Research and Advances
Computing Applications Contributed articles

How Offshoring Affects IT Workers

IT jobs requiring interpersonal interaction or physical presence in fixed locations are less likely to be sent out of the country.
  1. Introduction
  2. Data and Methods
  3. Results
  4. Occupational Attributes
  5. Discussion
  6. Acknowledgments
  7. References
  8. Authors
  9. Footnotes
  10. Figures
  11. Tables
  12. Sidebar: key insights
  13. Sidebar: Key Survey Questions
offshoring illustration

Though the outsourcing of IT services has long been a topic of academic interest,22 the potential for the global sourcing of IT services to have a long-term effect on the domestic IT work force continues to attract significant interest from the media, public, and academic community.1,3,6,15,19,24 Here, we use survey data collected in 2007 to characterize the effect offshoring has had on the U.S. IT work force (see the sidebar “Key Survey Questions“) and estimate how it will affect the demand for skills among U.S. IT workers in the future.

Understanding the effect of offshoring on domestic employment is potentially important for anticipating the training needs of existing and future IT workers and for enabling policymakers to frame initiatives that ease the transition to a global IT work force. However, our current understanding is limited by a paucity of data on firms’ offshoring activities. Most discussion of offshoring relies on anecdotes, press reports, and theoretical arguments. Indeed, the U.S. government acknowledges development of better offshoring data is a pressing policy concern.9

The primary contribution of this study is the collection and analysis of data describing how offshoring affects the U.S. work force. That data comes from two complementary, unusually large surveys carried out in late 2007, one involving 3,014 human resources managers and the other more than 6,000 U.S. workers employed in a variety of occupations. The data allows us to provide general statistics about the overall rate of U.S. IT offshoring and address two main questions: Do the rates of IT worker offshoring differ significantly from the offshoring rates for workers in other occupations? And is the pattern of IT offshoring consistent with the theory that jobs are less readily offshored if they require face-to-face contact with U.S.-based consumers or co-workers or require employees to perform hands-on work with U.S.-based assets?

Our interest in the second question was motivated by work suggesting that job characteristics (such as the need for customer contact or physical presence or information intensity) are closely related to the potential rate of offshoring.2,4,11,18 In the study, we combined data on offshoring-related displacement by occupation with Blinder’s classification4 of “offshorability” of various occupations to understand how job characteristics correlate with offshoring rates.

About 15% of all firms and 40% of technology firms we surveyed engaged in some offshoring activity, with about 30% offshoring IT workers. About 8% of IT workers reported having been displaced due to offshoring, more than twice the percentage of any other type of employee in the survey. However, this rate implies an annual displacement rate of about 1% per year, a relatively small fraction of the annual worker turnover rate in the U.S. economy. In addition, the offshoring of some IT occupations (such as programmers and software developers) was especially likely to be associated with domestic job displacement. Other occupations requiring more interpersonal interaction (such as systems analysts) were less likely to be offshored, and overseas employment in other occupations (such as sales and management) may have been directed at serving offshore customers and therefore were also less likely to be associated with job displacement in the U.S.

We make three separate contributions toward understanding how offshoring affects domestic IT workers: quantify the extent to which offshoring has affected IT workers; show a relationship between occupational attributes and offshoring-related displacement, providing empirical support for emerging theories of how offshoring drives the global disaggregation of skills2,4; and contribute to the literature demonstrating the growing importance of managerial and interpersonal skills for IT workers.13,18,21

Back to Top

Data and Methods

Our primary data comes from two separate questionnaires, both administered in the winter of 2007 by a third-party survey firm on behalf of one of the largest online recruitment and career-advancement companies in the U.S. The first focused on the offshoring practices of 3,016 individual firms, including whether and why they offshore and what types of work and to what countries they offshore. It was conducted within the U.S. among hiring managers and human-resource professionals employed full-time with significant involvement in hiring decisions. Respondents were also asked about firm characteristics (such as size and industry). The second was administered to individual workers and included questions relating to whether or not they had been displaced due to offshoring. It was also conducted within the U.S. (online) among 6,704 employees employed full-time and included both firm characteristics (such as size and industry) and employee characteristics (such as age, salary, and job level).

To test the hypothesis that job characteristics affect the likelihood of a job being offshored, we used probit models in which the dependent variable was 1 if an employee reported being displaced due to offshoring or an employer reported offshoring a particular type of work; we also included a measure of the importance of face-to-face contact or physical presence as an independent variable. Rather than restrict our sample to IT workers, we included all occupations in the analysis to increase the variation in the skill content of jobs, employing Huber-White robust (clustered) standard errors to account for possible random firm effects.

