Use this URL to cite or link to this record in EThOS:
Title: Three essays in the financial economics of conditional volatility
Author: Liu, Jingyi
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 2009
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This thesis concentrates on investigating three important questions on conditional volatility in financial markets: if volatility is forecastable, which method will provide the best forecasts? What economic behaviour is the reason behind conditional volatility, if any? What optimal statistical evaluation criterion conditional forecasts does the economic utility maximisation correspond to? The new ideas, viewpoints, methodologies and theoretical underpinning employed in this thesis endow the study on conditional volatility in financial markets with a deep and comprehensive understanding in the need for better controlling and modelling asymmetric and clustering volatility. Chapter 2 investigates the out-of-sample predictive ability of 73 competing time series models for the volatility of foreign exchange changes. The empirical results support the stylised facts of volatility. Historical volatility models are superior to ARCH class models. However, ARCH class models take predominance where over-predictions are more heavily penalised. The various model ranks are shown to be sensitive to the error statistics used to assess the accuracy of the forecasts. There is no single forecasting model suitable for all purposes. Chapter 3 presents evidence that a habit persistence version of the theoretical Lucas two-country model is capable of generating predictable conditional volatility in spot foreign exchange returns. We compare the “ARCH” properties of the ‘theoretical’ model to those of the empirical estimations. Our results show that the theoretical model fits the empirical behaviour of volatility. Chapter 4 proposes an optimal forecast error criterion for utility maximisation under an option trading rule. The empirical results show that for a more highly averse investor the optimal forecast error criterion is a weighted average of MAE and MSE but which weights MSE less heavily. The optimality forecast error criterion based on functions of forecast errors for utility maximisation under asymmetric loss provides a simple rule for making economic and financial decisions under uncertainty.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available