Introducing Dependent Microstructure Noise and Integrated Volatility Estimation from High-Frequency Data for after-work quants

Tick data look precise until microstructure noise contaminates every print. What if that noise is not i.i.d., but dependent over time—so standard integrated-volatility recipes break?

Dependent Microstructure Noise and Integrated Volatility Estimation from High-Frequency Data (Li, Laeven, Vellekoop, 2018) studies how to think about integrated volatility when noise dependence matters.

Fit for learners: Open-PDF classic/working paper on microstructure — useful for after-work Python quant learners building intuition before coding.

This Hub note is a plain-language intro. Not investment advice.

The intuition (without the full asymptotics)

• Observed high-frequency prices = efficient price + noise. Noise can be bid–ask bounce, rounding, latency artifacts, and more.
• Many textbook fixes assume noise is roughly independent. Dependence changes bias/variance trade-offs for realised volatility and cousins.
• The paper works in that harder regime and discusses estimators suited to it (see PDF for formal results).
• Practical habit for learners: never treat a 1-second realised-vol number as “truth” without asking about sampling and noise.

Read theorems and simulations in the PDF. Do not turn a volatility estimator into a trading signal by default.

Why after-work learners should care

Evening notebooks love np.diff(log_price)**2. Microstructure papers explain why that sum can lie. Before you chain volatility features into any model, you want this vocabulary: noise, dependence, sampling frequency, integrated volatility.

Healthy practice: downsample a public tick or 1-minute series, compute realised vol at several frequencies, and plot the famous volatility signature—label “learning only.”

One-evening reading checklist

1. Read until “dependent microstructure noise” is clear in your own words.
2. Skim what goes wrong for naive integrated-volatility estimators.
3. Note one proposed remedy or modelling assumption from the paper.
4. Write a disclaimer you would put above any personal realised-vol chart.

Download the PDF

Download PDF

Landing: arXiv:1704.08964

Important disclaimer

Not investment advice. Volatility estimators are research and risk tools. They are not instructions to trade options or futures.

What this is not

• Not CEF / subsidy framing
• Not Quant or Macro course CTA
• Not a volatility-product sales page

Next step

Single course CTA: SoloQuant Academy’s Introduction to Python Programming—arrays and plots first; high-frequency econometrics becomes readable later.