THE RISK PREDICTION MEASUREMENT OF TAIL-VALUE-AT-RISK (TVAR) FOR RISK AGGREGATE

• Main Author: MAWADDAH ABDAL (20815002), AINUN
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/20873
• Institution: Institut Teknologi Bandung

The risk aggregate is a widely used risk model for modeling financial losses. The risk aggregate used in this thesis involves two components of random variable having interdependent assumptions. Next, the measure of the risk are required to know the amount of the risk. Value-at-Risk (VaR) and Tail-VaR (TVaR) are the most common measure of the risk used for quantification of risk values. In this thesis, risk prediction measure is done by using risk ag- gregate models. Copula is used to facilitate in obtaining the joint distribution function of two random variables. VaR and TVaR can become tools to predict the value of loss by determining the quantity of the data distribution. Data <br /> <br /> <br /> <br /> simulations were performed using the generated data from the Exponential bivariate distribution and the Gamma-Exponential mixture distribution. Simulation results show that VaR and TVaR predictions are in uenced by the level of trust. If the chosen confidence level is higher then the predicted result of VaR and TVaR is also greater. It also found that TVaR prediction results are always greater than the results of VaR predictions. Since the accuracy test of VaR and TVaR prediction results are shown at the correct calculation value of both. The results show the percentage of the accuracy test that is almost equal to the level of the given confidence level.