•  
• Tweet
•  
•  
• 

Newedge's Brian Walls and Galen Burghardt discuss zero correlation, managed futures liquidity and CTA evaluation

Brian Walls and Galen Burghardt are considered among an elite group of managed futures researchers. The pair authored the book: Managed Futures for Institutional Investors: Analysis and Portfolio Construction. Both work in the prime brokerage unit of Newedge.

Mr. Walls is the Global Head of Alternative Investment Solutions and US Head of Prime Clearing Services at Newedge. As the Global Head of Alternative Investment Solutions, he leads all Investor Research, Capital Introductions and related brokerage services. He is the Chairman of the Newedge Index Committee, the AIS Global Executive Committee and the US Prime Clearing Services Executive Committee. In addition he sits on the US Credit and Risk Committee, Sponsored Direct Market Access Committee and the Newedge USA Executive Committee. He is co-creator and host of several innovative research symposiums that offer a unique opportunity for top fund managers to meet with and convey up-to-date research information to institutional investors in an academic yet informal atmosphere.

Mr. Burghardt is Director of Research for Newedge USA, LLC. Mr. Burghardt is lead author of The Treasury Bond Basis and The Eurodollar Futures and Options Handbook, which are standard texts for users of financial futures. He was an adjunct professor of finance at the University of Chicago's Graduate School of Business (now the Booth School). He was the head of financial research for the Chicago Mercantile Exchange, and gained access to the world of futures through his work in the Capital Markets Section of the Federal Reserve Board. His PhD in economics is from the University of Washington in Seattle.

This is the first of a two part interview.

Opalesque Futures Intelligence (Mark Melin): Tell me how you two work together conducting managed futures research.

Brian Walls: We work well together because we come from different perspectives. I am more of a practitioner and Galen is our academic. I want our work to be thorough, objective and robust, but useful. Galen brings in a career of proper and classic analysis and teaches us how to apply them to what appear to be new problems
As an example, in the book I get most excited about the introduction and the practical guide because I feel like when it comes to professional institutional investors, they might have their own opinions on portfolio construction. Many of those things that Galen covers are usable and practical in any portfolio exercise not just managed futures. But the introduction and the practical guide really break down some of the misconceptions. For instance, in the beginning of chapter one we discuss returns. In managed futures, there is no coupon, there is no dividend, there is no maturity, there is no true return. The return is just the trading P&L bolted on top of some denominator (used as a margin deposit).

"In a direct managed futures account investors can choose whatever denominator they want¢â‚¬¦ and can determine the volatility level."
Brian Walls

In a direct managed futures account investors can choose whatever denominator they want. The investment (margin deposit) can be denominated in gold or whatever currency they choose. They can also determine the volatility level of the investment to suit individual needs.

Audited Self Reporting

Another practical thing to understand is where to find CTA performance data. Industry wide CTA data is not like something pulled out of the Wall Street Journal like stocks traded on exchange; direct accounts are not mutual funds or ETFs. A direct managed futures account it is self-reported return and choosing to report is voluntary. However, it is important to understand the nature of self-reporting, is it audited? Is it non-audited? Not all managers report to these databases. There is a fairly big overlap in terms of the availability of data. There is not a 100% overlap; it is not like an apples to apples comparison sometimes.

So understanding where to find the data, understanding what returns are and trying to go through the frequently asked questions show how many managers they need, what is the difference between the fund of funds and a platform. If you are an institutional investor considering investing into this space, to some extent you could read the introduction and part one and feel much more competent in understanding the mechanisms of the industry and how these things will fit into your organization.

Mark Melin: This leads to the question: Is managed futures performance reporting to be trusted? We both know that the NFA audits not only the performance of members, but view the reporting of performance to databases to be promotional and subject to review as well. It would be fair to say then that the Newedge CTA Index is a reasonable reflection of the top CTAs in the industry without considerable study bias. The performance has been audited by the NFA, everyone is an NFA member. The Newedge CTA Index is similar in this regard to the Barclay BTOP50 and the Altegris 40, to name of few of the large cap indices.

