Handbook of Insurance

Huebner International Series on Risk, Insurance, and Economic Security

J. David Cummins, Editor The Wharton School University of Pennsylvania Philadelphia, Pennsylvania, USA

Series Advisors:

Dr. Phelim P. Boyle University of Waterloo, Canada

Dr. Jean Lemaire University of Pennsylvania, USA

Professor Akihiko Tsuboi Kagawa University, Japan

Dr. Richard Zeckhauser Harvard University, USA

Other books in the series:

Cummins, 1. David and Derrig, Richard A.: Classical Insurance Solvency Theory

Borba, Philip S. and Appel, David: Benefits, Costs, and Cycles in Workers' Compensation

Cummins, 1. David and Derrig, Richard A.: Financial Models of Insurance Solvency

Williams, C. Arthur: An International Comparison of Workers' Compensation

Cummins, 1. David and Derrig, Richard A.: Managing the Insolvency Risk of Insurance Companies

Dionne, Georges: Contributions to Insurance Economics Dionne, Georges and Harrington, Scott E.: Foundations of

Insurance Economics Klugman, Stuart A.: Bayesian Statistics in Actuarial Science Durbin, David and Borba, Philip: Workers' Compensation Insurance:

Claim Costs, Prices and Regulation Cummins, 1. David: Financial Management of Life Insurance

Companies Gustavson, Sandra G. and Harrington, Scott E.: Insurance,

Risk Management, and Public Policy Lemaire, Jean: Bonus-Malus Systems in Automobile Insurance Dionne, Georges and Laberge-Nadeau, Claire: Automobile Insurance:

Road Safety, New Drivers, Risks, Insurance Fraud and Regulation Taylor, Greg: Loss Reserving: An Actuarial Perspective

Handbook of Insurance

Edited by

Georges Dionne Risk Management Chair

HEC-Montrea1

SPRINGER SCIENCE+BUSINESS MEDIA, LLC

ISBN 978-0-7923-7911-9 ISBN 978-94-010-0642-2 (eBook)

DOI 10.1007/978-94-010-0642-2

Library of Congress Cataloging-in-Publication Data

A C.l.P. Catalogue record for this book is available from the Library of Congress.

Copyright © 2000 by Springer Sciencet-Business Media New York Originally published by Kluwer Academic Publishers in 2000 AII rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, mechanical, photo­copying, recording, or otherwise, without the prior written permission of the publisher, Springer Science+Business Media, LLC.

Printed on acid-free paper.

To:

J. David Cummins, University of Pennsylvania

Robert Lacroix, Universite de Montreal

Pierre Picard, Universite de Paris X-Nanterre

Jean-Marie Toulouse HEC-Montreal

for providing me significant leverages over the past twenty years.

TABLE OF CONTENTS

CONTRIBUTING AUTHORS

REFEREES

PREFACE Denis Kessler

INTRODUCTION Georges Dionne

PART I HISTORY

DEVELOPMENTS IN RISK AND INSURANCE ECONOMICS: THE PAST 25 YEARS Henri Louberge

1.1 Introduction 1.2 Insurance economics in 1973 1.3 Developments 1.4 New approaches: finance and insurance 1.5 Conclusion 1.6 References

PART II INSURANCE THEORY WITHOUT INFORMATION PROBLEMS

2 NON-EXPECTED UTILITY AND THE ROBUSTNESS OF THE CLASSICAL INSURANCE PARADIGM Mark J Machina

2.1 Introduction 2.2 Non-expected utility preferences and generalized expected

utility analysis 2.3 Individual demand for insurance 2.4 Pareto-efficient bilateral insurance contracts 2.5 Pareto-efficient multilateral risk sharing 2.6 Self-insurance versus self-protection 2.7 Outcome kinks and first order risk aversion 2.8 Extensions and limits of robustness

XVll

XIX

XXll1

XXVll

3

3 4 8

20 24 25

35

37

37

39 52 62 65 68 70 75

Vlll Table of Contents

2.9 Insurance as a source of non-expected utility preferences 79 2.10 Conclusion 82 2.11 References 91

3 OPTIMAL INSURANCE DESIGN: WHAT CAN WE DO WITH AND WITHOUT EXPECTED UTILITY? 97 Christian Gallier

3.1 Introduction 97 3.2 The basic framework 99 3.3 The case of linear transaction costs 101 3.4 Nonlinear transaction costs 109 3.5 Other reasons for partial insurance III 3.6 Conclusion 113 3.7 References 113

4 THE EFFECTS OF CHANGES IN RISK ON RISK TAKING: A SURVEY 117 Christian Gallier and Louis Eeckhoudt

4.1 Introduction 117 4.2 A simple model 118 4.3 Detrimental changes in risk 119 4.4 The comparative statics of changes in the controllable risk 121 4.5 The comparative statics of background risk 125 4.6 Extensions 127 4.7 Conclusion 129 4.8 References 129

5 THE THEORY OF INSURANCE DEMAND 131 Harris Schlesinger

5.1 Introduction 131 5.2 The single risk model 132 5.3 The model with multiple risks 142 5.4 Concluding remarks 150 5.5 References 150

