This paper examines the influence of risk aversion on the pricing policies of a market maker for securities. It is shown that a market maker's bid-ask spread can be decomposed into a portion for the known limit orders, a risk-neutral adjustment for expected market orders, and a risk adjustment for market order and inventory value uncertainty. It is demonstrated that a risk-averse market maker may set a smaller spread than a risk-neutral specialist. Finally, this paper demonstrates the pervasive role of inventory in affecting both the placement and size of the spread.

David Easley, Maureen O'Hara, Adverse Selection and Large Trade Volume: The Implications for Market Efficiency, The Journal of Financial and Quantitative Analysis, Vol. 27, No. 2 (Jun., 1992), pp. 185-208

Extending an empirical technique developed in Easley, Kiefer, and O'Hara (1996), (1997a), we examine different hypotheses about stock splits. In line with the trading range hypothesis, we find that stock splits attract uninformed traders. However, we also find that informed trading increases, resulting in no appreciable change in the information content of trades. Therefore, we do not find evidence consistent with the hypothesis that stock splits reduce information asymmetries. The optimal tick size hypothesis predicts that stock splits attract limit order trading and this enhances the execution quality of trades. While we find an increase in the number of executed limit orders, their effect is overshadowed by the increase in the costs of executing market orders due to the larger percentage spreads. On balance, the uninformed investors' overall trading costs rise after stock splits.

We examine the potential profits of trading on a measure of private information (PIN) in a stock. A zero-investment portfolio that is size-neutral but long in high PIN stocks and short in low PIN stocks earns a significant abnormal return. The Fama-French, momentum, and liquidity factors do not explain this return. However, significant covariation in returns exists among high PIN stocks and among low PIN stocks, suggesting that PIN might proxy for an underlying factor. We create a PIN factor as the monthly return on the zero-investment portfolio above and show that it is successful in explaining returns to independent PIN-size portfolios. We also show that it is robust to inclusion of the Pástor-Stambaugh liquidity factor and the Amihud illiquidity factor. We argue that information remains an important determinant of asset returns even in the presence of these additional factors.

Katrina Ellis, Roni Michaely, Maureen O'Hara, The Accuracy of Trade Classification Rules: Evidence from Nasdaq, The Journal of Financial and Quantitative Analysis, Vol. 35, No. 4 (Dec., 2000), pp. 529-551

Recently, exchanges have been directly selling market data. We analyze how this practice affects price discovery, the cost of capital, return volatility, market liquidity, information production, and trader welfare. We show that selling price data increases the cost of capital and volatility, worsens market efficiency and liquidity, and discourages the production of fundamental information relative to a world in which all traders observe prices. Generally, allowing exchanges to sell price information benefits exchanges and harms liquidity traders. Overall, our results suggest that regulations on selling market data can play an important role in improving market quality and trader welfare.