Key Takeaways
- A spread-out gauge: DSPX reflects how far apart S&P 500 stocks are expected to move.
- 30 days ahead: it is priced from index and single-stock options, like a VIX for dispersion.
- Volatility units: a reading of 20 means a 20% annualized cross-sectional standard deviation.
- The mirror of correlation: high dispersion usually means low implied correlation.
- Earnings lift it: single-stock options get bid before earnings season, pushing readings up.
DSPX is the Cboe S&P 500 Dispersion Index, a measure of how differently the options market expects the stocks inside the S&P 500 to move from each other over the next 30 calendar days. According to the S&P Dow Jones Indices and Cboe launch announcement, it is calculated from the prices of S&P 500 index options and single-stock options on selected index members, "using a modified version of the VIX Index methodology." A higher reading means the market expects stocks to go their own ways; a lower one means it expects them to move as a pack.
Think of it as the VIX's counterpart. The VIX prices how much the index will move. DSPX prices how much the stocks inside the index will move relative to each other, which is a separate question with a separate answer.
What the DSPX Index Measures
The short version: dispersion is the spread of returns across stocks, and DSPX is the version of that spread implied by option prices.
Dispersion describes how far apart individual stock returns land over a period. If every stock in an index rose about 2% in a month, dispersion would be near zero even though the index moved. If half the stocks rose 10% and half fell 10%, the index might barely move while dispersion was very high. Cboe's research note Demystifying Dispersion with the DSPX Index illustrates this with two months that each had the VIX averaging around 15: in one, the best and worst S&P 500 sectors ended 2.7% apart, and in the other they ended 14.7% apart.
That example is realized dispersion, measured from past returns. DSPX is forward-looking. Cboe's dispersion index page calls it "a forward-looking implied measure" and describes it as an indication of "the market's perception of the near-term opportunity set for diversification."
S&P Dow Jones Indices and Cboe announced the index on September 25, 2023, for launch on September 27, 2023, with S&P DJI as benchmark administrator and Cboe as calculation agent, per the launch release. The same release notes that, like the VIX, DSPX "is not expected to display a long-term trend for growth" and is instead expected to mean-revert toward a level that can drift over time.
The VIX prices how much the index will move. DSPX prices how much the stocks inside it will move differently from each other.
How DSPX Is Calculated
The core idea: compare the weighted average variance of the stocks with the variance of the index, then take the square root of the gap.
Cboe's research note describes DSPX as "the square root of the weighted single stock variances versus index variance." In practical terms, the calculation has three pieces, per Cboe's dispersion page:
- Single-stock variance. A market-cap-weighted 30-day implied variance for a basket of S&P 500 constituents, which Cboe publishes, converted to volatility, as the Cboe S&P 500 Constituent Volatility Index (VIXEQ).
- Index variance. The square of the VIX, which represents the S&P 500's expected 30-day variance. The page notes the VIX level is "an explicit input" to DSPX.
- The gap. Subtract index variance from single-stock variance and take the square root, which returns the answer to annualized volatility units.
Variance is volatility squared, and working in variance is what makes the gap meaningful. In theory an index's variance cannot exceed the weighted average variance of its stocks. The shortfall comes from two things: stocks not moving in lockstep, and stocks having different volatilities, so they move by different amounts even when they move together. That shortfall is the dispersion. Because DSPX is built from option prices on a basket of constituents, the implied gap can occasionally close completely.
A worked example with two stocks. Suppose a toy index holds XYZ and ABC at 50% each, both stocks carry 30-day implied volatility of 30%, and index options imply 24%:
- Weighted single-stock variance: (0.5 × 0.30²) + (0.5 × 0.30²) = 0.045 + 0.045 = 0.09.
- Index variance: 0.24 × 0.24 = 0.0576.
- The gap: 0.09 − 0.0576 = 0.0324.
- Implied dispersion: the square root of 0.0324 is 0.18, which an index quoted like DSPX would show as 18.
Now hold the index at 24% and suppose both stocks' implied volatility rises to 32%, for example because each has an earnings report inside the 30-day window. Weighted single-stock variance becomes 0.1024, the gap becomes 0.1024 − 0.0576 = 0.0448, and the square root is about 0.212, a reading of roughly 21.2. Nothing happened to the index; the stocks simply got more expensive relative to it.
The other direction is just as instructive. Suppose index options in the toy case jumped to 30%, matching the stocks. Index variance would be 0.09, the gap would be zero, and the reading would be 0: the options market would be pricing the two stocks to move as one.
How DSPX Differs From the VIX and Implied Correlation
The neighbors it gets confused with: the VIX and the implied correlation indexes.
- What it measures: the VIX estimates how much the S&P 500 will move over 30 days; DSPX estimates how widely its stocks will spread apart over the same window.
- Inputs: the VIX uses S&P 500 index options only; DSPX uses the VIX as an input plus single-stock option prices.
