Modeling the variability of the art markets using the extreme value theory and calculating their risk measures

Abstract

This paper attempts to measure the riskiest markets in the art domain. The goal is achieved by using the Extreme Value Theory. The EVT model attempt to use the extreme events happened in each variable to calculate the extreme market risk. Extreme Value Theory provides a well organized statistical model to calculate the extreme risk measures like the Value at Risk and the Expected Shortfall. In my project, I demonstrated that the EVT was successfully applied to the art market log returns, for predicting in the next step the VaR and ES. The study was conducted for 10 variables related to art market from April 1998 to January 2015. The major findings of the study are 1) the VaR estimation for the right tail distribution using the historical simulation shows that 19th century, drawings and sculpture had the highest VaR values at 0.5%, 1% and 5% tail probabilities, respectively. For the let tail, Modern art was the riskiest among the other art log returns at 0.5% and 1% tail probabilities and drawings at 5%.2) The EVT was successfully applied to the art market log returns, for predicting in the next step the VaR and ES. The risk estimation using EVT shows that drawings, old masters and modern art enclosed the highest VaR values.