ORIGINAL PAPER Does Gibrat’s law for cities hold when location counts?

Abstract According to Gibrat’s law, the growth rates of cities are expected to be independent of their sizes. In this study, we hypothesize that growth rates do depend on population size, but that the direction of the relationship depends on location, i.e., in unfavorable loci, growth rates are positively related to size, while the relationship is reversed in favorable loci, i.e., places with several location advantages. The present study examines that possibility, using 1990–2000 population growth data for two levels of geographic resolution—4,667 local administrative units (i.e., municipalities) and 2,189 contiguous urban areas in 40 European countries. According to our findings, when individual localities are considered, ‘proportionate ’ growth (expected under Gibrat’s law) emerges at the aggregate (system-wide) level, but ‘dissipates ’ when the settlement system is disaggregated into two urban sub-systems, formed by well-positioned localities and poorly positioned ones. Concurrently, for urban areas, a strong positive association between population size and growth emerges both before and after controlling for location attributes. However, this association between population size and growth is not especially strong, if favorably and unfavorably located urban areas are looked at separately