This post has been a LONG time coming, but a while ago I wrote about the rate of return for investing in the S&P 500. In that article, I compared the returns of someone executing a buy-and-hold strategy starting in a certain year and ending 10 years later. Unsurprisingly, the best time to start a 10-year investment was in 1990 or early 1991, as the peak of the DotCom bubble happened 10 years later and you could sell out at the top.

Figure 1: Return over 10 years of a $10,000 investment, assuming buy-and-hold strategy

But what about someone who wants a more sophisticated strategy than simple buy-and-hold? The reason people day-trade is that they hope to beat the market, not just match it. One strategy that I have seen genuine, peer-reviewed literature discussing is the so-called “momentum” strategy of buying while the market is going up and selling while it’s going down. In this way you should avoid big loses (like the DotCom bust) but still have big gains (like the DotCom bubble).

Now, a momentum strategy can be done in different ways. It can look at specific time periods, it can include shorting, it can include sector rotation, etc. But the simplest momentum strategy I found was to simply sell out whenever the market dropped by 20%, and then buy back in when it recovered 20% from the bottom. This is intended to stop loses on the way down and avoid FOMO-ing back in during a bull trap, only buying stocks during a true bull market.

I wrote a program to calculate the return on a $10,000, 10-year investment using that strategy.

Figure 2: Return over 10 years of a $10,000 investment, assuming 20% momentum strategy

The results are fairly discontinuous because of the rigidity of the 20% cutoff, but some patterns do emerge. The return is almost identical for people who invested in 1990, because for that 10-year period the market never dropped 20%. Once you get into 1991 however, this strategy would have allowed some people to avoid the worst of the DotCom crash, as they would have sold out when the market dropped hard. In that case they would have done better than a buy-and-hold strategy.

However that’s just an example of the strategy working at it’s best. I decided to compare the two strategies. I simply subtracted the two graphs from each other, creating the below figure as a result. Any dot that is on the zero line is a point in which buy-and-hold performed identically to momentum. Any dot below is where momentum performed worse, and the few dots above are where it performed better.

Here, we see some interesting patterns, the momentum strategy actually performed pretty poorly for anyone who started a 10-year investment in the 2000s. The peaks in the early 90s are people who sold out during the DotCom bust and missed the worst of the loses. The peak around 1999 is people who sold out during the Financial Crisis and missed the worst of the loses. But the declining valley during the 2000s is the result of people who would have sold out during the Financial Crisis, but then waited for the market to get above where they had sold before buying back in.

Remember that the momentum strategy involves selling when the market has lost 20% and only re-buying when it’s regained 20% off the bottom. Less than 20% off the bottom and you can argue (as some have this year) that it’s just a “bull trap” and the market still has “another leg down” ie much further to fall. This can result in standing on the sidelines with your cash while the market makes money without you. And using this momentum strategy, that’s exactly what can happen.

I use this to illustrate a point I’ve talked about before, it’s not usually smart to just sit on cash waiting for the market to fall further. Sure the market can fall further, but it can also rise and leave you behind. Time in the market beats timing the market. Furthermore, this experiment is as generous as possible to the momentum strategy: there are no transaction costs (the bid-ask spread is an unavoidable real-world cost) and we ignore dividends (which further rewards time in the market at the expense of timing the marker). If total returns were taken into account along with transaction costs, it’s debatable as to whether any 10-year momentum investment would have beaten buy-and-hold. Even as it stands now, only a very few lucky investment windows would have benefited from momentum strategies, most would do best with buy-and-hold.

Just for kicks, I reran this data with a 10% momentum strategy instead of 20%, and the results were even worse for momentum. Selling out at the first sign of trouble, FOMO’ing back in to the first recovery, and then losing all over again makes for a terrible strategy and that can basically be what momentum trading is.

I can go forward and look at more exotic momentum strategies some other time (for example short stocks that are falling and long stocks that are rising), but for now I think I’ve proven my point.