Date: Fri, 5 Jan 2007 00:22:56 -0800 (updated) From: Norm Matloff To: Norm Matloff Subject: new Duke study To: H-1B/L-1/offshoring e-newsletter Vivek Wadhwa is a former tech CEO who, finding himself with a lot of time on his hands, has become a Renaissance Man of sorts, with activities ranging from writing for BusinessWeek Online to an adjunct position at Duke University to--hey, this is for real--Bollywood film producer. Good for him. He's a nice guy, and I always admire people who can get involved in several diverse fields. I've alternately panned and praised him. On the one hand, a check of the Dept. of Labor's H-1B Web page shows that Wadhwa's firm hired an H-1B Software Engineer for $44,144, and a Computer Programmer for $31,933. Note that even new graduates in computer science get around $50,000, so these are very low by any standard. (The one making $44K had a Master's degree and some experience. See the details in http://heather.cs.ucdavis.edu/Archive/CaseStudyH1BUnderpayment.txt) On the other hand, he refused to buy the then-prevailing wisdom last year that China and India were producing six and five times the number of engineers per year that the U.S. is, and showed that "wisdom" to be wrong. More recently I praised him for his "out of the box" thinking in a study he did which debunked the notion that there is a tech labor shortage, in which he mentioned pointedly, his project "did not take a penny from anyone" in funding. This time, though, Wadhwa collaborated with a very much "inside the box" researcher, Prof./Dean AnnaLee Saxenian of UC Berkeley. Saxenian's pro-industry research is often cited by industry lobbyists (see http://heather.cs.ucdavis.edu/Archive/Saxenian.txt for more background), and I believe that the effect is quite visible in the present work. So, I'm not a fan of this work, as you'll see below. Before getting to the study, I should mention that Wadhwa says this in his latest BW column: It doesn't matter which side is right here [about H-1B]. The bigger issue is that if we do need workers with special skills, we should ask them to come and stay. Let's offer them permanent residence rather than short-term visas. Temporary workers can't start businesses, and we haven't given them the incentive to help us compete globally or to integrate into American society. They can't sink deep roots because their visas limit how long they can stay. We also want to keep the best and brightest students who complete their graduate studies in our universities. They come here with a global outlook and a hunger to succeed. They understand the culture and values of the countries we're going to be competing with in the new global economy. We want these young and brilliant minds on our shores. First, his characterization of H-1B is inaccurate. H-1Bs can indeed convert to permanent resident status if they are sponsored by their employers, and over the years, most have done so. But more importantly, I disagree with this "green cards, not H-1Bs" theme, which is currently popular in certain circles in DC and academia (not to mention IEEE-USA). It is largely based on the fact that 30-50% of the students in U.S. tech graduate programs are foreign. The proponents say, "Let's give them incentives to keep them coming to the U.S. for study, and to get them to stay." Note carefully the shift in attitude implicit in that statement. It used to be that the industry lobbyists would say, "We needs H-1Bs for the short term, but need to get more American students into the field." But now they seem to have abandoned even that, as they now realize what some of us have been saying for quite a while now: Currently it is a losing financial proposition for an American to pursue a PhD, since he would forego five years of industry salary for five years of a small graduate stipend. The only way to change that equation would be to raise the graduate stipends, and they don't want to do that. They're getting the cheap graduate labor from foreign students, so why change? It is basically support for the proposed F-4 visa in S.2611 and the SKIL bill of last year, which I have pointed out would have just as pernicious an effect on American engineers (U.S. citizens and permanent residents) as H-1B does. See my CIS article, at www.cis.org/articles/2001/back301.html and my recent article about the effect of importing foreign engineers on age discrimination in the industry, at http://heather.cs.ucdavis.edu/CLER.pdf Since F-4 would bring in mainly the young, it would exacerbate the already-severe effect that H-1B and employer-sponsored green cards are having on American engineers today. Moreover, Wadhwa's subtle distinction between H-1B and green cards will be lost on most. Even the Post reporter in the article enclosed below thought that Wadhwa was calling for an increase in the H-1B cap when she contacted me for an interview for the piece. The industry lobbyists have already started using Wadhwa's study for just such an end. As to attracting "the best and the brightest" from around the world, I strongly support the notion of bringing in "the best and the brightest" from all over the world, and have personally acted on that conviction by urging various employers (including my own) to hire foreign nationals (mainly Indian and Chinese) whom I consider to be brilliant. But most skilled immigrants are not in that category. Now, my comments on the study: 1. There is a troubling lack of apportionment of value. Look at the analysis of patents, for instance. Suppose a patent names five inventors, one of which is an immigrant. The study credits this entire patent to the immigrant column of the ledger, rather than 0.2 patent. 