An Application of Eigenvector Scaling Method to Currency Exchange Rate Data in Short-Term Forecasting

Article excerpt

INTRODUCTION

Exchange rate fluctuations and interest rate disparities represent one of the most challenging problems affecting consumers, businesses, and the world economy in general. They can create inefficiencies and distort world prices. Moreover, the long term profitability of any investment, export opportunities and price competitiveness imports are all impacted by long-term movements in exchange rates. Therefore, international businesses engaging in foreign exchange transactions on a daily basis could benefit from knowing some short-term foreign exchange movements.

Exchange rate economics however remains one of the least understood and most controversial areas of economics. It is often argued that it is futile to attempt to forecast foreign exchange rates because the foreign exchange market is an efficient market (Goodman, 1979) and therefore historical exchange rate data hold no information in predicting future values in the market. Most economists believe that there is no reliable method available to forecast exchange rates (Paul Krugman and Maurice Obstfeld, 1994). Nevertheless, the business community has to form expectations about the short-term trends of exchange rates, their underlying tendencies and patterns in order to reach good and sound business decisions.

LITERATURE REVIEW, EVIDENCE OF FOREIGN EXCHANGE MARKET EFFICIENCY, AND OBJECTIVE

The recurring question to researchers has been to agree whether or not the foreign exchange market is an efficient market. Early studies tested randomness of exchange rates (Poole, 1967) and found the theory of random walk to be inconsistent with the interest rate parity theory. Although, it remains true that time series for the major nominal exchange rates over recent float are extremely hard to distinguish from random walks (Michael Musa, 1984). Other empirical studies of exchange market efficiency have examined the profitability of trading or filter rules. Drawing from the concept of speculative runs, a currency that has risen significantly is likely to rise and a currency that has fallen significantly will continue to fall. Logue and Sweeney (1978) tested the profitability of filter rules. A simple k-percent filter rule consists in buying a currency after it has risen k-percent from its previous low and sell it after it has fallen k-percent from its previous peak. Results showed the rule yielding marginal profits as compared with a buy and hold strategy. A number of studies do indicate profitability of filter rules (Dooley and Jeffrey Shaffer, 1983, Levich and Lee Thomas, 1993). If rules that are consistently profitable can be found over a reasonable period of time, then the market is not efficient.

Current theories explaining the relationship between the values of two currencies can be revisited and each time pin down disagreements and limitations surrounding them as tests have been more or less unconvincing through out the literature. Economists do not possess reliable methods of forecasting exchange rates over short time horizons such as days or weeks. The numerous methods available for forecasting can be categorized into four general groups: (1) technical, (2) fundamental, (3) market-based, and (4) mixed. Technical analysis involves use of the historical data to predict future values. I t is based on systematic patterns in the exchange rate. If this pattern over time appears random, then the technical forecasting is not appropriate (Madura 1995). Fundamental forecasting, a model based, is one of the numerous methods available for forecasting exchange rates. It is based on relationships between economic variables and exchange rates. Current values of economic factors along with their historical impact on the currency can lead to an exchange rate projection that is useful in deciding whether an international business should hedge or not future payables and receivables in foreign currency. …