The elastic field energy change associated with a faulting process is indeterminate, because work may be done on distant bounding surfaces. Savage (1969) has shown that plastic strain, the source of the earth's self-stress, must be considered in the resolution of this question (Steketee's paradox). In a self-stressed earth, faulting will decrease the elastic field energy. Over a period of time following the faulting, distant viscoelastic regions can be expected to respond in such a way as to return stress to a value appropriate to steady flow. In the course of this response, work will be performed on the brittle seismogenic region to increase elastic field energy smoothly throughout the region. Steketee's result for a region bounded by a surface of constant stress is appropriate to the question of tectonic reloading, if account is taken of irregular slip due to past earthquakes. The energy increase due to tectonic reloading is not necessarily equal to energy released in the earthquake and has a different spectral decomposition. Although there is no energy balance in the seismogenic region for a single earthquake, there must be such a balance for a cycle of earthquakes of all sizes.