John P. Hussman, Ph.D.
Portfolio Notes – Particularly over the past few quarters, the Strategic Growth Fund has enjoyed muted volatility and positive returns during market declines, but also a moderately inverse relationship versus the S&P 500 during market advances. This behavior isn’t a general feature of our hedging approach, but rather the reflection of two factors that are currently in place. One is our significant « underweight » in financials, materials, cyclicals and other « risk on » sectors that we view as speculative, and where we find few candidates that satisfy the discounted cash flow criteria that we rely on for stock selection. While our stock selection has significantly outperformed the S&P 500 over time, our continued avoidance of financials does introduce some inverse behavior in our hedged investment position during runs of « risk on » speculation.
The other related factor is that the past several quarters have been a constant game of « hot potato » between recession risk and what we identify as an « overvalued, overbought, overbullish » syndrome. The result is that one or the other has generally kept us in a tightly hedged investment position. In our most defensive stance (conditions we identify as « hard negative »), we typically endure some decay in option value during market advances, because we raise the strike prices of our put options in order to defend against indiscriminate selling that often follows, as we saw during sharp market declines in 2010 and 2011. The overall result is that the Fund has typically enjoyed positive returns when the market plunges, but has experienced some erosion during periods of « risk on » speculation.
In order to eliminate this somewhat inverse pattern, we could take a new position in financial stocks and other « risk on » sectors here, and lower the strike prices on our defensive put options. My impression is that both of those would be hostile to our prospective returns, and of course, would also depart from our stock selection discipline in the process. One do-it-yourself method of closing down that pattern would be to take a position in financials, materials, and cyclicals, write call options on them in order to take in time premium, but accept all of the downside risk in those holdings. That would do it. For most of our shareholders, my guess is that that doesn’t sound like a brilliant idea here. This may offer some appreciation for why we continue to pursue our discipline, despite the occasional pressure we experience – I expect temporarily. Even if we ignore economic risks entirely, we presently observe overvalued, overbought, overbullish conditions that have repeatedly been resolved by steep declines, even in the past few years.
We often receive questions relating to the ensemble method that guides our hedging strategy. Along with last week’s comment (Notes on Risk Management ), the following section is intended to provide a broad overview.
One of the main approaches we use to estimate return and risk prospects is to group current market conditions among historical instances that are most similar. Each point in history is defined by various « features » based on a broad range of key factors, including valuations, trend-following indicators, market breadth, sentiment, credit spreads, economic factors, overbought/oversold measures, and so forth. In order to make the analysis less dependent on any particular historical period (e.g. postwar data, bubble-era data, Depression-era data), or any single set of indicators, we extend this analysis to a very large number of randomly selected sub-samples across history.
This sort of analysis is an example of an « ensemble method, » which has several benefits, the two most important being on measures of « accuracy » and « robustness. » It’s easy to fit a model to past data, but those models often break down quickly in new data. So to evaluate accuracy, we estimate return and risk on data that the model has not « seen » previously, and find that the ensemble approach generally performs better than alternative methods. Equally important, the ensemble is robust to very large changes in the underlying economic environment, because randomizing over numerous sub-samples of history reduces the likelihood that the model is « over-fitted » to a particular economic environment.
I wish I could say that we anticipated the depth of the 2008-2009 credit crisis so completely that I developed these ensemble methods in advance, already confident in how they would have performed even in Depression-era data. Unfortunately, that’s not the case, and shareholders are well aware of the challenges we went through in stress-testing our approach against other periods of credit crisis.
Though these weekly comments forewarned much of what actually occurred during the credit crisis, I certainly didn’t anticipate what I still consider to be terrible policy mistakes – particularly the absolute unwillingness to restructure bad debt, in preference for kicking the can down the road with public funds. It was a far cry from how U.S. regulators had responded to the S&L crisis, and how other international banking crises had been successfully addressed (for example, in the early 1990’s, the Swedish banking crisis was durably resolved by the government taking receivership of a large portion of the banking industry, wiping out existing shareholders, writing down bad assets, and then taking the banks public to recapitalize them under new owners).
For anyone who was responsible for investing the funds of others, the proper response to the 2008 crisis was to stress-test every method, though I’m not convinced that much of Wall Street has stress-tested anything at all. For us, stress-testing meant taking our models to Depression-era data, because it was clear that events of the time were largely « out of sample » from the standpoint of post-war data. At the time, we were basing our estimates of market risk and return on data since about 1950, which I had – incorrectly – believed was sufficient to capture « modern » market behavior.
While our existing hedging approach performed well in that Depression-era data overall, the occasional losses were far deeper than I was willing to risk for our shareholders. The result was what I called a « two-data sets » problem, which demanded that our hedging methods perform well, out-of-sample, and with tolerable drawdowns in data drawn from both post-war and Depression-era periods. We reached a satisfactory solution in 2010 with the introduction of our ensemble approach. For the full period, we avoided a significant portion of the market’s 2007-2009 downturn, but in hindsight, my decision to fully stress-test our methods led us to miss a rebound in 2009 that we should not have missed, had our present approach been already in hand.
In real-time, our hedging approach has repeatedly demonstrated value over complete bull-bear market cycles, both adding returns and defending against severe market losses (exceeding 50% downturns twice in the last decade). We’ll certainly have periods where we appear remarkably out-of-step with the prevailing trend of the market, particularly in overvalued, overbought, overbullish periods of speculation. But defending against losses in these periods is essential to risk management, despite the tendency of bulls to declare victory at halftime.
The period since 2010 has been largely characterized by a fragile underlying global economy coupled with a persistently overvalued stock market (though to varying degrees). We’ve seen little during this period but the effect of a hot potato being repeatedly passed from speculatively overvalued, overbought, overbullish market conditions driven by massive central bank interventions, to credit strains and emerging economic weakness nearly the instant those interventions are even temporarily suspended. As a result, by turns we’ve seen the repeated emergence of the same speculative « Aunt Minnies » that have historically accompanied major and intermediate market peaks, followed by the emergence of credit strains, economic pressures, and a flight to safe-havens.
The alternation is certainly not typical of market history. Nor is it typical of a complete market cycle or business cycle. As unsatisfactory as it may be, the market is presently in an extended game of hot potato which will be resolved by the eventual removal of both conditions.
When I was in college, I bought a stick-shift hatchback wagon that I used for years to haul equipment as the guitarist and lead singer for a rock band (the gig earnings went to buy time on the mainframe at Northwestern’s Vogelback computing center so I could run investment research). If you’ve ever been a beginner driving a stick, you know that if you don’t release the clutch just right, the car goes nowhere and then suddenly jolts forward once the gear engages. My impression is that this is largely what we’re seeing in the economy. Each time underlying credit strains emerge, demand backs off as consumers and businesses become averse to spending. Then, each time central banks launch some massive new intervention, there is a jolt of pent-up demand that is interpreted as sustainable growth. This was the result when the Fed launched QE2, and we’re seeing a replay as the ECB provides enormous loans to banks in return for « collateral » in the form of newly-created, unlisted bonds that European banks have simply issued to themselves.
But what if we are not, in fact, facing further economic weakness, and are instead on course for recovery? What indicators should we monitor, and how would our investment position change?
On the indicator front, as I noted last week, we use dozens of economic indicators and discriminators in practice, but a good, simple rule of thumb to gauge recession risk is to use the combination of the OECD leading indices for the US and total world, combined with the ECRI weekly leading index (WLI). The latest readings from the OECD come out this week. Given that the WLI is still negative, an upturn in both OECD measures – to about +2 on each – would help to relieve our economic concerns.
