Adverse retention: Strategic renewal of guaranteed renewable term life insurance policies

• Published date: 01 December 2021
• Author: Gene C. Lai,Hisashi Nakamura,Shinichi Yamamoto,Takau Yoneyama
• Date: 01 December 2021
• DOI: http://doi.org/10.1111/jori.12338

J Risk Insur. 2021;88:1001–1022. wileyonlinelibrary.com/journal/JORI

|

1001

Received: 29 April 2020

|

Revised: 20 January 2021

|

Accepted: 25 January 2021

DOI: 10.1111/jori.12338

ORIGINAL ARTICLE

Adverse retention: Strategic renewal of

guaranteed renewable term life

insurance policies

Gene C. Lai

1

|Hisashi Nakamura

2

|Shinichi Yamamoto

3

|

Takau Yoneyama

4

1

Department of Finance, Professor and

James J. Harris Chair of Risk

Management and Insurance, University

of North Carolina at Charlotte, Charlotte,

North Carolina, USA

2

Graduate School of Commerce and

Management, Hitotsubashi University,

Tokyo, Japan

3

Department of Economics, Ritsumeikan

University, Kyoto, Japan

4

Faculty of Business Administration,

Tokyo Keizai University, Tokyo, Japan

Correspondence

Gene C. Lai, Department of Finance,

Professor and James J. Harris Chair of

Risk Management and Insurance,

University of North Carolina at

Charlotte, Charlotte, NC 28223‐0001,

USA.

Email: glai@uncc.edu

Funding information

Summer Research Grant, Belk College of

Business, University of North Carolina,

Charlotte; Grant‐in‐Aid for Scientific

Research from the Japan Society for the

Promotion of Science,

Grant/Award Number: 23530553

Abstract

This study examines adverse retention using guaran-

teed renewable level term life insurance policies as a

sample. No additional underwriting or physical ex-

amination is performed at the time of renewal. Our

results show adverse retention exists in retention de-

cisions. Further, we also find that the degree of adverse

retention increases with age and policy amount and

decreases with time elapses. Our results are robust

with respect to the trend of before and after the re-

newal. During the prerenewal period, we find the

prerenewal mortality rate is not higher than the non-

renewal mortality rate which is opposite to the evi-

dence of postrenewal period. We also find correlation

between “residuals of lapse”and “residuals of fatal-

ities”is positive, also implying adverse retention exists.

Most empirical studies on adverse selection often face

problems such as the role of underwriting, entangle-

ment of adverse selection, moral hazard, and regula-

tions. Our study is free of these problems.

KEYWORDS

adverse retention, adverse selection, asymmetric information,

guaranteed renewable insurance, level term life insurance

© 2021 American Risk and Insurance Association

1|INTRODUCTION

Using a sample of guaranteed renewable level term life (GRLTL) insurance policies, this study

examines adverse retention. Our paper is inspired by the outstanding paper by Altman et al.

(1998), which is the first paper to examine adverse retention using health insurance data. They

find that high‐risk individuals are more likely to renew their policies, whereas low‐risk in-

dividuals are less likely to renew, and conclude that adverse retention exists. Unlike Altman,

Culter, and Zeckhauser, who use health insurance data, we use renewable term life insurance

data to examine adverse retention and extend Altman et al. (1998) by considering the impact of

policy characteristics on adverse retention.

We believe that a discussion on adverse retention should form part of the adverse selection

literature because GRLTL policies provide options for policyholders when renewing. Whether

adverse selection exists has been a major research question for decades, with no conclusive

answers to date.

In an excellent paper, Cohen and Siegelman (2010) review the literature about testing for

adverse selection in insurance markets and reach “generally negative”conclusions about life

insurance markets. For example, Cawley and Philipson (1999) find that the mortality rate is

lower for people who have life insurance than for those who do not have it. They also find that

unit premiums for life insurance with more coverage tend to be lower.

1

These results are

contrary to the prediction of adverse selection theory.

He (2009) questions some of Cawley and Philipson's (1999) results. He uses an example that

shows their sample is biased against detecting adverse selection and finds significant adverse

selection effects in life insurance markets. McCarthy and Mitchell (2010) compare mortality

rates for policyholders and others across several countries and find that purchasers of life

insurance (except in Japan) have substantially lower mortality risk than the population as a

whole.

2

They conclude that adverse selection does not exist and insurance companies do a good

job in underwriting and are effective at screening out bad risk. In summary, the evidence on

whether adverse selection exists in health and life insurance markets is not conclusive. This

paper has three goals. The first is to examine whether adverse retention or a broad sense of

adverse selection exists in life insurance markets. The second is to examine the impact of a

policy's characteristics on the degree of adverse retention. The third is to shed additional light

on the adverse selection literature using a unique data set that is free of underwriting, moral

hazard, and nonuniform regulation biases. Our evidence is important because the current

empirical evidence on whether adverse selection exists is not conclusive.

To empirically examine whether adverse retention or adverse selection exists, we use

GRLTL data from a Japanese life insurance company. We chose Japanese life insurance data for

many reasons. First, the data set has several distinct characteristics. GRLTL has two important

features: guaranteed renewability and a level‐premium term policy over the term (10 years).

Guaranteed renewability gives a policyholder the right to renew the contract at the end of the

insurance period without having to demonstrate good health. Although premiums often rise at

the time of renewal, no medical examination is necessary for renewal. In other words, GRLTL

provides a natural experiment for testing adverse selection. Some papers (e.g., Cawley &

Philipson, 1999; McCarthy & Mitchell, 2010) find no adverse selection and conclude that

1

Hendel and Lizzeri (2003) explore term life insurance but focus on the insurer's commitment. Yamamoto et al. (2014) analyze the data of Japanese term life

and whole life insurance and observe evidence consistent with the existence of asymmetric information.

2

They find a negative, rather than a positive, correlation between risk and coverage.

1002

|

LAI ET AL.