Settlement Populations and the Lognormal Distribution

New Evidence on Gibrat’s Law for Cities

“…Eeckhout (2004) was able to reconcile both results by demonstrating (as Parr and Suzuki (1973) claimed in a pioneering work) that, imposing size restrictions on the cities, taking only the upper tail, this biases the analysis. Thus, if all cities are taken, it can be found that the true distribution is lognormal, and that the growth of these cities is independent of size.…”

On City Size Distribution

“…This is the reason why the magnitude of ζ is highly sensitive to the truncation point of data. Related with this, some other authors argued that the best approximation of city-size distribution is a combination of a power law distribution (for large cities) and a lognormal distribution for the smallest cities (Parr and Suzuki, 1973;Levy, 2009).…”

Accessibility of Cities in the Digital Economy

“…Because of these attributes, the power form (3) better reflects spatial disparities such as those observed as a result of large agglomeration effects (Wilson, 1967). In addition to these two forms, the log-normal form (4) is also tested here, because of its attribute as a bridge between the above two distributions (Parr & Susuki, 1973). The common characteristic of the three forms is the existence of the distance-sensitivity parameters (b 1 -b 3 ), which can be useful for observing the aggregate behaviour of Internet backbone providers (IBPs), who design and control the topology and the capacity of this digital infrastructure.…”