High-frequency trading - the last gold mine in FX?

Speaking on active FX hedging, James said the high-frequency trading model has not been well explored but in some ways is more predictable as it is only concerned with internal market dynamics. "Exogenous variables have very little effect upon them, apart from the occasional shock, so I think this is the most model-able," she said.

The more often the model trades, the more chance it has to deliver smooth returns, said James. However, this also makes the model more difficult to implement and increases the sensitivity to trading costs.

"It is less likely to have anything to do with underlying fundamentals and is purely an endogenous system, apart from external shocks," she added.

She said many models that work on daily data are lost when it comes to high-frequency trading, such as differential forwards and trend. In fact, partly because many data series are only daily, models that trade several times in a day are the least explored part of the market, she explained.

James also discussed the benefits of using combinations of models. "The whole idea is diversification. Hopefully, as one model enters an inimical environment and loses money, other models will be making money and overall the portfolio should not suffer," she said.

Additionally, given that models typically work on only a limited number of currencies, having a 'stable' of models available increases the range of 'covered' currencies. "Thus, selecting models that are relatively anti-correlated should go some way towards achieving this goal," said James.

She warned, however, that there are pitfalls to be aware of when switching between strategies. "A common 'switch' is performance, where you get into model two if model one has done badly for time T or amount M. It is better, if possible, to have a switch that comes into operation before the model loses too much money," said James.

She added that a complex switching strategy could end up being a "big fat data mine". "It is possible to have good switch indications from other markets, like interest rates," she said.