PART III ASYMMETRIC INFORMATION: THEORY 153

6 OPTIMAL INSURANCE UNDER MORAL HAZARD 155 Ralph Winter

6.1 Introduction 155 6.2 The simplest model 157

Table of Contents IX

6.3 Self-protection and uncertain losses 162 6.4 Loss reduction and moral hazard 165 6.5 General assumptions on the distribution of losses 167 6.6 Extensions 171 6.7 References 181

7 ADVERSE SELECTION IN INSURANCE MARKETS 185 Georges Dionne, Neil Doherty and Nathalie Fombaron

7.1 Introduction 186 7.2 Basic assumptions and some fundamental results 187 7.3 Monopoly 190 7.4 Competitive contracts 202 7.5 Moral hazard and adverse selection 225 7.6 Adverse selection when people can choose their risk status 228 7.7 Concluding remarks: extensions to the basic models 233 7.8 References 237

8 THE THEORY OF RISK CLASSIFICATION 245 Keith J. Crocker and Arthur Snow

8.1 Introduction 245 8.2 Risk classification in the absence of hidden knowledge 246 8.3 Risk classification in the presence of hidden knowledge 249 8.4 Risk classification and incentives for information gathering 260 8.5 Competitive market equilibrium and extensions of the

basic model 270 8.6 Summary and conclusions 273 8.7 References 274

9 THE ECONOMICS OF LIABILITY INSURANCE 277 Scott Harrington and Patricia Danzon

9.1 Introduction 277 9.2 Legal liability, deterrence, and insurance 279 9.3 Limited liability, insurance, and deterrence 287 9.4 Liability insurance with correlated risk 291 9.5 Contract interpretation and litigation 293 9.6 The liability insurance crisis 294 9.7 Efficiency of the tort liability / liability insurance system 301 9.8 Conclusions 306 9.9 References 307

x Table of Contents

10 ECONOMIC ANALYSIS OF INSURANCE FRAUD 315 Pierre Picard

10.1 Introduction 315

10.2 Costly state verification: the case of deterministic auditing 317 10.3 Costly state verification: deterministic auditing with

manipulation of audit costs 324 lOA Costly state falsification 331 10.5 Costly state verification: the case of random auditing 336 10.6 Morale costs and adverse selection 343 10.7 The credibility issue 348 10.8 Collusion with agents 352 10.9 Conclusion 356 10.10 References 360

PART IV ASYMMETRIC INFORMATION: EMPIRICAL ANALYSIS 363

11 ECONOMETRIC MODELS OF INSURANCE UNDER ASYMMETRIC INFORMATION 365 Pierre-Andre Chiappori

1l.l Introduction 365 1l.2 Empirical tests of information asymmetries:

the theoretical background 367 1l.3 Empirical estimations of asymmetric information in the

static framework 377 1104 Dynamic models of information asymmetries 386 11.5 Conclusion 390 1l.6 References 391

12 THE EMPIRICAL MEASURE OF INFORMATION PROBLEMS WITH EMPHASIS ON INSURANCE FRAUD 395 Georges Dionne

12.1 Introduction 395 12.2 Measurement of residual adverse selection in the portfolio

of an insurer 399 12.3 Ex-ante moral hazard and choices of work contracts 402 1204 Ex-post moral hazard, demand for medical services, and

duration of work leaves 404 12.5 Insurance fraud 407 12.6 Adverse selection and the quality of the product in a market 412 12.7 Conclusion 414 12.8 References 414

Table of Contents Xl

13 INCENTIVE EFFECTS OF WORKERS' COMPENSATION: A SURVEY 421 Bernard Fortin and Paul Lanoie

13.1 Introduction 422 13.2 Theoretical work 423 13.3 Empirical work 432 13.4 Concluding remarks and policy discussion 452 13.5 References 454

14 EXPERIENCE RATING THROUGH HETEROGENEOUS MODELS 459 Jean Pinquet

14.1 Introduction 459 14.2 Tariff structures and experience rating schemes in the

insurance industry 462 14.3 Models with heterogeneity: definitions and examples of interest

for insurance rating 463 14.4 Heterogeneous models and prediction on longitudinal data

through a revelation principle 472 14.5 Heterogeneity, state dependence and prediction on

longitudinal data 478 14.6 Estimation and tests for heterogeneous models: a survey of

the literature 481 14.7 Score-based inference for linear and Poisson models with

heterogeneity 485 14.8 Examples of consistent estimators for heterogeneous models 489 14.9 Empirical results 493 14.10 References 498

PART V RISK MANAGEMENT 501

15 INNOVATION IN CORPORATE RISK MANAGEMENT: THE CASE OF CATASTROPHE RISK 503 Neil Doherty

15.1 Introduction 503 15.2 Why is risk costly to firms? 506 15.3 Globality, duality and four principle strategies 509 15.4 Catastrophe risk: insurance, reinsurance and

financial innovation 513 15.5 Innovation: market enhancement and technical efficiency 533 15.6 Some actual and potential strategies 536 15.7 References 538

