Paper: arXiv 2611.23495
Authors: Paul Glasserman, Mike Li, Dan Pirjol
Abstract
Call option prices, viewed as functions of strike and time-to-expiration, are totally positive of order 2 (TP2) if the ratio of the price of a higher-strike call to a lower-strike call increases with time-to-expiration. A corresponding property applies to put option prices. Prior work has developed conditions for these properties to hold for European options; here we establish these properties for American options. The key condition we require is that log returns on the underlying asset have a log-concave density. Gaussian densities provide an important special case. Within our framework, we show that total positivity properties hold more broadly for American than European options. Empirically, we find that violations of TP2 properties are infrequent, typically small, and reverse quickly. Trades that bet against violations achieve high hit rates. Violations are more frequent among European than American options, a pattern that is qualitatively consistent with our theoretical results.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 6.0/10 (capped from 7.5: no out-of-sample test; see penalty flags)
- Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 7.5)
Why this score: This paper presents a rigorous mathematical framework for total positivity properties in American options, extending prior work on European options. It combines advanced theoretical derivations with a solid empirical investigation, demonstrating both strong mathematical foundations and practical relevance. The novelty lies in applying these properties to the more complex American option structure and showing broader applicability.
Research Flowchart
flowchart TD
A["Research Goal: Establish TP2 for American Options"] --> B{"Methodology: Extend TP2 conditions from European to American Options"};
B --> C["Key Condition: Log-concave density of log returns (e.g., Gaussian)"];
C --> D{"Computational Process: Theoretical Derivations & Empirical Analysis"};
D --> E["Data/Inputs: American & European Option Prices; Underlying Asset Data"];
E --> F{"Key Findings: TP2 holds more broadly for American than European options under log-concavity"};
F --> G["Empirical Findings: TP2 violations infrequent, small, quickly reverse; more frequent for European than American options"];
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 establishes that American call and put option prices are totally positive of order 2 (TP2/RR2) under log‑concave return densities, and provides empirical evidence on the frequency and size of violations.
Source sentence
Prior work has developed conditions for these properties to hold for European options; here we establish these properties for American options. In this paper, we investigate total positivity properties of American options.
Assumes 3
- Log returns on the underlying asset have a log‑concave density. model
Source sentence
The key condition we require is that log returns on the underlying asset have a log-concave density.
- For discrete log‑return distributions, the probability mass function is log‑concave. model
Source sentence
Suppose (33) holds and suppose the Zi have a log-concave probability mass function.
- The theoretical analysis assumes a model without dividends or interest. market
Source sentence
1 Introduction In a model without dividends or interest, the absence of arbitrage requires that the prices of European call and put options increase with the time-to-expiration.
Data
SPY and SPX options · options · daily
- SPY option data options commercial
Source sentence
We mainly focus on options on SPY (January 2005 to August 2025), an exchange-traded fund that tracks the S&P 500, which allows us to compare American SPY options with European SPX options.
OptionMetrics IvyDB WRDS Cboe FRED
Costs
Unclear
Finds 5
- Violations of TP2/RR2 properties are infrequent, typically small, and reverse quickly. statistical
Source sentence
Empirically, we find that violations of TP2 properties are infrequent, typically small, and reverse quickly. Taken together with Table 7.1, these results show that TP2/RR2 violations are rare and tend to be small when they occur.
- Overall violation rates for SPY are less than 1% for calls and less than 2% for puts. statistical
Source sentence
For SPY, overall violation rates are less than 1% for calls and less than 2% for puts.
- Trades that bet against violations achieve high hit rates, indicating near‑arbitrage opportunities. performance
Source sentence
Trades that bet against violations achieve high hit rates. Trades that bet against violations achieve high hit rates, suggesting that these are “near arbitrage” opportunities.
- Violation rates are higher for European SPX options than for American SPY options, consistent with the theory. statistical
Source sentence
Violations are more frequent among European than American options, a pattern that is qualitatively consistent with our theoretical results. Violation rates for European SPX options are generally higher than for American SPY options.
- Ex‑dividend safe subsets show higher violation rates than the full dataset, largely due to composition of short‑term maturities. statistical
Source sentence
The ex-dividend safe subset experiences higher violation rates than the full dataset. However, the figure also shows that this difference is largely explained by the change in composition of maturities: the DTE ≤30 subset also experiences higher violation rates than the full dataset.
Checked by others
Supported by
- Hidden Risks and Optionalities in American Options
Compare the findings
This paper: Violation rates are higher for European SPX options than for American SPY options, consistent with the theory.
That paper: American options exhibit positive πA values, indicating they are less vulnerable to interest‑rate stochasticity than European options.
Code
No repository by the authors found in the paper.
Extracted 2026-11-19 · same card as JSON · A language model chose these items: check the source sentence before relying on one.