Combining Theory and Empirics for Causal Inference in Economics: Two Examples