[Geek’s Note: We use standardized values for all of these measures: WLI mean 2.2, std 7.6, OECD_US mean 2.8, std 5.1, OECD_W mean 2.8, std 4.3. Standardized values below -0.5 on all three of these are nearly always associated with recessions. A subsequent move above a standardized value of about -0.2 in at least two of these three has quickly marked new expansions in the past. The corresponding level to monitor on each of the raw indices, of course, is mean – 0.2*std, which translates to levels of about 0.7 on the WLI, 1.8 on the OECD US leading indicator, and 1.9 on the OECD total world index. Again, this is a rule of thumb, but it has a good record for a three-indicator model].
However, it bears repeating that even if we zero out our economic concerns (and the associated warning indicators in our ensembles) we still obtain very unfavorable expected return and risk estimates for the stock market here. This is because we presently observe a number of historically hostile syndromes that are almost uniquely associated with losses – not always immediately, but almost always large enough to make any intervening gains purely temporary. The stock market may very well enjoy a further advance from here. The likelihood of those gains being durable, however, is quite small.
We’ve heard a few objections that our concerns about market risk are inconsistent with the continued downtrend that we observe in new claims for unemployment. On that note, it’s true that there are a few useful indicators that can be derived from new claims data. For example, the stock market often suffers when new claims rise above their 5-year average, from being previously below that level. But even in those cases, the stock market has usually been falling already, because stocks are short-leading and new claims are at best coincident with the economy. So rising unemployment claims are a « bearish continuation » signal, but should not be expected to provide early warning. More generally, the trend of new claims actually has very little to do with oncoming market action.
A particularly instructive instance is 1987. The chart below shows the 4-week average of new claims for unemployment that year. Notably, the persistent downtrend in new unemployment claims provided no barrier at all to the October 1987 crash. I’ve chosen this particular counterexample because we presently observe the same unusually overextended market conditions that characterized the 1987 peak. These include an overvalued, overbought, overbullish syndrome, which has become increasingly familiar near both major and intermediate market highs in recent years, including the peaks we saw in 2007, 2010 and 2011. The 1987 peak also featured the same « exhaustion syndrome » I discussed a few weeks ago in Goat Rodeo (basically a recent « whipsaw trap » syndrome coupled with falling earnings yields).
If you spend any time at all with historical data, you’ll find a multitude of nearly equivalent ways to define an exhausted advance, and the average outcomes are almost always uncomfortable. When these conditions are coupled with any upward interest rate pressure at all, whether from corporate, Treasury bond, or T-bill yields, the outcomes are almost uniformly hostile.
As one of many ways to define « overvalued, overbought, overbullish, rising yield » conditions, consider the points in history when the S&P 500 was at a « Shiller » multiple of over 19 times 10-year inflation-adjusted earnings, the index was at least 8% over its 89-week moving average, within 2% of a 3-year high, with Investors Intelligence sentiment over 45% bulls, less than 30% bears, or both, and with at least one yield measure above its level of 26-weeks earlier (corporate, Treasury bond, or T-bill). This set of conditions produces a cluster restricted to about 8% of market history, and also self-selects for many of the worst times an investor could have chosen to buy stocks, based on the depth of the market’s decline within the following 18 months.
While the criteria above are loose enough to include several false signals, the periods also include late-1961 (-25%), early-1966 (-20%), late-1968 (-30%), late-1972 (-30%, and a nearly -50% loss extending beyond that 18 month window), mid-1987 (-33%), mid-1998 (-12% over the next 13 weeks), mid-2000 (-35%, and a loss of more than 50% beyond that 18 month window), and mid-2007 (-55%).
We’d never recommend this as an actual investment strategy, but it’s informative that even if an investor had simply avoided the market for a full 18-month period after every emergence of the foregoing set of conditions (including every false signal, and entirely regardless of what else happened over the next 18 months), that strategy would still have captured cumulative returns about 70% higher than a buy-and-hold since 1963, with periodic losses only about half as deep as a buy-and-hold approach. Again, we’d never recommend this in practice, but it underscores that although the market may move even higher over the near term, investors have achieved nothing durable from investing in overextended conditions like those we presently observe. On average, very weak market outcomes have followed for an extended period of time.
