The fundamental criterion for listing or delisting species under the United States Endangered Species Act (ESA) is acceptable extinction risk. But what is the meaning of “acceptable risk”? This concept has received little attention in conservation biology, and hence, the purpose of this paper is to explain acceptable risk within the context of the ESA. All complex environmental policy decisions involve trade-offs, and hence, choosing one policy from several policy options entails a comparison of the benefits, costs, and risks of all options. Acceptable-risk problems are decision problems requiring a choice among different policy options, and an acceptable risk is defined as the risk associated with the chosen policy. In general, people will accept the risk imposed by a particular policy when that policy yields sufficient net benefits. The ESA, its legislative history, and implementing regulations clearly state that species listing determinations must be based solely on science and “without reference to possible economic or other impacts” caused by the listing of that species. If economic considerations are prohibited from influencing listing decisions, then risk reduction is not constrained by cost, and therefore, the only rational choice for acceptable extinction risk is zero risk. This outcome, known as a “stopping-point problem”, is absurd, and consequently, federal agencies responsible for listing determinations are forced to issue opaque decisions that obscure their true rationale, which in turn, lead to inconsistency and improper political influence. Illustrations of acceptable risk in null hypothesis tests, the United States Clean Air Act, and minimum viable population size are also presented.