The Economic Value

of Information
Springer Science+ Business Media, LLC
David B. Lawrence

The Economic Value

of Information

With 45 Figures

Springer
David B. Lawrence
College of Business and Public Administration
Drake University
Des Moines, IA 50311
USA

Library of Congress Cataloging-in-Publication Data

Lawrence, David B., 1929-
The economic value of infonnation / David B. Lawrence.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-4612-7152-9 ISBN 978-1-4612-1460-1 (eBook)
DOI 10.1007/978-1-4612-1460-1
1. Business infonnation services. 2. Infonnation resources
management-Economic aspects. 1. Title.
HF54.5.L38 1999
658A'038-dc21 98-51179

Printed on acid-frec paper.

© 1999 Springer Science+Business Media New York

Originally published by Springer-Verlag New York Berlin Heidelberg in 1999
Softcover reprint of the hardcover I st edition 1999
Ali rights reserved. This work may not be translated or copied in whole or in part without the written
pennission of the publisher ( Springer Science+Business Media, LLC ), except for brief excerpts in
connection with reviews or scholarly analysis. Use in connection with any fonn of infonnation
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed is forbidden. The use of general descriptive names,
trade names, tradernarks, etc., in this publication, even if the fonner are not especially identified, is
not to be taken as a sign that such names, as understood by the Trade Marks and Merchandise Marks
Act, may accordingly be used freely by anyone.
Production managed by Allan Abrams; manufacturing supervised by Jacqui Ashri.
Photocomposed copy prepared using the author's Microsoft Word files.

9 8 765 432 I

ISBN 978-1-4612-7152-9
To Gail, Sarah, and Dan
Preface

The Scope of This Book

Popular culture often refers to current times as the Information Age, classifying
many of the technological, economic, and social changes of the past four deca:les
under the rubric of the Information Revolution. But similar to the Iron Age be-
fore it, the description "Information Age" suggests the idea that information is a
commodity in the marketplace, one that can be bought and sold as an item of
value. When people seek to acquire information yet complain about information
overload, and when organizations invest millions in information systems yet are
unable to pinpoint the benefits, perhaps this reflects a difficulty with the as-
sessment of the value of this commodity relative to its cost, an inability to dis-
cern the useless from the useful from the wasteful. The Information Age requires
us to assess the value, cost, and gain from information, and to do it from several
different viewpoints.
At the most elementary level is the individual who perceives a need for in-
formation-her current state of knowledge is insufficient and something needs to
be understood, or clarified, or updated, or forecast. There is a universe of al-
ternative information sources from which to choose, some more informative
than others, some more costly than others. The individual's problem is to evalu-
ate the alternatives and choose which sources to access.
An organization comprising many information-seeking employees and agents
must take a somewhat broader viewpoint. The organization's problem is to de-
sign and manage a system for information that can meet the needs of diverse
groups of individuals as those needs arise, while maximizing the organization's
net gain from the services the system provides.
This is a book about the evaluation and choice of information sources by in-
dividuals and the design and management of information systems by organiza-
tions. These topics are unified by the thesis that both information sources and
information systems are valuable to the extent they contribute to better decision
making. Since the incorporation of information is costly in terms of money,
time, and effort, optimal information use and system design involves classic
economic tradeoffs between value and cost. This book studies the determinants of
the value and cost of information, both to the individual and to the organization,
provides techniques for the assessment of the value of information and the com-
viii Preface

parison of informativeness among alternative sources, and presents principles for

