Skip to Content

Alex Proimos

Alex Proimos Macquarie Research Student
  • Title: Mr
  • Position: PhD Student - Department of Accounting and Corporate Governance

Contact Details

Student information

  • Load: PhD Student Part Time
  • Associate supervisor: Professor Neil Lawrence Fargher
  • Date of submission: 29/10/2015
  • 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.
  • Purpose: An empirical study of whether distress risk exposure is reflected in the cross section of equity return premiums.
  • Originality: It is common to study and measure risk premiums, such as that associated with the risk of financial distress, in terms of the differential between mean returns of the extreme portfolios (i.e. high and low distress risk). Not only are such approaches sensitive to model specification, distributional assumptions and the measurement error, they do not address directly the question of whether risk exposures and measured risk premiums are monotonically related - as would be expected if distress is viewed as a systematic source of risk. The current study applies the recent methodology of Patton and Timmerman (2010) as a direct test of the monotonicity hypothesis.
  • Findings: Using a formal test of monotonicity this paper finds little evidence of a systematic relationship between expected return and distress risk premiums. Statistically significant distress-related equity premiums appear observable at the extremes of exposure only.
  • Practical and social implications: This finding suggests that prior empirical results (see: Dichev, 1998; Griffin and Lemmon, 2002; Vassalou and Xing, 2004; and Campbell, Hilscher and Szilagyi, 2008) that find either positive or negative systematic distress risk premiums have been driven by measurement error and outliers within the lowest and highest distress risk (or extreme) portfolios.
  • Keywords: monotonicity, distress, default, risk, premium