The folks at Sunrise Capital are some of the smartest in the industry, if you ask us. They have long been leaders in the space, with a fierce and admirable dedication to systematic trading and the vigorous research it requires. We’ve got a thing for strong argumentation and debate, and if you’ve ever heard their advocates speak at a conference or event, you know they make a strong case for the systematic approach.
But it’s their most recent work that has our office talking. Revisiting Kat’s Managed Futures and Hedge Funds: A Match Made in Heaven, penned by Sunrise’s Director of New Strategies Development Thomas Rollinger, updates the 2001 work of Professor Harry M. Kat of the Cass Business School- one of the more thorough explanations of the benefits of managed futures in a traditional portfolio allocation. But it also goes beyond Kat’s initial work, providing an in-depth look at how kurtosis and skew can be altered in such a way that make the risk profile of the overall investment portfolio far more attractive.
So, this week, our newsletter will take a look at both Kat’s previous work and the new Sunrise paper, explaining the differences and the major takeaways from the research, explaining how this can apply to your own investment strategy. In a time of heightened economic uncertainty, it’s only fitting to emphasize how, exactly, managed futures is what you truly need in your portfolio.
The First Take: Kat
In 2001, Professor Kat circulated a working paper entitled Managed Futures and Hedge Funds: A Match Made in Heaven. It acknowledged the desire of investors to protect against major economic events that would send a traditional stock and bond portfolio spiraling downward, and the previous trend of the affluent in seeking allocations to hedge funds. Kat, however, argued that there might be another alternative to consider- managed futures.
Kat’s work was significant, and demonstrated a strong benefit to be derived from a managed futures allocation. However, it was not without flaw. To begin with, the data selection was not beyond reproach. It used the S&P 500 to represent stocks, but capped returns to “adjust” for a belief that the bull market of the 90’s wouldn’t be repeated. It used the Salomon Brothers Government Bond Index to represent a bond allocation- but it was not as comprehensive and dynamic as a cohesive bond allocation today might be. In the hedge fund category in particular, there was substantial selection bias, as Kat used a hand selected portfolio of 20 hedge funds out of the thousands in the investing universe. Finally, for managed futures, Kat utilized the Stark 300 Index, which was comprised of the 300 top performing managed futures programs in the Daniel B. Stark & Co. database. So, while Kat’s analysis was strong, the data set was not ideal for making a balanced argument.
The Sunrise paper addresses these insufficiencies in their own data selection, opting to use the S&P 500 Total Return Index for stocks and the Barclay U.S. Aggregate Bond Index to represent bonds. While these are more conventional indices to use today for such an argument, it was in the alternative space that the data selection improved the most. For hedge funds, Sunrise uses the HFRI Fund Weighted Composite Index, which boasts 2,300 constituents. All of these programs have at least $50 million in AUM and 12 active months of reporting, providing the continuity necessary for a legitimate analysis of returns, but the broad base of programs reporting also provides a more comprehensive view of hedge fund returns in general.
The managed futures data selection is what makes the Sunrise analysis most unique, though. As Rollinger explains, “[M]any investors generically say “managed futures” or “CTAs” when they more precisely mean “systematic CTAs who employ trend following strategies,” likely due to the fact that many of the largest and most successful trading managers employ some variation of a trend following strategy.” As a result, this paper opts to focus in on systematic traders instead of the more broad managed futures universe. While Attain believes that allocations across a variety of strategies can create the most balanced and diversified portfolio, we advise clients to allocate the bulk of their portfolio to trend following managers, so this research is still significant, even if a more narrow argument.
To facilitate this focus, Sunrise uses the Barclay Hedge Systematic Traders Index. Even with the more narrow scope, this is, in our opinion, a better data set than that used by Kat. With more than 488 constituents, this equally weighted index provides an even wider field of data than Kat was able to use, with less selection bias and a more targeted approach – a clear boon to the analysis. While the piece itself focuses on the Barclay index, it is noted that they repeated the analysis with several other indices, with little variation in results (lending credence to our argument that criticisms of CTA indices are poorly founded, but that’s a take for another day).
Using this data, Sunrise looks at what has transpired since Kat’s original research, focusing on the period between 2001 and 2011, but they also extend their research to cover the original 1994-2001 period, as well. To fully explore the opportunities presented by managed futures, Sunrise explains the necessity of analyzing four critical components of performance and risk:
When building portfolios using the Modern Portfolio Theory (MPT) framework, investors focus almost solely on the first two moments of the distribution: mean and variance. The typical MPT method of building portfolios appears to work well, as long as historical correlations between asset classes remain stable. But in times of crisis, asset classes often move in lock-step, and investors who thought they were diversified experience severe “tail-risk” events. By only focusing on mean return and variance, investors may not be factoring in important, measurable, and historically robust information.
