Cream Skimming: Innovations in Insurance Risk Classification and Adverse Selection

• Date: 01 September 2018
• Author: David A. Cather
• Published date: 01 September 2018
• DOI: http://doi.org/10.1111/rmir.12102

Risk Management and Insurance Review

C

Risk Management and Insurance Review, 2018, Vol.21, No. 2, 335-366

DOI: 10.1111/rmir.12102

PERSPECTIVE

CREAM SKIMMING:INNOVATIONS IN INSURANCE RISK

CLASSIFICATION AND ADVERSE SELECTION

David A. Cather

ABSTRACT

Wedemonstrate how innovations in insurance risk classification can lead to ad-

verse selection, or cream skimming, against insurers that are slow to adopt such

pricing innovations. Using a model in which insurers with insufficient pricing

data cannot differentiate between low- and high-risk policyholders and there-

fore charge both the same premium, we show how innovative insurers develop

new risk classification data to identify overcharged low-risk policyholders and

attract them from rival insurers with reduced prices. Less innovative insurers

thus insure a growingpercentage of high-risk customers, resulting in adverse se-

lection attributable to their informational disadvantage. Next, we examine two

cases in which “Big Data” innovations in risk classification led to concerns about

cream skimming among U.S. auto insurers. First, we track the rapid adoption

of credit-based insurance scores as pricing variables in personal auto insurance

markets. Second, we examine the growing popularity of usage-based insurance

programs like telematics, plans in which insurers use data on policyholders’

actual driving behavior to set prices that attract low-risk customers. Issues as-

sociated with the execution of such pricing strategies are discussed. In both

cases, we document how rival insurers quickly adopt successful innovations to

reduce their exposure to adverse selection.

INTRODUCTION

A recurring theme in scholarly research on insurance markets is the issue of adverse

selection, a pricing phenomenon in the industry that, left unchecked, can cause great

financial harm to an insurance company. Simply stated, adverse selection occurs when

an insurer’s premium revenues are insufficient to cover its insured losses because it

insures an increasing proportion of high-risk customers and a dwindling proportion of

low-risk customers. Adverse selection is a concern in insurance markets that rely upon

risk classification to sort insurance applicants into pricing categories with similar risk

characteristics, including most insurance products purchased by individual consumers,

such as personal auto, homeowners, or life insurance.

David Cather is in the Department of Risk Management, Smeal College of Business, Penn State

University,355 Business Building, University Park, PA 16802; phone: 814-863-5455; fax: 814-865-

6284; e-mail: dac32@psu.edu

335

336 RISK MANAGEMENT AND INSURANCE REVIEW

While experienced insurance executives are very familiar with its harmful effects, ad-

verse selection is not well understood by people outside the insurance industry.1This

article examines the adverse selection phenomenon, explaining how it is a negative by-

product of a greater and more positive force within the industry, the force of competition

among insurers within the industry. It is typically assumed that competition serves the

best interests of consumers by providing insurers a financial incentive to find new ways

to attract and retain customers, such as adopting innovations that reduce prices or im-

prove insurance products. One such innovation involves the integration of useful new

types of pricing data into insurance risk classification systems. Thus, if an insurer can

develop innovative insurance pricing methods to identify and attract low-risk customers

away from higher priced rival insurers, these customers will have a financial incentive

to switch to the innovative insurer to save money by paying the lower premium rates

charged to the new, low-risk group. One insurer’s gain is another insurer ’s loss, how-

ever, and the loss of these low-risk policy owners can ultimately erode the customer

base and the profits of less innovative insurers through this adverse selection process.

Competition has a comparable divergent effect on insurance consumers as well. As

insurers increase their ability to differentiate between applicants with a high chance of

suffering claims and those less likely to incur insured losses, they are better able to match

premium rates to risk levels. In this sense, improvements in insurance pricing accentuate

the differences in premiums charged tolow- and high-risk applicants, as applicants who

are more likely to suffer losses are often charged increasingly higher premiums than

low-risk consumers. This trend toward divergence in prices between low- and high-risk

consumers is often interpreted differently by each group. Low-risk consumers often

view their lower premiums as a more equitable pricing structure; high-risk consumers

often view their higher premiums as inequitable or, at the extreme, as unfair pricing

practices.

