E C O N O M I C S OF I N S U R A N C E

ADVANCED TEXTBOOKS
IN ECONOMICS

V O L U M E 29

Editors:

C.J. BLISS

M.D.INTRILIGATOR

Advisory Editors:

W. A . BROCK

D.W.JORGENSON

A . P. K I R M A N

J.-J.LAFFONT

J.-F. R I C H A R D

NORTH-HOLLAND
AMSTERDAM . LONDON . NEW YORK . TOKYO
ECONOMICS
OF INSURANCE

K.BORCHt

Norwegian School of Economics and Business Administration,

Bergen, Norway

Edited and Completed

after Professor Borch's Death by

Knut Κ. A A S E and Agnar S A N D M O

Norwegian School of Economics and Business Administration,

Bergen, Norway

NORTH-HOLLAND
AMSTERDAM · LONDON . N E W YORK · TOKYO
 E L S E V I E R S C I E N C E P U B L I S H E R S Β. V.
Sara Burgerhartstraat 25
P.O. Box 211,1000 A E Amsterdam, The Netherlands

ISBN: 0 444 87344 9

1st edition: 1990

2nd printing: 1992

Library of Congress Cataloging-in-Publication Data

Borch,KarlH. (Karl Henrik), 1919-1986.

Economics of insurance / K. Borch; edited and compiled by Knut Κ. Aase
and Agnar Sandmo.
p. cm. — (Advanced textbooks in economics; v. 29)
I S B N (invalid) 0 444 8 7344 9 ( U . S . )
1. Insurance. I. Aase, Knut Kristian. II. Sandmo, Agnar.
III. Title. IV. Series.
HG8026.B67 1990
368'.01—dc20 89-26556
CIP

® 1990 E L S E V I E R S C I E N C E P U B L I S H E R S Β.V. Allrightsreserved.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or
by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written
permission of the publisher, Elsevier Science Publishers B.V., Copyright & Permissions Department,
P.O. Box 521,1000 A M Amsterdam, The Netherlands.

Special regulations for readers in the U . S . A . - This publication has been registered with the Copyright
Clearance Center Inc. ( C C C ) , Salem, Massachusetts. Information can be obtained from the C C C about
conditions under which photocopies of parts of this publication may be made in the U . S . A . All other copy-
right questions, including photocopying outside of the U . S . A . , should be referred to the copyright owner,
Elsevier Science Publishers Β. V , unless otherwise specified.

N o responsibility is assumed by the publisher for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions or ideas contained in the material herein.

This book is printed on acid-free paper.

PRINTED IN THE N E T H E R L A N D S
This page intentionally left blank
PREFACE

When Karl Borch died in December, 1986, he was working on the manu-
script of a textbook on the economic theory of insurance, the area which
had been his main interest throughout his academic career. T h e manuscript
had not been finished; we know that he planned to add at least one ad-
ditional chapter, and it may well be that he would have expanded on the
chapters which had already been written. But even in its unfinished form
the manuscript does give a very readable and fascinating overview of the
field to wich Borch made so many pioneering contributions since his first
publications around 1960. W e have therefore decided to publish these chap-
ters more or less in the form in which they were to be found at Borch's
death; only minor editing has been undertaken.
Given the incomplete nature of the manuscript we have further decided,
in consultation with the series editors, to add a selection of Borch's articles
in the economics of insurance. W e have naturally chosen the articles with
a view to their complementarity with the textbook exposition, so that they
may be seen also as a selection of readings to be studied along with the
main text. W e hope that the book as a whole will serve as a stimulating
introduction to the economics of insurance.
W e are glad to be able to publish this book as a memorial to a scholar
who was not only a leader in his field, but also an important source of
inspiration and support to a large number of colleagues and students, both
at the Norwegian School of Economics and Business Administration and
at many universities and research institutions throughout the world.

