Bearish…With a Stop (Maybe)
March 30, 2016
- S&P 500: 2063.95 (3/30/16 close)
- 200dma: 2016.23
- 50dma: 1950.51
- Price/200dma: 102%
- 200dgr: -1.54%
- Schiller PE: 23.1 times (10y median EPS 89.29)
- S&P 500 Fair Value: 1562.58 (17.5 times Schiller PE)
- Valuation – EV to EBITDA and Median Price to Sales at record highs
- Positioning – Household/individual equity allocations at ‘peak cycle’ levels
- Global Economic Conditions – Global recession probability over 80%
- US Economic Conditions – US Initial Jobless Claims at lowest level since 1967, indicating business cycle maturity
- Monetary Conditions – Credit spreads in cautionary zone and in uptrend; fundamentals at or past peak
- Market Model – On SELL since 2/23/16
- 200-Day Growth Rate – S&P 500 200-day moving average ‘200-day growth rate’ (200dgr) on SELL since 3/11/16; minimum decline target of 15% from SPX 2022.19, or 1718.86.
- Valuation – Historically low interest rates justify high PEs
- Positioning – Household/individual equity allocations have room to rise to 2000 record highs
- Global Economic Conditions – International economic conditions may have bottomed in 1Q16
- US Economic Conditions – Weight of the evidence points to a ‘jump ball’; bulls have slight edge due to still rising employment-population ratio
- Monetary Conditions – Recent rise in spreads and fundamental deterioration indicative of cyclical downturn; cyclical international economic upturn + more robust US economic growth drive tightening
- Market Model – First failed signal since 1980’s (fifth SELL since 2000)
- 200-Day Growth Rate – Signal fails (decline from SPX 2022 less than 5%), market rallies at least 15% from 2022.19, or 2325.52.
Valuation and positioning – key long-term sentiment indicators – are terrible timing tools. In the near- to medium-term, they matter very little when the weight of the evidence is overwhelmingly bullish or bearish. They ‘only’ tell you the possible extent of a long-term market move. Take 2012 for example. The S&P 500 closed 2012 with a median price to sales ratio of around 1.8 times – very close to the all-time high set in 2007. The market went on to peak at over 2.1 times in early 2015, dragging the market from 1402.43 to 2110.30 on 2/20/15. Today on the other hand…
With two caveats, I believe the S&P 500 currently sits at the worst possible juncture: extreme overvaluation/over-allocation with the weight of the evidence decidedly bearish.
Caveat #1 is the US economy. Historically, 20%+ market bear markets occur while US Jobless Claims are rising 10-20% YOY. At present, Jobless Claims are falling 12% YOY. We could be more mid- to late-business cycle than very late, which would be a major blow to the bear case in the near- to medium-term. But while it typically pays to wait for the 10-20% level to kick in before getting bearish, three things. One, the global economy is just as important (if not more) to S&P 500 aggregate earnings power as the US – and global economic conditions, ex. US, are bleak. Two, at the end of January Jobless Claims were down just 1.6% YOY. And three, as a % of the civilian labor force Jobless Claims have never been lower. In the context of rising spreads, low corporate investment confidence, and poor organic growth, my conclusion is that Jobless Claims are set to surprise to the upside in the coming months and quarters.
Caveat #2 is near-term market ‘breadth’. The strength of the rally out of the 2/11/16 low has been impressive. Approximately 90% of stocks traded above their 50-day moving average; 10-day advancing volume outpaced declining volume by more than 2 to 1; and the McClellan Summation index recently hit the bullish ‘1,000’ level. But this bullish short-term breadth has yet to be confirmed by more durable, longer-term breadth indicators such as the % of stocks above their 200-day moving average.
In 2011 the market bottomed at approximately 88% of its rising 200-day moving average (11% 200 days over 200 days). In the initial rally off the bottom the % of stocks over their 200dma hit approximately 50%.
On 2/11/16 the market bottomed at approximately 90% of a 200dma that was up just .6% 200 days over 200 days. The 200dma is now falling at a rate of more than 1%, and the market has rallied more than 12%. This set-up is a relatively easy breadth hurdle for the market to clear, yet as of the 3/29 close just 51.4% of stocks are above their 200dma. This is objectively weak breadth, and why, despite setting a 65% ‘stop loss’ hurdle on my bearish stance, I say ‘maybe’ with regard to using it. This market, in my opinion, has a lot to prove if in fact we have entered a new cyclical bull market.
