Independent components in investor attention to energy market

Abstract

In this article, we investigate the correlation structure of the time series of investor attention as measured by relative search query volume of stocks in Google. Specifically, we explore - i) Whether the time series has a power law correlated dependence (long range memory) and how does it evolve over time? ii) How does this dependence vary with frequencies of sampled data? iii) Does a cross-correlation dependence exist between local and global investor attention? iv) What happens to this memory structure in case of volatility clustering periods of price and volume? We perform detrended fluctuation analysis and detrended cross-correlation analysis of the time series of investor attention of top 20 energy companies (by their market capitalization). The results confirm the existence of long range dependence in investor attention. The memory dynamics are characterized by persistent and mean-reverting behavior. There is a reasonably high positive cross-correlation dependence between local and global investor attention. Finally, we observe that volatility clustering has little effect on long range dependence structure of investor attention.