The market is a forward-looking beast and we are seeing some pretty strong signals from short-term and medium-term breadth indicators. The long-term breadth indicators, however, are still lagging and have yet to trigger. Today I will put a medium-term breadth model to the test and show how to improve results with a simple timing mechanism. This timing mechanism also offers a quicker exit alternative should the market turn south. I will also show a three indicator S&P 500 breadth model that matches buy-and-hold results with much better drawdowns and much lower exposure.
Note that I will update the “usual” breadth charts and models tomorrow.
FOMO and FOUP Hit the Markets
I get the sense that fear of missing out (FOMO) is hitting the markets and some of us. Even though the size of the rebound since mid March may seem irrational, we must keep in mind that markets can remain irrational a lot longer than we can remain solvent. This quote is usually attributed to Lord Maynard Keynes, but it seems to have originated from A. Gary Shilling. Regardless of its origin, it is a great quote that belongs on a post-it in the middle of our charts.
Mr. Market is usually not the irrational one at the table. Thinking that the markets are either rational or irrational is itself irrational. Or, as Spock might say: the market is fascinating, but hardly illogical. As technicians, we must simply weigh the evidence and act accordingly. Take bullish signals and increase when the market environment is favorable and avoid stocks and lower risk when the market environment is unfavorable. Furthermore, it is important to have a strategy and trade according to the rules of the strategy.
Fundamental reasoning can distort chart analysis, but I can see at least three fundamental reasons that could push the S&P 500 to new highs in the next few weeks (another 7% higher). Monetary policy is in full gear, the fiscal response continues to ramp up and the US is reopening. Thus, the economy and markets are getting juiced from all sides. Throw in fear of under-performing (FOUP) from under invested PMs and we have the recipe for further gains, maybe even a melt up.
Medium-term Breadth Thrust Strategy
There has been a lot of chatter on the medium-term breadth thrust, which is the percentage of stocks above the 50-day SMA moving above 90%. I decided to put this indicator to the test using Amibroker and data from Norgate. Instead of using 90%, I will consider a move above 80% as a bullish breadth thrust and a move below 20% as a bearish breadth thrust. I will use %Above 50-day SMA for the S&P 500, S&P MidCap 400 and S&P SmallCap 600 to form the indicator group.
Note that StockCharts has symbols for percent above 50-day SMA for the S&P 500 ($SPXR50A), Nasdaq 100 ($NDXR50A) and S&P 100 ($OEXR50A), but not for the S&P SmallCap 600 and S&P MidCap 400. There is no sense testing the first three as a group because there is too much overlap. All 100 stocks in the S&P 100 are in the S&P 500 and 82 Nasdaq 100 stocks are also in the S&P 500. In any case, a true breadth composite should be broad and feature mid-caps and small-caps.
The backtest period matters because the breadth indicators did not perform well during the bear market of 2001-2002. In general, there were bearish signals in September-October 1999 (way early), and there were some bad whipsaws during the decline (bad bullish signals). I should also note that the nature of breadth changed in January 2001 when the exchanges moved spreads from 1/16th of Dollar to pennies (decimalization). This greatly reduced the number of unchanged issues.
And finally, note that the values for breadth indicators from StockCharts do not always match the breath values from Norgate, even for the same indicator. I suspect that StockCharts calculates their breadth using adjusted data (with dividends) and Norgate uses unadjusted data (no dividends).
Oh, one more note. Ned Davis Research (NDR) uses 90% to define a bullish breadth thrust for the percentage of stocks above the 50-day SMA. A move above 90% is a serious breadth thrust, but there were no signals between 1/1/1993 and 1/1/2003. I am looking for an indicator that triggers more signals and also triggers bearish signals. For what it is worth, NDR testing shows good results for these signals. Here is a link to a tweet from @edclissold
Timeframe and Benchmarks
Given the change to decimalization on 29-January-2001, I am going to run this backtest from 1-February-2001 to 1-June-2020. For reference, the S&P 500 SPDR produced a 6.24% Compound Annual Return (CAR) during this period and there were three big drawdowns: 40% in 2002, 55% in 2009 and 33% in 2020. The Compound Annual Return includes dividends and the subsequent backtests also include dividends. Buying and selling SPY based on the 5/200 day SMA cross returned 5.8% per annum with a much lower Maximum Drawdown.
Starting in 2001 or 2003
Next we test the %Above the 50-day SMA for the S&P 500, S&P MidCap 400 and S&P SmallCap 600. The system buys when at least two of the three have crossed above 80% and sells when two of the three have dipped below 20%. The signals do not have to occur at the same time. We just need two “active” signals for a breadth signal.