As I have explained ad nauseum in various posts and investment letters, I utilize a ‘weight of the evidence’ approach to market analysis that I call the Market Mosaic. In summary, it is comprised of various economic, monetary, valuation and technical indicators that I assign various weights to (composite score of -100 to +100). In detail it is as follows:
- Market Model (40%) – Proprietary composite of valuation, momentum and moving average indicators
- Market Technicals (25%) – Price/200dma and 200dma Direction (trend + growth rate, equally), 12.5% each
- Economic Conditions (20%) – Global recession probability model (5%) and US Jobless Claims (15%)
- Monetary Conditions (15%) – HY spreads (trend + level, equally)
(Disclosure: I have reconfigured the Mosaic a bit to include the economic conditions component. The weight of the Technicals component was reduced accordingly.)
No model is perfect – obviously. And as with everything in the investment business there is a fair to high amount of ‘art’ involved. That said, I believe the Mosaic does a reasonable job of getting the ‘directional’ correct. For example, when I started 17 Mile on 6/15/14 the Mosaic was in maximum bullish mode of +100, and the market went on to rise from 1936.16 (6/13/14 close) to 2134.72 in mid-2015. There was volatility from point to point, but the model got me in line with the bulk of the trend.
The ‘art’ is supplementing the Mosaic with various third-party shorter-term sentiment and breadth indicators. For example, despite the relatively bullish configuration of the Mosaic in August 2015, the market had begun to ‘act’ poorly, and this was showing up in several non-Mosaic indicators. I reduced exposure accordingly in mid-August.
In near complete opposition to 6/15/14, the current Mosaic composite score is -41.25.
As in August 2015, third-party indicators could override this bearish Mosaic reading. But as noted in the Discussion section above, the market has a lot to prove.
Valuation. For its long-term relevance, it is rather amazing how much controversy broad market valuation garners. Bulls point to the fact that the S&P 500 has traded over its ‘fair value’ for the better part of two decades as proof historical valuation levels are no longer relevant; while bears point to the fact that we had an unprecedented two 50% bear markets in ten years as proof historical valuation is in fact highly relevant. I believe both camps are wrong.
Bulls are wrong to ignore the fact that two 50% bear markets in ten years is unprecedented in the post-WWII era (or to ignore the eerie similarities between today and the 1929-1937 period where there were two massive crashes following extreme over-valuation) and thus likely due to the power of valuation. Bears are wrong to ignore the fact that valuation is next to irrelevant when economic, monetary and technical conditions are favorable.
With the weight of the evidence now in its favor, I believe valuation’s time to shine has arrived. Courtesy of Ned Davis Research, Domestic Nonfinancial EV to EBITDA, and the S&P 500 Median Price to Sales:
* * * * * * * * * * * *
Bull Case: In my opinion, the most credible valuation point in the bulls’ favor is the level of interest rates. Naturally, with low returns available via alternative financial assets the market moves up a risk level in search of a more normal return level. But I have yet to see a credible correlation analysis for PEs and rates. If you plot PEs alongside rates in the following chart of Moody’s Corporate Baa Yields, you will notice that PEs expanded in the 60’s alongside rising rates, while they have declined since 2000 on falling rates. Me thinks rates may have a thing or two to do with economic growth and inflation. Mr. 30-Year Bond at 2.66% may have a thing or two to say about the implied growth rate embedded in near-all-time-high EV/EBITDA multiples.
Positioning. As with valuation, investor positioning must be looked at within the context of the ‘weight of the evidence’. A fully positioned market does not need to roll over if the economic and monetary tailwinds are strong, and if the trend is firmly upward. As the following charts demonstrate, investors’ allocation to equities reached near-record highs in the last two years, but have begun to roll over. Within the context of questionable economic conditions, deteriorating monetary conditions, poor breadth, and a well-defined downtrend, this ‘roll over’ is bearish. Federal Reserve Equities as % of Total Assets, and AAII Equity Allocation (both courtesy of Sentiment Trader):
* * * * * * * * * * * *
Bull Case: If the global economy picks back up, the US economy is closer to mid- to late-cycle, and the Fed remains on hold, it is tough to argue that equity allocations do not have room to move up closer to the all-time highs set in 2000.
Economic Conditions. Unfortunately I cannot post a global recession model chart, as it is proprietary to the provider; but at present the reading is 89%, and has been materially north of 80% for approximately six months now. Historically, the global economy has an 83% chance of entering recession when the model reads north of 70%.
One of the best leading indicators for US economic conditions is the YOY change in the 4-week moving average of Jobless Claims. This indicator has yet to warn of any oncoming US economic weakness. Full history and trailing 5-year charts (YOY deltas):
Before jumping to the conclusion that the global economy is out of the woods due to a non-recessionary US economy and a potentially improving international economy on the back of a not-so-strong USD, I would like to highlight the following.