Xli Table of Contents

16 ON CORPORATE INSURANCE Richard MacMinn and James Garven

16.1 Introduction 16.2 Basic model 16.3 Costly bankruptcy 16.4 Agency problems 16.5 Tax asymmetries 16.6 Concluding remarks 16.7 References

17 FINANCIAL RISK MANAGEMENT IN THE INSURANCE INDUSTRY J. David Cummins, Richard D. Phillips and Stephen D. Smith

17.1 Introduction 17.2 The rationale for corporate risk management: a survey of

recent literature 17.3 Corporate risk management: empirical evidence 17.4 Corporate hedging: multi period contacts, and private

information 17.5 Conclusion 17.6 References

18 LINKING INSURANCE AND MITIGATION TO MANAGE NATURAL DISASTER RISK Howard Kunreuther

541

541 544 548 550 557 560 562

565

565

570 575

579 586 589

593

18.1 Introduction 594 18.2 Insurers concern with insolvency 595 18.3 Linking mitigation with insurance 596 18.4 Role of building codes 604 18.5 Encouraging mitigation through other incentives and regulations 606 18.6 Future research directions 609 18.7 References 616

PART VI INSURANCE PRICING

19 APPLICATIONS OF FINANCIAL PRICING MODELS IN PROPERTY-LIABILITY INSURANCE J. David Cummins and Richard D. Phillips

19.1 Introduction 19.2 Insurance as risky debt

619

621

621 623

Table oj Contents Xlll

19.3 A simple CAPM for insurance pricing 625 19.4 Discrete time discounted cash flow (DCF) models 627 19.5 Option pricing models 633 19.6 Pricing CAT, call spreads and bonds 644 19.7 Continuous time discounted cash flow models 646 19.8 Conclusions 651 19.9 References 652

20 VOLATILITY AND UNDERWRITING CYCLES 657 Scott E. Harrington and Greg Niehaus

20.1 Introduction 657 20.2 The perfect markets model 659 20.3 Unexplained/predictable variation in underwriting results 662 20.4 Capital shocks and capacity constraints 669 20.5 Price cutting and soft markets 679 20.6 Regulatory influences 680 20.7 Conclusions 682 20.8 References 682

PART VII INDUSTRIAL ORGANIZATION OF INSURANCE MARKETS 687

21 ORGANIZATIONAL FORMS WITHIN THE INSURANCE INDUSTRY: THEORY AND EVIDENCE 689 David Mayers and Clifford W Smith

21.1 Introduction 689 21.2 Alternative organizational forms 691 21.3 Managerial discretion and alternative organizational forms 696 21.4 Corporate policy choices and organizational form 697 21.5 Organizational form and efficiency 704 21.6 Conclusions 705 21.7 References 705

22 INSURANCE DISTRIBUTION SYSTEMS 709 Laureen Regan and Sharon Tennyson

22.1 Introduction 709 22.2 Background 711 22.3 Direct writing versus independent agency 718 22.4 Agent compensation and resale price maintenance 730

22.5 The regulation of insurance distribution 737

XIV

23

24

25

Table of Contents

22.6 Concluding remarks 22.7 References

THE RETENTION CAPACITY OF INSURANCE MARKETS IN DEVELOPING COUNTRIES Jean-Fran~ois Outreville

23.1 Introduction 23.2 Measuring the economic importance of insurance market in

developing countries 23.3 The retention capacity of developing countries' markets 23.4 Market structure and the retention capacity 23.5 Comparative advantage and the retention capacity 23.6 A consolidated model explaining the retention capacity 23.7 Discussion 23.8 References

ANALYZING FIRM PERFORMANCE IN THE INSURANCE INDUSTRY USING FRONTIER EFFICIENCY AND PRODUCTIVITY METHODS 1. David Cummins and Mary A. Weiss

24.1 Introduction 24.2 The concepts of efficiency and productivity 24.3 Methodologies for estimating efficiency and productivity 24.4 Defining outputs and inputs 24.5 A survey of insurance efficiency research 24.6 Summary and conclusions 24.7 References

DEALING WITH THE INSURANCE BUSINESS IN THE ECONOMIC ACCOUNTS Tarek M. Harchaoui

25.1 Introduction 25.2 Insurance in the system of national accounts: a macroeconomic

approach 25.3 A microeconomic approach of the insurance business in the

system of national accounts 25.4 Concluding remarks 25.5 References

743 745

749

749

751 754 755 757 759 761 764

767

768 770 777 788 799 823 825

831

831

832

842 864 866

Table ot' Contents

PART VIII LIFE INSURANCE, PENSIONS AND ECONOMIC SECURITY

26 DEVELOPMENTS IN PENSIONS Olivia S. Mitchell

26.1 Introduction 26.2 How do pensions influence the risks workers confront

26.3

26.4 26.5 26.6 26.7

in retirement? What explains the global trend toward defined contribution pension plans? What other factors influence the market for pensions? How can pension performance be judged? Looking ahead References

27 LIFE INSURANCE Bertrand Villeneuve

27.1 Introduction 27.2 Possibilities and needs in life insurance 27.3 A contract theory of life insurance 27.4 Conclusion 27.5 References

