Estimation Risk in Value at Risk Backtesting

Summary

The document explains estimation risk in the context of Value at Risk backtesting. VaR is calculated from a model and inputs that rely on assumptions, estimated parameters, or random quantities. Because those choices may not perfectly represent the underlying risk, the resulting VaR can differ from the true but unobserved risk. A hedge based on the estimate may consequently leave exposure to that difference.

The answer treats estimation risk as a broad issue that applies to VaR and other calculations: it is the uncertainty and potential error introduced when estimated values stand in for actual values. The discussion is conceptual and brief. It gives no formula, backtesting procedure, empirical example, or method for separating estimation risk from other sources of VaR model error, so it serves as a definition rather than a practical assessment framework.

Key ideas

• VaR depends on assumptions and inputs that must be estimated.
• Differences between estimated and actual risk create estimation risk.
• Hedges sized from an estimated VaR can retain exposure when the estimate is inaccurate.
• The concept applies broadly to models and calculations, but the document gives no measurement procedure.

Tags

# What is Estimation Risk - VAR Backtest


# What is Estimation Risk - VAR Backtest












Simple Question. Can someone explain please: What is Estimation Risk in Value at Risk Backtesting

## Answer by kris123456 (score 2)

https://quant.stackexchange.com/a/26092

Even in calculating VAR, you have certain assumptions / constants / random numbers being used. Hence, even your VAR calculation is not 100% correct. So, you are estimating VAR and you hedge similar portion of risk, however your Estimations aren't 100% correct. This is Estimation Risk.

Estimation risk is a generic term. It could be applied to models, VAR, and all other calculations. It highlights the actual vs estimated values differences and the risk associated with this difference.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.