VaR Estimation and Backtesting Using R

Abstract

In modern risk management, risk identification is the premise of risk management, risk control is the means of risk management, and risk measurement is the foundation and core of the whole risk management system. The goal of this thesis is to validate different VaR models by retrospectively testing selected time series. However, in order to provide a detailed description of the backtesting process in the empirical section, we first discuss the general estimation methods of value at risk and the theory of backtesting methods. The thesis consists of five chapters. The first part is the introduction, which explains the purpose and structure of the thesis. Chapter two describes the basic ideas behind VaR and gives some background and history on the topic. Chapter three focuses on the backtesting procedures. The fourth chapter describes the empirical research of the thesis, which can be considered as the core of the thesis. Chapter 5 summarizes and reviews the most important results of the theoretical and empirical parts.