We captured the importance of face-to-face contact and physical presence in our regression models by including index values computed in a study of the offshorability of various occupations.4 Blinder’s index is derived by placing jobs into categories depending on whether they require face-to-face interaction (such as child-care workers) and whether they require workers (such as in the construction trades) to be in a particular location. To maintain consistency with Blinder’s classification, we adopted the term “personally delivered” or “personal” services to describe tasks requiring customer contact or physical presence and “impersonal” services to describe tasks requiring neither of these characteristics. Higher values on this scale indicate workers in these jobs provide fewer personally delivered services, or “impersonal” jobs, and are therefore more likely to be offshored, all else being equal.

We also included additional variables in our regressions to control for other factors that might affect an employee’s chances of being displaced due to offshoring. Since the relative benefit of offshoring a particular worker depends on the cost of the worker to the firm, we included a measure of employee salary (coded in discrete levels). Employees are less likely to be displaced if they have more firm-specific knowledge or experience with the firm. Though we did not have access to organizational tenure variables, we included the individual’s job level, coded in discrete levels. We included demographic variables for employees, as there is evidence that such factors as race, age, and gender influence displacement; see Kletzer12 for a review of the job-displacement literature. We also included the number of employees at the firm to control for firm size, as well as a dummy variable indicating the industry in which the firm competes. Finally, in some regressions, we also included the state within the U.S. in which the firm operated in 2007, to control for regional differences.

Back to Top


Here, we present some statistics and results from the regression analyses aiming to identify the factors that most affect offshoring:

Employer statistics. Table 1 reports the overall incidence of offshoring by industry in 2007. The proportion of firms that reported offshoring any type of work across all industries was 15.2%. However, within technology-services and telecommunications industries, over 40% of firms in the sample reported offshoring some type of work. The hypothesis that offshoring is more common in high-tech industries than in other industries is significant at the p<.01 level (X2(1)=100.5). The figure here shows that offshoring rates vary significantly by job type. Over 30% of respondents reported offshoring computer programmers and software developers, but only about half of them, or 15.5%, reported offshoring systems analysts. About 24% of employers offshore customer service, and a smaller percentage (less than 10%) offshore management, sales, and marketing functions. A test of the hypothesis that employers offshore IT workers more than other types of workers is significant at the .01 level (X2(1)=86.6). Among IT workers, the hypothesis that computer programmers and software developers are offshored in greater numbers than systems analysts is significant at the p<.01 level (X2(1)=30.9).

Employers also reported offshoring different types of work for very different reasons. A test of the hypothesis that occupation and reason for offshoring are independent is rejected at the p<.01 level (X2(36)=165.2). Table 2 lists correlations between type of work and reasons for offshoring it. Jobs involving close interaction with the markets they serve (such as management and sales) are offshored for quality reasons in firms that are expanding geographically, while firms appear to offshore computer and technical work and customer service jobs primarily for cost savings and for access to skills. These correlations suggest that offshoring may have more direct implications for U.S.-based IT workers than the offshoring of other types of workers.

Table 3 further supports these ideas, showing firms’ offshoring destinations, as well as correlations between destination and type of work being offshored. A test of the hypothesis that occupation and offshoring destination are independent is rejected at the p<.01 level (X2(171)=298.2), indicating that particular types of work are best suited for offshoring only to certain countries. IT work and customer-service work appear to be much more concentrated than sales, management, and marketing, which are spread over a larger number of countries, consistent with the idea that jobs involving personally delivered services are often co-located with overseas customers. In 2007, India was the most popular destination for offshoring any type of work, especially for IT work, and, of all the countries in our sample, offshoring to India was most associated with cost savings.

Employee statistics. The findings from the employer data are supported by statistics from the employee surveys. Table 4 includes the percentage of workers, by occupation, who reported having been displaced due to offshoring. Across all occupations, slightly over 4% of workers reported having been displaced due to offshoring. Of occupations with at least 100 observations in the sample, engineers, machine operators, and IT workers reported the highest rates of offshoring-related job displacement. Of the five occupations with the highest displacement rates, all but machine operators were technology-related. Furthermore, unlike computer jobs, which in 2007 were increasing as a proportion of employment worldwide, machine-operator employment numbers have declined, suffering from unusually high displacement rates.10,16 These results support the common perception that U.S. IT workers have experienced higher rates of offshoring-related displacement than other U.S. workers.

Table 5 compares the displacement frequency of IT workers with that of all other types of worker. At about 8%, IT workers were (in 2007) displaced at twice the rate of other workers. A test of the hypothesis that displacement rates differ between IT workers and non-IT workers is significant at the .01 level (X2=27.5, p<.01). However, because these numbers reflect the percentage of workers who have ever been offshored, an 8% displacement rate implies an annual average offshoring-related displacement rate of about 1% for IT workers, assuming that many U.S. firms began offshoring in 2000.a Surveys conducted 1995–2005 suggest average IT staff turnover rates vary from 10% to 15%.7,8

Among occupations with at least 100 observations, we observed the lowest displacement rates among sales representatives and nurses. Moreover, a number of occupations with fewer observations (such as real estate agents, veterinarians, professors, and religious professionals) reported no offshoring displacement. In addition to providing a basis for comparison with the IT worker population, data from these other occupations provides preliminary support for the hypothesis that employees providing more personally delivered services are generally less vulnerable to offshoring than occupations that need not be in a fixed location (such as computer programmers and machine operators).

Back to Top

Occupational Attributes

We provide a more rigorous test of the hypothesis that “impersonal” jobs are more vulnerable to offshoring. Table 6 reports the results of some regressions using the survey data. The unit of observation in the regression is the employer-occupation offshoring combination, taking a value of 1 if the employer reported offshoring a particular occupation and 0 otherwise. Column (1) lists the results from a regression that includes all firms in our sample. Employers were less likely to offshore jobs in which employees provide personal services (t=9.0). The size of the employer (t=2.0) and the local cost of doing business (t=3.5) for the employer both significantly increase the probability that it will offshore a particular job. Column (2) lists results from a regression that includes only the firms that reported offshoring work in 2007. The extent to which an employee provided personal services is still significantly and negatively associated with whether a job was off-shored (t=14.0).

Column (3) adds a covariate indicating whether the employer was exapanding geographically, along with an interaction term between geographic expansion and personal interaction. Our estimates indicate that firms expanding geographically are more likely to hire offshore workers (t=6.71). Furthermore, geographic expansion moderates the type of work being offshored, consistent with the hypothesis that provision of personal services must be co-located with the markets being served. If a firm’s customers are all located in the U.S., it will offshore jobs that do not require personal interaction with the U.S. market. However, if a firm does business in overseas markets, employers will also hire offshore workers who can provide personal services directly to overseas customers.

Table 7 reports the results of the primary regressions from the survey data relating the personal services provided in one’s occupation to the likelihood of offshoring-related displacement. The results in column (1) support the hypothesis that employment in a job providing personal services significantly decreases the likelihood of being displaced due to offshoring (t=3.5). Somewhat surprisingly, the coefficient estimate on salary level is negative and significant, suggesting workers with higher salaries are less likely to be offshored (t=2.09). However, in the absence of human-capital data, the salary term in the regressions also reflects human-capital variables (such as education and experience). We therefore interpret the negative coefficient as indicating that, conditional on job level, workers with more human capital are less likely to be offshored, an effect that dominates any direct gains from offshoring more expensive workers. The results also suggest that older workers (t=4.9), males (t=2.21), and workers in simpler jobs (t=2.15) requiring less firm-specific capital are likely to be offshored.

After including industry dummies in column (2), the coefficient estimate on gender is no longer significant, indicating our earlier estimate on gender may have reflected high offshoring intensity in such industries as IT with a higher fraction of men and low offshoring intensity in such industries as health care with a higher fraction of women. However, the estimates on the other coefficients remain significant. Dummy variables for state and race in column (3) do not significantly alter the coefficient estimates on any other variable.

Column (4) shows the marginal effects of the estimates from our baseline regression in column (1) where all variables are standardized so effect sizes are comparable. A one standard deviation increase in our personal-services index measure decreases the probability of offshoring-related displacement by about 1%, a 25% increase over the U.S. national base rate of 4% in 2007. The effect of a one standard deviation decrease in our personal-services index appears to be similar in magnitude to a one standard deviation increase in age, also increasing the likelihood of offshoring-related displacement by slightly over 1%. Older workers who do not provide personal services are thus particularly vulnerable to offshoring-related displacement.

The sample in column (5) is restricted to IT workers. The measure on personal interaction is insignificant because there is little variation in this index within the IT workers in our surveyed population. Of the remaining variables, only age is significant, suggesting that among IT workers, older workers are at the greatest risk of offshoring-related displacement (t=3.10).

