Abstract
We employ linear and nonlinear ARDL models using monthly data from January 1985 to April 2025 to examine whether domestic and foreign economic policy uncertainty (EPU) exert asymmetric effects on exchange rates. The analysis focuses on five major economies relative to the United States: Australia, Canada, Japan, the United Kingdom, and the Euro Area. The baseline linear ARDL model, excluding EPU variables, fails to establish a stable long-run relationship between exchange rates and economic fundamentals. In contrast, an extended linear ARDL model with domestic and foreign EPU identifies stable long-run relationships across all countries. The nonlinear ARDL (NARDL) model confirms long-run relationships and reveals significant short- and long-run asymmetries in EPU shock effects. Our findings demonstrate the importance of including domestic and foreign EPUs in exchange rate models, and the benefits of using nonlinear models to capture asymmetries between exchange rates and economic fundamentals. The Global Financial Crisis (GFC) reinforced the long-term relevance of EPU in exchange rate determination, while the COVID-19 pandemic introduced heightened short-run volatility and modest long-run structural adjustments, particularly in countries more vulnerable to external shocks.