Beyond Right and Wrong: The Power of Effective Decision Making for Attorneys and Clients

Beyond Right and Wrong Randall Kiser Beyond Right and Wrong The Power of Effective Decision Making for Attorneys and...

Author: Randall Kiser

95 downloads 949 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Beyond Right and Wrong

Randall Kiser

Beyond Right and Wrong The Power of Effective Decision Making for Attorneys and Clients

Randall Kiser DecisionSet1 550 Hamilton Avenue, Suite 100 Palo Alto, CA 94301 USA [email protected]

ISBN: 978-3-642-03813-6 e-ISBN: 978-3-642-03814-3 DOI 10.1007/978-3-642-03814-3 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2009941064 # Springer-Verlag Berlin Heidelberg 2010 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. This book does not contain and is not intended to be a substitute for legal advice. The ideas and strategies contained in this book may not be suitable for a reader’s circumstances, and readers should consult with a professional and conduct independent research before taking any specific action. The views expressed in this book are those of the author and do not necessarily reflect those of any organization or business with which he is affiliated. Cover design: WMXDesign GmbH, Heidelberg, Germany Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Acknowledgements

This book reflects thousands of interactions with judges, clients, attorneys, law professors, insurance company executives, law students and mediators. I have a sense of gratitude for every person who shared minutes or hours of their time to relate their experiences and impart their insights. Jeffrey Rachlinski expressed an early interest in my empirical research regarding attorney-client decision making, and I recall distinctly his initial, lengthy email sent at dawn on a Sunday morning. His analytical rigor and constructive criticism refined the research later described in my Journal of Empirical Legal Studies (JELS) co-authored article, “Let’s Not Make a Deal: An Empirical Study of Decision Making in Unsuccessful Negotiations.” Some of the key concepts in the JELS article are carried over into Chapter 3 of this book, although I am solely responsible for this substantively different treatment of legal decision making. The legal community’s immediate and broad interest in the JELS article provided fresh impetus to complete this book, a task started in 2004 and completed about a year after the article’s publication. I thank each individual who found the article useful, thought the subject of attorney-client decision making deserved more extensive attention, and encouraged me to finish the book. Special thanks are due to Samantha Cassetta and three anonymous reviewers for their review of and comments on portions of this book. I also thank Wiley-Blackwell and John Wiley & Sons, Inc. for permission to reprint some sentences excerpted from the JELS article. All errors, of course, are mine. The editorial and production departments at Springer-Verlag greatly facilitated the metamorphosis from manuscript to book. Anke Seyfried, in particular, was invariably efficient, knowledgeable, direct, and enthusiastic. My wife, Denise, made a major contribution to this book. Her steadfast and selfless support during the years spent on research and writing turned a daunting endeavor into an enjoyable challenge.

v

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Purposes and Premises of this Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Organization and Philosophy of this Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 What Attorneys Think About Other Attorneys’ Decision-Making Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Part I 2

3

1 3 4 6

Evidence

Prior Research on Attorney-Litigant Decision Making . . . . . . . . . . . . . . . 2.1 The Paradox of Copious Lawyers and Scant Data . . . . . . . . . . . . . . . . . . . 2.2 Empirical Legal Research on Judge, Jury and Attorney Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Judge-Jury Agreement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Punitive Damages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Judges’ Assessments of Juries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.4 Attorney-Jury and Attorney-Attorney Agreement . . . . . . . . . . . 2.2.5 Attorney-Litigant Negotiation Positions, Assessments and Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.6 Disparities In “Same Case” Evaluations and Outcomes . . . . . 2.2.7 Comparisons of Predictions and Outcomes . . . . . . . . . . . . . . . . . . 2.2.8 Damages Award Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.9 Overview of Judge, Jury and Attorney Decision Making . . . 2.2.10 Attorney-Litigant Decision Making in Actual Cases . . . . . . . . 2.2.11 Kiser, Asher and McShane Study of Attorney-Litigant Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11 11 15 17 19 19 20 20 21 22 23 24 24 27 27