Table 8 explores how the level of personal interaction in a prior job affects outcomes for workers after being displaced due to offshoring. Employees in occupations providing fewer personal services are more likely to be separated from their employers, while those providing more personal services are more likely to be retained for other positions (t=2.0). This suggests that firms might retain and move workers with interpersonal or management skills not as easy to source globally. Column (2) shows how much change in the personal-delivery measure affects retention. A one standard deviation increase in the personal-skills index increases the chance of being retained by a firm by about 6%.

Back to Top


Although about 15% of firms in the U.S. offshored in 2007, firms in hightech industries offshored at rates higher than 40%, and IT work was the most commonly offshored type of work. IT workers in the U.S. have experienced offshoring-related displacement at a rate of 8%, more than double the percentage in other occupations. Firms offshore for a number of reasons, but IT workers appear to be offshored primarily for cost or access to skills. Therefore, compared to sales workers offshored to provide customer contact to overseas markets, the offshoring of IT workers should lead to greater displacement of U.S.-based IT workers. Our results also provide empirical support for the hypothesis proposed in earlier work4 that employees in jobs requiring face-to-face contact or physical presence in a fixed location are less likely to be offshored. This suggests that IT workers are especially vulnerable to offshoring because IT jobs generally require less customer contact or interaction with fixed physical assets.

Our estimates imply an average displacement rate of about 1% per year for U.S.-based IT workers. However, as offshoring grows more popular, our findings, which suggest that workers who do not provide personal services are being displaced at a higher rate, are consistent with emerging work providing evidence for a potentially significant long-term shift in the relative demand for skills within the IT labor market.23 These results suggest that technical occupations reliant on skills that can be delivered with relatively little face-to-face contact are more easily offshored. Other scholars have noted that interpersonal or managerial skills are increasingly valuable for IT workers,13,17 so our findings suggest that offshoring will continue to drive a secular increase in the direction of this trend. IT workers concerned about offshoring-related displacement may find more robust career paths in IT professions that require personal delivery.

These results also have policy implications. First, the relatively low level of offshoring suggests any policy prescription for addressing the adverse consequences of offshoring should be concerned with the potential growth of offshoring rather than the existing level of offshoring-related displacement. Annual rates of offshoring-related displacement in the survey were on the order of 10% of aggregate IT-worker turnover. While unclear at what level offshoring shifts from a trend affecting mostly individual workers to a concern for all workers in an occupation, the trends should be measured and monitored.

Proposed policy interventions attempting to reduce the adverse effects of worker displacement (such as worker retraining and government compensation to offset wage losses associated with moving to new industries) could focus on specific occupations. Furthermore, training programs could focus on the movement of displaced workers toward work that combines existing skills with those that involve elements of personal delivery. Private or public educational institutions can potentially adjust their curricula to address this emerging need. Our findings are consistent with broader calls from education scholars who have advocated (in response to recent waves of technological change) emphasizing “softer” skills (such as complex communication)14 in the U.S. educational system; for example, educators could interweave existing material in the IT curriculum with projects that promote teamwork, negotiation, and presentation skills.

Over 30% of respondents reported offshoring computer programmers and software developers, but only about half of them, or 15.5%, reported offshoring systems analysts.

In the future, this area of research would benefit from improved offshoring data, including more fine-grain measures of the task content of individual jobs. Data at the task level would allow researchers to test more nuanced models of which attributes make a job vulnerable to offshoring (such as those considering the modularity, codifiability, or information intensity of a worker’s task set). These tests would also provide insight into how jobs and educational programs can be designed so U.S.-based workers maximize the value they provide to the global economy. Furthermore, although our survey data was unique because it allowed us to capture fine-grain outcomes, a limitation of the data was its reliance on self-reported responses from employees and hiring managers that might be subject to bias. Evidence from other data sources could therefore be useful in validating these results.

Finally, although our study focused on job displacement, offshoring may also affect workers through reduced wages. A more comprehensive understanding of the full effects of offshoring on IT workers and the demand for particular skills could therefore be provided through analyses of job displacement and wage effects.

Back to Top


We thank Peter Cappelli, Eric Clemons, Lori Rosenkopf, and three anonymous reviewers for their guidance and comments.

Back to Top

Back to Top

Back to Top

Back to Top


UF1 Figure. Percent of firms reporting offshoring by worker type.

Back to Top


T1 Table 1. Percent of surveyed firms by industry reporting offshoring work.

T2 Table 2. Correlations between occupation and reasons for offshoring.

T3 Table 3. Correlations between offshoring destinations and type of work.

T4 Table 4. Worker displacement levels by occupation.

