[The empirical evidence on the cross-sectional relation between idiosyncratic risk and expected stock returns is mixed. We demonstrate that the omission of the previous month's stock returns can lead to a negatively biased estimate of the relation. The magnitude of the omitted variable bias depends on the approach to estimating the conditional idiosyncratic volatility. Although a negative relation exists when the estimate is based on daily returns, it disappears after return reversals are controlled for. Return reversals can explain both the negative relation between value-weighted portfolio returns and idiosyncratic volatility and the insignificant relation between equal-weighted portfolio returns and idiosyncratic volatility. In contrast, there is a significantly positive relation between the conditional idiosyncratic volatility estimated from monthly data and expected returns. This relation remains robust after controlling for return reversals.]

We present empirical evidence that short sales contribute to market efficiency by increasing the speed of price adjustment to not only private/public firm-specific information but also market-wide information. Shortable stocks are characterized by weaker trade continuity and stronger quote reversals. They adjust faster to new information than non-shortable counterparts. These findings remain robust even in an “up” market condition in which short sales are not binding. The amount of information incorporated in each trade is also significantly higher for shortable than non-shortable stocks in both “up” and “down” market conditions. After controlling for firm size, trading volume, liquidity, price and option trading, short sales stand out as one of the significant factors that speed up the price adjustment.

The empirical evidence on the cross-sectional relation between idiosyncratic risk and expected stock returns is mixed. We demonstrate that the omission of the previous month's stock returns can lead to a negatively biased estimate of the relation. The magnitude of the omitted variable bias depends on the approach to estimating the conditional idiosyncratic volatility. Although a negative relation exists when the estimate is based on daily returns, it disappears after return reversals are controlled for. Return reversals can explain both the negative relation between value-weighted portfolio returns and idiosyncratic volatility and the insignificant relation between equal-weighted portfolio returns and idiosyncratic volatility. In contrast, there is a significantly positive relation between the conditional idiosyncratic volatility estimated from monthly data and expected returns. This relation remains robust after controlling for return reversals.