The last decade has been an extremely difficult period for stock market investing. Buying and holding the S&P 500 Index over this period has resulted in a loss of 0.95 percent as of the end of 2009, according to iShares.com. Because of these extremely frustrating results, many investors have turned away from buy-and-hold investing, which worked so well in the 1980s and 1990s. They have turned to a variety of strategies, from diversifying into alternative asset classes to fundamental or technical analysis, to improve performance.
One technique widely employed by professional and individual investors alike is the moving average. Moving averages have been embraced for good reason, as investors could have saved themselves substantial losses in the last bear market by using them to determine when to hold and when to sell broad-market indexes. Applying a 50-day or 200-day moving average to the S&P 500 in 2008 would have produced a loss of 3.14 percent and 3.47 percent, respectively. By contrast, a buy-and-hold position in the S&P 500 would have returned a negative 38.49 percent.
These losses will eventually be recouped. However, many do not have the nerves of steel necessary to “hang in there” after watching their hard-earned nest egg decimated. For these investors, a disciplined investment approach is needed to reduce the volatility in their portfolios, while preventing them from bailing out at the bottom. One such method may be the use of moving averages.
In this paper, we will explore three of the more popular moving averages—the 200-day, the 50-day and a 50-day/200-day crossover. We will analyze the risks and returns of each, comparing them with each other and also with buy-and-hold investing. We will study the results of trading these systems on the S&P 500 stock index, starting in 1971, the first year data is readily available. Special emphasis will be given to the three major bear markets occurring in this period.
Moving Averages Defined
What is a moving average? Investopedia defines it as “An indicator frequently used in technical analysis showing the average value of a security’s price over a set period.” For example, to calculate a 10-day simple moving average, one would collect the prices (usually at close) for the security over the past 10 days, add them together and divide by 10. The following day, one would include the price for the most recent day and drop the price for the first day. Hence, the average is dubbed “moving.”
The advantage of using moving averages is that they are great in defining whether an investment is trending up or down. The problem though, is that they are by definition slow to adapt when a trend has changed. As Robert D. Edwards and John Magee wrote in their classic work “Technical Analysis of Stock Trends”:
The trouble with a Moving Average (and which we discovered long since, but keep bumping into from time to time) is that it cannot entirely escape from its past. The smoother the curve (longer cycle) one has, the more “inhibited” it is in responding to recent important changes of trend.
To make a moving average somewhat more responsive to recent data rather than older data, many investors use an exponential moving average, which calculates the moving average geometrically. In this study, we will use exponential averages exclusively.
200-Day Moving Average
Perhaps the most well known is the 200-day moving average, which averages prices over the past 200 days. Generally, the longer an investor’s time frame, the longer the moving average used. For example, a day trader has no use for a 200-day moving average. A long-term investor, in contrast, has no use for an hourly one. The 200-day moving average is the standby for long-term investors.
For the purposes of our study, one would be invested in the S&P 500 when the price was trading above the 200-day moving average, and out of the market when the price was below the average. The buy trigger is pulled when the day’s low price rises above the average. A sell signal is flashed when the day’s high price falls below the average. Most market technicians rely on closing data rather than intraday highs and lows to indicate trades. When I analyzed the data, however, this resulted in an excessive amount of whipsaw trading. Thus, in this paper, intraday highs and lows are employed to generate the trade signals.
The performance results for this strategy compared to buying and holding the S&P 500 beginning Dec. 31, 1970 and ending Dec. 31, 2009 are in Figure 1.
The total returns for both methods are virtually the same. The big difference lies in how those returns were achieved. Using the 200-day moving average, investors would have decreased their annual volatility by 26 percent. In other words, one would have achieved the same return with only three-quarters of the risk.
Figure 2 quantifies the returns during the bear markets of 1973-74, 2000-02, and 2008. In every one, buy-and-hold investors lost approximately half their capital.
Figure 3 depicts the results of using the 200-day moving average system during these bear markets. The portfolio was valued on the S&P 500’s highs just prior to these bear markets. The portfolio was then adjusted based on the number of trades during the bear market. In the ’73-’74 downturn, this system gave six false buy signals on the way down. In the 2000-02 bear, it gave three false signals. In the 2008 decline, it gave four. Still, overall results were much better than buying and holding.
The 200-day moving average strategy works best during trending markets, whether up or down. It does not do well during periods in which the market moves more or less sideways. In the entire period, the system dictated 87 round-trip trades (174 total). Although this works out to only 4.5 trades per year, the majority of these trades occurred during choppy trading ranges. Approximately half appeared in the 1970s alone. During these sideways markets, it was not uncommon to do trades on a weekly or biweekly basis. Just 36 of the 87 trades were profitable. Thus, nearly 60 percent of the time, the positions were closed at a loss. This would have been frustrating for most, increasing the likelihood that one would “cheat” on using the system. Also, an investor would have incurred trading costs and tax consequences if trading in a nonqualified account (which are not included in this analysis). Because of these drawbacks, many investors do not use the 200-day moving average solely to dictate portfolio activity. They will use other indicators in conjunction with it on which to base their investment decisions. (A study of these other tools is beyond the scope of this paper.)
It should be noted that the return calculations for this paper do not include dividends, which would have increased returns substantially. Investors would have been in the market approximately 70 percent of the time (28 of the 39 years studied) trading the 200-day strategy. Therefore, they would have received many dividend payments, but obviously not as much if they had simply held the S&P 500 for the entire time instead. It seems prudent to remain in cash for a few weeks upon receiving a sell signal, and then when it appears that a whipsaw is less likely, invest the portfolio in Treasury bonds. This may mitigate the loss of dividends when not holding stocks.
50-Day Moving Average
The 50-day moving average is generally defined as one used by intermediate-term investors. The buy and sell criteria remain the same as that for the 200-day moving average study—the only difference being the moving average itself.
The performance for this approach versus buy-and-hold appears in Figure 4. The return offered by the 50-day moving average system was substantially less than that of the buy-and-hold approach. However, the standard deviation was also much less for the moving average strategy.
When zooming in on the three bear markets using this moving average, the results get interesting. Figure 5 shows a big difference from the buy-and-hold approach (see Figure 2) in the 1973-74 and 2008 downturns. The disparity is not so stark in the 2000-02 bear market. While the S&P 500 declined by almost half, an investor would have still watched a third of his portfolio evaporate. The reason for this disappointment was whipsaw trading. Prices crossed the 50-day moving average so much during these three years that 29 trades were executed. Most of these transactions resulted in small losses, which added up over time.
Comparing the 50-day with the 200-day moving average, we see that the former bested the latter in the first bear market. The 200-day lost far less in the second bear. Results were practically identical in the last downturn.
As one might expect, using a shorter moving average generates more buy and sell signals. While the 50-day average certainly indentifies a change in trend sooner, many of these signals proved to be false. The number of round-trip trades was 218, or 436 total. This means 11.2 trades per year; almost one per month. Thus, this strategy generates 2.5 times as much activity as the 200-day approach. During trendless markets, there were times when an investor would have been buying one day, and selling the next. Just 30 percent of trades were profitable. Time spent in the stock market was also less than the longer moving average, at approximately 64 percent (25 of the 39 years studied). It appears that for most long-term investors, utilizing a 200-day moving average may be a better choice than the 50-day.
50-/200-Day Moving Average Crossover
By now, this paper has established that using moving averages can reduce portfolio volatility. There are drawbacks, however, including false buy and sell signals and a high percentage of losing trades. The former increases trading costs, and possibly taxes as well. The latter can be psychologically damaging. As veteran investors realize, controlling one’s emotions is half the battle.
To address these shortcomings, the moving average crossover system was developed. Using this approach, one employs both the 50-day and the 200-day moving averages. The trade trigger is pulled when one moving average crosses over the other. If the 50-day average crosses above the 200-day, a buy signal is given. If the 50-day line crosses below the 200-day moving average, a sell signal is generated.
This is a popular method in the literature on technical analysis, where the crossing of the 200-day and 50-day averages is referred to as a “golden cross.”
Performance against a passive buy-and-hold approach is shown in Figure 6.
This moving average system generated a respectable 0.6 percent larger annual return than the passive approach, but with an amazing 33 percent less volatility. It also had a higher return with lower standard deviation than either the 50- or 200-day strategies.
Figure 7 analyzes the bear market performance of the 50-/200-day crossover.
The method lost more than the 50-day and 200-day moving average systems in 1973 and 1974 ($15,300 vs. $7,000 or $10,400, respectively); however, during the last two bear markets, the crossover system worked better than the other two.
Taking a closer look at the three bear markets, we do not see many wrong signals. Investors had just one whipsaw trade in the 1973-74 bear. They had no such trades in either the 2000-02 or the 2008 period (see Figures 8-10).
One can also observe by reviewing these three charts why solely using either the 50-day or the 200-day moving average results in so many incorrect trades. The blue line is the 50-day average. The orange line is the 200-day average.
While perhaps not as quick to discern a change in trend, the strategy’s prognostications were much more accurate. Seventy-six percent of all round-trip trades were profitable. The system generated only 17 round-trip trades, 34 in all—less than one per year. An investor using this method would have been invested in the market 72 percent of the time—just slightly more than using the 200-day moving average by itself. All three moving-average strategies are compared with buy-and-hold in Figure 11.
This study indicates that an investor can reduce risk in his portfolio by enlisting the help of moving averages. Using the S&P 500 as a proxy investment, it is clear that the 50-day/200-day crossover system is superior to the 50-day or 200-day moving averages by themselves. While this may be true for the broad stock market, results may vary for different indexes.
There are many ways an investor can use this information. One tactic would be to simply trade an S&P 500 exchange-traded fund or index fund based on the signals generated. Another would be to invest in other attractive equity investments that one thinks will outperform the market during an upturn. One would remain in those holdings as long as the 50-day average is above the 200-day average on the S&P 500 stock index, and liquidate those positions when the sell signal is generated.
It goes without saying that investors should not rely solely on any one technique. However, applying moving-average strategies in conjunction with portfolio diversification and prudent money management may reduce one’s risk substantially. If nothing else, this will lead to better sleep if the next decade is anything like the last one, and it may lead to a larger nest egg.