March 19, 2012
An Angry Army of Aunt Minnies
.
John P. Hussman, Ph.D.
As of Friday, the S&P 500 was within 1% of its upper Bollinger band at virtually every horizon, including daily, weekly and monthly bands. The last time the S&P 500 reached a similar extreme was Friday April 29, 2011, when I titled the following Monday's comment Extreme Conditions and Typical Outcomes . I observed when the market has previously been overbought to this extent, coupled with more general features of an "overvalued, overbought, overbullish, rising yields syndrome", the average outcome has been particularly hostile:
.
"Examining this set of instances, it's clear that overvalued,
overbought, overbullish, rising-yields syndromes as extreme as we observe today
are even more important for their extended implications than they are for market
prospects over say, 3-6 months. Though there is a tendency toward abrupt market
plunges, the initial market losses in 1972 and 2007 were recovered over a period
of several months before second signal emerged, followed by a major market
decline. Despite the variability in short-term outcomes, and even the tendency
for the market to advance by several percent after the syndrome emerges, the
overall implications are clearly negative on the basis of average return/risk
outcomes."
.
As it happened, April 29, 2011 turned out to mark the exact
high of the S&P 500 for the year, and was followed by a steep intermediate
market plunge. My impression is that despite the recent run of speculation the
market has enjoyed - largely reflecting a reprieve in European debt concerns and
what appears to be a drawing-forward of jobs into the first quarter due to
unseasonably favorable weather - the extended implications of present market
conditions remain decidedly negative.
.
If you examine the components of the S&P 500
individually, you'll quickly find that the majority of those stocks are also at
or through their own upper Bollinger bands. In overvalued, overbought,
overbullish, rising-yield conditions, those extensions are often resolved in
unison, which is what produces the characteristic "air pocket" where the index
can give up weeks or sometimes months of upside progress in a handful of
sessions (though we often see a knee-jerk reaction to buy that initial dip
before more serious follow-through occurs).
.
In recent weeks, the market has faced what I've called an
"angry army of Aunt Minnies" (see for example Warning: A New Who's Who of
Awful Times to Invest and Goat Rodeo), These are
indicator syndromes that are generally (and sometimes singularly) followed by
steeply negative market outcomes. When I present these in the weekly comments,
I'll often give specific thresholds (such as a Shiller PE of 18, 27% bearish
sentiment, and so forth) in order to give an idea of where the "border" of a
given cluster of data tends to be, but these aren't magic numbers. The objective
is to capture a syndrome of conditions that is characteristic of some
particular diagnosis, but the outcomes are generally robust to small variations
in how they are defined. For example, with as few bears as we have at present,
there's little distinction between say, 45% bulls and 44% bulls. As I noted last
May:
.
"We can define an 'overvalued, overbought, overbullish,
rising-yields syndrome' a number of ways. The more general the criteria, the
better you capture historical instances that preceded abrupt market weakness,
but the more you also encounter 'false positives.' Still, as long as the
criteria capture the basic syndrome, we find that the average return/risk
profile for subsequent market performance is negative, almost regardless of the
subset of history you inspect."
.
The steepest market plunges on record (e.g. those following
the 1973-74, 1987, 2000 and 2007 peaks, among others) have generally followed an
overvalued speculative blowoff coupled with divergent interest rate pressures.
This is why we take the "overvalued, overbought, overbullish, rising yields"
syndrome so seriously. Indeed, the outcomes are usually negative on average even
without rising yields, but the yield pressures tend to add immediacy. Notably,
the emergence of this syndrome has provided accurate warning of on coming losses
both historically, and also as recently as 2010 and 2011.
.
These syndromes are useful because the combination
of several conditions often carries far more information than any of them
individually. To ignore syndromes like the ones we've increasingly observed in
recent weeks is like having nausea, sudden lower-right abdominal pain -
especially following a period of dull pain in the upper abdomen, coupled with an
inability to even consider jumping, and shrugging it off as a stomach
ache instead of recognizing that, in all likelihood, your appendix has just
burst.
.
Two concepts of risk - market risk and tracking
risk
.
Our approach has always been to accept the risk of market
fluctuations in proportion to the average return that we can expect per
unit of risk, given prevailing market and economic conditions. So we are willing
to take a higher exposure to market risk when the prospective return/risk
profile is favorable (as it was in 2003 for example). In contrast, we limit our
exposure when the profile is weak, and hedge very tightly against market risk
when we are confronted with classic signs that have almost invariably preceded
major market setbacks (such as the "overvalued, overbought, overbullish,
rising-yields" syndrome we observe today). That word "average" is important -
our exposure to market risk tends to be proportional to the average
return that we expect from that risk, given market conditions. From a
full-cycle investment perspective, this is actually optimal strategy.
