Inter-firm mobility and return migration patterns of skilled guest workers

Abstract

Two concerns central to the debate over skilled guest worker programs in the USA are that (1) guest workers are restricted from inter-firm mobility and are “effectively tied” to their firms, and (2) guest workers provide cheap and immobile labor that crowds out natives, especially during times of heightened unemployment. We address these concerns by using a unique dataset of employee records from six large Indian IT firms operating in the USA. We find that the guest workers in our sample exhibit a significant amount of inter-firm mobility that varies over both the earnings distribution and the business cycle. We also find that these workers exit the USA during periods of heightened unemployment. These findings provide new evidence on the implications of the institutional features and debate surrounding guest worker programs.

Appendices

Appendix A: Institutional details and data construction

The H-1B and L-1 visa programs are intended for highly-skilled guest workers. Individuals who receive H-1B visas are required to possess skills in a “specialty occupation” while holders of L-1 visas are expected to possess “specialized knowledge.” An annual cap of 65,000 was initially placed on the number of H-1B visas available. The American Competitiveness and Workforce Improvement Act of 1998 increased the H-1B visa cap to 115,000 for 1999 and 107,500 for 2000. The American Competitiveness and Worker Investment Act for the 21st Century of 2000 increased the cap to 195,000 through 2003, after which the number of visas reverted to 65,000. Additional changes allowed another 20,000 recipients of post graduate degrees obtained in the USA to receive this visa. Employers of H-1B workers must pay the annual prevailing wage which can be prorated into monthly or more frequent regular salary payments. Hourly workers must meet the prorated annual minimum each pay period (they have to be paid for non-productive time), as described in paragraph (c)(7)(i) of of 20 CFR Section 655.731, which is why firms in this industry typically pay a prorated annual salary. See http://www.law.cornell.edu/cfr/text/20/655.731#c_2. While regulations prohibit workers on an L-1 visa from switching jobs in the same manner as workers on an H-1B, L-1 workers are able to switch jobs if they find a new employer who will sponsor them for an H-1B visa, or if they are able to obtain permanent residency. Legislation enacted in 2004 restricted the use of L-1 visas by “job-shops” that contract L-1 workers out to other firms, and starting in 2009, the Obama administration began a crackdown on L-1 usage, especially with regard to applications from India (National Foundation for American Policy, 2012). While we do not know whether the guest workers in our data are on an L-1 or an H-1B visa, we believe most are on the H-1B and consult several sources to confirm. USCIS provided information on large H-1B and L-1 visa sponsoring firms to U.S. Senator Charles Grassley. A link to this data is available here: https://web.archive.org/web/20070627215621/http://grassley.senate.gov/releases/2007/062620072.pdf. This document gives the number of H-1B visas issued to the top 20 users of H-1B visas in fiscal year 2006, which falls during the middle of our period of study. It also reports the number of L-1 visas issued to these firms. From this list, we exclude seven domestic corporations: Microsoft, Deloitte and Touche, IBM, Oracle, Cisco, Intel, Ernst and Young, and Motorola. The remaining firms are Indian IT firms, and accounted for 23,293 H-1B visas and 11,712 L-1 visas. Thus, this information suggests that there is nearly a 67 to 33 % split between H-1B and L-1 visas in large firms in this industry in fiscal year 2006. As we do not know the identity of the six firms in our data, we must rely upon this industry level measure. One potential concern with this approach is that by focusing on the largest users of H-1B visas, we may be overestimating the importance of the H-1B visa in this industry. To examine this issue, we turn to a separate report which gives the number of L-1 visas issued to the Top 20 L visa-using firms. See https://web.archive.org/web/20070627215549/ http://grassley.senate.gov/releases/2007/06262007.pdf. In this report, we note that 12 firms from the previous document also appear. Of these, non-domestic firms account for a total of only 945 visas. Even if these firms sponsored no H-1B visas, this would lower our estimate of H-1B workers in this industry by only two percentage points. Finally, using data from government reports (available at https://web.archive.org/web/20130215115750/http://www.travel.state.gov/visa/statistics/nivstats/nivstats_4582.htmlas FY1997-2012 NIV Detail Table (Excel spreadsheet)), we see that H-1B visas account for 67.8 % of combined H-1B and L-1 visas issued to Indian nationals between fiscal years 2003 and 2011. We believe that guest workers who quit in the USA separate to other employment, and that other separations are returns, for the following reasons. Firms are required to repatriate workers according to Section 8 C.F.R. 214.2(h)(4)(iii)(E). We note that guest workers who quit without an approved change of employer are heavily penalized if they move to informal employment; the law requires an employee who quits or is dismissed to find a new job. Failing this, they will immediately become “out of status,” and thus become an “illegal alien.” If a worker left their job but did not obtain new employment, the worker would be “out of status” and in violation of immigration law. Furthermore, time limits for workers are not reset when a H-1B or L-1 