ABSTRACT: This paper develops a nonparametric test to investigate whether any of two classes of “simple” models can rationalize observed option prices. These classes are (i) processes with independent returns and (ii) univariate Markov processes. The practice of recalibration is mimicked by only imposing minimal stability requirements on the candidate pricing models. The main finding is that processes with independent returns are inadequate even for the purpose of fitting the cross-section and term-structure on a single day. Univariate Markovian processes, however, perform much better. This is due to their ability to capture leverage.