If the number of invested sectors drops to three, those sectors will be invested equally in the market at 25% each, with the remaining 25% going to cash. And each further dropped sector is allocated to cash.
At the climax of the financial crisis in October 2008, AlphaSector went to 100% cash, and stayed entirely in cash until the first week of April 2009. "That was an investment of conviction. That was purely quantitative, objective. As it turned out, that was a good decision," Present says.
The investment process involves a momentum or trend-following approach. It looks at each ETF's historical returns. It also calculates a proprietary measure of volatility, including intraday and closing price volatility.
F-Squared treats volatility not as a sell signal, but as a risk factor, using a technique Present calls a "dynamic volatility window." This means that when a sector's volatility is low, the team can look to the longer end of the intermediate horizon. But as volatility increases, the horizon shortens, heightening the analytical engine's sensitivity to current events and the speed with which it will signal a sell decision.
For example, when F-Squared sold out of financials in July 2007, that sector had been declining for nearly three months during an extended low-volatility investment period. A year later, volatility skyrocketed. F-Squared owned energy, which peaked between the first and second week of July 2008. It sold out of the sector just four and a half weeks later. "We moved so much faster in 2008 for energy than we did for financials in 2007 because the volatility had gotten so much higher," says Present.
Thanks to the engine's dynamic flexibility, the strategy lost less than 2% in 2008, and beat the market in 2009 and 2010. In 2011, it finished up 1.7% versus the S&P's gain of 2.1%, but with roughly half the volatility of the index for the full year because the strategy went aggressively to cash in the second half.
Present acknowledges that the strategy may perform less than ideally under certain conditions. It will take a hit if there's an external shock to the system, such as the 9/11 attacks.
The strategy will also lag anytime there's a hard rally off a market bottom because it doesn't try to time the market. For example, when the strategy was 100% in cash between January 1 and March 7, 2009, it was up 25% on the S&P because the index was down 25%. Then, on March 9, the index took off. "We didn't chase it right away," says Present. "Our goal is to wait for the market to de-risk enough for it to be safe to get back in." F-Squared maintained its 100% cash position in March, and averaged about 50% cash in April. Then in May, it had no cash. "Because of that, we lagged in March and April."
And the strategy will lag during a sector bubble: If the best-performing asset class in the S&P is the largest or one of the largest weighted asset classes, the strategy will lag because the ETF portfolio is equal weighted. Recall 1999 when technology, representing 40% of the S&P, was up 65%. If the company had been around then, "we would have gotten the 65% from the tech sector, but we would have had only a 12% to 13% allocation. Clearly, we would have lagged on the upside."
Present says that whenever the index adds more than 25% over a 12-month rolling period, clients are told the strategy will fall behind. "We're not going to apologize for this. It's not about returns in the traditional context of returns. Our sweet spot is when the market goes down. Anytime the S&P has a 12-month period when the market is negative, we do extremely well. That's our value proposition," Present says.