A Current Assessment of Attorney-Litigant Decision Making In Adjudicated Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.1 The Fifty Percent Implication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

vii

viii

Contents

3.2 New Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 The Four Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 VerdictSearch Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Case Database Selection Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Attorneys in Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Concepts and Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Negotiation Disparities and Decision Error . . . . . . . . . . . . . . . . . . . 3.3.2 Underpricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3 Overpricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4 Negotiation Disparities Without Decision Error . . . . . . . . . . . . . . 3.3.5 Effect of Negotiation Disparity on Decision Error . . . . . . . . . . . . 3.4 Overall California Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 Costs of Decision Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2 Negotiation Disparities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 New York Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 40-Year Historical Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1 Historical Decision Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.2 Historical Cost of Decision Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7 Attorney-Mediator Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.1 Attorney-Mediator Decision Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.2 Attorney-Mediator Negotiation Disparities and Settlement Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.3 Tentative Conclusions About Attorney-Mediators . . . . . . . . . . . . 3.8 Predictor Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.1 Context Variables Trump Actor Variables . . . . . . . . . . . . . . . . . . . . 3.8.2 The Five Major Context Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.3 Two Secondary Context Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.4 The Major Actor Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.9 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Part II 4

32 32 33 34 35 36 38 39 39 40 41 42 42 44 45 46 46 47 48 49 51 51 52 53 54 71 76 85

Causes

Psychological Attributes of Decision Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.1 Perceptions of Adversaries and Conflicts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 4.1.1 Fundamental Attribution Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 4.1.2 Selective Perception and Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 4.1.3 Self-Serving Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.1.4 Reactive Devaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 4.1.5 A Practical Example Of Overcoming Self-Protective Biases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 4.2 Evaluations of Risk and Reactions to Perceived Risk . . . . . . . . . . . . . 108 4.2.1 Framing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 4.2.2 Anchoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Contents

5

4.3 Reactions to Threatened Changes in Position and Status . . . . . . . . . . 4.3.1 The Endowment Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2 Status Quo Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.3 Overconfidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.4 Confirmation Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.5 Representative and Availability Heuristics . . . . . . . . . . . . . . . . . . . 4.3.6 Hindsight Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.7 Discounting Of Future Payments and Costs . . . . . . . . . . . . . . . . . . 4.3.8 Sunk Cost Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

120 120 122 124 126 129 132 133 136 139

Institutional Impediments to Effective Legal Decision Making . . . . . 5.1 Law School Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.1 Separation of Legal Education from Legal Practice . . . . . . . . 5.1.2 Testing Law Students’ Reasoning Skills and Moral Judgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.3 An Example of Law Student Decision Making . . . . . . . . . . . . . . 5.1.4 Deficiencies in the Case Method of Teaching . . . . . . . . . . . . . . . 5.1.5 Attempts to Change Law School Curriculum . . . . . . . . . . . . . . . . 5.2 Law Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Conflicts Between Efficient Problem Solving and Billable Hour Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2 The Consequences of Avoiding “The Big Picture” . . . . . . . . . . 5.2.3 “Due Process” and the Elevation of Process Above Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.4 Competitive Market Pressures, Undue Deference to Client Expectations and Inappropriate Client Involvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Mental Impairment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 The Disappearing Civil Trial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.1 Settling Without Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.2 Causes and Motivations for Pre-Trial Settlements . . . . . . . . . . . 5.5 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

141 143 144

Part III 6

ix

145 148 150 156 164 166 169 172

174 182 188 189 192 195

Consequences

Legal Malpractice Liability For Settlement Counseling and Decision Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Malpractice Claims Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Competing Policy Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Malpractice Claims Arising from Settled Cases . . . . . . . . . . . . . . . . . . . 6.3.1 Inadequate Advice Regarding Settlement and Trial Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.2 Client Coerced into Settlement by Attorney . . . . . . . . . . . . . . . .

