This paper discusses Bayesian inference in change-point models. Current approaches place a possibly hierarchical prior over a known number of change points. We show how two popular priors have some potentially undesirable properties, such as allocating excessive prior weight to change points near the end of the sample. We discuss how these properties relate to imposing a fixed number of change points in the sample. In our study, we develop a hierarchical approach that allows some change points to occur out of the sample. We show that this prior has desirable properties and handles cases with unknown change points. Our hierarchical approach can be shown to nest a wide variety of change-point models, from time-varying parameter models to those with few or no breaks. Data-based learning about the parameter that controls this variety occurs because our prior is hierarchical.