Significance tests of probability non-stationarity of security price returns

This dissertation explores the explanation that a primary for markets to exist is to solve a nonintegrability problem that society faces in allocating its scarce resources. Existence evidence in the form of regulation, institutions and other agents is presented. These entities have no economic rationale to exist under the assumption of current asset pricing models. Asset pricing anomalies are documented for which nonintegrability offers a unified explanation. Indirect empirical evidence for the existence of the nonintegrability problem is supplied by testing for specific models whether or not the stochastic error term follows the identical distribution over time. This dissertation employs distribution-free norm test statistics for the investigation. This paper finds strong evidence of probability non-stationarity for Brownian asset pricing models.