(CONDITIONAL) VALUE-AT-RISK PREDICTION ON RETURN HETEROSCEDASTIC MODEL

• Main Author: Xavier Setiawan, Timotius
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84155
• Institution: Institut Teknologi Bandung

Risk in financial sector could be predicted using risk measure. Two risk measures that are commonly used, especially in finance, are Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR). In practice, Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR) are used for predicting return that has special properties such as fat-tailed and heteroscedasticity property. These properties, however, could be used for modelling return volatility. Two most commonly used for volatility modelling are Autoregressive Conditional Heteroscedasticity (ARCH) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH). These model will be used for VaR and CVaR future prediction. Distribution assumptions for predictions are normal distribution and tstudent distribution. Evaluation on these predictions shows that t-student distribution is more appropriate to predict extreme value as it has a heavy tail property.