199 200 202 204 206 209

x

Contents

6.3.3

Attorney’s Mistakes Prevented Client from Obtaining a Better Settlement or Prosecuting Case to Trial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.4 Attorney’s Delays Caused Client to Forego More Favorable Settlement Terms . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.5 Conflict of Interest, Fraud and Collusion with an Adverse Party . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.6 Attorney Did Not Transmit Settlement Proposals to Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.7 Failure to Conduct Adequate Legal Research, Discovery and Investigation Before Settlement . . . . . . . . . . . . 6.3.8 Attorney Not Authorized to Consent to Settlement Agreement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.9 Settlement Agreement Defectively Drafted . . . . . . . . . . . . . . . . . 6.3.10 Client Misunderstood the Settlement Agreement . . . . . . . . . . . 6.3.11 Failure to Advise of Uncertainty of Law and Anticipate Judicial Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Malpractice Claims in Adjudicated Cases . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.1 Attorney Remiss In Failing To Initiate Settlement Negotiations, Solicit A Pre-Trial Settlement Offer Or Otherwise Effectuate Settlement . . . . . . . . . . . . . . . . . . . . . . . . 6.4.2 Client Inadequately Apprised of Risk of an Adverse Verdict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Defenses to Settlement Malpractice Claims . . . . . . . . . . . . . . . . . . . . . . . . 6.5.1 The Client’s Consent Bars a Challenge to the Adequacy of the Settlement Agreement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.2 The Client’s Ratification of the Settlement Agreement . . . . 6.5.3 The Client’s Failure to Prove Reliance on the Attorney’s Advice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.4 The Judgmental Immunity Rule and the California Model Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.5 The Client Cannot Prove Damages Proximately Caused by the Attorney’s Negligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.6 Another Attorney’s Negligence as an Intervening or Superseding Cause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.7 Reduction of Malpractice Awards by the Amount of Attorneys Fees the Client Otherwise Would Have Paid the Attorney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

211 214 215 218 219 223 225 226 227 231

232 235 237 238 240 240 241 243 245

246 247

Ethical Implications of Attorney-Client Counseling and Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 7.1 A Profile of Disciplinary Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

Contents

xi

7.2 The Duty to Communicate all Material Facts and Events to Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 The Duty to Exercise Independent Judgment and Render Candid Advice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 The Duty to Provide Adequate Advice to Enable Clients to Make Informed Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 The Duty to Identify and Protect Clients with Diminished Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 The Duty to Competently, Independently, Diligently and Expeditiously Represent Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.7 The Duty to Abide by Client Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.8 The Duty to Prevent Conflicts of Interest in Aggregate Settlements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.9 The Duty to be Candid and Truthful in Communications with Clients, Opposing Counsel and the Courts . . . . . . . . . . . . . . . . . . . 7.10 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Part IV

252 257 260 262 266 270 272 274 279

Solutions

8

Obstacles to Becoming an Expert Decision Maker . . . . . . . . . . . . . . . . . . . 8.1 Defenses and Barriers to Sound Decision Making . . . . . . . . . . . . . . . . . 8.1.1 Defenses to Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.2 Distortions of Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.3 Attorney Belief System Defenses . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Myths and Misconceptions About Decision Making Expertise . . . . 8.2.1 Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2 Education and Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.3 Peer Ranking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.4 Intuition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

283 284 285 288 293 295 296 298 302 303 307

9

Personal Expertise in Legal Decision Making . . . . . . . . . . . . . . . . . . . . . . . . 9.1 Phase One: Finding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.1 Still The Messenger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.2 Bottom-Up Decisions Beat Top-Down Decisions . . . . . . . . . . 9.1.3 Challenge Your Perceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.4 Give Vivid Pictures Time to Fade . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.5 Credit Randomness its Due . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.6 Deal with Attribution Errors Early . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.7 Diversify the Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.8 Time Does not Take Sides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.9 Align Client Objectives and Attorney Incentives . . . . . . . . . . . 9.1.10 Consider Appointing Separate Settlement Counsel . . . . . . . . .

