Dupire Realized Volatility Estimators from Daily Price Ranges

Summary

The document discusses two realized volatility estimators attributed to Bruno Dupire and asks what makes an estimator tradable. It reports a proposed estimator that excludes overnight movement, constructed from the squared distances between the day's high, close, and low. It also asks whether an overnight-inclusive version can be inferred by combining those intraday range components with the gap from the previous close to the current open.

The text provides one formula attributed to a discussion participant and a second formula only as a conjecture; it does not establish that either is correct or explain how to trade or validate the estimates. It mentions a presentation whose formulas were unavailable to the author, so readers should treat the overnight expression as unverified. No performance evidence, assumptions, or comparison with standard realized volatility measures is supplied.

Key ideas

• The document asks how a realized volatility estimate can be made tradable.
• One proposed estimator uses squared high-to-close and close-to-low distances, excluding overnight moves.
• An overnight-inclusive estimator is suggested using the previous close to current open gap.
• The overnight formula is a conjecture, and the document provides no validation or performance evidence.

Tags

# Dupire's tradable realized volatility estimators


# Dupire's tradable realized volatility estimators












It seems to have vanished off the web but a few years back Dupire published two tradable realized volatility estimators in 2015, I even asked a question around it here. What makes a realized vol estimate "tradeable"?

However, the only presentation I could find on the web has the formulas removed here: https://www.math.cmu.edu/CCF/CCFevents/shreve/abstracts/B.Dupire.pdf

Does any have the formulas for the two estimators or know why they are no longer available on the web?

Update:

Thanks to Pleb, we have one one of formulas, Dupire Realized Vol without night:

$\sigma_i = \frac{(H_i - C_i)^2 - (C_i - L_i)^2 }{2}$

Is it safe to assume Dupires formula with overnight is a follows?

$\sigma_i = \frac{(H_i - C_i)^2 + (C_i - L_i)^2 + (O_i - C_{i-1})^2}{3}$

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.