The Application of Gumbel Copula to Estimate Value at Risk With Backtesting in Telecommunication Stock

Abstract

The Value at Risk (VaR) method refers to a statistical risk measurement tool used to determine the maximum loss of an investment, while the distribution that must be met is the normal distribution. This is not in line with the actual situation, because the distribution of the return value is found to be not normally distributed but depends on market conditions that occurred at that time, thus invalidating the VaR estimate and resulting in greater portfolio risk. Therefore, in this study, the estimation of risk value will be carried out using the Gumbel Copula method which can model the dependency structure between stocks and is flexible enough to model financial return data from https://finance.yahoo.com/. The parameter estimates produced by the Gumbel Copula method are then used to calculate the VaR at 90%, and 99% confidence levels. The resulting VaR values ​​are 0,076 and 0.231. To test the feasibility of the VaR model, backtesting was carried out and concluded that the VaR value obtained was valid and suitable for use in the risk assessment of PT. XL Axiata Tbk and PT. Telkomunikasi Indonesia Tbk.

Loss Model of Climate Insurance Based on Effect of Growing Degree Days Index

“…The calculation method related to determining the risk of crop failure based on the growing degree days (GDD) index used is a combination of the time series method and the deterministic element of trend and seasonal which depends on time by estimating parameters using least squares [12], [13]. This study modeled the average daily temperature associated in determining the amount of loss calculated using the combination but determining its time dependence with the copula [5], [14], [15]. This is done because there are two different locations which are assumed to have independent dependence on each other.…”