Secular Bear Markets - Volatility Without Return
By John Hussman
December 10, 2012
Just a note – Last May, my friend Mike Shedlock’s wife Joanne passed away from an aggressive form of ALS (amyotrophic lateral sclerosis). Mike ("Mish") is an investment analyst whose work can be found on the Global Economic Trend Analysis site. In Joanne's memory, and to benefit the Les Turner ALS Foundation, Mish has organized a conference in Sonoma California, on April 5, 2013. I don’t often speak at conferences or schedule media interviews, but that day, Mish, Chris Martenson, Michael Pettis, James Chanos, John Mauldin, and I will be speaking – for a very good cause. The Hussman Foundation has also committed a $100,000 matching grant to the Les Turner ALS Foundation, to encourage donations that will support patient services and research for individuals with ALS. So with winter approaching, if you’re already looking forward to springtime, I hope you’ll join us in California in April. Thanks - John
Present market and economic conditions highlight a fairly dramatic disparity between continued economic and valuation headwinds (particularly on a “cyclical” horizon of 18-24 months) and complacent short-term conditions that rest on the continuation of massive monetary and fiscal imbalances. It’s obvious even from a casual observation of economic conditions that these imbalances are inconsistent with a healthy economy; short term interest rates near zero, monetary base at 18% of nominal GDP (more than twice the level that would be consistent with short-term yields at even 2%), and a Federal deficit near 10% of GDP. Because it is an accounting identity that the deficit of one sector must be the surplus of another, and neither consumers nor our trading partners are running surpluses, it follows that our massive Federal deficit has temporarily driven corporate profit margins to historic highs about 70% above their norms even while wages as a percent of GDP have reached a record low.
Despite these imbalances, as long as Wall Street collectively closes its ears and hums, everything seems to be just fine. Sure, valuations are rich on the basis of normalized earnings, but stocks have performed well in hindsight. Sure, short-term interest rates are at zero, but investors have found what they believe is value in the higher interest rates available on junk debt. Sure, the labor force participation rate has plunged back to 1980 levels while every cohort of the population has lost jobs in the past 3 years except workers over the age of 55, but the payroll figures remain positive to-date. Sure, Europe is already in recession, with a largely insolvent banking sector, but for now, words have been enough to talk investors down from concern about any of that.
For our part, we continue to focus on the prospective market return/risk profile that has been associated with prevailing market conditions (including not only post-war and Depression-era data, but also data from the most recent cycle). The fact is that market conditions vary measurably over the course of the bull-bear market cycle. It's true that repeated “kick-the-can” interventions during the most recent cycle have created more variability in the lag between unfavorable conditions and subsequent market losses. Still, even this did not prevent significant corrections in 2010 and 2011. The corrections in 2012 have been fairly shallow despite fairly extreme conditions from the standpoint of our own return/risk estimates, but we also observed that in 2000 and 2007 – when it would have been tragic to confuse the postponement of bad outcomes with an escape from them. I have little doubt that the coming market cycle will provide extended opportunities for an unhedged and even aggressive exposure to market risk. At present, however, I am convinced that investors are mistakenly reaching for yield in credit-sensitive debt, misled by temporarily elevated profit margins in the stock market, overconfident about government safety nets, complacent about European risks, and still likely to be blindsided by a U.S. recession.
Secular Bear Markets – Volatility without Return
We continue to view the stock market as being in a “secular bear” – a period that includes a series of separate “cyclical” bull and bear markets, with the defining characteristic that successive bear market troughs move toward increasingly depressed levels of valuation. Secular bears begin from elevated valuations – generally Shiller P/E’s well above 18 (the ratio of the S&P 500 to the 10-year average of inflation-adjusted earnings), and typically end about 14-18 years later, at depressed valuations and after a number of separate market cycles. There are certainly many periods during a "secular" bear market when it makes perfect sense to take moderate and even aggressive "cyclical" risks, but it is worth noting that the average "cyclical" decline in a "secular" bear has wiped out about 80% of the prior bull market advance.
As a severe example of a secular bear period, the Shiller P/E reached 25 in 1929, and plunged to less than 5 by 1932, but despite significant gains off of the 1932 low, valuations eventually settled back to a Shiller multiple of just 7.5 by 1942. From the standpoint of prospective returns, we estimate that valuations were consistent with negative 10-year prospective returns on the S&P 500 at the 1929 peak, but with prospective returns near 20% annually at the 1942 low (see the chart in A False Sense of Security).