Skewness and kurtosis, the third and fourth moments of the distribution, can offer vital information about the real-world return characteristics of asset classes and investment strategies. The concepts of skewness and kurtosis are paramount to this study.
Skewness is a measure of symmetry and compares the length of the two “tails” of a distribution curve. Kurtosis is a measure of the peakedness of a distribution — i.e., do the outcomes produce a “tall and skinny” or “short and squat” curve? In other words, is the volatility risk located in the tails of the distribution or clustered in the middle?
All of this data may be useful in general, but it is particularly critical when examining the relationship between hedge funds and a portfolio, in particular. As the paper explains,
To understand how vital these concepts are to the results of this study, we revisit Kat’s original work. Kat states that when past returns are extrapolated, and risk is defined as standard deviation, hedge funds do indeed provide investors with the best of both worlds: an expected return similar to equities, but risk similar to that of bonds. However, Amin and Kat (2003) showed that including hedge funds in a traditional investment portfolio may significantly improve the portfolio’s mean-variance characteristics, but during crisis periods, hedge funds can also be expected to produce a more negatively skewed distribution. Kat (2004) adds, “The additional negative skewness that arises when hedge funds are introduced [to] a portfolio of stocks and bonds forms a major risk, as one large negative return can destroy years of careful compounding.”
Kat’s finding appears to be substantiated in Koulajian and Czkwianianc (2011), which evaluates the risk of disproportionate losses relative to volatility in various hedge fund strategies:
“Negatively skewed strategies are only attractive during stable market conditions. During market shocks (e.g., the three largest S&P 500 drawdowns in the past 17 years), low-skew strategies display:
- Outsized losses of –41% (vs. gains of +39% for high-skew strategies);
- Increases in correlation to the S&P 500; and
- Increases in correlation to each other”
Skewness and kurtosis may convey critical information about portfolio risk and return characteristics, something which should be kept in mind when reading this study.
The final important consideration in difference between the two papers is the type of investor considered. Kat looked at two kinds of investors: the 50/50 investor, and what he referred to as 33/66 investors. The 50/50 investors, by Kat’s explanation, would typically have half of their portfolio in stocks, and the other half in bonds, with any allocation to alternatives like hedge funds or managed futures reducing each allocation proportionally. The 33/66 investor would allocate in a similar fashion, but the split would be 1/3 of their portfolio in stocks and 2/3 in bonds. The Sunrise paper focuses exclusively on the 50/50 investor, which, in our opinion, is likely sufficient and more appropriate, as the investor focused on a 2/3 bond allocation is likely far more risk-averse than the 50/50 investor, making them an unlikely candidate for alternative investments.
So now that we understand the why and how behind the research, what exactly did they find, and why does it matter?
The first major comparison between the Kat and Sunrise papers comes in the comparison of the asset classes as standalones. While the data sets are different, the comparison is still eye opening:
(Past performance is not necessarily indicative of future results.)
You’ll notice that the mean for stocks, hedge funds, and managed futures are all lower, while bonds have increased. The standard deviation has increased for stocks while decreasing for the remaining asset classes. Correlations have remained largely the same, though- so what happens when you combine these into a portfolio?
To line up the results with those provided by Kat, the Sunrise paper explained:
In order to study the effect of allocating to hedge funds and managed futures, we form a baseline “traditional” portfolio that is 50% stocks and 50% bonds (“50/50”). We then begin adding hedge funds or managed futures in 5%-allocation increments. As in Kat’s original paper, when adding in hedge funds or managed futures, the original 50/50 portfolio will reduce its stock and bond holdings proportionally. This produces portfolios such as 40% stocks, 40% bonds, and 20% hedge funds, or 35% stocks, 35% bonds, and 30% managed futures. (Note: All portfolios throughout the paper are rebalanced monthly.)
And what were the results? Well, unlike the Kat analysis, Sunrise initially ignored the more lofty aspirations of a 75% allocation to hedge funds of managed futures, looking instead at a 0 to 50% allocation, and what it would do to mean, standard deviation, skew and kurtosis.
(Click to embiggen. Past performance is not necessarily indicative of future results.)
