This man is the most brilliant mind in economics today. God bless him!
Investors who believe that history has lessons to
teach should take our present concerns with significant weight, but
should also recognize that tendencies that repeatedly prove reliable
over complete market cycles are sometimes defied over portions of those
cycles. Meanwhile, investors who are convinced that this time is
different can ignore what follows. The primary reason not to listen to a
word of it is that similar concerns, particularly since late-2011,
have been followed by yet further market gains. If one places full
weight on this recent period, and no weight on history, it follows that
stocks can only advance forever.
What seems different this time, enough to revive
the conclusion that “this time is different,” is faith in the Federal
Reserve’s policy of quantitative easing. Though quantitative easing has
no mechanistic relationship to stock prices except to make
low-risk assets psychologically uncomfortable to hold, investors place
far more certainty in the effectiveness of QE than can be demonstrated
by either theory or evidence. The argument essentially reduces to a
claim that QE makes stocks go up because “it just does.” We doubt that
the perception that an easy Fed can hold stock prices up will be any
more durable in the next couple of years than it was in the 2000-2002
decline or the 2007-2009 decline – both periods of persistent and
aggressive Fed easing. But QE is novel, and like the internet bubble,
novelty feeds imagination. Most of what investors believe about QE is
imaginative.
As Ray Dalio of Bridgwater recently observed, “The
dilemma the Fed faces now is that the tools currently at its disposal
are pretty much used up. We think the question around the effectiveness
of QE (and not the tapering, which gets all the headlines) is the big
deal. In other words, we’re not worried about whether the Fed is going
to hit or release the gas pedal, we’re worried about whether there’s
much gas left in the tank and what will happen if there isn’t.”
While we can make our case on the basis of fact,
theory, data, history, and sometimes just basic arithmetic, what we
can’t do – and haven’t done well – is to disabuse perceptions. Beliefs
are what they are, and are only as malleable as the minds that hold
them. Like the nearly religious belief in the technology bubble, the
dot-com boom, the housing bubble, and countless other bubbles across
history, people are going to believe what they believe here until
reality catches up in the most unpleasant way. The resilience of the
market late in a bubble is part of the reason investors keep holding
and hoping all the way down. In this market cycle, as in all market
cycles, few investors will be able to unload their holdings to the last
of the greater fools just after the market’s peak. Instead, most
investors will hold all the way down, because even the initial decline
will provoke the question “how much lower could it go?” It has always
been that way.
The problem with bubbles is that they force one to
decide whether to look like an idiot before the peak, or an idiot
after the peak. There’s no calling the top, and most of the signals
that have been most historically useful for that purpose have been
blaring red since late-2011.
As a result, the Shiller P/E (the S&P 500
divided by the 10-year average of inflation-adjusted earnings) is now
above 25, a level that prior to the late-1990’s bubble was seen only in
the three weeks prior to the 1929 peak. Meanwhile, the price/revenue
ratio of the S&P 500 is now double its pre-bubble norm, as is the
ratio of stock market capitalization to GDP. Indeed, the median
price/revenue ratio of the S&P 500 is actually above the 2000 peak –
largely because small cap stocks were much more reasonably priced in
2000 than they are today (not that those better relative valuations
prevented wicked losses in small caps during the 2000-2002 decline).
Despite the unusually extended period of
speculation as a result of faith in quantitative easing, I continue to
believe that normal historical regularities will exert themselves with a
vengeance over the completion of this market cycle. Importantly, the
market has now re-established the most hostile overvalued, overbought,
overbullish syndrome we identify. Outside of 2013, we’ve observed this
syndrome at only 6 other points in history: August 1929 (followed by
the 85% market decline of the Great Depression), November 1972 (followed
by a market plunge in excess of 50%), August 1987 (followed by a
market crash in excess of 30%), March 2000 (followed by a market plunge
in excess of 50%), May 2007 (followed by a market plunge in excess of
50%), and January 2011 (followed by a market decline limited to just
under 20% as a result of central bank intervention).
These concerns are easily ignored since we also observed them at lower levels this year, both in February (see A Reluctant Bear’s Guide to the Universe)
and in May. Still, the fact is that this syndrome of overvalued,
overbought, overbullish, rising-yield conditions has emerged near the
most significant market peaks – and preceded the most severe market
declines – in history:
1. S&P 500 Index overvalued, with the Shiller
P/E (S&P 500 divided by the 10-year average of inflation-adjusted
earnings) greater than 18. The present multiple is actually 25.
