Evaluating asymmetric effect in skewness and kurtosis

• Autores: Sonia Benito Muela
• Localización: Notas técnicas: [continuación de Documentos de Trabajo FUNCAS], ISSN-e 1988-8767, Nº. 744, 2014
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • This paper shows evidence that the higher moments parameters (skewness and kurtosis) of the distribution of financial returns are time-varying. This means that the distribution of financial returns is not i.i.d. as many approaches for portfolio risk management assume. Therefore it may be preferable to assume that the stochastic process for returns has time-varying conditional distributions. The possible dependence of these parameters on the sign and the size of unexpected returns have also been analyzed. The Engel and Ng. (1993) test was used to do so. The tests carried out provide some evidence that the skewness and kurtosis parameters respond asymmetrically to shocks of different signs and sizes. This result suggests that if we are interested in modeling the dynamic behavior of these parameters we should take asymmetric GARCH specifications into account.

    Besides, it has been detected that in volatile periods, when the size of unexpected returns is bigger, the tails of empirical distributions grow fatter, and the distribution is not as peaky. If the empirical distribution is skewed to the left skewness rises, otherwise it goes down.