Last week we discussed the possibility that “momentum is finished.“
Astute readers of the blog generally agreed with the concept, but also pointed out that risk management techniques can potentially save momentum strategies.
Other observers–to include a Wharton finance professor who has more brains than all of us combined–highlighted recent work by Kent Daniel:
I decided to conduct some back of the envelope tests for some of these claims: 1) risk management techniques can save momentum, and 2) Momentum suffers from “predictable crashes.”
Conclusion: Maybe momentum isn’t dead.
First question: Do risk management techniques help save momentum?
To conduct these tests I created two portfolios:
1) A monthly rebalanced portfolio that goes long the highest momentum stocks and short the lowest momentum stocks (note, these are decile portfolios. This strategy is not the size/bm adjusted momentum portolios on French’s website).
2) The portfolio in 1), but I apply a simple moving average timing rule that invests in the strategy if the 2 month moving average is above the 10 month moving average, and sits in cash otherwise. Symbolically, this is represented as MA(2,10). This classic idea has recently been brought to the forefront by Meb Faber over at http://www.mebanefaber.com/timing-model/.
Results: equal-weight market, L/S mom, and risk-managed L/S all do about the same…but the risk profiles of each are entirely different!
Next, we look at various bull/bear cycles. The striking difference comes in the latest bull market rally, where momentum gets crushed. The MA rule saves you (as it does in the 30′s crashes).
The drawdown analysis is another way of pointing out that risk-managed momentum certainly seems superior to regular momentum.
And our favorite MBA 101 chart. The chart shows a definite trade-off for small-caps and large caps (as captured by equal weight vs. value weight). However, the chart doesn’t show a sensible tradeoff for the straight L/S momentum strategy and the risk-managed L/S momentum strategy–the risk-managed version has similar return, but much less standard deviation.
If you’d like to dig in to the year by year returns of the various strategies, here is your opportunity:
Second question: Can we use ‘predictable crashes’ to save momentum?
A key insight from Kent’s analysis (as well as others who have studied momentum) is that momentum gets destroyed following major market blow outs. Two prime examples of momentum failing are in the sporadic bull markets of the Great Depression and during the March 2009 bull market rally of our current Great Depression (v 2.0).
I developed a simple rule that attempts to save momentum from blowing up. I create a ‘bear market indicator’, which is equal to 0 if the past 24-month return on the market is positive, and equal to 1 if the past 24-month return on the market is less than zero. The trading strategy invests in l/s momentum when there is no bear market signal (recent market performance has been above 0%), and sits in cash when we have been in a recent bear market (recent performance sucked). Simple enough. Now how does it perform?
Looks good for the bear market indicator risk management technique
Still looking good…
The drawdowns using the bear market indicator are rough, but better than the straight-up L/S momentum…
And our favorite MBA 101 chart again. The bear market indicator does very well!
Here are the detailed year by year returns–not bad at all.
These indicators are completely data-mined by previous researchers and may not have any predictive value in the future. But you know what? Is the risk of being in a data-mined trading rule higher than staying long and strong in the rough and tumble buy-and-hold mantra?
Another caveat: I’ll admit it, I was brainwashed to believe that technical rules can’t work because Warren Buffett, Ben Graham, and a whole slew of academics have sworn off technical analysis as the enemy of the world–I still might believe them. Nonetheless, the more I dig into the results from using simple technical analysis rules, the more I realize I need to think outside the box more often. Even if the rules don’t add “alpha”, they do help manage risk–especially the rip-your-face-off drawdowns of 50%+.