PREVIOUS EMPIRICAL WORK
Empirical studies of urban wage variation have focused on estimating the
wage ‘‘gradient’’ hypothesized by Moses. Early studies, such as Segal 22 and
DiMasi and Peddle 3 , used average wage levels for certain occupational
categories as a means of testing for differences between central city and
suburban wages. Using aggregate wage data can lead to bias if other worker
characteristics vary systematically across different locations, which they most
surely should. Eberts 5 sought to overcome this problem by focusing only on
municipal workers, and further, only those within relatively homogeneous
occupations, such as firefighters. While an improvement, the problem of using
aggregate wage data remained. Later studies, such as Ihlanfeldt 9 , Madden
14 , and McMillen and Singell 12 , avoided the problem by using microdata
on individual workers. The advantage of this approach is that a range of human
capital factors which determine wage levels, and which may be correlated with
location, such as experience, education, occupation, and industry, can be
controlled for. This allows for a more precise isolation of the portion of the
wage which is due to work place location. The first contribution of this paper
then is to use extensive microdata to study intra-urban wage variation.