The peer effect in adverse selection: Evidence from the micro health insurance market in Pakistan

• Published date: 01 December 2023
• Author: Xia Du,Wei Zheng,Yi Yao
• Date: 01 December 2023
• DOI: http://doi.org/10.1111/jori.12447

Received: 16 July 2022

|

Revised: 10 August 2023

|

Accepted: 14 August 2023

DOI: 10.1111/jori.12447

ORIGINAL ARTICLE

The peer effect in adverse selection:

Evidence from the micro health insurance

market in Pakistan

Xia Du

1

|Wei Zheng

2,3

|Yi Yao

2,3,4

1

Risk Management and Insurance

Department, School of Finance, Nankai

University, Tianjin, China

2

Risk Management and Insurance

Department, School of Economics,

Peking University, Beijing, China

3

China Center for Insurance and Social

Security Research, Peking University,

Beijing, China

4

Institute for Global Health and

Development, Peking University, Beijing,

China

Correspondence

Yi Yao, Risk Management and Insurance

Department, School of Economics,

Peking University, Beijing, China.

Email: yao.yi@pku.edu.cn

Funding information

The Ministry of Education of the PRC,

Grant/Award Number: 14JZD027; Young

Faculty Research Seed Fund of School of

Economics, PKU; Bill and Melinda Gates

Foundation, Grant/Award Number:

2022YFAG1004; National Natural

Science Foundation of China,

Grant/Award Number: 72261147759

Abstract

The peer effect may amplify adverse selection in social

networks, hampering the sustainable operation of micro-

insurance.Thispaperusesdatafromamicrohealth

insurance program in Pakistan to test for the peer effect in

renewal decisions and the role it plays in amplifying

adverse selection within social networks. The paper finds

evidence supporting that insurance renewal decisions are

similar among peers in the same network, and the peer

effect is stronger among households of the same risk type

than households of different risk types, indicating that the

heterogeneous peer effect acts as an amplifier for adverse

selection. The paper provides policy implications for

effective ways to mitigate the peer effect and adverse

selection, based on the results of heterogeneity analyses.

The policy recommendation is to enforce a minimum

group enrollment rate requirement of at least 60% for large

groups to mitigate the peer effect.

KEYWORDS

adverse selection, peer effect, social networks

1|INTRODUCTION

Micro health insurance programs aim to serve low‐income people in developing countries by

offering coverage and services tailored to their needs. At the same time, the microinsurance market

strives to build business models that enable a sustainable segment for commercial or cooperative

Journal of Risk and Insurance. 2023;90:1063–1100. wileyonlinelibrary.com/journal/JORI

|

1063

© 2023 American Risk and Insurance Association.

insurers (Churchill, 2006). Different from the customers in the traditional insurance market, the

low‐income population is usually poorly educated and has little experience with insurance.

Adapting to the target population, micro health insurers often design simple policies with a flat

premium, universal coverage, and simplified underwriting and claim procedures. Due to concerns

about holding down costs and improving accessibility, there is no individual risk screening or

restriction on the pre‐existing conditions of the insured. These features lead to adverse selection in

the microinsurance market, which means that high‐risk individuals tend to self‐select for more

generous coverage (Einav et al., 2010; Rothschild & Stiglitz, 1976). Since micro health insurance in

Pakistan, which we will investigate in this paper, provides only one contract with pooled

premiums to people with different risk types, people cannot choose between different levels of

coverage. In this paper, adverse selection is interpreted as low‐risk individuals opting out of the

insurance market because the pooling contract is less attractive to them (Zweifel et al., 2021),

while high‐risk individuals are more likely to stay in this market, which could be demonstrated by

a positive correlation between enrollment (or renewal) and individual health risks.

Moreover, there are fewer options for dealing with adverse selection in the microinsurance

market (Brau et al., 2011). Micro health insurance must minimize administrative costs to make

the premium affordable for the target population while maintaining sustainable operations,

which limits the effectiveness of traditional methods to mitigate adverse selection and impairs

the sustainable operation of microinsurance (Brau et al., 2011; Yao et al., 2017).

