Dynamic GAS Based Modeling for Predicting and Assessing the Value at Risk of Tehran Stock Exchange Index and Gold

This research has been written with the aim of modeling a new criterion for measuring risk in order to eliminate the shortcomings of traditional models in the field of investment risk management.

In the present study, with a practical purpose, to estimate the value at risk of daily bitcoin price data (2,707 views) in the years 2013 to 2020 and the data of the total stock exchange index (2,753 views) 2011 to 2020 has been used in two groups of education and test (500 views). In order to estimate the value at risk using the nonlinear method and the generalized variable self-fitting time (GAS) method, modeling was performed by learning from the data of the training group and the accuracy of the model was determined by the data of the experimental group.

The results showed that for the total stock index, only two models, GAS and GARCH, are suitable risk estimators. On the other hand, for Bitcoin cryptocurrencies, only two models, GAS and GARCH, are suitable risk estimators, which GARCH model is more preferable.

Originality / Value: 

Findings showed that the new GAS model is a preferential estimator for the total stock market index than other nonlinear models. This is due to the variable time feature as well as the dynamics of the GAS model, which is able to respond to market turbulence conditions unlike traditional models in the short run. These results also help investors and active financial institutions to manage risk in their trading systems.