The IKOS FX Fund is purely systematic and has had positive returns even during the past three years of major turbulence in the financial markets as Table 1 and Fig.1 show. The fund is up more than 17% year-to-date. It has a low correlation to other CTAs in part because we are not trend-followers.
There is still a lot of mystery surrounding what purely systematic funds do and this mystique has been further fostered by the managers themselves. My goal is to demystify this. The essence of systematic trading is to find some structure, some pattern, some pocket of predictability in noisy time-series. This structure is discovered by statistical, econometric or other techniques or observing the markets. Once a structure has been found, the goal is to teach the machine to recognise the structure, put the trade on and manage the risk. Easier said than done, but that is the goal. What makes the problem so hard is that as a first approximation markets do behave like noise. But a closer examination shows relationships between time-series, auto-correlations and more, with big doses of non-stationary behavior and the occasional extremely large moves that can be the undoing of the entire fund. Managing this tail risk of extreme moves is crucial not just to the performance but the survival of the fund as the last few years have shown.The IKOS FX Fund consists of three broad families of models, traded in three different holding periods – high frequency (few minutes to hours), daily frequency and long term strategies – with overall average holding periods of 10 days. The three families of models are as follows:
Systematic Global Macro Models: Are the short rates driving the movements in currencies? Or is it the other way around? Are stock indices a barometer of the economy and the currency or does a strong currency boost the stock markets as is often the talk in Japan? Which way does causality flow? The answer depends on the situation. The beauty of systematic trading is that we don’t start with preconceived notions coloured by our judgment. Instead we let the data speak for itself. If a relationship exists, we exploit this relationship and put on our trades. If none exists, we don’t trade.
Carry Models: When interest rates in Japan are close to zero and those in Australia are north of 5% it is tempting to borrow from Japan and invest it in Australia and pocket the interest-rate differential. This is the essence of the carry trade. The trouble is that everybody knows this. So during moments of market panic such trades come to a violent end wiping out the entire year of P&L. The cleverness comes in being able to detect risk-aversion in the market place prior to the crash and exiting the trade while keeping the gains.
Trend-following Models: These models were added in 2008 and form a small part of our suite of strategies.
Risk Management: Our long-term carry and trend-following models are susceptible to giving up gains during periods of risk-aversion in the market place. Knowing this, we have created a risk-management indicator that measures risk-aversion in the market place and in an algorithmic manner, depending on the level of the indicator, reduces the capital allocated to these models and increases the allocation to the high-frequency models that tend to do well during these turbulent times. Fig.2 charts the values of the risk indicator. Note, the values close to zero correspond to times of crises.
There are also times when many of our models lose small amounts every day. In situations like this, when the peak-to-trough drawdown reaches a certain threshold, the entire portfolio is gradually deleveraged in a formulaic way and when the fund starts recovering the losses the leverage is brought back to normal.
With the hundreds of models competing for capital, we have an optimiser that, besides netting trades between models, ensures concentration limits on individual currencies are respected and the models that forecast a higher return on a trade get more capital than a model that forecasts less.
In addition to these risk-management tools, emphasis is placed on execution algorithms to minimise the market impact of our trading, trying as far as possible to be liquidity providers.
Why IKOS?
IKOS has been successfully managing FX strategies for a very long time. There is a growing research group dedicated to improving our strategies, developing new ones as well as improving the execution algorithms. Many in this group have PhDs in the hard sciences. The firm and its funds are regulated by the US SEC, US CFTC/NFA, CySEC and other global regulatory bodies. IKOS has a strong culture of corporate governance and its funds are highly liquid, which also means it is well-suited for the launch of FX strategies in a UCITS wrapper.
Dr. Ravi Chari joined IKOS in January of this year. He is a member of the portfolio management team at IKOS, which is comprised of four portfolio managers, and he is a member of the IKOS Investment Committee, which looks after global portfolios of equities, futures and FX traded at IKOS. Ravi has a PhD in Probability Theory from Michigan State University.
