Skip to content
All library documents

PRIIPs Cornish-Fisher VaR as a Historical Real-World Measure

Article Quant Q&A · Author: Thegamer23
Read the original

Summary

The document asks whether the Cornish-Fisher value-at-risk measure referenced in PRIIPs should be interpreted as a real-world measure, especially when the calculation appears to use a zero average return. The question highlights a distinction between the return assumption inside a risk calculation and the source of the data used to estimate risk.

The answer points to a nearby regulatory provision that appears to base the calculation on five years of historical returns, and on that basis characterizes it as real-world. The document does not explain the Cornish-Fisher adjustment, show the regulation’s full methodology, or resolve how a zero-return assumption fits with historical estimation. Its conclusion is therefore limited and depends on the cited provision; it provides no calculation or empirical comparison.

Key ideas

  • The question concerns the interpretation of Cornish-Fisher VaR under PRIIPs.
  • A zero assumed average return does not by itself establish whether a risk measure is real-world or risk-neutral.
  • The answer relies on a regulatory provision that appears to use five years of historical returns.
  • The brief response does not detail the methodology or fully explain the role of the mean-return assumption.

Tags

Risk managementStatistics
Full text
# Risk neutrality in the recent PRIIPs regulation


# Risk neutrality in the recent PRIIPs regulation












In the recent regulation (link pag 31 point 12) is reported the VaR according to Cornish Fisher. Is this a real world measure?

By making computation I have seen that they assume an average return that is null. But this formula is still real world? Why?

## Answer by Adam N. (score 1)

https://quant.stackexchange.com/a/33312

Point 9. above seems to imply it's a real world measure based on 5 years worth of historical returns.

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.