2. S&P 500 Index overbought, with the index
more than 7% above its 52-week smoothing, at least 50% above its 4-year
low, and within 3% of its upper Bollinger bands (2 standard deviations
above the 20-period moving average) at daily, weekly, and monthly
resolutions. Presently, the S&P 500 is either at or slightly
through each of those bands.
3. Investor sentiment overbullish (Investors Intelligence),
with the 2-week average of advisory bulls greater than 52% and
bearishness below 28%. The most recent weekly figures were 55.2% vs.
15.6%. The sentiment figures we use for 1929 are imputed using the
extent and volatility of prior market movements, which explains a
significant amount of variation in investor sentiment over time.
4. Yields rising, with the 10-year Treasury yield higher than 6 months earlier.
The blue bars in the chart below depict the complete set of instances since 1970 when these conditions have been observed.
Our investment approach remains to align our
investment outlook with the prospective market return/risk profile that
we estimate on the basis of prevailing conditions at each point in
time. On that basis, the outlook is hard-defensive, and any other
stance is essentially speculative. Such speculation is fine with
insignificant risk-limited positions (such as call options), but I
strongly believe that investors with a horizon of less than 5-7 years
should limit their exposure to equities. At this horizon, even
“buy-and-hold” strategies in stocks are inappropriate except for a
small fraction of assets. In general, the appropriate rule for setting
investment exposure for passive investors is to align the duration of
the asset portfolio with the duration of expected liabilities. At a 2%
dividend yield on the S&P 500, equities are effectively instruments
with 50-year duration. That means that even stock holdings amounting
to 10% of assets exhaust a 5-year duration. For most investors, a
material exposure to equities requires a very long investment horizon
and a wholly passive view about market prospects.
Again, our approach is to align our outlook with
the prospective return/risk profile we estimate at each point in time.
That places us in a defensive stance. Still, we’re quite aware of the
tendency for investors to capitulate to seemingly relentless
speculation at the very peak of bull markets, and saw it happen in 2000
and 2007 despite our arguments for caution.
As something of an inoculation against this
tendency, the chart below presents what we estimate as the most
“optimistic” pre-crash scenario for stocks. Though I don’t believe that
markets follow math, it’s striking how closely market action in recent
years has followed a “log-periodic bubble” as described by Didier
Sornette (see Increasingly Immediate Impulses to Buy the Dip).
A log periodic pattern is essentially one where
troughs occur at increasingly frequent and increasingly shallow
intervals. As Sornette has demonstrated across numerous bubbles over
history in a broad variety of asset classes, adjacent troughs (say T1,
T2, T3, etc) are often related to the crash date (the “finite-time
singularity” Tc) by a constant ratio: (Tc-T1)/(Tc-T2) = (Tc-T2)/(Tc-T3)
and so forth, with the result that successive troughs come closer and
closer in time until the final blowoff occurs.
Frankly, I thought that this pattern was nearly
exhausted in April or May of this year. But here we are. What’s
important here is that the only way to extend that finite-time
singularity is for the advance to become even more vertical and for
periodic fluctuations to become even more closely spaced. That’s
exactly what has happened, and the fidelity to the log-periodic pattern
is almost creepy. At this point, the only way to extend the singularity
beyond the present date is to envision a nearly vertical pre-crash
blowoff.
So let’s do that. Not because we should expect it, and surely not because we should rely
on it, but because we should guard against it by envisioning the most
“optimistic” (and equivalently, the worst case) scenario. So with the
essential caveat that we should neither expect, rely or be shocked by a
further blowoff, the following chart depicts the market action that
would be consistent with a Sornette bubble with the latest “finite time
singularity” that is consistent with market action since 2010.
To be very clear: conditions already allow a
finite-time singularity at present, the scenario depicted above is the
most extreme case, it should not be expected or relied on, but we
should also not be shocked or dismayed if it occurs.
Just a final note, which may or may not prove
relevant in the weeks ahead: in August 2008, just before the market
collapsed (see Nervous Bunny),
I noted that increasing volatility of the market at 10-minute
intervals was one of the more ominous features of market action. This
sort of accelerating volatility at micro-intervals is closely related
to log-periodicity, and occurs in a variety of contexts where there’s a
“phase transition” from one state to another. Spin a quarter on the
table and watch it closely. You’ll notice that between the point where
it spins smoothly and the point it falls flat, it will start vibrating
uncontrollably at increasingly rapid frequency. That’s a phase
transition. Again, I don’t really believe that markets follow math to
any great degree, but there are enough historical examples of
log-periodic behavior and phase-transitions in market action that it
helps to recognize these regularities when they emerge.
Risk dominates. Hold tight.