- Units: both are quoted in annualized volatility points, so a DSPX of 20 and a VIX of 20 are on the same scale.
- Direction with the VIX: holding everything else constant, a higher VIX lowers DSPX, because index variance is subtracted. In practice they often rise together, because single-stock variance can climb faster than index variance.
Cboe's dispersion page says DSPX and the VIX "follow similar trends over time and their absolute levels are positively correlated," which is why the two are best read together rather than as substitutes.
Implied correlation is the closest relative. Our explainer on COR1M, Cboe's implied correlation index, uses the same toy inputs (two stocks at 30%, index at 24%) and gets an implied correlation of 28. Here those inputs give a dispersion of 18. Both numbers describe closely related gaps between index and single-stock option prices: COR1M expresses it as an average correlation among the 50 largest stocks, and DSPX expresses it as an expected spread of returns. When one rises, the other usually falls.
Why DSPX Matters to Options Traders
What it explains: why single-stock options can stay expensive while the VIX looks calm, and the reverse.
Cboe's research note interprets the level directly: "A DSPX index level of 20%, for example, implies an expectation that the cross-sectional standard deviation of the S&P 500 stock returns will average 20% over the next month." For example, 20% annualized works out to roughly 5.7 percentage points over a 30-day window (20 × the square root of 30/365). A high reading says the market expects big gaps between winners and losers inside the index. A low reading says it expects stocks to move together, so the index and its components should behave alike.
That matters when you choose between index and single-stock exposure. If you sell premium on individual stocks, a high DSPX means those options are priced for large stock-specific moves, as our guide to how earnings change implied volatility explains. If you hold index options as a hedge, DSPX helps explain why a calm index can sit on top of turbulent stocks, a pattern we also see when market concentration lets a few mega-caps steer the S&P 500.
Cboe's note frames the VIX as a measure of broad market risk and DSPX as a measure of stock-specific, fundamental risk, so reading them together can help separate a macro shock from a change in company fundamentals. For traders who build positions around the gap itself, our guide to correlation and dispersion trading covers how those multi-leg trades are structured and where they break down. Whether any of it fits depends on your risk tolerance, account size, and experience with multi-leg positions.
Two months can share the same VIX and still look completely different underneath. DSPX is the number that tells them apart.
Edge Cases and Gotchas
Where the simple reading breaks down:
- Earnings season inflates it. Cboe's research note found that DSPX rose an average of about 4.3 points in the month before earnings season and ran about 3.5 points higher during earnings periods than outside them. A high reading in late January, April, July or October may be the calendar talking.
- It can hit zero. In Cboe's DSPX daily history, the index closed at 0.00 on February 8, 2018. Cboe's note explains that index volatility surged during that month's Volmageddon sell-off while single-stock volatility stayed largely stable, erasing the gap.
- Early history is back-tested. The history file starts in June 2014, and Cboe's research note labels all data before the September 27, 2023 launch as "backtested hypothetical data," so long-run comparisons lean partly on calculated values.
- A rising VIX can push it either way. Because index variance is subtracted, a VIX spike with steady single-stock volatility lowers DSPX, while a sell-off driven by company fundamentals can lift both. Cboe's note contrasts February 2018 with the fourth quarter of 2018 on exactly this point.
- The ceiling is not fixed. The highest close in the history file is 58.87, on March 18, 2020, during the early pandemic sell-off, when single-stock volatility outran even a very high VIX.
FAQ
These answers cover what traders usually ask after seeing DSPX quoted next to the VIX and wondering what a given reading implies.
What Is a Normal DSPX Reading?
There is no fixed normal, and the range has shifted over time.
In Cboe's daily history from mid-2014 through 2025, which includes back-tested values from before the 2023 launch, the median close was about 24 by our calculation, and from 2023 through 2025 it was about 29. Treat any band as historical context.
Can DSPX Go to Zero?
Yes. Because the index is a square root of single-stock variance minus index variance, it reads zero when the VIX reaches or exceeds VIXEQ, so implied index variance is at or above the weighted single-stock variance and there is no positive gap left to take a square root of.
Cboe's history shows a close of 0.00 on February 8, 2018, during the Volmageddon spike in index volatility.
How Is DSPX Different From COR1M?
Both compare index option prices with single-stock option prices, but they express the gap differently. COR1M solves for an average correlation among the 50 largest stocks, while DSPX expresses the gap as an expected spread of returns in volatility units over 30 days.
Can You Trade DSPX Directly?
DSPX is published as an index. Cboe has said it may serve as the basis for listed futures and options, so check Cboe's product listings for what currently trades. Traders who want dispersion exposure usually build it from index and single-stock options.
Where Can I Find the DSPX Index?
Cboe publishes current values and charts on its DSPX index dashboard, along with a downloadable daily history file. Market data vendors also carry it, and the exact symbol varies by platform.