From a psychological perspective, however, it may
not be. Specifically, we've had a number of conversations with shareholders of
the following form: "I know that the indicators you're looking at are negative,
and I do agree that eventually we'll see a bear market that wipes out a lot of
these market gains, but here and now, the market is going up. Can't you take a
larger exposure to market risk - not huge - but enough to participate a little
bit until things turn down?" The basic answer is, yes we could, but we know from
historical evidence that we would not actually expect to add to our net
returns over time.
.
Instead, we might gain some amount in the event that the
market was to advance further, but we would also expect to lose that amount
without any predictable expectation of "locking it in." This is because we would
be taking a positive position despite negative expected returns, so there would
then be no trigger to sell except at the point where a new downtrend
was already established, most likely at similar or lower prices, and very
possibly on a large and abrupt decline where good price execution would be
difficult.
Still, I recognize that this sort of strategy might be
psychologically preferable to some investors - despite not adding
anything to long-term returns. Participating some amount during an advance might
simply feel better, even if there is no expectation of actually
retaining that short-term gain. This is an important issue because it highlights
the difference between two kinds of risk - one being market risk itself, and one
being tracking risk (the difference between one's own performance and
the performance of the market).
Our own investment strategy places very little weight on
tracking risk, because we are focused on the complete bull-bear market cycle.
Likewise, we don't particularly care whether we take our drawdowns during
periods when the market is rising or when it is falling, provided that those
drawdowns are relatively muted over the full cycle (particularly compared with
the losses of over 50% that the stock market has periodically experienced).
Indeed, from a diversification perspective, it's better for us to take our
drawdowns when the major indices are not, and vice versa. Accordingly, if a
given amount of risk isn't expected - on average - to have a positive return,
given prevailing market conditions, we simply don't take it. Yet it's clear,
especially at times like these, how uncomfortable it can be psychologically to
be defensive during an exuberant rally.
.
While I am convinced - even adamant - that the present time
would be the wrong moment to make any shift toward greater market risk, I do
understand the psychological discomfort that tracking risk can create. Reducing
tracking risk involves a tradeoff - you tend to move somewhat more in line with
the market, but you also tend to experience more frequent drawdowns. It's very
difficult to exit with further gains once the average return/risk profile goes
negative - even if various trend-following measures are still positive. We're
open to refinements in our approach that might reduce tracking risk somewhat -
provided that we don't materially increase large drawdowns, or reduce expected
long-term returns in the process. But our main objective is always focused on
investment returns over the complete market cycle (bull market plus bear market)
with significantly smaller periodic losses than a buy-and-hold approach. Within
the limits of that objective, we'll work to find ways to diminish the periodic
discomfort we ask our shareholders to accept. Again however, I am convinced that
now would be precisely the wrong time to give in to that discomfort by accepting
greater market exposure.
On the subject of investment performance over the complete
market cycle (bull peak to bull peak, bear trough to bear trough), a few
observations may be useful. In the Strategic Growth Fund , we view the period
from the 2000 market peak to the 2007 market peak (peak-to-peak), and the period
from the 2002 market trough to the 2008-2009 market trough (trough-to-trough),
as fairly representative of typical full-cycle application of our investment
approach. In these complete market cycles, the Fund widely outperformed the
S&P 500, with far smaller periodic losses.
As we've discussed in previous weekly comments, annual
reports, and briefly below, the 2009-early 2010 period was notably
unrepresentative of our typical investment strategy. Unfortunately, as
a result of that period, the total return of the Fund from the 2007 market peak
(10/9/07) through Friday's peak is - at least temporarily - a cumulative 11.6%
behind the S&P 500, though we've experienced much less volatility and far
smaller intervening losses.
The simple fact is this. Despite anticipating much of the
financial crisis and the accompanying market losses, I misjudged the likely
policy response, expecting what I called a "writeoff recession" where reckless
lenders would be expected to take losses, rather than a "kick the can" approach
- defending bondholders at public expense, changing accounting rules, and
mounting massive central bank interventions, all of which promise to create
recurring crises for years to come (see the Swedish banking crisis in the 1990's
for an example of a proper and durable response). As the crisis deepened, I
insisted on stress-testing our hedging approach, with the requirement that it
should perform well in validation data from both post-war and Depression-era
periods. There was simply no way to "average in" Depression-era evidence without
getting consistently negative expected return/risk estimates.
