Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.820513
Title: A financial analysis of new econometric techniques
Author: Calonaci, Fabio
ISNI:       0000 0004 9355 6066
Awarding Body: Queen Mary University of London
Current Institution: Queen Mary, University of London
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
The predictability of asset returns is one of the most controversial topic in nancial literature other than critical issue for portfolio managers. Over the last decades, the topic has been investigated from two di erent points of view: one more empirical with new models and one more theoretical with new estimation techniques. Understanding the complexity of the topic, we decided to contribute to both of these approaches. On one side we propose a new model of nancial volatility, while on the other we develop a new econometric methodology for better capturing the risk of portfolios, which are both key elements in the analysis of returns. The rst part of the Thesis suggests a compressive approach for capturing the volatility of assets returns. It extends the Heterogeneous Autocorrelation model, Corsi (2009), HAR, decomposing the volatility into its principal features: slow decay of the autocorrelation function, asymmetric behaviour with returns and all the facets of volatility jumps. The empirical forecasting exercise shows remarkable improvements in the precision of the forecasts. The second part of the Thesis introduces a new hierarchical methodology for estimating dynamic pricing models and increasing their performances. The method, based on the Fama and MacBeth (1973) approach and developed in the classical kernel regression framework, employs a exible algorithm for the selection of the bandwidth. Wide empirical evidence is provided in support of the methodology and its ability to produce a more accurate description of the systematic risk. The last part of the Thesis sheds light on the relationship among di erent long memory models. The chapter investigates the role of long memory parameters in addition to the classical HAR approach nding that its importance varies across the assets. The HAR restricted ARFIMA model seems to be a good approximation of the dynamic structure of several Realized Volatility series.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.820513  DOI: Not available
Share: