Very Noisy Option Prices and Inferences Regarding Option Returns Jefferson Duarte, Christopher S. Jones, and Junbo L. Wang * October 2019 Abstract We show that microstructure biases in the estimation of expected option returns and risk premia are large, in some cases over 50 basis points per day. We propose a new method that corrects for these biases. We then apply our method to real data and produce three main findings. First, the expected returns of straddles and delta- hedged options written on the S&P 500 Index are smaller than previously estimated in the literature. Second, delta-hedged options and straddles written on individual stocks have negative expected returns. Third, the price of individual equity volatility risk is about 45% of the price of market volatility. These findings show that the stylized finding that volatility is not priced in individual stock options is due to microstructure biases. Keywords : Options; Liquidity; Microstructure bias * We thank and Greg Eaton, Ruslan Goyenko and seminar participants at the 2019 Latin America FMA Conference, 2019 CICF. Duarte is with the Jesse H. Jones School of Business at Rice University (jeffer- son.duarte@rice.edu). Jones is with the Marshall School of Business at University of Southern California (christopher.jones@marshall.usc.edu). Wang is with the Ourso College of Business at Louisiana State Uni- versity (junbowang@lsu.edu).