We investigate the underlying risk exposures of ETFs compared with their indices using a Principal Component Analysis approach. Then, we test whether ETFs’ tracking errors can capture the risk exposure difference between ETFs and their underlying benchmarks. We document a significant positive relation between tracking error and differences in risk exposure between ETFs and their corresponding indices. Even modest increases in tracking error are associated with economically meaningful divergences in risk exposure between an ETF and its benchmark. These findings suggest that comparisons of tracking error across index ETFs when making investment decisions may be misleading for investors seeking benchmark-consistent risk exposure.
Alfnaisan et al. (Wed,) studied this question.