Value-at-risk models: a systematic review of the literature

Reem Shayya, Maria Teresa Sorrosal-Forradellas and Antonio Terceño

Abstract

This paper presents a systematic review of the literature (SRL) on value-at-risk (VaR). More specifically, we review the models that have been applied to estimate VaR, with the following two aims: to find the most used models in the literature and to verify whether their popularity has changed since the 2007–9 financial crisis. The SRL is based on Scopus for the period from 1996 to 2017. Our results show that (generalized) autoregressive conditional heteroscedasticity models and extreme value theory, together with Monte Carlo simulation, historical simulation and variance–covariance, were the most used models. Since the crisis, the autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive conditional heteroscedasticity (GARCH) models have clearly been the most popular, while no significant difference has been found in the percentage of articles on the other models. This study can be considered the first SRL on VaR models because (to the best of our knowledge) no previous work of a similar nature has been carried out on this topic. This study provides a rich background for researchers and professionals interested in the topic, contributing detailed information about the papers published, classifying them by, for example, the model used, author(s), citation count, journals and year published.