Brian Walls: Yes, in effect the same. We use BarclayHedge as a calculation agent, so every manager sends their returns independently to BarclayHedge and independently to Newedge. We both calculate an index value and then compare it and if we come up with a difference, we find it that day and then in addition, we both reconcile the daily return to the manager's reported monthly returns every month, ensure that our dailies add up to the reported monthly, because it is the monthlies that are audited.

We do this work at Newedge as well so that we have two independent verifications. It is important to note, however, that daily returns reported by Newedge are just estimates, they are not audited. So you have to know that a daily reporting is not done with as much care as a monthly audited number.

OFI: I find it interesting the benchmarks Newedge has developed. On page 124 of your book you have the Newedge Trend Indicator. How is that used in practice?

BW: It is a good question. The Newedge Trend Indicator is something we developed as a benchmark for trend following. We found that many investors were just choosing other trend followers for a comparison basis. We thought this was a flawed process because you do not know how those managers are doing, what their strategy is, you do not have full transparency on the strategy or the process or the weighting mechanism. In other words, you are comparing them to another unknown.

We decided to distill the very essence of a trend following model, which is fairly well documented and published, but we started from scratch just because we learned a lot through the experience. We concatenated the futures contracts to create price history. We had to understand how roll rules work, how you roll from one contract market to the next, and the rules to be consistent across all markets. Then we volatility weighted the markets in the portfolio in a way that managers tend to weight it which is to maximize diversification as much as possible. After the ideal weights, we looked at which markets would need to be constrained given the size of the markets and like I said pretending we had $2 billion. Lastly we tried to select a single parameter which would be consistently the highest correlation to the big known managers. We ended up choosing the 20 day, over 120 day moving average crossover. It is a single parameter, always-in system, weighted approximately the way the industry managers are weighted and has a fairly high correlation to them. It is running at like a 0.7 to 0.74

It gives you a pretty good idea of approximately how they are positioned and approximately where their P&L is coming from. It is not optimized for returns, so it significantly under-returns the managers. It is not optimized for transaction expense and it is not optimized for roll cost, and we never really put any effort into trying to make it to have high returns, the goal was to have a high correlation.

So that way managers, investors and the industry can use it as a transparent standard benchmark. We have received great feedback. A lot of people are using it and asking questions about it.

OFI: When you are talking about performance analysis, I like consider how a particular strategy or CTA performs against a benchmark. So having the ability to compare a trend following CTA to this benchmark I would imagine is useful.

BW: It is useful especially if you have daily data on the two of them over time, you get a better feel. The problem with it is you just need so much data to really perform those analyses. You are kind of almost better off just getting a better understanding of the manager's research department because I do not know if you live long enough to have an update to come up with a trustworthy comparison.

OFI: Galen, tell me how you look at your book. If you could have re-written any aspect of the book, what would it be?

Galen Burghardt: I wish I could write the introduction to the book over again - at least the part that dealt with how much an investor should invest in CTAs. In the book, we used historical returns, volatilities, and correlations to see what a good portfolio would look like, and it contained a lot of bonds, almost no stocks, and a pretty substantial allocation to CTAs. But this was because bonds had done so well, stocks had done so badly, and CTAs has done pretty well and really were uncorrelated with bond and stock returns.

When we were preparing a talk for the CERN Pension Fund Conference last year, though, Brian argued that we really ought to start thinking about the world from the standpoint of the pension funds. What were they starting with? What could they reasonably expect conventional assets to do? We began with a 60/40 stock/bond portfolio, which is pretty standard in the pension fund world. Sharpe ratios for stocks and bonds that were more in line with what one might reasonably expect. From there, we considered allocations of anywhere from nothing to 50% and asked, at each level, just how good CTAs would have to be for each allocation to be optimal? The reason we stopped at 50% was that it was not until we reached that level that CTAs' risk-adjusted returns would have to be better than they really were.

