Information in stock prices and some consequences: A model-free approach

TL;DR

This paper introduces a model-free method to derive risk-neutral probabilities from stock prices, clarifies their relation to market efficiency, and explores implications for option pricing, including conditions where Black-Scholes-Merton applies.

Contribution

It provides a closed-form, model-free way to determine risk-neutral probabilities and links market information to arbitrage-free pricing, extending understanding of incomplete markets and option valuation.

Findings

Risk-neutral probability ${\

}^*$ may be a mixture.

${\

Abstract

The price of a stock will rarely follow the assumed model and a curious investor or a Regulatory Authority may wish to obtain a probability model the prices support. A risk neutral probability ${\cal P}^*$ for the stock's price at time $T$ is determined in closed form from the prices before $T$ without assuming a price model. The findings indicate that ${\cal P}^*$ may be a mixture. Under mild conditions on the prices the necessary and sufficient condition to obtain ${\cal P}^*$ is the coincidence at $T$ of the stock price ranges assumed by the stock's trader and buyer. This result clarifies the relation between market's informational efficiency and the arbitrage-free option pricing methodology. It also shows that in an incomplete market there are risk neutral probabilities not supported by each stock and their use can be limited. ${\cal P}^*$-price $C$ for the stock's European call option expiring at $T$ is obtained. Among other results it is shown for "calm" prices, like the log-normal, that i) $C$ is the Black-Scholes-Merton price thus confirming its validity for various stock prices, ii) the buyer's price carries an exponentially increasing volatility premium and its difference with $C$ provides a measure of the market risk premium.