Abstract
The derivation of the gravity model from utility theory is reformulated and generalized in order to show (1) how both the number of trips to individual destinations and the total travel budget may be determined simultaneously, and (2) how the effect of both distance and destination quality on trip distribution and total budget may be analyzed. Results are compared with the revealed space preference approach and are found to be superior in dealing theoretically with trip frequency and the effects of spatial context. For empirical applications, however, the revealed preference approach is advantageous.