Date: Wed, 8 Jul 2009 22:19:50 -0700 From: Norm Matloff To: Norm Matloff Subject: the Kerr/Lincoln patenting study To: H-1B/L-1/offshoring e-newsletter Back in February, I reviewed a working paper by Prof. Bill Kerr of HBS and Prof. Wm. Lincoln of the University of Michigan, a review which you can read at http://heather.cs.ucdavis.edu/Archive/KerrLincoln.txt Bill (Kerr) sent me a detailed response to my analysis, which I am sharing with his permission, at http://heather.cs.ucdavis.edu/Archive/KerrResponse.pdf He and I also met when I was in Cambridge in April, and had a nice exchange (though only partly on his paper). I promised readers of this e-newsletter that I would reply here to Bill's message when I got a block of time large enough to go into detail. Unfortunately, due to family and other issues, I've not had a chance to do so, especially recently, and indeed have posted material to my e-newsletter only sporadically in the last few months. However, I am now able to resume, and finally here is my commentary on Bill's response. By the way, I will also be reviewing the recent Hunt paper in the near future as well, as well as some other items. The best way to begin here is to quote the abstract of the Kerr/Lincoln paper: This study evaluates the impact of high-skilled immigrants on US technology formation. Specically, we use reduced-form specifications that exploit large changes in the H-1B visa program. Fluctuations in H-1B admissions levels signicantly influence the rate of Indian and Chinese patenting in cities and firms dependent upon the program relative to their peers. Most specifications find weak crowding-in effects or no effect at all for native patenting. Total invention increases with higher admission levels primarily through the direct contributions of ethnic inventors. (In my original posting on the paper, I objected to the paper's characterization of immigrants as "ethnic" and its referring to natives as "English." But to avoid insertion of numerous disclaimers below, I will simply use the authors' terminology.) Here is a summary of the first main conclusion, in plain English: Increased admissions of H-1B tech workers, who are mostly Indian and Chinese, are associated with increases in patents by Indians and Chinese. That of course is completely unsurprising, and it should have no implications for the H-1B debate. Yet the authors claim such implications in the Introduction section of their paper, in which they repeatedly assert that the H-1B program has been good for U.S. innovation, and moreover that the paper's findings should be useful in guiding the H-1B debate: The H-1B visa program governs most admissions of temporary immigrants into the US for employment in patenting-related fields. This program has become a point of significant controversy in the public debate over immigration, with proponents and detractors at odds over how important H-1B admission levels are for US technology advancement and whether native US workers are being displaced by immigrants. In this study we quantify the impact of changes in H-1B admission levels on the pace and character of US invention over the 1995-2006 period. We hope that this assessment aids policy makers and business leaders by informing their current discussions about appropriate admission rates in the future. Of more general interest, the variations induced by changes in H-1B admissions are an attractive laboratory for studying whether immigrant scientists and engineers (SEs) crowd-in or crowd-out native SE workers. Identifying this native response is important for assessing the extent to which aggregate US invention is promoted by more flexible high-skilled immigration policies. The link between immigration policy and new innovation may appear tenuous at first, but immigrant SEs are central to US technology formation and commercialization. In terms of levels, immigrants represented 24% and 47% of the US SE workforce with bachelors and doctorate educations in the 2000 Census, respectively. This contribution was significantly higher than the 12% share of immigrants in the US working population... So, the authors make it clear--they regard H-1B as crucial to U.S. tech innovation. Yet, as I said above, their finding regarding ethnicity-specific patenting should have no bearing on the H-1B debate. Where, then, is the resolution? How can the authors reach such a strong conclusion about the value of the H-1B program from their finding regarding only ethnicity-specific patenting? The answer to these questions is in the next-to-last sentence of the abstract: Most specifications find weak crowding-in effects or no effect at all for native patenting. In other words, the authors are saying that increases in H-1B ethnic patenting are not associated with decreases in native patenting. Therefore--they assert--we have more patents than we would if there had been no immigration, a supposed net overall gain. It is that conclusion that is the basis for my main objection to the paper. When Bill sent me the article back in December, I replied to him as follows: I'll take a look, thanks. But from just reading the abstract it's clear that you haven't accounted for displacement of natives and non-H-1B immigrants. In Bill's reply, he chides me for coming to such a quick conclusion based on just the abstract: I was certain you would dislike the paper when I sent it to you.   