Implied Pricing Kernels: An Alternative Approach for Option Valuation

摘要

This study proposes a new estimation approach for option valuation (an implied pricing kernel-based approach), which estimates model parameters under the physical probability measure (P-measure) using a pricing kernel implied by the GARCH option pricing model. Analyzing the dataset on the KOSPI 200 options market, we examine the empirical performance of the implied pricing kernel-based approach and compare it with the performance of the classical GARCH option valuation approach (i.e., a pricing model-based approach) that estimates model parameters under the risk-neutral probability measure (Q-measure). As proposed in this study, the implied pricing kernel-based estimation approach requires approximation and discretization in order to derive the functional form of the pricing kernel. Regardless of this approximation, however, when it comes to pricing OTM calls, the implied pricing kernel-based approach performs slightly better than the pricing model-based approach that has been traditionally used for option valuation. Additional analysis based on the put option sample indicates that the implied pricing kernel-based approach clearly dominates the classical pricing model-based approach during the early stage of emergence of the KOSPI 200 options market (1999-2000) when the market was immature. During the recent global financial crisis (2007-2009), the pricing kernel-based approach also yields smaller pricing errors for OTM puts than the classical approach does. These empirical results imply that the new approach suggested in this study can be advantageous for option valuation, particularly when the information embedded in options prices is not sufficient for estimation and/or the market is speculative and volatile.