The first line on the table below shows results from 2/1/2001 to 6/1/2020 and the second shows results when starting two years later (skipping the bear market and the bad signals). The Compound Annual Return (CAR) since 2001 is just 5% and the Maximum Drawdown (MDD) is a hefty 37%. Unsurprisingly, results improve dramatically when we exclude the first three negative signals. Not sure we can get away with that though!
The next image shows the buy and sell signals. The first indicator window is the %Above 50-day SMA composite (Thrust Signals). +1 means 2 of 3 are on bullish signals, +3 means all three are on bullish signals, -1 means 2 of 3 are on bearish signals and -3 means all three are bearish.
Improving Results with a Timing Mechanism
Overall, the performance for the %Above 50-day SMA thrust is not so great because the Compound Annual Return is low and the Maximum Drawdown is high. There is, however, a simple solution to reduce the drawdowns. Buy when the breadth indicator is net bullish AND the 5-day SMA for SPY is above the 200-day SMA. Sell when the breadth indicator turns net bearish OR the 5-day SMA for SPY moves below the 200-day SMA. Selling when the 5-day SMA moves below the 200-day SMA will insure an exit before a decline gets out of hand.
The table below restates results without timing on line 1 and then shows results with this simple timing mechanism. There were just two extra signals and they were both losses (whipsaws). However, the Compound Annual Return increased to 6% per year, the Maximum Drawdown fell to 17.2% and the Profit Factor exceeded 5. This is a huge improvement.
The next chart shows the current situation for this breadth indicator trio and the 5/200 cross. Breadth triggered bullish on May 18th when two of the three breadth indicators had exceeded 80%. The 5-day SMA of SPY crossed above the 200-day SMA on May 27th. Thus, this system is bullish until the breadth composite turns bearish OR the 5-day SMA crosses below the 200-day SMA.
A 3-Indicator S&P 500 Strategy
I realize that not everyone has access to all these indicators so I put together a simple strategy using three StockCharts indicators for the S&P 500: %Above 50-day SMA ($SPXA50R), %Above 150-day SMA ($SPXA150R) and %Above 200-day SMA ($SPXA200R). This strategy is more in line with my Index Breadth Model because it has two long-term indicators. It turns bullish when two of the three indicators are on active bullish signals and bearish when two of the three are on bearish signals.
The image below shows these indicators from StockCharts with signals over the last five years. Note that I softened the bullish and bearish thresholds as the moving average periods increased. 70% and 30% are the bullish and bearish thresholds for %Above 150-day SMA. 60% and 40% are the thresholds for %Above 200-day SMA. This indicator group is still net bearish because neither one of the long-term indicators has triggered bullish.
The next table shows backtest results using Norgate data, which differs from StockCharts data. Again, this is probably because StockCharts uses dividend-adjusted data and Norgate uses unadjusted data. The results with the market timing mechanism are the best so far with a 6.3% Compound Annual Return and a Profit Factor above 7.
Why Bother with Market Timing?
So you might be thinking why even time the market when the Compound Annual Return for the last strategy was 6.29% and the Compound Annual Return for buy-and-hold is 6.24%? Well, if you can stomach the drawdowns, be my guest. Also note that the strategy achieved the 6.29% return by being invested just 67% of the time. Buy-and-hold, in contrast, was invested 100% of the time with three large drawdowns along the way. The journey matters as much as the end result.
The chart above shows the equity curve for the SPX Trio strategy with the timing mechanism in green and buy-and-hold for SPY in blue. They both end up at similar levels, but their journeys are much different. Buy-and-hold had negative equity in 2002-2003 and again in 2008-2009. The breadth strategy never went negative and equity hit a new high by the end of 2009. Also notice that the strategy outperformed buy-and-hold most of the time and the drawdowns were much more tolerable.
Conclusions and Signals
I use a market timing model to tell me when to be long stocks and when to be out of stocks. The market environment is favorable when the breadth models are bullish. This is when I look to increase exposure (risk) with bullish setups and bullish signals. The market environment is unfavorable when the breadth models are bearish. This is when I look to raise cash and look for stock alternatives.
At this point, the 5-day SMA for SPY is above the 200-day SMA, the short-term breadth thrust indicators are bullish and the medium-term breadth thrust indicators are bullish. The 10-day EMAs of Advance-Decline Percent are the short-term breadth thrust indicators. The long-term indicators are still bearish though (%Above 150-day SMA, %Above 200-day SMA and High-Low Percent). The evidence is about as mixed as it could be, but this could change with another rip-roaring day.
Adding the 5/200 day SMA cross as a simple timing mechanism can reduce the pain of a whipsaw and capture gains should the trend extend. As with all trend-following indicators, we never know which signals will result in extended trends and which will result in whipsaws. It could be March 2019 (extended trend) or it could be October 2007 (whipsaw).