The Department of Homeland Security's proposed new rules on the H-1B program are deeply disappointing as they:
- Do nothing for U.S. workers displaced by this needlessly large program;
- Lavish millions of dollars in savings on the employers of H-1Bs;
- Seem to invite huge additional numbers of applications, which will just give industry new arguments for higher ceilings;
- Create a new way of doing business that seems to be an invitation to gaming the system; but
- Do add a long overdue and deft adjustment to the current system, a good, medium-sized move that is initially hard to understand and is explained badly by DHS.
The announcement of the proposed rulemaking came Friday afternoon, November 30, and was accompanied by a 41-page document in the December 3 Federal Register.
The announcement proclaimed it was: "introducing a more meritorious selection of beneficiaries."
This turns out to be a puzzling process that, by definition, will do no good to any American worker, and the "meritorious selection" process (which only relates to the selection among aliens) is, in effect, a slightly different and better way of loading the dice used in two government-run lotteries, a subject to which we will return.
Unless I missed it completely — and this is a long document, one version runs 148 pages — there is nothing in the proposed regulations that can be construed as helping U.S. residents recapture any of the nearly one million jobs currently held by H-1Bs. (I saw nothing in the document as useful as an estimate of the reach of the program; the nearly one million estimate is my own.)
On the other hand, there are numerous pages that deal lovingly with how DHS proposes to save H-1B employers millions in administrative costs. One of these tabulations (on p. 62420 of the Federal Register) seems to indicate that for employers that secure one or more H-1B lottery tickets the new system will cost them all an extra $6.2 to $10.3 million, but for those that do not get selected it will save them $47.3 to $75.5 million, for a net savings of $41.1 million to $65.2 million. Since all but the smallest employers will both lose some of their applications and win others, the net range is the important one.
Currently, a would-be employer of one, or thousands of, H-1B workers(s) must file a complete Labor Condition Application (LCA, form ETA-9035) for each H-1B job with the U.S. Department of Labor, and a nonimmigrant worker petition (I-129) with the Department of Homeland Security. There are usually more than twice as many such application packages filed as slots available (65,000 for those with at least a bachelor's degree, and 20,000 for those with at least a master's degree). DHS first holds a lottery among the master's degree applications; those applications that are not accepted then go into a second lottery along with the bachelor's degree applicants.
The proposed new process would cut down on the paperwork by 1) setting up an electronic registration system, and 2) by running the lotteries first, and then asking those employers who won in the lotteries to file applications for the jobs that were selected. Thus employers would need only to submit the application packages in cases where they had already won the lottery, thereby saving themselves millions in administrative costs as a result. The non-handling of the losing applications would also save the government some money, and that is useful.
The proposed procedure would mean, in effect, that only employers that had won the lottery would have to buy tickets, and losers would never have to pay for them, an Alice in Wonderland situation.
This brings up yet another savings for the employers, which I did not see spelled out in the proposal, though it may be hiding in the text.
There are a range of fees laid on each of the applications to DHS (but none on the ones that go to DoL.) These run from $1,710 to $7,210 with the H-1B-dependent firms paying $4,000 a pop more than other firms. H-1B-dependent firms (a major factor in this program) have:
- 25 or fewer full-time-equivalent employees and at least eight H-1B workers; or
- 26-50 full-time-equivalent workers and 13 or more H-1Bs; or
- 51 or more full-time-equivalent employees of whom 15 percent or more are H-1Bs.
Currently the fees must be paid when the applications are filed, and are returned later, maybe months later, for applications not selected. Under the new system, the fees would only be applied to already accepted applications.
Under the old system, the employers who did not win the lottery were making a substantial, no-interest loan to Uncle Sam. This will no longer be the case. There is to be no fee for the initial registration of an application; the fees will be paid only by those winning the lottery.
While the "loans" to the government were small change for the big H-1B employers, to many of the smaller scale ones they must have really complicated their finances.
One of the unwelcome oddities of the new regulations is its obsession with what I call "not letting any petitions go to waste". For example, we find this on p. 62415 of the December 3 Federal Register: "If the number of registrations received during the initial registration period is fewer than the number of petitions USCIS projects are needed to meet the regular cap ... ." (Emphasis in original.)
Needed? Doesn't USCIS know the difference between an upper limit, the ceiling as set by law, and an allocation? Congress, or at least previous congresses, did not want too many H-1Bs, but that concept has apparently been lost.
Similarly, on p. 62414 that worry is expressed in a different context:
DHS is particularly interested in preventing circumstances where petitioners submit large numbers of registrations but never follow up with the filing of H-1B petitions for the selected beneficiaries.
Under these circumstances there would be fewer H-1Bs taking jobs that U.S. residents could take, and that, in USCIS eyes, seems to be a tragedy.
Probable Consequences
More Applications. My sense is that the combination of completing all the forms and lending the feds thousands of dollars per application serves to reduce the number of applications filed under the current system. With neither of these brake-like functions in place, I suspect that the number of registrations under the new system will easily outnumber the applications filed under the old system.
Why not register for lots of jobs, when you know that more registrations will equal more chances for H-1B workers? On the other hand, if you only register for what you think you want, you will only get a fraction of the slots that you desire.
