Adjusted Hurst exponent evaluations for equity and energy markets

This study introduced an alternative method in long memory volatility financial time series evaluation using the autoregressive conditional heteroscedasticity models. Instead of direct measurement for long memory Hurst exponent parameter, this approach suggested a short memory filtering procedures using the conditional heteroscedastic specification. The stationary time series which is free from short memory is later evaluated using two heuristic long memory estimations. It is found that this method is able to eliminate the possible spurious long memory in the selected equity and energy markets. The findings of this study are important in long memory estimation and market efficiency analysis.