Higher Degree Research EXPO 2009
Alex Proimos
 |
Formal Name: Alexander Proimos Personal Title: Mr Organisational Unit: Department of Accounting and Finance Load: PhD Student Part Time Supervisors:
Submission Date: 29th October 2015 Email: alex.proimos@mq.edu.au |
Is Default Risk Priced? Empirical Evidence in the Presence of Fat Tails
Abstract
Whilst financial market turbulence and bestselling popular books (Black Swan and Fooled by Randomness by Nassim Taleb) have served to underscore the importance of accounting for “fat-tails” in pricing and managing risk, most popular models, and empirical tests of models, begin with the assumption that returns are well described by a normal distribution. Normality of returns is often assumed for purposes of mathematical or computational convenience, despite the weight of published empirical evidence suggesting its inadequacy.
Recent work by Kan and Zhou (2006) suggests that the important empirical characteristics of monthly equity returns are well accommodated by an appropriately specified multivariate t distribution, and that modelling returns using a multivariate t in place of multivariate normal can have a substantial impact on the associated economic inferences.
Using econometric advances in the estimation of parameters under the multivariate t (pioneered by Dempster, Laird and Rubin (1977), Liu and Rubin (1995)), the current thesis addresses several empirical controversies using the approach advocated by Kan and Zhou (2006). The focus of the presentation will be on the first of the thesis papers: Is Default Risk Priced? Empirical Evidence in the Presence of Fat Tails.