It is shown, from analysis of typical reinforced concrete beam sections, that current design practice, which assumes beam stiffness is independent of reinforcement ratio but equal to a constant fraction of gross section stiffness is inappropriate. The analyses indicate that effective beam yield curvature can be considered constant, when non-dimensionalized by beam depth and yield strain, indicating that beam stiffness is proportional to strength. Based on this observation, a simple expression for yield drift of frames is proposed and is calibrated by comparing with results of a large number of beam/column subassemblage experiments. Good agreement is obtained. It is pointed out that current estimates of frame stiffness are generally too high. A consequence is that simple calculations show that the vast majority of frame buildings will be unable to achieve code design ductility levels before exceeding code drift limitations.