2. Even more importantly, there is the issue of substitutability. In my hypothetical example here of a patent filed by four natives and one immigrant, the study's analysis tacitly assumes that if that immigrant hadn't been working on the project, the invention never would have occurred. In most cases this is incorrect; without the immigrant, someone else would likely have been hired for the project. Of course, without data we can't tell which of the inventors were the main sources of ideas on the project. But data in my University of Michigan article (see below) indicate that the immigrants are in general no more innovative than the natives, and in fact tend to be less innovative, as they come from cultures which are less conducive to innovation than that of the U.S. 3. There are similar problems with the study's data on entrepreneurship. If an immigrant and a native cofound a firm, the study counts this as one firm in the immigrant column of the ledger, instead of 0.5, and presumes that if the immigrant hadn't been here, the native would not have come up with someone else as a partner. 4. There is no distinction made as to what contribution, if any, the immigrant founded/cofounded firm makes to the technological advance of the industry. The first Saxenian study found, for instance, that 36% of the Chinese-owned firms were in the business of "Computer Wholesaling," meaning that they were simply assemblers of commodity PCs. That certainly doesn't square with the present study's bottom line that "The key to maintaining U.S. competitiveness in a global economy is to understand our strengths and to effectively leverage these. Skilled immigrants are one of our greatest advantages." To put this in an even more relevant context, the study fails to even mention, much less quantify, how many of those immigrant founded/cofounded firms have as their business function the offshoring of American tech work. This is in stark contrast to the study's introductory statement that "The [authors] were concerned about the growing momentum in outsourcing and its impact on U.S. competitiveness--and sought to understand the sources of the U.S. global advantage as well as what the U.S. can do to keep its edge. To better understand the contributions of skilled immigrants to the competitiveness of the U.S. economy, they decided to expand and update Saxenian's study." In other words, the study's claims that immigrant entrepreneurship, especially by Indians, is helping offset our losses to offshoring has it exactly backwards. On the contrary, many of these "entrepreneurs" serve specifically as FACILITATORS of offshoring. This is also true of the immigrant workers they employ; Prof. Ron Hira's work showed that the H-1B and L-1 work visa programs play a key role in facilitating offshoring. (R. Hira, U.S. Immigration Regulations and India's Information Technology Industry, Technological Forecasting & Social Change, 2004.) In this light, compare the graphs on pages 12 and 25. The former says that the Indian immigrants have founded/cofounded twice as many "tech" businesses as the immigrant Chinese, yet the latter says that the Chinese in the U.S. have filed/cofiled 12.5% more patents than their ethnic Indian counterparts. This undoubtedly is due to the presence of all those Indian-owned offshoring companies, which presumably are not producers of many patents, if any. 5. The study implies that the immigrants are, by nature, personality and talent, more entrepreneurial than the natives. Yet its own data show that the immigrants and natives to be pretty much on par with each other. The study says that 53% of the tech workforce in Silicon Valley are foreign-born and that 52.4% of the tech firms were founded/cofounded by immigrants. In other words, the immigrants are not founding/cofounding a disproportionately large number of tech firms. 6. Returning to my issues of substitutability and contribution, the study implies that without the immigrants the U.S. tech industry simply would not be so dominant as it is today. This is a non sequitur. No immigrant-founded/cofounded firm (or native one, for that matter) has been pivotal to the technological evolution of the industry. The industry lobbyists love to use Intel and Google as examples, yet they are in fact nonexamples. Putting aside the fact that Intel founders were all natives (Andy Grove, a Hungarian refugee often cited by the industry lobbyists as founding Intel, was not a founder or cofounder of the firm), Intel's processor chip did not advance the industry one bit (no pun intended). Intel's big impetus of growth was IBM's selection of Intel as its vendor for the IBM PC, but IBM could have chosen other chips from Motorola, NEC etc. (which the IBM engineers would have preferred). As to Google and its immigrant cofounder, Google is cute and I use it constantly, but there are lots of other search engines one can use to the same effect. The study says that the immigrant founded/cofounded companies have employed 450,000 people in the last 10 years. Putting aside the fact that those workers themselves tend to be H-1Bs (the Indian companies tend to hire Indians, the Chinese companies tend to hire Chinese, etc.), this again falsely assumes that if not for the immigrant entrepreneurs, these workers would not be employed in the industry. Yet a counterargument could be made that without the immigrants, the percentage of native workers in the industry would be much higher, and THEY would be the ones to get the VC funds, thus founding the "same" companies and employing the same workers. Another counterargument could be made that without these firms the remaining firms would simply be bigger, and would employ those same people (or slightly less, due to economies of scale). 