The Effects of Urbanization on Insurance Consumption—The experience of China

• Published date: 01 September 2016
• Author: Ping Wang,James Barrese,Ji Zhang
• DOI: http://doi.org/10.1111/rmir.12063
• Date: 01 September 2016

Risk Management and Insurance Review

C

Risk Management and Insurance Review, 2016, Vol.19, No. 2, 285-306

DOI: 10.1111/rmir.12063

THE EFFECTS OF URBANIZATION ON INSURANCE

CONSUMPTION—THE EXPERIENCE OF CHINA

James Barrese

Ping Wang

Ji Zhang

ABSTRACT

Cross-country studies of insurance find little support for a hypothesized

urbanization–insurance consumption link. Unlike the literature, we argue and

find that the relationship is positive. Our study refines the literature relating

economic development and financial service activities. The empirical evidence

in the existing literature is based on cross-country samples that are criticized

as unrepresentative because they are disproportionately composed of highly

developed economies and consequently have limited range in the variables

considered; this limits confidence in the resulting evidence. The current study

overcomes these empirical difficulties by focusing on one large country with a

unique recent development pattern. Like the existing studies, we find a positive

insurance–economic development link. In contrast to cross-country studies, we

find that insurance consumption is influenced by urbanization and that the

country’s World Trade Organization entry had a differential impact on its in-

surance markets.

INTRODUCTION

The role of urbanization in comparative insurance consumptions studies is discussed

in the insurance-economic development literature. The set of hypotheses relating insur-

ance activity to economic development typically regress insurance activity on a set of

variables that include a measure of economic activity for which a positive coefficient

is interpreted as evidence of an insurance-economic development link. The general

hypotheses considered in the insurance-economic development literature include:

Insurance Activity =fIncome,Financial Development,LossProbability/

Awareness,CulturalVariation].

James Barrese, School of Risk Management, Insurance, and Actuarial Science, St. John’s Univer-

sity, 101 Astor Place, New York, NY 10003; e-mail: barresej@stjohns.edu. Ping Wang, School of

Risk Management, Insurance, and Actuarial Science, St. John’s University.101 Astor Place, New

York,NY 10003; e-mail: wangp1@stjohns.edu; Ji Zhang, School of Insurance Economics, Univer-

sity of International Business and Economics, Beijing, China; e-mail: zhangji1030@126.com.

285

286 RISK MANAGEMENT AND INSURANCE REVIEW

In the history of the literature on this topic, each of the concepts described has been

measured in a variety of ways. For example, economic wherewithal—incomeor wealth—

has been variously measured using GDP per capita, national income, and a UN index

of national wealth. With the exception of the UN wealth index, the relationship is

consistently found to be positive. Financial development—the notion that the more

sophisticated the population with respect to financial tools, the more likely insurance

will be one of those tools—is considered in the literature using ratios of monetary supply

measures. The estimated relationships are not typically significant. Loss probability and

loss or risk awareness is based on the hypotheses that humans are generally risk averse,

so areas with a higher probability of loss would demand higher levels of insurance;

the related requirement is that those with loss exposures be aware of their exposure in

order to act. The many measures of loss likelihood in an area include life expectancy (for

studies of life insurance activity) and the urbanization percentage (for studies of property

insurance activity). A proxy for risk awareness is the level of tertiary education.

Although the existing evidence of a loss probability–insurance demand link is significant

for life insurance, where life expectancy is the loss probability proxy, the results are not

significant for the property–liability insurance industry,where the urbanization percent-

age is used as a loss probability proxy (Browne et al., 2000). However, the literature on

which these studies are based have a variation in urbanization that is small compared

to the variation of China in recent years. China’s urban population percentage grew

from a 1990 urbanization rate of 26.2 percent to 49.9 percent in 2010. By comparison,

the Organisation for Economic Co-operation and Development (OECD) average, a set

of countries that represents the focus of many of the studies in the insurance–economic

development literature, rose from 73 percent in 1990 to 79 percent in 2010. Given the

empirical reality that higher rates of volatility give more reliable relationship estimates,

we expect the China data to provide a better sense of the impact of urbanization policies

than is found in the existing literature. We also present a more complete description of

why one would expect strong links between urbanization and insurance demand.

In the discussion of the urbanization–insurance demand link above, it is argued that a

merit of this article is the wider range of values for the variables in the China sample

than in, for example, studies considering only OECD countries. The range of values of

China is considerably larger than would be true for a study across the U.S. states. For

instance, the per capita GDP in the richest U.S. state is twice that of the poorest state; the

ratio for China’s provinces is five-to-one in 2011. Because urbanization is a focus of this

study, the following section provides a summary of the recent urbanization history of

China. In addition to providing information pertinent to the current study, we hope this

broadens the research familiarity with China. Much of the insurance literature makes

use of a comparison across the U.S. states; this article mimics that approach but relies

on data from Chinese provinces rather than U.S. states. Millo and Carmeci (2015) take a

similar approach by examining variations in life insurance market across areas of Italy.

Although, for reasons described in the next section, the period starting in 1980 would be

interesting, data availability constraints restrict the empirical focus of this study to the

period from 1997 to 2011.

The remaining sections of the article are organized as follows. After summarizing the

evolution of China’s urbanization and insurance market since 1980, a section on the

factors affecting insurance consumption reviews the variables of interest in the literature.