Is There a Season for Happier Returns?
Our author is Tim Merryman, who draws on research by BarclayHedge and Attain Capital to suggest that over the years commodity trading advisors' monthly returns show a persistent tendency to perform better later in the year. Besides presenting the evidence, he explores possible reasons for this intriguing seasonal pattern.
As we go to press, there is tentative evidence that managed futures overall is doing better than it did earlier this year. Credit Suisse/Tremont estimates, based on 54% of assets in funds reporting to the database, that CTAs returned 0.79% in August. This is the second positive month for CTAs so far in 2009 after a 0.85% gain in May. But June and July brought losses.
Credit Suisse/Tremont says trend followers may to be beginning to show profits as models gain more traction. Year to date, manages futures is down 7.1% while global macro is up 6.2%.
Mr. Merryman is managing director at Interconti Ltd. and Opalesque Futures Intelligence editorial advisor.
Year-to-date performance by commodity trading advisors and systems traders has been disappointing. The reasons for this are not entirely clear. Managed futures tends to benefit from market trends, but in that case CTAs should have done better because equities and, by inference, the stock indices that now make up a large part of the financial futures sector that many CTAs trade, have been trending for several months.
There is another factor that may be germane. Some industry observers have long suspected the existence of a seasonal pattern to CTA performance, with a tendency to do better late in the year. Is this a fact or merely a hope that performance will improve after all in 2009?
I discussed the question with Sol Waksman of BarclayHedge, who has been thinking about the subject of CTA "seasonality" for some time now. He noticed a few isolated tendencies over the last few years but had not yet researched these in any detail. So we decided to delve into the possibility to see what we could find.
Sol's data analysis runs from 1998 through 2008 and includes 158 of the CTAs in the BarclayHedge database with a minimum 10-year performance record. Over this period, March and July are the only two months showing consistent drawdowns (see Fig 1 in the attached pdf). July is by far the worst month of the year for CTAs, with July losses averaging 60-to-65 basis points.
Walt Gallwas and his team at Attain Capital provided some answers. One, long-term trends tend to emerge more clearly in the latter part of the year as investors flock to what's working and dump what's not-causing trends to extend. This is backed up slightly by the S&P Diversified Trends Indicator, which on average shows an annualized gain of 10.3% in the second half versus an annualized gain of 8.4% in first half over the years.
It is possible that stock markets become less efficient in the second half of the year, as market volume goes down in the summer months under the long-standing "sell in May and go away" rule. That would result in a greater opportunity for systematic models.
Another plausible explanation is that volatility tends to pick up in the September-November period each year, possibly in anticipation of another October 1987 crash. There is some support for this argument in the average monthly VIX - the CBOE Volatility Index - going back to the 1990s (Figure 2 in the attached pdf). In the first half of the year, average monthly VIX reading is 18.39, whereas the average for the second half of the year is 20.82..
The VIX, a measure of sentiment on U.S. equities derived from S&P 500 index option prices, reflects investors' consensus view of stock market volatility over the next 30 days. Because investors buy protective puts when the market is in a steep decline to protect from further declines in their portfolio's value, a high VIX reading usually represents increased investor fear and occurs during times of market turmoil. A low VIX reading implies that investors are getting complacent about the need to buy protection against declines.
Note that this year volatility has smoothed substantially from the wild rodeo we saw in the second half of 2008. The analysts at Attain argue that July's closing VIX level, representing a 74% drop in the index since its October 2008 peak, is a somewhat significant number. The VIX declined that much only two other times in the past 20 years. Both times signaled a volatility spike in the not too distant future.
The first time there was a spike of 156% just five months later and volatility steadily increased for the next eight years. The second historical example of a steep decline in the VIX had a more complicated sequel. Although there was a 114% increase 18 months later, volatility basically stayed around low levels for two-and-a-half years. Then came the dramatic spike to the all time highs in October 2008.
On the basis of these examples, odds are there will be a spike but when it will happen is anyone's guess.
Large Changes in the VIX
On the whole, the data indicates that futures money managers and systems traders may benefit from longer-lived trending tendencies normally experienced in the third trimester and volatility tends to spike during the same period, creating more opportunity. But that's all in the rearview mirror. What will the future bring?
One mustn't lose sight of the reality that the current cyclical correction has been and will probably continue to be unique in modern history. That said, it is likely that historical tendencies have not been entirely negated even in these extreme conditions. It is interesting that volatility spiked on the 17th of August, as can be seen in Figure 4 (attached pdf). Another big move in volatility will favor managed futures, whether as a hedge or as an alpha-enhancer.