Comparing City Size Distributions: Population vs. Economic Activity

This paper compares the size distribution of cities when they are measured in both demographic and economic terms. In doing so, we have exploited more recent and accurate nighttime lights data than those previously used to proxy urban economic activity. Our results for a sample covering 12,852 urban centres in 100 countries show that economic activity is more unevenly distributed than population. There is weak evidence for Zipf's law in aggregate urban night lights, especially in medium and low income countries. Our findings do not support a Pareto distribution for city sizes measured in economic terms.