Occupational Classification system and are not
comparable with figures for prior years.
Appendix Table 2. Average Salary Offers to Bachelor’s Degree
Candidates, by Major Field of Study and by Job Function
Major field of studya
Software design & development
Hardware design & development
Systems analysis & design
Source: National Association of Colleges and Employers (NACE). Salary Survey, various fall
Note: The September 2000 salary survey, for example, reports on information the NACE received
September 1, 1999 and August 7, 2000 about starting salary offers made to students graduating with
bachelor’s degrees between September 1, 1999 and August 31, 2000. The offers are a representative
sample of actual job offers made to new college graduates during the recruiting year. The survey is
published four times a year.
Graduates may have entered any occupation.
Graduates may have had any major field of study.