In this paper the correlation dimension is used to detect determinism in an economic time series generated by speculative behavior of market participants during a currency crisis. The time series are from the emerging market of Lebanon, which experienced episodes of severe currency fluctuations in the years 1985-1992. The detection of determinism and nonlinearities in the stationary time series is done through the estimation of the correlation dimension. The time series is shown to have low and finite dimensionality. Furthermore, a significance test for the correlation dimension is applied using Fourier-transformed surrogate data. The results show strong evidence of a nonlinear structure, which lends support to the hypothesis that the currency crisis in Lebanon was driven by deterministic speculative activity.
展开▼