An approach to estimation and inference in econometrics in which the uncertainty about the value of an unknown parameter is expressed in terms of a probability distribution. This is in contrast with the classical approach in which parameters are fixed in repeated samples. In the Bayesian framework data is treated as a fixed set of additional information used in updating the prior beliefs of the analyst about the distribution of the parameters. The revised beliefs, or the posterior distribution of the parameters, given the current data, are obtained from the prior distribution using Bayes theorem. Therefore, the prior distribution of the parameters and the likelihood function of the data must be fully specified for estimation to proceed.