Structural empirical work either results in models with stochastic error terms (i.e. models which don’t fit), or places weak deterministic bounds on the objects of interest (i.e. bounds which are “sharp” in the sense of exhausting the empirical content of the economic theory, but which are close-to-useless for any practical purpose). This talk looks at an alternative: statistical interpolation based on random fields. The talk will cover the general idea and look at an application to individual labour supply