How long before we get swallowed by a whale?
John P. Hussman, Ph.D.
Last week, the S&P 500 advanced the extra 1% required to re-establish virtually every “overvalued, overbought, overbullish, rising-yields” syndrome that we define – syndromes that have appeared at or close to the beginning of what investors can easily recall as the singularly worst set of market instances in history, including the 1973-74, 1987, 2000-2002, and 2007-2009 plunges. With some minor imputation (estimating bullish and bearish sentiment as a function of the extent and volatility of prior market movement), we can verify that these syndromes also emerged just prior to the 1929-1932 collapse.
The S&P 500 is presently near or through its Bollinger band (2 standard deviations above its 20-period moving average) at daily, weekly and monthly resolutions, the Shiller P/E (S&P 500 divided by the 10-year average of inflation-adjusted earnings) is above 22, Investors Intelligence reports lopsided sentiment at 53.2% bulls versus 22.3% bears, the S&P 500 is well into a mature bull market and Treasury bond yields have advanced measurably.
The blue lines over the chart of the S&P 500 below are even sparser than our typical charts of “overvalued, overbought, overbullish, rising-yield” events, and identify the points where the S&P 500 was within 3% of its upper Bollinger bands, at least 7% above its 52-week smoothing, and over 50% above its 4-year low, with bulls above 52% and bears below 27%, a Shiller P/E above 18, and 10-year Treasury yields above their level of 6-months earlier. As I often note, there are numerous ways of defining syndromes like this. The conditions here are essentially a way of identifying what have normally been the most strenuously overvalued, overbought, overbullish, rising-yield points within already mature market advances. Investors who are willing to accept present, record profit margins (about 70% above historical norms) as permanent will reject the notion that stocks are overvalued, but I trust that the evidence we’ve presented over time underscores the deeper basis for that conclusion (see last week’s comment Declaring Victory at Halftime).
For optimists, there is a false one-week signal in 1997 – during the internet bubble – that was not associated with a negative follow-through. That is, as long as one ignores that the S&P 500 was essentially unchanged 5-years later, underperformed Treasury bills for the next 12 years, and even through last week’s advance, has outperformed Treasury bills by less than 2% annually in the nearly 16 years since then (all of which would be surrendered by even a run-of-the-mill bear market decline – and then some). There is also a fairly uneventful signal in mid-1964 that was not followed by near-term losses, though the market was still lower two years later and would underperform Treasury bills for the next 20 years.
Notably, during the recent bull market advance, we’ve seen conditions as extreme as at present only once – in early 2011, just prior to a nearly 20% market loss, though not rivaling the 30-50% plunges that this syndrome has also captured.
My concerns here are understandably easy to dismiss given that the S&P 500 is now more than 5% higher than it was in March of last year, when our estimates of prospective return/risk (on a smoothed horizon from 2-weeks to 18 months) first plunged into most negative 1% of market history. Yet despite the intervening monetary heroics, history suggests not that these conditions will persist without resolution, but instead that their resolution is likely to be that much worse. Anyone who followed me in 2000 or 2007 will easily recall a similar frustration before the bottom fell out of the market on each occasion. I can’t assure that the same will occur in the present case, but I believe it would be reckless to assume that the Fed has these risks covered. With margin debt over 2% of GDP as it was on three prior occasions – 2000, 2007 and early 2011 – it’s clear that investors are all-in when it comes to faith in the Fed. Still, when an investment thesis becomes so universally embraced and so apparently easy to follow that anyone who resists is considered foolish (as was the case with tech stocks in 2000), my risk aversion needle hits the red zone.
If the defensiveness of a supposed permabear seems hard to swallow even with the S&P 500 pushing its upper bands in a mature bull market, here’s what I would suggest. Carefully identify what has ultimately (not temporarily) been unlikeable or incorrect about my investment views over the years. For most people who follow my work, the main answer will be my insistence on remaining defensive in 2009 (despite the success of our approach to that point) until I was certain that our methods were robust to “out of sample” Depression-era data. There are also several points in recent years when trend-following measures were favorable – but without the presence of hostile “overvalued, overbought, overbullish” syndromes – where some added emphasis on trend-following measures would have helped in the face of unprecedented monetary intervention (unfortunately, constructive trend-following measures have historically been of little help to prospective returns once overvalued, overbought, overbullish syndromes have kicked in).
The ensemble methods that resulted from that 2009-early 2010 “two data sets” challenge, and criteria to increase the frequency of constructive positions when our trend-following measures are favorable (absent hostile syndromes, but even when our return/risk estimates are negative), represent the two material changes we’ve made to our approach since 2000 (see Notes on an extraordinary market cycle). This past cycle was extraordinary because it featured not just risk (the negative portion of a reasonably known range of outcomes), but also extreme uncertainty (the lack of knowledge about the range of outcomes that are even possible). Our 2009 “miss” did not result from applying our present methods, but as I noted at the time, resulted from the need to address a “two data sets” problem in a world where Depression-era outcomes had become possible. It’s not entirely clear whether future cycles could invite Depression-like outcomes, rapid inflation, or unforeseen fiscal strains. What I do believe is that our approach (incorporating ensemble methods) is robust to “out of sample” data, and can engage those uncertainties – if they emerge – without further changes. It’s clear that what I saw as a necessary fiduciary response in 2009 ended up injuring not only my record, but my reputation, and the trust of those who rely on my work. I do believe it made us far stronger, but time will tell, and I don’t expect that we will require a great deal of it.
