The empirical evidence that stocks with high aggregate shorting demand subsequently underperform is a puzzle given that the information is public. In this paper, we first confirm the return predictability of monthly short interest is not driven by new private information. Second, we show that highly shorted stocks tend to be very liquid and part of the return anomaly is attributable to inadequate liquidity adjustment. Third, the monthly significant negative returns (i.e., -0.7 to -0.6% per month) associated with high short interest are concentrated in stocks with high loan fees (i.e., high shorting costs). However, high shorting demand and high shorting cost do not necessarily imply withheld negative information, because (i) there is a significant presence of institutional investors in the highly shorted stocks and (ii) there is no evidence of informed selling by these investors following high short interest. By considering the loan fees as an implicit dividend income, the abnormal returns become economically and statistically insignificant among the highly shorted stocks with high lending fees. Thus, the significant negative return associated with high public short interest reflects return mis-measurement rather than an anomaly due to short-sale constraints.