Identification and estimation problems concerning simultaneous equation models which have random parameters are considered, and some results are derived. For instance, a reducibility condition is derived under which the conditions for identification of such a system are identical to those that would be relevant if the parameters were not random. This condition is then weakened to one which relates to the identification and estimation problems of a particular equation in the model. Examples are given. Further generalizations are also considered but only conditional results are given. Further work is suggested.