From Steben & Company – Our study shows that style factors including volatility targets, speed and sector exposures explain many of the historical short-term performance differences among trend followers
Also see
- Q3 2016 hedge fund letters
[drizzle]
Managed futures have become an alternative asset class that is widely used by investors seeking overall portfolio diversification and absolute returns independent of the direction of broad equity and bond markets. The most common managed futures trading strategy is trend following, a strategy that attempts to exploit momentum in more than 200 global futures markets (including commodities, equities, fixed income, and currencies) by taking long positions in rising markets and short positions in falling markets.
While investors have embraced the potential benefits of managed futures, the causes of the large performance dispersion among trend following commodity trading advisors (“CTAs” or “managers”) are not well understood given that their trading programs are conceptually similar. The research team at Steben & Company set out to find answers.
From 2006 to 2015, the annual performance gap between the year’s top and bottom quartile trend followers averaged 28 percentage points, a very wide margin (Chart 1). Furthermore, in 2008, when investors were most reliant on trend followers to deliver performance to offset stock market losses, the performance gap between top and bottom quartile trend followers grew to 55 percentage points. With managed futures, we believe, manager selection is paramount.
A comparison of historical performance is an obvious starting point for investors seeking to evaluate trend followers. However, investors will find that returns varied significantly from year-to-year. The question then is whether the stronger performing manager in a particular year had a superior trading system that could lead to better than average returns in future years. Unfortunately, the answer is often no. As you can see in the chart (Chart 2) below, the top quartile trend following program in one year did not usually stay in the top quartile in the next year. There was little short-term persistence in the performance rankings.
Despite the lack of persistence, there are major differences in the quality and sophistication of trading systems among trend followers, and these can manifest themselves in long run performance differences. Managers with a stronger focus on research, risk management and trade execution are more likely to have better performance when measured over multiple market cycles. But in a single year the wide dispersion in managed futures fund performance cannot be explained solely by differences in manager skill or “edge.”
Instead, we found that style factors explain most of the performance differences between managers in a given year. Just as long-only equity funds may have a particular style tilt (value vs. growth, large cap vs. small cap), trend following CTAs also have style biases. In the traditional investing world, a small cap growth equity fund manager may outperform a large cap value manager in a particular period, not because he or she has greater stock selection skill, but simply because small caps outperformed large caps and growth outperformed value. Similarly, in managed futures, a trend following program’s style choices can be the key drivers of short-term relative performance compared to its peers.
In trend following strategies, we believe the three most important style differences are volatility targets, speed and sector exposure.
We will explore each of these style factors in turn, and then conclude with our thoughts on how investors should consider allocating capital to trend followers given the different styles that are available.
Style Factor 1: Volatility Targets
One of the most fundamental differentiators within any investment strategy is the amount of risk assumed by a manager. Traditional long-only strategies generally do not target specific risk levels, but passively bear the amount of risk generated by the broader market. For example, an equity mutual fund is generally fully invested in stocks and takes on the volatility of the broad stock market. In contrast, trend following strategies typically target a specific level or range of volatility as an inherent element of the investment process. They actively adjust the size of their futures positions in inverse proportion to changes in market volatility in order to achieve that target.
The key aspect is that different trend followers target different levels of volatility. This target is a purely subjective choice on the part of the manager. A higher volatility target is achieved through higher leverage. The chart (Chart 3) below shows the very wide range of manager volatilities, as represented by their annualized standard deviation. The median annualized standard deviation of trend following programs was 14.6% (close to the S&P 500’s volatility of 15.1% over the same time frame). While the most popular range for volatility is 10%-15%, there were plenty of programs with an annualized standard deviation in excess of 30%.
Any given manager can increase a program’s target rate of return (assuming it is positive) by increasing leverage and hence increasing volatility. The next chart (Chart 4) presents returns of two simulated trend following programs. Manager A exhibits a higher return between 2006 and 2015. Some investors might consider Manager A superior to Manager B, however, these simulations are identical in every way, with one important exception. The simulation for Manager A targets 20% annualized volatility whereas Manager B runs the identical program with an annualized 10% volatility target. Manager A generated its outperformance from simply taking twice as much risk over the life of the investment. This emphasizes the importance of looking at risk-adjusted returns in making any comparison between trend followers.
In managed futures, there is a close relationship between leverage and volatility. For a given trading program, doubling the leverage generally doubles the standard deviation of returns, all else equal. Managers have a lot of flexibility in setting their leverage levels since futures are traded on margin. In managed futures, leverage is usually measured by the margin-to-equity ratio, which is essentially the percentage of the net capital of the fund that is being put up as margin to support its positions in futures contracts. The higher the margin-to-equity, the higher the leverage and hence the higher the volatility. A reasonable rule of thumb for a diversified trend following program is that x% margin-to-equity translates into approximately the same x% expected annualized standard deviation of returns (give or take a couple of percent). So 10% margin-to-equity would imply roughly a 10% standard deviation; 20% margin-to-equity would imply a 20% standard deviation, and so on.
The charts on the following page (Charts 5 and 6) reflect the performance and volatility of every trend following program in the Barclay CTA database during an example of a strong and a weak period for managed futures. In good periods for trend following (such as 2008), more volatile managers tended to do better. In contrast, in difficult periods for trend following (such as the period from May 2011 to April 2012), more volatile managers generally lost more.
For investors, the appropriate level of volatility depends on their tolerance for losses. If an investor wants to allocate to a particular trend following program, but feels its volatility target is too high, they have the option to allocate a smaller amount and hold cash against it, so the blended CTA plus cash position hits the desired volatility level. For a portfolio of multiple managed futures funds, it can be helpful to size manager allocations in inverse proportion to their volatility to balance the risk contribution of each manager. Again, if the mix is more volatile than desired, cash can be held alongside the position to bring the blended volatility down.
Article by John Dolfin, CFA & Christopher Maxey, CAIA – Steben & Company, Inc.
See the full PDF below.
[/drizzle]