Bayesian analysis of nonlinear regression with equicorrelated elliptical errors

For a nonlinear regression model with an intercept, it is shown that, under diffuse priors of the intercept and of the error precision, the assumption of equicorrelated and jointly elliptical errors (instead of uncorrelated Normal ones) has no new consequences on posterior inference about the (nonlinear) regression parameters (except for the constant term) and on predictive inference.