Momentum and trend following
Securities that did well recently did better than others on average over the following months too. The effect is well documented, but has violent crashes and has weakened since it became known.
The basic idea of momentum is very simple: securities that rose more than others over the last three to twelve months tend to rise more over the following months too. And the other way round: what was weak recently tends to stay weak.
It's one of the best researched patterns in all of financial research. In 1993 the economists Narasimhan Jegadeesh and Sheridan Titman showed that a strategy that buys the strongest stocks and shorts the weakest earned clearly more than the market. Since then the result has been found again and again in most of the world's stock markets and in other asset classes.
Related, but not the same, is trend following. Here you don't compare different securities with one another. You only look at which direction a single market is moving. A study by the fund company AQR found positive results for trend following in stocks, bonds, commodities and currencies in almost every decade since 1880.
The effect has an unpleasant side. Momentum strategies sometimes suffer very violent, short crashes, especially when a market suddenly turns after a crisis. In spring 2009, the previously weakest stocks rose many times over within a few months. Anyone who had shorted them lost in a few weeks much of what they'd earned over years with the trend.
Jegadeesh and Titman (1993) showed that buying past winners and short selling past losers over holding periods of three to twelve months earned about one percent extra return per month. In a follow up study in 2001 they showed that this extra return didn't disappear in the years after the first publication either. That argues against it having been found in the data by chance.
Two camps face each other on the explanation. Behavioral researchers say investors react too slowly to new information, and later trend followers amplify the move further. Others argue that past winners carry higher risk than past losers and are paid for it. Cooper, Gutierrez and Hameed (2004) and Daniel and Moskowitz (2016) show that the extra return depends heavily on how the market has done recently. In recoveries after crises it can even turn clearly negative. That's called a momentum crash.
For trend following across whole markets, Hurst, Ooi and Pedersen (2017, AQR) studied 67 markets over more than a hundred years, with consistently positive results and low correlation with traditional investments. Important: both effects are measured in studies with realistic but simplified trading costs. And in Jegadeesh and Titman (2001) the extra return reversed two to five years after the end of the holding period. That suggests markets overshoot in the long run and the overshoot later unwinds.
Data as text
Cumulative profit is +12.2 percent after 12 months and −0.4 percent after 60 months.
Summary
- Securities that were strong recently did better on average over the following months too.
- The effect is documented over decades and in many markets, but it isn't without risk.
- Momentum strategies sometimes suffer violent crashes, especially in recoveries after crises.
Did you get it?
What did Jegadeesh and Titman show in 1993?
That it paid to buy the winners of the last three to twelve months and sell the losers.
What is a momentum crash?
A violent, short crash of momentum strategies, usually in a recovery after a crisis.
Which two camps face off in explaining the momentum effect?
One side says investors react too slowly and trend followers amplify that, the other says past winners carry more risk.
Sources and further reading
- Jegadeesh, N. and Titman, S. (1993), Returns to Buying Winners and Selling Losers, Journal of Finance
- Jegadeesh, N. and Titman, S. (2001), Profitability of Momentum Strategies, Journal of Finance View source ↗
- Hurst, B., Ooi, Y. H. and Pedersen, L. H. (2017), A Century of Evidence on Trend-Following Investing, AQR Capital Management View source ↗
Related
- Mean reversionStrategy Types
- Pairs trading and statistical arbitrageStrategy Types
- Grid-trading and DCA botsStrategy Types