…some investments take on the status of an asset class simply because the managers of these assets promote them as an asset class. They believe their investors will be more inclined to allocate funds to their products if they are viewed as an asset class rather than merely as an investment strategy.
Monday, May 09, 2011
Is Managed Futures an Asset Class?
A few weeks ago we posted on Mack Frankfurter of Cervino Capital’s recent discussion of managed futures and whether or not it is an asset class. Mack argues that it is not, resting his case on the logical fallacies prevalent on the other side of the debate. When considering why managed futures were often described as an asset class, Mack’s explanation was simple: people are lazy.
Mack’s Cervino Capital is one of the CTAs we track, and we have a lot of respect for him and his program, so his comments gave us pause. Mack is not alone in this assessment. As Mark Kritzman of Windham Capital Management speculated in 1999,
We frequently refer to managed futures as an asset class on our blog and in our newsletters, so we had to wonder… were we just being lazy? Is “asset class” the wrong descriptor for managed futures?
We decided to research the idea to make our own, informed determination on the subject, but realized as we began our quest that a key ingredient was missing from the equation. It was absent from almost every piece we found, including Mack’s.
We were missing a good definition of an asset class.
Defining an Asset Class
There are a couple of problems with pervasive definitions of asset classes. For one, half of them define asset classes while using the word “asset,” and any good debater will tell you that this practice leads to confusing and vague explanations. The other half of available definitions seem to parametricize asset classes to stocks, bonds and cash, and while most figures in the financial industry will identify these common investments as asset classes in their own right, examples do not a definition make.
If you search beyond the surface definitions, you will find more nuanced and complicated explanations. Even here, finding a warranted explanation was difficult. Many authors simply asserted qualifications for asset class consideration without providing any kind of logic as to why those qualifications were pertinent to the discussion.
To save you from a 200 page literature review, we’ll generally define an asset class as: A grouping of investment opportunities that behave and are treated in a similar manner.
The explanations for this definition presented by Anson, Fabozzi, and Jones (2010), Focardi and Fabozzi (2004), Gibson (2008), Winograd (2004), Hall (2004), Considine (2004), and Swensen (2005) give versions of the following specific qualifications of an asset class:
1. The asset class should be subject to similar regulatory guidelines.
2. The asset class should have a significant amount of capital allocated to it.
3. The performance of the asset class should not be able to be replicated by other established asset classes.
4. Risk and return characteristics of all portions of the asset class should be similar.
Why these qualifiers? We selected these 4 test points for three reasons. First, traditional asset classes were able to pass all four tests. Second, a plurality of authors described each of these points or a variation therein as a defining factor of an asset class, providing a comfortable consensus for each point. Third, each of these four points, unlike many others proposed in the literature base, had multiple explanations attesting to their significance as a test of value, which gave them more merit than the qualifiers that were merely asserted.
In order to determine whether or not managed futures was an asset class, we evaluated whether or not it met each of these four standards
Testing the Definition
1. Are all portions of ‘Managed Futures’ subject to similar regulatory guidelines?
Managed futures has this one locked up pretty good, with managed futures comprised of those programs managed by professional money managers who are registered as Commodity Trading Advisors (CTAs) with the Commodity Futures Trading Commission (CFTC) and National Futures Association (NFA). This registration mandates strict levels of disclosure and reporting, and is applied to nearly all forms of managed futures programs, no matter what strategy they may implement(some very large programs and foreign programs may not register).
But despite some exceptions, no matter if a manager is trading stock index options or Corn spreads, they are subject to the same regulations.
Does managed futures pass this definition of an Asset Class? YES
2. Does ‘Managed Futures’ have a significant amount of capital allocated to it?
BarclayHedge puts assets under management (AUM) at the end of 2010 for Managed Futures at $267.6 Billion. This is after an explosion in managed futures investing following the 2008 financial crisis, and leaves managed futures as the single largest investment strategy amongst hedge fund strategy types (merger arbitrage, global macro, private equity, etc)according to the BarclayHedge data (although some larger funds labeled as managed futures programs in the BarclayHedge database are not solely dedicated to the space)
[The data here is, in part, from new investment, and, in part, from investment gains]
Since 1980, AUM for managed futures has increased over 863%, with an average increase of 27.7% per year. Some managed futures commentators are even predicting that AUM for CTAs will pass $3 trillion by 2021. Whether that happens or not, the data shows that Managed Futures have this requirement easily covered.
Does managed futures pass this definition of an Asset Class? YES
3. Can the performance of ‘Managed Futures’ be replicated by other established asset classes?
Managed Futures are a bit of a slam dunk on this front as well, with their whole purpose in life (seemingly) to perform when other asset classes are not. If they don’t follow what other established asset classes do, they can’t be replicated by those asset classes. For proof of this, one need only look to 2008, when managed futures were up 18.33%, compared to bond returns of 10.77%, and stock market losses of 38.49% [past performance is not necessarily indicative of future results].
