Typically papers like this find some instrument. The fact they don't have one might explain why they went for PNAS and not an economics journal, which are usual

Typically papers like this find some instrument. The fact they don't have one might explain why they went for PNAS and not an economics journal, which are usually more rigorous (or more anal?) That said, their results on trust fit their mechanism pretty well. The big possible confounder might be that the inquisition correlates with Muslim ("morisco") or Jewish presence. There's a big literature arguing that ethnic heterogeneity weakens trust, and that might persist over time even when the ethnic heterogeneity has vanished. But: "adding a control for pre-expulsion morisco population has little impact on the results."