309 310 311 311 313 314 315 316 317 318 319 321

xii

Contents

9.2 Phase Two: Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1 Start with Ideals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.2 Switch Sides to Debias Judgment . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.3 Think Divergently . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.4 Stop Pattern Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.5 Work Well with Others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.6 Consider Whether a Litigation Attorney or a Trial Attorney is Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Phase Three: Solving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.1 Don’t Follow Your Gut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.2 Search for Disconfirming, Discrepant Facts . . . . . . . . . . . . . . . . 9.3.3 Pay Attention to Base Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.4 Prepare to Justify Your Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.5 When in Doubt, Act it Out . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.6 Step Off the Information Treadmill . . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Phase Four: Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.1 Find Your Inner BATNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.2 Separate Facts from Theories, Values and Beliefs . . . . . . . . . 9.4.3 Enlarge the Pie Before Cutting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.4 Subjective Fairness Matters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.5 Think and Communicate Affirmatively . . . . . . . . . . . . . . . . . . . . . 9.4.6 Depressed People Make Depressing Deals . . . . . . . . . . . . . . . . . 9.4.7 Fatigue Stifles Creative Problem Solving . . . . . . . . . . . . . . . . . . 9.4.8 Use Email Carefully . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.9 Get a Grip on Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5 Phase Five: Choosing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.1 Perform a Premortem on Overconfidence . . . . . . . . . . . . . . . . . . . 9.5.2 Take the Outside View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.3 Keep Positions Aligned with Facts . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.4 Separate the Primary Decision from the Secondary Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.5 Assumptions Were Made to be Explicit and Tested Continuously . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.6 Walk Around the Sunk Cost Trap . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.7 Past Performance Is No Guarantee of Future Results . . . . . . 9.5.8 Funny Things Happen on the Way to the Forum . . . . . . . . . . . 9.5.9 Linear Thinking Leads to Impasse . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.10 Appeals are Part of the Settlement Equation . . . . . . . . . . . . . . . 9.5.11 Moderate the Mediator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6 Phase Six: Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6.1 Pin Yourself Down for Some Real Feedback . . . . . . . . . . . . . . . 9.6.2 Don’t Just Provide Feedback – Discuss it . . . . . . . . . . . . . . . . . . 9.6.3 Learn from Surprises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.7 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

322 323 324 325 326 327 328 330 330 332 333 334 335 336 337 338 339 340 341 342 343 344 345 347 348 349 350 351 353 354 354 355 356 357 358 359 362 363 364 365 366

Contents

xiii

10

Group Expertise In Legal Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1 Deficiencies in Group Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1.1 Elements of Defective Group Decisions . . . . . . . . . . . . . . . . . . 10.1.2 Group Polarization and Groupthink . . . . . . . . . . . . . . . . . . . . . . . 10.2 Characteristics of Effective Decision-Making Groups . . . . . . . . . . . 10.2.1 High Reliability Organizations (HROS) . . . . . . . . . . . . . . . . . . 10.2.2 Expert Teams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Steps to Improve Group Decision Making . . . . . . . . . . . . . . . . . . . . . . . . 10.3.1 Ask For Multiple Opinions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.2 Cross-Pollinate the Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.3 Proliferate Team Leaders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.4 Appoint a Devil’s Advocate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.5 Seed the Brainstorm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.6 Promote a Good Fight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.7 Build Trust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.8 Reach a Consensus, Don’t Build One . . . . . . . . . . . . . . . . . . . . . 10.3.9 Schedule a Last Clear Chance Meeting . . . . . . . . . . . . . . . . . . . 10.4 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

367 368 370 371 376 377 383 385 385 386 387 389 389 391 392 394 395 396

11

Peer Review, Client Evaluations and Law Firm Audits . . . . . . . . . . . . . 11.1 A Brief History of Quality Management in Law Firms . . . . . . . . . . 11.2 Peer Review in the Medical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2.1 The Inception of Medical Peer Review . . . . . . . . . . . . . . . . . . . 11.2.2 The Modern Medical Peer Review System . . . . . . . . . . . . . . . 11.2.3 Confidentiality of Medical Peer Review . . . . . . . . . . . . . . . . . . 11.3 Peer Review in Law Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3.1 Priorities in Law Firm Peer Review . . . . . . . . . . . . . . . . . . . . . . 11.3.2 Confidentiality of Attorney Peer Review Proceedings . . . 11.3.3 Professional Ethics and Attorney-Client Privilege . . . . . . . . 11.3.4 The Role of Confidentiality in Peer Review . . . . . . . . . . . . . . 11.3.5 The Structure of Law Firm Peer Review . . . . . . . . . . . . . . . . . 11.4 Client Evaluations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4.1 Challenges of Evaluation Design and Analysis . . . . . . . . . . . 11.4.2 Sample Questions to Probe for Decision-Making Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5 Assessments and Audits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6 Chapter Capsule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

397 399 400 401 402 403 404 404 405 407 410 412 413 414

12

418 419 422

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

Chapter 1

Introduction

Let us endeavor to see things as they are, and then enquire whether we ought to complain. Whether to see life as it is, will give us much consolation, I know not; but the consolation which is drawn from truth if any there be, is solid and durable: that which may be derived from errour, must be, like its original, fallacious and fugitive. Samuel Johnson, Letter to Bennet Langton (1758)

