Hussman analytically describing the process leading to a crash. It hasn’t happened yet, so it won’t happen. Right?
From an investment standpoint, market conditions remain characterized both by obscene valuations and still-negative market internals. It’s that combination that continues to suggest potentially vertical downside risks.
When people think about crashes, they tend to think about an event – as if some massive, grotesque, red, scaly, fire-breathing, razor-toothed catalyst should be obvious beforehand. But we know from history that that’s not the way it works.
Instead, the sequence goes like this: the conditions that create vulnerability to a crash emerge first (elevated valuations coupled with deterioration in market internals and/or widening in credit spreads), the crash emerges second, and catalysts are then identified – often just flashpoints that were consistent with speculative breakdown. Many investors think that “Lehman” caused the global financial crisis, but the mortgage crisis was already unfolding well before that. Lehman and Bear Stearns before it were only symptoms, not causes.
The cause is always speculative distortion that was well-known for quite some time: elevated valuations, often accompanied by speculation and new issues of low-quality stocks representing some “new economy” theme, or yield-seeking speculation and heavy issuance of low quality debt. The main reason investors don’t believe that such speculation will end in a crash is simply that a crash hasn’t happened yet.
In shorter-term market action, we see a general tendency toward distribution, for example, declines on expanding volume coupled with low-volume recoveries on mixed breadth and narrowing leadership (which was also the pattern last week). We note that prior to Friday, the Dow Jones Industrial Average had gone 40 trading sessions without setting a 20-day high or low.
If we look across history for periods of extended range-bound activity in overvalued markets where: a) the DJIA had gone more than a month without setting a 20-day high or low; b) the DJIA was confined to a range of less than 6%; c) the DJIA was within 10% of a 2-year high; and d) the Shiller P/E was 18 or higher, there are only 7 clusters that fit the bill (1929, 1937, 1965, 1973, 1999-2000, 2007-2008, and today).
While the full-cycle resolution was repeatedly brutal, I should note that the short-term resolution was not very informative at all, and didn’t depend on whether the initial break out of the range was higher or lower. What’s interesting about the general pattern is that range-bound action often coincides with distribution on price-volume measures. In October 2000, I mentioned similar measures of distribution, such as one that Peter Eliades called the “sign of the bear” based on range-bound market breadth. As I also observed then, neither elevated valuations, nor internals, nor distribution patterns ensure a market crash, but we don’t like the probabilities. Still, we aren’t terribly impatient about the near-term resolution of what we see as a likely topping process here.
You’ve been warned.
Read Hussman’s Weekly Letter