I’m Gonna Finish this White Paper
Something I first read a long time ago, but that took a while to understand and appreciate was this special feature from the June 2000 Bank for International Settlements Quarterly Review. Called “Evaluating changes in correlations during periods of high market volatility,” it makes the case, using math, that when things become more volatile, they also become more correlated.
Now, if you’re among the legions of people more intuitively numerate than I am, you might be saying, “No duh.” But for me this was not intuitive. That’s because it seems like one way to think about volatility would be as difference. For example, if lots of things are zigging and zagging all the time, it seems like it should be the case that all of that zigging and zagging would be more random and therefore less correlated.
But another (and the right) way to think about volatility is as magnitude. In other words, something that’s volatile isn’t necessarily zigging and zagging a lot, but rather has the potential for explosive zigs and catastrophic zags. And as the math in that paper shows, while there can be a lot of relative difference between a lot of events in a clustered distribution, extreme events are fewer and a lot more alike relatively speaking precisely because they’re extreme and so they are much more highly correlated.
This is a problem for business and investing because Modern Portfolio Theory posits that you can earn better returns with less risk by doing things that are uncorrelated. But if it’s true that things become more correlated when they start going really badly, then the benefits of being diversified are reduced at precisely the time you need them the most.
And this is a reality that has proved to be true during significant market dislocations.
This National Bureau of Economic Research working paper, “Long-Term Global Market Correlations,” was published in November 2001, just as the bursting of the tech bubble and 9/11 terrorist attacks were wreaking havoc on global markets. It remarked that “the diversification benefits to global investing are not constant,” that “diversification potential today is very low compared to the rest of capital market history,” and that “periods of poor market performance, most notably the Great Depression, were associated with high correlations, rather than low correlations.” Global correlations increased again with the 2009 financial crisis, and during Covid everything went way down then way up together.
In other words, it’s clear that when people panic, they’re likely to panic about everything all at once.
And while there is less academic research into how this propensity manifests itself in small operating businesses, I can tell you from experience that when something starts to go poorly at a small operating business, a lot more is likely to go poorly alongside it. A business that loses a bunch of money on a project, for example, might see performance suffer at other projects as it repurposes its best people to deal with the crisis. Then its controller, stressed out by the situation, might decide to retire a year earlier than expected, leaving a massive hole on the leadership team right when accurate numbers are most important. And seeing deteriorating prospects and a lack of appetite for growth, a star salesperson might then leave for a competitor. Heck, it’s when morale is low that office supplies even start to disappear.
When it rains, as they say, it pours.
So if the lesson is that the world is more correlated than we think, an implication is that we can’t engineer our way to stability when constructing an investment portfolio or building a business. Instead, we have to accept that there will be instability and that when it happens, there will be a lot of this at once. The two ways to handle this are (1) temperament, i.e., know ahead of time what you’ll need to do to keep your cool and (2) structure, i.e., your agreements with others and liquidity requirements can never turn you into a forced actor.
-Tim
