Market participant behaviour and equity market dynamics

Jackson, Andrew (2003) Market participant behaviour and equity market dynamics. Doctoral thesis, University of London: London Business School.

Abstract

This thesis consists of three studies examining the interaction between market participant behaviour and equity market dynamics. Chapter 2 examines the relationship between sell-side analysts and investors in the presence of repeated interaction and trade generation incentives. I find that optimistic analysts generate more trade for their brokerage firms, as do high reputation analysts. Sell-side analysts face a conflict between telling the truth to build their reputation, and misleading investors to generate short-term increases in commission. When the investor is unsure of the analyst's type and when short sales constraints are present, I show using a simple model that equilibrium forecast optimism can exist, even when investment-banking affiliations are removed. Chapter 3 examines the behaviour of individual investors. Behavioural models generally assume that investment decisions of irrational investors aggregate in a systematic way. Using a unique dataset of individual investor trades I find aggregate individual investor trades exhibit strong systematic patterns, including negative feedback trading and substantial persistence. In addition the trades of a large number of independent retail brokerage firms are highly contemporaneously correlated. However I do not find that the net trades of retail investors consistently predict future returns in a negative fashion. While small investors do act in a highly systematic fashion, their actions may not, at least in the short run, be classed as irrational. Chapter 4 examines the behaviour of individual and institutional investors to assess whether there is evidence of noise trader risk in a broad equity market context. I find that increases in institutional participation are associated with future excess volatility and excess correlations between stocks. I also find evidence of a significant priced noise trader factor in returns. I conclude that there is evidence of noise trader risk, however this risk is not associated with individual investors, but instead with institutional investors. I argue that institutional frictions are a much more plausible source of noise trader risk than is individual investor sentiment.