Working Paper: CEPR ID: DP5235
Authors: Andrew K. Rose
Abstract: If one ranks cities by population, the rank of a city is inversely related to its size, a well-documented phenomenon known as Zipf's Law. Further, the growth rate of a city's population is uncorrelated with its size, another well-known characteristic known as Gibrat's Law. In this paper, I show that both characteristics are true of countries as well as cities; the size distributions of cities and countries are similar. But theories that explain the size-distribution of cities do not obviously apply in explaining the size-distribution of countries. The similarity of city- and country-size distributions is an interesting riddle.
Keywords: distribution; empirical; Gibrat; growth; logarithm; mean; rank; size; Zipf
JEL Codes: F00; R12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Rank of a city (R12) | Size of a city (R12) |
| Rank of a country (O57) | Size of a country (R12) |
| Size of a city (R12) | Growth rate of a city's population (J11) |
| Size of a country (R12) | Growth rate of a country's population (J11) |
| Size distributions of cities (R12) | Size distributions of countries (D39) |