A recent announcement by SSARIS Advisors, LLC., an affiliate of SSgA, about eliminating commodities from their managed futures funds, reported in the Attain Capital Managed Futures blog, was partly justified in the context of the high standard deviation of daily returns most commodity futures contracts have in comparison to stock index, financials and currency futures contracts. In this article I first investigate the validity of this claim by calculating the standard deviation of daily returns of 19 popular futures contracts covering financials, currencies, metals, energy and stock indices. I then argue that the link between low standard deviation of daily returns and managed futures funds performance is elusive and rests on a rather unfounded assumption about the structure of the markets. Since most managed futures funds employ trend following strategies, I argue with the use of examples that limiting trading to markets with low standard deviation of daily returns may not be a sufficient condition for satisfactory performance because standard deviation is a lagging indicator.
The standard deviation for a period of N days was calculated using logarithmic daily returns. Charts show the 60-day rolling standard deviation of logarithmic daily returns in percent. Some charts also show 60-day rolling correlation between various futures contracts based on the difference between two consecutive closing prices.
Standard Deviation of Daily Returns of Futures Contracts
Table 1 shows the 60-day and 250-day standard deviation of daily logarithmic returns of 19 futures contracts as of the close of October 19, 2012:
60-day stdev %
250-day stdev %
Stock indices, bonds
Energy and Metals
Table 1. 60 and 250-day standard deviation of daily returns of 19 futures contracts as of the close of October 19, 2012
It is evident from Table 1 that stock index (SPY), bond (TLT) and currency futures (FXE) have the lowest standard deviation of logarithmic returns as compared to the other futures contracts included. Bonds show the smallest values but gold futures show values comparable to those of the Nasdaq-100 futures contract (QQQ). I just make a note of this here and the explanation I can offer is the ongoing process of financialization of commodities, such as gold, through the availability of popular ETFs like GLD (GLD). However, the same does not hold for silver and SLV (SLV), its popular ETF. It is surprising however that DBC (DBC), the commodity ETF, has 60 and 250-day standard deviations of log daily returns in percent of 0.87 and 1.02, respectively. Apparently, fund managers can now trade these ETF products that exhibit much lower variability of daily returns than their underlines.
The conclusion from just looking at the numbers in Table 1 is that the premise that the standard deviation of daily returns of stock index, financials and bond futures contracts is lower than that of the other commodity contracts (except in the case of gold futures as noted) is true. But we know that true premises do not always guarantee sound conclusions unless the rules of logic that are used are valid and any other premises involved are also true.
Standard Deviation of daily returns is a lagging indicator
The standard deviation is calculated based on past prices and as a result it is a lagging indicator. Lagging indicators are not entirely worthless but their use can result in missing trading opportunities and in other undesirable effects to be discussed below, especially when used in the context of asset allocation. For example, the S&P 500/10-Year Note correlation in the beginning of 2011 was close to 0% as shown on the lower pane of the chart below (vertical blue line mark). But then the correlation decreased steadily and by the end of the year it was at -0.77, i.e. the two contracts became anti-correlated. Stocks ended almost flat for the year but any gains of managed futures funds from the strong bond uptrend may have been wiped out because of the anti-correlation with stocks.
click to enlarge images
The Barclay Systematic Trader Index for 2011 fell -3.83%, as shown below, in a year during which bond prices gained close to 30%:
Source: BarclayHedge, Ltd
Despite the low standard deviation exhibited by stock index and bond futures contracts, the combination proved not to be very beneficial to most futures managed funds in 2011. Bond futures gains were mostly wiped out by stock index futures losses. Thus, trading futures contracts with low standard deviation of daily returns is neither a necessary, nor a sufficient condition for profitability. As a matter of fact, it can hurt profitability, as in the case just discussed.
Timing Is (Almost) Everything in Managed Futures
The lower pane of the chart of corn futures shows that at a time when the Corn-S&P 500 correlation was near 0 (blue vertical mark), meaning that these two assets were perfect candidates for a diversified portfolio, the standard deviation of corn daily returns was at 2.16%:
By excluding corn from a futures fund, one would miss a substantial rally in prices that started in June of this year. Some could argue that the choice of instruments to trade depends on the risk profile of a manager and his objectives. But the manager also faces any consequences from missing trends and generating negative absolute returns or returns below those of a benchmark index. Things like standard deviation and correlation are important but there is nothing better in managed futures than catching a good trend. This is where the skill lies in generating alpha for investors. As a matter of fact, I strongly believe that investors in managed futures want their managers to take calculated risks, not trying to find ways to avoid them. Since such investors are for the most part sophisticated, if they want low risk they can find other vehicles, including short duration bond funds.
Another example that shows the type of opportunities that funds that exclude commodities may be missing is the spectacular multi-year downtrend in natural gas futures prices shown on the chart below:
Here we are also looking at a market with low correlation with S&P 500 index futures (shown on the bottom pane) but, nevertheless, with a high standard deviation of daily returns.
The Elusive Connection
The link between standard deviation of daily returns and managed futures fund performance is elusive because at the highest level it relies on an unfounded assumption about market structure and operation. Markets do not exist above and beyond their own participants, which is a common misconception. When fund managers use indicators as part of their strategy, then the values of these indicators are in turn affected by the strategy that employs them in adverse ways, often invalidating it. Specifically, when participants flock to certain markets, volatility increases as everyone tries to anticipate the action of the other participants and their timing. In my opinion, the move by certain big funds away from commodities is more of a marketing campaign directed towards risk averse investors who are not familiar with the benefits of diversification. Those funds should not give up on trying to devise better ways for market timing and instead turn to low risk/correlated markets because as a result of their actions, the risk for trading these markets may increase and so the variability of their performance. In my opinion then the answer to recent dismal managed futures performance is not to lower risk by brute force market selection because risk indicators are both lagging and non-stationary but a move away from traditional trading models and development of novel approaches with higher win rates that are compatible with the new realities of the markets.
Additional disclosure: Charting program: Amibroker. Futures data source: Pi Trading.