28 THE DIVISION OF LABOR BETWEEN PRIVATE AND SOCIAL INSURANCE Peter Zweifel

xv

871

873

873

877

881 886 890 894 898

901

901 903 918 926 928

933

28.1 Introduction and overview 933 28.2 Factors determining the division of labor between private

and social insurance 935 28.3 The four challenges confronting the existing division of labor 949 28.4 Improving the division of labor between private and social

insurance 953 28.5 Suggestions for an improved division of labor between private

and social insurance 28.6 References

INDEX

960 964

967

CONTRIBUTING AUTHORS

Pierre-Andre Chiappori University of Chicago, USA pchiappo@midway.uchicago.edu

Keith 1. Crocker University of Michigan, USA kcrocker@umich.edu

1. David Cummins University of Pennsylvania, USA cummins@wharton.upenn.edu

Patricia Danzon University of Pennsylvania, USA danzon@wharton.upenn.edu

Georges Dionne Ecole des Hautes Etudes Commerciales, Canada georges.dionne@hec.ca

Neil Doherty University of Pennsylvania, USA doherty@wharton.upenn.edu

Louis Eeckhoudt Faculte Universitaire Catholique de Mons, Belgium eeckhoud@messagefucam.ac.be

Nathalie Fombaron Universite de Paris X-Nanterre, France nathaliefombaron@u-parisl Ofr

Bernard Fortin Universite Laval, Canada bfor@ecn.ulaval.ca

James Garven Louisiana State University, USA jim@garven.com

Christian Gollier Universite de Toulouse I, France gollier@cictfr

Tarek M. Harchaoui Statistique Canada, Canada harctar@statcan.ca

Scott Harrington University of South Carolina, USA scottnet@darla.badm.sc.edu

Howard Kunreuther University of Pennsylvania, USA kunreuther@wharton.upenn.edu

Paul Lanoie Ecole des Hautes Etudes Commerciales, Canada paul.lanoie@hec.ca

Henri Louberge Universite de Geneve, Switzerland henri.louberge@ecopo.unige.ch

Mark 1. Machina University of California, San Diego, USA mmachina@weber.ucsd.edu

Richard MacMinn University afTexas, USA macminn@mail.utexas.edu

xv 111 Contributing Authors

David Mayers University of California. Riverside. USA david. mayers@ucr.edu

Olivia S. Mitchell University of Pennsylvania. USA mitchelo@wharton.upenn.edu

Greg Niehaus University of South Carolina gregn@darla.badm.sc.edu

Jean-Fran90is Outreville Office des Nations Unies, Switzerland }outreville@unog.ch

Richard D. Phillips Georgia State University. USA rphillips@gsu.edu

Pierre Picard Universite de Paris-X Nan terre. France pierre.picard@u-paris1 Ojr

Jean Pinquet Universite de Paris-X Nanterre, France pinquet@u-paris10jr

Laureen Regan Temple University, USA lregan@thunder.ocis.temple.edu

Harris Schlesinger University of Alabama, USA hschlesi@cba.ua.edu

Clifford W. Smith University of Rochester, USA smith@simon.rochester.edu

Stephen D. Smith Georgia State University, USA sdsmith@gsu.edu

Arthur Snow University of Georgia. USA snow@rigel.econ.uga.edu

Sharon Tennyson Cornell University, USA st96@cornell.edu

Bertrand Villeneuve Universite de Toulouse 1, France villeneu@cictjr

Mary A. Weiss Temple University. USA mweiss@thunder.temple.edu

Ralph Winter University of Toronto, Canada rwinter@chass.utoronto.ca

Peter Zweifel Zurich University. Switzerland pzweifel@sozoec.unizh.ch

D. Arnaud Universite d'Angers, France

R. Arnott Boston College, USA

M.M. Boyer

REFEREES

R.A. Devlin University of Ottawa, Canada

G. Dionne Ecole des Hautes Etudes Commerciales, Canada

Ecole des Hautes Etudes Commerciales, N. Doherty Canada University of Pennsylvania, USA

P. Boyle L. Eeckhoudt University of Waterloo, Canada Universite Catholique de Mons,

Belgium

R. Butler University of Minnesota, USA

B. Caillaud Ecole Nationale des Ponts et Chaussees, France

P.A. Chiappori University of Chicago, USA

A.P. Contandriopoulos Universite de Montreal, Canada

K. Crocker University of Michigan, USA

D. Cummins University of Pennsylvania, USA

M. Crouhy CIBC, Canada

K. Dachraoui

C. Fluet Universite du Quebec a Montreal, Canada

N. Fombaron Universite de Paris X-Nanterre, France

B. Fortin Universite Laval, Canada

R. Gagne Ecole des Hautes Etudes Commerciales, Canada

1. Garven Louisiana State University, USA

H. Geman ESSEC, France

C. Gollier Universite de Toulouse I, France

Ecole des Hautes Etudes Commerciales, A. Gron Canada Northwestern University, USA

