Why is H-1B Use Skyrocketing among Bay Area Tech Giants?
Even as the White House began cracking down on U.S. work visas, major Silicon Valley technology firms last year dramatically ramped up hiring of workers under the controversial H-1B visa program, according to newly released data. The data show the importance of H-1B workers to the tech industry, which has long lobbied to increase the number of highly skilled foreign workers. Menlo Park-based Facebook in 2017 received 720 H-1B approvals, a 53 percent increase over 2016, according to the National Foundation for American Policy, which obtained federal government data. Mountain View's Google received 1,213 H-1B approvals, a 31 percent increase. The number of H-1B approvals at Intel in Santa Clara rose 19 percent and Cupertino-based Apple received 673, a 7 percent increase.
The NFAP (National Foundation of American Policy) has posted a study which contains this data at this link. As it happens, the USCIS (U.S. Citizenship and Immigration Services) has posted similar data at this link. The following table contains data from both of these sources:
NFAP: 2015-17 % Change USCIS: 2015-17 % Change
----------------- ----------- ----------------- -----------
Company 2015 2016 2017 16-17 15-17 2015 2016 2017 16-17 15-17
--------------- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
Amazon .... 1416 2515 78% ... 1858 2739 4767 74% 157%
Microsoft .... 1145 1479 29% ... 2523 3556 4069 14% 61%
Intel .... 1030 1230 19% ... 1358 1873 2625 40% 93%
Google .... 924 1213 31% ... 2023 2517 2758 10% 36%
Facebook .... 472 720 53% ... 894 1107 1566 41% 75%
Apple .... 631 673 7% ... 1586 1992 2055 3% 30%
As can be seen, the USCIS data shows much higher approval numbers than the NFAP study. The reason for this difference may be due to the following reason given on page 10 of the NFAP study:
The USCIS data on initial employment for FY 2017 show that it can be misleading to rely on alternative H-1B statistics absent the proper context. For example, as noted, Cognizant had 3,194 new H-1B petitions for initial employment approved in FY 2017. However, because of extensions for existing H-1B employees and, in particular, a USCIS policy change in 2015 (Matter of Simeio Solutions) that requires companies to file amended petitions when work will take place in a new metropolitan statistical area (MSA), a recent USCIS report lists Cognizant with 28,908 approved petitions in FY 2017. Those 28,908 approved petitions - 9 times the number of new H-1B professionals approved for the company in FY 2017 - do not represent 28,908 new workers but reflect compliance with government rules when existing workers are moved to new locations. (13)
Following is footnote 13 which the prior excerpt references:
13. The 29,908 figure is contained in a USCIS report on its website on Approved H-1B Petitions by Employer, FY 2017. Unlike the data contained in this NFAP report that focus on initial petitions, the USCIS report contains a combination of initial petitions, amended petitions and extensions.
Table 6 on page 9 of the NFAP study gives similar data for Indian-based companies. Below is the data from that table, along with the same data from USCIS. The section in which this table occurs is titled "Indian-Based Company H-1B Numbers Continue to Fall" and the data shows the numbers falling for all but two companies from 2016 to 2017 and all but one company from 2015 to 2017.
NFAP: 2015-17 % Change USCIS: 2015-17 % Change
----------------- ----------- ----------------- -----------
Company 2015 2016 2017 16-17 15-17 2015 2016 2017 16-17 15-17
--------------- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
TCS 4674 2040 2312 +13% -51% 7936 11295 14697 30% 85%
Infosys 2830 2376 1218 -49% -57% 7989 12780 13408 5% 68%
Wipro 3079 1474 1210 -18% -61% 5968 6819 6529 -4% 9%
HCL America 1339 1041 866 -17% -35% 2548 3492 4392 26% 72%
Larsen & Toubro 830 870 479 -45% -42% 1277 1558 1864 20% 46%
Tech Mahindra 1576 1228 2233 +82% +42% 2553 3344 4931 47% 93%
Mindtree 464 327 150 -54% -68% 751 1103 967 -12% 29%
As can be seen, there is again a large difference in the numbers between NFAP and USCIS. As before, this could be due to the inclusion of petition extensions and amendments as stated in the NFAP study. Still, it seems very strange that all of the USCIS numbers show an increase from 2015 to 2017 and all but 2 show double-digit increases from 2016 to 2017 rather than the double digit declines seen in all but 2 in the NFAP data. The USCIS policy change in 2015 could be a factor for the 2015 to 2017 change but it would not seem likely to have so large an effect from 2016 to 2017.
In fact, this would seem to show a problem with many studies that use data that is obtained via a FOIA (Freedom of Information Act) request. It seems that such data is often provided only to the requester and is not made public. The data does not appear to be available on the NFAP website though I plan to check with them to verify that the data was obtained via an FOIA request and see if they will release it. In any event, the non-public status of much FOIA data means that studies based on such data cannot be verified and arguably should not be used to set any public policy. Unless it is publicly released, it would seem that the analysis of such data can only be used by the requester.
One other point that this brings up is that the approval numbers in the USCIS data that has been made public at this link in the section titled "H-1B Datasets" is of little use. If approvals include every extension and amendment, it would seem that the number of approvals is not of much use except perhaps to estimate the workload required to process them.
Still, these USCIS files do give information about salary and the degrees of the H-1B workers in the petitions. If you go to the R/Shiny app at this link and set "Minimum Approved" to 300, you'll get the 89 employers for whom 300 or more H-1B petitions were approved in 2017. If you then change "Sort by" from "Approved" to "Advanced", it will sort them by the percentage of the petitions that are for advanced degrees (Master's or higher). Following are the lines for the 13 companies listed above:
N X TaxID Employer Approved Salary Advanced -- --- ----- ------------------------------ -------- ------ --------- 2 234 8068 AMAZON WEB SERVICES INC 402 112338 100.00000 12 75 2743 INTEL CORPORATION 2625 104691 91.83206 21 70 3581 GOOGLE INC 2758 134419 73.32849 24 41 6545 AMAZON CORPORATE LLC 4767 118637 71.60546 27 95 5019 FACEBOOK INC 1566 144812 70.24266 31 83 4110 APPLE INC 2055 142974 66.92721 44 51 4442 MICROSOFT CORPORATION 4069 130259 58.57319 56 19 4401 WIPRO LIMITED 6529 75082 39.08381 57 46 5035 HCL AMERICA INC 4392 87978 38.02143 62 33 2696 TECH MAHINDRA AMERICAS INC 4931 78443 35.54156 70 129 5091 MINDTREE LIMITED 967 75677 28.74871 74 13 235 INFOSYS LTD 13408 85717 24.33400 78 5 9806 TATA CONSULTANCY SVCS LTD 14697 73505 19.89924 79 89 4303 LARSEN & TOUBRO INFOTECH LIMIT 1864 78737 18.24034As can be seen, the 6 high tech companies (there are two entries for Amazon) had petitions where 58 percent or more were for advanced degrees. The 7 Indian-based companies, on the other hand, all had petitions for which 39 percent or less were for advanced degrees. That would mean that 61 percent or more are for workers with bachelor's degrees or less. It would seem difficult to view these workers as having special skills that cannot be filled by U.S. workers.
Also visible is the fact that the 6 high tech companies all have petitions with average salaries above $104,000 whereas the Indian-base companies all have petitions for workers with average salaries below $88,000. Hence, unless there is a demonstrable shortage of workers with bachelor's degrees, it would seem many of many of those workers may be getting requested because of their lower pay.
Comments
Post a Comment