Cuantificarea riscului de piață prin utilizarea distribuțiilor non-gaussiene. O abordare din perspectiva Value at Risk
Quantification of market risk by using non-Gaussian distributions. An approach from the perspective of Value at Risk

Summary/Abstract: This paper aims to model market risk through non-Gaussian distributions, the statistical method being through Value-at-Risk. Estimates of VaR and CVaR were made at both one day and 10 days for 10-year US government yields, the effectiveness of the results being subsequently tested by a back-testing procedure over another period. The results mostly indicate an underestimation of the risk for one-day yields, while the fitting for 10-day yields seems to be much more accurate. The estimated values largely covered the risk in the out-of-sample period, but this could be attributed to the fact that the period in question was free of events generating major price fluctuations.

• Details
• Contents