Date: Mon, 31 May 2010 23:29:56 -0700 From: Norm Matloff To: Norm Matloff Subject: "the value of a green card" To: H-1B/L-1/offshoring e-newsletter I'll be reporting here on an academic paper by Amy Kandilov, a Vivek Wadhwa blog posting, and relating both to the recent Mithas and Lucas study. I've been stating for years that one doesn't need any data (though plenty exists) to see that H-1B workers make less money than comparable Americans (with "comparable" being a key word): Since it's difficult for H-1Bs to switch jobs, and basically out of the question if they are being sponsored for a green card, they have a narrower range of opportunities than do those who are U.S. citizens or permanent residents. Thus, on average H-1Bs will be underpaid, relative to comparable Americans (the latter a shorthand, though technically inaccurate, term here for U.S. citizens and permanent residents). When I explain the mathematical issue to students, I call it the Big Bag, Little Bag Principle. Say you have two bags, filled with bills of various denominations. The big bag has every bill the little bag has, AND MORE. You are allowed to choose a bag and then choose a bill from that bag, with a full view of everything. Which bag should you choose? Of course, the big bag! I also tell my students not to make decisions on the basis of a single study. Look at multiple studies, and most importantly, see if they jibe with what you know qualitatively about the subject. So, in addition to the many studies showing that H-1Bs are underpaid relative to comparable Americans (including two congressionally-commissioned employer surveys in which the question was asked directly), one also has the qualitative backing--the Big Bag, Little Bag Principle, applied to the lack of mobility of the H-1Bs. In my recent reviews here of a new paper by Mithas and Lucas (see http://heather.cs.ucdavis.edu/Archive/MithasLucas4.txt), I emphasized that although I consider their data set to be invalid for the topic at hand (H-1Bs in the computer industry) and their general conclusions are quite different from mine, Mithas and Lucas do agree on my point about the Big Bag, Little Bag Principle. They cite this as the reason why they found that green card holders got higher salaries than H-1B workers. Vivek Wadhwa, a proponent of immigration programs for tech workers, recently cited the Mithas/Lucas paper as supporting his positive views on such programs, but also states that he too agrees with me on the Big Bag, Little Bag Principle. He says; The study also highlighted an issue that I have written about: these workers are earning less than they should. Because of a flawed immigration system, the U.S. has admitted large numbers of workers on temporary visas, but has never increased the number of permanent resident visas (also called green cards) to enable these workers to make America their home. Waiting times for green cards can be longer than a decade, and while workers wait in queue, they can’t easily change jobs. Employers often take advantage of the situation. Indeed, Lucas and Mithas found that H-1B visa holders earned 6.1% less than equivalent green-card holders. I'll have much more to say about Vivek's blog later in this posting, but for now let's stick to Big Bag/Little Bag point. Vivek points out that the M/L data would indicate a 6.1% premium accruing to H-1Bs when they get their green cards. But 6.1% is not very much. Again, I don't think the M/L data are representative, for the reasons, I've given, but compare this figure to the others that research has found: 15-20% in my studies, 33% in the UCLA study, etc. See my University of Michigan article for details, available online at http://heather.cs.ucdavis.edu/MichJLawReform.pdf Well, the other day, I came across a 2007 study by Amy Kandilov, then a doctoral student (coincidentally at the University of Michigan), titled The Value of a Green Card: Immigrant Wage Increases Following Adjustment to U.S. Permanent Residence. This study is different from those done by the rest of us (M/L, myself, UCLA, etc.), in that the Kandilov study analyzes data for new green card recipients, Princeton's New Immigrant Survey. Moreover, the data state which of the respondents got their green cards through employer sponsorship. This is very valuable data. The NIS data is not perfect either. It is far too coarse in its handling of region, it mixes data sets and so on. Most importantly, it does not have before-and-after salary data. Instead, it has the salary of the new immigrant and the salary he made when first coming to the U.S. For that reason, Kandilov had to match the immigrants with similar Americans (she stuck to natives, not even naturalized citizens), using propensity analysis and the Current Population Survey data. That way she could compare how much the immigrants gained in salary over the period from entry to the U.S. to the time of acquiring a green card, compared to what salary increase they WOULD have had during that period if they had had full labor market mobility like Americans do. Here's what Kandilov found, as she writes in her abstract: The results from the nearest neighbor propensity score matching indicate that becoming a permanent resident is accompanied by at least an 18 percent wage increase for employer-sponsored immigrants. Additionally, I use kernel matching to confirm that immigrants who receive green cards experience a 25 percent gain in wages between their first U.S. job and the job they hold after receiving permanent residence. That 18-25% range is consistent with the previous studies I mentioned. Moreover, it is consistent with the special-skills argument employers make in justifying their green card sponsorship. In the computer industry, for instance (20% of Kandilov's sample), here were typical salary premiums for special skills around 2002: Oracle DBMS tool 24% SAP 24% HP UNIX 20% Visual C++ 20% Java 16% Now, as I've detailed