Nobel prize winner and ex-Chicago economist Myron Scholes came and spoke at the Chicago GSB today. He is perhaps most famous for coming up with the Black-Scholes option pricing formula back in the day, and for his ill-fated Long Term Capital Management hedge fund which blew up so spectacularly in 98.
It was clear that between 98 and today he's devoted all his time to trying to figure out what happened in financial markets when Russia defaulted on its rouble debt. I was fortunate(?) enough to have a front row seat through all of this as I was working for a hedge fund at that time. The Russian default caught people by surprise, and some folks lost lots of money. To recover from their loss, they had to sell some of their holdings, which depressed the price of those holdings. This made other people lose money, forcing them to unwind their positions into falling markets as well. Which depressed prices further. Long story short, things got very ugly. Prices that were uncorrelated began to move in synch, and all the liquidity dried up (read: became fantastically expensive). Everyone's models broke, prices went way out of wack, and lots of people lost their shirts.
Myron pointed out the following things:
1) We need to reconsider the distinction between idiosyncratic and systematic risk. Standard portfolio theory tells us that since idiosyncratic risk can be diversified away, no one should expect to be compensated for carrying it. But during shocks, risk that used to be idiosyncratic becomes systemic as unrelated prices start moving together. So what do we do now?
2) The price of liquidity also changes, becoming very expensive during shocks. This suggests taking out liquidity insurance by buying it cheap in unrelated markets ahead of time. But what does it mean if no one is buying or selling, even at (insanely) wide spreads?
3) Models drive buying and selling decisions. When the models break, people need to come up with new ones and figure out what went wrong with the old ones. This takes a really long time and there will be no trading activity while this is happening. Good teams will be able to reenter markets sooner.
4) Asset allocation and risk management needs to think of risk optimization, not minimization. I had no sense of what risk optimization meant, and I'm not sure Myron's sure either.
5) Speculators exist only to provide liquidity and bear risk. They need to get paid for this, even though people don't want to pay them. Risk transference is changing the banking business from a matching service (lender, meet borrower) to a factory where they produce appropriate financial instruments for a fee and they pay specialists to take them off their books.
His point about the way capital structures financing esoteric ventures (including most third world development, btw) can switch from being uncorrelated to moving together during a shock is well taken, and a huge problem. The Economist recently recommended that developing countries switch entirely to FDI (foreign direct investment -- factories) because the volatility short term capital flows created were too damaging. The alternative would be for the country to take on complex short positions against other third world countries, but I don't know if this would work out to be cheaper. Tricky.