On the Use of Information in Oligopolistic Insurance Markets

• DOI: http://doi.org/10.1111/j.1539-6975.2012.01490.x
• Author: Iris Kesternich,Heiner Schumacher
• Date: 01 March 2014
• Published date: 01 March 2014

© The Journal of Risk and Insurance, 2014, Vol. 81, No. 1, 159–175

DOI: 10.1111/j.1539-6975.2012.01490.x

159

ONTHEUSE OF INFORMATION IN OLIGOPOLISTIC

INSURANCE MARKETS

Iris Kesternich

Heiner Schumacher

ABSTRACT

We analyze the use of information in an oligopolistic insurance market with

costly market entry. For intermediate values of entry costs, an equilibrium

exists that is profit maximizing for incumbents and in which companies do

not discriminate between high and low risks. The model therefore provides

an explanation for the existence of “unused observables,” that is, information

that (1) insurance companies collect or could collect, (2) is correlated with

risk, but (3) is not used to set premiums.

INTRODUCTION

Starting with Rothschild and Stiglitz (1976), the theoretical literature on insurance

markets is centered on the question of how companies can extract their customers’

private information. However, thereis substantial evidence of “unused observables,”

that is, information that (1) insurance companies collect or could collect, (2) is corre-

lated with customers’ risk, but (3) is not used to set premiums. For example, Finkel-

stein and Poterba (2006) use data on annuity purchases to show that the information

on the annuitant’s residential location does not influence the insurance premium

although it would help to predict future mortality risk. Information on gender is usu-

ally collected by default, but not used for pricing in long-term care and automotive

insurance markets.1In both markets, the expected costs for the insurer differ substan-

tially for men and women. Brown and Finkelstein (2007), Ivaldi (1996), and Makki and

Somwaru (2001) provide further empirical evidence on unused observables. This con-

trasts sharply with standard insurance theory where companies exploit any available

Iris Kesternich is with Ludwig-Maximilians University Munich, Germany.Heiner Schumacher

is with Goethe-University Frankfurt, Germany. The authors can be contacted via e-mail:

iris.kesternich@lrz.uni-muenchen.de and heiner.schumacher@econ.uni-frankfurt.de, respec-

tively. We thank Amy Finkelstein, Guido Friebel, Michael Hoy, Matthias Polborn, Monika

Schnitzer, Ray Rees, Andreas Richter, Achim Wambach, and Joachim Winter for helpful com-

ments; Anna Gumpert and Sebastian Kohls for excellent research assistance; and Georges

Dionne (the editor) and two anonymous referees for superb advice. Financial support through

SFB-TR15 is gratefully acknowledged.

1A report by the European Commission (2010) provides a good overview on the use of certain

observables, such as age and sex, by the European insurance industry.

160 THE JOURNAL OF RISK AND INSURANCE

risk-relevant information. A monopolistic insurance company will use it to determine

the profit-maximizing insurance premiums. In a competitive market, companies will

charge fair premiums to customers (and thereby discriminate between risks).

In this article, we analyze the circumstances under which there exists an equilibrium

with unused observables in an oligopolistic insurance market that is profit maximiz-

ing for companies.2Companies interact repeatedly and can discriminate between

high and low risks. We model the underlying game as a Bertrand oligopoly where

collusive equilibria exist when companies are sufficiently patient. However, if com-

panies are able to sustain collusion, then an equilibrium with unused observables is

(in general) not profit maximizing. We therefore assume that there is a limit to profit

maximization. Such a limit may have several causes.

Social Conventions

Kahneman, Knetsch, and Thaler (1986) show that many individuals have a clear idea

about what price is “fair” and when an increase in a company’s profits is “acceptable.”

A fair price does not necessarily correspond to the price that is obtained in a market

equilibrium. By now a large experimental literature reports a substantial willingness

to punish unfair behavior (see Camerer, 2003, for a survey). Fairness considerations

therefore may limit the price that firms can profitably charge to customers.

Antitrust Regulation

Unusually high industry prices and profits may create the suspicion that insurance

companies have formed a cartel. Cartel members therefore have to take into account

how their prices influence the likelihood of detection by the antitrust authority or

customers. In order to avoid detection and punishment, firms may thus choose to

maintain their pricing schedule even if higher collusive prices and profits could be

established (see Block, Nold, and Sidak, 1981, and the subsequent literature on cartel

pricing under the threat of detection; see Harrington and Chen, 2006, for a review).

Market Entry

The threat of market entry may limit the incumbent companies’ ability to maximize

profits (see Gilbert, 1989, for a review). A substantial literature in industrial organiza-

tion shows how firms can deter market entry. One mechanism is to set a limit price,

which is just low enough so that entry becomes unprofitable. When entry barriers

are modest, the limit price is lower than the monopoly price, but still larger than the

perfectly competitive price.

2Empirical evidence suggests that the common market structure in most insurance markets

is oligopolistic. Concentration indexes for the top five insurance companies in the nonlife

business in Europe in 2002 ranged from 27 percent in Germany to 89 percent in Finland

(Buzzacchi and Valletti, 2005). Concentration measures in the life insurance sector in most

developed nations in the 1990s have been constantly high. Even in the United States, the

least concentrated market, concentration indexes for the top five insurance companies have

been above 25 percent. Market concentration is also reflected in insurance premiums (Dafny,

Duggan, and Ramanarayanan, 2010).