Today I think we have a much, much more compelling case, one that does not depend on history that is unlikely to be repeated. The problem we wrote is we have just finished something like 30 years of a seemingly unstoppable bull market in bonds. And the past decade has been truly horrible for stocks - two 50% drawdowns that lasted several years each. Any reasonable investment thinking has to allow for the fact that stocks should do better than they have, and that is no way that bonds can repeat what they have done over the past two or three decades.

OFI: In the case of bonds, not at these low yield levels.

GB: That's right. What I like about the framework we put together for the CERN conference is that it lays out really clearly the assumptions you have to make when you decide how to allocate your investments. Investors might not believe the 0.4 Sharpe we used for bonds, and you might not believe the 0.25 Sharpe we used for stocks. But no one is going to think we were delusional when we made them up. Moreover, we don't really need the Sharpe for CTAs to be as high as it has been for a large allocation to make sense. But as we learned, CTAs don't have to deliver spectacular performance for an allocation of 50% to make a ton of sense.

"CTAs don't have to deliver spectacular performance for an allocation of 50% to make sense."
Galen Burghardt

OFI: Well, that is interesting, 50%; that is the first time I have heard that number¢â‚¬¦ publically anyway. Equity investments are very risky at this point in economic history, but that is my own personal bias.

GB: They are. Actually, they are worse than risky. But that is my own personal bias.

OFI: I would imagine that is when the correlation started to come in the play too, once you get up to 50%.

GB: Exactly right. The entire case rests on low correlations combined with some expectation of positive returns. We did a very neat piece that was published in Futures Industry Magazine called Managed Futures and Pension Funds: A Post-Crisis Assessment. Even if I do say so myself, it's a highly readable and important piece. The foundation of that note - and for that matter, the foundation of our work for CERN - was really the financial collapse of 2008 and 2009. One often hears that true correlations only appear in times of crisis, and we found that CTA returns were just as uncorrelated with stock and bond returns during those two years as they had been in the previous 15 or more years.

"CTA returns were just as uncorrelated with stock and bond returns during those two years (of crisis) as they had been in the previous 15 or more years."
Galen Burgdhardt

BW: And I think we have avoided a little bit because to start saying numbers like 30%, 40% of our portfolio did sound impracticable. Who is going to approve them? ButEwan Kirk at Cantab had a phrase: "You should know what optimal is and compare your performance to it." That is a good point, you should know what this practical constraint is costing.
GB: Brian's right. Ewan's remark is really telling. The institutional investment field is very inertial. In the US, it would be professionally reckless for a pension fund manager to allocate 30% or 40% to CTAs. Their evaluation horizons are too short, and their return requirements are simply too high to be achieved. At the same time, if they want to stick with a more conservative path, they should keep track of where the other portfolio - the one with a heavier allocation to CTAs - would have taken them.

OFI: What do you consider to be holding managed futures back from more widespread adoption?

GB: Some people argue that it's the name, or the names. Commodity Trading Advisors sounds as if the industry is trading pork bellies, wheat, and crude oil. And managed futures, which is what we use in the title of our book, sounds awfully narrow and possibly exotic.

But I don't think that's the name. Futures have been around for a long time. They provide access to all the world's asset markets and commodity markets. And there are a slew of very, very bright and experienced people who are working with some very major funds who understand what's behind the names. Even a small allocation from a subset of these funds could double the size of the CTA industry. It must be something else.

"I think the major roadblock for managed futures is that it is a difficult strategy to explain. Let us be sympathetic to the investor."
Brian Walls

BW: I think the major roadblock for managed futures is that it is a difficult strategy to explain. Let us be sympathetic to the investor. It is not easy to understand the general risk premium that these strategies are capturing and why they should persist. You have very accessible arguments for things like equity, private equity, venture capital and emerging market equity and debts, very accessible, simple sounding things about macroeconomics and China that seem really quick to digest and really kind of hit the right notes. It may be poor selection criteria, but it resonates with a lot of people.

Also, there is still a misunderstanding because the way the returns are constructed which is bolting a P&L on to a denominator, I think there is an uncomfortability around the risk measurement and construction of the portfolios. It just sounds riskier because of the funding arrangements. I think those are two big obstacles.

