The study uses Theil's concept of certainty equivalence and bias to analyse the production decisions of the purely competitive firm when some of its inputs are controlled and others are non-controlled and subject to random influences. Conditions are outlined under which a surrogate procedure based upon the expected values of the non-controlled variables leads to certainty equivalence and others under which it leads to bias in the controlled variables. General factors which determine the direction of the bias are discussed. The managerial optimality of the surrogate procedure is investigated and also its welfare implications. The analysis is of particular relevance to agriculture because of the prevalence there of random noncontrolled inputs in production.