Urban Systems: Understanding and Predicting the Spatial Distribution of China's Population

• Published date: 01 July 2021
• Author: Pengfei Li,Ming Lu
• Date: 01 July 2021
• DOI: http://doi.org/10.1111/cwe.12380

©2021 Institute of World Economics and Politics, Chinese Academy of Social Sciences

China & World Economy / 35–62, Vol. 29, No. 4, 2021 35

*Pengfei Li, Post-doctoral Fellow, Antai College of Economics and Management, Shanghai Jiao Tong University,

China. Email: lipfchn@163.com; Ming Lu (corresponding author), Professor, Antai College of Economics and

Management; China Institute for Urban Governance, Shanghai Jiao Tong University, China. Email: luming1973@

sjtu.edu.cn. The authors acknowledge research support from the National Natural Science Foundation of China

(Nos. 72073094 and 71834005), Shanghai Institute for National Economy, Shanghai Institute of International Finance

and Economics, and the China Merchants Charitable Foundation.

Urban Systems: Understanding and Predicting the

Spatial Distribution of China’s Population

Pengfei Li, Ming Lu*

Abstract

W ith urbanization and population migration, some Chinese cities fall into decline

whereas others prosper. Using nighttime light data, we redefine the city based on

economic function and evaluate the city size distribution in representative countries.

The results provide evidence not only for Zipf’s law, but also for a distortion in

China’s current city size distribution. This study proposes a feasible method to predict

urban population distribution based on the role of geographical factors in regional

development, following the idea of spatial equilibrium. This prediction suggests that

the divergence of city size in China tends to be pronounced, with inter-regional income

disparity being narrowed and the city size distribution following Zipf’s law. The Chinese

government should further relax restrictions on population infl ow into large cities and

prepare for more migration in the future.

Key words: geographical factors, population distribution, urban system, Zipf’s law

JEL codes: O21, R12, R23

I. Introduction

Predicting future city size is difficult. In practice, incorrect predictions may cause

two directions of problems. One is the insuffi cient supply of infrastructure and public

services, restricting city development. The other is excessive public investment, which

places a heavy burden on local public finance. In China, the former problem occurs

in large cities with continuous population infl ow, and the latter happens in small- and

medium-sized cities with population outfl ow, especially in inland provinces.

The reasons for the above phenomena are the following. First, there is no simple

Pengfei Li, Ming Lu / 35–62, Vol. 29, No. 4, 2021

©2021 Institute of World Economics and Politics, Chinese Academy of Social Sciences

36

and feasible method of predicting population distribution in economics and geography.

Future urban development is difficult to predict, especially in countries experiencing

rapid urbanization. Second, the potential population growth in each city is mainly

determined by the overall population of a country. However, in practice, city size is

usually planned by local government, which ignores inter-relationships across cities.

Third, the central government often ignores the agglomeration economies of large cities,

and sometimes prefers an even population distribution. This preference makes it diffi cult

for the government to make urban development policies in line with migration trends.

Fourth, local governments in cities with population outfl ow have a strong motivation

to attract migrants, whereas governments in cities with population infl ow worry about

urban problems, such as congestion, which are seemingly related to rapid population

growth. In China, urban planning is largely affected by the heritage of a planned

economy with strong government intervention, resulting in serious practical problems,

such as overbuilding infrastructure in cities with population outfl o w.

Most existing literature concerning city size distribution provides empirical

evidence for whether Zipf’s law holds and focuses more on administrative cities

(Anderson and Ge, 2005; Kausik and Basu, 2009). However, in China’s statistics, there

are mainly two levels of administrative cities: four provincial-level municipalities

directly under the central government, and 293 prefectural-level cities. A municipality

or a prefectural-level city usually consists of an urban district (shiqu) and several

conuties or county-level cities. There are 394 county-level cities, which are relatively

economically independent from the administrative cities that govern the counties, so the

administrative cities could not refl ect the economically integraded cities. In this paper,

a city is redefi ned based on the economic function of city according to nighttime light

data, and the city size distribution is discussed accordingly. A “nighttime light city”

(NLC) refers to connected urban areas measured by nighttime light. The boundary of

the NLC is determined by economic activities and does not coincide with administrative

boundaries. Some large NLCs may refer to metropolitan areas that cover a large

municipality/prefecture and the surrounding small- and medium-sized counties. Most

of the NLCs correspond to economically independent counties, which are disconnected

from the urban districts of the prefectures. Based on this concept, we examine city size

distribution in representative countries (including Japan, South Korea, India, Indonesia,

and China) and find a distortion in China’s city size distribution. Compared with the

existing literature, this paper also provides a simple and feasible method to predict

future population distribution. First, we establish an econometric model to study the

effects of geographical factors on urban economic growth, and then use it to predict the

future GDP growth r ate of each city and the spatial distribution of economic activities