Your negative reply within five minutes of my hitting send, based  upon reading an abstract that by custom is limited to 100 words,  confirmed that. I probably deserved his chiding. However, after reading through his entire paper several times, meeting with him in person, reading his written response and so on, that "hasty" initial response of mine above remains my core criticism of the paper. In short, I don't agree that his and Prof. Lincoln's analysis implies that immigration has given the nation a net increase in patent activity, as they don't account for the extensive displacement of Americans from the patent-producing jobs. A key point in that regard is that one must distinguish between short-term and long-term trends. The Kerr/Lincoln study looks at short-term effects, but the displacement of Americans is long-term, thus not picked up in the study. Here are the issues. As I pointed out in my original posting on Bill's paper, the H-1Bs who file patent applications typically have graduate degrees, and typically have earned those degrees at U.S. universities. The industry lobbyists use this fact to portray these as the "good" H-1Bs, meaning the ones who don't displace U.S. citizens and permanent residents. The lobbyists point out that Americans don't pursue graduate degrees in large numbers (fewer than 50% U.S. doctoral degrees granted in computer science are earned by American students), and use this fact to asser that there is a genuine shortage of workers with this education, hence no displacement. But the reality is that such displacement has in fact occurred quite extensively over the years. As I've mentioned often, this displacement was actually planned for by the National Science Foundation, the federal government's central agency for science. In my original posting, I wrote: Also on the crowding-in/out issue, the authors correctly recognized that patents typically arise from the work of people with graduate degrees, and thus raised the question as to whether the presence of the foreign students has displaced Americans from university graduate programs. The authors cite the work of fellow Harvard professor George Borjas that "natives are crowded-out from graduate school enrollments by foreign students, especially in the most elite institutions, and suffer lower wages after graduation due to the increased labor supply." The authors say that Borjas' findings don't jibe with other work on this topic, but in fact the papers they cite don't address the Borjas issue at all... Moreover, the National Science Foundation (one of the funders of the Kerr/Lincoln study) itself stated at the time Congress was considering instituting the H-1B program that the program would indeed crowd OUT the Americans. In the paper I've quoted often here, the NSF said that "A growing influx of foreign PhDs into U.S. labor markets will hold down the level of PhD salaries...[If] doctoral studies are failing to appeal to a large (or growing) percentage of the best citizen baccalaureates, then a key issue is pay...A number of [the Americans] will select alternative career paths." That of course is exactly what happened in the subsequent years; enrollment by Americans in tech grad programs has gone way down, and tech PhD salaries have not kept pace with those of comparable professions, exactly as Borjas found. So this is, in my view, the major flaw in the Kerr/Lincoln paper. When Profs. Kerr and Lincoln say they find that H-1B doesn't in the short term crowd-out American workers from patenting, they are missing the fact that the H-1B program has, over the long term, pushed Americans out of the patent-generating jobs. Thus, instead of absolving the H-1B program of guilt as the authors do, a more thorough analysis would have investigated H-1B as the culprit. Bill does mentions the short-term/long-term distinction in his letter to me: In paragraphs ten through twelve, you express concern with how I  present George Borjas’ work and technique.  I have loads of  respect for Borjas and have learned a lot from his work.  For the  purposes of this dialogue, I again hope that the paper speaks for  itself.  I am very up‐front that this is a short‐run analysis of  changes in the H‐1B program.  I speak at several points about how  long‐run effects can be different due to the issues you highlight. I cite the existing work on  these long‐run effects with equal  weight, and I hope that readers draw their own conclusions.  