This article presents an econometric analysis of parameter estimation for continuous-time affine term structure models driven by latent Markovian factors. In this setting either methodology, frequentist or Bayesian, is confronted with two major problems: First, each parameter set implies a time series of latent factors the transition densities of which determine the likelihood of the parameters themselves. Thus, an estimation procedure has to be capable of dealing with data that changes for each likelihood evaluation. Second, in contrast to the continuous-time model formulation, data are available only in discrete time and formulae for transition densities are known only for a very small subset of the affine term structure family.
展开▼
