In their reply to my comment (Walker 2008) Grinols and Mustard explained that in their original study (2006) their interest was “in the costs to the host county associated with a change in crime from whatever source” (Grinols and Mustard 2008, p. 22). In this rejoinder, I explain that the estimated costs of crime attributable to casinos will be overstated if the estimated crime effects are based on the “undiluted” crime rate used by Grinols and Mustard (2006). I also discuss why this issue is important, in the context of “social cost of casino gambling” estimates that are frequently quoted in political debate and by the media.