The model of compensating differentials in regional labor markets was developed by Roback (1982). The model interprets regional differences in constant quality wages and rents as compensating firms and residents for inter-regional differences in amenities. The model assumes that the costs of relocating to a new city are zero. The results hold in the presence of moving costs for the marginal migrant. This paper extends the Roback model to allow for moving costs which vary among a city’s residents and businesses. This modification of the model generates new interpretations of regional differences in rents and (to a lesser extent) wages. The importance of amenities is retained, but housing supply becomes the main other determinant of regional rents. Housing supply was for the most part ignored in the literature following on Roback’s initial insight. The new perspective also provides a bridge between the neo-classical perspective implicit in Roback’s approach and the newer literature on agglomeration economies.