Are You Estimating the Right Thing? An Editor Reflects

A key aspect of applied economics over the last decade has been the status accorded to research designs that may credibly identify causal effects. Without such designs, one may not estimate the right thing and publishing options may be limited. Yet situations where reported results are not for parameters that researchers think they have identified are potentially more widespread than what is usually considered in discussion of research designs or identification strategies. I give five examples of applied practice where what is estimated may not identify what is wanted. Each practice came up in submissions I handled as co-editor of an agricultural economics journal, and these issues seem common in applied work. A narrow focus on identification strategies may lead to the neglect of these examples, and of other productive ways to improve the credibility and usefulness of applied economics research.