Attorneys and clients make hundreds of decisions in every litigation case. From initially deciding which attorney to retain to deciding which witnesses to call at trial, from deciding whether to file a complaint to deciding whether to appeal a verdict, attorneys and clients make multiple, critical decisions about strategies, costs, arguments, valuations, evidence and negotiations. Once made, these decisions are scrutinized by an opponent intent on exploiting the consequences of any mistake. In this intense and adversarial arena, decision-making errors often are transparent, irreversible and dispositive, wielding the power to bankrupt clients and dissolve law firms. Although attorneys and clients may regard sound decision making as incidental to effective lawyering, sound decision making actually is the essence of effective lawyering. An attorney’s knowledge, intelligence and experience are inert resources until the attorney decides how to deploy those skills to serve the client’s interests. Those decisions, in turn, largely determine a case’s course and outcome. Very few cases are lost because attorneys and clients do not understand the law; losses are more often traceable to poor quality decisions than poor quality research. The unfortunate consequence is that legally meritorious claims and defenses, advanced by technically competent attorneys, can be lost through bad decision making. As one major law firm declares in its Wall Street Journal advertisement, “Being a good lawyer takes more than being a good lawyer.”1 In most cases with disappointing results, there is a point where an effective decision could have averted an adverse financial outcome. The ability to identify and seize that pivotal opportunity separates novice decision makers from experts. 1

(2007, December 3). The Wall Street Journal, p. A8.

R. Kiser, Beyond Right and Wrong, DOI 10.1007/978-3-642-03814-3_1, # Springer-Verlag Berlin Heidelberg 2010

1

2

1 Introduction

An effective decision’s capacity to circumvent a financial disaster, in litigation phases ranging from pre-trial settlement negotiations to new trials in remanded cases, is illustrated in the actual cases briefly described below.2 Each case presented at least one opportunity to insert a protective decision in front of a startling outcome: l

l

l

l

l

l

A plaintiff demands $13 million to settle a breach of contract case and refuses to accept the defendant’s settlement offer of $500,000. At trial, the plaintiff recovers nothing and the defendant is awarded $22 million under its crosscomplaint against the plaintiff. An arbitrator issues an award against a defendant in the amount of $175,000. The defendant rejects the award and exercises its right to a new trial before a jury. The jury returns a verdict of $2,025,000 against the defendant, an amount nearly 12 times larger than the arbitration award that the defendant rejected. A plaintiff declines a defense settlement offer of $100 million in a securities class action case. After a four-week trial, the jury takes less than two days to render a verdict in favor of the defendant. A defendant employer rejects the plaintiff employee’s offer to settle a sexual harassment case for $75,000 and a job transfer. Five years later, an appellate court upholds a $2 million award in favor of the plaintiff employee. In a legal malpractice action, the plaintiff demands $325,000 to settle. The defendant law firm does not make an offer to the plaintiff until the day of trial, at which time it offers $50,000. The plaintiff declines the $50,000 offer and the jury later renders a verdict of $7 million against the law firm. Including interest, the amount ultimately paid by the law firm to satisfy the judgment is $10 million. A defendant successfully appeals from a $675,000 award entered against it. As the defendant requested, the appellate court reverses the lower court’s award and remands the case for a new trial. Upon retrial, the jury finds against the defendant and awards the plaintiff $2.2 million, roughly triple the amount of the original award from which the defendant appealed.3

In each of these cases, attorneys and their clients passed a decision inflection point and proceeded to a major, yet entirely avoidable, adverse outcome. Looking back on cases that went awry, clients have claimed “our lawyers did not do what they were supposed to do,” attorneys have blamed “stupid jurors” and “runaway juries,” and both clients and attorneys bemoan the apparent vagaries of the civil justice system. For readers whose reaction to these adverse outcomes is anything other than “tough luck,” this book presents compelling data, concepts and 2

Many decisions, of course, are high quality decisions with bad outcomes, i.e., good processes accompanied by bad results. The emphasis here on effectiveness promotes closer scrutiny of both poor quality decision making and arguably good quality decision making with adverse outcomes. This emphasis also shifts attention from fault-finding to improvement. 3 Each case scenario is based on an actual case on file with the author. The outcome of subsequent appeals, motions, and settlement negotiations, if any, and the existence and importance of noneconomic factors are unknown.

1.1 Purposes and Premises of this Book

3

correctives that could prevent their own cases from becoming exemplars of catastrophic decision making.

