Under minimum assumptions on the stochastic regressors, strong consist
ency of Bayes estimates is established in stochastic regression models
in two cases: (1) When the prior distribution is discrete, the p.d.f.
f of i.i.d. random errors is assumed to have finite Fisher informatio
n I=integral(-infinity)(infinity) (f')(2)/f dx < infinity; (2) for gen
eral priors, we assume f is strongly unimodal. The result can be consi
dered as an application of a theorem of Doob to stochastic regression
models. (C) 1996 Academic Press. Inc.