Author Topic: Duke studies of offshoring  (Read 2501 times)

0 Members and 1 Guest are viewing this topic.

Offline davkakach

  • Pro JTFer
  • *****
  • Posts: 508
Duke studies of offshoring
« on: November 13, 2006, 01:26:00 PM »
http://www.numbersusa.com/news?ID=6866

Duke studies of offshoring
November 13 , 2006 -- by Prof. Norm Matloff

There seems to be a bit of an academic civil war developing at Duke University over the offshoring issue, with the recent release of two dueling studies.
 
On the one hand, there is the study by Arie Lewin and Carine Peeters of Duke's Fuqau School of Business, and Mahadeva Mani and Vinay Couto of the consulting firm Booz, Allen and Hamilton. You can download it at https://offshoring.fuqua.duke.edu/pdfs/globalization_of_white_collar.pdf
 
On the other hand, there is the study by Vivek Wadhwa, an "academic without portfolio" (his actual title is Executive in Residence) in Duke's School of Engineering. In this e-newsletter I've previously described Wadhwa as the rare (former) tech executive who is capable of
 
"thinking outside the box" on the offshoring issue; see http://heather.cs.ucdavis.edu/Archive/VivekWadhwa.txt for my comments. Wadhwa's new study is at http://memp.pratt.duke.edu/news/?id=793
 
So, what do these two studies say? In short, it is about shortages (pun intended). BAH says that firms are offshoring tech work not just to save money but also to remedy a tech labor shortage; the latter is no problem, BAH says, since Americans couldn't fill those jobs anyway.
 
Wadhwa says, "No, there is no tech labor shortage. We surveyed tech firm HR people and the like, and they told us there was no problem."
 
It so happens that I'm teaching a freshman seminar course on immigration this quarter (http://heather.cs.ucdavis.edu/~matloff.frs.html). I'm really enjoying it, since it has been many, many years since I've taught nonengineers (and for that matter, freshmen). One of the things I hope to impart to them is critical thinking (much ballyhooed in the high schools, but almost totally ignored in actual educational practice). And the very first thing I brought up with them was that they should be quite wary of what any "study" says, or of what any seemingly-neutral person in authority (e.g. academics) says. There is quite often an agenda, sometimes hidden.
 
In our case here, BAH appears to be the financial sponsor of the Fuqau study. That of course is a danger signal from the beginning; as John Miano pointed out to me, "Who ever heard of an academic study that devotes a full page to listing its sales offices?" BAH concludes that offshoring is both necessary for business and good for the U.S. This may be meant to be a message to the Democrats not to get in the way. (They needn't worry about that, as I've pointed out before, e.g. in my exposure of presidential candidate John Kerry's duplicity on offshoring in 2004.) But it is also likely also meant to support the industry's current push for an increase in the H-1B visa cap.
 
The Wadhwa study did not have commercial funding.
 
Both studies consists of business surveys, but they are poles apart, not just in their conclusions but also in terms of methods. The Wadhwa survey states exactly who was surveyed, exactly what they were asked, and even how the integrity of the data was preserved. The BAH does none of this.
 
Much more importantly, the Wadhwa survey asked employers precisely how their hiring is going. How long does it take to fill a job? What is the quality of the people hired?
 
Clearly the conclusions of the Wadhwa study, which I'll detail below, are more in line with the way I view things. But as longtime readers of this e-newsletter know, I've never hesitated to criticize studies, claims and so on which supported my point of view but were based on sloppy thinking and questionable data.
 
In that light, I must say that there is simply no comparison between the two studies. Even someone who knows little about the subject matter would find that the Wadhwa study is more concrete and precise, whereas the BAH report consists of one unsupported claim after another.
 
A case in point is the BAH study's claim that part of the offshoring trend is due to a lack of American workers with Master's degrees in engineering. They offer no evidence that this is a motivator for employers who offshore engineering work. As I mentioned, they do not provide a list of questions in their survey, but it is fairly clear that they did NOT ask employers, "Do many of the workers you employ overseas have a Master's degree?" I say it's fairly clear that they didn't ask this, because all the authors of the study offer as "evidence" is a vague "correlation" (their word, but again not quantified or studied) over time between the numbers of Master's degrees granted to Americans and the amount of offshoring.
 
The starting salaries for Master's graduates in electrical engineering and computer science, adjusted for inflation, have been FLAT or FALLING since 1999 (www.cis.org/articles/2001/back301.html). Employers admit that many of the engineers they are laying off have Master's degrees. And based on my experience and those of others I've talked to, I am reasonably sure that only a tiny fraction of engineers who do work offshore for U.S. firms have a Master's degree. So the whole thing is a phony issue; the industry lobbyists know that Pushing the Education Button is a great way to distract attention from the real issue, which in this case is cheap labor.
 
One more word on the BAH study: In case some of you readers are wondering, there really are precious few rules, or even customs, in academia to safeguard against undue influence from a research sponsor, in this case BAH. At any major research university, research funding is sought with a maniacal obsession; any safeguards would be attacked as "bad for business."
 
