In recent days, I’ve heard a lot of people talking about correlation across financial markets. This article, from today’s NYT, sums up the issues, which I find interesting and perhaps more far-reaching than simply the herd reacting to volatility.
The conventional wisdom is that in the midst of economic upheavals—the bursting housing bubble, Lehman, the Great Recession, Europe—investors make similar moves, like fleeing to safety of US T-bills. And when that happens, the correlation between asset prices rises. In such markets, diversification provides less insulation from market swings: once everything moves together, there’s no place to hide and this amplifies large market swings, volatility, and risk.
Makes sense, and while the figure below reveals a fairly strong up-and-down cycle of the correlation between stocks, it’s a cycle around a rising trend:
The rise in correlation between individual stocks, but also between completely separate asset classes like stocks and gold or stocks and oil, “has been one of the big themes of the investment climate this year,” said Marc Chandler, a market strategist at Brown Brothers Harriman in New York.
Source: NYT, see link above.
But it seems to me that similar correlations have shown up in some unexpected places, and the problems this engenders may thus go deeper than a temporary response to volatile economics.
For example, consider the securitization of mortgage debt during the inflating of the housing bubble. The argument—I recall Greenspan going on about this—was that such securitization was a great way to diversify and thus diminish risk. The high risk inherent in the subprime tranche of the MBS would be offset by the low risk in the higher, AAA-rated tranches.
But in too many cases, the opposite occurred: the tranches were correlated, and when one defaulted, the others did too. Part of this was faulty evaluation by the credit rating agencies, but there’s something else in play here: interconnectedness stoked by information.
The cost of information has fallen sharply and its velocity has massively increased. What investors know is itself more highly correlated now than in the past. Yes, there are still arbitrage opportunities, but they are the stuff of flash trades tapping micro-second price differences. It’s not what I know or my analysis of a firm or a country’s economic conditions. It’s my quant’s algorithm against your quant’s algorithm.
I’m just theorizing, but if I’m even partially right—if globalization and the democratization of market information has raised the structural (i.e., not just the cyclical) correlation between asset prices—that explains a) the increasing tendency to underprice risk, b) the increased volatility of markets, as the herd lurches from flights to the safety of cash and bonds (Merkel frowned!) back to equities (wait…she smiled!), and even the “shampoo cycle”—bubble, bust, repeat—that’s defined the pattern of US economic growth in recent years.