1.1

Purposes and Premises of this Book

This book is written for attorneys who aspire to become better decision makers, clients who seek realistic guidance in making legal decisions and law students who wish to spare clients the ordeal of trial and error training. Its objective is to teach attorneys, clients and law students to make effective decisions in resolving civil litigation cases. Its underlying premises are that ample room exists for improvement in attorney-litigant decision making, trial outcomes can be predicted with greater accuracy than is presently achieved, decision-making errors about case strategies and pre-trial settlements can be reduced, and tough decisions about cases are best made within an analytical framework rather than behind a courtroom counsel table cornered by intuition, hunch, instinct, and hope. To obtain maximum benefit from this book, attorneys may need to recognize that their experience in decision making is not equivalent to expertise in decision making, clients may need to acknowledge that their confidence in decision making is different from proficiency in decision making, and law students may need to discover that their knowledge of the law does not automatically impart competence in decision making. Effective decision making, in short, is a distinct skill. Contrary to popular perceptions, effective decision-making skill has little relation to experience, intelligence, education and professional reputation. As Oliver Wendell Holmes observed, “some of the sharpest men in argument are notoriously unsound in judgment. I should not trust the counsel of a smart debater, any more than that of a good chess-player.”4 Technically competent attorneys, therefore, are not necessarily effective decision makers, and many effective decision makers are not recognized as experts in any particular practice area. Knowing “what” and selecting “how” are independent yet complementary skills. In endeavoring to become expert decision makers, attorneys, clients and law students inevitably will shift their focus from how to prevail in a trial to how to resolve a case through settlement. This shift follows from the fact that about 95% of all civil litigation cases are resolved without a trial. Making decisions about whether to settle and the terms on which to settle, consequently, is more important in the vast majority of cases than an attorney’s trial skills. Although many clients initially resist the idea of settling a case and prefer to vindicate their positions at trial, the reality is that nearly every case is involuntarily dismissed or eventually settled. In the vast majority of cases, clients have a greater likelihood of making a devastating settlement decision in a mediation session than watching their attorney 4

Holmes, Oliver Wendell. (1858). The autocrat of the breakfast-table (pp. 16–17). New York: Dutton, Everyman’s Library.

4

1 Introduction

conduct a devastating cross-examination at trial. Because a settlement is the most likely result in civil litigation, the critical factor that separates successful litigants from unsuccessful litigants often is the quality of their decision making. Contrary to legal stereotypes, the party most likely to win a case is not the one that can afford the best trial attorney but rather the party that forms the best attorney-client decision-making team. This book’s emphasis on decision-making skills also promotes the ethical practice of law and enables attorneys to fulfill their professional obligations, as envisioned by the American Bar Association (ABA). In its Model Rules of Professional Conduct, the ABA demarcates four roles attorneys assume when representing clients: advisor, advocate, negotiator and evaluator.5 Only one of those roles (advocate) requires conventional courtroom skills and tactics, while the other three roles (advisor, negotiator and evaluator) mandate proficiency in the broader skill set that underpins decision-making acumen.

1.2

Organization and Philosophy of this Book

Like any distinct skill, decision-making acumen is acquired by objectively assessing one’s performance, identifying the impediments to superior performance, evaluating the consequences of continued suboptimal performance and improving performance through a rigorous and testable regimen. This book, accordingly, is organized to address four questions critical to developing expert skills in legal decision making: l

l

l

l

Do attorneys and their clients make financially sound decisions about pre-trial settlement offers in civil litigation cases? What psychological and institutional factors affect decision making in civil litigation cases? What are the legal and professional consequences of making ineffective decisions about the settlement or trial of civil cases? How can attorneys and clients improve their decision-making skills in all phases of civil litigation?

Stated differently, this book examines the quality of decisions made by attorneys and clients, explains why attorneys and clients make both effective and ineffective decisions, outlines the legal malpractice and ethical implications of ineffective decisions and shows how to make better decisions. Part One of this book reviews prior research on attorney-litigant decision making and the disparities between the predictions of attorneys and clients and their actual case outcomes. It then summarizes recent research results regarding nearly 11,000 5

Center for Professional Responsibility. (2007). Model rules of professional conduct (p. 1). Chicago, Illinois: American Bar Association.

