Why Practitioners Should Use Bayesian Statistics

Abstract

A 2017 paper by Harris argues that financial returns are a mixture distribution and that the dominant distribution is the truncated Cauchy distribution. Both the mixture nature and the poor properties of the truncated Cauchy distribution functionally exclude both the Frequentist and the Likelihoodist methodologies. In addition, restrictions known to exist in the parameters cause the available methods to be inadmissible, biased solutions. Only the Bayesian methodology results in statistics and predictions that are admissible and coherent.

Keywords: Bayesian methods, Frequentist methods, heavy tails, Cauchy distribution, Stable distribution

JEL Classification: G00, G28