When and how should we encourage network providers to mitigate the harm of security and privacy risks? Poorly designed interventions that do not align with economic incentives can lead stakeholders to be less, rather than more, careful. We apply an economic framework that compares two fundamental regulatory approaches: risk based or ex ante and harm based or ex post. We posit that for well known security risks, such as botnets, ex ante sanctions are economically efficient. Systematic best practices, e.g. patching, can reduce the risk of becoming a bot and thus can be implemented ex ante. Conversely risks, which are contextual, poorly understood, and new, and where distribution of harm is difficult to estimate, should incur ex post sanctions, e.g. information disclosure. Privacy preferences and potential harm vary widely across domains; thus, post-hoc consideration of harm is more appropriate for privacy risks. We examine two current policy and enforcement efforts, i.e. Do Not Track and botnet takedowns, under the ex ante vs. ex post framework. We argue that these efforts may worsen security and privacy outcomes, as they distort market forces, reduce competition, or create artificial monopolies. Finally, we address the overlap between security and privacy risks.