@article{LIU2026, 
author = {Long LIU and Fangling REN},
title = {VaR calculation method based on normal inverse Gaussian distribution},
year = {2026},
journal = {Journal of Capital Normal University (Natural Science Edition)},
volume = {47},
number = {3},
pages = {74-79},
keywords = {normal inverse Gaussian distribution, peak thick tail, VaR, Kupiec failure frequency check},
url = {https://www.sciopen.com/article/10.19789/j.1004-9398.2026.03.006},
doi = {10.19789/j.1004-9398.2026.03.006},
abstract = {The normal inverse Gaussian distribution can accurately describe the distribution of financial asset returns. Firstly, introduce the normal inverse Gaussian distribution and its parameter estimation methods. Secondly, provide two methods for calculating VaR values under the normal inverse Gaussian distribution. Then, the empirical analysis of the Shanghai Stock Exchange internet financial index is carried out, and the VaR value under the normal inverse Gaussian distribution is calculated and compared with the VaR value under the normal distribution. Finally, conduct a Kupiac failure frequency test to determine the reasonableness of the calculation. Empirical analysis shows that the normal inverse Gaussian distribution can reflect the peak and fat tail characteristics of the distribution of financial asset returns, and the calculation of VaR values at high confidence levels is more accurate than the normal distribution. Using the normal inverse Gaussian distribution for financial data analysis can effectively avoid underestimating financial risks.}
}