Robust Volatility Index Calculation with OTM Option-implied Probability

TL;DR

This paper introduces a new method for robustly estimating volatility indices from OTM options by constructing arbitrage-free, continuous option pricing functions that align with observed bid-ask spreads, addressing real-market data gaps.

Contribution

It proposes a novel, less parameter-dependent approach to create arbitrage-free option pricing functions from discrete market data, improving volatility estimation robustness.

Findings

Method effectively handles low-liquidity markets.

Ensures arbitrage-free, monotonic, and convex option pricing functions.

Provides more reliable volatility indices from real-world data.

Abstract

In financial markets, accurately measuring the risk of future fluctuations in asset prices is of paramount importance. Studies such as Carr and Madan have shown that the expected value of the quadratic variation of log prices can be expressed as an integral of European option prices over a continuum of strikes. This has led to the widespread estimation of model-free volatility (implied variance). However, this theoretical calculation assumes that options are continuously traded across all strike prices, which creates a fundamental gap with real-world market environments where options are only traded at discrete strikes. How to appropriately address this gap and robustly estimate volatility is a crucial issue for both practitioners and academics, and is the primary objective of this paper. Focusing on the fact that volatility indices are primarily calculated from the prices of out-of-the-money (OTM) options, this paper proposes a novel method for constructing a continuous European option pricing function that is consistent with the bid-ask spreads of observed OTM options and strictly satisfies arbitrage-free conditions (such as monotonicity and convexity). Although previous studies have attempted to construct arbitrage-free option pricing functions from bid-ask spreads, the construction method proposed in this paper requires fewer market parameters than existing methods. This makes it possible to robustly calculate volatility indices while maintaining theoretical consistency, even in markets with extremely low liquidity.