7. In addition to the very serious flaws cited above, there is a major error of omission: The study analyzes all the "gains" to the system--the immigrants--but completely ignores the losses--the Americans, both native and earlier-immigrant, who are displaced by high-skilled immigration. High-skilled immigration is a major driving force in the rampant age discrimination in the industry. In contrast to "the good old days," in recent years software engineers have found that they start to become less attractive to employers around age 35 or so, and (counter the industry lobbyists' claims that it is an issue of outdated skills), this primarily is because older workers cost more, in salary, benefits and so on. Most of the skilled immigrants are young--the median age for the H-1Bs is 27.4--and thus they enable employers to avoid hiring the older Americans. (See N. Matloff, The Adverse Impact of Work Visa Programs on Older U.S. Engineers and Programmers, California Labor and Employment Law Review, August 2006, and N. Matloff, On the Need for Reform of the H-1B Nonimmigrant Work Visa in Computer-Related Occupations, University of Michigan Journal of Law Reform, Fall 2003.) Moreover, the well-publicized loss of jobs to offshoring and H-1Bs is causing many bright young people to avoid the computer field in the first place. Computer science enrollment is down 50% or more at most major universities. In other words, immigration and offshoring (and again, keep in mind that immigration facilitates offshoring) is causing an INTERNAL BRAIN DRAIN from the tech fields in the U.S. This is not addressed at all in the study. The study may be downloaded at http://memp.pratt.duke.edu/downloads/americas_new_immigrant_entrepreneurs.pdf The Washington Post article follows below. Norm http://www.washingtonpost.com/wp-dyn/content/article/2007/01/03/AR2007010301678_pf.html Immigrants a Driving Force Behind Start-Ups, Study Says Tech Industry Clamors to Get More Visas for Foreign Workers By Krissah Williams Washington Post Staff Writer Thursday, January 4, 2007; D05 About 25 percent of the technology and engineering companies launched in the past decade had at least one foreign-born founder, according to a study released yesterday that throws new information into the debate over foreign workers who arrive in the United States on specialty visas. The report, based on telephone surveys with 2,054 companies and projections by researchers at the University of California at Berkeley and at Duke University, found that immigrants -- mostly from India and China -- helped start hundreds of companies with estimated sales of nearly $50 billion. It was written by a former technology executive who was an immigrant himself. Technology-industry lobbyists have already cited the study in a push to persuade Congress to increase the annual allotment of H-1B visas, which allow U.S. companies to sponsor temporary workers in specialty occupations, such as computer programming and systems analysis. The companies say they cannot find enough Americans to fill jobs; other proponents contend that globalization requires U.S. companies to import talented workers. "This research shows that immigrants have become a significant driving force in the creation of new businesses and intellectual property in the U.S. -- and that their contributions have increased over the past decade," wrote Vivek Wadhwa, the study's author, who immigrated from India with his family as a young man. Another study will be released next month by the Center for Immigration Studies, which supports low levels of immigration. That report says most specialty visa holders come to the United States to do low-level professional jobs for relatively low pay. Wadhwa's study looked at founders of engineering and technology companies started from 1995 to 2005, and analyzed the World Intellectual Property Organization Patent Cooperation Treaty database. About 25 percent of international patents filed in the United States in 2006 were submitted by immigrants. Scott McNealy, chairman and co-founder of Sun Microsystems, is among the advocates for an expanded visa program, writing editorials, calling members of Congress and supporting political action committees. McNealy noted that immigrants Vinod Kosla of India and Andy Bechtelsheim of Germany co-founded Sun. The company "created tens of thousands of jobs that have generated billions of dollars in exports and has created thousands of patents and intellectual-property positions," McNealy said. "Why would you have any arbitrary number on smart people?" Last year, the industry raised the issue in the national debate over immigration reform, but Congress ended its session without acting on the Securing Knowledge, Innovation and Leadership Act. The bill would increase the annual quota on the H-1B visas to 115,000 from 65,000, eliminate green-card caps for some advanced-degree holders and streamline the processing of employment-based green cards. Tech lobbyists want to revive it. "We are working on that new piece of legislation that will hopefully be a great fix for a lot of our companies," said Andrea Hoffman, vice president of government and political affairs for TechNet, an industry lobby backed by hundreds of technology companies, including Apple Computer, Microsoft, and Google. Those who favor low levels of immigration and oppose expanding the specialty-worker programs contend that foreigners accept lower pay and depress wages. Jessica M. Vaughan, an analyst at the Center for Immigration Studies, said an increase in the cap would amount to "a subsidy for business because it allows them to bring cheaper labor from overseas." It is unknown how many of the immigrants who founded technology companies had H-1B visas. At least two Northern Virginia tech companies were founded by former H-1B holders. Sudhakar V. Shenoy, founder and chief executive of Reston-based Information Management Consultants, immigrated to the United States in 1970 after graduating from the Indian Institutes of Technology -- known informally as the "MIT of India" -- and attending graduate school in Connecticut. In 1974, he was offered an H-1B visa, and a manufacturing company sponsored his green card in 1977. Four years later, he founded IMC, which has 350 employees in Reston and 125 in Pune, India. Peter Harrison came from Britain on the specialty visa and later became chief executive of GlobalLogic (formerly Induslogic), a Vienna-based software development company founded in 2000 by two men from India, who were also H-1B holders. The company has grown rapidly and employs 1,600 people in the United States, India and Ukraine. Only a few dozen of them have H-1B visas. "They are very, very hard to come by," Harrison said. "We are always at a challenge to recruit people."