Abstract: Over the last few decades, empirical economists have experienced what Angrist and Pischke (2010) call a “credibility revolution”. Today, empirical economists take the quality of research designs far more seriously than twenty years ago. Unfortunately, however, many policy practitioners tend to overstate the value of experimental studies that are ill-designed, overlooking the pitfalls of experimental approaches and the advances in other areas of empirical research. This article reviews the pros and cons of structural versus reduced-form approaches to causal inference in a simple conceptual framework that is easy enough for practitioners. I argue that an informed policy decision requires credible evidence not only on causal relationships but also on the economic mechanisms underlying such relationships. My view is that economic theory plays a non-marginal role in virtually all empirical studies in economics, including those that rely on field experiments, and structural and reduced-form approaches are complements in producing credible evidence toward better-informed policy making. I reinforce this view by reviewing some of the recent empirical studies in the environmental economics literature.
Key words: Causal inference, Evidence-informed policy making, Experimental/Quasi-experimental approach, Reduced-form estimation, Structural estimation