the optimal design and management of information systems.
Decision theory is a ready-made methodology for assessing the economic
value of information. Since many activities can be cast into a decision making
framework, applications of decision theory are widespread, ranging from the no-
torious [the Ford Pinto gas tank decision, Gioia (1992)] to the nutritious [the
foraging decisions of bumblebees, Real (1993)]. Applications of decision theory
to the evaluation of information appear in such diverse scientific disciplines as
accounting, economics, engineering, environmental science, geology, informa-
tion science, management information systems, medical science, meteorology,
operations research, psychology, and statistics. As Chapter 1 demonstrates,
nowadays the applications are becoming more multidisciplinary in nature-
decisions to study the banning of a potential carcinogen or the consequences of
global warming necessarily require contributions from many fields.
Unfortunately, much of the existing literature tends to be highly specialized
and technical, with little notational uniformity. Even basic words such as data,
knowledge, information, and information system have different meanings in dif-
ferent contexts. An integrative understanding of the many aspects of information
evaluation, choice, and system design requires consistency and precision in the
definition of terms. One of the purposes of this book is to provide students and
researchers with a unified and coherent notation and approach to this diverse mul-
tidisciplinary literature.
Research described in Chapter 9 indicates that as individuals and teams face
more complex and multidisciplinary decision problems, the role of the informa-
tion system becomes more critical. It is my belief that decision theory has not
yet achieved its potential as a tool for corporations, governments, and other or-
ganizations to use when evaluating their information systems. One of the goals
of this book is to improve the integration of the decision-theoretic approach to
information value with the knowledge from information science on the design,
management, and cost of cooperative information systems.
This book concentrates on a subset of a vast literature known collectively as
the economics of information. The focus is on the evaluation and incorporation
of information by a single decision maker, and on the cooperative facilitation of
that task by an organization; with only a few exceptions, strategic and noncoop-
erative uses of information, along with the market consequences of differential
information, are excluded.

A Reader's Guide

The basic plot of this book is to move from the assessment of the economic
value of a specific source, to the choice of source or sources, to the design of an
organization's system for information. Although the mathematical prerequisite
 Preface ix

for the material is no more than undergraduate courses in calculus and probability
theory, there may be different ways to read this book depending upon the reader's
background and interest. A student who desires to learn decision theory and its
application to the evaluation of information will find Chapters 1 through 4, Sec-
tions 5.1, 5.2, 5.4, 6.1, 6.2, 7.1, 7.3, and 8.1 to be most helpful. The reader
interested primarily in information systems will find Chapters 1 and 9 most per-
tinent. Indeed, the book is written so that those interested primarily in informa-
tion science can skip directly from Chapter 1 to Chapter 9 without loss, picking
up the material in between as need be. Researchers who are already familiar with
decision theory and who are interested in a particular application or in the more
theoretical aspects of information value will find this in Chapters 5 through 8.
The primarily technical and mathematical material of this book is constrained to
these four chapters; there is a detailed symbol glossary beginning on page 365 to
help with the notation. Subsections that are marked with an asterisk (*) contain
either peripheral or highly technical material that can be bypassed without loss
of continuity.
The ideas and techniques are illustrated throughout the book by means of ex-
amples which are set off from the text by black diamonds (.). Since they may
continue through several chapters, the examples are numbered according to the
section in which they first appear: Example 4.2 first appears in Section 4.2.
Each example hones in on the substance of the most recently introduced topic;
the purpose is to fix ideas and not to provide immediate generality. The proxi-
mate text, however, presents citations for models and theories that offer more
general, technical, complex, or realistic applications.
The book holds to a number of conventions throughout. The first definition
of an important term is indicated by placing that term in italics. In equations,
parentheses 0 always denote functional dependence, and brackets [] and braces {}
always indicate separation for arithmetic operation. Finally, the protagonist in
this book is the decision maker. The decision maker is considered to be a person
and it is convenient to use a pronoun for reference; in many languages such pro-
nouns are indicative of gender. We can use that to our advantage by adopting the
following convention from the principal-agent literature: in this book the deci-
sion maker is the principal and so is always referred to as "she," and the informa-
tion source, when human, is the agent and so is always "he."

Acknowledgments
In the several years it has taken to write this book, many individuals have pro-
vided me with valuable information, advice, and encouragement. lowe special
gratitude to Stuart Klugman, along with Newton Bowers, John Rozycki, Rahul
Parsa, C. Kenneth Meyer, Harry Wolk, Rick Trieff, and Brent Stuart. Numerous
researchers have contributed to the pure theory of the economic value of informa-
x Preface

tion, but because of depth, rigor, and clarity, in my view the greatest contributor
is Jacob Marschak (1974). "It's all in Marshall" students of microeconomic the-
ory often say; students of the economics of information and organization can
equally say, "It's all in Marschak!" I would also like to thank Karl A. Fox, who
introduced me to the topic, Anne Mayere of ENS SIB , whose 1995 conference on
the economics of information stimulated me in several directions, and Drake
University, whose generous sabbatical policy gave me the time to do this work.
Special appreciation and love go to Gail, Sarah, and Dan, who tolerated a lot and
to whom this book is dedicated. Finally, there are the people who put me on the
right track from the very beginning: Alpha, Louise, George, Jackie, George,
Annette, and Doug. Thank you.

