assessment was later tempered by arguments emphasizing the benefits of spatial ag-
glomerations, which help consumers by economizing on shopping costs in terms of
shopping time, transportation cost (Eaton and Lipsey, 1976, 1979) and uncertainty
reduction (Wolinsky,1983; Dudey, 1990). Following numerous papers on efficiency
gains from agglomeration, however, new negative assessments appeared, for exam-
ple Dudey (1993) on welfare-decreasing agglomerations. More recent theoretical and
empirical work emphasizes the role of physical distance in the production function—
indivisibility in production (Kanemoto, 1990), labor market pooling (Rosenthal and
Strange, 2001), and complementarity between workers and firms (Andersson, Burgess
and Lane, 2007)—to better characterize efficiency of agglomerations. Rather than try-
ing to harmonize contradictory normative theories (Fischer and Harrington, 1996),
this paper attempts to exploit the mixed normative message in the agglomeration lit-
erature. This follows Gigerenzer et al (1999) and Gigerenzer and Selten (2002), whose
normative approach—referred to as ecological rationality—seeks to analyze when de-
cision procedures are well matched, or badly matched, to decision environments. The
matching concept underlying ecological rationality stands in contrast to universal,
context-free and content-blind normative criteria for evaluating rationality, such as
transitivity, the Kolmogorov axioms of probability theory, or the Savage axioms of
expected utility theory.1
1 Context-dependent normative analysis does not imply relativity, as there remain many compelling reasons other
than violations of consistency axioms for policy makers to be concerned about behavioral underpinnings of spatial
agglomerations. Berry (1961), for example, argues that steepness of city-size distribution curves is inversely related
to economic development and, therefore, that policy makers interested in economic development have good reason
to be concerned with different forms of agglomeration as a primary issue in planning. Muiz, and Galindo (2005)
present evidence on suburban agglomerations and environmental impacts. And Anas and Rhee (2007) demonstrate
the sensitivity of normative evaluations of policies that concern spatial agglomerations to apparently innocuous as-
sumptions such as exogenous agricultural land rents in areas surrounding cities. Similarly, Turner (2007) shows that
small coordination and free-rider microstructures lead to a large divergence between equilibrium and socially efficient
spatial distributions.