To some extent, the expansive results of the no-application, no-fee policy, unless you have won a slot or a thousand of them, are balanced by a good new provision, which prevents individual employers from entering an individual worker's name more than once in the registration process.
More Gaming of the System. With minimal paperwork, and no fee-paying, H-1B employers will be able to game the system at will. Let's take five relatively small H-1B outsourcing operations, each known to the other four; each would like 20 new H-1Bs from the next lottery. If all five file for 20 each they are likely to get only four of the workers that they want (assuming a 5:1 overage of applications over winners).
On the other hand, the five might get together before the registration period, share their lists of potential H-1B workers, pooling their some 100 candidates, and then each would file for all 100 of them. If the lottery produced 20 winners and 80 losers for each of the conspirators, on average, then the five could do some trading among themselves, with each getting approximately the people that they had wanted all along. If there was an excessive number of winners, then those workers could be rented out, at a profit, to H-1B users who had not oversubscribed their wants.
The new regulations seem to say that an individual employer cannot file more than one piece of paper per potential worker, but there does not seem to be any bar to a potential worker being on more than one employer's want list.
The One Deft Change
One of the changes in the new rules that makes good sense — though it does not help U.S. workers — is altering the sequence of the lotteries. Currently, the first lottery is for the 20,000 slots set aside for aliens with U.S. master's degrees or better (usually in computer sciences) and then there is a second lottery; all not yet selected are included in the second lottery for 65,000 workers with bachelor's degrees or better.
The proposal would reverse the sequence of the two lotteries. This did not seem to make much sense to this non-mathematician. I then reread (on p. 62417) the government's not-very- helpful explanation of its rationale:
As described, the current lottery system does not provide an optimal mechanism for achieving that aim [a higher percentage of H-1Bs with U.S. master's degrees] because it dilutes the candidate pool in a manner that greatly diminishes the possibility of adding 20,000 such H-1B nonimmigrants beyond those that would be admitted without the advanced degree exemption allocation.
Still puzzled, I reached out to Professor Ron Hira of Howard University, the academic with the best grasp of H-1B in the nation. That turned out to be a good move on my part. Hira, unknown to me, testified before a U.S. Senate committee back in 2013, calling for such a shift in the lottery sequence. And then he presented an illustration to me of how it would have worked with the current (spring 2018) set of applicants, using the term "master's +" for those candidates with either a master's degree or a doctorate:
[There] are 190,098 total applicants, 95,885 of which are master's +. The current USCIS lottery sequence first gives 20K to master's applicants from the master's cap. So the remaining pool would be 170k, with 76k being master's+. That means that the remaining master's+ would have a 45 percent chance of winning the 65k cap; 28,998 master's+ would win the 65k lottery. The net result of the current system is that 48,998 master's+ are selected.
In the system I'm proposing, the random lottery would be re-sequenced. Run the lottery on the 65k base first. In this scenario, master's+ would have a 50.4 percent chance of winning the base cap lottery. That would translate into 32,786 master's+ winning. Then allocate the remaining 20k to the master's cap. The net result of my proposed system is that 52,786 master's+ are selected. This is a modest but material difference favoring U.S.-earned master's + applicants.
The assumption is that an H-1B worker with an American master's degree or a doctorate would be better trained than someone with either just a bachelor's degree or someone with an overseas master's or doctorate.
So in that instance, there would be 3,788 more H-1B workers admitted for that year who had a superior education to the educational mix than we would get under the current system. That would be useful, and it would be even more useful if DHS had spelled it out as Hira has done.
The government should have taken that step years ago.
The truth about the H-1B program is that it allows U.S. employers (and many big ones operating here, but based in India) to secure alien talent at a major bargain. That shuts hundreds of thousands of U.S. resident workers out of these jobs. What the government could have done, and should have done, was to make the program more expensive, thus pushing at least some of the employers into working with the American labor market.
That central truth, of course, is denied by industry routinely and, in this case, by DHS. That agency charged off in the opposite direction, making the program less expensive than before, and thus more damaging to U.S. workers. The concept of justice for U.S. workers was junked in favor of higher profits for the bosses (and stockholders).
One totally appropriate way of increasing the costs, and bringing better salaries to the foreign workers, would have been to adopt an auction, rather than a lottery. Or rather two auctions, one for the 20,000 ceiling and one for the 65,000 ceiling. The H-1B slots would go to employers who offered the highest salaries.
Another way would be to levy a non-returnable fee for, say, $250, for the privilege of joining the lottery. If there were 200,000 applications that would bring $50 million to DHS, roughly balancing the savings that would be created by the proposed new rules (remember the $41.1 million to $65.2 million range mentioned earlier). It would have been even better to set that fee at $500, making the new program more expensive than the old, despite the savings in administrative costs.
But no, DHS decided what needed to be done was to make the program (so detrimental to U.S. workers) even more attractive to employers.
Because of the need both to cope with comments on it, and to create the necessary software, the proposed new rule will not go into effect until after the next filing season, in the spring of 2019.
Readers are encouraged to react, in writing, to the H-1B proposal. The deadline is January 2, 2019, and details on how to communicate with DHS can be found here.