1.2 Organization and Philosophy of this Book

5

pre-trial settlement decisions made by attorneys and clients in California and New York. Part Two examines psychological factors that contribute to the decision-making shortcomings described in Part One and considers how institutional factors (law school education, law firm culture, and the judicial system) may affect attorneys’ forecasting and problem-solving skills. Part Three explains the legal and ethical consequences of inadequate or inaccurate legal advice, showing how poor quality counseling about settlement prospects can become actionable malpractice and a breach of professional ethics. Lastly, Part Four describes why attorneys find it difficult to learn better decision-making skills, how individual attorneys and clients can improve these skills, what techniques groups employ to develop superior decision-making skills, and how law firms can utilize peer review, evaluations and audits to enhance their attorneys’ decision-making capabilities. This book differs from other books and articles on settlement negotiations in that it places greater weight on scientific evidence than the war stories of attorneys, mediators and judges; it assumes that empirical studies are more instructive than anecdotes and statistics are more dependable than surmise. The overall philosophy of the book is to bump, when possible, the legal field from the narrative to the empirical, from qualitative conjecture to quantitative proof. As a result, this book may be less entertaining than popular books on negotiation and litigation and actually will require considerably more work on the reader’s part. For the determined reader, the additional cognitive effort, hopefully, will be rewarded by a more durable understanding of what really happens in litigation decision making and what has proven effective in improving its quality. This book defers to the time demands placed on busy, hyper-scheduled attorneys, clients and law students. Each chapter may be read without reading the prior chapter, and the summary at the end of each chapter can be used as a snapshot of that chapter. Attorneys who want to read only about improving their decisionmaking skills, for example, may move directly to Chapter 9. Reading the book in a piecemeal or abbreviated manner conveys the key points to readers with very limited time, but it is not recommended. Nevertheless, some readers have less than an hour to read the material most pertinent to their needs, and this book is structured to accommodate the narrowly focused as well as the broadly inquisitive reader. Two important clarifications are necessary. First, the term “decision making” used throughout this book is a compact substitute for the more expansive set of cognitive skills identified by psychologists as judgment, decision making and problem solving.6 Non-psychologists might call these skills “good sense,”

6

“Decision making has been defined as ‘the ability to gather and integrate information, use sound judgment, identify alternatives, select the best solution and evaluate the consequences.’” Salas, Edward, et al. The making of a dream team: When expert teams do best. In Ericsson, K. Anders, et al. (Eds.). (2006). The Cambridge handbook of expertise and expert performance (p. 441). New York: Cambridge University Press. Cf. Tichy, Noel M., and Bennis, Warren G. (2007). Judgment (p. 287). New York: Penguin Group. (“We make a distinction between judgment and decision making”).

6

1 Introduction

“practical,” “good judgment,” or simply “wisdom.” Second, although this book emphasizes empiricism over anecdotes, readers do not need a background in statistics, mathematics or psychology to understand it. This book deliberately excludes decision-making models, tables and charts that require familiarity with probability theory, regression analysis, game theory, decision tree algorithms, t-tests, p-values and Bayesian analysis. These complex methods and tests are highly valuable tools for decision makers, but they are excluded here for a simple reason: attorneys generally don’t like them, don’t understand them and won’t use them. Readers seeking a more scientific or statistical analysis of attorney-litigant decision making may wish to review the author’s article, “Let’s Not Make A Deal: An Empirical Study Of Decision Making In Unsuccessful Settlement Negotiations,” co-authored with Martin A. Asher and Blakeley B. McShane of The Wharton School, and published in the Journal of Empirical Legal Studies, Vol. 5, No. 3, pp. 551–591 (September 2008).

1.3

What Attorneys Think About Other Attorneys’ Decision-Making Skills

If attorneys question the importance of decision making to clients or doubt that the quality of decision making varies among attorneys, they may be surprised to see what their colleagues say about the profession’s decision-making capabilities. Recent advertisements in The Wall Street Journal, placed by the nation’s leading law firms, appear to capitalize on the perceived inadequacy of their competitors’ decision-making skills: l

l

l

l l

l

l

“We believe that what separates us from the pack is not what we do, but how we do it – aggressive not conservative, team players not one-man-bands, problem solvers not just legal practitioners.” “I don’t need theories from my lawyers. I need answers. Ever get a three page memo from your lawyer when you’re looking for quick, to-the-point advice? At Nixon, Peabody LLP, we know that you prefer simple, clear and practical to rambling and theoretical. Your world is complicated enough.” “Major litigation is rarely straightforward. Working with your law firm should be.” “The best attorneys know how to balance aggression with delicate handling.” “I don’t need lawyers who win at all costs. I need them to win, but calculate the costs.” “You need lawyers who will simplify the process – not complicate it further. At Winston & Strawn we’re committed to helping our clients find the most direct route to a successful outcome. When you’re faced with complex litigation, choose a law firm that will help you chart the right course.” “If your lawyers seem more concerned about enumerating your options than helping you choose among them, you might wonder whose interests are really

