Graduate admissions, human capital and globalization

I'm on the graduate admissions committee this year. We're in the middle of reviewing over 200 applications for about a dozen places in our PhD program. About 3/4 of the applicants are foreign, the vast majority of those from China, while less than a quarter are domestic. We're a middle-level PhD program at a flagship public university -- not ranked among the top 20 physics departments, but probably not very different from the next 20 departments, with several very strong research specialties.

Through some independent research, I've learned that the top 5% or so of students from the very best universities in China (e.g., Beijing University or Tsinghua University) can expect to be admitted to the top 10 programs in the US, so they might not apply at Oregon. A typical student admitted here from China will be in the top 10-20% of their class at one of the top 5-10 departments. The modal GRE quant score is a perfect 800 (>94 percentile) and a typical GRE subject score of an admitted Chinese applicant is above 90th percentile. Very few American applicants score above the 90th percentile on the subject test (the top ranks are dominated by foreigners), and those individuals usually end up at top US grad schools.

Our yield (ratio of students who enroll to those who are accepted) is about 10%, probably because most of these applicants are accepted at multiple programs of similar quality. Among our applicants I've already spotted winners of the Chinese national physics and math Olympiads (these teams often score 5 golds at the world competition) and numerous graduates of the "Special Youth" program which admits young talents to university early. One applicant is 19 and took general relativity at age 18. (As I mentioned already, the very strongest Chinese applicants are only applying to the top 5 or 10 US graduate programs, so we may not see them in our pool.)

How do these students perform in graduate school? Typically very well in the courses and on the comprehensive exams. The top scorers on the PhD candidacy exam, both here and when I was a professor at Yale, are disproportionately from China (and rarely Americans, although I think when I was a grad student myself at Berkeley there were often Americans at or near the top). The harder question is whether, in the long term, these Chinese students are as productive researchers as the domestic students. I haven't seen any systematic studies of this, but I suppose they are at least comparable since a lot of professors these days are PRC immigrants who went through graduate school here. I suspect that mastering English might be harder for many of these students than the physics itself.

One thing I am sure of -- without these foreign applicants the strength of US graduate programs in physics would be reduced significantly (I am sure the situation is the same in engineering and other sciences). We're reaping the benefits of a very well organized system for training students up to the BS and MS levels; most of these students want to stay in the US after finishing their educations. I suppose more and more good students are staying at home for their PhD, rather than coming to the US, but at the moment we can still draw on a huge pool of human talent from abroad.