Skip to content
All library documents

Estimating Volatility from Transaction Prices and Bid-Ask Bounce

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

Summary

The discussion considers whether volatility can be estimated using execution prices alone. One proposed approach is to estimate the bid-ask spread from transaction-price changes using Roll’s method, then adjust observed volatility because bid-ask bounce adds noise to the underlying price variation. In principle, separating that component can help infer a less noisy volatility estimate.

A second response says transaction data or mid-quotes can support realized-volatility measurement, while Level II and Level III order-book data may introduce substantial noise from cancellations and rapidly placed orders. The material offers suggestions, not a complete estimator or empirical comparison. It does not specify sampling choices, assumptions, or adjustment equations, so applying the idea requires further methodological detail and care about market microstructure effects.

Key ideas

  • Bid-ask bounce can inflate volatility measured from transaction prices.
  • Roll’s method is suggested as a way to estimate the spread from execution prices.
  • An estimated spread may be used to adjust observed volatility toward underlying price variation.
  • Order-book data can contain cancellation and order-placement noise that complicates volatility measurement.
  • The discussion does not provide a full estimation procedure or validate the suggestions empirically.

Tags

VolatilityStatisticsMarket microstructure
Full text
# Measuring Volatility from Execution Prices


# Measuring Volatility from Execution Prices












I was told of a way of measuring the volatility of a stock by looking at the reported execution prices (from Level III or Level II data.) I'm well aware of how to measure volatility by looking at the mid-quote or similar but I have never heard of a way that would use solely the prices of the executions. Unfortunately, I cannot find any paper that would either use such method or define it. Can anyone point me in the right direction?

## Answer by Alex C (score 1)

https://quant.stackexchange.com/a/18425

Possibly you might be able to first estimate the bid-ask spread from execution prices, using the method of Roll (1984), and then adjust the volatility for this.

Essentially the bid-ask bounce adds to the underlying volatility, so knowing an estimate of the b/a and the apparent volatility, the underlying volatility could be recovered by subtraction.

Roll(1984)

## Answer by Jianxun Li (score 0)

https://quant.stackexchange.com/a/18431

Do you mean the "realize measure" of volatility using the intraday Transaction-and-Quote data? If that's the case, just trade data, or mid-quote would be sufficient.

Looking at Level II and level III data really introduces a lot of noises (hudge cancellation rates, orders placed by HFT blindly to gain time-priority advantage). Using those data to calculate some weighted-average of price is not a sound approach to me.

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.