In my view, we are likely witnessing the peak of the third equity valuation bubble in the past 14 years, the first two which saw major indices plunge by at least 50%. It’s important to recognize that market peaks are a process, not an event. Internal deterioration has actually been developing since early July, and became measurable in early August (see A Hint of Advance Warning). This process has been quite like what we observed in 2007, when deterioration became measurable in July of that year (see Market Internals Go Negative). Despite an initial selloff, the major indices recovered to a marginal new high in October 2007 before continuing lower.
As conditions presently stand, the equity market remains vulnerable to the same kind of abrupt air-pocket we observed several weeks ago. We don’t need to make forecasts however – if market internals and credit spreads improve, the immediacy of our concerns will ease considerably. It’s just that we presently observe a combination of evidence that has generally been disastrous over time (see A Most Important Distinction for a chart of the cumulative performance of the S&P 500 under conditions matching what we presently observe).
On the valuation front, based on measures that we find most reliably correlated with actual subsequent market returns (and many widely-quoted measures are not), equities are more overvalued than at any point in history except 2000, at more than double their historical norms, and the best-correlated ones are within 15% of the 2000 peak (see Ockham’s Razor and the Market Cycle to review a variety of these). With market internals still struggling and credit spreads widening, elevated valuations have now been joined by a subtle but measurable shift toward increasing risk aversion. In prior cycles, that shift has typically discriminated between overvalued markets that continued higher from overvalued markets that fell like a stone. Again, if that evidence eases, the immediacy of our concerns will ease as well. That wouldn’t make stocks any less overvalued, so a certain line of defense will be important in any event.
Here and now, the simple fact is that we observe conditions that have historically been associated with very little but air-pockets, free-falls, and crashes. In the past few months, trend-sensitive components of our discipline (involving credit spreads, market internals, and other features of market action that I sometimes describe collectively as “trend uniformity”) have been deteriorating, suggesting a shift toward risk aversion in the face of extreme overvaluation, even as the major indices have been pushing to marginal new highs.
That combination of rich valuations, lopsided bullish sentiment, overbought conditions, and importantly, deteriorating market internals, has been critical in identifying major equity market risk over time. As I noted at the October 2007 peak (see Warning: Examine All Risk Exposures):
“If we partition history into ‘buckets’ according to the Market Climate that was prevailing at the time, and look at how the market performed over say, the next week, month, or quarter we find that every bucket includes periods that were followed by advances, as well as periods that were followed by declines. In other words, we can rarely predict that the market will reliably advance or decline over the next week or month or quarter. What we can say is that the average return/risk profile varies substantially across buckets. Given observations over at least one complete market cycle, we regularly find that the strongest average return/risk profile was associated with periods that one could identify in advance as having favorable valuation and (already) favorable market action, and that the poorest subsequent return/risk profile was associated with the bucket of periods having unfavorable valuation and unfavorable market action. Over the complete market cycle, this knowledge has generally been enough to achieve strong full-cycle returns with moderate risk.”
The main challenge to a hedged investment approach in recent months has not been a runaway bull market that is violating historical regularities. Rather, the main challenge is the short-term effect of those internal breakdowns, as the more democratic equity indices remain below their July highs (see for example the Russell 2000 and the NYSE Composite), while the mega-cap weighted indices such as the S&P 500 and Nasdaq 100 have advanced about 5% and 10% further, respectively. That sort of “basis widening” is uncomfortable for hedged strategies, but in the context of overvalued market conditions, it also tends to be a precursor to major market losses.
While internal dispersion has negative implications for an overvalued market, it’s also somewhat welcome from a stock-selection perspective, because it creates new opportunities. As I observed in January 2008 (see Minding the Hinges on Pandora’s Box), before the market collapsed: “While I remain very concerned about overall market conditions, I am increasingly enthusiastic about the emerging dispersion in valuations that we're starting to observe among individual stocks and industries… You can observe this dispersion in the fact that equal-weighted market indices have been declining more sharply than the capitalization-weighted indices. This can produce a bit of short-term discomfort because we do use those cap-weighted indices to hedge, but I see it as a very good development in terms of intermediate and long-term prospects for investment returns.”
