Use this URL to cite or link to this record in EThOS:
Title: Long memory in volatility, modelling and forecasting in the wavelet domain
Author: Noutsos, Apostolos
ISNI:       0000 0004 2673 0677
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2009
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis examines the long memory property of financial time series specifically returns and volatility. In our analysis we use wavelets and their decorrelating property to model long memory processes. For returns, it has been observed that their long memory parameter is very close to 0. Based on improvements we have proposed for a wavelet estimation method, and on Monte Carlo tests for the power of several semi-parametric methods in distinguishing long memory, we re-examine the long memory property of returns of several financial assets. For volatility a Bayesian estimation method in the wavelet domain is proposed for long memory stochastic volatility models with shorter memory dynamics as well, in the form of autoregressive or moving average parameters. Although estimation of these types of models can be unstable, we apply them to data describing various financial assets where we find that the dynamics involved might be more complicated than expected, since both long and short memory parameters are found to be statistically significant. Finally, we compare on the basis of density forecasts, for the first time, the long memory stochastic volatility model with GARCH-type models for data describing a range of financial assets.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available