| Title |
A quantitative analysis of risk measures in finance: volatility, value at risk (var), and expected shortfall |
| Translation of Title |
Kiekybinė finansinės rizikos matų analizė: kintamumas, rizikuojamoji vertė (VaR) ir tikėtinas deficitas. |
| Authors |
Kabir, Md Humaun |
| Full Text |
|
| Pages |
34 |
| Keywords [eng] |
Financial Risk, Volatility, Value at Risk (VaR), Expected Shortfall (ES), Heavy-Tailed Distributions, GARCH(1,1) Model, Tail Risk, Volatility Clustering, Financial Time Series, Quantitative Risk Analysis |
| Abstract [eng] |
The study provides a quantitative description of significant financial risk measures, i.e. volatility, Value at risk (VaR) and Expected shortfall (ES) in one probabilistic model. The motivation of the study is that the traditional measures of dispersion used are constrained in explaining extreme market behaviour especially when the distributions of returns are believed to have heavy tails. The analysis commences with the formulation of a strict mathematical basis to the asset based financial risk where asset returns are modeled as the random variables in a probability space. Logarithmic returns are computed and utilized to estimate empirical properties, such as, mean, variance, skewness and kurtosis. The findings show that its normality assumptions are grossly violated, and are skewed and overly kurtosis, which substantiates the fact that the financial data have heavy tails. A GARCH(1,1) process models the volatility and conditional time-dependent variance and volatility clustering. Although it has been found to be useful in the process of measuring the dynamics of dispersion, volatility has been found to be inadequate in the endeavor to measure downside risk. VaR is then historically and parametrically tested. Whereas VaR gives a sensible amount of the possible losses, it is less on the heavy-tailed conditions and not receptive in the magnitude of the maximum losses. Proposed alternative is Tail-Sensitive and consistent: Expected Shortfall. Both parametric and historical ES measure are more extreme market behavior sensitive and give a more detailed description of tail risk. The results show that ES beats VaR in the occurrence and magnitude of extreme losses in heavy tailed environments. Altogether, the article underlines the significance of tail-sensitive risk indicators and shows the inefficiency of the traditional methods to the correct quantification of the financial risk under the real market conditions. |
| Dissertation Institution |
Vilniaus universitetas. |
| Type |
Master thesis |
| Language |
English |
| Publication date |
2026 |
