Analysis of local and global instability in foreign exchange rates using short-term information entropy

Abstract A new method for estimating the short-term information entropy for analyzing foreign exchange rates is presented. The short-term information entropy measure presented in this paper was constructed by reducing variations in foreign exchange rates to binary symbolic dynamics. This method was used to automatically extract periods of local instability from a single time series of exchange rates. This method was also used to automatically extract periods of global instability from multiple time series of exchange rates. Short-term mutual information was used to quantify the temporal correlation between foreign exchange rates. The short-term information entropy and short-term mutual information were both found to be high over an extended period after the 2008 financial crisis.