Similarly, at the 1965 valuation peak, the Shiller P/E reached 24, plunging to less than 9 by the 1974 low. But while 1974 marked the low in price, it was not the low in valuation. The secular bear continued into August 1982, when the Shiller P/E reached 6.5, at which point the prospective 10-year return for the S&P 500, by our estimates, was again nearly 20%. Future outcomes may be different, but the historical record suggests that extended "secular bull" periods typically emerge from durable valuation troughs, and those troughs are generally associated with prospective 10-year S&P 500 total returns in the range of 15-20% annually. At present, our estimates hover around 4.5% annually even without assuming future undervaluation.
The problem with the late-1990’s market bubble was that it took valuations well beyond those of 1929 or 1965, to a Shiller P/E of nearly 44 by the 2000 market peak. The 1.3% annual total return, including dividends, achieved by the S&P 500 from that point through last Friday’s close, was a fairly predictable result of that overvaluation. What is undoubtedly difficult for investors to accept, however, is that even the 2009 market low took the Shiller P/E only to 12 (and a prospective 10-year return only slightly over 10% by our estimates). The multiple is presently above 21 again.
Now, if the S&P 500 had achieved significantly different returns over the period since 2000 than we actually projected on the basis of normalized valuations, one might be inclined to disregard our valuation concerns as somehow outdated or unreflective of new realities. But we’ve observed no such disparity. Moreover, there's no evidence that our valuation methods are locked in a bearish mode. Even our own estimates of prospective 10-year returns (which are based on a variety of normalized fundamentals well beyond Shiller multiples) indicated – I believe correctly – that stocks were modestly undervalued at the 2009 low, though I certainly don’t believe that 2009 represented a secular low.
On a related note - however one views the need to contemplate Depression-era data in the 2009-2010 period, our valuation approach was not the cause of my concerns in that period. Rather, market action and trend-following components that had performed well in post-war data turned out to allow intolerably large drawdowns in Depression-era data in several instances. Once we had to contemplate the possibility that Depression-era outcomes might be relevant, very robust methods were required to adequately deal with those potential outcomes. We insist on stress-testing, and on testing our approach against out-of-sample validation data. The need to incorporate Depression-era data in our approach is an event that occurred, and will only occur, precisely once - I'm fairly certain that we will never have to contemplate data from the South Sea Bubble or the Dutch Tulip Mania. In contrast, the defensiveness we have at present is an event that I would expect to occur in every market cycle under similar conditions.
I believe that normalized valuations present an accurate view of prospective market returns here. There is enormous risk, in my view, in the temptation to accept zero interest rates and low single-digit prospective market returns as an enduring characteristic of the financial markets while ignoring the unsustainable distortions that have produced this environment. In the short-run, there may be near-term returns available in reaching for yield by accepting greater credit risk, or speculating on periodic relief rallies even at present valuations. Those prospects, however, don’t change our view that stocks remain in a secular bear market.
A bit of arithmetic is instructive here. Suppose that future economic growth remains similar to past growth, and normalized earnings grow by about 6% annually. Assume also that at some point, let’s say 12 years into the future, the Shiller P/E simply touches 10 – still above the 5-7 multiples reached in prior secular bears (excluding the bubble period since the late-1990's, the historical norm for the Shiller P/E is less than 15). Given a dividend yield of 2.3%, and a present Shiller P/E about 21, we can estimate the 12-year prospective return on the S&P 500 on those assumptions at:
(1.06)*(10/21)^(1/12)+.023*(21/10+1)/2 – 1 = 3.2% annually.
As for the past 12 years, we started at a Shiller P/E of 44, moving to 21 at present, and began at a dividend yield around 1.2%. Our arithmetic indicates we should have observed an annual total return of about:
(1.06)*(21/44)^(1/12)+0.012*(44/21+1)/2 = 1.5% annually (Check. The actual return came in at 1.3%).
For optimists, assuming that the Shiller P/E falls no lower than 16, and not until 5 years from now, we would still expect a 5-year prospective total return of just 3.1% between now and that touch-point. For pessimists - or equivalently, for optimists that a secular bull market will begin sooner rather than later - a Shiller P/E of 7 reached 4 years from today would result in a prospective market loss of -14.9% annually over that 4-year period. There are certainly many intermediate possibilities, but unless one expects valuations to remain richindefinitely, with no retreat to historically normal or undervalued levels as far as the eye can see, the eventual resolution of this period of secular overvaluation leaves little to be desired for long-term investors. I have little doubt that better opportunities will be available even over the course of the present market cycle.
