What is the best moving average period?

What is the best moving average period?

The big question when using moving averages is: what time period should I use? I could find very little about this on the web so I decided to research it.

I analyzed what simple moving average (SMA) would have given the best return, historically speaking. I used the S&P 500 index as a benchmark since it encompasses a large portion of the stock market and can therefore give a general idea of what SMA period works best.

Unlike other technical indicators, for example RSI, the moving average gives extremely clear signals. There is really no interpretation required which means it can be easily computerized. I put an SMA on the historical prices of the S&P 500 (from 1950 to 2010) and let Excel calculate the average return that would have been achieved over this 60-year period if stocks were bought and sold exactly as the moving average dictated. This was done for many different SMA periods. Results are shown in the table. Note that brokerage fees are ignored.

As you can see, the period length makes a big difference in the return. For example, a 130-day SMA generates a return of 6.6%, whereas a 175-day SMA generates a return of 7.5%. That 1% makes a big difference when compounded over time.

It actually turns out the 2-day moving average gives the highest return, 12.2% annually! The annualized market return for the same period was only 6.9%.

However, with a 2-day SMA, trades happen roughly every two days, resulting in high brokerage costs.

I redid the calculation incorporating a 0.5% brokerage fee at every transaction. As can be seen in the second table, it’s amazing how devastating such a seemingly small fee can be for the average return.

The highest return was reached with an SMA period between 360 and 365 days. When using these averages the return maximizes at 5.6% annually. The market generated a return of 7% annually in the same timeframe. This is somewhat of a disappointment as it shows that you can’t beat the market (or at least, the underlying security) using only simple moving averages purely because of brokerage fees.

As the SMA period becomes longer, the average reacts slower and there will be fewer trades. At some point the average is so slow that there is only one trade, a buy signal at the beginning. After that there will be no more trades and essentially this is the same as a buy and hold strategy. The return that is reached is equal to that of the market. Essentially the most profitable SMA period is [some large number; say 5000 days] since the greater the SMA period the closer the return will be to that of the market.

Conclusion

Historically speaking a buy and hold strategy would have beaten the return of any moving average. Yet the highest yielding moving average (of period bellow 1000 days) was the 361-day SMA. Its annualized return was 5.6% compared to the market return of 7%. Had there been no brokerage fees the highest returning SMA would have been the 2-day moving average with an annualized return of 12.2%.