David B. Lawrence
Des Moines, Iowa
August, 1998
Contents

1. Introduction and Overview

1.1 Information in Decision Making
1.1.1 The Basic Framework for Decision
1.1.2 Information and Information Sources 2
1.1.3 Statistical and Pragmatic Information 5
1.1.4 Ex-Post Value and Gain 6
1.2 Information in the Organization 7
1.2.1 Information Systems 7
1.2.2 User Information Processing 8
1.2.3 Information System Accounting II
1.3 Economics of the Incorporation of Information 12
1.3.1 The Criterion 12
1.3.2 Ex-Ante Components of Decision Making 13
1.4 The Expected Value of an Information Source 15
1.4.1 Information Structure and Informativeness 15
1.4.2 The Measurement of Information Value:
Chapters 2 and 3 16
1.4.3 The Assessment of Information Value:
Chapters 4 and 5 18
l.4.4 Case Studies of Information Value 20
1.4.5 The Theory of Information Value: Chapters 6 and 7 25
1.5 Information Value, Cost, and Procurement 28
1.5.1 The Optimal Information Structure 28
l.5.2 Procuring Direct Information 32
1.5.3 Case Studies of Information Cost, Value, and Choice 33
1.6 System Design for Information Gain 35
1.6.1 Decision Theory, Information Science,
and System Design 35
1.6.2 Information Use Environments 37
1.6.3 A Model of System Design and User Behavior 38
1.7 Information Social Science 39
1.7.1 The Information Society 39
1.7.2 How Monumental Are Recent Events? 41

2. The Value of the Informed Decision 45

2.1 Elements of a Decision Problem Under Uncertainty 45

Xl
xii Contents

2.1.1 Utility of Outcome 45

2.1.2 Decision Problems 46
2.1.3 Decision Making 49
2.2 The Framing of the Decision Problem 53
2.2.1 Practical Issues in Framing 53
2.2.2 Framing the State Space and Initial Knowledge 55
2.2.3 The Incorporation of Information 57
2.3 Useful Facts About Statistical Information 58
2.3.1 Statistical Properties of the Information Structure 58
2.3.2 The Finite Model 59
2.3.3 Some Specific Expectations 60
2.4 Value of the Informed Decision 65
2.4.1 Extensive Form Analysis 65
2.4.2 Perfect Information 69
2.4.3 Worthless Information 70

3. Measures of the Value of Information 73

3.1 Measures of the Value of a Message 73
3.1.1 Cash-Equivalent Values of a Decision 73
3.1.2 Value of Decisions Posterior to the Message 78
3.1.3 The Simplest Special Case 81
3.1.4 Two Examples 82
3.1.5 The Need for Preposterior Evaluation 88
3.2 Measures of the Value of a Source 89
3.2.1 Incremental Value from Incorporating Information 89
3.2.2 The Range of the Expected Value of Information 91

4. The Assessment of Statistical Information 95

4.1 Coherent Assessment of Probability Distributions 95
4.1.1 Coherence and Consistency 95
4.1.2 Families of Probability Densities 97
4.2 Assessment of Beliefs and Foreknowledge 100
4.2.1 Assessing a Univariate Distribution 100
4.2.2 Obtaining Qualitative Foreknowledge of Information 103
4.2.3 Expert Resolution 106
4.2.4 Analysis of a Track Record 107
4.3 Simplifying the Assessment 113
4.3.1 The Preposterior Mean 114
4.3.2 The Calibrated Noiseless Model 115