.
Indeed, during the
Depression, the same valuations that we saw at the 2009 lows were
followed by a further loss of two-thirds of the market's value. Until I
was certain that our methods were robust to both data sets, we remained hedged.
We ultimately tackled that "two data sets" problem with
ensemble methods, which capture distinctions that we were frankly unable to make
without them. We were very open about the challenges we faced during 2009 and
early-2010 in solving that problem, because the need to properly integrate
Depression-era data forced a defensive stance that was not representative of our
"typical" investment strategy. In short, the 2009 - early 2010 period is
distinct in that it is not indicative of the hedge position that can be expected
of our strategy in future market cycles, even under identical conditions and
evidence.
These considerations may go some way in understanding why the
Strategic Growth Fund remains my largest single investment holding (with nearly
all of my remaining liquid assets invested in the other three Hussman Funds).
The S&P 500 has underperformed Treasury bills for what is
now a 13 year period, and we certainly expect that future cycles will feature
more historically normal valuations and prospective market returns. In that sort
of environment, we would expect to take far more aggressive investment positions
than we've been able to take since 2000. That said, our investment actions from
2000 through 2008, as well as our actions since early 2010, are representative
of what we could be expected to do again under identical conditions and
evidence.
.
Frankly, this includes the fact that our investment
stance has been largely defensive since early 2010. In that regard, it's notable
that the S&P 500 gained no net ground from April 2010 through November of
2011. Of course, the market has advanced significantly in the past few months.
Our defensiveness in these recent months has been driven by overlapping
syndromes of historically negative evidence, with increasingly extreme warning
flags. The day-to-day discomfort that we experience here - being defensive in an
overvalued, overbought, overbullish advance where various warning flags abound -
may be frustrating, but is fairly typical of what our investment strategy will
experience in similar conditions, as we did in 2000, 2007, and prior to the less
severe 2010 and 2011 declines.
Undoubtedly, our recent defensiveness would feel less
uncomfortable had I realized that we would be forced to contemplate
Depression-era data, and had our current ensemble methods been in place to
capture significant gains during that 2009 - early 2010 period. It's clearly my
job to minimize or eliminate the impact of that "miss" over the present cycle.
Again, however, I remain adamant that now would be one of the worst possible
times to accept market risk in hopes of "catching up" - when valuations are rich
instead of reasonable, prices are strenuously overbought instead of oversold,
investor sentiment is exuberant rather than fearful, the growing technical
divergences are negative rather than positive, corporate insiders are
frantically selling instead of buying, stocks are grasping at speculative highs
instead of multi-year lows, and risk premiums are razor-thin instead of
satisfactory.
Economic Notes
.
On the economy, our broad view is based on dozens of
indicators and multiple methods, and the overall picture is much better
described as a modest rebound within still-fragile conditions, rather than a
recovery or a clear expansion. The optimism of the economic consensus seems to
largely reflect an over-extrapolation of weather-induced boosts to coincident
and lagging economic indicators -- particularly jobs data. Recall that seasonal
adjustments in the winter months presume significant layoffs in the retail
sector and slow hiring elsewhere, and therefore add back "phantom" jobs to
compensate. While my remarks on seasonal adjustment have been co-opted by a few
conservative blogs, I don't view these adjustments as manipulative. It's just
important to understand their impact. It seems likely that particularly in
retail and construction, favorable weather has brought forward hiring that would
normally occur in the spring months, and the seasonal adjustment factors have
added to those anyway. The question - still unresolved - is whether the jobs
brought forward from the spring months will take anything out of the job
creation numbers we see in the months ahead, particularly in the April-May time
frame.
There's slightly more dispersion in various economic Aunt
Minnies we track, but the overall picture is mixed at best. On the side of
concern is a syndrome that uses the standardized values (zero mean, unit
variance) of the OECD US, OECD Total, and ECRI Weekly leading indices. A strong
discriminator of actual recessions and recoveries can be defined as follows:
When all three are below -0.4, and the average is less than -1, the economy has
always been approaching or in recession.
.
Conversely, when at least two of these
three improve to -0.2 or better, the economy has always been entering a
recovery. It's possible to vary the criteria slightly (e.g. 0.1 tighter or
looser), but that seems to invite some false positives or false negatives,
particularly due to data revisions. This is still only one measure, and as such
should be used as a rule of thumb in the context of dozens of other
indicators (which is how we use it in practice). Still, it's an instructive
example of the coordinated indicator movements we try to capture with
Aunt Minnies.
