Saturday 10 May 2014

Profiting from momentum strategies - Part 1

Momentum is arguably the most well-known trading strategy. A simple strategy of buying stocks that have done well over the past 6 months ("winners"), and shorting stocks that have done badly ("losers"), earns a 1%/month return over the next 6 months. While other trading strategies stop being profitable once they have been discovered (because investors start exploiting them, removing the profit opportunity), momentum has remained surprisingly lucrative ever since Jegadeesh and Titman (1993) first documented it. 

A momentum strategy is attractive because it is market-neutral - since you're buying some shares and shorting others, it can make money in up markets and down markets. Thus, it is relatively immune to market risk. The Sharpe Ratio (a measure of the risk-adjusted return to a trading strategy) of momentum is about 0.6, compared to 0.3-0.4 for just holding the market. I attempt to exploit momentum myself, through the AQR Momentum ETF (ticker AMOMX).

Momentum is also pervasive - it works not only in stocks, but also bonds, commodities and exchange rates as shown by Asness, Moskowitz, and Pedersen (2013). That we see it in so many assets suggests that momentum is due to investors making mistakes - popularized by a branch of research known as "behavioral finance". The main psychological explanation is that investors are slow to react to information - thus, good news takes time to be incorporated in prices, and ditto for bad news. 

However, even though the momentum strategy does well on average, there are some periods where it does very badly, such as in the recent hedge fund crisis - some hedge funds went under because they followed momentum strategies that tanked. For example, between March and May 2009, the "losers" generated 163% returns, but the "winners" generated only 8% returns. Thus, a momentum strategy is somewhat like selling options - it makes money on average, but sometimes does really badly.

This new paper by Kent Daniel of Columbia GSB, a former Managing Director in Sachs Asset Management and Toby Moskowitz of Chicago Booth, a former winner of the Fischer Black Prize for outstanding contributions to finance research by someone under 40, shows you when to get out of momentum strategies - and thus how to make momentum even more profitable. 

The answer is surprisingly simple - get out of the momentum strategy in times of market stress, when 1) the market has recently declined, and 2) market volatility (measured by the VIX volatility index) is high. Here's a simple intuition. If the market has recently declined, the "loser" portfolio must have declined much faster than the broader market. Thus, it has a high beta (= sensitivity to the market). The "winner" portfolio has a relatively low beta, which is why it didn't decline so much. After times of market stress, the market typically recovers. Thus, the "loser" portfolio, which has high beta stocks that are sensitive to the market, does especially well in the market recovery, and so you want to get out of the momentum strategy. Kent and Toby find that this surprisingly simple enhancement to the momentum strategy doubles the Sharpe ratio from 0.6 to 1.2.

6 comments:

  1. Hi Alex,

    Interesting subject. You might also be interested in a recent paper from the University College Cork which looks to explore when these strategies do well and do badly (Hutchinson & O'Brien (2013) - 'Is this time different? Trend following and financial crises').

    I have two quibbles with the overall thrust of the post though which I'd be interested in your take on:

    1) Is it that momentum strategies only do badly post a financial crisis, or is it that they only in fact do well in the run up to a financial crisis, in a classic late stage bull market? and

    2) Isn't the particular downturn since 2009 as much a function of money following the strategy? Burghardt, Kirk & Liu (2013) have written an interesting paper with stats on this, but most obviously the trend-following CTA universe has increased assets c. 17x over the last twenty years to $200-300m, whilst futures markets have only grown 5-6x. Capacity seems to be $150-200m and to have been breached a few years ago. It is interesting that most funds that follow these strategies now seem to spend more time limiting slippage to transaction costs etc. than actually optimising their strategies.

    I'm increasingly of the view that a combination of the two means this strategy will disappoint for a while yet.

    ReplyDelete
  2. Thanks James, appreciate your interest!

    1) It's more the former - poor performance of momentum is the exception, rather than the rule. In most market conditions, momentum does well, but only in rare cases (times of market stress) does it underperform - and underperform significantly.

    2) You are right that another reason for poor performance of momentum strategies is many investors piling into these strategies, i.e. too much money chasing it. This is actually the topic of my Part 2 of "Profiting From Momentum Strategies", hopefully next week - stay tuned!

    ReplyDelete
  3. hi Alex
    great blog

    does momentum work across countries ? ie if i go long countries w high momentum
    and short those w low momentum? or does it only work for instruments within those markets (stks, bonds, commodities, fx etc)

    also what about momentum long only? how does this impact the strategies performance?

    many thx!

    ReplyDelete
  4. you may also want to check aqr's latest paper - "fact, fiction and momentum investing "....

    ReplyDelete
  5. Thanks for your interest and kind words! I only know of Asness, Moskowitz, and Pedersen (2013) who study instruments within those markets - I don't know of any academic study that looks at investing in winner countries and shorting loser countries. (If any reader does, please enlighten me!) There is the less informal "Henry Hindsight" study in The Economist (e.g. http://www.economist.com/node/281885) which shows that investing in the country / asset class that did best over the last year doesn't generate high returns.

    Both the long and short size contribute significantly to momentum profits, so a long-only strategy will be less profitable. Shorting involves higher transaction costs, but Korajczyk and Sadka (2004) (http://onlinelibrary.wiley.com/doi/10.1111/j.1540-6261.2004.00656.x/abstract) showed that momentum profits are robust to trading costs. Thanks for notifying me about this new AQR paper.

    ReplyDelete
  6. thx alex
    these two links might be useful.
    the first seems to substantiate long only momentum investing. the second seems to show anomalies persist at the country level . what do you think?

    http://optimalmomentum.blogspot.com/2013/10/momentumthe-only-practical-anomaly.html?m=1

    http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2375902

    thx!

    ReplyDelete