before, the legal prevailing wage is NOT the market wage, due to a number of gaping loopholes. So, ironically, the employer can say "I need to sponsor Mary for a green card, because she knows SAP," but that same employer does not have to factor in Mary's SAP skills when he calculates the legal prevailing wage for her. So he gets a 24% "discount" (more, if Mary has several of these skills) in hiring her as an H-1B, all perfectly legal. At any rate, Kandilov's finding does jibe with these other figures. Now, concerning other aspects Vivek's blog posting, he too comments on the nature of the Mithas/Lucas data set, but with a very different spin: The data Lucas and Mithas analyzed did not include all tech professionals: Information Week readers are typically I.T. workers in large and midsized American corporations. The researchers say that they compared their sample with that of other studies and believe that this is more representative of the tech industry than any other research. And unlike other studies, this compares apples with apples (previous research has looked only at aggregate data, not at individual education and skills). The Lucas/Mithas study clearly did not account for the shoddy body shops that break the law and underpay their workers. But it is likely to be representative not only of the Information Week readership, but also of foreign-born I.T. workers in Silicon Valley. (Silicon Valley tech workers usually have the same advanced educational background and receive the same competitive salaries as their brethren in the I.T. world.) These are the skilled immigrants that American industry needs—to remain globally competitive. As I have shown in detail in recent postings, InformationWeek readers (the source of M/L's data) are not relevant to the issue of H-1Bs in the computer field, because IW's audience is largely semitechnical. I've mentioned, for instance, IW's job ads from a given issue: % D. E. Shaw Research seeking Chief of Staff in New York, NY. % Switch and Data seeking Sr Product Marketing Manager in Tampa, FL. % Kadrmas, Lee and Jackson seeking Network Architect in Bismarck, ND. % Osram Sylvania seeking Benefits Specialist in Danvers, MA. % AccuWeather seeking Business Development Manager in Atlanta, GA. Only one of these, the network architect position, is a technical job. InformationWeek is NOT typical reading by software engineers, say at Google or Apple. As I also mentioned, the M/L mean years of experience for H-1Bs in their sample is much larger than those of H-1Bs in general, etc. It's just a different population. But Vivek is completely wrong in saying that the underpaid H-1Bs are mainly in the "shoddy body shops that break the law and underpay their workers." First of all, as I said above, one can underpay H-1Bs and yet be in full compliance with the law. Vivek has said this too (http://www.american.com/archive/2008/july-august-magazine-contents/america2019s-other-immigration-crisis): I know from my experience as a tech CEO that H-1Bs are cheaper than domestic hires. Technically, these workers are supposed to be paid a "prevailing wage," but this mechanism is riddled with loopholes. Indeed, a couple of years ago Vivek graciously consented for me to write up the details of his employment of H-1Bs, in a posting to this e-newsletter (http://heather.cs.ucdavis.edu/Archive/CaseStudyH1BUnderpayment.txt). He for example was able to hire an H-1B of top talent, with a Master's degree and some experience, for about 15% less than he would pay an American of ordinary talent, only a Bachelor's degree, and no experience--again, fully legally. Secondly, it's not just the "shoddy body shops." Presumably Vivek doesn't consider the two companies he founded to fall in that category, and more importantly, the mainstream, household name industry firms use H-1Bs for cheap labor too, as I've shown before in detail. By the way, the bodyshops would not be in Kandilov's sample, as they sponsor very few of their H-1Bs for green cards. Finally, one more point on Big Bag/Little Bag: Vivek, and many others, conclude from this that the "remedy" is to give fast-track green cards instead of H-1B work visas to foreign graduates of U.S. universities. Although this would address the issue of exploitation, it would NOT protection for American workers (natives, naturalized citizens, green card holders). Again, one of the most overlooked issues about H-1B as cheap labor is that, even with salary parity, the H-1B program provides employers with a vehicle for rampant age discrimination. It expands the pool of younger (thus cheaper) workers, whom employers hire in lieu of older (thus more expensive) Americans. The beneficiaries of the fast-track green card proposals would, by definition, be young, and thus would produce the same problem for older (age 35+) American workers. I've enclosed the Wadhwa blog below. Norm http://wadhwa.com/blog/2010/05/29/why-policy-makers-should-review-the-facts-before-marching-to-the-drumbeat-of-the-xenophobes Vivek Wadhwa Tech Entrepreneur, Academic, Researcher and Writer May 29, 2010 Post Under TechCrunch Why Policy Makers Should Review the Facts Before Marching to the Drumbeat of the Xenophobes The H-1B visa has become the beachhead in the battle against the legal immigration of skilled workers. This is because it allows highly educated, skilled workers from abroad to take employment in the U.S. and eventually become citizens. Anti-immigrant groups believe that they can close the door to foreigners by restricting or abolishing the visa. So they have been trying to convince lawmakers that H-1Bs depress wages and take jobs away from American workers. To prove their point, they highlight examples of unscrupulous body shops that underpay their workers, and they cite questionable research published by other anti-immigrants. But a new peer-reviewed study, published in Management Science, a top academic journal, challenges these