Rush Hours: The Real Impact of Trading Activity on Volatility. Evidences from the Dow Jones Industrial 30.

Abstract

This paper examines the influence that trading activity has over volatility at intraday level. The correlation between the waiting times between consecutive trades and the amplitude of price movements has been analyzed, confirming the results found by Engle (2000), and Doufur and Engle (2000), according to which volatility is positively correlated with the frequency of the trades, and so, negatively correlated with the waiting times. However, this study demonstrates that these findings are strongly influenced by the presence of deterministic patterns in both waiting times and volatility. In fact, both time series are highly non-stationary and autocorrelated. In order to analyze an unbiased correlation between the two quantities, a detrending procedure with different orders of the polynomial has been used. The results show that, with an appropriate detrending order, the correlation between trading activity and volatility is reduced during the central part of the day, while completely disappearing at the beginning and at the end of the trading session. The explanation may lie in the activity of informed vs uninformed traders. As Admati and Pfleiderer (1988) pointed out, in the central part of the day the liquidity traders are staying away leaving a high proportion of informed traders, which may translate in a direct information impact from the trades to the price formation process. On the other side, the higher activity of noise traders in the first and in the last 20 minutes of the trading session may destroy the relationship between transactions’ frequency and price movements.