Momentum vs Mean Reversion


Most traders are probably familiar with these two main techniques- momentum trading and mean reversion.


As the names suggest, these two techniques represent two very different approaches towards trading. Momentum trading is one which you place trades on stocks which are having positive momentum in prices. High can go higher. You look out for several technical indicators to understand if a breakout has happened/going to happen from the price movement of the stocks. If yes, you will then be looking into initiating/adding your position into these stocks. In the past year, momentum trading has worked very well for tech stocks, especially Tesla.


On the contrary, mean reversion is an entirely different technique. It is one which you believe that the prices of stocks follow a certain mean. If the prices are too high, there is a certain belief that the prices will revert to the mean. The same thing happens if the prices are too low. You believe that the prices will eventually rise and revert to the mean. In fact, this has large similarities with value investing. You look out for stocks which are undervalued and invest in them. This hasn't exactly worked out too well in the past few years as compared to momentum trading when viewed broadly and hence value stocks didn't perform as well as expected. You may refer to this article which I wrote previously on value investing.

Given different trading time horizons though, both techniques could actually be useful as they each has their own advantages. In short time horizons (eg. days), mean reversion could actually prove to be useful as very few stocks go up straight for days in a row without any pullback. Hence, mean reversion strategies do work out in some sense as the stocks do either pullback if they have been going up or go up if they have been oversold. In that sense, short time horizons could favour mean reversion technique with the possibility of allowing the traders to accumulate and compound many small wins/profits. The risk usually happens when there is a sudden breakout in the price charts of the stocks which results in a large loss which might erode all the small wins/profits accumulated.


Sometimes, momentum trading works better in longer time horizons. Breakouts are likely not going to happen in short time horizons, even for stocks with positive price momentum. By holding the stocks for a longer period of time, the chances of getting a breakout in the prices increases. Of course, trading frequency associated with this strategy will be less than the mean reversion strategy and hence you will not be looking at accumulating or compounding many small wins/profits. A large win coupled with many small losses is usually the characteristic of this strategy. There a well written article on this comparison on quora and you might like to check it out.


If both strategies are advantageous in different situations, could we accurately pinpoint when to use which strategy in the market from a data point of view. Well, the answer isn't as straightforward as it is.

Generally, the movement of the data could fall into three broad categories- mean reverting, random, trending. The categorisation of data into these three categories depend on the Hurst exponent value of the graphs. It generally represents the autocorrelations of the time series. If the Hurst exponent is less than 0.5, the data belong to the mean-reverting category. If the Hurst exponent is more than 0.5, the data belong to the trending category. Of course, if the Hurst exponent is 0.5, the data is then truly random. Theoretically speaking, it should be easy for us to understand which strategy to be using on a long term basis if the market data (eg. historical prices of S&P 500) falls under either the mean reverting category or the trending category. Unfortunately, it is very likely that the market data falls under the last category when viewed in the long term horizon- that is the random category. Hence, things aren't as straightforward as we want them to be.


So what could happen in 2021? Are we going to see a mean reversion after a relatively bull year in 2020 or is this strong trend likely to continue with more momentum? To be honest, I don't have a good answer to this and I don't think anyone has. However, I have some materials here which represent either side of the viewpoints and you might like to check them out.

In this particular article on Seeking Alpha, the author has made an analysis based on 100 years of S&P 500 data. Based on his analysis, it seems likely that certain reversion to mean might happen as the market now enters 2021 with a very high Shiller PE ratio of 33. When you enter the market with such a high Shiller PE ratio, it is likely that your future returns are likely to be lower (and hence a representation of a reversion to mean). I personally find this article insightful as it is littered with various tables and graphs to give you good visual representation of the data.


On the other hand, there is also this article on ofdollarsanddata.com which shares the view that breaking an all-time high is a bullish sign as it is very likely that you will see a continual breakthrough of all-time high in the near future- a sign of more momentum gains. This is backed with interesting data such as the days between all-time highs for S&P 500 over the past decades.


To be honest, it could be a situation whereby both could be right. We might be seeing more all-time highs in the near future yet the overall returns might be dismal when viewed in a longer time horizon- eg. 10 and 20 years. Of course, that is just how I interpret it and you are free to interpret them and have your views too.


Feel free to follow me on Twitter and share with me your thoughts.


Last but not least.. please scroll down and subscribe for regular content if you like what you read!

407 views0 comments

Recent Posts

See All
  • Facebook

©2019 by datascienceinvestor