semiparametric value-at-risk estimation of portfolios

This MATLAB-code reproduces the results of Xu, Jiahua (2019). Semiparametric Value-At-Risk Estimation of Portfolios. A replication study of Dias (Journal of Banking & Finance, 2014). International Journal for Re-Views in Empirical Economics, Vol3(2019-6).

Abstract: This paper aims to replicate the semiparametric Value-At-Risk model by Dias (2014) and to test its legitimacy. The study confirms the superiority of semiparametric estimation over classical methods such as mixture normal and Student-$t$ approximations in estimating tail distribution of portfolios, which can be credited to the model’s uniqueness in combining strengths of both extreme value theory (EVT) models and other multivariate models. The author however discovers, in one instance, the infeasibility of the Dias model, and suggests a modification.

Data and Resources

Suggested Citation

Xu, Jiahua (2019): Semiparametric Value-At-Risk Estimation of Portfolios. Version: 1. International Journal for Re-Views in Empirical Economics. Dataset. http://dx.doi.org/10.15456/iree.2018304.055009