Estimation of VaR in conditional heteroscedastic models for principal‐protected notes

Abstract

Purpose

The aim of this paper is to examine the accuracy of GARCH and provide a comparison of GARCH‐type and the other time series models in financial commodity markets.

Design/methodology/approach

First, a model fitting is performed to choose suitable models with conditional volatility for principal‐protected and path‐dependent notes by means of Akaike information criterion (AIC) and Schwartz Bayesian information criterion (SBC). Second, this paper adopts the backtesting criteria and the Diebold and Mariano test to compare the performances of the selected time series models.

Findings

The empirical results show that the performance of GARCH is significantly worse than EGARCH(1,1) based on the Diebold and Mariano test criteria. By backtesting test criteria, the null hypothesis that a given confidence level is the true probability in ARCH(4) cannot be rejected. The interesting results are different from past studies.

Originality/value

There is little literature of principal‐protected notes that focuses on the downside risk for investors. But, managing downside risk is important for individual and institution investors. This paper offer new insight into the literature of principal‐protected notes.

Citation

Chen, F. (2008), "Estimation of VaR in conditional heteroscedastic models for principal‐protected notes", Journal of Risk Finance, Vol. 9 No. 5, pp. 492-501. https://doi.org/10.1108/15265940810916148