There are tighter ways to define conditions that exclude false signals but also miss some major declines (see A Who’s Who of Awful Times to Invest and Extreme Conditions and Typical Outcomes ), and looser ways to define conditions that capture even more historical plunges, but also invite more false signals. Regardless, the benefit of avoiding the major plunges nearly always swamps the cost of the occasional false signals.
The S&P 500 has lost more than half of its value on two separate occasions since 2000, and the value of avoiding major losses in a decline generally offsets missed gains very quickly (a 20% loss wipes out a 25% gain, a 30% loss wipes out a 43% gain, a 40% loss wipes out a 67% gain, and a 50% loss wipes out a 100% gain). My argument is certainly not that stocks will decline immediately, nor that they cannot advance further from present levels. Rather, the point is that even if such an advance emerges, the likelihood of those gains being retained by investors over the course of the full market cycle is exceedingly small.
This cycle of hot potato will end, and we will have opportunities to accept moderate or even significant amounts of risk, with proportionately high expected returns. As recession uncertainties resolve, and we observe a normal ebb-and-flow of investor sentiment and short-term price movement, I expect that we’ll observe at least some of these opportunities in the months ahead. A major positive shift in our investment stance would most probably accompany a significant improvement in valuations, confirmed by improving market internals (a sequence that is characteristic of early bull market advances). With the market presently at (normalized) valuations that are associated with poor long-term prospective returns, with short-term conditions overbought and overbullish, and with intermediate-term conditions characterized by an exhaustion syndrome that has historically produced disproportionately weak returns over the next few quarters, I do not believe that we are faced with such an opportunity here. This will change, and we will respond accordingly.
Meanwhile, it’s worth repeating that most bear markets wipe out more than half of the gains achieved during the prior bull market. There is very little chance, in my view, that gains from present levels will be retained by investors over the completion of the current cycle. There is equally little chance that investors who are willing to accept significant risk now will be prompted to reduce their risk later, until they encounter a market decline that is – by then – nearly impossible to act upon. Have we not seen this movie before?
As of last week, the stock market remained characterized by an overvalued, overbought, overbullish, rising-yields syndrome, coupled with an « exhaustion » syndrome that has historically been followed by declines on the order of -25% over the ensuing 6-month period. Our return/risk estimates remain « hard negative » here. We’ve been relatively slow in raising our put option strike prices toward the prevailing level of the market, but I continue to expect some modest « inverse » behavior of Strategic Growth relative to the major indices, largely owing to our lighter holdings of financial stocks, cyclicals, and high-beta stocks here. Though day-to-day movements in individual holdings can affect day-to-day Fund values (as one of our largest and most profitable holdings did last week on a pullback), our stock selections have performed well relative to the indices in recent years, and also since inception. Most of the « basis risk » from weighting our holdings differently from the major indices has tended to be short-lived. Strategic Growth and Strategic International are tightly hedged here.
In Strategic Total Return, we reduced our precious metals holdings further early last week. Despite longer-term inflation risks, precious metals shares clearly dislike upward pressure on long-term interest rates, particularly when that pressure does not reflect simultaneous pressure on observed inflation rates. While gold stocks tend to decouple from the broad stock market during economic downturns and inflationary periods, that doesn’t always happen during significant selloffs, particularly when there is any upward pressure on interest rates. I expect that we will significantly expand our holdings in precious metals shares (currently less than 2% of assets) in the event of price weakness in this sector, particularly if observable economic pressures and renewed downward pressure on interest rates reappear. Presently, given the volatility of this sector, we’ve taken some profits and remain on watch. The Fund continues to hold a duration of about 4.5 years in Treasury securities.