A Hedge Strategy to Exploit Sector ETF Momentum?
Posted in Momentum Investing
July 9, 2010
Scott Oyen of Scott’s Investments asked: “Regarding your 6-1 sector ETF momentum backtests, have you tested hedge strategies such as:
Purchasing the top sector based on six-month momentum and shorting the bottom sector based on six-month momentum?
Hedging an equal-weighted portfolio of the top three sector ETFs based on six-month momentum with a short position in the S&P 500 Index (via ProShares Short S&P500 – SH)?
By adding a short position, volatility and drawdowns could decrease.” Using the same data as in “Simple Sector ETF Momentum Strategy Performance”, we find that:
Regarding the first question, the following chart shows the cumulative values through June 2010 of $10,000 initial positions initiated at the end of July 1999 for three strategies:
- At the end of each month, take a long position in the sector ETF with the highest total return over the past six months (Long 6-1 Winner).
- At the end of each month, take a short position in the sector ETF with the lowest total return over the past six months (Short 6-1 Loser).
- At the end of each month, take a long position in the sector ETF with the highest total return over the past six months and a matched short position in the sector ETF with the lowest total return over the past six months (6-1 Winner – Loser), rebalanced monthly. In other words, the monthly return is the winner total return minus the loser total return.
While incorporating a 0.25% per switch trading friction, the comparison ignores the costs of shorting in strategies 2 and 3 and the costs of rebalancing in strategy 3. The comparison also ignores differences in capital requirements among the three strategies.
Results indicate that shorting the momentum loser among sector ETFs is unprofitable over the sample period. In fact, the sector ETF with the worst lagged six-month return is not the worst next-month performer among sector ETFs (see the last chart in “Simple Sector ETF Momentum Strategy Performance”).
The long/short strategy is also unprofitable, suffering from an unprofitable short position, additional trading frictions and a damaging rebalancing effect.
Regarding the second question, it would take quite a bit of work to test a short position in S&P Depository Receipts (SPY) or long position in SH hedging a same-size, equal-weighted position in the three sector Exchange-Traded Funds (ETF) with the highest lagged six-month momentum, rebalanced monthly. Instead, consider hedging just the above Long 6-1 Winner strategy.
The next chart shows the cumulative values through June 2010 of $10,000 initial positions initiated at the end of July 1999 for two strategies:
- At the end of each month, take a long position in the sector ETF with the highest total return over the past six months (Long 6-1 Winner).
- At the end of each month, take a long position in the sector ETF with the highest total return over the past six months and a matched short position in SPY (6-1 Winner – SPY). In other words, the monthly return is the winner total return minus the SPY total return.
While incorporating a 0.25% per switch trading friction, the comparison ignores the costs of shorting and rebalancing in strategy 2. The comparison also ignores differences in capital requirements between the two strategies.
Results indicate that, over the sample period, the hedged strategy produces a comparable outcome at lower volatility, but costs of shorting and rebalancing would lower the performance of the hedge strategy.
For reference, the average monthly return of the Long 6-1 Winner (6-1 Winner – SPY) strategy during 7/99-6/10 is 0.48% (0.46%), with standard deviation of monthly returns 6.00% (4.84%).
The next chart replaces the above short position in SPY with a long position in SH, which has been available only since June 2006 (6-1 Winner + SH). Results again indicate that, over this shorter sample period, the hedged strategy produces a comparable outcome with perhaps lower volatility, but costs of rebalancing would lower the performance of the hedge strategy. Confidence in this inference is extremely low because of the extremely short sample period (only eight independent six-month intervals).
For reference, the mean monthly return of the Long 6-1 Winner (6-1 Winner + SH) strategy during 7/06-6/10 is -0.14% (-0.02%), with standard deviation of monthly returns 5.97% (4.71%).
A strategy based on the three sector ETFs with the highest lagged six-month momentum would probably further reduce volatility, at the cost of some mean return. However, a three-sector strategy also involves additional trading frictions for sector rotation and maintenance of equal weighting among the three selected ETFs.
The sample periods (especially for SH) are short for this kind of analysis. The available history for the sector ETFs is short for testing alternatives (data snooping bias accumulates quickly with the number of alternatives tested in small samples).
In summary, evidence from simple tests over a limited sample period suggests that hedging a sector ETF momentum strategy with a broad market ETF (the loser sector ETF may improve (harms) risk-adjusted performance.