xx

T. Harchaoui Statistique Canada, Canada

S. Harrington University of South Carolina, USA

1. Hendel Princeton University, USA

M. Huberman Universite de Montreal, Canada

M. Hoy University of Guelph, Canada

B. Jullien Universite de Toulouse I, France

E. Karni The Johns Hopkins University, USA

A. Kleffner University of Calgary, Canada

R. Klein Georgia State University, USA

H. Louberge Universite de Geneve, Switzerland

M. Machina University of California, San Diego, USA

D. Mayers

Referees

P. Ouellette Universite du Quebec a Montreal, Canada

J.F. Outreville Office des Nations Unies, Switzerland

M. Pauly University of Pennsylvania, USA

S. Perelman Universite de Liege, Belgium

J.E. Pesando University of Toronto, Canada

L. Posey Penn State University, USA

R. Puelz Southern Methodist University, USA

B. Salanie INSEE, France

e. Smith University of Rochester, USA

S. Spaeter Universite de Strasbourg, France

P.K. Trivedi

University of California, Riverside, USA Indiana University, USA

J. Meyer Michigan State University, USA

O.S. Mitchell University of Pennsylvania, USA

B. Venard ESSCA, France

J.e. Vergnaud Universite de Strasbourg, France

Referees XXI

B. Villeneuve M. Weiss Universite de Toulouse I, France Temple University, USA

D. Vittas D. Zajdenweber The World Bank, USA Universite de Paris X-Nanterre, France

PREFACE Denis Kessler

President of the Federation Fram,:aise des Societes d' Assurances

As the new century begins, the question of risk and theories as to how it should be understood and measured are at the forefront. We know that risks evolve, and that today's risks are different from those of the past. The characteristics, frequency, intensity and, above all, the very nature of risks are changing radically.

If risks are becoming increasingly complex, it is because they are related to tech­nological developments, new economic and social activities, and the aggregate effects of multiple factors. At the same time, risks have become more integrated, since a given event may have a series of consequences, giving rise to yet other risks. Risks are often interrelated, and sometimes the combined effect is greater than the sum of the indi­vidual parts. Smoking poses a risk, contact with asbestos poses another, and we know that the combined effect of these two risks does not merely increase the threat to health incrementally, but also exponentially.

The risks of today and tomorrow are also becoming more foreseeable as knowl­edge increases and we upgrade our statistical databases. This trend poses a major challenge to the insurance industry. Advances in genetics and more sophisticated knowledge of weather patterns, for example, have an impact on insurance techniques.

Risks are also becoming more endogenous to the behaviors of economic agents as we improve our understanding of such risks, as well as of the aggravating factors or, conversely, the preventive measures whose efficacy can be measured. A pedestrian who is hit on the head by a falling flowerpot is not responsible for his misfortune. A smoker, on the other hand, engages in behavior that greatly increases his likelihood of contracting a number of health problems. The risk run by the smoker can be quan­tified and priced.

Finally, the new risks tend to unfold more gradually, are more spread out in time and space, and often surface well after the causal event. This is true of environmen­tal impairment, health risks and long-term care. Some risks are long-lasting or irre­versible, making the analysis of appropriate preventive measures a complex matter. This leads us to the now famous "principle of precaution."

Given the changing face of risks, insurance techniques must evolve radically, and a new paradigm of risk and uncertainty must be constructed.

At the end of World War II, the rise of game theory and the first precepts of economic uncertainty theory initiated research in the field, laying the theoretical foundations that will be vital to the practice of insurance tomorrow. In the 1960's and 1970's, theoretical work in these areas made spectacular strides possible, in

XXIV Preface

the form of sophisticated economic risk models and the creation of new financial instruments.

This theoretical body of work, enriched by developments to come, should be applied to the science of risk. Broken down into their smallest component parts, these risks can be hedged by relatively simple instruments, such as the "plain vanilla" options of finance.

New and better-targeted insurance products can be designed to better meet the needs and expectations of economic agents while more effectively dealing with the issues of moral hazard and adverse selection. Consequently, the insurability of risks will expand significantly.

However, this conceptual progress will not translate into better pricing of risks in general, and of previously uninsurable risks in particular, unless practical risk man­agement tools and the means of relaying information are also improved. The issue of access to information needs to be resolved, as does the question of methods used to monitor both insureds and claims.

Naturally, these changes will affect the organization and processes of the insur­ance industry. We are likely to see broader integration of the various segments of the insurance offer, as the development of new risk management techniques further blurs the lines between various specialties and between insurance products.

In the face of new risks, and given our ability to better understand known risks, the nature of insurance demand is bound to change. It will be more closely aligned with the desire on the part of economic agents for protection against risks, which appears to increase with the level of democracy and wealth.

We can expect a strong rise in the demand for insurance on the part of corpora­tions, which is strongly correlated to their ability to generate lasting profit. Private individuals are expected to demand more personalized products and greater freedom of choice in terms of the insurance products available to them. It is very likely that the days of cumbersome, constraining and uniform offers in health insurance, for example, are over.

This transformation of risks and the attendant demand for insurance will ulti­mately and necessarily lead to major changes in the insurance industry. This will be seen on the level of product distribution as well as organizational and strategic choices, in particular the relationship with banking services.

