Skip to content
All library documents

Interpreting Hurst Exponent Tests with Bid-Ask Bounce in Mind

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

Summary

The document considers whether a Hurst exponent test can identify whether financial prices are trending, mean reverting, or closer to a random walk. The question raises bid-ask bounce as a possible source of distortion and asks whether the test is too simplistic for practical use, as well as what alternatives might better suit financial data.

The answer notes that the referenced article uses several statistical tests that view the question from different angles. It adds that bid-ask bounce is not significant when using daily data for liquid, heavily traded instruments. This is a brief, qualified response rather than a detailed evaluation: it does not name the tests, compare their performance, or establish how results vary across sampling frequencies, liquidity levels, or instruments. Its practical reassurance is limited to the daily-data setting it describes.

Key ideas

  • The Hurst exponent is used to assess persistence patterns such as trending or mean reversion.
  • Bid-ask bounce can be a concern when interpreting price-series tests.
  • The response says bounce is not significant for daily data on liquid, heavily traded instruments.
  • Using several statistical tests can provide different perspectives on the time-series question.

Tags

StatisticsMean reversionTrend following
Full text
# How to know if a time series is trending or mean reverting?


# How to know if a time series is trending or mean reverting?












I came across Michael Halls-Moore article on using the Hurst exponent test to determine if a price time-series is mean-reverting, trend-following or closer to a random walk, but doesn't this disregard the fact that financial instruments often exhibit a "bid-ask bounce affect" ? In which case, is his approach of using the Hurst exponent a bit too naive to use in practice?

Are there other ways to determine if a price time-series is mean-reverting or trend following specifically for financial instruments that more applicable in real-world settings?

## Answer by onlyvix.blogspot.com (score 3)

https://quant.stackexchange.com/a/35269

The article that you linked mentioned several statistical tests, that approach the question from different angles. However, you should know that bid-ask bounce is not significant if you're looking at daily data on liquid (heavily traded) financial instruments.

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.