causal

Short-Run Data Too Noisy For Development Verdicts

Short-run development data (e.g., national child mortality estimates for Ethiopia) are notoriously noisy and unreliable—different sources can imply anything from a huge decline to an increase—so using year-by-year improvements to render verdicts or credit a dictator is a sloppy, non-empirical error; reliance on long-run trends is far more reliable because unbiased measurement error averages out over time.

causalpending

Speaker

Guest

Evidence Quote

“the short run data are just too unreliable. What those data are useful for are really to give you much more of a long run verdict.”

Source

William Easterly on the Tyranny of Experts— EconTalk
Created: 6/13/2026, 12:08:27 AM

Your notes

Loading notes...