I find that a presidential executive order that took back military equipment from local police had no detectable impact on crime or officer safety. I then use 3.6 million archived inventory records to show previous studies that find favorable effects for military equipment are likely biased by unobserved attrition and weapons transfers.