Skip to content →

The Worlds of Yesterday

Written for the Risk Magazine 15th Anniversary Book on Risk Management.

I never expected to be in the business world, not even a few years before I finally entered it. I got here by a series of languid apparently low-volatility Brownian diffusions within one little world interspersed with sudden jumps to another. The transitions weren’t truly unexpected; their probability built up inside me up like a bubble, slowly and predictably.

Growing up in Cape Town, where the college system required that you specialize immediately, I knew I wanted to be a scientist. Somewhat regretful to leave the more expansive world of other things, I nevertheless registered for a Bachelor of Science degree.

I began by studying physics, chemistry, pure and applied math. I liked theory, was bad at lab work. In my second year I dropped chemistry and the next year I abandoned pure math, steadily mean-reverting to theoretical physics and applied math. Each course lasted a full academic year, and terminated in a massive final closed-book three-hour exam preceded by a several-week-long study period. You had to remember everything you might need in the large examining room filled with hundreds of people taking several different finals – Fourier and Mellin transforms, indefinite integrals, the works. We learned thoroughly and repetitively, so that after three or four years what little I knew I knew very well.

Physics was entrancing, precise and cosmic at the same time; it seemed a good way to spend one’s life. At the start of my Honours year in Cape Town I found myself applying for scholarships overseas, and ended up at Columbia University doing theoretical particle physics in the late 1960s. Though we didn’t know it then, particle physics was nearing the end of its classification of the hordes of newly discovered particles and their symmetry-violating interactions. It was also the beginning of the era of unified quantum field theories. I wrote my 1973 thesis on how to test the Weinberg-Salam electro-weak model’s prediction of small amounts of parity violation in electron-nucleon scattering, an effect that was then discovered at SLAC in the late 1970’s in an experiment that set the final seal of approval on the model.

From 1973 to 1980 I continued to do research in particle physics at various universities around the world. Many applicants, few jobs; exhilarating ups, dispiriting downs. Post-doc jobs, if you found them, usually lasted no more than two years, and permanent jobs were harad to find, so you had to publish a quick paper and then start looking for your next musical chair. My wife was an academic too, and we lived peripatetically, each of us in different places half the time.

By 1980, I was an assistant professor in Boulder, Colorado with a wife and son in New York City, so I left physics and took a job at Bell Laboratories in Murray Hill N.J. I felt treasonous, for I thought then of the academic world as some great nirvana where you devote each day to the transcendental rather than the mundane. That was partly true, I suppose, but it was also part illusion, a powerful one.

I worked for 5 years at Bell Labs in a Business Analysis Systems Center, where, despite its name, I learned mainly about computer science. What intrigued me most during that time was writing parsers and lexical analyzers, because until then I had thought of computing mainly as a numerical activity. I divided my time at the Labs more or less equally between the pleasure of creating little UNIX worlds that users could control via the little modeling languages I wrote and between the displeasure of living in the little world of large bureaucratic organizations. Though I didn’t learn much about finance during this time, UNIX and computer science turned out to be a near perfect background for a Wall Street where, when I arrived, quants still had to do all their own systems-building.

Headhunters came calling. Slowly I grew accustomed to the idea of leaving the Labs. Finally, in late 1985, I jumped to Goldman Sachs’ fixed income research group, and I found I loved the realm of financial modeling. No-one stood on ceremony, talent of any kind was almost always a virtue, work was informal and spontaneous, and it combined computing with the spirit of theoretical physics. Everything I had learned in those areas came together there.

Nowadays financial engineering is a discipline; in 1985 on Wall Street it was amateur heaven, a fluidly makeshift field filled with retreads from other fields who could learn quickly, solve equations and write their own programs. I liked it that way. There were only two textbooks I knew of — Jarrow and Rudd, and Cox & Rubinstein — and the only derivatives meeting I went to each year was the annual spring meeting of the Amex. Now there are tens of thousands of books, thousands of conferences and hundreds of degree programs.

I spent my first four years in the world of fixed-income derivatives, where the dominant issue of the 1980s was how to extend and calibrate the replication methodology of Black, Scholes and Merton to interest rates. Working with the fixed-income bond options traders was wonderful; it was a small world of motivated quants and traders, the latter of whom were eager to embrace whatever new model you could create.

I spent the next ten years in equity derivatives, where the dominant issue was the smile. The group I ran, Quantitative Strategies, was full of hands-on people who were trying to understand risk at the front lines and then build systems to control it, not always in that order. What I liked most, when you could find it, was working closely with little worlds, small groups of quants, traders and salespeople interested in valuation and in doing something about it, without too much bureaucracy or too much rigor. We were lucky to be theorists living in the experimenters’ laboratory, the first to hear about new irregularities. I liked the mix of business and theory. I enjoyed thinking about things slowly, figuring them out together with my colleagues, and then explaining them by talking or writing.

I spent my last two-and-a-half years in the bigger world of firmwide risk management, and then, in mid-2002, I left Goldman Sachs to take a break and write a book. I still like the world of financial research with its mix of theory and experiment, science and sociology, models and computer systems, doing and teaching, and expect to return to it.

What did I learn from all this?

I started out in 1985 thinking of quantitative finance as a branch of the natural sciences – I imagined you could search for an all-encompassing theory that would explain everything. Over the years, I’ve come to see less evidence for universal laws in prices, and progressively more evidence of the vagaries of human behavior. I am still amazed and awed that anyone trades anything on the basis of a model. But I’ve learned that trading with a model is not the simple procedure academics imagine. Intelligent traders iterate between imagination and model use in a way that belies easy categorization and testing. It’s not a clockwork universe out there, and therefore, I learned to avoid needless rigor and axiomatization.

It pays to build models as though you were working in the natural sciences, but you have to keep reminding yourself that you aren’t. It’s unbelievably hard to build a truly successful model. I learned what a once-in-a-lifetime thing a discovery like the Black-Scholes model is – not necessarily right, ” but easily embraceable because it asks just enough of you, but not too much.

I learned that building trading systems that make it easy to use models is at least as hard as building the models themselves, and takes many more people. I learned to try to avoid middlemen: there’s no substitute for communicating directly with traders and other model users.

Most importantly, I learned the value of colleagues. I’ve seen how easy it is to go down wrong paths when you work alone, and how just one remark from someone can set you straight.

Finally, I kept relearning that I like little worlds better than big ones, and that for me, the path to bigger issues is through the details of little ones.

I’ve been lucky to have been in this business at a time when it was still developing, when small contributions could exert disproportionate leverage. I wish, of course, that I’d been in this field even earlier.

Published in other