An Explicit Solution to Black–Scholes Implied Volatility
Wolfgang Schadner a · 2026
Read An Explicit Solution to Black–Scholes Implied Volatility by Wolfgang Schadner a (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-24480.txt.
Introduction
Schadner shows Black-Scholes implied volatility is a quantile. The normalized option price acts as a probability on the variance scale, with the inverse Gaussian distribution linking price and volatility. This yields exact formulas for implied volatility from quantile functions, with only observable option inputs on the right. The result rewrites the price-volatility map as a distributional transform, gives implied volatility a first-passage interpretation, and identifies variance as the inversion coordinate. The formula reaches machine precision faster than a leading solver in the benchmark considered.
Catalog record
| CREATOR | Wolfgang Schadner a |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 13654 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.24480
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
An Explicit Solution to Black–Scholes Implied Volatility Wolfgang Schadner a Abstract This paper observes that the Black–Scholes call price can be written as the survival probability of an inverse Gaussian distribution, equivalently as a probability in variance space. Inverting this representation yields an analytically explicit formula for implied volatility in terms of the corresponding inverse Gaussian quantile function, with volatility on the left-hand side and only observable option inputs on the right-hand side. Numerical tests recover implied volatility to machine precision and, in a controlled setting, show the formula to be faster than a state-of-the-art benchmark. Keywords: option markets, implied volatility, Black–Scholes, inverse Gaussian JEL Classification: G13; C58 1 The Insight The Black–Scholes formula is explicit from volatility to option price. Market practice, however, usually needs the reverse map, from an observed option price to the implied volatility that reproduces it. Since 1 , implied volatility has become the standard convention for quoting and comparing option prices across strikes, maturities, assets, and dates. It is used, for example, throughout derivatives markets, empirical asset pricing, volatility forecasting, risk management, volatility-surface modelling, option-return studies, and tests of volatility risk premia. A formula for implied volatility therefore concerns one of the standard coordinates in which option markets are measured. Despite this importance, implied volatility is usually obtained by numerically inverting the Black–Scholes pricing formula. Given an option price, one searches for the volatility \sigma for which the Black–Scholes price matches the observed price. The literature has therefore developed explicit approximations and asymptotic formulas such as 9 , analytical upper and lower bounds ( 11 ) , highly efficient iterative algorithms with 6 ; 7 as a state-of-the-art benchmark, and exact but implicit representations through infinite series ( 4 ) . 5 showed that implied volatility is not D -finite, ruling out a broad class of usual closed-form representations. The representation derived below is explicit in a different sense. It does not express implied volatility through elementary or D -finite functions, but reduces the Black–Scholes inversion exactly to the evaluation of a specified inverse Gaussian quantile. For a call option with price C , strike K , forward F , risk-free discount factor D , and maturity T , the result for K\neq F is \displaystyle\sigma(K,C)=\frac{2}{\sqrt{T}}\left[\mathcal{F}^{-1}_{IG}\left((1-c)/m;\,\tfrac{2}{|k|},\,1\right)\right]^{-1/2},\qquad m=\begin{cases}1&K>F,\\ K/F&K<F\end{cases} (1) where c:=C/(DF) is the normalized call price, k:=\log(K/F) is forward log-moneyness, and \mathcal{F}^{-1}_{IG} denotes the inverse Gaussian quantile function. At the forward strike, the equation collapses to \displaystyle\sigma(K=F,C)=\frac{2}{\sqrt{T}}\Phi^{-1}\left(\frac{c+1}{2}\right) (2) The link to the spot price is S=DFe^{\delta T} with continuous dividend yield \delta . The formula turns the conversion from price to implied volatility into an inverse Gaussian quantile evaluation. It places implied volatility on the left-hand side and observable option inputs on the right-hand side. The only non-elementary operation is this quantile function. Thus, the formula removes the outer Black–Scholes root search and shifts the numerical task to the stable and accurate evaluation of that quantile. The underlying distribution is well studied, and the required quantile is available in standard numerical computing libraries. In the numerical tests below, the formula recovers implied volatility to machine precision. It is also about 3.4 times faster than 7 ’s implementation, a high-accuracy benchmark that is itself known for extremely fast Black–Scholes implied-volatility inversion ( 10 ; 12 , see, e.g.,) . The identity also gives a probabilistic interpretation. The normalized call price is the probability that a Brownian motion with drift k/2 has not yet reached level 1 by the clock time 4/v^{2} . Equivalently, total implied variance is the corresponding generalized-inverse Gaussian quantile. Thus, implied volatility is not merely the output of a volatility root search, but can also be read as a distributional transform of the market option price. 