Bergen, March 1989

Knut Κ . Aase Agnar Sandmo

This page intentionally left blank
CONTENTS

Preface vii

C h a p t e r 1. I n s u r a n c e a n d Economics 1

1. Insurance and Economic Analysis 1

2. Insurance in the National Economy 6
3. Different Markets for Insurance 11
4. T h e Place of Uncertainty in the Theories of the
Austrian School 16

C h a p t e r 2. I n s u r a n c e a n d U t i l i t y T h e o r y 29

1. T h e Ordering of Insurance Risks 29

2. T h e Development of Reinsurance 35
3. T h e Reinsurance Market 39
4. Some Elements of a Theory of Reinsurance 47
5. T h e Safety Loading of Reinsurance Premiums 61
1. Introduction 61
2. A model of the Reinsurance Market 62
3. Equilibrium Price in a Market 68
4. Discussion of a Special Case 71
5. Market Equilibrium under Uncertainty 78
6. Equilibrium in a Reinsurance Market 81
1. Introduction 81
2. A Model of the Reinsurance Market 83
3. T h e Price Concept in a Reinsurance Market 86
χ Economies of Insurance

4. Existence of an Equilibrium Price 91

5. T h e Models of Allais and Arrow 98
6. T h e Problem seen as an η-person Game 103
7. T h e Theory of Risk 105
1. Historical Note 105
2. T h e Classical Theory of Risk 106
3. T h e Collective Risk Theory 113
4. T h e Modern Risk Theory 120
8. T h e Utility Concept Applied to the Theory of Insurance 134
1. Introduction 134
2. T h e Theory of Risk 135
3. Measurable Utility 137
4. Application to Reinsurance 140
5. Conclusion 142
9. Optimal Insurance Arrangements 144
10. Objectives and Optimal Decisions in Insurance 150
1. Introduction 150
2. Some Simple Formulations of the Objectives 151
3. A More Realistic Model 157
4. Conclusion 160

C h a p t e r 3. I n s u r a n c e a n d C o m p e t i t i v e E q u i l i b r i u m 163

1. Principles of Premium Calculation 163

2. Insurance Premiums in the Market 168
3. Insurance Premiums and Asset Prices 174
4. Insurance and the Theory of Financial Markets 181
1. Introduction 181
2. A Simple Model 182
3. Discussion of the Simple Model 187
4. Generalization of the Simple Model 189
5. Additive Insurance Premiums: A Note 192
6. Static Equilibrium under Uncertainty and
Incomplete Markets 198
1. Introduction 198
 Contents xi

2. Arrow's state Model 200

3. T h e Traditional Approach 203
4. Incomplete Markets 207
7. A Theory of Insurance Premiums 209
1. Introduction 209
2. A Model of an Insurance Market 211
3. Examples and Applications 215
4. A Dynamic approach to Utility 222
5. A Modification of the De Finetti's Model 225
6. T h e Models and the Real World 227

C h a p t e r 4. Life I n s u r a n c e 231

1. Early Development of Life Insurance 231

2. Elements of Actuarial Mathematics 233
3. Life Insurance, Consumption and Saving 240
4. Optimal Life Insurance 244
1. Introduction 244
2. T h e Simplest Saving — Consumption Models 245
3. Models with Uncertainty and Insurance 249
4. Life Insurance 252
5. Insurance for the Benefit of Survivors 257
5. Life Insurance and Consumption 261

C h a p t e r 5. Business I n s u r a n c e 265

1. Insurance bought by Business Enterprises 265

2. Risk Sharing in Business 268
3. Captive Insurance Companies 273
4. Application of Game Theory to some Problems in
Automobile Insurance 276
1. Discussion of a Numerical Example 277
2. A more General Case 281
3. Another Numerical Example 288
4. Conclusion 289
xii Economics of Insurance

C h a p t e r 6. H o u s e h o l d I n s u r a n c e 293

1. Administrative Expenses 293

2. No-Claim Bonus and Merit Rating 299
3. T h e Market for Household Insurance 304
4. T h e Optimal Insurance Contract in a Competitive Market 310

C h a p t e r 7. U n i n s u r a b l e Risks 315

1. Conditions that a Risk is Insurable 315

2. Adverse Selection 319
3. Moral Hazard 325
4. T h e Monster in Loch Ness 331
5. Ethics, Institutions and Optimality 335
1. Introduction 335
2. Presentation of the Problem 337
3. Some Examples from Economics 340
4. Examples from Insurance 342
5. Concluding Remarks 346
6. T h e Price of Moral Hazard 346
7. Insuring and Auditing the Auditor 350
1. Introduction 350
2. T h e Sampling Approach — A Simple Model 351
3. A Two-person Game 354
4. A Three-person Game 356
5. Other Applications of the Model 360

C h a p t e r 8. R i s k T h e o r y a n d G o v e r n m e n t Supervision 363

1. Is Regulation of Insurance Companies Necessary? 363

1. T h e Basic Model 363
2. A Naive Dividend Policy 365
3. T h e Form of the Optimal Dividend Policy 367
4. Regulation and the Ruin Problem 371
5. A Numerical Example 372
6. Final Remark 375
 Contents xiii

2. Risk theory and Serendipity 375

1. Introduction 375
2. A model of the Insurance Market 376
3. Examples and Applications 380
4. A Dynamic Approach to Utility 387
5. T h e Model on the Real World 392

Author Index 395

Subject Index 397

This page intentionally left blank
CHAPTER 1

INSURANCE AND ECONOMICS

1. Insurance a n d Economic A n a l y s i s

1.1 It is not possible to give a definition of insurance which is short and

precise, and at the same time completely satisfactory. A number of defi-
nitions can be found in the insurance literature, but they tend to be long
and involved, and usually one can find some kind of insurance that does
not quite fit the definition.
For the present purpose it is sufficient to consider an insurance contract
as described by two elements:

(i) Ρ = the premium paid by the insured when the contract is

concluded.
(ii) χ = the compensation which the insured receives if specific
events occur when the contract is in force. Clearly χ is
a random variable, and must be described by a proba-
bility distribution F(x).
T h e essential objective of a theory of insurance is to determine the rela-
tionship between the two elements, i.e. how the premium Ρ depends on
the properties of the probability distribution F(x). A pair (P,F(x)) can
obviously also be interpreted as a gamble, or as a risky investment. This
means that results in the theory of insurance may be relevant in the study
of uncertainty in general economics and finance. It also means that meth-
ods developed in economic and financial analysis may have applications in
insurance.

1.2 Insurance is an economic activity, and an important one. Most authors

of general economic texts discuss insurance, usually in a separate section.
In his Wealth of Nations, (Book I , Chapter 10) A d a m Smith (1776) writes
that the insurance "premium must be sufficient to compensate the common
losses, to pay the expense of management, and to afford such a profit as
might have been drawn from an equal capital employed in any common
2 Economies of Insurance

trade". This is a very good statement of how insurance premiums must be

calculated, and we shall return to it in later chapters.
On the effect of insurance A d a m Smith writes (Book V , Chapter 1): "The
trade of insurance gives great security to the fortunes of private people, and
by dividing among a great many that loss which would ruin an individual,
makes it fall light and easy upon the whole society. In order to give this
security, however, it is necessary that the insurers should have a very large
capital". He does not say how large this capital should be, nor does he
discuss how much of his money an entrepreneur would be prepared to risk
in the insurance business. A d a m Smith does however observe (Book I ,
Chapter 10) "the universal success of lotteries", and compares it with the
"very moderate profit of insurers". In modern terms this means that A d a m
Smith considered people's desire to gamble as a more important element
in the economy than their "risk aversion", i.e. their willingness to pay to
get rid of risks.

1.3 T h e passages quoted above show that more than 200 years ago A d a m
Smith had a good insight into the essentials of insurance. In the following
century there was a rapid development in economic theory, which did not
seem to have led to any deeper understanding of insurance. T h e three main
centers of this development were Cambridge, Lausanne and Vienna. A t the
time the theories developed at these centers were considered as different
"schools". Today the differences between them seem less fundamental, and
it is convenient to refer to the three schools as the "neo-classical" theory.