(1) As a % of the civilian labor force, Jobless Claims have never been lower in absolute terms, at .17%. And as a % of the former high – 41.1% – only 1989’s 40.4% was lower.
(2) Until February 2016, the YOY Claims delta was in fact declining, at just -1.5% YOY on 1/31/16. I will explain the similarities between now and 1973 in greater detail later, but just prior to a US recession and large-scale bear market, Claims were at a similar YOY level (prior to the February plunge in YOY rate). I am not calling for a 1974-style bear market, but simply pointing out that near the end of a business cycle, low rates of YOY Claims growth fail to act as a ‘leading’ indicator. Looked at in conjunction with the record low Claims to Labor Force percentage and deteriorating credit cycle, I believe Claims are set to rise in the coming months and quarters.
(3) The Wells Fargo Economic Group believes there is a 25% chance of a US recession in the next six months based on their ‘Probit’ model. While not as high as prior to the 2000 and 2008 recessions, it is materially higher than any readings since the Great Recession.
* * * * * * * * * * * *
Bull Case: If the USD has topped, and monetary stimulus is sufficiently stimulative, international economic conditions very well may have bottomed in 1Q16. Regardless of international economic conditions, however, the key to the bull case is that the US does not enter a recession, as even a mild recession could ignite a large-scale bear market (i.e. the 2000-2002 recession was quite mild) given the extreme valuation and positioning. I believe the state of US economic conditions is a bit of a ‘jump ball’ at the moment, perhaps between Shaq (bulls) and and Kobe (bears…but with hops). In the bull camp is the fact that the Civilian Employment-Population Ratio continues to rise (chart below). The case can be made (and is made by Ned Davis Research) that we are in a secular bull market similar to the 1980’s to 1990’s that sees a large-scale influx of high-earning employees into the workforce, driving big gains in earnings, profits and thus stock prices. Tough to predict, but the following chart will likely need to roll over at least for a bit to get any real downside traction in equities.
Monetary Conditions. BB and High Yield Master option adjusted spreads, while down from January/February highs, remain in a medium-term uptrend. While not an absolute concern, since 1996 current levels have led deteriorating economic conditions; and within the context of a ‘peaky’ business cycle, I rate monetary conditions neutral to favorable. Also included below is a CDS Index (courtesy of Sentiment Trader), that shows similar level/trend readings.
Again – as with economic conditions, given extreme overvaluation and investor positioning, I do not believe a large-scale bear market requires a 2008-like meltdown in credit markets.
Regarding the fundamentals underpinning the credit cycle, the BAML High Yield Group has stuck by its bearish stance in the face of massive retail inflows since the surprisingly dovish March FOMC meeting. They argue that non-commodity HY has had an unhealthy 80% correlation with oil YTD, which they believe demonstrates the unfounded ‘technical’ nature of recent market movements. They believe that fundamentals have unequivocally rolled over, and that the HY market could face a steep decline once retail money attempts to exit en masse.
* * * * * * * * * * * *
Bull Case: The bull case on monetary conditions is that this recent spike in spreads, and downturn in non-commodity fundamentals, are indicative of a cyclical bottom. This would align with a fresh international cyclical upturn, and more robust US economic growth.
Market Model. From my analysis of market history it appears the nature of the market has changed since the 1991 recession/bear market. It appears to move in much longer, smoother cycles. I am unclear if this is due to longer business cycles, the decline in rates, or the Fed ‘put’. What I do know is that long-term oriented models such as the Market Model and the 200dgr indicator have performed brilliantly since 1991, with highly accurate long-term readings. Until I studied the history of the 200dgr, I chalked it up to the depressed valuation readings in the 70’s and 80’s, as the Market Model was largely on BUY from the early 70’s thru late 80’s, even thru the ’74 bear market. But the 200dgr itself has produced only three SELL signals since 1991: March 2001, May 2008, and March 2016.
The Market Model has several high frequency components that will respond to medium-term changes in market conditions, so it has been a bit more volatile than the 200dgr indicator since 1991. Accurate, but volatile. Some history:
- BUY/HOLD from 1991 thru a 2000
- SELL/HOLD 2000-2003
- BUY/HOLD from 2003 thru early 2008
- SELL/HOLD from early 2008 thru late 2008
- BUY/HOLD from late 2008 thru early 2010
- SELL in early 2010
- BUY in mid-2010
- SELL in early 2011
- BUY in mid-2011
- HOLD from 2012-2013
- BUY from early 2014 thru March 2016
The Market Model is guaranteed to fail at some point. Perhaps now is that time. But when I line up the weight of the evidence – firm downtrend in place (as determined by the 200dma 200-day growth rate), ‘peaky’ equity and business cycles, poor international economic conditions, worsening monetary conditions, and poor broad market breadth (% of stocks over 200dma) – I believe the current SELL signal warrants significant attention.
