This study tests the effect of turnover shocks on the asymmetric autoregressive behavior of index returns. The methodological approach adopted in this study is based on the relationship between market return and trading volume. This study first uses a vector autoregression (VAR) to model two market trading volume series by controlling for the variation associated with the sign and magnitude of both week t and week t − 1 or week t + 1 market returns. The generated market-adjusted relative turnover (MRTO) series are then plugged into the bivariate asymmetric AR GARCH-t model (asAR-GARCH-t) using weekly data from the Shanghai and Shenzhen A-share stock markets. The results show that turnover shocks have a material effect on asymmetric autoregressive behavior and thus on asymmetric persistence of past returns.