The momentum factor is one of the most hotly contested yet natural factors to exist. Momentum is a classic phenomenon for stocks - winners beget winners and losers can find it tough to recover losses. Trend-following is a popular technical approach favoured by many investors, both professional and retail and makes intuitive sense.
Since the global financial crisis, Central Banks acted in a coordinated way and each embarked on large-scale quantitative easing (QE) campaigns. As a result, nearly all asset prices inflated over the course of a decade and a momentum-style follow the trend strategy worked very well. Going long equities and long bonds, similar to the risk-parity strategy, was a very successful investment and both markets rallied.
Momentum can also occur on a sector basis. The technology sector rally from 2010 onwards, whilst based on a fundamental shift in productivity and usage, can also be put down to this fear of missing out as technology winners turned into bigger winners and the price action was a self-fulfilling prophecy.
Traditional stock investors are human and exhibit human traits and patterns. Quantitative investment firms were quick to develop and build trading signals around this behaviour and classify the momentum premium as a bona fide style factor. Some general reasons why the momentum factor can exist are given by
These types of patterns can most easily be quantified by using historical returns to analyse which stocks have performed the best - and this is exactly what academics and quantitative firms did in order to come up with a definition of the momentum factor.
The momentum factor, as seen in the graph below, tends to historically have a higher rolling volatility than all other style factors. This is quite natural as it is arguably the closest style factor to the overall market index. On average the long-short momentum factor has a 1y rolling volatility of 13% since 2010. Since covid began in 2020, the volatility of momentum significantly increased, up to approximately35%, but then sharply retreated back to its average once the Fed turned on the stimulus.
The momentum factor as we saw in the value factor post is negatively correlated to the value factor, which overall makes sense. The relationship with the quality factor also intuitively infers a positive relationship. Strong companies should perform well and leverage that performance to build market share and keep performing strongly i.e increase momentum. However these relationships are time-varying and can ebb and flow as we can see, currently momentum and quality have close to zero correlation and could go negative.
The momentum factor overall has proved to be a consistent performer across time and geography. Momentum tends to do well in environments where there is a consistent theme and volatility is relatively contained - this makes it easier for the status quo to continue and past winners to keep doing well.
We now plot the Fama-French style momentum factor, which is long-short market neutral and based on recent price performance. This is a very simple but useful trend measure. Momentum much like value, is a more volatile factor and we can see this clearly in the yearly performance plots.
Anecdotal evidence suggests that there is some kind of momentum factor reversion that occurs in January. This also comes from some industry intuition that factor performance over the year-end turn can revert as asset managers and investors start the new year by trying to invest in the underperformers of last year and sell the best performers of the prior year.
Momentum over the last two decades has become quite a well-respected and harvested alpha factor. There has been some resistance in the traditional academic world in accepting it as a bona fide factor but there is now ample evidence that momentum exists. Carhart in 1997 did indeed extend the Fama and French 3-factor model to include a fourth momentum factor. Momentum has been shown to have the following properties:
Regardless of these properties, the resistance in academia may be down to the more human/behavioural reasons for its existence and a lack of financial explanation. Fama and French for example when they extended their 3-factor model chose to include two quality metrics (profitability and investment), rather than add a momentum factor, as their factors are easier to justify by referencing financial and economic logic.