reviews previous research on industrial location decisions, pointing to per-
ceived problems with CLM and Poisson models. Section 3 proposes solu-
tions to these problems in econometric location modeling. Section 4 offers
the empirical illustration, providing evidence for location factors affecting
locational choices among U.S. counties. Section 5 summarizes the paper
and points to directions for further research.
2 Previous Research on Industrial Location Deci-
sions
Most recent empirically based interregional location papers have relied on
the CLM. The virtue in this line of research is a profit maximizing model
linking the site selection decision to specific area characteristics. The proba-
bility of a new plant being opened at a particular site depends on the relative
level of profits that can be derived in this site and hence on the site’s at-
tributes compared with those of all other alternatives. As stated in the
introduction, this approach was pioneered by Carlton (1983), who modeled
the location of new branch plants across standard metropolitan statistical
areas (SMSAs) in the United States. With the exception of Hansen (1987),
Woodward (1992), and more recently Guimaraes, Figueiredo & Woodward
(2000), who also relied on narrowly defined spatial choice sets, subsequent
research has modeled location choices among highly aggregated regions, such
as U.S. states [Bartik (1985), Coughlin, Terza & Arromdee (1991), Friedman,
Gerlowski & Silberman (1992), Friedman, Fung, Gerlowski & Silberman
(1996), Head, Ries & Swenson (1995), Levinson (1996), Head, Ries & Swen-