In any event, one can hardly embrace our 2000-2002 and 2007-2009 defensiveness (not to mention our intervening bullish shift in early 2003), and at the same time reject our present defensiveness, because what our approach indicates today is what it would have correctly indicated in similar conditions in numerous other cycles. None of this ensures that stocks will not advance to further extremes in the near-term. Regardless, I have very little doubt that investors will observe better – or spectacularly better – points at which to accept market risk, when the evidence does not ask them to rely on the hope that a strenuously overvalued, overbought, overbullish market will become more so.
The U.S. fiscal deficit remains near 10% of GDP. The Federal Reserve continues a well-advertised campaign of quantitative easing, on its way to push the monetary base toward 27 cents per dollar of nominal GDP (the last time we even hit 17 cents was the early 1940s, which was not unwound by subsequent tightening, but instead by a near-doubling in the consumer price index by 1952). Conditions in Europe have stabilized due to massive easing by the European Central Bank, coupled with assurances that the ECB will do whatever it takes to backstop sovereign debt obligations (though these words are actually still untested, as is the meaning of “strict conditionality.”)
What’s fascinating is that in the presence of what are not thin strings, but massive cables supporting the economy like a puppet, the only response that Wall Street can muster is “Hey! He’s walking!” – as if the puppet is capable of motion without being propped up to a nearly reckless extent.
I’ll preface these economic comments by repeating that our concerns about market risk here are independent from our economic concerns, and are driven by conditions that have typically been negative for stocks regardless of the economic context. Still, to the extent that part of Wall Street’s enthusiasm here rests on the idea that Pinocchio has become a Real Boy, it’s useful to review the evidence.
At this point, the economic data are wrestling between two likely possibilities and a third less likely one. The first of the likely ones remains that the U.S. already entered a recession in the third quarter of 2012. On that possibility, as statistician Nate Silver has noted, in data since 1965 the average revision from the initial estimate of quarterly GDP growth to its eventual figure has been 1.7%, and the 95% confidence interval has been plus or minus 4.3%. While I expect the full third-quarter GDP figure of 3.1% to remain positive post-revision, it’s not at all clear that fourth-quarter GDP (estimated to come in about 1.5%) will survive those eventual revisions – ditto for the marginal bounce in industrial production.
The second likely possibility is that the enthusiasm about QEternity (combined with a positive jolt to personal income from special dividends to front-run the fiscal cliff) represented another successful round of “kick-the-can” to push a weak economy from the verge of recession for another few months. When we look at the broad evidence from a variety of good leading and coincident indicators, that’s actually the possibility that I am starting to lean toward.
The chart below offers a reasonable picture of present conditions. Note that in recent months, we’ve enjoyed a bounce in the overall, new order, and order backlog components of a variety of regional and national economic surveys from the Federal Reserve and Institute of Supply Management. Interestingly, the employment component has persistently deteriorated. Present levels are largely indistinguishable from what we observed when the economy was already in recession in early 2008, but in view of the series of stronger, more successful can-kicks the Fed has achieved in recent years, we have to allow for the possibility that the most recent one has bought a few months of recession avoidance.
The unlikely possibility, in my view, is that the economy has started to walk on its own (much less sing and dance with Jiminy Cricket). There is not nearly enough improvement in leading measures, coincident/lagging measures, or other metrics to infer a meaningfully positive shift in the broad economy. As I’ve noted before, joint upward leaps in the Philadelphia Fed index and the new orders component of the Chicago Fed report would be strong indications that we should back off from our recession concerns. As things stand, the Philadelphia Fed numbers were revised lower for prior months, and jolted to the downside in last week’s report.
I should note that we did see a large drop in initial claims for unemployment last week, but it’s important to understand the dynamics of initial claims. As a rule, the second week in January experiences a massive upward spike in actual, non-seasonally adjusted initial claims, as retail jobs are lost following the holiday season. To smooth that spike away, the seasonal adjustment for the second week of January is aggressively downward. But as a further result, in years where that upward spike in actual job losses doesn’t concentrate in week two, the seasonal adjustment turns out to be far too large, so the seasonally-adjusted figure seems remarkably low – as it did last week. In those cases, the normal post-holiday job losses still occur, just distributed more broadly through about mid-March. But because the seasonal adjustment process “thinks” it already took care of that spike, the seasonal adjustment isn’t aggressive enough as those jobs actually come off, and what follows is several weeks where the seasonally-adjusted initial claims numbers are tens of thousands higher than they otherwise would have been. We’ll know that this dynamic is operating if initial claims inexplicably begin to rise toward 400,000 between next week and mid-March. We’ll suspect that a recession dynamic is also in play if the figures rise materially beyond 400,000.
The foregoing comments represent the general investment analysis and economic views of the Advisor, and are provided solely for the purpose of information, instruction and discourse. Only comments in the Fund Notes section relate specifically to the Hussman Funds and the investment positions of the Funds.