claims. This research finds that foreign-born I.T. professionals on temporary work visas actually earn more than their American counterparts; and that limits on H-1B visas cause the salaries of foreign workers--rather than of Americans--to increase. This, along with research completed by my colleagues at UC-Berkeley, Duke, and Harvard, confirms what most people in Silicon Valley already know: that foreign-born I.T. workers complement American professionals and make the pie bigger rather than take jobs away. This new study was completed by University of Maryland professors Hank Lucas and Sunil Mithas, using data from a survey of 50,000 I.T. professionals that InformationWeek and Hewitt Associates conducted from 2000 to 2005. After adjusting for educational qualifications, work experience, and other individual characteristics, the researchers found that I.T. professionals without U.S. citizenship earned 8.9% more than U.S. citizens. Tech workers on temporary visas were paid 6.8% more than those with U.S. citizenship; green-card holders took home 12.9% more than their American-born counterparts. In years when Congress increased the numbers of visas available, salaries of foreign workers fell. The salary premium for H-1B holders was 10.6% in 2000; 8.1% in 2001; 17.5% in 2004; and 4.7% in 2005. This corresponds to a visa cap of 115,000 in 2000; 195,000 in 2001; 65,000 in 2004; and 85,000 in 2005 When the number of visas rose, the corresponding salaries shrank. To explain why firms pay more to foreign I.T. professionals, the researchers speculate that such foreign workers supply intangible human capital: knowledge of markets and cultures outside the U.S.; social relationships with colleagues at firms located in their home countries; access to networks in home countries and the ability to spot, circulate, and mix ideas and skills from different parts of the world. Foreign-born I.T. professionals are also usually willing to travel for extended periods. The researchers say that because the proportion of the sales and revenue that companies receive from abroad is increasing, they need to hire more I.T. professionals who are familiar with foreign culture; language; and business practice. UC-Berkeley dean and professor AnnaLee Saxenian refers to the foreign-born professionals in Silicon Valley as the "new Argonauts". She says that over the past three decades, the region's immigrant professionals have helped open foreign markets; identified overseas sources of talent and innovation; and pioneered long-distance partnerships. Far from being zero-sum, this process continues to expand economic opportunities for both Silicon Valley and its collaborators. She says that foreign nationals are clearly a complement to the native workforce, their diverse perspectives and global networks being catalysts for local innovation. The data Lucas and Mithas analyzed did not include all tech professionals: Information Week readers are typically I.T. workers in large and midsized American corporations. The researchers say that they compared their sample with that of other studies and believe that this is more representative of the tech industry than any other research. And unlike other studies, this compares apples with apples (previous research has looked only at aggregate data, not at individual education and skills). The Lucas/Mithas study clearly did not account for the shoddy body shops that break the law and underpay their workers. But it is likely to be representative not only of the Information Week readership, but also of foreign-born I.T. workers in Silicon Valley. (Silicon Valley tech workers usually have the same advanced educational background and receive the same competitive salaries as their brethren in the I.T. world.) These are the skilled immigrants that American industry needs--to remain globally competitive. The study also highlighted an issue that I have written about: these workers are earning less than they should. Because of a flawed immigration system, the U.S. has admitted large numbers of workers on temporary visas, but has never increased the number of permanent resident visas (also called green cards) to enable these workers to make America their home. Waiting times for green cards can be longer than a decade, and while workers wait in queue, they can't easily change jobs. Employers often take advantage of the situation. Indeed, Lucas and Mithas found that H-1B visa holders earned 6.1% less than equivalent green-card holders. There is a far more serious problem brewing: we're not letting the foreign-born workers start companies and boost the economy, so they are leaving. Consider that 52% of Silicon Valley companies started in 1995-2005 were founded by immigrants. There are now hundreds of thousands of highly educated and skilled workers who also could be starting companies that are stuck in "immigration limbo". Many are getting frustrated and returning to their home countries. Others languish in the same jobs they had when they stated their green-card-application process; just as they can't change companies, they can't receive a promotion, so if they started as a computer programmer, they can't become an architect or manager. Policy makers are used to the constant drumbeat of the xenophobes. To educate policy makers on the facts and to make them aware of the green-card backlog, a group called Immigration Voice is organizing Advocacy Day on June 7 and 8 in Washington, D.C. This group represents the foreign-born I.T. workers sampled in the Lucas/Mithas study and more than 250,000 doctors, PhDs, research scientists, scholars, engineers, and professors also stuck in the green-card backlog. Hundreds of Immigration Voice members will be travelling to the nation's capital to tell their stories. They have set up more than 400 meetings with members of Congress and the Senate. Let's hope they get through and introduce a little logic into the immigration debate.