2 Derivation Let v=\sigma\sqrt{T} be the total implied volatility. The normalized Black–Scholes European call price is \displaystyle c_{BS}(k,v)=\Phi\left(-\frac{k}{v}+\frac{v}{2}\right)-e^{k}\Phi\left(-\frac{k}{v}-\frac{v}{2}\right). (3) The case k=0 is handled by Eq. ( 2 ). It remains to consider k\neq 0 . First take an out-of-the-money call, so k>0 . Write \mathcal{F}_{IG}(\cdot;\mu,\lambda) for the inverse Gaussian distribution function with mean parameter \mu and shape parameter \lambda , following the standard parametrization in 3 . Its survival function is \displaystyle 1-\mathcal{F}_{IG}(x;\mu,\lambda)=\Phi\left(-\frac{\sqrt{\lambda x}}{\mu}+\frac{\sqrt{\lambda}}{\sqrt{x}}\right)-e^{2\lambda/\mu}\Phi\left(-\frac{\sqrt{\lambda x}}{\mu}-\frac{\sqrt{\lambda}}{\sqrt{x}}\right). (4) With \lambda=1 , \mu=2/k , and x=4/v^{2} , this becomes \displaystyle c_{BS}(k,v)=1-\mathcal{F}_{IG}\left(\frac{4}{v^{2}};\frac{2}{k},1\right),\qquad k>0. (5) Inverting Eq. ( 5 ) gives \displaystyle v(k,c)=\frac{2}{\sqrt{\mathcal{F}^{-1}_{IG}\left(1-c;\frac{2}{k},1\right)}},\qquad k>0. (6) Finally, \sigma=v/\sqrt{T} . For k<0 , the call is in the money. Its out-of-the-money counterpart has log-moneyness |k|=-k and normalized call price \displaystyle\tilde{c}=1+\frac{F}{K}(c-1),\qquad 1-\tilde{c}=(1-c)\frac{F}{K}. (7) Combining this transformation with Eq. ( 6 ) gives Eq. ( 1 ), the analytically explicit Black–Scholes implied-volatility representation. The corresponding put formula follows from put–call parity, \displaystyle c-p=1-\frac{K}{F}=1-e^{k}. (8) For a put with normalized price p=P/(DF) and K\neq F , the probability argument in the inverse Gaussian quantile is (e^{k}-p)/m . Hence \displaystyle\sigma(K,P)=\frac{2}{\sqrt{T}}\left[\mathcal{F}^{-1}_{IG}\left((e^{k}-p)/m;\tfrac{2}{|k|},1\right)\right]^{-1/2},\qquad m=\begin{cases}1&K>F,\\ K/F&K<F.\end{cases} (9) Probabilistic interpretation. The identity also gives a direct probabilistic interpretation of the call price itself, analogous in spirit to the familiar probability reading of Black–Scholes delta. For an out-of-the-money call with k>0 , \displaystyle c_{BS}(k,v)=\mathbb{P}\left(Y_{k}>\frac{4}{v^{2}}\right),\qquad Y_{k}\sim IG\left(\frac{2}{k},1\right). (10) This representation has a first-passage-time interpretation. Let B_{t} denote standard Brownian motion and define \displaystyle\tau_{k}:=\inf\left\{t>0:B_{t}+\frac{k}{2}t=1\right\}. (11) Then \tau_{k} is the time it takes a Brownian motion with drift k/2 and diffusion volatility 1 to reach level 1 . For this drift and boundary, \tau_{k}\sim IG\left(\frac{2}{k},1\right) , so Y_{k} is identical in distribution to \tau_{k} . Therefore, the normalized call price can also be interpreted as the probability that this process has not yet reached level 1 by the variance-implied clock time 4/v^{2} . Equivalently, with Z_{k}:=4/Y_{k} , \displaystyle c_{BS}(k,v)=\mathbb{P}\left(Z_{k}<v^{2}\right),\qquad Z_{k}\sim GIG\left(\frac{1}{2},\frac{1}{4},k^{2}\right), (12) where GIG denotes the generalized inverse Gaussian distribution ( 8 , see, e.g.,) . This form moves the probability statement into variance space. Since k can be thought of as the log-return required to reach the strike, Z_{k} is the corresponding variance-space hurdle. The normalized call price is therefore the probability that implied variance v^{2} clears this moneyness-induced barrier. Thus, total implied variance is identified as a quantile of a generalized inverse Gaussian distribution, \displaystyle v(k,c)^{2}=\mathcal{F}_{GIG}^{-1}\left(c;\tfrac{1}{2},\tfrac{1}{4},k^{2}\right),\qquad k>0. (13) This differs from return-space probability readings. For example, delta can be viewed as the probability that the terminal return clears the strike hurdle. 2 show that call prices can also be given a return-space probability interpretation under a suitable change of measure. The representation here is different because it assigns the probability statement to variance space rather than return space. The additional insight is that the market option price is not only a value obtained by averaging terminal payoffs, but also a probability level on a moneyness-specific variance scale. Implied volatility is then obtained by reading off the corresponding quantile on that scale. 3 Numerical Evaluation I evaluate the formula on a deterministic grid of total implied volatilities and Black–Scholes call deltas. Total volatility is set to v\in\{0.01,0.05,0.10,\ldots,2.00\} , and call delta is set to \Delta\in\{0.05,0.20,0.30,0.45,0.55,0.70,0.80,0.95\} , giving 328 test cases. For each pair (v,\Delta) , the corresponding forward log-moneyness is obtained from the Black–Scholes delta relation, \displaystyle k=v\left(\frac{v}{2}-\Phi^{-1}(\Delta)\right). (14) The normalized call price is then generated from Eq. ( 3 ). The implied volatility formula is applied to this price, and the recovered total volatility is compared with the original input value. Figure 1: