Below I will show the results of a backtest of a "risk off" timing strategy using SPDR S&P 500 Trust ETF (NYSEARCA:SPY) from 1993-2016 with a cash position when out of equities. The buy/sell signals are based on the crossovers of the SMA 50/200 (the 50-day and 200-day simple moving averages), popularly known as the "Golden Cross" and "Death Cross." The thesis is checked by testing monthly data from 1953-1973 and 1973-1993. This is the first in a series of articles on this kind of strategy.
A Touchy Subject for Investors
I'm an avid student of mathematics and also the son of a financial advisor. During the early 2000s, I naturally overheard several conversations with clients that all roughly followed the same script. The key refrain always was simply that, "You can't time the market."
Coming off the long bull run of the 1980s and 1990s, any investor certainly could relate to that thesis. Investors who did anything but "buy and hold" felt very little but pain over lost gains as they watched the market march on from any correction within a few months. The thesis seemed justified.
The 1930s might as well have been ancient history for all the impact they had on investor thinking. The long upward march of U.S. stock market history seemed like it might go on unimpeded forever, but the major bear markets of the early 2000s and 2008 proved otherwise. The story of Japan's "lost decade" also provided investors with some reason for caution.
The market timer, however, still hasn't fully shaken his bad reputation, especially because the financial news cycle seems to operate at hypervelocity as compared to the actual business cycle. As the old joke goes, economists have predicted 8 of the last 2 bear markets.
For those willing to have the patience to go through several "selling fire drills" after "false alarms," as long as they have discipline and a clear strategy, it is essentially a matter of time until history rewards them with outsized long-term performance from the avoidance of a large drawdown. This should make such strategies of interest not only to the risk-adverse but also to those who have long periods of accumulation still ahead of them.
SMA 50/200 Strategy with SPY and Cash, 1993-2016
A simple moving average is a lagging indicator, taken from the average of a certain number of market closing prices. Its primary purpose is to smooth out data to reduce noise and reveal trends. The lag is also a feature for the SMA 50/200 indicator, as the "crossover" of the 50-day average above the 200-day average indicates that the market had started an uptrend, while the crossover of the 50-day average below the 200-day average indicates that the market had started a downtrend. Naturally, such a trend does not always continue and can reverse quickly.
(Chart from 11/11/1993 to 1/15/2016. Copyright © Peter Kirby.)
This simple backtest was performed on the daily dividends-reinvested SPY adjusted close data, from 11/11/1993 to 1/15/2016. The Buy & Hold investor is left with $60,938 for an annualized return of 8.48% with a 55% maximum drawdown. The SMA 50/200 long-term trend-following investor is left with $81,312 for an annualized return of 9.9% with 19% maximum drawdown.
One of the biggest risks when backtesting strategies is that one has overfitted the strategy to the data tested, resulting in imaginary returns that could never be repeated. Another risk is that the time series chosen has been cherry picked and that different economic conditions dramatically affect the results.
There are a few main ways to mitigate these risks. One is to ensure that the strategy is simple, with few factors (or just one), as every economic variable watched creates more opportunities for curve-fitting. Another is to ensure that the strategy is robust and that adjusting the parameters slightly does not dramatically affect the backtested returns. A third way, possibly the most important, is to "forward test" or, generally, to use data outside of the sample used when creating the strategy, making sure to include data from a range of different historical periods.
Is It Simple And Robust?
The SMA 50/200 has the benefit of being well-known, closely watched, and relatively simple. Moving averages have been observed by market participants for a very long time, with recent advocates for using them as an indicator when making medium-to-long-term calls such as Meb Faber. Such a general strategy also can provide positive results with a variety of other investment options, both equities and non-equities (though not all, as I will show in another article).
When applied to SPY over the period from 1993 to 2016, the indicator doesn't require the timing to be exactly right. Indeed, over this time period, returns could actually be increased by reducing the noisiness of the indicator and requiring slightly more significant breakdowns or rallies as measured by the SMA 50/200 before exiting or entering the market. This backtest, for example, compares against buy & hold with an exit point of the 50-day SMA at 1% below the 200-day SMA and an entry point of the 50-day SMA at 1% above the 200-day SMA.
