As of today, a considerable amount of urban land is devoted to parking, especially in downtown areas. Fully autonomous vehicles (FAVs) – also referred to as self-driving car or driverless car – can self-park in less-expensive areas or even have no need to park at all and so can free city centers from parking lots and thereby relieve downtown land for more valuable purposes such as housing, production, recreation. This paper analyzes whether the technological capability of FAVs to reposition could also be economically viable under a wide range of differentiated constellations, paying particular attention to the heterogeneity of (downtown) parking pricing schemes (hourly parking vs. daily parking vs. employer-paid parking). We develop an economic speed choice model and calculate the utility maximizing (cost minimizing) driving speed of an empty electric FAV returning home. It is shown that compared to human driven cars, individual welfare losses when deviating from the privately optimal speed are rather small, implying that congestion effects would hardly affect economic viability of ‘return home’, as opposed to feasibility. We estimate feasible ‘return home’ options of an empty FAV depending on stylized parameters such as the average duration of stay of the traveler and distance between origin (home) and destination (e.g. work). To identify economic viability, we then juxtapose the net benefit of vehicle repositioning with the cost of downtown parking. Our analysis suggests that vehicle repositioning in the form of ‘return home’ may be economically viable for a wide range of constellations, in particular when account is taken of the benefits of intra-household car-sharing made possible by vehicle relocation. Our findings highlight the potential of FAVs to revolutionize the world’s transportation system and urban structure and stress the need to carefully fine-tune transportation policies to tackle the potential adverse societal effects of induced vehicle miles traveled owing to driverless vehicle repositioning.