Can Monthly Crossovers Help Predict Turning Points?

You knew it would take some major market mayhem to bring the Yinzer Analyzer out from hiding but with the markets on the cusp and with investors running for the exits, he’s starting to wonder if now is finally the time to be looking to switch to a defensive playbook. We’ve talked a lot in the past about how this market’s major turning points have revolved around the Fed and its various QE programs and while taking down the beta before the next FOMC meeting seems like a good idea, the Yinzer Analyst is all about timing. Having been too early to call the end of the bull market before, I can tell you that having the right call and being early is the exact same thing as being wrong. So to make sure that this time, it really is different, I’m dusting off one of my favorite technical trading rules, the 2/10 monthly crossover rule (simplifying it as 2/10) to see whether or not the long promised pullback is finally upon us!  And with it shooting off signals like fireworks at a redneck Easter, the time to turtle up might be nigh.

The Rule:

The 2/10 was one of the first indicators I learned when I started my career in asset management and it’s about as basic an indicator as you could ask for. When used correctly it can offer solid risk-adjusted returns but investors are rightly wary of any technical indicator that promises easy returns thanks to back tested nightmares like the Hindenburg Omen. Unlike the Hindenburg, the 2/10 crossover doesn’t rely on multiple indicators or complicated formulas. To see the rule in action, look at this chart of the S&P 500 from the beginning of 1995 to the end of 2009 where the 2 month simple moving average is in blue, the ten month in red. According to the rule, when the 2 month crosses the 10 month from above, it signals a reversal in momentum meanings investors should be wary and when it crosses from below, you should be getting ready to buy with both hands.

95-2009 OV

At heart, the 2/10 is simply another way of expressing momentum like many of the other indicators we’ve talked about and can be used with data in any time frame but when looking at daily or weekly charts, it’s likely to throw off a tremendous number of buy and sell signals and investors can get whipsawed with heavy trading fees. That’s why for long-term investors it’s best to use the rule only with monthly data to help determine whether you should be shifting to offense (buying the dips) or defense which could include holding on to excess cash, looking at defensive equities or even adding inverse positions to your portfolio to cushion the potential downside. And while it seems so simple, let’s take another look at that chart. You can see that while the 2/10 seems to almost perfectly time the tops and bottoms of the market, it also throws off a bunch of false positives like it did in both the late 90’s and the mid 2000’s bull markets. Even if investors using the rule went right back into equities once the all clear was signaled, any investors who used a 2/10 crossover for market timing were left with major performance issues thanks to transaction costs.

Does that mean investors are doomed to suffer through volatility or instead wind up with performance like something out of a Dalbar study? The Yinzer Analyst is going to try some modifications of the 2/10 rule and show you that with some basic modifications, this most basic of indicators can deliver some real value to your portfolio.

The Experiment:

Men have long been ruled by a very simple philosophy of “screw with it till it’s broke” because no matter how perfect or well-crafted something might be, it can surely be made better by tweaking it just enough to guarantee it’ll never work the same way again. While the 2/10 crossover might seem to be the pinnacle of trading perfection, the only way to put it to the test is to generate signals to study so the first step in the experiment is the easiest…we’re going to pull it apart and try to reassemble it. Don’t worry; we’ll take tons of pictures with my phone.

Actually, the first step was downloading the daily return history for the S&P 500 from Yahoo Finance and then calculating rolling simple 2 and 10 month moving averages so we can be more precise as to when the crossovers happen. Readers who compare my figures and charts with Stock Charts will notice a slight discrepancy between my moving averages and there’s that could be around a half percent difference. Say that last trade day was August 7th, according to Stock Charts the 2 month moving average would be the close on the 7th plus the close on the last day of July divided by 2. So even if there’s only one trade day, it would be considered the monthly close. My data uses a rolling 2 months divided by 2 or the rolling 10 months divided by 10; little more cumbersome but it has the advantage of making it easy to check the markets history to determine when exactly a crossover occurred, how long did it last for and most importantly what were the results. But because Stock Charts makes much nicer looking charts, we’re going to keep using there’s instead of relying on Excel.

