Topological Properties of Network of Global Currency Exchange Rates

Abstract

We study time series of exchange rates of 37 global currencies expressed in terms of US dollars. We take US dollar is as the base currency because it is one of dominant currencies, which is used almost in 66 countries as their currency. The average volatility is computed from returns using overlapped rolling window technique. To study the effects of crisis on the structure and dynamics, we consider three sub periods; before crisis, during crisis and after crisis. Different statistical properties and network properties in three sub periods. From analysis of currency network at different thresholds, we find change in the structure of network in the period of crisis. We find the highly correlated and weakly correlated currencies in the calm and crisis period using threshold networks, which can help the investors in portfolio management. The group of most correlated currencies in the crisis period is different from that in before and after crisis period. Different centrality measures can differentiate the currencies according to their geographical location.