But this transformation in the nature of risks and insurance demand will also affect pricing policy and risk management in insurance companies. Here again, the development of new risk management and diversification techniques through rein­surance or direct access to the financial markets should offer insurers more options in terms of portfolio management, which will benefit end customers.

The new generation of derivative products, indexed to meteorological data for example, lays the foundations of the risk management tools that will support this enlarged insurance offer. Not only do such products allow for better coverage of risks whose complexity or magnitude previously rendered them all but uninsurable, they

Preface xxv

also contribute to optimizing insurance company investments, asset liability manage­ment and accounting for underwriting cycles.

The work accomplished by Georges Dionne and the authors who contributed to this project is of vital importance to anyone who is interested in the development of insurance. It lays the foundations of a new knowledge base-practically a new disci­pline: that of the science of risks, which is likely to lie at the heart of our future.

INTRODUCTION

It was the article "Uncertainty and the Welfare Economics of Medical Care" by Kenneth Arrow (American Economic Review, 1963) that first drew my research atten­tion to risk, uncertainty, insurance, and information problems. This article proposed the first theorem showing that full insurance above a deductible is optimal when the premium contains a fixed-percentage loading, provided there are no information prob­lems. It also suggested economic definitions of moral hazard and adverse selection. It generated many doctoral dissertations, my own included.

During the 1970s, different contributions proposed theorems regarding optimal insurance coverage, security design, and equilibrium concepts for situations with imperfect information. The 1980s were characterized by several theoretical develop­ments such as the consideration of more than one period; of many contracting agents; of multiple risks; of non-expected utility; of commitment; and of several information problems simultaneously. Other economic and financial issues such as underwriting cycles, price volatility, insurance distribution, liability insurance crisis, and retention capacity were addressed by academics and practitioners during that period. Hierarchi­cal relationships in firms and organizations and organizational forms were also studied, along with the pricing and design of insurance contracts in the presence of many risks.

The empirical study of information problems became a real issue in the 1990s. These years were also marked by the development of financial derivative products and large losses due to catastrophic events. The last months of 1999 were again cat­astrophic for South America and Europe. Alternatives to insurance and reinsurance coverages for these losses are now currently being proposed by financial markets.

The aim of this book is to provide a reference work on insurance for professors, researchers, graduate students, regulators, consultants, and practitioners. It proposes an overview of current research with references to the main contributions in different fields. It contains twenty-eight chapters written by thirty-five collaborators who have produced significant research in their respective domains of expertise. It can be considered as a complement to the previous books I edited for the S.S. Huebner Foundation of Insurance Education in 1992: Foundations of Insurance Economics­Readings in Economics and Finance (with S. Harrington) and Contributions to Insur­ance Economics.

Each chapter is presented with an abstract and keywords and each can be read independently of the others. They were (with very few exceptions) reviewed by at least two anonymous referees.

HISTORY AND FOUNDATIONS OF INSURANCE THEORY

The first chapter is concerned with history. H. Louberge relates the evolution of insurance research since 1973. One important message from this contribution is

xxv III Introduction

that the significant developments of insurance economics during the last 25 years are exemplified by those in the economics of risk and uncertainty and in financial theory.

We next turn to the foundations of insurance theory in th;: absence of informa­tion problems. M. Machina's chapter investigates whether or not some classical results of insurance theory remain robust despite "departures from the expected utility hypothesis." His analysis covers insurance demand; deductible and co-insurance choices; optimal insurance contracts; multilateral risk-sharing agreements; and self-insurance vs self-protection. The general answer to the above question is posi­tive although other restrictions are necessary since the technique of "generalized expected utility analysis" is broader than that of the classical, linear expected-utility model.

C. Gollier concentrates on comparisons among optimal insurance designs. He shows that three significant results can be obtained without the restriction of linear expected utility: (1) at least one state of the world is without insurance coverage; (2) the indemnity schedule is deterministic; and (3) the optimal contract contains a straight deductible. However, the hypothesis of linear expected utility generates addi­tional results when transaction costs are nonlinear.

The ways in which changes in risk affect optimal-decision variables is a difficult and elusive research topic. The major problem is that risk aversion is not sufficient to predict that a decision-maker will reduce his optimal risky activity (or increase his insurance coverage) if an exogenous increase in risk is made in the portfolio. Usually, strong assumptions are needed regarding the variation of different measures of risk aversion or regarding distribution functions, in order to obtain intuitive compara­tive static results. C. Gollier and L. Eeckhoudt increase the level of difficulty by adding a background risk to the controllable risk. They propose restrictions on first- and second-order stochastic dominance to obtain the desired results. They also consider restrictions on preferences.

H. Schlesinger has contributed to many articles on market insurance demand, particularly as related to deductible insurance. He first presents the classical results related to changes in optimal coinsurance and deductible insurance with respect to initial wealth, loading (price), and risk aversion. Comparisons with self-protection and self-insurance are given and the basic models are extended to account for default and background risks.

ASYMMETRIC INFORMATION

The book then moves on to asymmetric information problems, which have often been introduced into economics and finance journals through examples of insurance allo­cation problems. Two sections ofthe book are devoted to this subject. The first reviews the main results related to ex-ante and ex-post moral hazard (fraud), adverse selec-

Introduction XXIX

tion, liability insurance, and risk classification. The second studies the empirical significance of these resource allocation problems.

