Less Volatile Value-at-Risk Estimation Under a Semi-parametric Approach

In this study, we propose a two-step, less-volatile value-at-risk (LVaR) estimation using a generalized nearly isotonic regression (GNIR) model. In the proposed approach, a VaR sequence is first produced under the generalized autoregressive conditional heteroskedasticity (GARCH) framework. Then, the VaR sequence is adjusted by GNIR, and the generated estimate is denoted as LVaR. The results of an empirical investigation show that LVaR outperformed other VaR estimates under the classic equally weighted and exponentially weighted moving-average frameworks. Furthermore, we show not only that LVaR is less volatile, but also that it performed reasonably well in various backtests.