In my book, the best long-term market statistician is mutual fund manager John Hussman, whose free weekly column is more valuable than any high-priced newsletter.
Using a century of data, there is a very strong pattern of declines from a simple set of conditions. I’ll let Hussman explain them:
In the chart below, the bars indicate points in the past 20 years where the following conditions were true.
- NYSE 52-week highs and lows both greater than 2.5% of total issues traded (composite highs work better than equity-only, because dispersion of interest-sensitive issues is often meaningful);
- New highs no greater than twice the number of new lows;
- S&P 500 greater than its level of 10-weeks earlier (some versions use the NYSE composite, but our index of interest is generally the SPX, and we are less interested in signals that might occur when the market is already down substantially);
- McClellan Oscillator (the 19-day minus the 39-day smoothing of daily advances minus declines on the NYSE) below zero, which is another indication of dispersion;
- Two signals within 36 trading sessions, which is helpful for reducing one-off noise.
Ominous? Not necessarily. Worth considering in the context of a much more troubling syndrome of overvalued, overbought, overbullish, rising-yield conditions? Sure.
Furthermore, the extreme in valuation (Shiller PE over 23 on record margins – Shiller PE on normalized margins would be closer to 29) means that regardless of the near-term outcomes, the returns of future years have been cannibalized. Because earnings grow remarkably steadily over the long-term (6% nominal is the average) stock market returns depend overwhelmingly on the price paid for a claim on these earnings. With the brief exception of the mid 1990s, 10-year returns have never been favorable when the S&P500 trades over 20x 10-year average earnings.
Furthermore, the Shiller PE appears to still be correcting from the late-90s bubble. We should welcome this trend, since valuations have in the past overshot to the downside. By the way, the previous examples of seriously undervalued markets all occurred after high inflation during the later years of secular bears (1917-1920, 1940s, 1977-1982).
A Shiller PE of 10 within a few years would fit nicely in this chart, though it would not be a pleasant experience getting there. Today, Shiller earnings are about $70.
Sentiment and price action are extreme. Combined, they make a very strong case for a rebound.
Gold stocks are also very cheap relative to the broad indices, as well as gold itself (which is also the most under-loved since it was $300 per ounce).
Images from stockcharts.com
Gold itself is still very expensive relative to other hard assets and financial assets on an historical basis, but trader sentiment is bleak, a contrary indicator. From 2001 to the top in 2011, we never had such a long down streak on such bearishness, which may indicate, despite its short-term bullish implications, that the bull market may have topped at $1920 per- in August 2011. We had parabolic moves on extreme bullishness in gold (peaking 8.11), silver (5.11) and platinum (3.08), and moves like that tend to burn up all reserves of excitement and mark tops for years to come.
We now have a syndrome that has rarely been followed by significant market gains, and almost always followed by a decline:
- Overvalued: Shiller PE above 18 (it’s 23)
- Overbullish: High “bullishness” and low “bearishness” readings in DSI, NAAIM, Investors Intelligence, VIX, etc
- Overbought: SPX near upper Bollinger bands (daily, weekly & monthly), over 50% above 4-year low, over 7% above 52-week running ave.
- Rising yields: Treasury bonds have recently declined, 10-year yields above level of 6 months ago.
- Developing choppiness: A smooth rise of less than 1% per day for several weeks, followed by several days of increased market volatility and sideways price action.
- Declining RSI: Indicates momentum is lost.
The first four of these points are the work of John Hussman* (though I may cite here different sentiment indicators or signs of an overbought market than he does). The last two are my own, and serve to identify a top with greater precision, as Hussman’s syndrome, while an excellent intermediate-term indicator, can be sustained for several weeks as the market continues to rise. When the last two are present, stocks rarely make further headway before declining, though they may churn sideways for weeks.
* Here is Hussman’s uber-bearish syndrome in his own words, with his own chart showing points in history when this has occurred:
- S&P 500 Index overvalued, with the Shiller P/E (S&P 500 divided by the 10-year average of inflation-adjusted earnings) greater than 18. The present multiple is actually 22.6.
- S&P 500 Index overbought, with the index more than 7% above its 52-week smoothing, at least 50% above its 4-year low, and within 3% of its upper Bollinger bands (2 standard deviations above the 20-period moving average) at daily, weekly, and monthly resolutions. Presently, the S&P 500 is either at or slightly through each of those bands.
- Investor sentiment overbullish (Investors Intelligence), with the 2-week average of advisory bulls greater than 52% and bearishness below 28%. The most recent weekly figures were 54.3% vs. 22.3%. The sentiment figures we use for 1929 are imputed using the extent and volatility of prior market movements, which explains a significant amount of variation in investor sentiment over time.
- Yields rising, with the 10-year Treasury yield higher than 6 months earlier.
The blue bars in the chart below identify historical points since 1970 corresponding to these conditions.
The average active manager is now leveraged long, according to NAAIM’s weekly survey:
Sustained bullishness is bearish, especially once the market starts to trend sideways. We don’t yet have that choppy sideways action on declining RSI that has been a death knell for rallies, but sooner or later it will emerge. If the market starts sidedays, this would complete the most bearish syndrome possible, though we already have a market that is overbought and overvalued, with overbullish sentiment and rising bond yields (John Hussman’s bearish syndrome that has nailed most major tops for decades).
In economic news, Q4′s negative GDP print supports the thesis that we entered a recession in the 2nd half of 2012, as leading indicators had been suggesting for months. It also comes right as the Citi Economic Surprise Index is again on the downward slope of its regular cycle, meaning surprises are more likely to be to the downside.
John Hussman does the best long-term statistical analysis of the broad equity market, bar none. He has identified a set of four conditions that has appeared at or just before significant tops in the stock market:
Overbought: S&P 500 within 3% of its upper Bollinger bands, at least 7% above its 52-week smoothing, and over 50% above its 4-year low
Overbullish: Investors Intelligence sentiment survey shows bulls above 52% and bears below 27%
Overvalued: Shiller P/E above 18 (it’s currently 23)
Rising yields: 10-year Treasury yields above their level of 6-months earlier.
This condition also appeared in 1929 (followed by a crash and 20 year bear market in real terms) and 1964 (stocks peaked in ’66 before going down 80% in real terms over the next 16 years). When stocks are overbought and overvalued, treasuries have fallen, and most investors are bullish, it is to your great advantage to eliminate market risk (sell your stocks or hedge them).
Chart from Doug Short, Advisor Perspectives:
Platinum has had a sharp rally in recent sessions, while gold has been only slightly higher, bringing the two heavier precious metals (specific gravity of 19.30 for Au and 21.45 for Pt) to parity for the first time since April. The ratio remains high, and has been elevated since gold left the more-industrial platinum in the dust in the summer of 2011:
This has been one of the longer periods of inversion in recent history. Selling gold and buying platinum in equal weights when the ratio is well over 1.0 has been a dependable money-maker for patient traders, as the ratio tends to revert to well under 1.0.