A wavelet analysis of inter-dependence, contagion and long memory among global equity markets

Abstract

This study attempts to investigate into the structure and features of global equity markets from a time-frequency perspective. An analysis grounded on this framework allows one to capture information from a different dimension, as opposed to the traditional time domain analyses, where multiscale structures of financial markets are clearly extracted. In financial time series, multiscale features manifest themselves due to presence of multiple time horizons. The existence of multiple time horizons necessitates a careful investigation of each time horizon separately as market structures are not homogenous across different time horizons. The presence of multiple time horizons, with varying levels of complexity, requires one to investigate financial time series from a heterogeneous market perspective where market players are said to operate at different investment horizons. This thesis extends the application of time-frequency based wavelet techniques to: i) analyse the interdependence of global equity markets from a heterogeneous investor perspective with a special focus on the Indian stock market, ii) investigate the contagion effect, if any, of financial crises on Indian stock market, and iii) to study fractality and scaling properties of global equity markets and analyse the efficiency of Indian stock markets using wavelet based long memory methods.