.
The OECD leading indices were released again last week, as was the
latest reading in the ECRI index. The standardized values of all three remain
below -0.2. That may change in the weeks ahead, but we have not seen it yet. An
improvement on this Aunt Minnie would be helpful in alleviating the concern we
presently have about leading indicators, particularly if we see some strength in
real consumer spending as well.
.
Here's what this particular Aunt Minnie looks like (shown by
the regions outlined in red). Actual US recessions are shaded blue. The most
recent data would not trigger this syndrome today, but it does not relieve the
existing signal either. Our best interpretation is that the leading data is
mixed - not deteriorating at present, but certainly not "all clear" yet.
.
.
On the brighter side, the recent speculative run does give us
fewer recession warning signs from financial market variables like stock prices
and credit spreads. Recessions tend to be associated with stock price weakness,
widening credit spreads, a reasonably flat yield curve, and fairly tepid ISM
readings, among other factors. Generally speaking, we use a 6-month look-back
(again, not a magic number, but simply a way to make the syndrome operational).
Presently, credit spreads would have to widen by about 40 basis points and the
S&P 500 would have to decline by about 15% in order to re-establish this
particular syndrome of recession warning conditions, which we observed last
summer. Given the unusually extreme stock market conditions and narrow risk
premiums here, I certainly wouldn't rule out fresh deterioration in the S&P
500 and corporate debt. But for now, suffice it to say that despite continued
concerns from leading economic measures, financial variables would have to
deteriorate in order to confirm those concerns more unanimously.
Market Climate
As of last week, the Market Climate for stocks remained
characterized by an unusually hostile set of overvalued, overbought,
overbullish, rising-yield conditions. The record of steeply negative market
outcomes that have followed these conditions has nothing to do with my
opinion but instead reflects objective historical evidence. The outcome
in this particular instance may be different, and we have no problem
with investors who are willing to invest on that expectation. We are not. It's
that simple.
At present, we estimate the likely total return for the
S&P 500 over the coming decade to be about 4.1% annually (nominal). While
this may seem adequate compared to a 10-year Treasury yield of 2.3%, the
comparison entirely ignores risk. You don't just "lock in" a 10-year return -
you ride it out. The volatility of stocks is dramatically higher than the
volatility of a 10-year bond. So the proper question is not whether stocks are
priced to achieve a greater 10-year return than bonds, but what happens if
investors eventually demand even modestly higher prospective returns.
The answer is that the impact of changes in valuation multiples would swamp any
reasonable expectation for growth in fundamentals. Satisfactory returns on
stocks now require strong assumptions for GDP growth (about 6% nominal
annual growth) and sustained profit margins (presently over 60% above the
historical norm), in addition to the requirement that valuations remain rich and
prospective returns stay indefinitely depressed. This may occur over the short
run, but beyond that it strikes us as pure speculation.
.
.
Strategic Growth Fund and Strategic International Fund remain
fully hedged here. About 50% of the holdings in Strategic Dividend Value are
presently hedged - which is the Fund's most defensive stance. In Strategic
Growth, our avoidance of "high beta" financial stocks, coupled with some option
price decay in our higher-strike puts, does create a moderate tendency for the
Fund to move inversely to the market - especially during "risk on" days when
financials lead.
.
Given that option volatilities have retreated and we have not
been aggressive in raising our put option strike prices, that tendency has
declined a bit in recent weeks. Moreover, our stock holdings have a strong
record of outperforming the major indices over time, so we remain comfortable
with our avoidance of certain sectors. Our present investment position is
well-aligned with the return/risk profile implied by the market evidence here,
but we'll continue to take opportunities to mute the effect of "risk on"
speculation whenever possible. In the meantime, I expect our present sensitivity
to these periodic spikes to more a short-term nuisance than a longer-term
feature of our investment stance.
.
In Strategic Total Return, our precious metals position
remains at less than 3% of assets, as the defensive shift we saw in the Market
Climate has persisted despite the recent price weakness. Likewise, we cut back
our duration weeks ago in our Treasury holdings, currently to about 3.5 years.
While we would expect Treasury bonds to rebound considerably on any fresh
economic deterioration, our present duration is adequate from the standpoint of
present evidence - the expected return/risk estimates from our ensembles have
not rebounded much despite the recent increase in yields. Overall then, we are
maintaining fairly modest risk exposure here across the board.
0 comments:
Publicar un comentario