For the more statistically inclined, performance replication, in this instance, is best measured in terms of correlation. As a refresher, correlation is a statistical figure with values which range between -1.00 and +1.00, meant to show how inter-related two sets of data (in this case monthly % returns) are. If they have a correlation of 1.00, they are perfectly correlated, meaning when one market rises 1%, the other will do the exact same, and when one loses -1%, so will the other. If they are at -1.00, they are exactly opposite; with one making the exact opposite amount the other loses each day, and vice versa.
Given that the only three asset classes that everyone seems to be able to agree on are stocks, bonds and cash, we calculated the rolling correlation coefficient of managed futures against the monthly performance of each of these classes over the last 15 years to see how managed futures stack up for this definition of an asset class.
[Past performance is not necessarily indicative of future results; Data Source: Stocks = S&P 500; Bonds = Citiworld Bond Index; Managed Futures =BarclayHedge Managed Futures Index]
As we expected, the calculation of 12 month rolling correlations going back over the past 15 years showed the wildly oscillating lines typical of non-correlation. In fact, over the 15 year period, we found managed futures had a correlation of -0.078 to Stocks, -0.140 to Cash and 0.296 to bonds. Statistically speaking, this tells us that the performance of managed futures cannot be explained by the performance of traditionally accepted asset classes. In other words, you cannot replicate managed futures performance with these other asset classes.
Does managed futures pass this definition of an Asset Class? YES
4. Are the risk and return characteristics of all portions of ‘Managed Futures’ similar?
This is the hardest definition for managed futures to meet, as managed futures spans investment strategies as diverse as long term trend following of bond and currency futures, to naked selling of stock index options, to spread trading of grain and energy markets.
Statistically speaking, many of those different types of managed futures investments ARE NOT similar. In fact, many pride themselves on their non-correlation with the main managed futures indices, and market that fact.
At the same time, many of these programs are also not correlated with other asset classes, especially during stress periods for the traditional asset classes. This is because, despite their non-correlation with each other, they generally still share the most common denominator in terms of risk and reward characteristics for managed futures programs - their long volatility profile.
This long volatility profile is what allowed the majority of managed futures programs we track at Attain to post profits in the incredibly volatile markets of 2007 and 2008. One strategy type which does not share this long volatility profile (by design), is option trading. But does one outlier on one of four definitions of an asset class nullify all of managed futures as an asset class? We say no, others may say yes… But consider that not all stocks perform similar, either; nor all bonds (just look at Portugese debt versus US debt right now). If we could kick just kick option sellers out of the managed futures space, and give them their own registration category and regulatory body – we would solve a lot of the issues in meeting this definition.
For more of an outside the box look at managed futures having the same risk and reward characteristics – we can look at the liquidity and transparency of managed futures program (no matter what type of strategy they employ) as key component which links them as an asset class. The ability to see position level detail, to have individually managed accounts, and liquidate an account in hours to days are common themes amongst all managed futures programs (and indeed several of those characteristics are the result of the programs being under the same regulations and rules of conduct).
Does managed futures pass this definition of an Asset Class? MAYBE
At this point, we are ready to confirm managed futures is an asset class, meeting three of the four criteria handily, and suffering only a single outlier on the fourth definition. But we were of the opinion it was an asset class before writing this, and this is a hotly debated topic, so we don’t mind addressing some possible arguments we’ve seen out there against them being an asset class.
Objection 1: Not all managed futures programs react the same way to market events.
This is the ringer for the non-asset class camp, as we discussed above, as, at the base level, there is no way to defend it. Stock index selling managed futures programs, for example, don’t react the same to market crisis periods as traditional managed futures programs, and are more highly correlated with stock market performance than that of the managed futures indices.
Our best rebuttal is that not all stocks and bonds (asset classes by nearly all accounts) react the same to market events, either. For example, airline stocks may fall after a spike in Oil while oil companies stocks rally on higher revenue projections. In bonds, you can easily imagine a bond backed by the revenues of a toll bridge, versus a bond backed by the profits of a shipping company – reacting very differently to market events, and indeed, being not very correlated with one another. How correlated have Greek government bonds and US government bonds been over the past two years, for that matter? We would assume the correlation is at about -0.98 (out of a maximum of -1.00).
Yet despite obvious examples of outliers in the stock and bond asset classes, nobody questions whether they are an asset class.
Objection 2: Managed futures trade stock indices, bonds, and currencies – not just commodities; meaning they are more a part of the hedge fund asset class (essentially global macro funds) than their own.
This is a hasty generalization, in our opinion, that rests upon assumptions about hedge funds even being an asset class. There are two responses here. First, the performance of hedge funds and whether or not they are an asset class has nothing to do with whether or not managed futures are an asset class. Given the fact that managed futures meets the tenets outlined in our definition, our conclusion, we feel, is well-warranted. Second, if you look at hedge funds, one particular tenet of our analysis indicates why they are not included in the asset class, technically.
Hedge funds have been difficult to peg. They operate in a realm where disclosure is often optional, and as such, sufficient data from which to draw conclusions is difficult to find. That being said, Professor Harry Kat of the Cass Business School at City University, London, published a paper in 2007 that collected every shred of data available, and he concluded that hedge funds, more often than not, “have a very high correlation with the stock market and therefore make lousy diversifiers.”
