Central counterparties carry out stress tests using historical and hypothetical stress scenarios to assess the adequacy of their default management resources. Parametric models for the construction of hypothetical stress scenarios require expert judgment in parameter setting and substantial model assumptions. We propose a nonparametric method for the generation and/or validation of hypothetical stress scenarios using the vine copula. Our method is superior to others because it allows for the modeling of individual marginal distributions of multiple risk factors independent of joint distribution structures as well as capturing nonlinear tail dependence and fat tails. We show that the method can be extended for generating coherent stress scenarios for multiple central counterparties or clearing services, and that the generated scenarios are “extreme but plausible”. Our method is also pragmatic: use of the vine copula makes the method scalable for large numbers of risk factors, and we propose a fast screening algorithm to reduce computational requirements by quickly identifying small numbers of stress scenarios from a large number of simulations.We also demonstrate the plausibility of scenarios generated using the proposed method.