Forecasting Exchange Rate Methodology
Economists around the world do not possess reliable and convenient methods of forecasting fluctuations and changes in exchange rates that happen over a short period of time such as days or weeks. However, the expectations about short-term changes in the exchange rates are made depending on the market sentiments. But, several times these expectations tend to change according to the changing economic prospects of the country.
According to the PACIFIC exchange rate service, there are two sets of trends for forecasting exchange rates. According to these, the exchange rates are forecasted based on certain “technical” trading techniques based on 30-day and 60-day intervals. These methods are not dependent on the complex computational methods and economic fundamentals but are intended to reveal market trends. One has to always keep in mind that these trends are not actual forecasts. They only study the market trends and predict accordingly.
The 60-day interval method is based on combined quadratic trend model and an autoregressive model called stepwise autoregressive method. In this method, there are two main components called the trend component and autoregressive component. The trend component studies the long-term changes and the autoregressive component is based on the short-term behaviors of the market. Both these components are sequentially computed.
The 30-day interval method is basically called as the heuristic trending technique based on “runs” of exchange rate movements within a tunnel. This tunnel should be long and as narrow as possible. There are two different softwares available that are used for computing these trends. The 60-day interval method is computed using PROC TREND and the 30-day interval method is computed using PROC IML software. Both these software have been designed using SAS language.