Long Memory in Indian Equity Market : De-trended Fluctuation Analysis

The presence of diverse investor categories of varying sophistication and the role of wealth funds along with other characteristics has left the Indian equity market susceptible to inefficiencies in the price discovery process. This study examined the long memory possibilities in the Indian market across various indices representing market coverage, market capitalization, and liquidity based portfolios, both in the pre and post financial crisis period of 2008. The null hypothesis that prices follow a pure random process was tested against the long-range dependence. The results based on de-trended fluctuation analysis technique indicated relatively weaker long range dependence in all the indices and higher level of persistence in their volatility. The study also observed varying degrees of persistence across the market cap and liquidity based portfolio indices and the persistence was found to be higher in the post-crisis period. Broadly, the nature of the dependence was found to be different both in price indices and in their volatility. Across market cap portfolios, the temporal dependence was found to be negatively associated, that is, large cap and medium cap security price indices movements evinced lower long range dependence compared to that with small capitalization portfolio price indices. Contrary to this, long range dependency in their volatility was found to be increasing with the market size. Linkage between level of persistence and market liquidity price indices was found to be weak. NSE indices evinced more efficiency than BSE indices. In short, price formation in Indian market evinced long memory, and this information was decisive, especially in pricing of index linked derivatives and other funds traded in the Indian market. It also indicated the relevance of non-linear asset pricing models in the Indian equity market.