What we see here is something that we’ve been writing about for some time – that an increased allocation to managed futures- instead of hedge funds- provides better risk adjusted performance. As the Sunrise paper pointed out, “Adding managed futures exposure increased the mean return and simultaneously increased the negative skewness of -0.76 of the traditional portfolio to a positive 0.05 at the 45% allocation level. The standard deviation dropped more and faster than it did with hedge funds, and kurtosis also improved, dropping from 2.23 to -0.21 at the 40% allocation level.”
But, to be fair, investors don’t have to choose either or – they can combine managed futures and hedge funds in their allocation. As much as we may dislike hedge funds as an alternative to managed futures, the Sunrise paper makes a compelling argument for a blend, stating, “Due to their positive skewness and significantly lower kurtosis, adding managed futures to hedge funds appears to provide a substantial improvement to the overall risk profile. With 40% invested in managed futures, the standard deviation drops from 1.87% to 1.55%, but the expected return only declines by 2 basis points. At the same allocation to managed futures, skewness increases from -0.93 to 0.07, while kurtosis drops noticeably from 1.98% to -0.39. Hedge funds are impressive on their own, but managed futures demonstrate that they are the ultimate teammate by improving the return characteristics of the overall portfolio.”
Still, the meatiest part of the piece comes from the visualization of individual metrics when considering, not only various allocations to alternatives (a blend of managed futures and hedge funds, in this instance), but at what blend of alternatives you see which results.
(Past performance is not necessarily indicative of future results.)
Talk about some beautiful charting. The cliff notes version of the results? While alternatives exposure does decrease the average monthly returns of an overall investment portfolio, it increases skewness, particularly when the alternative exposure blend includes a hefty allocation to managed futures, and a similarly positive impact on kurtosis, decreasing it to a negative level. Translation? Much lower volatility with only marginally reduced returns.
Perhaps our favorite graph, though, was their version of the efficient frontier. It may not have the color scheme of the others, but what it lacks in complexity it makes up for in punch. We have, in the past, demonstrated what an efficient frontier might look like when considering stocks, bonds and managed futures, but when you add hedge funds into the alternatives mix, the results might take you aback:
(Past performance is not necessarily indicative of future results.)
Yes, you read that correctly. One of the most efficient investment portfolio blending these asset classes includes only a ten percent allocation between stocks and bonds.
The Sunrise paper does an excellent job of updating the research, and for many, the conclusions that alternatives should be the bulk of one’s portfolio may be surprising. But there are a few things to consider regarding that conclusion. For one, this data is from 2001 through 2011 – a period not exactly kind to stocks and a short period of analysis at that (however it was building on a study covering the 7 years prior to that). The second, more nuanced point, is that labeling hedge funds as “alternatives” is a little misleading. As we have frequently written about in the past, hedge funds tend to be highly correlated to stocks, which makes them suspect as an alternative investment in our minds (and is further articulated by the skewness analysis above), but at the same time makes them a great substitute for stocks.
So what we’re really seeing in the Sunrise work and the Kat work before it is an argument for replacing your stock exposure with more efficient stock exposure through hedge funds. It’s not about doing a large allocation to alternatives as much as it is about making a smarter stock allocation. Instead of buy and hold, use hedge funds which give the stock exposure with better risk controls, in our opinion. The managed futures portion, meanwhile, remains the true alternatives allocation and diversifier of the portfolio; at levels we have seen as valid in our own efficient frontier analysis.
Finally, some who look closely at the efficient frontier graph and conclusions of this type of work may ask what the point is. The ideal portfolio in this work results in a portfolio with annualized returns just 1% or so higher, and risk in terms of volatility just 3% lower – not the type of earth shattering difference likely to push many investors one way or the other. It’s not like we’re talking about half the risk and twice the return here. All we’re talking here is a game of inches, which likely matter to billion dollar investors, but maybe not so much to less well heeled investors.
But risk in terms of volatility (and skew and kurtosis as Sunrise points out) is just one part of the equation. What is likely to push investors more so than those small differences are larger differences in more telling risk statistics like maximum drawdown. This is where pushing more of a portfolio into managed futures and hedge funds really pays dividends in our opinion. Consider the following table which compares a ‘normal’ portfolio as reported by Kat with the optimal portfolio found by Sunrise.
That’s a smaller maximum drawdown AND larger compound ROR- something real life investors are likely to care a lot more about than an incremental increase in return and almost too small to notice reduction in volatility of the portfolio.
No matter where you fall in the investing spectrum – whether you are an alternatives enthusiast or someone just beginning to consider dipping their toe in – this data should give you something to think about. We know it got our minds spinning. Thanks again to Sunrise for the excellent research – they’re certainly a program to watch.