Wednesday, August 16, 2006

Portrait of a quant II

Those funsters at Businessweek have another goofy bit about quants in their latest issue. (See their previous article Math Rules! and here for previous quant-related posts.)

Well, they have the general idea right. If you can't read the text in the figure below, it says " from MIT, Caltech or India in applied mathematics, physics, computer science or all of the above", "800 math SAT de rigueur" and "math tournament rock star" :-)

The Quintessential Quant: James H. Simons, the former math professor who founded the $12 billion quantitative shop Renaissance Technologies Corp., pocketed an estimated $1.5 billion last year. That was thanks to the 5% in fees and nearly 44% of profits that Renaissance docks its investors (vs. traditional hedge funds' typical "2 and 20"). Clients don't complain; Renaissance's leading fund has returned 35%, after fees, since 1989. And D.E. Shaw & Corp., the brainchild of ex-Columbia University computer science professor David E. Shaw, with $23 billion in capital, has netted investors 21% a year for 17 years, without a single losing 12-month stretch.

Landing a job at either of these shops can be insanely lucrative -- and even more insanely competitive. "Using a self-consciously obnoxious term, we're looking for superstars, the kinds of people who would be extraordinarily good at nearly anything," says Nicholas P. Gianakouros, head of global recruiting for New York-based D.E. Shaw.

He is being euphemistic. The handful of quant and programming geniuses who get into the toughest mathematics, physics, and computer science PhD programs on the planet are already best in class. So screening for the 5 or ten very best of that best means establishing a whole new set of prerequisites. "The quant shops are a different animal," says Alison Seanor, vice-president at Glocap Search, a Manhattan hedge-fund recruiter. What is the "it" factor that distinguishes the crème de la crème? All Seanor will say is, "I know it when I see it."

One obvious filter is that liberal arts students -- or even bankers and stock jockeys -- need not apply. What you will need is a nosebleed grade-point average in applied mathematics, physics, or computer science at an elite school like the Massachusetts Institute of Technology, California Institute of Technology, or Indian Institutes of Technology. Many of these students are published and have won high math honors such as the Putnam Fellowship. Often, their names are already so well known in the field that the quant funds make the first approach.

Another must-have: an 800 math SAT score (even if you sat for that exam in your awkward adolescence). Although the funds diplomatically claim the number is "just another data point," it's pretty well understood to be a critical credential.

The quant shops want malleable intellect untainted by Wall Street dogma -- i.e., not "buy, sell, or hold" types. "They're not really looking to make money on corporate events like takeovers," says Emanuel Derman, director of the financial engineering program at Columbia University and head of risk for quant house Prisma Capital Markets. "They're looking to make money on mathematical models." Top funds often advertise in esoteric scientific journals. "You'll not likely find our ads in a dentist's waiting room," says D.E. Shaw's Gianakouros.

If yours is one of the lucky 1% to 3% of résumés to survive an exhaustive initial culling, you can look forward to an hour-long phone interview peppered with thought problems and brain teasers. Pass that test and you will then be summoned as many as three times to undergo up to a dozen grueling interviews. "Every interviewer uses a different approach," says Gianakouros, citing programming problems and math proofs. Expect to be asked to build an intricate Excel model on the spot. Whatever the case, advises Derman, "don't say anything unless you're ready to be quizzed on it."

The firm will then solicit references for areas in which a candidate may appear weak. Ultimately, it takes a consensus among everyone who has met the candidate to extend a coveted offer. D.E. Shaw says that out of every 500 candidates who got the initial callback, only one makes the final cut. Many agree it's even harder to get into secretive Renaissance, which would not comment for this story.

A typical offer, say sources, starts with a base salary of around $250,000, plus a guaranteed annual bonus that could double that. The best can command a cut of a fund's upside -- beaucoup bucks when you consider the multibillion-dollar asset pots. All this, yet, says Seanor, "most of these guys have never even had a real job."


Anonymous said...

Given a field of applicants with essentially perfect quantitative skills, I wonder when assets like communication skills, or something even less "quantitative", begin to take on value to the interviewer/company. At some point, sheer analytical brain power is no longer your limiting factor.

Steve Hsu said...

Umm... there is no maximum value for math ability (the normal distribution has a long tail :-)

Now, the question is whether returns to math ability are diminishing, or perhaps increasing nonlinearly :-) In the case of Perelman and the Fields medal, I think they are clearly monotonically increasing...

In the case of finance, it is quite likely that after a while it is some combination of common sense, market intuition or psychological insight, combined with math ability, that determines how good you can be. People skills are of course important if you are working on or leading a team, but then again top traders on Wall St. are notoriously aggressive personality types.

I doubt either Buffet or Soros, to take two examples, are really top level mathematical talents. Although they are probably quantitatively reasonably strong they are certainly not in, e.g., Simons class.

Anonymous said...

A few comments:

* Although many top math,CS, physics students end up working as quants, I still think the very best still stay in academia. That does not mean anyone who gets to stay in the academia are better than those who left(often that is not the case, actually). I just don't see the work being as interesting to the geniuses.

I don't see Perelman or anyone who is a "shoe-in for tenure" caliber person ever working as a quant. Incidentally, the rumor is that Perelman will decline the Fields medal and his Millon dollar part of the Clay price.

* Much more impressive than the 800 SAT math scores is performance at the Math Olympiad(or Putnam later). Interestingly, Perelman got the perfect score.

* Amazingly, they don't mention Chinese students(say Tsinghua?). Have you seen the Math Olympiad scores in recent years? It is interesting to note Chinese doing very well(including perfect scores) in recent times. In a few years, there will be Fields medallists from there...

* I think it is more impressive being a successful entrepreuner and setting up startups than working for someone else. Not to mention doing both academia and startups ;-)

* What is it about being able to build an "intricate Excel model"??? :-)


Anonymous said...

Businessweek is more interested in selling magazines than accurately representing reality. An 800 math SAT is something that probably sounds very impressive to your typical Businessweek reader, but at the level we're talking about here looking at SAT scores would be like looking at a candidate's report cards from elementary school. I think that half of my class at Caltech had 800 math SATs, and the other half were chemistry majors.

Steve Hsu said...

MFA and Anonymous: Ha! you are both very perceptive. To the journalist, Excel models and 800 SATs are impressive, but not to the cognoscenti :-) (Actually, Excel is used a lot, but real coding in finance is done in C++. I think 800 on the current SAT is the modal score at Caltech these days.)

MFA, I think I posted an article some time ago (NYTimes) which mentioned that the top traders for PBOC (Chinese central bank) are recruited for their offscale math abilities. It's true -- the Chinese teams post a lot of perfect scores these days.

At the top levels of finance (i.e. at hedge funds, or managing directors at banks) I see a lot more Indians here in the US than east Asians. I think it's English communication as well as some other cultural things at work.

Anonymous said...

Thanks for posting the caricature, as I was unaware of the "flat-front vogue" in dress pants. Now I can avoid it more carefully.

Seth said...


Chinese Fields medalists won't be new ... only that they live and work in China rather than emigrating as with S-T. Yau.

Anonymous said...

Professor Hsu,
Surely with your large web of math contacts, you can come up with some information as to what securities and methods Simons uses in his math models? Any info at all, no matter how vague, would be useful.

Blog Archive