Modeling of Financial Losses Due to Traffic Accidents with Copula Gaussian Model Utriweni Mukhaiyar(a*), M. Akhirul R. (b), Novita Lestari P. (b), Lailatul Maziyah W.M. (b), dan I Gede Maha Hendra P. (b)
a)Statistics Research Division, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Jl. Ganesa No.10 Bandung 40132, Indonesia.
b)Master Program in Actuarial Science, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Jl. Ganesa No.10 Bandung 40132, Indonesia.
Abstract
Abstract. Traffic accidents cause financial losses. The losses model will be used to estimate the future total losses. In this research, the modeling of the total loss due to traffic accidents in Indonesia in 2022 uses the distribution model without dependence and the Copula Gaussian model. The data used are total loss (Million Rupiah), severity amount of loss (Million Rupiah), and the number of accidents from 1992 to 2019 with monetary value adjustments using the concept of future value. The Goodness-of-Fit test shows that the Gamma, Weibull, and Geometric distribution can describe the total loss, severity loss, and the number of accidents, respectively. Furthermore, the correlation between the severity and frequency loss is highly negatively correlated. From the results of the Total Loss modeling, the calculation of VaR_(95%) and TVaR_(95%) for total loss without-dependence and Copula Gaussian model are respectively 369,015 and 341,817 billion Rupiah. The results of TVaR_(95%) are 425,462 and 386,498 billion Rupiah. The total loss in the without dependence distribution model shows greater estimation results in the calculation of VaR_(95%) and TVaR_(95%) under stochastic dominance theory, thus Copula Gaussian model is the least dangerous model at level 95%.