Though the Kerr/Lincoln paper does mention the distinction between short-run and long-run effects, e.g. on pages 4 and 30, it is not in the context I cited. Instead, the authors discuss the long-term issue of an economy adjusting to changing needs. On p.4, for instance, they say, However, there are often significant adjustment costs when workers move across occupations and sectors, particularly in moving into research-oriented occupations. Ryoo and Rosen (2004) and others have documented the significant costs and "time to build" inherent in increasing the supply of ideas workers, particularly in SE. Shortages of sufficiently trained labor may exist even if forward-looking workers are rationally responding to incentives. These costs and slower adjustments open up the possibility for supply shocks to US innovation through shifts in immigration policy. Indeed, public debates about the H-1B visa often turn on whether a shortage of high-skilled workers exists or not. This is very, very different from the point that I'm making here and that I made to Bill in my e-mail message to him last December. We're not talking about the economy's evolving toward unanticipated labor needs. Instead, the NSF advocated a deliberate policy that it said would push Americans out of research jobs. There are other factors to worry about regarding displacement. I've noticed that, due to their superior English and knowledge of U.S. culture relative to the foreign students, American PhDs with good research abilities are often hired into non-research positions, involving work in areas such as strategic overalll planning of the firm's research activities, interfacing with national standards organizations, and so on. Some quick comments on the bulleted remarks in Bill's letter (I more or less agree with the remarks not covered below): bullet 2: My posting's phrasing regarding the term "shock" was too brief and was possibly misunderstood by some readers. I should have explained that it's a standard economic term. bullet 3: Bill says, The results find that Indian and Chinese invention rose as a share of US patenting with higher H-1B populations. Given that the first-order effect [of] the program is for letting more of these immigrants into the US for work, we should expected to see relatively higher responses from inventors with these ethnic names. And that is what the paper finds. This of course is my point at the beginning of my remarks above. H-1B brings in more Chinese and Indian techies, so we should expect that Chinese and Indian patenting rises, and sure enough, it did. No surprise, and as I said above, no implication for the H-1B debate. He goes on to say that he does not claim that the Chinese and Indian immigrants are more innovative. This is a key point, when coupled with the displacement issue. If they had been more innovative, then displacement of the Americans by the immigrants would produce a net increase in innovation. But again, Bill is not making that claim, and I would add that the common perception is that the Chinese education system inhibits innovation. (The governments of China and Taiwan have engaged in efforts to remedy that problem, as have Japan and S. Korea.) He cites the Hunt/Gauthier-Loiselle paper. I'll be commenting on that paper and a subsequent one by Hunt in a few days, lots of interesting material to talk about. (Added later: I subsequently reviewed the more extensive Hunt paper; see http://heather.cs.ucdavis.edu/Archive/JenniferHunt.txt) bullet 7: Bill makes lots of disclaimers here, but I must disagree, because most people will read only the Introduction section, or (oh, the horror!) only that 100-word abstract. bullet 10: Bill says, ...you ask about the counterfactual of giving H-1Bs to native workers [sic]...I believe it is the wrong question...the US government does not direct people to work in one capacity or another. The policy lever is the H-1B worker admissions... Bill's point is that we don't have a managed economy in which the government can push Americans into jobs the government thinks are worthy, and thus H-1B is available as a remedy if too few Americans go into those jobs. That indeed is the ostensible reason for having the H-1B program. BUT THE REALITY is that government acted quite opposite to this. The US government used that H-1B lever knowing (as it explicitly stated) that it would DISPLACE Americans from the patent-generating jobs. Bill should be outraged at this. bullet 11: Bill did say in our meeting that he regrets the terms "ethnic" and "English," and he agrees in his letter that there are serious methodological problems in the fact that his analysis is based on surnames. During I meeting, I suggested some analyses that might show that the effects of such errors may not be so large after all. Norm