I think the wind is at our back in terms of there is a general comfort level in the liquidity and transparency and pricing mechanism of the futures market, especially coming through 2008.

It was not just that performance was great in 2008 but also I knew the value of my portfolio was worth at every minute of the day and I could get out of it whenever I wanted. I think that worked well. I think that the fact that no managed futures funds got side pocketed or gated during the 2008 crisis period has kind of reminded people about the liquidity of the strategy. I think that there is a lot of sophisticated people who are understanding the value of low to zero correlation and how it improves their total portfolio.

"The fact that no managed futures funds got side pocketed or gated during the 2008 crisis period has kind of reminded people about the liquidity of the strategy."
Brian Walls

We do have things working our way. A lot of these options reach strike price worth exercising if the industry prints a big number; has a lucky run like up 5 to 10% over 12 months and then suddenly we will get a lot of converts.

GB: It's a matter of horizon and memory. We're human. It is almost impossible to work in the here and now and remember, at the same time, that we are aiming for a prize that will be won 10 or 20 or more years from now. Just last month I was getting ready for something in New York, and the person who invited me to give the talk (and who should have known better) asked me why I thought CTAs had been struggling recently. My immediate reaction was to pull up 20 years or so of CTA returns and calculate some trailing returns. What I found was that the past year or two look exactly like four or five earlier episodes. I also found that the trailing 5-year returns don't look that bad.

One of the things I hate about a lot of CTAs' marketing material is that they are selling CTAs as if they were negatively correlated with the stock market - as a hedge against the stock market tanking. This is a total pain because the real argument rests on zero correlation. The problem with zero correlation, though, is that it's just zero and isn't sexy.

" tthe real argument rests on zero correlation. The problem with zero correlation, though, is that it's just zero and isn't sexy."
Galen Burghardt

Roy Niederhoffer gives a talk that reminds of something that Woody Hays - the old head football coach at Ohio State - once said about the forward pass. The way Woody saw it, only three things could happen with a forward pass and two of them were bad. As Roy argues, it sucks to be the uncorrelated asset in the portfolio. Imagine the four things that can happen. The portfolio's up and you're up. The portfolio's down and you're down. In either case, who needs you? Now then the portfolio's up and you're down. Now you're the goat. And finally the completed pass comes into play. The portfolio's down and you're up. Woohoo!

And what are the chances that you hit that one time in four at the same time the investment committee is reviewing your performance?

BW: It is about the total portfolio striving for a state of zero correlation. Managed futures does not hedge a specific risk.

GB: Or hedge anything. When investors look at marketing materials often they find arguments that somehow there is negative correlation or you have got tail hedges or things like that. What if, in fact, the the more solid argument in favor of managed futures is zero correlation.

BW: It gets a little dry but we tend to look at things in terms how does adding an asset lower the average pairwise correlation of the portfolio. The average pairwise correlation or the average correlation of the portfolio is very persistent. Returns are very random, but the average pairwise correlation to a portfolio was very persistent.

Anything you can do by adding an asset whether that is a manager, strategy, or market, anything you can do that lowers the average pairwise correlation of the portfolio makes that portfolio more robust and more persistent from one period to the next.

OFI: Interesting. Now, when you present that information, which to me seems so logical to a professional investor when they react negatively or they do not understand it where do they miss the point?

BW: I have been reading a lot of Daniel Khanemen's book lately. It seems like a behavioral bias. If you start by asking the question of what is the return today or what is the return year to date or you ask the question of how does that single asset relate to my portfolio? It is very hard to get them off of it. It is very hard to reframe it into what we need to look at. We have a whole research note that Galen often call Superstars vs. Teamwork.

OFI: I would say the "Teamwork" report is probably one of your better known research notes.

BW: It is and it really is the kind of the idea that anytime you try to have a strategy that is looking at recent returns and trying affect your portfolio, you will underperform a portfolio that is driven by trying to manage the average correlation of the portfolio.

Part two of this interview will appear in the next Opalesque Futures Intelligence.