Yesterday we profiled correlations; today, we update volatility.
In preview, vol is up, which is to say that it's no longer down, as it was for several years previous. The bear market in volatility ended in late 2006/early 2007, as our chart below reminds. As it happened, the windup in vol's decline preceded the start of the bear market in equities. If you're thinking that volatility's nadir was a clue of things to come, you're right. In fact, we considered no less in the recent past, including this post from January 2007, when we advised, after a long spell of falling standard deviations: "Higher volatility is probably coming, one day, and history reminds that sometimes higher volatility is forged by falling prices."
But that was then. What of the future? As always, we can only guess. Fortunately, we can also learn from history. No, not everything is revealed by looking backward, and certainly not with full clarity about what's coming. Still, Mr. Market leaves a few crumbs of insight about the morrow.
On that front, volatility at the moment has little to tell that we don't already know. Yes, standard deviations in the major asset classes are rising, as we suspected they would after reaching unsustainable lows in late 2006/early 2007. Based on the above chart, one may expect that vol still has a ways to go, which suggests that the selling isn't about to dry up just yet.
The broader point is that standard deviations, like much else in finance and business, are subject to cycles. Alas, the exact nature and timing of the cycles are fully transparent only in hindsight. But don't despair: vol cycles aren't completely obscure in real time either. What's more, the primary value of this metric--other than considering the historical context for building informed portfolios--is taking the hint when standard deviations are at extremes, high or low.
Yes, there'll always be a debate about whether the current levels are at extremes, or not. But that's where history comes in handy, if only to provide some guidelines. We can start from the premise that vol won't fall to zero. Stepping out a bit on the speculative path from there we can say that a precipitous drop in vol--particularly if it's been several years in the making--will run out of momentum. When it does, the shift in trend is a warning signal that higher volatility may be coming, which almost surely means falling prices.
Generally, falling vol is a byproduct of rising prices; meanwhile, rising vol comes by way of falling prices. This knowledge/assumption, combined with an understanding of the mechanics and implications of correlations and valuation (dividend yields, p/e ratios, etc.) adds up to a powerful tool kit for coming up with an informed guess about what's coming and building diversified portfolios that have some chance of meeting expectations, or at least not disappointing too extensively.
Why go to all this trouble? One reason is that prices are almost impossible to forecast directly, at least in the short term, which we define as something less than two years. Volatility, correlations, and other non-return measures are a bit more amenable to analytical predictions, which open the door to price predictions by reverse engineering through the other metrics.
Does it work? Sometimes, although one can't be sure until the future arrives. In short, there's still a fair amount of risk embedded in even the most enlightened investment strategies. No surprise, since the presence of risk implies that there's also a risk premium lurking about.
Yes, Virginia, there's an order in all the chaos, and so some of Mr. Market's patterns may be visible to the naked eye. Sometimes.