Essays on model uncertainty in financial models

This dissertation studies model uncertainty, particularly in financial models. It consists of two empirical chapters and one theoretical chapter. The first empirical chapter (Chapter 2) classifies model uncertainty into parameter uncertainty and misspecification uncertainty. It investigates the impact of model uncertainty on bond pricing with affine term structure models, and finds the significant impact of misspecification uncertainty. As an extension, the second empirical chapter (Chapter 3) proposes a prediction interval incorporating misspecification uncertainty, and studies the funding ratio in a defined-benefit pension system. The last chapter (Chapter 4) proposes a time-consistent method to incorporate model uncertainty in an investment-consumption problem, and studies portfolio rules and asset pricing in theory. It shows the necessity to consider misspecification uncertainty to explain the equity risk premium puzzle. Overall, this dissertation demonstrates that model uncertainty, especially misspecification uncertainty, is non-negligible in applying and constructing financial models.