Ten Sigma

Money & Markets · Greenwich, 1998 · LTCM, Merton & Scholes

TL;DR: Long-Term Capital Management had two Nobel laureates, returns of 40%, and models that priced its August 1998 losses as so unlikely the universe wasn't old enough to expect one. The Fed had to convene fourteen banks to unwind it before it took the system down.

What actually happened

LTCM's convergence trades were individually tiny edges, leveraged roughly 25-to-1 (far higher including derivatives) into billions. The models assumed markets stay liquid and panics stay local; Russia's default in August 1998 made every 'uncorrelated' position correlate at once, because the common factor was LTCM-style leverage itself.

The fund lost $4.6 billion in weeks. The New York Fed brokered a $3.6 billion private recapitalization: not a bailout with public money, but an admission that one hedge fund's book had become systemic. Merton and Scholes had won the Nobel for the option-pricing mathematics the year before.

Ferguson's framing, kept in the episode: the models weren't wrong about the past. They were only wrong about how much past there was.

The play to remember

Leverage converts being right eventually into being insolvent now. Ten-sigma events are usually one-sigma flaws in the model.

Sources & fact flags: Lowenstein, When Genius Failed; Ferguson, The Ascent of Money; Fed testimony (1998).

New cases drop daily on TikTok and Instagram. The first 200 on the waitlist get 3 months free at launch.

Join the Vellum waitlist