From duration analysis to GARCH models – An approach to systematization of quantitative methods in risk measurement

Abstract

The development of scientific research has led to the very dynamic growth of methods in the area of financial risk management. This refers particularly to risk measures in which quantitative methods are applied. The paper provides a discussion on a systematization of different risk measures proposed in scientific literature and used in practice. There are four criteria proposed in the paper. The first is the concept of risk applied by distinguishing negative and neutral concept. The second criterion is the character of the risk variable, either discrete or continuous. The third criterion makes the distinction between high frequency, low severity events, corresponding to standard (normal) type of risk, and low frequency, high severity events, corresponding to extreme risk. Finally the fourth criterion distinguishes between the risk variable expressed in monetary values and risk variable expressed in time units. Using these criteria the most common groups of risk measures are discussed. The final part of the paper gives a synthetic discussion on model risk which is a risk resulting from the erratic model used in a real world. In the paper three main sources of model risk are presented and the methods to evaluate model risk are given.