(Chart from 11/11/1993 to 1/15/2016. Copyright © Peter Kirby.)
Here the returns are $85,077 for an annualized return of 10.12% and a 19% maximum drawdown.
Performance Outside the Initial Data Set (1953-1993)
This chart is based on monthly data for 1973-1993 from Robert Schiller, before expenses, with dividend reinvestment for the S&P 500 approximated. The SMA 50/200 is converted to a 2-month / 10-month simple moving average appropriate to the monthly data.
(Chart from 11/1973 to 10/1993. Copyright © Peter Kirby.)
Here the buy & hold investor ended the 20-year run with $189,056 for an annualized return of 15.83% with a 36.5% maximum drawdown. The 2-month/10-month SMA long-term trend-following investor ended with $226,567 for an annualized return of 16.89% with a 14.7% maximum drawdown.
Here is the same test again, from 1953 to 1973.
(Chart from 11/1953 to 10/1973. Copyright © Peter Kirby.)
Here the buy & hold investor ended the 20-year run with $72,857 for an annualized return of 10.44% with a 25.7% maximum drawdown. The 2-month/10-month SMA long-term trend-following investor ended with $108,138 for an annualized return of 12.64% with a 5.4% maximum drawdown.
Slightly Weaker During Booms, Sharply Stronger During Busts
The worst cases of underperformance against the S&P 500 index uniformly occur in decades of strong growth without any major, prolonged bear markets, during the 1950s, 1980s, and 1990s. Even then, most of the performance of the index is achieved, with double-digit annual gains (in absolute terms) during the good times that should help quell doubts about this kind of strategy in periods of underperformance against the benchmark.
This doesn't mean that the strategy is for everyone. For example, some investors have shorter time horizons, while other investors don't have the time or the discipline to find a way to implement any strategy with moving parts like this one.
The strategy should appeal the most to long-term investors who already follow financial news and can become fearful about their investments but have no real plan for making their exits and entries during a bear market.
The strategy will completely break down if an investor "plays trader" only when they let their emotions take control. Bias against turning paper losses into real ones will prevent them from executing the exits early enough, while the same lack of discipline might mean that they wait too long to get into the market again. This is, of course, the all-too-well-known cycle of greed and fear that robs many investors of their long-term returns. The strategy requires discipline (or automation) and must be adopted in the spirit of "cautious optimism," seeking long-term gains, not simply from fear stemming from the latest correction.
Is It "Too Late" Right Now to Exit Equities?
Nonetheless, if someone is still heavily invested into equities and is considering a revision of their strategy, it is not yet an absurd time to exit the stock market, whether today or after waiting a little for a short-term rally. We are still within small single-digit percentage points from the price level at which the simple SMA 50/200 strategy would have made its exit. Keep in mind that this is always a lagging indicator in the first place and never times the peaks perfectly, often missing them by about 10%. Yet it still can be used for long-term outperformance.
However, the S&P 500 is currently testing the lows from the last market correction last year, and it may quickly become a lot more than a 10% correction at this point. There is a considerable likelihood at this point of a deeper correction (or a bear market), and there is no way to outperform long-term by waiting for a major correction to happen before pulling money out of the market.
Whether you stay in or out of the market right now, discipline is still key. Those who hold must continue holding despite larger losses on paper, or they've abandoned their strategy in a way that practically guarantees underperformance. Those who exit must be ready to jump back in when the rally proves to have begun, or they lose the advantage that they attempted to create by the exit.
If you sell equities during a correction to avoid the risk of further deterioration, it is not necessary to be infallible or precise. Attempts at either actually create psychological insecurities that will adversely affect your decision-making. Buffett said it well: "It is better to be approximately right than precisely wrong."
Future articles in this series will, among other things, discuss options for improving the performance of a "risk off" portfolio (out of equities) and a "risk on" portfolio (exposed to equities).
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.