ŽMuth 18 and Mills 16 . More recently, authors have noted that if firms are .
located at different points within a metropolitan area, then transportation costs
must be capitalized into wages as well. As Moses 17 pointed out, if ‘‘local’’
firms employing adjacent workers coexist with firms in a Central Business
District, then the wages such local firms pay must decline with distance just as
rents do. With differences in both wages and rents it is possible for workers to
achieve equal utility across residence zonesregardless of whether they work
‘‘locally’’ or ‘‘centrally.’’ In the longer run, whether firms tolerate these wage
differences depends on whether their production function generates cost or
productivity differences Wheaton and Sivitanidou 27 .Ž .
This paper uses a linear programming model to demonstrate more generally
how commuting costs will be capitalized into both labor market wages and land
market rents. Specifically, variation in commuting costs of individual workers
employed at the same location but living at different locations will be capitalized into land rents. Variation in the average commuting costs between those
employed at different work locations will be capitalized into wages. This paper
seeks to empirically validate the conclusions of this model by focusing on three
questions. First, do wages vary within metropolitan areas by workplace location? Second, is this variation correlated with commuting times? Third, do
larger employment centers have longer average commutes and hence higherŽ
wages ? The paper leaves open the question as to whether larger employment. Ž .
locations with longer commuteshigher wages have offsetting agglomeration
economies, or whether such differences are simply temporary and eventually
will be equilibrated away through firm mobility