* * * * * * * * * * * *
Bull Case: There is no bull case with regard to this indicator. Since 1991 there has not been a failed reading. And prior to 1991, ‘failed’ SELL signals simply resulted in a flattish market for an extended period of time; and at that, those SELL signals could have been overridden by other breadth, economic and monetary data.
200-Day Growth Rate. The 200dgr is a component of the Market Model, but when I created the Mosaic I broke it out as a component of the larger 200dma ‘Direction’ component. Until recently, I had not studied its history as a pure standalone indicator. As with most investment-related exploration, I happened down the rabbit trail.
* * * * * * * * * * * *
Since calendar year-end 1929, the S&P 500 has seen 21,661 trading days (as of 3/24/16 close). Only 22 of those trading days (including a 3/11/16 signal) have seen the SPX’s 200-day moving average 200-day growth rate (200dgr) breach the -1% level. Twenty-two.
The 200dgr is simply the current 200dma versus where it stood 200 days ago, in growth rate form [(200dma present / 200dma 200 days ago) – 1]. It is a slow-moving, long-term oriented technical indicator designed to detect the primary market trend in place. While only one indicator out of hundreds available, its standalone effectiveness is striking. And while there are certainly failures (as I will discuss), as with the Market Model, when the weight of the evidence lines up, it warrants close attention.
Each period is measured from the first time the ‘-1%’ level is breached on the downside to the first time the ‘+1%’ level is cleared on the upside.
- 1930’s: 4
- 1940’s: 3
- 1950’s: 2
- 1960’s: 4
- 1970’s: 2
- 1980’s: 3
- 1990’s: 1
- 2000’s: 2
- 2010’s: 1 (3/11/16 signal)
By drawdown (3/11/16 not included):
- 0-5%: 7
- 5-10%: 2
- 10-20%: 5
- 20%+: 7
(1) First and foremost, this indicator always triggers during weak market periods (obviously), so when it is wrong it is wrong big. It needs a stop loss. Returns from signal date for drawdowns of 0-10%:
- 1939: 3.2%
- 1949: 22.5%
- 1953: 17.9%
- 1960: 19.7%
- 1962: 20.4%
- 1966: 22.9%
- 1984: 12%
- 1988: 16.8%
- 1990: 19.6%
The minimum decline from signal date during these periods was .9%. The current signal has produced a .31% decline (two days after the 3/11/16 signal date), which would be the smallest decline from signal on record since 1930.
(2) For the most part, 10%+ declines are concurrent with Jobless Claims rising YOY and/or weak ‘breadth’ readings (i.e. the McClellan Summation). But four problems with that observation:
- Claims and Summation data only go back to the early 60’s
- The 1990 signal flashed with Claims up 30% YOY
- The 1973 signal flashed with Claims down 4% YOY
- The S&P 500 is far more sensitive to global economic conditions now than it was prior to 1990
(3) As mentioned in the economic conditions segment, the similarities between today and the 1973 200dgr SELL signal are striking. 1973 saw Claims falling 4% YOY and a Summation of over 800. Today, Claims are falling by more than 10% YOY, and the Summation is a hair over 1,000. 1973-1974 and 2016 trailing 3-year Summation (courtesy of Sentiment Trader):
(4) As mentioned in the Market Model section, the nature of the market appears to have changed since 1991. Including the 3/11/16 signal, there have been three signals since 1991. Three in 25 years! From 1930 thru 1980, there 2-4 signals per decade.
Perhaps we are breaking out of the ‘new market era’ and the 3/11/16 SELL signal will be dead wrong. But given that the high valuation/low and declining interest rate/activist Fed policy condition set remains in place, I would not, and am not, betting that it is wrong. I will let longer-term breadth signals ‘stop’ me out of a bearish stance.
* * * * * * * * * * * *
Bull Case: We have either seen the lows, or will decline by no more than 0-10% from 2022.19 before rallying more than 20% from the ultimate bottom.
Stop Loss. The ‘maybe’ stop-loss on my bearish stance is 65% of stocks trading over their 200dma. I say ‘maybe’ because I would like to see this confirmed by a higher reading of the percentage of stocks with 50dma > 200dma (source: NDR).
In Summary. The rally out of 2/11/16 is either a massive bull trap (‘massive’ due to the high Summation reading), or the beginning of a big cyclical bull market. In my opinion, the weight of the evidence points to the former.