Insurance executives often refer to this adverse selection process as cream skimming,

since the innovative insurer targets high-quality, low-risk insurance buyers who, like

cream in a container of fresh milk, rise to the top of the pool of all policyholders in

an insurance company. This article examines the relationship between cream skimming

and competition among insurers to develop improved risk classification systems. The

article consists of two parts. First, we develop a simple pooled premium model based

on the risk classification systems used in the personal insurance markets to describe

how innovations in insurer risk classification can lead to competition-induced adverse

selection. Based on the model, we discuss the conditions under which risk averse low-

risk policyholders choose to leave their insurers in favor of the lower priced products

of innovative insurers. In the second part of the article, we describe two recent real-

life examples that demonstrate the relationship between innovative risk classification

and adverse selection. First, we examine the introduction of credit-based insurance

scores (CBIS) in the pricing of personal auto insurance in the 1990s. The credit scoring

case demonstrates how insurers can integrate additional pricing variables into their

traditional pricing systems to identify low-risk consumers who are paying too much for

1For example, see Vaughan (2009), a former state insurance commissioner, who notes that many

consumers and legislators are unfamiliar with the economic characteristics of the insurance

industry that create an environment in which adverse selection occurs.

CREAM SKIMMING 337

insurance, and thus improve upon the current practices used in the classification of risk.

In the second example, we examine the growing popularity of setting auto insurance

premiums based on a driver’s actual driving activities, using data collected using GPS

tracking devices or data collection hardware installed in vehicles. Both innovations

are examples of the “Big Data” movement in business, in which businesses focus on

gleaning useful business information from expansive amounts of customer data, such

as a continuous stream of information about a customer’s financial transactions or daily

driving activities.

This article contributes to the literature on adverse selection in several important ways.

First, it expands upon previous commentary on adverse selection by insurance prac-

titioners (State Farm Insurance Companies, 2005), developing a model based on risk

aversion to explain how innovations in risk classification can lead to cream skimming

in competitive insurance markets. Second, it describes this type of competition-induced

adverse selection in a straightforward, nontechnical manner. Thus, key stakeholders

involved in public policy decisions who are unfamiliar with insurance risk classification

systems can better understand why competition often leads to the rapid adoption of

pricing innovations in insurance markets. Additionally, it examines two recent cases

that demonstrate how innovations in risk classification systems can abruptly change

insurance markets, detailing the key strategic decisions that are important to an insurer

that wishes to maximize its return on investment from the costly development of such

innovations. Because these key decision variables are vital to the successful execution of

an insurance company’s competitive strategy,insurers are generally reluctant to discuss

them in public, viewing them as trade secrets (Harrington, 2003). Thus, by identifying

and discussing these considerations, this article adds to the limited academic literature

that is focused on the strategic role of pricing and technology in competitive insurance

markets.

Perhaps most importantly, this article demonstrates why efforts to restrict the use of inno-

vative risk classification variables in the pricing of insurance may be counterproductive

to the best interests of many insurance consumers. Over the years, insurance executives

have come to recognize that innovations in risk classification are a common strategic

tactic in competitive insurance markets. The industry has also defined standards that es-

tablish the characteristics of appropriate risk classification variables,2standards that are

increasingly relevant as insurers develop a wider variety of new pricing data in today’s

information-driven economy. Unfortunately, the general public is unfamiliar with this

literature and often views insurance pricing innovations with skepticism. By chronicling

two highly publicized examples of recent innovations in the personal auto insurance

markets, this article demonstrates to a broader audience how such pricing innovations

can lead to a redistribution of premium charges between low-risk and high-risk insur-

ance policyholders. Additionally, it explains why insurers react to pricing innovations

with urgency, quickly adopting successful new risk classification schemes in an effort to

avoid adverse selection.

2See Actuarial Standards Board, 2005, “Risk Classification” (Actuarial Standard of Practice

No. 12), http://www.actuarialstandardsboard.org/asops/risk-classification-practice-areas/

(accessed June 26, 2017) for a description of the characteristics of pricing variables that are

recognized as being suitable for insurance pricing.