Using principal component analysis, we identify a common component driving a panel of 15 monthly bilateral exchange rates against the US dollar. We find this common (first principal) component is related to the fundamentals commonly used in exchange rate determination models, such as US nominal and real macroeconomic variables, financial market variables and commodity prices. We obtain the relevant set of fundamentals using the Lasso (least absolute shrinkage and selection operator) technique and find that this set changes over time.
- Exchange rate models
- Principal component analysis