Nonparametric Asset Pricing with Conditioning Information

Sammanfattning: This study sets out to be the very first in introducing the notion of a nonlinear pricing kernel in conditional asset pricing for the Swedish equity market. By implementing a flexible nonparametric methodology, we are able to conduct tests that are completely free from functional form specifications of time-varying betas, risk premia and the stochastic discount factor. The test provides a refined view about the empirical performance of asset pricing models with conditioning information. In this paper, we investigate the most general versions of the conditional Fama and French (1993) model and the CAPM ever examined in Sweden. We demonstrate that both these models significantly outperform their static counterparts, that they price stocks very well and that nonlinearity in the first and second moment of the market return is important.