Paper: arXiv 2309.00943
Authors: Evgenii Vladimirov
Abstract
This paper proposes the option-implied Fourier-cosine method, iCOS, for non-parametric estimation of risk-neutral densities, option prices, and option sensitivities. The iCOS method leverages the Fourier-based COS technique, proposed by Fang and Oosterlee (2008), by utilizing the option-implied cosine series coefficients. Notably, this procedure does not rely on any model assumptions about the underlying asset price dynamics, it is fully non-parametric, and it does not involve any numerical optimization. These features make it rather general and computationally appealing. Furthermore, we derive the asymptotic properties of the proposed non-parametric estimators and study their finite-sample behavior in Monte Carlo simulations. Our empirical analysis using S&P 500 index options and Amazon equity options illustrates the effectiveness of the iCOS method in extracting valuable information from option prices under different market conditions. Additionally, we apply our methodology to dissect and quantify observation and discretization errors in the VIX index.
Complexity vs Empirical Score
- Math Complexity: 7.0/10
- Empirical Rigor: 7.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced mathematical tools including Fourier-cosine series, characteristic functions, and asymptotic theory with derivations, indicating high mathematical complexity. It also demonstrates strong empirical rigor with extensive Monte Carlo simulations, real-world data applications (S&P 500, Amazon, VIX), and explicit discussion of estimation errors, making it backtest-ready.
Research Flowchart
flowchart TD
A["Research Goal: Non-parametric Estimation of RND & Option Pricing"] --> B["Input: Observed Market Option Prices"]
B --> C["Methodology: iCOS Framework"]
C --> D["Step 1: Estimate Option-Implied<br>COS Coefficients"]
C --> E["Step 2: Reconstruct Risk-Neutral<br>Density via Inverse COS Transform"]
D & E --> F{"Computational Process"}
F --> G["Compute Option Prices & Sensitivities<br>Extract VIX Information"]
G --> H["Key Findings & Outcomes"]
H --> I["Model-Free, Non-Parametric Estimation"]
H --> J["Proven Accuracy via<br>Monte Carlo & S&P 500/Amazon Data"]
H --> K["Quantification of<br>Observation & Discretization Errors"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes the option‑implied Fourier‑cosine (iCOS) method for non‑parametric estimation of risk‑neutral densities, option prices, and option sensitivities.
Source sentence
This paper proposes the option‑implied Fourier‑cosine method, iCOS, for non‑parametric estimation of risk‑neutral densities, option prices, and option sensitivities.
Assumes 2
- The smallest and largest strike prices are fixed while the number of options n increases as the strike mesh shrinks. data
Source sentence
we assume that the smallest and the largest strike prices, K and K, are fixed, and the number of options n with strike prices between them goes to infinity by shrinking the strike mesh.
- The solution to the stochastic process ST is homogeneous of degree one in the initial stock price S0 (∂ST/∂S0 = ST/S0). model
Source sentence
Assumption 4 The solution to the stochastic process ST is homogenous of degree one as a function of initial stock price S0, i.e., ∂ST∂S0 = STS0 .
Data
- SPX options options public
Source sentence
In the first empirical illustration, we consider options on the S&P 500 stock market index (SPX) obtained from the Chicago Board Options Exchange (CBOE)
- AMZN options options unknown
Source sentence
In our second application, we examine equity options on Amazon with a very short time‑to‑maturity of T = 1 day. These options are traded on the Earning Announcement Day (EAD) of April 26...
Costs
Not applicable
Finds 4
- Pricing errors of the iCOS estimates are typically between $‑0.05 and $0.05, i.e., within two ticks, and most lie within half the bid‑ask spread. pricing-calibration
Source sentence
most pricing errors fall within half of the bid‑ask spread, indicating a very good pricing performance of the method.
- Monte Carlo simulations show no significant bias for option prices across strikes and maturities, and the estimated asymptotic standard deviations match the simulated standard errors. statistical
Source sentence
the asymptotic standard deviations, defined as the square root of the average estimated asymptotic variance, roughly correspond to the Monte Carlo standard errors.
- The iCOS RND estimator is asymptotically unbiased and exhibits small variance when the number of expansion terms grows slower than n^{-1/6}. theoretical
Source sentence
when this condition is met, the option‑implied RND remains asymptotically unbiased.
- Compared to kernel smoothing, iCOS yields insignificant biases with smaller variances for RND and delta estimates, while kernel smoothing can produce large variances and arbitrage‑violating RND shapes. performance
Source sentence
able to simultaneously capture the shape of the observed option data, RND and deltas with insignificant biases and small variances.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 25 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.