In our previous article, (see here) we documented that over short periods of time stock index markets such as the S&P 500 ETF (NYSEARCA:SPY) have a tendency to experience price reversals: Market strength is on average followed by market weakness, and Market weakness is on average followed by market strength. We noted that this counter-intuitive behavior seems tailor-made to throw off the timing of the average investor.
But what about over longer periods of time? Look at a long term chart of the S&P 500 index or the Dow Jones Industrial Average, and it becomes clear that there have been large benefits to owning stocks over the very long term. However, this has also entailed suffering through long periods where stock prices fell and there was substantial downside volatility.
Is there any way to improve on buy-and-hold results? If history can be used as a guide, it appears that this is the case. Simple, very long term trend-following models appear to work quite well when applied to stock indexes.
Let's create a long term trading system and apply it to S&P 500 data. Here is our rule set:
- If the close this month > the close "X" months ago, buy
- If the close this month < the close "X" months ago, sell
- When out of the market, earn risk-free rate on cash. For this study, we will use historic T-Bill rates that correspond to the appropriate month.
That is it! No fancy rules or calculations. We will look at results relative to buy and hold and cash using 8 to 16 month look back periods stepping two months at a time.
Here are the results using a 10 month look-back period. This was the first parameter I tried, and interestingly it creates the highest total return relative to the other parameters I evaluated for this study.
Using the 10 month look-back period, the average compound return increases about 1.1%, and the maximum peak-to-valley drawdown (looking at monthly closing data) is reduced from 53% to 30%.
Let's look at the historical performance for look-back periods between 8 and 16 months, stepping two months at a time:
Please note that the legend on the far right is stacked in the same order as the results appear on the chart. For example, cash is the lowest return and is both the lowest line on the chart, and the lowest label on the chart legend.
All look-back values between 8 and 14 months handily beat buy and hold. However, it looks like too short of a look-back period does not give the market enough room, while too long of a look-back period benefits less from exiting the markets during negative-return periods.
One weakness of this study is that we are using Index Data, and not an investable security. If you would like to see how this concept has worked on the SPDR S&P 500 ETF from 1993 to present, I have posted the results here.
Will simple trend-following systems continue to work in the future? That of course is an unknown. However, our studies suggest that a long term trend-following approach does have strong potential to provide return and diversification benefits.
The question is often asked about such trading methods, "If it works so well, why can't everyone use this approach?"
Unfortunately, the markets (and life) don't work that way. If too many people try to get on one side of a seesaw or trade the same approach, the markets will adjust and take away the potential benefits. Superior results rely on a trading method maintaining its status as a minority viewpoint. When too much money tries to capitalize on any one method, results tend to degrade until enough people "drop out."
All investment and trading strategies have a carrying capacity beyond which they stop working. This is a very important concept, because it suggests that in order to get superior results, the investor must act before an idea becomes conventional wisdom.
Disclosure: I 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. I have no business relationship with any company whose stock is mentioned in this article.