This paper addresses the question of whether transaction costs affect stock prices. This question, in the intersection of market microstructure and asset pricing, has no supportive causal evidence to this date, which may explain the omission of transaction costs in mainstream asset pricing theories. To answer this question, we benefit from the Tick-Size Pilot Program, a laboratory-like experiment conducted by the Securities and Exchange Commission. This large pilot changed the tick size for approximately 1,200 randomly chosen stocks. We show that there is a decrease in market capitalization of $6.3 billion for (treated) stocks affected by the larger tick size relative to a control group after the start of the pilot program. We then study the channels for the price effects. Despite the increase in spreads for the treated stocks relative to the control group that increased transaction costs, the increase in transaction costs accounts for a small percentage of the price decrease. We therefore turn to channels of price variation due to changes in expected returns: investor horizon channel, information risk channel, and liquidity risk channel. Amihud and Mendelson (1986) argue that stocks with higher transaction costs attract a clientele of investors with longer investor horizons that in equilibrium require higher expected rates of return. Using institutional investors’ holdings data, we find that the investment horizon of institutional investors increases for the treated stocks relative to the control group after the tick size increased, consistent with Amihud and Mendelson’s (1986) model. In Easley and O’Hara (2004) and O’Hara (2003), when prices are noisier and reflect less information, informed investors are able to trade with less price impact. Thus, ex-ante there is more adverse selection leading to increased information risk and higher expected returns. Using several measures of price efficiency, including measures of the high-frequency speed of market response to company-related news, we show that prices are noisier and take longer to incorporate information, which are consistent with an increase in information risk. Following Acharya and Pedersen (2005), we construct several firm betas that capture liquidity risk, including a beta describing how firm liquidity co-moves with aggregate liquidity. We find a statistically insignificant decrease in liquidity risk for all test stocks. Overall, the evidence suggests that trading frictions not only affect liquidity, but most importantly, firms’ cost of capital