R. Winter extends his 1992 survey by presenting the development of optimal insurance under moral hazard over the past twenty-five years. He shows how the insur­ance context manages to introduce some structural devices. For example, optimal insurance contracts vary when effort affects the frequency (deductible) rather than the severity (coinsurance above a deductible) of accidents. The author also discusses dynamic contracts and contract renegotiation.

The chapter by G. Dionne, N. Doherty, and N. Fombaron proposes an extension of Dionne and Doherty (1992). Many new subjects are added to the classical one­period models of Stiglitz (Monopoly) and Rothschild and Stiglitz (Competition). Much more attention is paid to the recent developments of multi-period contracting. A section on the endogenous choice of types before contracting was added and another one treats moral hazard and adverse selection simultaneously. Finally, the last section covers various new subjects related to adverse selection: risk categorization and re­sidual adverse selection; various types of risk aversion; incomplete symmetrical infor­mation; principals better informed than agents; uberrima fides and adverse selection with multiple risks.

The literature of risk classification was strongly influenced by K. Crocker and A. Snow. Risk classification may increase efficiency when certain conditions are met but it may also introduce adverse equity in some risk classes. The authors revise the theory of risk classification in insurance markets and discuss in detail its implications for efficiency and equity. They show how the adverse equity consequences of risk clas­sification are related to economic efficiency through their treatment of the social cost of risk classification.

S. Harrington and P. Danzon study the basic relationships between liability law, liability insurance, and loss control. They study what implications limited wealth and liability have for the demand for liability insurance and accident deterrence. They discuss many other subjects such as correlated risks and liability insurance markets; liability insurance contract disputes; tort and liability insurance crises in the 1980s; and the efficiency of the U.S. tort liability/liability insurance system.

Insurance fraud is now a significant resource-allocation problem in many coun­tries. It seems that traditional insurance contracts are not efficient to control this problem. In fact, there is a commitment issue involved, since audit costs may become quite substantial for different claims. P. Picard surveys the recent development of two types of models: costly state verification and costly state falsification. In the second type, the insured may use resources to modify the claims, whereas in the first he simply lies. Other subjects include adverse selection; credibility constraints on anti-fraud policies; and collusion between policy-holders and insurers' agents.

The empirical measurement of information problems is a recent research topic. Many issues are considered in the two chapters written by PA. Chiappori (11) and G. Dionne (12). P.A. Chiappori puts the emphasis on empirical models that test for

xxx Introduction

or evaluate the scope of asymmetric information in the insurance relationship, whereas G. Dionne discusses insurance and other markets such as labour and used cars. P.A. Chiappori suggests that empirical estimation of theoretical models requires precise information on the contract: information available to both parties on performance and transfers. He provides many examples of articles that approximate such conditions. G. Dionne concentrates his review on adverse selection and moral hazard in different markets. He concludes that efficient mechanisms seem to reduce the theoretical dis­tortions due to information problems and even eliminate some residual information problems. However, this conclusion is stronger for adverse selection. One explanation is that adverse selection is related to exogenous characteristics, while moral hazard is due to endogenous actions that may change at any point in time. Finally, he shows how some insurance contract characteristics may induce insurance fraud!

B. Fortin and P. Lanoie review the major contributions on workers compensation, focusing on empirical measurement of the incentive effects of different workers com­pensation regimes. They also discuss the theoretical issues raised concerning the effects of such insurance on the individual's behaviour. They show how workers compensation can influence the frequency, duration, and nature of claims. They also examine what impact workers compensation has on wages and productivity. Finally, they show how workers compensation can be a substitute for unemployment insur­ance, a subject on which they have contributed in the literature.

The last paper on the empirical measurement of information problems presents statistical models of experience rating in automobile insurance. 1. Pinquet shows how predictions on longitudinal data can be performed via a heterogeneous model. He also offers consistent estimations for numbers and costs of claims distribution. Examples are given for count-data models and empirical results from the portfolios of insurers in France are presented.

RISK MANAGEMENT AND INSURANCE PRICING

Risk management in insurance is now linked to the financial management of differ­ent risks. N. Doherty points out that the recent financial innovations in managing catastrophe risk may be interpreted as a response to the problem of insurance and reinsurance capacity brought on by the catastrophes having occurred over the last ten years. His chapter starts by showing why risk is costly to firms. The structure developed shows how reinsurance, financial instruments, insurance policy design, leverage management and organizational structure are linked to managing the different risks.

The role of corporate insurance demand has not received much attention in the literature, although we observe that insurance contracts are regularly purchased by corporations and do have their importance in the management of corporate risk. The model developed by R. MacMinn and 1. Garven focuses on the efficiency gains

Introduction XXXI

of corporate insurance to reduce bankruptcy costs, agency costs, and tax costs. In fact, insurance is simply another risk management tool much like corporate hedging.

