We propose to model endogeneity bias using prior distributions of moment conditions. The estimator can be obtained both as a method-of-moments estimator and in a Ridge penalized regression framework. We show the estimator’s relation to a Bayesian estimator.