Inferring Financial Bubbles from Option Data

Abstract

Financial bubbles arise when the underlying assets market price departs from its fundamental value. Unlike other bubble tests that use time series data and assume a reduced-form price process, we infer the existence of bubbles nonparametrically using option price data. Under no-arbitrage and acknowledging data constraints, we can partially identify asset price bubbles using a cross section of European option prices. In the empirical analysis, we obtain interval estimates of price bubbles embedded in the S&P 500 Index. The estimated index bubbles are then used to construct profitable momentum trading strategies that consistently outperform a buy-and-hold trading strategy.

Keywords: asset price bubble, fundamental value, risk-neutral probability measure, state price distribution, tail truncation, partial identification, nonparametric estimation, local polynomial

JEL Classification: C14, C58, G12