The findings of this research indicates that profits from Foreign Exchange trades can be generated from utilising one of the simplest form of technical analysis, single and twoperiod moving averages, ARIMA or Moving Average with Support and Resistance lines. The finding of the research confirms that technical analysis is a useful tool to create a trading system that is profitable even after accounting for funding cost.

It is especially true in trending markets (upward or downward) where moving average method proves to be one of the best simple methods available. The best Moving Average period in terms of highest ’Total Profit’ varied from one time series data to another. When volatility is higher or trend is more apparent, then the longer period moving average is desirable. The use of two moving averages seems to outperform that of single moving average, due to its ability to eliminate unnecessary trades and capture the existing trends. The introduction of a filter, which is used to eliminate less profitable trading signals, of 0.5% outperforms other techniques where no filters or higher percentage use of filters are utilised. The reasons for the higher percentage i.e. 1.0% or 1.5% used tend to eliminate many profitable trades.

This is particularly true in the foreign exchange market where the movement of the rates can be relatively small and the volatility is higher. The model that produces the highest Total Profit of $0.4417 is ARIMA with filter of 0.5% being used. The next best model that produces second highest Total Profit of $0.4352 is the Two Moving Averages with filter of 0.5%. Support and Resistance used in the model seem to contribute profitability to the single moving average model even though the overall profitability does not outperform the previous two moving averages and ARIMA models.

Support and Resistance lines are not very accurate predictors of when and where prices will reverse, but rather these lines are tools that can be used to alert traders to areas that need a closer examination. The optimisation model used for the above test recognises some limitation, which does not guarantee success. The main problem with optimisation is the need to constantly reoptimise every so often. Changing market conditions may cause these optimised numbers to change over time. Another limitation is the fact that these methods have not been used in other instruments such as securities and commodities.

Future Research

The trading system created in the research suggested continuous buying and selling every time certain conditions are satisfied and the system does not facilitate any trader to hold the foreign currency for more than one period unit at any given time. Another limitation is that system allows buying and selling only one unit of foreign currency. In real life, profit can be maximised by holding the foreign currencies for more than one period unit (which can be daily, weekly or monthly) and ideally a system should encourage more units to buy or sell when the signals appear to be very strong. In other words, this research does not look at any money management technique.

Money management technique or strategy that is worth considering is ’Optimal f’, introduced by Ralph Vince [1990]. Optimal f is a money management strategy that can be used to improve and maximise system performance by finding the best percentage of capital to invest in each trade. The ’optimal f’ can also be extended to optimise the timing of the investment. Future research will be conducted to investigate how long traders should hold their foreign currency exposure and how much he or she should buy or sell every time a signal is generated. Another possibility of future works will be to use the same or similar models and optimisation techniques using other historical time series data or other financial instruments, such as securities and commodities.

Prof. Clarence N W Tan and Herlina Dihardjo

Next: References

Summary: Index