Impact of trades on return comovements and asymmetric volatility in dual listings

We empirically investigate the impact of the lag of volume on time-varying return comovements (continuation vs. reversal) and asymmetric volatility (monotonic inverse relationship with return) of 175 American depository receipts (ADRs) listed on the New York Stock Exchange (NYSE) market and their underlying securities from 27 developed and emerging markets in the world. We denote volume by the number of traded shares that reveals information, liquidity, and momentum (feedback) motives and thus accounts for the asymmetry in return comovements and asymmetric volatility, the inverse relationship between volatility and return. We classify trading days into positive, negative, or zero momentum days based on a joint distribution of volume and return and hence determine if and how such a classification may partially explain asymmetric return comovements and asymmetric volatility. Empirically, we use a vector auto regression (VAR) setup to forecast volumes of cross-listed securities after adjusting for their comovements and denote the residuals as unanticipated volumes, which we use in all subsequent analyses. We estimate return spillover, serial auto and cross correlation coefficients along with those corresponding to two interaction terms, lag of volume times return which is the total number of shares traded on any day and return is the difference between the log (price) on two consecutive trading days and volume times return= volume*return, for each ADR and its underlying home share and find those are perfectly symmetric between ADRs and their underlying home securities. Both ADRs and their underlying home shares show positive auto- and cross-correlation coefficients for volume times return indicating continuation. Our momentum classification shows continuation (reversal) for positive (negative/zero) momentum days and further shows how firm size and liquidity contribute to the asymmetry in return continuation or reversal for the sample firms and for subgroups, for example, developed vs. emerging markets securities. We also find that the effect of auto- and cross-volume times return on ADRs and their home shares volatility is asymmetric across momentum classes. Our evidence suggests that volume times return denotes momentum revealed in trades. Many of our findings either contradict existing results or are new to the literature on cross-listed securities.