And what’s the hypothesis that we’re testing? That the 2/10 Crossover rule ISN’T an effective system for signaling when the market is about to run into periods of volatility that could spell trouble for your portfolio. Fortunately for yours truly who’s been sitting on cash waiting for a chance to buy some ProShares Short S&P 500 Fund (SH), the results show that it can add value when used correctly!

The Results:

Glancing at the results of the study would seem to confirm every bad feeling some investors have about using technical indicators. From the beginning of 1990 through August 7th 2015, there were 1689 days were the 2 month moving average was beneath the 10 (including the last three trading days), a little more than 26% of the time and we have forward returns for the 30, 60, 90, 180, 360 day periods for 1686 of and believe it or not, the returns are typically positive whether using the mean or median. In fact, the forward one year returns are positive 62% of the time with a median return of 11.53%. Would seem that jumping in and out of the market based on using the 2/10 crossover would be a sure fire way to lose a lot of money until you do a deeper dive into the data.

And while the 2/10 does fire off a warning signal without the market pulling back on a somewhat frequent basis if you use daily data, sticking with monthly provides a clear pattern that investors can use to minimize risk. Generally in each bull cycle, the 2/10 will send out three signals over the life of the rally, with two false signals before a final and correct signal is provided. By studying the last three bull cycles (including the current one that began in 2009) we can learn what signals and supporting data to look for to better time buys and sells.

The First Stage:

Look at this chart of the S&P 500 from 1993-2003 for a better example where the market sent off two false signals before the great capitulation in 2000. The first one was in March of 1994 after a steady advance of nearly 21% from the fall of 92 to February of 94 had pushed the market nearly to within overbought territory and the trading had become choppy as investors turned cautious. If you sold the index the first morning after the indicator flashed a sell signal and then bought it back after it flipped to buy, you would’ve missed a 2.1% gain which palled in comparison to the 21% gain you would’ve made just by being 100% long the S&P 500 for the next year after you got back in!

93-03

Intermittent pauses after strong advances are common as you can see from this chart showing 2003 to the end of 2009. If you waited until the 2 month crossed the 10 month from below and held for a full year, you would’ve made 25% but by June of 2004 the easy and early money had been made and investors were feeling cautious, leading to more selling pressure in July and August.

03-09

A similar pattern emerges in 2010 when the S&P 500 lost its footing as QE1 came to an end.

09-15

Thanks to market history repeating itself, the general observation that we can take from studying these three cycles is this: volatility is high coming off the end of a bear cycle and even minor profit taking can send the 2 month moving average tumbling through the 10 month that’s just beginning to recover. In all three cases, the long term 14 month RSI was still well outside the overbought reading of 70, the purchasing power oscillator was still positive and the 20 month McClellan showed accumulation. In short, there was no supporting evidence to indicate that the market was either slipping into or had never left a bear market.

In both cases, the forward returns during the crossover events in even the 60 day period were generally positive while the forward 360 returns were typically double digits. Remember that momentum trading strategies are inherently backward looking so in the early stages of a bull market, a 2/10 crossover signals profit taking and generally only fires AFTER the worse of the selling pressure. The signal is still important, but more to indicate that buying pressure has been easing and setting the stage for smart investors to pick up shares at better prices. In 2004 the returns even in the next 30 days were generally positive.

The Second Stage:

The 2/10 crossover typically fires off a second time as a further warning after a strong advance to indicate that some long-term bulls are starting to grow fearful of the easy profits they’ve made and coinciding with the Oppenheimer Framework for evaluating market cycles. You can see a clear example of it in 1998 just after the S&P 500 first crested 1100 after making over 100% in less than three years. The market began moving sideways but began to lose ground after hitting an intraday high on July 17th before a steep sell-off in the last week of August fired off a 2/10 crossover. The market opened down 70 points lower at the open on 9/1 (957.28), the first day after the crossover and if you didn’t buy back in until the 1st day after

it crossed back again two months later, you missed out on an 18.65% gain. Investors were afraid to be left behind after that and although the 2 month slide below the 10 several times in 1999, it never stayed there for more than a few days and always resulted in another move higher.

