Realized volatility of stock returns results from a smooth stochastic process and discontinuous jumps. The jump component is often measured by subtracting bipower variation of high frequency data from realized variation. However, measurement of the bipower variation depends on trading volume. Generally, when the stock return follows a Wiener process, so does the trading volume. It turns out the relative trading volume affects the size of return fractality, and thus, stock returns appear mostly characterized by jump elements. To the extent that the stock's return fractality is measured by its fractal dimension, the paper also offers an alternative approach to computing fractal dimension.
展开▼