Red-Blooded Risk: The Secret History of Wall Street

Three Paths

Nevertheless, many of the ideas in my dissertation were validated by the crash. And there were lots of new ideas from other people in the air. The question was how to identify arbitrage strategies and run them so they survived crises. As usual, we split into three groups.

The frequentist card counters who ran small hedge funds or invested their own money zeroed in on the concept of capital. You had to have the right amount of capital for the set of strategies you ran. This was a theoretical calculation you could make, using observations of past frequencies plus market data. For example, suppose looking over the past you discovered that one time in a thousand, a bond rated single-A at the beginning of a year would default during the year. Then suppose you estimated, also using past frequencies, that your strategies would be down $1 million or more at their low point in one year, with probability 1 in 1,000. Thus, if you held $1 million of capital to run your strategies, you had one chance in a thousand of going broke, so your strategies were the equivalent of a single-A-rated bond. You could observe the market yield on single-A-rated bonds; that was your cost of capital.

In practice you didn't use only single-A capital. You wanted a mix of capital, just like any operating business. Some would be safe and cheap, AAA-bondlike, and some would be risky and expensive, equitylike. By constructing your capital carefully, you could design a dynamic business that could survive ...