D. Cummins, R. Phillips, and S. Smith survey the finance literature on corporate hedging and financial-risk management and show how it applies to insurance. They also present empirical results on corporate hedging. They then develop a theoretical model to explain why insurers manage risk. They emphasize that the main motiva­tion is to avoid shocks to capital that may trigger liquidations. The chapter by H. Kunreuther examines the role of insurance in managing risks from natural disasters, by linking insurance to cost-effective risk mitigation measures. The author outlines the roles that private markets (financial, institutions, and real estate developers) and municipalities can play in encouraging the adoption of cost-effective risk mitigation measures.

We then attack insurance pricing. Two complementary chapters treat this subject: the first discusses financial-pricing models, while the second introduces underwriting cycles. D. Cummins and R. Phillips propose a comprehensive survey of financial pricing for property-liability insurance and propose some extensions to existing models. Financial-pricing models are based on either the capital-asset pricing model; the intertemporal capital-asset pricing model; arbitrage theory or option pricing. Also presented are approaches using internal rate of return and insurance derivatives such as catastrophic-risk-call-spread and bonds.

After reviewing evidence that market insurance prices follow a second-order autoregressive process, S. Harrington and G. Niehaus present different theories that try to explain the cyclical behaviour of insurance prices. Capital shocks may explain periods of high insurance prices, while moral hazard and/or winners-curse effects can explain periods of low insurance prices. The potential effects of price regulation are also summarized.

INDUSTRIAL ORGANIZATION

The section on the industrial organization of insurance markets starts off with the two researchers who have most influenced this area of research, D. Mayers and C. Smith. They stress the association between the choice of organizational structure and the firm's contracting costs. They analyze the incentives of individuals involved in the three major functions of insurance firms: the executive function, the owner function, and the customer function. They also examine evidence on corporate-policy choices by alternative organizational structures: executive compensation policy; board com­position; choice of distribution systems; reinsurance decisions; and use of participat­ing policies. The relative efficiency of different organizational forms are reviewed.

Insurance distribution systems are analyzed by L. Regan and S. Tennyson. Their chapter focuses on three major economic issues: (1) the choice of distributive system(s) by an insurer; (2) the nature of insurer-agent relationships; and (3) the

xxxii Introduction

regulation of insurance distribution activities, including regulation of entry and dis­closure of information to the consumer. IF. Outreville studies the retention capacity of insurance markets in developing countries. He analyzes which factors may affect the aggregate-retention capacity of a country and provides statistical results obtained from a data base published by the United States Conference on Trade and Develop­ment of the 1988-1990 period.

Measuring the efficiency and productivity of financial firms is very difficult, since the definitions of output are multidimensional. Cummins and Weiss review the basics of modern frontier methodologies; discuss input and output measurement for insur­ers; and review the significant contributions made on these topics. As pointed out by the authors, modern frontier efficiency and productivity methodologies represent the state of the art in measurement of firm performance. Measures of efficiency and pro­ductivity based on these methods are useful in testing economic hypotheses about market structure, corporate governance, organizational form and other important topics. The measurement of efficiency and productivity also is useful in informing regulators about the firm's performance and in comparing performance across countries. So, they hope that more economists will use these methodologies for the insurance industry. The contribution of T. Harchaoui reviews the treatment of the insurance business in the system of national accounts, with a focus on measurement of productivity analysis. He first shows that the macroeconomics approach is very limited in many aspects. A more desaggregated approach allows for better under­standing of the delineation of insurers' lines of business; the measurement of their activity; and their interactions with the economy.

LIFE INSURANCE, PENSIONS, AND ECONOMIC SECURITY

The book ends with life insurance, pensions, and economic security. It is well known that pension institutions face difficult years. One development, studied by 0. Mitchell, is that defined contribution plans are now very popular, often at the expense of pension benefits. This changes the risk and rewards for participants and for government regulators. The expenses associated with pension management have become an im­portant issue. For the author, reforms will be necessary to restore government social security programs to solvency.

B. Villeneuve analyses the micro-foundations of life insurance markets. He starts with the well-known life-cycle hypothesis and builds on contract theory to highlight the main issues in life insurance design. He shows how the trade-off between flexi­bility and opportunistic behaviour is an equilibrium outcome of actual life insurance contracts.

P. Zweifel proposes two types of reasons for the existence and growth of social insurance: (I) possible enhancements of efficiency and (2) public choice related to the interests of governments and politicians. Empirical evidence suggests that the second

Introduction XXXlll

one is significant to explain the choice between private and social insurance. It also shows that individuals in the United States and Gennany are subject to excess asset variance. Four proposals to improve the interplay between private and social insur­ance are fonnulated.

ACKNOWLEDGEMENTS

I wish to acknowledge all the authors and the referees for their significant contribu­tions. The preparation of this book would not have been possible without the gener­ous collaboration of Claire Boisvert who managed all the correspondence and spent many hours on all stages of the production process. I also thank Rose Antonelli and Jill Strathdee from Kluwer for their efficient collaboration. The preparation of the book was financed by the Buebner Foundation of Insurance Education, BEC-Mon­treal, the Federation Franfaise des Societes d'Assurance (FFSA), FCAR (Quebec), and CRSB (Canada).

Georges Dionne Risk Management Chair

BEC-Montreal