Investigating of using fuzzy credibility theory for measuring value at risk

Value at risk is a new risk measure for measuring risk in financial institute. In this paper, a practical model base on fuzzy credibility theory for measuring value at risk is introduced. For this purpose, return of assets are considered in shape of triangular fuzzy numbers and value at risk is estimated by credibility distribution for triangular fuzzy variables. Then, value at risk for an investment company with different approaches is calculated for analyzing the results of this modeling approach and obtaining the appropriate time window. For this purpose, value at risk is estimated by three methods which are using fuzzy credibility theory with the time windows of 6 months and 4 months for parameters estimation and the traditional method of simple variance-covariance and the results are compared by Bernoulli unconditional coverage test. The result shows that value at risk by fuzzy credibility theory which use the four-month window of time for estimating of parameters provide more accurate specifying. Because return and risk of assets are uncertain variables in real world, using this model can transfer calculations to the uncertain environment and researchers are leaded to correct conclusions. Moreover, introduced model reduce volume of computations dramatically and value at risk is estimated easier than conventional methods such as methods base on GARCH family.