Measures of financial market integration and their application to the NYSE

Recent advances in the finance literature have introduced several useful tools for exploring the structure of asset prices and any relative price misalignments. Specifically, Chen & Knez (1995) introduced unconditional measures of price integration and misalignment based on price kernel representations. The current document builds on their work. It explores the potential of such methods in investigating questions of economic interest and in augmenting current asset pricing methods. The first chapter of this document considers issues surrounding the interpretation and estimation of Chen & Knez's first measure. A value-based normalization is proposed which is argued to provide a more economically relevant measure of price misalignment. Methods and issues surrounding estimation are then considered, and Monte Carlo simulations are performed to study the sampling behaviour of these estimators. The second chapter extends these unconditional measures to dynamic measures of price integration in a discrete time framework. The dynamic processes are then used to investigate the magnitude and dynamics of violations of the Law of One Price in the NYSE. Specifically, the dynamic integration between four 2-digit level industry groups are examined at the daily horizon between 1962 and 2002. The aim is to provide a different perspective on the issue of market efficiency and the speed with which information propagates through prices. Results suggest that while misalignments are not as large and long-lasting as documented in some of the case study literature; there is an economically relevant degree of market inefficiency in the industries studied. The findings motivate further research in a number of directions.