TIME-VARYING HURST EXPONENT FOR THE BUCHAREST EXCHANGE MARKET

We investigate the behavior of the Bucharest Exchange Market (BEM) with the Hurst exponent, employing two alternative techniques, rescaled range analysis (R/S) and detrended fluctuation analysis (DFA). First we estimate the global Hurst exponent for BET (1998-2009) and RASDAQ-C (1999-2009). Because we notice a persistence behavior, we perform a moving windows analysis with R/S, to see the temporal variations of the scaling exponent for each time series. According to the results, the local time-dependent Hurst exponent HI c displays an erratic dynamic with some alternating episodes of low and high persistent behavior. The global exponent gives us an overview of the time series comportment, but the sliding-window analysis is essential for understanding the underlying process. In the end we discuss the statistical significance of our result.