5. Models with Convenient Assessment and Interpretation 117

5.1 Models with Payoff Quadratic in the Action 117
5.1.1 The General Case 118
 Contents X1lI

5.1.2 The Bivariate Normal Case 119

*5.1.3 The Quadratic Team 121
5.2 Models with Payoff Linear in the State 122
5.2.1 Raiffa and Schlaifer's Approach 123
5.2.2 Application to Statistical Decision Problems 127
5.2.3 Application in a Dichotomy 131
5.2.4 The Inventory Problem 133
5.3 Models with Concave-Exponential Utility 136
5.3.1 Information Value with Separable Outcome 136
5.3.2 Applications with the Bivariate Normal Information
Structure 138
5.4 Models with Nonseparable Outcome; Betting and
Investing Models 141
5.4.1 Nonseparability and Information Value:
Two Examples 141
5.4.2 Arrow's Contingent Securities Model 145
5.5 Models with Multicategorical State Description 147
5.5.1 Multicategorical Information Structures 148
5.5.2 Bidding Models 148
5.6 Dynamic Models and the Role of Timeliness 155
5.6.1 Dynamic Decision Problems 156
5.6.2 Delay 159
5.6.3 The Value of Timeliness and Accuracy 160
*5.6.4 Assessing the Joint Distribution of Messages and
States Via the Kalman Filter 163
5.6.5 Stochastic Control Theory 166
5.7 Approximations and Bounds for the Value of Information 166
*5.7.1 Use of the Taylor Series Expansion 167
5.7.2 Bounds for Stochastic Programming Problems 169
5.7.3 Approximating Information Value by Sampling 173

6. Statistical Determinants of Information Value 177

6.1 The Normal Form of Decision Analysis 177
6.1.1 Randomized Courses of Action 178
6.l.2 Decision Rules 179
6.1.3 Utility Possibilities and Admissibility 184
6.1.4 The Finite Model in Normal Form 190
6.1.5 Convexity in the Probabilities 192
6.2 Comparative Informativeness 197
6.2.1 The Relation "More Informative Than" 197
6.2.2 Blackwell's Theorem 199
6.2.3 Special Cases 204
xiv Contents

6.2.4 Constructing a Sequence of Structures Ranked

by Informativeness 209
6.3 Informativeness, Prior Knowledge, and Value 213
6.3.1 The Value of Statistical Informativeness:
The Binary Case 214
6.3.2 The Value of Systematic Redistribution of the
Probability: Stochastic Dominance 218
*6.3.3 The Differential Approach to the Value of Statistical
Informati veness 224
*6.3.4 Environmental Uncertainty 226

7. Stochastic Preference and Information Value 231

7.1 Prospects and Attitude Towards Risk 231
7.1.1 Prospects Induced by Actions 232
7.1.2 Cash Summarizations for Optimal Prospects 234
7.1.3 Attitude Towards Risk 242
7.2 Risk Preference and Information Value 247
7.2.1 Comparison of the Risk A verter' s Optimal Prospects
with the Risk Neutral Benchmark 248
7.2.2 Bounds on the Comparative Demand Value
of Information 248
7.3 Nonlinear Models and Information A version 256
7.3.1 The Expected Utility Criterion 256
7.3.2 Troubles with the Expected Utility Model 259
7.3.3 Aversion to Information 261

8. Information Demand and Procurement 265

8.1 The Demand for Information 265
8.1.1 The Value of the Costly Informed Decision 265
8.1.2 Economic Consequences of Differential Information 270
8.1.3 Optimal and Suboptimal Information Possession 272
8.2 Information Procurement 276
8.2.1 The Informant as Agent of the DM 277
8.2.2 The Framework and Primary Assumptions 279
8.2.3 The Contract Design Problem 281
8.3 Optimal Incentive Contracts to Procure Information 284
8.3.1 The Expected Value of Information 284
8.3.2 The Optimal Contract Design 287
8.3.3 Implementation 290
8.3.4 Examples 292
*8.3.5 Proofs of Lemmas 2, 3, and 4 296

9. Economics of Valuable Information Systems 299

9.1 Information Use Environments 299
 Contents xv

9.1.1 Business Environments 300

9.1.2 Scientific and Technical Environments 303
9.1.3 User Assessment of Information Value 305
9.2 System Design and Information Gain 307
9.2.1 The Gain-Producing Activities of an
Information System 308
9.2.2 System Data Processing 310
9.2.3 Design and Performance 317
9.3 Experiments in Ex-Ante Information Behavior 319
9.3.1 Normative and Descriptive Economics 319
9.3.2 Experiments in Information Acquisition 321
9.3.3 Experiments in Value Assessment 325
9.4 A Decision-Theoretic Model for System Design 328
9.4.1 The Framework 328
9.4.2 A Theory of Information Production and Cost 330
9.4.3 Information Gain in a Canonical Decision Problem 333
9.4.4 Reported Results from a Survey 336
9.4.5 Characterization of the Organization's Optimal System 340
9.5 Postlude 342

References 345

Acknow ledgments 365

Symbol Glossary 367

Author Index 375

Subject Index 383