We recently reposted commentary from Welton Investment Corporation in this space as well, which did an excellent job of explaining this diversification problem, showing that most hedge fund strategy types correlate more closely with equities (the stock asset class), that any other asset class. Interestingly, they found that the Global Macro hedge fund strategy type, which those who object to managed futures as an asset class argue should be the category managed futures is assigned to (instead of their own asset class), is more closely linked to managed futures than other hedge fund types.
It’s as if managed futures and hedge funds should do a trade, Global Macro for Option Sellers, putting both of those investment types into asset classes which more closely mirror their risk and return profile.
We can hear the objections now. “You’re saying hedge funds can’t be considered asset classes because they invest in a traditional asset class and, as such, have a high correlation. Couldn’t that be said of commodities and managed futures?”
Objection 2a: Managed Futures invests in commodities and as such is highly correlated (i.e not different) than the commodity asset class.
If you look at commodity performance, the only objective way to do so is to look at it from a long-only perspective as well. This is another way in which managed futures separates itself from the pack, because they can go long and short, making correlation between managed futures and commodities next to impossible. At times managed futures will be correlated with commodities (when they are long commodities), but at times they will be nearly perfectly negatively correlated (when they are short).
Just take correlation between managed futures and the S&P GSCI- which measures returns across commodity sectors from a long only perspective- over the past 15 years. The average correlation we found was only 0.228.
[Past performance is not necessarily indicative of future results]
Objection 3: Managed Futures performance is not easily replicated within the managed futures space.
While replication of the asset class within the asset class was not a definition we happened upon, it is an objection we come across from time to time. And it makes sense. After all, we can pretty easily buy S&P 500 futures, a total stock market mutual fund, or a sector ETF to replicate the stock asset class, and bond futures, funds, or ETFs to replicate the fixed income asset class.
But there aren’t managed futures futures (for those reading at the CME – Attain gets the royalty if you come out with this), and despite the attempts by an increasing number of firms, there haven’t been mutual funds or ETFs which have had success matching the performance of the managed futures indices.
From an individual investor standpoint, this takes on more meaning, as they may be invested in a $50,000 program such asClarke Global Basic- which is down for the year- while managed futures as an asset class is up. Or they may have invested in managed futures in 2008 via an option selling program, only to find steep losses at the end of the year while managed futures as an asset class were up.
This frustration is real, and this objection to managed futures as an asset class is the most damning, in our opinion, as we know how hard it can be for a single investor without a lot of capital to replicate the performance of the managed futures indices in any one month, quarter, or year. The less capital an investor has, the harder it usually is for them to replicate managed futures performance as a whole – because they are forced to rely on a single strategy type. The odds of any one strategy type varying from managed futures as a whole are pretty large.
Our rebuttal to this would be that the same thing can happen in all asset classes, where the performance of a single stock may greatly lag the general market, or a single bond sees its value go to zero on a corporate default while bonds in general post decent returns. The one point we don’t have a rebuttal to is the easy access to the whole asset class via an affordable ETF or mutual fund, like you can get with the SPDR.
The status of managed futures as an asset class has long been in debate, and it is unlikely that this publication, or any other, will put the issue to rest anytime soon. However, based on the literature we’ve reviewed, the data we’ve compiled, and the analysis we’ve conducted, we feel confident in our assessment of managed futures as a distinct asset class.
And despite our point by point methodology outlined above, the best argument for it being an asset class is usually the rather simplistic statement – if it isn’t an asset class, then what is it? We have yet to hear a good response to that question.
Does the conclusion fly in the face of what people understand to be true, intuitively, about asset classes? Did we skip some points and objections such as benchmark analysis? Sure, but until academia comes up with a better definition for what an asset class is, managed futures, as we understand it, is an asset class.
The real question here is what that conclusion means for potential investors. If managed futures are indeed an asset class, then they should not just be the realm of sophisticated investors and those with risk capital. They should be right up there with the seemingly carved in stone 60/40 stock and bond split that so many of us have been taught and told time and again as the optimal asset allocation mix.
The implication of managed futures being an asset class is that we need to get out the stone grinder and change that stodgy old 60/40 into a more reasonable 33/33/33. We aren’t the only ones who feel this way. In fact, Opalesque recenty published an article that argues that managed futures could actually replace bonds in the traditional portfolio.
When it is all said and done, the whole point of the debate is to decide whether managed futures actually add anything to a portfolio, or whether they are just another way to get the risk and reward profile you get with the traditional asset classes. We’re here to say with confidence that they do a lot more than the traditional asset classes. They will have losing periods, they will run in tandem with the other asset classes at times, but they do meet the definitions of a separate asset class, in our opinion, and, as such, demand consideration as part of a diversified portfolio.
While your stock broker and investment advisor will likely go on and on about the risks involved in futures (and there are substantial risks), we will say that the bigger risk lies in ignoring managed futures as an asset class.
Posted by Bud Fox at 4:19 PM