Lotka–Volterra signals in ASEAN currency exchange rates

The 1997 Asian financial crisis stimulated interest in whether there are regional currency linkages that can transmit exchange rate fluctuations and pave the way to a single regional currency. Most such investigations to date are statistical analyses such as calculating cross-correlations rather than applications of nonlinear statistical models such as the Lotka–Volterra model. However, statistical significance is not the same as meaningful significance, and exchange rate time series appear to be noisy. How can we determine whether we are looking at temporally correlated signals–or at temporally correlated noise? We utilized the Lotka–Volterra model to extract and reconstruct signal from noisy long-term time series of daily exchange rates of ASEAN-5+3 currencies. The Lotka–Volterra model revealed hitherto undetected relationships with meaningful significance between certain currency pairs and provided four interpretable parameters for each pair-wise interaction. Statically significant fits were obtained for eleven of the sixty-four pair combinations. There is no evidence for trading blocs.