The animal spirits weren’t as strong in 2004-2007; the market’s enthusiasm had more to do with surviving the tech crash and the early stages of the war on terror, not to mention Fed largesse through lower interest rates. Investors using the monthly rule would’ve missed out on October of 2004 (but the indicator fired again for a few days anyway) but after that would’ve been in the market until the end of July 2006 and made another 13% for their troubles. Not that the indicator didn’t fire, it did on numerous occasions throughout 2005 and 2006, but generally intra-month and only for a few days at that but after the sell-signal ended in August of 2006 investors would’ve been in the market up until almost until the very end as the next extended crossover didn’t take place until November of 2007 and didn’t flip back to buy until May of 2009. And it wasn’t like that final crossover of that bull cycle came without warning; the sell signal was sent off several times in August and September of 2007 indicating that trouble was ahead.

In both cases you witnessed several other indicators flashing trouble around the same time, a pattern than was repeated in late 2011. In all three cases the RSI (14) suffered a multi-month pullback while the purchasing power oscillator pulled back into negative territory and the CMF score at least flat lined as the market closed at or close to the lows of the month indicating that selling pressure had begun to rise. The current bull market experienced Stage 2 in late 2011 as the launch of Operation Twists duration adjustment could only slightly life the negative market outlook following the end of QE2. Like the prior two bull cycles, the adjustment was short-lived as rumors about QE3 began to percolate through the market and the smooth transition from Twist to the new program prevented another break in the action.

The Third Stage:

Where the 2/10 crossover rule really pays off is in the third stage, when the market has gone from being way too suspicious of its overwhelming success to starting to believe its own hype. It’s around that time when hair dressers tell value investors they don’t get it and index annuity products start taking over a big chunk of the market. Fortunately it’s easy to spot when you’ve reached the third stage:

  1. Longevity: It’s called the third stage because it comes last! Generally there has to be a strong multi-year uptrend that’s pushed the market to new highs and well into overbought territory.
  2. Momentum has clearly rolled over: Measuring momentum by some other indicator like the Purchasing Power Oscillator or simple RSI shows a clear uptrend followed by a steady downtrend. Both the PPO and RSI feel in 1998, made up some lost ground in early 1999 before steadily weakening after that. A similar picture emerges in late 2007 with a larger gap before the second and third stages.
  3. Distribution: I like to use the Chaikin Money Flow because it’s easy to interpret whether buying or selling pressure predominates. I use a long period, 20 months, and you can see in late 1998 that the reading gets close to perfect 1, meaning that the market had closed at or very near the high EVERY month for 20 months. After the second stage crossover, investors get anxious and begin to sell almost two full years before the bull ends. In 2007-2008, the distribution is more coincidental with the loss of momentum and the crossover.

So where does that leave us now? Call it an exercise in simple diagnostics.

  1. This bull is very much long-in-the-tooth and with earnings growth now negative and rate hikes imminent, it’s hard to support this valuation.
  2. Momentum has rolled over: RSI is backing off from oversold levels while the PPO went negative in late 2014. And after today’s performance (8/20), the Yinzer Momentum model is back to it’s lowest levels since the start of the Evans Rally last fall.
  3. Distribution: Hasn’t reached epic proportions yet, but the recent monthly closes have been getting progressively weaker.

I’ve been writing this post in stages on and off for the last few days and since the 2/10 crossover rule first fired (using daily data) on July 8th, it’s gone off again 4 times. First was for three consecutive days from 8/5-8/7 then again on 8/20.  Even the traditional formula calculations from Stockcharts.com shows the 2 and 10 month moving averages are just a point apart! Baring a major development from the Federal Reserve about putting monetary tightening on hold, it seems highly likely that August will see the first close of the 2 month below the 10 since late 2012? What do you plan to do when the next bear comes?

The Yinzer Analyst plans to hold onto his small inverse position and add accordingly but a future topic will be exploring what sectors or investments have held up in the past and could possibly do so again in the future.

And thanks again for checking back in, I promise I won’t let this go again!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s