Author: Brandon Eller
Degree: M.S. in Financial Engineering
Advisory Committee: Dr. Dragos Bozdog, Dr. Ionut Florescu
Abstract: Exchange rates are used by almost every business. How they move and what affect them are questions that have been asked for a long time. In this thesis, we take a look at where the BEER model came from and how it can be used to estimate equilibrium exchange rates by looking at past macroeconomic data such as net foreign assets, producer price index, consumer price index, inflation, output gaps and more. Using this model to observe how different variables affect exchange rates can bring more insight to traders, businesses, and others on what else they can look at when observing exchange rates. The main focus of this paper is to observe how the BEER model calculates the equilibrium exchange rates and the misalignment between the actual and the calculated exchange rates By using the BEER model to estimate three exchange rates for USD/JPY, JPY/CHF, CHF/USD and optimizing certain model parameters we can see how it is used when looking at different currency pairs. The misalignments are calculated from the difference between the actual and the calculated rates. The thesis also briefly looks at what happens to one dollar when it is exchange across all three-currency pairs. In an ideal world this process would return a value of one, not including exchange rate fees. From this process it was observed that the rates do not exchange exactly back into one dollar when using the BEER model. This opens up the possibility to future study to further optimize the model parameters based on the final results of the multiplication of the three estimated exchange rates to obtain more accurate results. By optimizing the results based on the exchange on the dollar across the three pairs the error calculated between the actual and calculated exchange rates should be reduced. Another potential candidate for further study would be to optimize the parameters more frequently and observe how it would affect the overall fit of the model, which in turn should reduce the error when performing the exchange rates.