Clarke's Program Design
Clarke Capital Management, incorporated in 1993, is one of the longest-lasting
businesses in managed futures. The firm has a number of programs with
distinctive performance characteristics, each consisting of multiple models.
Some of the programs are among the industry's perennial top performers-see Top
Ten in this issue.
Founder Michael Clarke is well-know in the industry. Before starting Clarke
Capital, he spent several years in options arbitrage. He developed his software
tools and futures trading strategy from 1989 through 1992.
Differences between the Clarke programs can be pronounced. The program with the
most assets, Worldwide, made more than 36% in 2010 (through November). In 2009
it lost 11%. Another Clarke program, Millennium, lost a lot in 2009-48%.
However, Millennium too survived and came back in 2010.
Below, Mr. Clarke talks about how different programs are designed and what he
expects in markets.
"We designed enough risk parameters so that when conditions are
difficult the models still have a chance to bounce backâ€¦"
Opalesque Futures Intelligence: What is your strategy?
Michael Clarke: Trend following is what all our different programs do, but they
contain multiple models. The models vary in what they look for and the time
frame they operate in. For instance, our two Global programs are relatively
short term, focusing on a 15-day horizon. That is the shortest we go.
OFI: Do the programs all have the same number of models?
MC: No. Some programs are less diversified than others-Global Basic has only
four models so it is more affected by a few great or bad trades.
OFI: Which markets do you trade?
MC: The Global programs trade only about 20 products, which are the most liquid
contracts. By contrast, Millennium trades 40 to 50 products. The original
programs did not include equity index futures but the newer ones we introduced,
like Jupiter, include equity indexes.
OFI: How can the same models apply to markets as different as equity indexes and
MC: We test new models against a huge database that goes back to the 1940s and
contains a great number of commodities. The model has to test extremely well
against all those commodities. The testing gives us very robust models that do
well with a variety of products. So the trend-following models we use are
generalist and apply to all the products that a program trades.
OFI: Why do programs like Worldwide and Millennium differ so much?
MC: We designed Worldwide to be conservative whereas Millennium is designed to
be aggressive. But also, it is normal to get variation across models. Just
randomly you would get variation.
OFI: What do you mean by aggressive versus conservative?
MC: Aggressive programs trade more models and commodities. They have a higher
ratio of margin to equity because they trade more. Millennium trades a number of
interest rate products and also holds positions longer. Because of these
characteristics Millennium has more drawdowns in very choppy markets. Clients
can pick the programs that fit their needs.
OFI: What happened in 2009?
MC: It was a period of rapid change of opinions. Coming off the global crisis,
confidence was lacking and individual views of the future were widely uncertain.
People were indecisive about how bad things were and changed their minds easily.
There were many divergences of intermediate duration, whether in bonds or
currencies. There would be a sell-off for a couple of weeks, then a sharp turn
around. We need relatively sustained trends; whipsaw markets are very difficult
for us. So almost all our programs suffered.
OFI: How did 2010 turn out?
MC: There was still some uncertainty in 2010 but it was a more normal
environment in which most of our models made money. For instance, interest rates
fell for six to eight weeks. That was a good trend for us.
OFI: How does the future look?
MC: 2009 was the worst year we're going to see in a long time. A crisis that bad
probably comes every 50 years and its aftermath was really turbulent. We've
survived it and from here on we should do well.
OFI: What accounts for your ability to stay in this business 17 years?
MC: It is going to be 18 years in September 2011. Our models are superior to the
average CTA's. So in a way, it is simple! You just need to develop good models.
We designed enough risk parameters so that when conditions are difficult the
models still have a chance to bounce back, as Millennium did after the '09
OFI: What would you advise a young person interested in becoming a CTA?
MC: It requires a combination of skills and discipline. To develop their own CTA,
they need to have a strong computer background as well as trading skills.
Developing models is very difficult but it can be learned. It's tough work, but