Researchers have provided extensive evidence of adverse selection in the micro health insurance

market (Ito & Kono, 2010;Sheth,2021;Yaoetal.,2017; Zhang et al., 2021). Yet, most of the existing

literature focuses on the individual's risk characteristics and insurance decisions, without

considering the role played by the social networks that the insureds share. Specifically relevant to

our research, Yao et al. (2017) show the existence of adverse selection by establishing a positive

correlation between households' risk types and their renewal probabilities. Their underlying

assumption is that a household would make an independent renewal decision, without considering

theroleofthepeereffect.WebuildonYaoetal.(2017) by providing insights into how the peer effect

impacts peer households' insurance decision‐making, using the same data set from Pakistan.

Although the aim of microinsurance is to protect low‐income people, the organizations,

whether for profitor not for profit, must build a sustainable business model. Thus, understanding

and mitigating adverse selection with cost‐effective techniques are key for the long‐term success

of microinsurance.

A social network is composed of multiple social actors and relationships among interacting

units (Wasserman & Faust, 1994). The existing literature provides abundant evidence showing

similarities in behaviors among peers in the same social network, which is also known as the

peer effect (Bertrand et al., 2000; Bramoullé et al., 2020; Manski, 1993). It is well recognized

that the peer effect plays an important role in changing the individual's risk behavior and/or

insurance decision (Giné et al., 2013; Karlan et al., 2014). This is because micro health

insurance programs are often provided to people living in remote areas where access to formal

or online information is restricted, and it is usually difficult for the target population to

understand this information based on their education and experience (Cai et al., 2015; Fischer

et al., 2023). Thus, we expect a similar pattern of renewal decisions among peers in the same

social network.

However, the peer effect in renewal decisions could also play an important role in

amplifying adverse selection when we take an individual's risk profile heterogeneity into

1064

|

DU ET AL.

consideration. On the one hand, interpersonal information sharing helps people understand

the concept and value of insurance by learning from peers in the same network, and thus

improves their ability to make strategic renewal decisions by taking advantage of asymmetric

information. On the other hand, the heterogeneous peer effect may enable individuals to mimic

the strategic behaviors of their peers to maximize insurance benefits according to their risk

type, even if they do not have any additional information about insurance. In this scenario,

high‐risk people may be more likely to follow their high‐risk peers to renew than to follow their

low‐risk peers to withdraw. In other words, there is a larger correlation between risk and

insurance demand and thus a greater scope of market failure if the peer effect in renewal

decisions is heterogeneous over individual's risk.

In this paper, we explore whether the peer effect amplifies adverse selection via social

networks in the context of micro health insurance using data from a program implemented in

the Northern Areas of Pakistan. We construct a linear‐in‐means model with instrumental

variables (IVs) to identify the heterogeneous peer effect in renewal decisions, which is

demonstrated by a stronger pattern of similarities in the insurance renewal decisions among

households of the same risk type in the same social network.

Our results support the existence of the peer effect in micro health insurance. We find

evidence supporting that insurance renewal decisions are similar among peers in the same

network, and the peer effect is stronger among households of the same risk type than

households of different risk types, indicating that adverse selection can be amplified by the

heterogeneous peer effect. Peer households have roughly 8% more impact on the renewal

decisions of households with the same risk type as themselves than on the renewal decisions of

those with different risk types. We also find that the peer effect is significantly smaller in larger

groups with higher group enrollment rates. We further provide policy implications on effective

ways to mitigate adverse selection in a market where the peer effect exists.

Our work contributes to the literature in the following aspects. First, we study the role of

the peer effect in a household's renewal decision in a group health insurance policy in the

microinsurance market. Second, we are among the first to explore how the peer effect changes

the dynamics of adverse selection. Third, we discuss the heterogeneity of the peer effect by the

characteristics of the social network and provide policy implications for ways in which adverse

selection in microinsurance can be effectively and efficiently mitigated.

The paper is organized as follows. Section 2summarizes the related literature on adverse

selection in micro health insurance and reviews the literature on the peer effect in insurance

demand. Section 3provides a brief introduction to the program and the data. Section 4uses a

simple two‐type model to demonstrate the amplification of adverse selection by the

heterogeneous peer effect and puts forward the basic hypotheses. Section 5discusses the

identification strategies. Section 6presents the empirical results and heterogeneity analyses.

Section 7concludes with policy implications.

2|LITERATURE REVIEW

Our work is closely related to two streams of literature—empirical studies that test the

existence of adverse selection in the micro health insurance program, and studies of the peer

effect in insurance demand.

DU ET AL.

|

1065