Essays on the relation between stock price movements and orders

Abstract

This thesis explores different aspects of the relation between orders submitted (or the trades they generate) and price changes over time. The second chapter provides descriptive evidence of the order flow in the French stock market. I define a rigorous measure of order flow imbalance based on limit order data. This imbalance is positively autocorrelated as similar orders tend to be clustered. I then analyze the price impact of individual orders on the price, which is a concave function of each order's volume. I then consider aggregated order flows, which are also highly correlated with stock returns, with an R2 around 50% for the average stock. This price impact of orders does not appear to be reversed later, at least in the next few months. I also present results on the multivariate structure of the order flow, where a market component is found, which is highly correlated with the market component for the return. Additionally, the order flow from other stocks does not affect stock i's return beyond the effect of stock i's order flow, although their return does affect stock i's return. The third chapter provides additional evidence to help interpret the observed return order flow correlation. I first argue that causality goes from orders to prices. I then distinguish between two causal interpretations: private information and uninformed price pressure by looking at the implications of a private information model. For idiosyncratic returns, where one would expect private information to be important, and the R2 to be high, the R2 is indeed around 41%. However, for the common market return, where one would expect private information to be minor, the R2 is even higher at 70%. The high R2 on the market suggests that private information does not fully explain the co-movement of orders and prices.

(cont.) This points toward a bigger role for uninformed price pressure than is usually assumed. The fourth chapter addresses the implications of concavity for the volume volatility relationship. Whereas Jones, Kaul, and Lipson (1994) find that trade size has no effect on volatility and that only the number of trades is important, we establish that trade size is important, but not in a linear form. One has to take a concave function of each trade size to maximize the relation with volatility. When this is done, the number of trades becomes irrelevant.