1.4 In Vienna Carl Menger, the founder of the Austrian School does not
seem to have anything important to say about insurance. His most brilliant
student and successor, Eugen Böhm-Bawerk did however write his disser-
tation (Habilitationsarbeit, 1881) about the value of contingent claims.
Apparently it did not occur to him that the obvious application of this
theory was to insurance premiums. T h e owner of an insured ship has a
claim against the insurer if the ship is lost or damaged at sea. T h e value,
or the price he pays for this contingent claim is clearly the premium de-
manded by the insurer. Böhm-Bawerk does not seem to have returned to
the problems of his dissertation at any later stage. His purpose was appar-
ently to show that values, or "certainty equivalents" could be computed for
contingent claims, and hence that one could confidently proceed with the
development of a theory based on complete certainty.
It was obviously a fairly difficult mathematical problem to compute the
value of a complicated contingent claim, and the solution was probably
 Insurance and Economies 3

beyond the mathematics of Böhm-Bawerk. In the second half of the nine-

teenth century Austrian and German actuaries developed a "risk theory".
T h e purpose of this theory was essentially to determine the capital which
an insurer had to hold in order to give adequate security to the buyers of
insurance contracts. If actuaries could solve this problem, Böhm-Bawerk
could with some justification ignore the uncertainty inherent in any con-
tract involving contingent claims. T h e effect of this division of labour was
however that an economic theory of insurance would have to be based not
only on standard economic analysis, but also on a "risk theory" beyond
the grasp of most economists.

1.5 In Lausanne Leon Walras (1874) saw insurance as a device for remov-
ing the uncertainty inherent in all other economic activities. This made it
reasonable to develop a theory for general economic equilibrium under full
certainty, leaving the insurance sector as a special subject to be studied
separately.
This is of course essentially the same conclusion as the one reached by
Böhm-Bawerk, although with a slightly different interpretation. T o Walras
the link between the two sectors is the "prime d'assurance", which a modern
business economist will recognize as the "cost of capital" to business of
different risk classes.

1.6 In Cambridge Alfred Marshall came close to develop an economic the-

ory of insurance. In his Principles (1890) he discusses insurance premiums
as the price one has to pay to get rid of the "evils of uncertainty". In
the Mathematical Appendix to the book he mentions the work of Daniel
Bernoulli 150 years earlier as an "interesting guess", and he seems to rec-
ognize that the Bernoulli principle may be the key to the problem of insur-
ance premiums. Marshall did not himself develop this idea, and nobody in
Cambridge seemed to take the hint.
Marshall wrote about the "evil of risk", and believed that people were
willing to pay to get rid of this evil. He noted that businessmen paid in-
surance premiums "which they know are calculated on a scale sufficiently
above the true actuarial value of the risk to pay the companies' great ex-
penses of advertising and working, and yet to yield a surplus of net profits".
This means of course in modern terms that Marshall believed that the im-
portant decision makers in the economy were "risk averse".
It is interesting to compare Marshall's views with those expressed by
A d a m Smith more than 100 years earlier. A t Marshall's time people were
no longer allowed to organize lotteries for their own profits. Gambling
4 Economies of Insurance

seems to have flourished in the Victorian era, and had at least a snob ap-
peal, even if it was not quite respectable. People's willingness to gamble
was however not worthy of the study by serious economists, who focused
their interest on the prudent risk averse investor. This attitude does not
seem to have changed during the last 100 years, and this is in a way surpris-
ing. Governments in many countries have made increasing use of people's
propensity to gamble to increase revenue by state lotteries and to borrow at
low interest rates by premium bonds. T h e revenue obtained by exploiting
"risk lovers" is already a fairly important element in public finance, and
the attitudes of these people seem to deserve serious study.

1.7 T h e Bernoulli principle and its implications will be discussed in the

next chapter, but it may be useful to give a brief outline already at the
present stage. In the early days of the calculus of probability it was taken
as granted that the value, and hence the fair price of a gamble was the
mathematical expectation of the gain. If the probability of a gain χ is
/ ( # ) , the fair price would be

oo
E{x) = χ = ^ χ f(x)
x=0

Applied to insurance this means that the fair premium for a risk described
by the probability distribution f(x) would be

oo

Ρ = χ = Σχί(χ)
x=0

T h e counter example given by Bernoulli (1738) has become known as the

St. Petersburg Paradox. He considered a game in which a coin is tossed
until it shows heads. If the first head appears at the n'th toss, a prize of
n
2 is paid. T h e expected gain in this gamble is

oo
n n
E{x} = ^(l/2) 2 = oo
71=1

Bernoulli argued that no rational person would be willing to pay an ar-

bitrarily large amount for the right to participate in this gamble. A s an
alternative Bernoulli suggested that the person would assign the "moral"
 Insurance and Economies 5

value or utility of log χ to a physical gain of x. T h e value of the gamble is

then the moral expectation:
oo oo
n n
£{logx} = £ ( l / 2 ) l o g 2 " = l o g 2 ^ n ( l / 2 ) = 2 log 2
n=l n=l

