Advances in statistical inference for econometric diffusion models

Abstract

Due to their analytical tractability, continuous-time models have become a centerpiece in the financial literature. The goal of this thesis is the development of new goodness-of-fit test for continuous-time diffusion models, considering stochastic differential equations with deterministic and stochastic volatility and Itô diffusions as functional time series. Notwithstanding the importance of goodness-of-fit tools, latent factors and a continuous-time setting with observations occurring at discrete time points challenge the estimation of the models. Therefore, the estimation problem is addressed, as it hinders the goodness-of-fit procedures, discussing the intricacies of different estimation implementations prior to the methodological contribution of the test procedures.