Forecasting value-at-risk for cryptocurrencies
Journal
International Review of Finance
Date Issued
September 2025
Author(s)
DOI
10.1111/irfi.70029
Abstract
Value-at-Risk (VaR), the primary measure of downside risk in market risk management, relies heavily on the accuracy of volatility forecasts produced by risk models. This paper shows that, for forecasting the VaR of cryptocurrencies, the time-heterogeneous Student's t autoregressive model outperforms standard models commonly used by practitioners.