Econometrics Editing and Proofreading Services

The coefficient is significant, the standard errors are clustered, the specification includes fixed effects and controls, and there are eight robustness tables. None of that establishes that the estimate is causal. That rests entirely on one argument in prose about why the variation being used is unrelated to everything else that moves the outcome — and that argument is the paper.

We edit what econometricians and applied economists write — identification strategy sections, instrumental variables and exclusion restriction arguments, difference-in-differences and event study designs, regression discontinuity papers, panel and fixed effects specifications, robustness and sensitivity sections, structural estimation documentation, referee responses, and job market papers and theses. Our editors work on the identification argument and the threats to it.

The identification section is where an econometrics paper earns the word "effect", and its failure is a technical description standing in place of an argument. Clustering and fixed effects are choices, not identification; identification is a claim about where the variation comes from. We work through these so the source of variation is named in one sentence a reader can evaluate, since a paper that cannot say what is generating the comparison has not identified anything; so the assumption is stated in its own terms rather than as a formula, because an exclusion restriction written only as an orthogonality condition hides the substantive claim it makes about the world; so each threat is named and addressed with evidence, given that the honest list of threats is short, specific and known to every referee in the field; so pre-trends, placebo periods and balance are shown rather than asserted, as a parallel trends assumption is testable in the pre-period and readers will look; so the estimated parameter is defined as what it actually is — a local, intention-to-treat or treatment-on-the-treated effect — since these answer different questions; so the population the estimate applies to is stated; and so what would falsify the design is on the page. Papers written this way get read past the abstract.

Everything you send is treated in confidence, including working papers, restricted-access data descriptions and referee correspondence. We are editors rather than econometricians, and we offer no view on identification, specification or results. What we can do is make the identification argument legible.