Skip to content
All library documents

Using Range-Based Estimators for Volatility Forecasting

Article Quant Q&A · Author: Porsche Tan
Read the original

Summary

The document asks whether Parkinson, Rogers–Satchell, and Yang–Zhang estimators are used to forecast future volatility. The response says they can be used, and highlights Yang–Zhang’s reported comparison with close-to-close volatility estimation: it is described as less biased and less variable. It also notes that high-frequency data may be a better input and is used by many practitioners.

The answer points to a high-frequency volatility estimator as further reading, but does not explain the estimators’ formulas, forecast horizons, assumptions, or how to evaluate them in a forecasting task. It offers no dataset or out-of-sample results. The practical takeaway is limited: range-based estimators are relevant candidates, while high-frequency measurements are another commonly used approach; the document does not establish which performs best for a particular market or application.

Key ideas

  • Parkinson, Rogers–Satchell, and Yang–Zhang estimators can be used in volatility forecasting.
  • Yang–Zhang is described as less biased and less variable than close-to-close estimation.
  • High-frequency data is presented as another widely used source for volatility measurement.
  • The document gives no forecasting setup, comparative test, or guidance on choosing an estimator for a specific use.

Tags

VolatilityStatisticsBacktesting
Full text
# Volatility Forecasting


# Volatility Forecasting












I would like to clear a certain doubt about volatility forecasting: Are models like Parkinson, Rogers Satchell and Yang Zhang used for predicting future volatility?

## Answer by herminat0r (score 3)

https://quant.stackexchange.com/a/76300

From what I have seen yes. Yang & Zhang showed that their estimator is less biased and less volatile compared to the close-to-close estimator. However, as mentioned, high-frequency data is probably better and is used by many. One good paper on this is the HEAVY-estimator from the Oxford-Man Institute (link here https://scholar.harvard.edu/files/multiHeavy.pdf).

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.