This article was first published in Institutional Investor on 9 April 2020 here.
There’s a flexible new trend following strategy that is designed to offer equity downside protection across a range of timescales – even short ones.
A few months before the market shocks driven by the Covid-19 pandemic, II spoke about trend following with Philip Seager, Head of Strategy, Quantitative Investment Solutions at Capital Fund Management (CFM). At the time, Seager noted that there was a misconception on the part of some investors that the strategy’s primary purpose is to provide downside protection on equities. During that interview, Seager mentioned that his firm’s core trend following strategy provided “some equity protection against protracted downward moves” and that “at CFM, we have been developing a second trending program that tries to maximize the amount of protection and hedge that we get from trend following.”
In the light of recent market developments that have left investors wondering about their strategies and protection, II recently caught up with Seager to further discuss the status of the new strategy, what the motivation behind it is, and its benefits to investors during an equity market drawdown.
What was the driving force behind the development of this new strategy?
Seager: The motivation came from talking to investors who wanted a product that provides equity downside protection. There are two camps around trend following: those who look to it as an absolute alpha generator, and those who view it as a protective strategy. For the original strategy, everything is set up to maximize the Sharpe ratio to fully exploit the behavioral bias that exists. We realized based on our experience that we could offer an additional strategy to maximize the amount of protection, but to do that you’re going to reduce the Sharpe ratio – there’s no free lunch.
How has the new strategy evolved since its initial concept stage, and how does it differ from CFM’s original trend following strategy?
We set out to build a strategy that offers downside equity protection on a range of timescales. That means you’re looking at the performance of your program in the context of it being affected by a sizeable downside move over a timescale of a year, six months, a quarter, a month, and a week. When you are a mid- to long-term trend follower, you tend to get that protection over the timescale of your trend – and that’s reflected in our original strategy in providing some protection on longer timescales. We wanted to fill in the gaps, to go to a shorter term – even down to a week, where trend following normally doesn’t provide any protection because you have to wait for the crisis to persist in order to see the convexity.
The program has evolved to where it now has four layers. The first layer is our core trend following program – a solid foundation. That’s mid- to long-term, and it’s applied to a universe of 100-plus futures contracts and FX pairs. We expect that strategy to have a forward-looking Sharpe ratio of 0.5.
The next step is to apply a beta cap at zero – always an ex ante beta cap because we can’t guarantee that the beta in the future will be capped to zero and only allowed to go negative. However, based on the past, we can try and forecast what the forward-looking beta will be. In doing so, we’re looking at the beta of our portfolio relative to the equity benchmark that we’ve chosen, in our case one in which we have equal allocation to each equity index future that is in our 100-plus universe of futures contracts. It could, however, be anything – it could, for example, be the S&P 500. The client can come to us and say, I want to have a beta cap on this equity benchmark, and we have the flexibility to build that for them.
How is that constraint applied?
We impose the cap to make sure the beta is always below zero in the form of a constraint in our portfolio construction. The beta cap is likely activated when the markets are going up. In that scenario, the optimization procedures look for opportunities to bring the portfolio equity beta down by looking for things that are trending downward and are positively correlated with equities – and things that are trending upward and negatively correlated with equities – and it will try to skew all of the positions in the portfolio in order to maximize the gain with this constraint.
When markets are falling, you probably don’t need to impose this constraint because the equity beta of the portfolio is likely already below zero – since you’re trending on equity markets you have short positions. The constraint is only applied when equity markets are, generally speaking, going up, in which case such an equity beta cap reduces the sensitivity of the portfolio to sharp down moves in equity markets.
We covered two of the four layers you mentioned. What’s next?
As I mentioned, we’re trying to get convexity or protection on timescales that are shorter than mid- to long-term. To try to fill that gap, we included shorter timescales, on an approximately two-month timescale, but only on a selection of the most liquid equity index futures and the most liquid, defensive bond futures in order to reduce the execution costs of trading more quickly.
There are a number of very liquid futures on U.S. government bonds, and we also apply the short-term trend to German bonds – both are very defensive, and you do get protection from trending on those contracts. We also apply this layer to a handful of the most liquid equity index futures, such things as the E-mini S&P, with a constraint that we cannot go long on any equity index futures in this layer.Option to Options box out 2
And the final piece of the pie is …?
Trying to get protection on as short a term as we can get, covering that one-week timescale. We do that through a long position in VIX futures, an interesting market as these volatility futures are only exposed to vega risks rather than gamma risks. Vega describes the exposure to moves in the implied volatility, while gamma measures the exposure to big moves in the underlying.
In this case, strictly speaking, the VIX future P&L should only be exposed to moves in the implied vol. But there is something called the leverage effect that means that when markets go down, the volatility tends to go up. So, a dynamically hedged long VIX position gives you something that doesn’t lose too much money, and has the additional feature of rising when the market crashes.
You can think of it as having some of the protection of options with less of the negative drift. We think that the protection we get outweighs the potential forward-looking losses we would get from holding that position.
So, it sounds a bit like having an option to options?
We hope it’s a more cost-efficient option to options, yes. In the history of CFM, we have used options in this way. The problem with options is that you’re paying a premium for that protection along with the rest of the market, so the more people who buy the protection, the higher the cost of options goes and the more you will lose in terms of the negative drift. The price for that protection is very, very high.
What we’re proposing, on the other hand, is a product that’s a competitor or a complement to long vol options-type programs and strategies. Our strategy has a mechanical convexity feature to it, and there are reasons to believe the modestly positive performance of trend following is persistent. You’re really arbitraging and exploiting an effect that seems to be ingrained in markets – the markets do trend. When you contrast that to long vol options-types of programs, it is an efficient option to options.
Any description or information involving investment process or allocations is provided for illustration purposes only. There can be no assurance that these statements are or will prove to be accurate or complete in any way. All figures are unaudited. This article does not constitute an offer or solicitation to subscribe for any security or interest.