Recent market action includes a concurrent expansion in the number of individual stocks setting both new 52-week highs and new 52-week lows. This is also an indication of growing internal dispersion. As I’ve noted before, instances where both new highs and new lows exceed, say, 2.5% of issues traded are relatively common, and represent practically useless information. Observers incorrectly identify these as “Hindenburg” signals, which get a bad rap as a result. A better definition of a Hindenburg requires additional signs of dispersion – strength in the indices coupled with weakness in the internals, and multiple signals within a few weeks of each other in order to filter out one-off outliers. Those signals, though more useful, are also somewhat spotty. However, when they occur in the context of rich valuations and negative trend uniformity on broader measures, we sit up and take note. Aside from one instance in August 2013 that was followed by just a quick 5% pullback, the only points in recent years we’ve seen that combination turned out, in hindsight, to be the 2000 and 2007 peaks.
As for economic data, remember that economic changes unfold sequentially, typically with changes in real sales and consumption first, then in production, then in personal income, and finally employment (a clear lagging indicator) much later. While Friday’s employment report was encouraging, it is also most informative about what was broadly happening in the economy 4-6 months ago, not about the future. Pay more attention to changes in real sales and consumption, then industrial production and other sensitive measures of activity such as materials prices, and then personal income. Employment is a lagging confirmation of what happened a while ago, not an indication of what’s coming next.
Again, if market internals and credit spreads improve, the immediacy of our concerns will ease quite a bit, despite what we view as extreme overvaluation. We need not rule out further extremes in valuation, nor do we dread them in the belief that we must remain locked into a hard-defensive position. At current levels, a safety net is necessary regardless of other conditions, but there is still a range of possible investment stances between hard-negative and constructive with a safety-net. Presently, we remain hard-negative based on the specific constellation of evidence we observe. Our outlook will change as the evidence does.
Observations on active vs. passive investing
I exchanged a few notes with my friend Bob Huebscher at Advisor Perspectives last week, who later suggested that I share part of that discussion more broadly. Bob began by asking my opinion of a defense of my work written by David Horn, a professor of value investing at Columbia University. Though it’s somewhat humbling to need a defense in the first place, I certainly appreciated David’s piece. Conversely, I don’t take various criticisms personally – I’ve earned some of it, and I know what’s true and what isn’t. We’ll live out what’s true in our actions over time.
Having addressed what we’ve viewed as the two unique challenges of the recent half-cycle since 2009, I think it’s important to note that we don’t look at the possibility of a continued bull market advance with dread or anxiety. As I’ve noted frequently, but still seems rather misunderstood, both our pre-2009 methods of classifying market return/risk profiles, as well as our current methods, are effective in data from market cycles across a century of history. That includes the present cycle and the half-cycle since 2009. The difficulty was not in our pre-2009 methods nor in our present ones, but in the awkward, one-time transition between them.
What seems lost in nearly every discussion of our challenges in the half-cycle since 2009 is the recognition of this transition. It’s evidently much more satisfying to ignore our very open discussion of those challenges, and instead to portray this period as a representative sample of our discipline. Anyone who reads these comments knows (a hundred times over) that this transition comprised: 1) my necessary but unfortunately-timed insistence in 2009 on stress-testing our methods against Depression-era data and, 2) several trend-sensitive overlays that were present in our pre-2009 methods, weren’t sufficiently captured in the ensemble methods that came out of the stress-testing, and essentially had to be reinstated as a result of QE-induced yield seeking. I detailed the last of those adaptations in June (see Formula for Market Extremes), and I view this transition as complete. Based on how our present, adapted methods could have navigated market cycles across history, including both the bubbles of recent years and the worst of the Depression era, we feel the need for neither hope, nor dread, about future market direction.
Frankly, I don’t really know how to communicate this more clearly, without demonstrating it through our actions over time. That’s probably the best way forward.
