Car ownership forecasters working in the United States have followed a different course from their British counterparts. At an early stage of development of land use/transportation planning it became clear that some method of car ownership prediction, other than simple linear extrapolation, was required if this critical element was to be effectively modelled at regional and area level. In an attempt to achieve more reliable base year data and future year projections at the zonal level the study team introduced the now familiar form of category analysis. Oddly, this work contains no discussion of prediction procedures and, in spite of a heavy emphasis on the development of multiple regression equations for trip generation, the problem of stability of car ownership coefficients over time is not highlighted. Partly in response to the relative importance of the automobile industry in the United States there has been a long history of the development of econometric models.