worker switches employment. A summary is available at http://www.klaskolaw.com/articles.php?action=view&id=50 See 8 C.F.R. 214.2(h)(13)(iii) as well as http://www.uscis.gov/USCIS/Resources/C2en.pdf for further details. We acknowledge that there is an Indian born population of undocumented and unauthorized workers in the USA, and that H-1B workers moving between jobs without proper authorization may be among these workers. Estimates from the Pew Hispanic Center (http://www.pewhispanic.org/2014/12/11/unauthorized-trends/) suggest that, from 2005 through 2011, there were an average of 360,000 unauthorized Indian Born individuals in the USA, compared to a average total unauthorized population of around 11.5 million. Combined with population counts from the 2005–2011 American Community Survey, this suggests that around 17.5 % of Indian migrants and 28.5 % of all migrants were unauthorized. ACS and Pew data (http://www.pewhispanic.org/2009/04/14/iv-social-and-economic-characteristics/) reveal that only 15 % of unauthorized immigrants have a Bachelor’s degree, compared to 30 % of all immigrants. Assuming that, among Indian Immigrants, the unauthorized are also half as likely to have a Bachelor’s degree, we can apply Bayes’ law to find that the probability that a college educated Indian born worker is undocumented. This calculation suggests the rate is a relatively low 8.75 %. It is unlikely that the workers in our dataset face mobility frictions due to a pending application for permanent residence. Hira (2010a, b) constructs an immigrant yield measure that compares the number of immigrants sponsored by firms for temporary worker visas to the number that the same firms sponsor for permanent residence. The largest ten offshore outsourcing firms, including the Indian IT firms, sponsored only 6 % as many workers for permanent residence as for H-1B visas in 2008, compared to 64 % yields for traditional technology firms such as Microsoft, Cisco, Oracle, Qualcomm, Google, and Intel. Hira discusses three reasons why Indian IT firms do not hire US workers: to facilitate knowledge transfer to India, to have an inexpensive labor source in the USA, and to train workers who will return to India and continue to support operations remotely. When studying mobility between firms using CPS data, we employ the CPS-March Supplement variable “numemps,” which reports the number of employers that a survey respondent had in the previous year. Using data from 2003 to 2011, we compute the percentage of IT workers who had more than one employer in a given year. We limit our sample to IT workers who are currently employed and working at least 40 h per week on the main job in order to avoid counting workers who are working more than one job at a time, rather than switching jobs. Our sample is of two occupations: CPS “occ1990” codes 229 and 64, which represent “Computer Programmers” and “Computer Systems Analysts and Scientists,” respectively. Likely, Indian guest workers are, as previously defined, non-citizens born in India who have been in the USA for less than 8 years. We would expect guest workers to have a Bachelor’s degree as is required by the H-1B and L-1 programs. In fact, 99.7 % of the likely guest workers in the CPS data do. Since no such degree requirement exists for native born workers in this industry, only 60.9 % of workers have a degree. To make the samples more comparable, we drop workers without a Bachelors in the analysis. All calculations are done using survey weights, a mobility gap of less than 1 % exists when not using survey weights, and in both cases, a Chi-square test of significance fails to reject the null hypothesis of equivalent quit rates. When studying previous quit elasticity estimates, we use 25 estimated quit elasticities reported by Manning (2011) in Tables 6 and 7 of his book chapter. When estimates for multiple groups were reported, we took the raw average of the reported estimates. When ranges were given, we took the midpoint. For a paper reporting one-sided bounds, we used the bound itself as the estimate. All reported elasticities in these tables were obtained by estimating the relationship between earnings and separations. Rather than report the implied supply elasticities, as Manning and the original authors did, we instead report minus one half of his numbers, i.e., the raw separation elasticity results that were used to generate the implied supply elasticity numbers, and are thus most comparable to the numbers presented in our study.

Appendix B: Tables and figures

Table 6 Comparison of LCA and firm data

Table 7 Comparison of trimmed firm and ACS data

Table 8 Quit to another firm

Table 9 Return to India

Table 10 Elasticity estimates: income trim check

Table 11 Robustness check: quit to another Firm

Table 12 Robustness check: return to india

Table 13 Unemployment rate descriptive statistics

Table 14 Unemployment rate correlations

Table 15 Quit elasticity estimates: assumptions about share of H-1B holders from the top two-thirds (bottom two-thirds) of earnings distribution

Fig. 6

Distribution of tenure

Fig. 7

Heterogeneity in quit elasticities over the business cycle

Fig. 8

Heterogeneity in return elasticities over the business cycle

Fig. 9

Unemployment rates faced

Appendix: References

Ron Hira. Bridge to immigration or cheap temporary labor? The H-1B & L-1 visa programs are a source of both. Briefing Paper 257, Economic Policy Institute, Washington, DC., 2010. http://www.epi.org/publication/bp257/.

National Foundation for American Policy. Analysis: Data reveal high denial rates for L-1 and H-1B petitions at U.S. citizenship and immigration services. NFAP Policy Brief, February 2012.