Limiting experiments for panel-data and jump-diffusion models

This work concerns the theory of limiting experiments and its use in econometrics. In Chapter 2, we consider jump-diffusion models and we compare, by means of the limiting experiment, the statistical information contained in continuous-time observations with that contained in discrete-time observations sampled in high frequency. In Chapter 3, we establish the Local Asymptotic Quadratic condition for bivariate hidden Ornstein-Uhlenbeck models using continuous-time observations. We assume that the hidden process is highly persistent and, using the limiting experiment, we discuss some inference procedures. Chapter 4 provides the power envelope for tests of the unit root hypothesis in Gaussian panel data models with cross-sectional dependence. And, it proposes a test statistic which attains the power envelope.