The costs of demolishing a vacant building are often justified on the grounds of crime reduction. I explore this claim by estimating the spatial and temporal effects of demolitions on reported crime in the city of Saginaw, Michigan. To do so, I estimate a model that uses within‐block group variation to compare crime after a demolition occurs to before the permit for that demolition was issued. Results indicate that demolitions reduce crime by about 8 percent on the block group in question and 5 percent on nearby block groups, with the largest impact concentrated one to two months after the demolition occurs.