Abstract:
The paper derives and proves results of Gaussian maximum likelihood estimation of constant unknowns (coefficients, covariances) and time-varying unknowns (factors, disturbances) of static and dynamic factor models and, thereby, extends the statistics and econometrics literatures on estimation and statistical evaluation of estimates of the unknowns. The paper presents a new, general, unified, and one-step-comprehensive method for simultaneously estimating and statistically evaluating all constant and time-varying unknowns of static and dynamic factor models.