One way to think about the effect of a secular bear market is to compare the absolute amount of volatility experienced by the market to the total distance it travels. In the chart below, the blue line represents the sum of absolute weekly percentage changes in the S&P 500 over the preceding 4-year period, divided by the absolute overall change in the S&P 500 over that period. A “spike” in that line indicates that the market experienced a great deal of week-to-week volatility over a 4-year period, without much net movement overall (Geek’s note – this calculation is related to the concept of “fractal dimension” - the spikes are singularities where the ratio is undefined because the 4-year change is close to zero).
An extended period of blue spikes is the hallmark of secular bear markets. These periods have reliably followed periods of elevated valuations as we have observed, with little respite, since the late-1990’s. A great deal of distance traveled, with little to show for it overall. Present valuations provide little reason to believe that this period is behind us. Things will change, and this period of distortion will be behind us. The transition is likely to be unpleasant for the market, but again, I expect that we’ll observe good opportunities to accept significant market exposure even in the coming market cycle.
I continue to view the U.S. economy as being in a recession that began in the third quarter of this year. The data has been very mixed, with the latest manufacturing Purchasing Managers Index (a survey of purchasing managers by the Institute for Supply Management) slipping below 50, but overall, I would characterize the most recent data as being negative but stable – not accelerating to the downside. The recent month-to-month fluctuations are too small to be very meaningful – it is more informative to focus on the overall levels and the persistence of those levels (particularly since all of these measures tend to soar early in true economic recoveries). Presently, the overall profile of national and regional Federal Reserve and purchasing managers surveys is about where it was at the beginning of the last recession. The chart below presents the average standardized value (mean zero, unit variance) of regional and national Fed and purchasing managers surveys. Recessions are shaded.
On the employment front, the November non-farm payroll gain of 146,000 on the establishment survey was surprisingly positive, and it would be uncharacteristic if we don’t see it revised significantly lower in the months ahead. The September and October payroll figures are already being revised lower, consistent with the historical tendency (noted by the Economic Cycle Research Institute) for data to be revised higher prior to the start-date of a recession and then lower after the start-date. The unemployment rate declined despite a drop of 122,000 jobs in the household survey, as 350,000 more workers left the labor force. We didn’t observe any anomalies on the seasonal-adjustment front.
The chart below updates the relationship between the average value of the overall, new order, order backlog and employment measures of Fed and purchasing managers surveys (blue line) and the total change in non-farm payrolls over the following quarter. Note that the present average of these surveys (which is better correlated with subsequent payroll growth than the employment component alone) is consistent with a contraction of employment in the months ahead.
Finally, as multiple measures of employment conditions are generally more informative than a single measure taken by itself, the chart below presents the average change in payroll employment over the most recent 3-month period, the 8-week average of new unemployment claims (scaled), the employment subindex of various Fed and purchasing managers surveys (scaled) and the average of overall, new orders, order backlog, and employment components of those surveys (scaled). Overall, we observe a significant weakening of these measures, which is consistent with our expectation of employment losses in the months ahead. As I noted last week, a material improvement in various coincident-to-lagging measures would be among the first signs of a robust turn in economic prospects (see How to Build a Time Machine).
A final note to our colleagues in finance – maybe it’s just a carry-over from my years as a professor, but I continue to view proper attribution as a basic reflection of intellectual honesty. Last week, several investors forwarded a recent asset allocation report issued by a well-known investment company, noting that it reads like a point-by-point summary of a dozen of our weekly comments, including the original observations of a colleague at another investment company as well. This should not be difficult - when you use someone's content, or build on someone’s concept, acknowledge the person or the concept, and then add your own work. Don't just lift models or material. Lifting someone else's work just says that you don't believe in your own intellect. Bill and I do our best to always include attribution when we draw on other people's work (send us a note if we miss one), except when we criticize someone's approach, when we may omit names unless the person appears to be damaging or misleading investors. This isn’t a business matter, it’s an ethical one.
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.
Past performance does not ensure future results, and there is no assurance that the Hussman Funds will achieve their investment objectives. An investor's shares, when redeemed, may be worth more or less than their original cost. Current performance may be higher or lower than the performance quoted above. More current performance data through the most recent month-end is available at www.hussmanfunds.com. Investors should consider the investment objectives, risks, and charges and expenses of the Funds carefully before investing. For this and other information, please obtain a Prospectus and read it carefully.
(c) Hussman Funds