Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.637084
Title: Modelling and forecasting stock and stock market volatility
Author: Gower, C. P.
Awarding Body: University of Wales Swansea
Current Institution: Swansea University
Date of Award: 2000
Availability of Full Text:
Access through EThOS:
Abstract:
This thesis examines the volatility of forty-five FTSE 100 Companies and the index itself for the period 4/1/1988 to 7/9/1998. The generalised autoregressive conditional heteorscedasticity model and the threshold and component derivatives are used to model volatility. The stochastic volatility model performs on a par with the generalised autoregressive conditional heteroscedasticity type models. The data is extended to 22/2/1999 in order to carry out volatility forecasts. The forecasting ability of the model is assessed by comparing the volatility forecasts to those of the historical mean, random walk and exponential smoothing models. For the daily forecasts, the historical mean, random walk and exponential smoothing models. For the daily forecasts, the historical mean model strongly outperforms the other models on mean absolute error, median absolute percentage error and median squared error forecast evaluation criteria. The exponential smoothing model provides the best forecasting model for the root mean square error forecast error measure. For the four-weekly volatility forecasts, the generalised autoregressive conditional heteroscedasticity type models and stochastic volatility models perform relatively better. For forecasting purposes, a risk averse investor should choose the threshold generalised autoregressive conditional heteroscedasticity model, while a more risk neutral investor should use the exponential smoothing model. The analysis is also extended to the volatility characteristics of six major and four emerging market stock indexes. A risk averse investor should choose the threshold generalised autoregressive conditional heteroscedasticity model while a more risk neutral investor should choose the random walk model. Only in the first forecast period should a risk neutral investor choose the stochastic volatility model. For the four weekly forecasts, a risk averse investor should use the GARCH/GJR model. For an investor who is more risk neutral the choice is less clear cut, with the CGARCH/ACGARCH, stochastic volatility and historical mean models proving equally valid choices.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.637084  DOI: Not available
Share: