John Hussman: Formula For Market Extremes

Includes: DIA, IWM, QQQ, SPY
by: John Hussman

Excerpt from the Hussman Funds' Weekly Market Comment (6/16/14):

Market extremes generally share a common formula. One part reality is blended with one part misguided perception (typically extrapolating recent trends as if they are driven by some reliable and permanent mechanism), and often one part pure delusion (typically in the form of a colorful hallucination with elves, gnomes and dancing mushrooms all singing in harmony that reliable valuation measures no longer matter).

The technology bubble was grounded in legitimate realities including the emergence of the internet and a Great Moderation of stable GDP growth and contained inflation. But it also created a misperception that it was possible for an industry to achieve profits while having zero barriers to entry at the same time (the end of that misperception is why the dot-com bubble collapsed), a misperception that technology earnings would grow exponentially and were not cyclical (as we correctly argued in 2000 they would shortly prove to be), and the outright delusion that historically reliable valuation measures were no longer informative. Meanwhile, the same valuation measures we use today were projecting – in real time – negative 10-year nominal total returns for the S&P 500 over the coming decade, even under optimistic assumptions (see our August 2000 research letter).

The housing bubble was grounded in legitimate realities including a boom in residential housing construction and a legitimate economic recovery that followed the 2000-2002 bear market (which we responded to by shifting to a constructive stance in April 2003 despite valuations still being elevated on a historical basis). But the housing bubble also created a misperception that mortgage-backed securities were safe because housing prices had, at least to that point, never experienced a major collapse. The delusion was that housing was a sound investment at any price. The same delusion spread to the equity markets, helped by Fed-induced yield-seeking speculation by investors who were starved for safe return. Meanwhile, the same equity valuation measures we use today helped us to correctly warn investors of oncoming financial risks at the 2007 peak (see A Who’s Who of Awful Times to Invest).


As usual, the delusion is that this Fed-induced reach for yield is enough to make equities a sound investment at any price. Ironically, the Fed itself is in the process of reversing course on this policy. Meanwhile, based on the same valuation methods that correctly projected negative 10-year returns at the 2000 peak, correctly gave us room to shift to a constructive position in early 2003, correctly helped us warn of severe market losses at the 2007 peak, and correctly identified the shift to market undervaluation in late-2008 (which those who don’t understand our stress-testing narrative may not recognize), we currently estimate prospective S&P 500 nominal total returns of just 2% annually over the coming decade, and negative returns on every horizon shorter than about 7 years.

In short, investors who are reaching for yield in stocks as an alternative to risk-free assets are most likely reaching for a negative total return in stocks between now and about 2021.

Lessons from the recent half-cycle

The half-cycle since 2009 has certainly challenged our reputation in recent years, but it has not challenged our confidence about the future a bit. To understand that, it may be helpful to look under the hood at what has been different about this period. I’ve discussed the impact of my 2009 stress-testing decision in prior comments (see Setting the Record Straight), but I’m not sure that I’ve been clear about what has legitimately been “different” about the recent half-cycle compared with historical market cycles, and how we’ve adapted to that.

The chart below may be helpful in that regard. Each line shows the cumulative total return of the S&P 500 Index under a variety of conditions we identify. The chart is presented on log-scale. Moving higher by one horizontal line represents a doubling. Moving lower by one horizontal line represents a 50% loss. I presented a similar chart last September (see The Lesson of the Coming Decade), but log-scale gives a much more obvious picture of what’s been so challenging in the past few years.


During the late-1990’s, observing that stock prices were persistently advancing despite historically overvalued conditions, we introduced a set of “overlays” that restricted our defensive response to overvalued conditions, provided that certain observable supports were present. These generally related to features of market action that I called “trend uniformity.” Though the ensemble approach we introduced in 2010 was more effective in classifying market return/risk profiles than every other method we’ve examined, similar overlays were required to restrict the response to overvalued, overbought, overbullish conditions.

In identifying proper restrictions, the best practice is that a) they should not contradict historical evidence, and b) they should testably result in improved discrimination of subsequent market return and risk, not only in the most recent cycle, but in prior historical cycles and in “holdout” validation data not used in the design of those restrictions.

For example, imposing a restriction that amounts to “Don’t fight the Fed” contradicts historical evidence because there are too many periods across history, most recently the 2000-2002 and 2007-2009 plunges, when the Federal Reserve was persistently and aggressively easing policy but stocks experienced profound losses anyway (see Following the Fed to 50% Flops). Imposing a restriction that amounts to “Don’t fight an expanding Fed balance sheet” also contradicts the evidence. A restriction that amounts to “Don’t fight quantitative easing” is arbitrary because it can’t be tested in historical data prior the recent cycle, and so on.

As it happens, the most reliable and testable restrictions have a great deal to do with market internals, price behavior, and interest rates, but they are far less general than investors seem to believe. Aphorisms about not fighting the tape or not fighting the Fed aren’t really bad advice, particularly in combination. But they are certainly weaker advice over historical market cycles than investors seem to imagine.