1.8 Bernoulli gave his own justification for setting the moral value of a
gain of χ equal to l o g x , but from his correspondence it is clear that he
was willing to replace log χ with some other concave function, such as
u(x) = y/x, or log(c + # ) . This means that the utility assigned to a gamble
described by a probability distribution f(x) would be

oo
E{u(x)} = ^u(x)f(x) (1)
x=0

where u{x) is an arbitrary increasing concave function, i.e. the inequalities

u'{x) > 0 and u"{x) < 0 hold.
From Jensen's inequality it follows that

E{u(x)} < E{x)

Hence a person will pay less than the mathematical expectation for the
right to play a gamble. Similarly an insurer will demand a "risk premium"
in addition to the expected loss in order to cover a risk.
Bernoulli assumed that the function u(x) was concave. This implies that
the person considered is risk averse, and that he will pay less than the
expected gain for a lottery ticket, and more than the expected loss for an
insurance contact covering the risk. If u(x) is convex, i.e. u"(x) > 0, the
person will be a "risk lover", and he will buy lottery tickets even if the
price is higher than the expected gain. He will however not buy insurance
if the premium is above the expected loss. If u(x) is linear, the person will
be "risk neutral".

1.9 Marshall in Cambridge was aware of these possibilities of generalizing

the Bernoulli Principle, and he was not the only one. Bernoulli's work,
originally published in Latin, was translated into German in 1896 by a
prominent mathematician, Alfred Pringsheim. T h e translation was pub-
lished with an introduction by the philosopher Ludwig Fick, and had the
title "Grundlage der modernen Werthlehre", (Foundation of the Modern
Theory of Value).
6 Economies of Insurance

This title was however premature, by about 50 years. T h e economists

who developed the neo-classical theory of value, continued to ignore the
Bernoulli hypothesis. They put it to work only after von Neumann and
Morgenstern (1947) had proved the hypothesis as a theorem. In fact they
proved that if a person had a consistent preference ordering over the set of
all games described by probability distributions, there exists a function u(x)
such that (1) gives the utility assigned to the gamble described by f(x).
This result has become known as the Expected Utility Theorem, and it is
indeed the foundation of the modern theory of value.

1.10 It is a little surprising that economists did not apply the Bernoulli
hypothesis before it had been proved as a theorem. It had been endorsed as
an interesting guess by the leading Cambridge economist, and Austrian and
German mathematicians, i.e. Czuber (1904) had shown how it could be used
to determine insurance premiums. Usually economists have been willing
to explore the implications of heroic assumptions which seem reasonable
on intuitive ground. It is not easy to see what held them back when it
came to insurance, although one may guess that they had accepted that
this activity required special methods of analysis.
T h e breakthrough was made by Arrow (1953), who showed how the
values of contingent claims were determined by market forces, and that
these made demand for risk-bearing services equal to the supply in the
market. Arrow's model is quite general, but the application he has in
mind is the prices of securities in a stock market. It is however clear
that the model can also be interpreted as an insurance market, and this
interpretation will be discussed in the following chapters.

2. Insurance in the N a t i o n a l E c o n o m y

2.1 In all industrial countries insurance is an important sector of the econ-

omy. In the U.S.A. there were in 1984 about 5000 insurance companies.
They collected more than 250 billion dollars in premiums, and provided
employment for about 2 million people.
Insurance is about equally important in other industrial countries. In-
ternational comparisons are difficult, since the statistical definition of in-
surance varies from one country to another. In most countries the primary
source of insurance statistics is the reports of the governmental supervi-
sory authority, which naturally tends to focus on the statistics it needs to
check that the insurance companies under supervision are solvent. In other