Abstract:
This paper presents an analysis of the basis target zone model, using data from three core EMS members. Recent research har argued that the model may not perform empirically as bad as previous investigation have suggested. In order to investigate this point, we first estimate the model by a nonparametric density matching procedure, and extract the implied drift and diffusion function for the exchange rate process. This allows us to view the target zone model essentially as a time series model, with a clearly defined diffusion process governing target zone exchange rates. We then compare the implied drift and diffusion function with their nonparametric estimates. We conclude that exchange rates under consideration are actually more mean reverting than the target zone model predicts. When very close to the intervention limits, the data are far from the target zone model's implications. We perform a Monte Carlo study, which documents that our methodology is indeed able to capture the potential non-linearities in the drift and diffusion predicted by the theory.