More broadly, what’s interesting is how the challenges we’ve had are used as talking points in arguments where I personally have no dog in the fight. In particular, I’ve become kind of a lightning rod and a fairly convenient object of focus in the debate over the merits of active versus passive investing. I suppose I’d be more comfortable being used as a poster child for active investing if my performance in the recent half-cycle was representative of active strategies in general, or even my own strategy either pre-2009 or at present. The intervening period has admittedly been an awkward transition for us, but again, I view that transition as complete.
As for my own view of the active/passive debate, I have no objection to passive investment strategies, provided that they are pursued with discipline and a full understanding of the periodic risks they unavoidably include. My view is that even passive investors should still be careful to align the effective duration of their portfolio with the duration of their future spending plans. At a historically normal 4% dividend yield on the S&P 500, the effective duration of equities works out to about 25 years, but at the present 2% dividend yield, it works out to about 50 years, and importantly, the mathematics of duration are not affected by variations in the growth rate of earnings. Accordingly, I believe that passive investors are being given a very large but possibly fleeting opportunity here to rebalance their duration. Even without any view about future market direction, investors with a 25-year future spending horizon should not have much more than half of their portfolios in equities here.
On the active side of investing, there seems to me to be a great difference between “active and undisciplined,” and “active, disciplined, value-conscious, risk-managed, and historically-informed.” Among active investors who have have outstanding long-term records (and in stock-selection, I believe we are included in that group, despite challenges we’ve had on the hedging side during the half-cycle since 2009), my impression is that nearly everyone worth their salt falls into the second category.
To be fair, there are a few active momentum investors who are extremely disciplined and have strong records, despite the fact that they don’t share the “value-conscious” and “risk-managed” components of that formula. The difference is that, wow, do they take a beating during bear markets. Nevertheless, they stick to a historically-informed discipline. From my perspective, that adherence to a historically-informed discipline is the central lesson.
Understand that discipline doesn’t mean blind adherence to an ever-unchanging strategy. Part of a good discipline is continuous research, so if something isn’t right, the challenge is addressed in a systematic way. That process doesn’t always go painlessly, but you address what needs to be addressed, you validate it as broadly as possible over history, and then you adhere to that discipline in order to take the very best next action you can identify. I think that good active investors address challenges in a way that is consistent with the lessons of history. Too many investors simply throw discipline to the wind and repeat at the peak of every cycle that “this time is different.”
Those who follow a passive investment strategy have no less need for discipline than an active investor. It’s unfortunate – and we saw this too often in prior cycles – to see an investor who becomes frustrated at missing returns in an overvalued market, finally capitulates in the belief that the effectiveness of “buy and hold” has been proven by a very long run of positive market returns, and who then encounters a market loss, holds most of the way down, ultimately cannot actually tolerate a 50% portfolio loss (as occurred in 2000-2002 and again in 2007-2009), and bails out at the lows. That’s why we encourage passive investors to understand the actual extent of historical market losses over time, and to properly align the duration of their portfolio to match their expected spending needs.
David Horn is exactly right that even when an active strategy has experienced far less downside risk than a passive one, the active strategy can be viewed as unforgivable if those losses occur in a period where passive strategies are gaining. Part of the discipline required to successfully pursue both active and passive strategies is to understand them from the standpoint of a complete market cycle. Quite often, investors jump between them depending on which has done best over some portion of that cycle.
My own view is that stocks are vulnerable to the risk of deep losses over the completion of the present cycle not unlike those it experienced in the two most recent cycles, and are likely to post total returns from present valuations of only about 1.4% annually over the coming decade. Still, we have no intention of drawing other investors into that view, and certainly no desire for others to abandon well-considered disciplines that at least allow for these risks.
For both active and passive investors, it helps to have discipline, historical perspective, and a risk-management approach that considers one's own investment horizon and tolerance for loss, as well as the actual losses that various markets have demonstrated over time. As long as those components are present, I don’t see any need for tension between the two approaches.
© Hussman Funds