Return to search

Empirical tests of the predictive ability of asset pricing models and of stock market overreaction in the U.K

This thesis considers two major issues in the context of empirical research into the U K stock market: (i) what is the ability of five models (the naive market return, the market model, the CAPM, the APT, and the LAPT) for predicting the U K stock market price behaviour and (ii) does the U K stock market have long-term overreaction Chapter 2 reviews the literature of some financial topics In Chapter 3, an asset pricing model called the Leveraged Asset Pricing Theory (LAPT) is developed Unifying the Arbitrage Pricing Theory and the Modigliani and Miller Theory of capital structure, the model allows the changes in the underlying leverage variable of each company at time t-l to have immediate impact on its beta estimated at time t. The predictive experimental procedures are designed, in Chapter 4, to examine the ability of the LAPT and other conventional models with different beta estimates to predict U K equity returns Through the estimation procedures, the Trade-to-Trade and the Discount Weighted Estimation methods, based on Bayesian Forecasting, are used to avoid the problems of the nontrading effect and variation in parameters, respectively. The results, in Chapter 5, showed that when the year 1987 is added to the test, the predictive ability of both the APT and the LAPT becomes higher and the LAPT, which makes explicit the leverage factor in its structure, does even better job than the APT in market valuations around that period as more common factors are extracted for the LAPT Based on the controversial work of De Bondt and Thaler (1985), Chapter 6 examines the long-term overreaction behaviour of the U K stock market for the period 1965 to 1993. After relating the findings of this empirical study to the predictive ability of those benchmarks (the naive model, the market model, the CAPM, and the size-adjusted CAPM) used, the results indicate that the apparent evidence for overreaction depends upon the benchmark employed The better the benchmark in terms of high predictive power and low statistical measurement error the less we are unable to reject the null hypothesis of no overreaction We find that we are unable to reject the hypothesis of U K stock market efficiency with respect to the Contrarian Investment Strategy of De Bondt and Thaler.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:307991
Date January 1996
CreatorsJang, Woan-yuh
PublisherUniversity of Warwick
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://wrap.warwick.ac.uk/109063/

Page generated in 0.0015 seconds