T5 Table 5. Offshoring-related displacement rates for IT and non-IT workers.

T6 Table 6. Probit analysis, employer offshoring.

T7 Table 7. Probit analysis, worker displacement.

T8 Table 8. Probit analysis of outcomes for all displaced workers.

Back to Top

Back to Top

    1. Amiti, M. and Wei, S. Fear of service outsourcing: Is it justified? Economic Policy 20, (Apr. 2005), 308–348.

    2. Apte, U. and Mason, R. Global disaggregation of information-intensive services. Management Science 41, 7 (July 1995), 1250–1262.

    3. Aspray, W., Mayadas, F., and Vardi, M. Globalization and Offshoring of Software: A Report of the ACM Job Migration Task Force. ACM Press, New York, 2006.

    4. Blinder, A. How Many U.S. Jobs Might Be Offshorable? CEPS Working Paper No. 142, (Mar. 2007).

    5. Blinder, A. Offshoring: The next industrial revolution. Foreign Affairs 85, 2 (Mar./Apr. 2006), 113–128.

    6. Carmel, E. and Tjia, P. Offshoring Information Technology: Sourcing and Outsourcing to a Global Workforce. Cambridge University Press, Cambridge, England, 2005.

    7. Ferrat, T., Agarwal, R., Brown, C., and Moore, J.E. IT human resource management configurations and IT turnover: Theoretical synthesis and empirical analysis. Information Systems Research 16, 3 (Sept. 2005), 237–255.

    8. George, T. Report: IT turnover high despite concerns over job stability. InformationWeek (May 6, 2002);

    9. Government Accountability Office. Current Government Data Provide Limited Insight into Offshoring of Services. GAO-04-932. Washington D.C., 2004.

    10. Herz, D. Worker displacement still common in the late 1980s. Monthly Labor Review 114 (May 1991), 3–9.

    11. Jensen, J. and Kletzer, L. Measuring Tradable Services and the Task Content of Offshorable Services Jobs. Working Paper, 2007.

    12. Kletzer, L. Job displacement. Journal of Economic Perspectives 12, 1 (Winter 1998), 115–136.

    13. Lee, D.M.S., Trauth, E., and Farwell, D. Critical skills and knowledge requirements of IS professionals: A joint academic/industry investigation. MIS Quarterly 19, 3 (Sept. 1995), 313–340.

    14. Levy, F. and Murnane, R. Teaching the New Basic Skills: Principles for Educating Children to Thrive in a Changing Economy. Free Press, New York, 1996.

    15. Mann, C. What global sourcing means for U.S. IT workers and for the U.S. economy. Commun. ACM 47, 7 (July 2004), 33–35.

    16. Mark, J. Technological change and employment: Some results from BLS research. Monthly Labor Review 110 (1987), 26–29.

    17. Mithas, S. and Krishnan, M.S. Human capital and institutional effects in the compensation of information technology professionals in the United States. Management Science 54, 3 (Mar. 2008), 415–428.

    18. Mithas, S. and Whitaker, J. Is the world flat or spiky? Information intensity, skills, and global disaggregation. Information Systems Research 18, 3 (Sept. 2007), 237–259.

    19. Panko, R. IT employment prospects: Beyond the dotcom bubble. European Journal of Information Systems 17 (2008), 182–197.

    20. Phillips, D. and Clancy, K. Some effects of 'social desirability' in survey studies. The American Journal of Sociology 77, 5 (Mar. 1972), 921–940.

    21. Ramasubbu, N., Mithas, S., and Krishnan, M.S. High tech, high touch: The effect of employee skills and customer heterogeneity on customer satisfaction with enterprise system support services. Decision Support Systems 44, 2 (Jan. 2008), 509–523.

    22. Slaughter, S. and Ang, S. Employment outsourcing in information systems. Commun, ACM 39, 7 (July 1996), 47–54.

    23. Tambe, P. and Hitt, L. Now I.T.'s 'Personal': Offshoring and the Shifting Skill Composition of the U.S. Information Technology Workforce. Working Paper, 2010.

    24. Zwieg, P., Kaiser, K., Beath, C. et al. The information technology workforce: Trends and implications 2005–2008. MIS Quarterly Executive (2006);

    a. We computed this average estimate by dividing the total displacement rate by the number of years since 2000, because many companies viewed the potential Y2K problem as a trigger. The annual displacement rate is slightly less if firms engaged in substantial offshoring before 2000. Due to data limitations, we can say little about how the actual displacement rate changed from year to year 2000–2007.


Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More