Risk-Neutral Density: Extracting State Prices from Option Quotes

For educational purposes only; not investment advice.

Direct answer

A risk-neutral density assigns pricing weights to possible underlying prices at a specified future date so that discounted expected payoffs match current option prices. It is also called an option-implied or state-price density after normalization. It describes prices under a pricing measure, not investors’ literal beliefs or the statistical frequency of future outcomes.

For one underlying and maturity, a sufficiently smooth, arbitrage-consistent curve of European Call prices across strikes contains the distribution. The first strike derivative identifies a discounted tail probability; the second strike derivative identifies the discounted density. In real chains, sparse strikes and noisy, non-synchronous quotes make this an estimation problem rather than a direct reading from one option.

From Call prices to a distribution

For a European Call with strike K, maturity T, and continuously compounded rate r:

C(K,T) = e^(−rT) E^Q[(S_T − K)⁺]

Under regularity assumptions, differentiating with respect to strike gives:

−e^(rT) ∂C/∂K = Q(S_T > K)

q(K) = e^(rT) ∂²C/∂K²

Here Q is the risk-neutral measure and q(K) its density at terminal price K. Call prices should decline with strike and be convex; a negative second derivative would imply a negative density and violate the corresponding butterfly no-arbitrage condition.

The calculation requires same-maturity, same-settlement European-equivalent prices. Analysts commonly convert quotes to implied volatilities, fit a smooth arbitrage-controlled smile, convert back to prices, then differentiate. Rates, dividends or forward prices, exercise style, settlement, quote time, and tail extrapolation must be consistent. American equity options cannot be inserted without treating early exercise and dividends.

Three-strike finite-difference example

Suppose same-maturity European Calls have a strike spacing h = $5, six months to maturity, and rate 4%:

Strike	Call price
`$95`	`$9.80`
`$100`	`$6.40`
`$105`	`$4.00`

The central second-difference estimate at strike $100 is:

∂²C/∂K² ≈ [C(95) − 2C(100) + C(105)] / 5²

= ($9.80 − 2 × $6.40 + $4.00) / 25 = 0.0400 per dollar

With rT = 0.04 × 0.5 = 0.02:

q(100) ≈ e^0.02 × 0.0400 = 0.04081 per dollar

Multiplying the point density by the $5 strike interval gives a rough local probability mass of 0.04081 × 5 = 20.4%. This is only a coarse approximation around the node, not an exact probability for a defined interval. A proper estimate uses a full, smoothed, arbitrage-consistent curve and integrates the density between explicit boundaries.

The central first difference is (4.00 − 9.80) / 10 = −0.58, implying a rough risk-neutral tail probability Q(S_T > $100) ≈ e^0.02 × 0.58 = 59.2%. Bid/Ask choices or a small quote error can materially change both estimates, especially the second derivative.

Extraction checklist

Use one underlying, expiration, settlement method, observation time, and consistent forward convention.
Prefer executable Bid/Ask-aware filters; Mid prices can create a smooth-looking but untradeable curve.
Remove stale, crossed, zero-bid, erroneous, and corporate-action-mismatched quotes under documented rules.
Account for dividends, rates, carry, and American early exercise before treating prices as European.
Fit option prices or total variance with explicit calendar and butterfly no-arbitrage controls.
Inspect monotonicity, convexity, nonnegative density, unit integral, and a forward-consistent first moment.
Test interpolation, strike spacing, bandwidth, and boundary choices.
Report observed-strike support separately from extrapolated tails; far-tail density is weakly identified.
Repeat the extraction using Bid-, Mid-, and Ask-based surfaces to show quote uncertainty.
Preserve the raw chain, timestamp, filters, curve parameters, rate and dividend inputs, and code version.
Do not compare maturities without accounting for horizon and forward changes.
Separate pricing-measure conclusions from real-world probability and expected-return claims.

Common misconceptions

“Risk-neutral means investors are indifferent to risk.” It is a pricing representation that embeds risk premiums in state prices.
“A 20% risk-neutral probability means a 20% objective chance.” Physical probabilities require a pricing kernel or additional assumptions and data.
“One option reveals the distribution.” A cross-section across strikes and one common maturity is needed.
“Implied volatility is the density.” IV is a quoting transformation; the strike shape must be converted to prices and differentiated.
“More precise differentiation is always better.” Second derivatives amplify quote noise and interpolation artifacts.
“Negative density proves an arbitrage anyone can trade.” It may reflect stale or non-synchronous Mid quotes; executable prices and costs matter.
“The tails are observed.” Sparse extreme-strike quotes make tails heavily dependent on extrapolation.
“The distribution predicts the direction of the underlying.” It summarizes current pricing weights and can move with risk aversion, hedging demand, and liquidity.