Long memory in PIIGS economies- An application of wavelet analysis

Abstract

This paper studies long-memory properties of stock prices in PIIGS economies (Portugal, Ireland, Italy, Greece and Spain) and measures the evo lution of their long-memory phenomena over time. We use the Hurst exponent as a measure of long-range dependence in stock prices. We apply wavelet analysis (based on Haar, Daubechies-4, Daubechies12 and Daubechies-20 wavelets) for computation of Hurst expon~nts. In addition, we utilize the semiparametric Local Whittle approach to test the robustness of results obtained from wave let analysis. Our findings support the dynamic nature of efficiency, characteristics of stock prices in PIIGS economies. Portugal and Greece show long-range dependence properties, whereas Ireland and Italy are weakly inclined towards mean reversion and Spain shows martingale behavior in its stock price.