Skip to content

The VIX Volatility Index: A Very Thorough Look at It: summary and key takeaways

The paper rebuilds the VIX from CBOE option quote data and tests which criticisms of the index hold up. It finds that the zero-bid tail cutoff can shift the VIX materially, that the Taylor approximation matters little, and that the VIX settlement value differs from the daily index by design.

Osterrieder, J.; Vetter, L.; Röschli, K. · 2019 · SSRN Electronic JournalSummary updated 20 November 2026
Published2019Working paper
Asset focusVolatility/VIXVIX overview
TypeResearchVolatility & VIX
Reading time4 minthis summary

The VIX Volatility Index: A Very Thorough Look at It at a glance

The question
Which of the claims made against the VIX formula and its derivative settlement process are justified when the index is replicated from raw option data?
Dataset & period
SPX option end-of-day quotes and trades from January 2015 to March 2018 (3:15 p.m. and 2:45 p.m. Chicago time); one-minute intraday SPX option data from January 2018 to March 2018; daily VIX open/high/low/close from January 2004 to March 2018; intraday VIX values at 15-second granularity from January 2017 to March 2018. All from CBOE; Treasury bill rates for the risk-free rate (set to zero in most tests). Sample counts of options per day are not stated as a single figure.
Method
The authors replicate the CBOE VIX formula from the white paper and check it against the white paper example, where it matches exactly. They derive the formula from variance-swap theory and isolate two approximations: a Taylor expansion of the log term and the rule that drops options after two consecutive zero bids. They then re-run the index without each approximation, vary the number of options included, and examine how each input could be moved by a market participant.

Key findings of The VIX Volatility Index: A Very Thorough Look at It

  • On 5 February 2018 the VIX closed at 37.32, a one-day rise of 115%, while products tied to short-term VIX futures were wiped out.
  • The additive correction term in the formula is tiny, at most 0.6% of near-term variance, so the weighted sum of option quotes accounts for at least 99.4% of the value.
  • Replacing the Taylor-based correction term with the exact expression changes the simulated VIX by at most 0.3%, and the CBOE version understates volatility slightly.
  • Cutting the strike range after two consecutive zero bids changed the VIX by more than 7.5% on some days, with up to 100 options excluded, mostly puts.
  • Deep out-of-the-money puts carry the largest weight in the formula, so quotes placed in the tails have the biggest leverage on the index.
  • The options closest to the forward-based strike K0 carry most of the information: the nearest ten options account for up to 60% of the VIX, and five puts plus five calls explain between 50 and 70 percent.

Robustness, as reported in the paper

  • The replication reproduces the CBOE white paper example exactly.
  • A move of the risk-free rate by an absolute 5% changes the VIX by at most 0.3%, which is why the rate is set to zero in most tests.
  • Replication outliers of up to 6% against the published close are traced partly to timestamp gaps of up to 15 seconds and to high-volatility days.
  • Rolling 30-day correlations show that index versions with fewer options still track the full VIX closely.

Caveats and limits

  • Part of the option data could not be obtained from CBOE, leaving gaps in the time series, and intraday data cover only a short window.
  • Manipulation is analysed through formula sensitivity and simulation; the paper does not show that any particular manipulation occurred.
  • Interest rates are set to zero in most analyses, which adds replication error (partly acknowledged by the authors).
  • Most results are shown in figures, so exact numerical results beyond those in the prose are hard to verify (inferred).

Items marked (inferred) are our reading of the design, not statements by the authors.

Why it matters

Anyone trading VIX futures, options or related ETPs is exposed to an index built from illiquid far-out-of-the-money quotes. The paper explains why the settlement value can differ from the prior-day VIX and which formula rules matter most for the level of the index.

Questions about this paper

Does the VIX settlement value jump from the previous day because of manipulation?

The paper argues the gap is mostly structural. Settlement uses a single SPX or SPXW expiry and opening prices, while the daily VIX blends two expiries using mid-quotes.

How much does the zero-bid rule matter for the VIX level?

The paper reports that including options past two consecutive zero bids changed the VIX by more than 7.5% on some days, with up to 100 options excluded.

Which options matter most for the VIX value?

The options nearest the forward-based strike carry most of the information, but deep out-of-the-money puts have the highest per-quote weight in the formula.

Where can I read the original paper?

The original is at its publisher: https://doi.org/10.2139/ssrn.3311727. We link to the paper and do not host the PDF.

Follow new quant research

Download MacroFinance market data as CSV or Excel, or pull it by API, and test the ideas in these papers yourself. New summaries are added as we review them.

Create a free account
Summaries are written by MacroFinance from the published papers, in our own words, for information only. They are not investment advice and may omit detail. Papers are not hosted here: each page links to the original at its publisher.