I thought it worth elaborating on last week's post regarding rule-based trading according to moving-average momentum. In particular, if that strategy provides greatly reduced volatility with somewhat less reduced returns, then that begs the question of whether or not there is a way to generate only slightly reduced volatility with non-reduced returns? In other words, buy&hold levels of returns with less volatility than the buy&hold portfolio. [By the way, if I ever find a formula like this, or better yet, one with greater returns than the buy&hold portfolio and less volatility - I may try making a living off it before I disclose it in this space.]
As a straightforward non-creative attempt at providing an answer, I've tweaked last week's analysis as follows: rather than going to cash when the moving-averages are trending down, see what happens if you short the market at those times. The results of this test are shown in the chart above.
As expected, this strategy produces the same volatility as the buy&hold portfolio. The reason is because, every day, long or short the market, your portfolio will bounce around one way or another in proportion to how the market moves. Also expected, the Beta of this strategy is in the ballpark of zero. To understand the reason, simply imagine being long the market 50% of the time and short the market 50% of the time. When you're long, your Beta is 1.0; when you're short, your Beta is -1.0. Mathematically, 50%*1.0 + 50%*(-1.0) = 0.
However, the returns from this strategy don't beat the buy&hold portfolio. They don't even do well enough to provide a superior Alpha in comparison to last week's strategy of going to cash, rather than shorting the market. I can't really provide a full explanation for why this is, except to say two things: 1) transaction costs are doubled because not only do you have to buy and sell, you also have to sell and buy (to short), and 2) the market is very (although maybe not perfectly) efficient, which causes a high degree of randomness.
Overall, I have to conclude this buy&short strategy is inferior to last week's buy&sell strategy because you have higher volatility and (slightly) lower returns.
Side note: check out 'Black Monday' (and the days following in Oct-'87) in the chart above. Wow.
Technical Notes:
1. One reader commented last week that the 6% returns for the buy&hold portfolio looked low. In other words, everyone tends to think stocks provide 10% returns in the long run. A few reasons for the discrepancy: first, the returns shown in each decade are geometric, rather than a straight average of the ten years; second, the returns in each decade exclude the results of the first 200 days in order to first calculate a 200-day moving-average prior to beginning the analysis; third, these returns are for the Dow Jones Industrial Average (historical dividend adjusted pricing provided by yahoo finance), which may vary from what an outfit like Ibbotsson may deem to be the 'stock market'.
2. It's interesting how the Beta can be near zero and yet the trading portfolio appears to somewhat follow the buy&hold portfolio when viewed on a 10-year chart. This is something to keep in mind when considering any statistic that's calculated based short term data (e.g. daily). Many paradoxes in finance (life?) can be resolved by rigorous attention to the time frame under discussion. In many cases, the small deviations from the short-term statistic (be it Beta, an average, or whatever) accumulate in one direction over the long-term. For instance, people like to point out that when the U.S. market declines, foreign stocks tend to decline as well, thereby nullifying the diversification benefit. However, when they say this, they are mainly thinking in terms of the short-run (i.e. days), when the diversification benefit is a actually a long-term phenomenon. Foreign stocks are less correlated with U.S. stocks in the long run mainly due to long run factors like demographics, political regimes, etc.
Quote of the Week: "Not needing wealth is more valuable than wealth itself." - Epictetus (AD 55–AD 135) Greek Stoic Philosopher.
No comments:
Post a Comment