As noted, I do believe that the Wadhwa study was done with rigor, and is a valuable contribution. Here are the main points:
 
Those surveyed were "presidents, division heads, managers and senior HR representatives"-- quite a contrast to getting "data" that has been spun by PR people. The job category focus of the study was "computer science and computer technology."
 
The survey began by asking about various measures which might indicate a labor shortage. One of these measures was rather ingenious, what the survey termed "acceptance rates"--the percentage of job offers which are accepted by the offerees. If there were a labor shortage, job seekers would have many offers, and thus the acceptance rate would be low. The survey reports:
 
# We asked how acceptance rates have changed over the past three to five years. 80% of respondents reported that acceptance rates had stayed constant or increased (see Figure 3 below).
 
This would indicate that the industry lobbyists' claim that the job market has rebounded is not correct.
 
The survey then reports data showing that 80% of the jobs are filled within four months. I think the authors' spin on this is that this too contra-indicates a labor shortage. Some people may disagree. But I wish to add that that datum concerns how long it takes to hire someone AT THE PRICE THE EMPLOYER WANTS TO PAY. If they were to offer a much higher wage, they'd get someone right away. This is a key point; remember, whenever the industry claims a shortage, they mean a shortage of cheap labor. Now if the industry were to say, "Gee, we can't make a
 
profit unless we turn to cheap labor, either by offshoring or by importing H-1B visa workers," I would have very little comment. My sole objection all along has been the gross deception of the public and Congress by the industry lobbyists.
 
The survey asked a series of questions about the employers' assessment of the quality of their American engineers. Here are some highlights:
 
# On the question "what changes have you seen in the capabilities or skills of engineers that you have recently hired from those that you hired 3-5 years ago?", the top response was that graduates had better technology and programming skills.
 
# On the question "what capabilities do your US engineers have that make it advantageous to keep their jobs in the US", the response was that U.S. engineers have a very good understanding of U.S. consumer needs, culture, business, and better communication and interpersonal skills. Respondents also stated that U.S. engineers were more creative, excelled in problem solving, risk taking, networking and had strong analytical skills. Plus they could work on high security applications and had the advantage of proximity to resources.
 
# We asked "how similar or different are the types of engineering jobs that your company performs in the U.S. from those that your company has offshored?". 44% said that their company's U.S. engineering jobs are more technical in nature vs. 1% that said that their offshore engineering jobs are more technical in nature. 33% said that jobs were equivalent. When asked to compare the productivity of the engineering workforce between their U.S and offshore facilities, 37% of respondents stated that U.S. engineering employees are more productive, while 24% stated that U.S. and offshore engineering teams are equivalent in terms of productivity. See Figure 6 below. We also asked companies to compare the quality of engineering work between their U.S. and offshore facilities. 38% said their U.S. engineering employees produced higher quality work and 40% reported that work quality was equivalent between U.S. and offshore facilities. 1% reported that their company's offshore engineering employees produce higher quality work. A full breakdown of these statistics can be found in Figure 6 below.
 
# We asked if the current U.S. engineering workforce met their business needs for entry-level engineers. 75% of companies who expressed an opinion said that India had between an adequate and large supply of well-qualified entry-level engineers vs. 59% for the U.S and 54% for China (see Figure 8).
 
(Point added by NM: A BusinessWeek analysis found that Bachelor's salaries at the entry level have been flat for several years, just as with the Master's-level analysis I did.)
 
# To the question, "What are the key reasons why either the U.S., Indian or Chinese entry- level engineers might not be able to meet your needs?", the responses were as follows: US: Respondents cited salary demands as a key issue, limited supply of available people, and lack of industry experience. Some respondents stated that there were no issues. A small minority of respondents raised issue with inclination towards non-technical work, an unwillingness to relocate and poor work ethics. China: The top response was that Chinese engineers lacked communication skills. Other issues included visa restrictions, a lack of proximity, and inadequate experience. A few respondents also cited lack of loyalty, cultural differences, intellectual property concerns and a limited "big picture" mindset. India: Inadequate communication skills, a lack of specific industry knowledge, and proximity/visa restrictions were the top responses. Other issues raised were lack of domain experience, limited project management skills, high turnover and cultural differences.
 
# We asked, "what are the relative strengths or advantages of U.S, Indian or Chinese entry- level engineers when compared to each other?". The responses were as follows: U.S.: Strong communication skills, an understanding of U.S. industry, superior business acumen, strong education/training, and a sense of creativity and desire to challenge the status quo. A few respondents cited strong technical skills, proximity to work centers, and a lack of cultural issues as advantages. China: Many respondents stated that the key advantage of hiring Chinese entry-level engineers was cost. A few respondents cited strong education/training, work ethics and a willingness to work long hours.
 
Norm


Dr. Norman Matloff is a professor of computer science at the University of California at Davis, and was formerly a professor of mathematics and statistics at that university. He is a former database software developer in Silicon Valley, and has been a statistical consultant for firms such as the Kaiser Permanente Health Plan. He has published numerous research papers in computer science and in theoretical and applied statistics, in fields such as parallel processing, data security, data mining, computer networks, and statistical regression analysis. He is the author of one textbook in computer science and another in statistics.
Tolerance becomes a crime when applied to evil.   --Thomas Mann