1.3 What Attorneys Think About Other Attorneys’ Decision-Making Skills

7

being served. At Nixon Peabody LLP, options are clear, advice is candid, and your needs always come first.”7 The message these major law firms deliver is compellingly persuasive to clients who feel corralled by legal naysayers, precisionists and deconstructionists: we’re not like other law firms; we communicate directly, don’t waste time on papering the file, give direct advice, avoid complicated and theoretical discourses, know better than to use heavy-handed litigation tactics indiscriminately, provide counsel about a realistic set of alternatives, and expeditiously guide clients to resolutions. They appeal to the client declaiming in one advertisement, “I need lawyers who are more concerned about managing my risks than their own.”8 When law firms direct advertisements like these to the nation’s largest and most sophisticated consumers of legal services, criticizing the decision-making and problem-solving skills of attorneys in other law firms, what weaknesses in the profession are they trying to exploit? Why do these advertisements highlight client dissatisfaction with the problem-solving capabilities, judgment and decisionmaking skills of their attorneys, while avoiding any criticism of their legal knowledge? What information do these law firms have that indicates clients are frustrated with attorneys who are long on legal theories and short on practical solutions, quick to generate alternatives but unable to rank them? These questions, central to a profession ethically required to advance and protect clients’ interests, are addressed in this book.

7

(2007, February 5). The Wall Street Journal, p. A4. (2007, September 13). The Wall Street Journal, p. A4. (2007, September 18). The Wall Street Journal, p. A6. (2008, March 31). The Wall Street Journal, p. A4. (2008, May 20). The Wall Street Journal, p. A17. (2008, September 23). Palo Alto Daily News, p. 6. 8 (2008, May 20). The Wall Street Journal, p. A17.

Part I Evidence

Chapter 2

Prior Research on Attorney-Litigant Decision Making

Nothing is more dangerous to a new truth than an old error. Johann Wolfgang von Goethe, Proverbs in Prose (1819)

Despite the filing of 15 million new civil cases every year, little attention has been given to the decisions made by attorneys and their clients in initiating, prosecuting, defending and attempting to resolve those cases.1 The perceived “litigation explosion” has not ignited a commensurate investment in empirical studies to describe the underlying reasons and motivations for filing and maintaining civil actions, the psychological and financial obstacles to conflict resolution, and the economic utility of decisions about settling cases or bringing them to trial. Academicians have analyzed these subjects, but funding for their research is miniscule relative to the impact of litigation on the nation’s economy. As Lela Love, a law professor and the chair of the American Bar Association’s Section of Dispute Resolution, notes, “We really know very little about conflict and its dynamics.”2

2.1

The Paradox of Copious Lawyers and Scant Data

The dearth of data about litigation is perplexing in a country with the highest concentration of attorneys per capita and a tradition of giving every citizen her day in court. Although the legal services industry is a major force and business in the American economy, affecting nearly every aspect of risk assessment from automobile design to pharmaceutical research, from kindergarten field trip waivers 1

Ostrom, Brian J., Kauder, Neal B., and LaFountain, Robert C., Eds. (2001). Examining the work of state courts 2001: A national perspective from the court statistics project (p. 16). National Center for State Courts. (January 2008). Appendix – Workload of the Courts. The Third Branch, 40(1). Administrative Office of the U.S. Courts Office of Public Affairs. 2 Love, Lela P. (March 2007). Voice of experience. Just resolutions eNews. ABA Section of Dispute Resolution. Available June, 2007 at http://abanet.org/dispute/voice.html


11

12

2 Prior Research on Attorney-Litigant Decision Making

to Fortune 500 companies’ earnings guidance, the quality of data regarding one of America’s greatest growth industries is distressingly poor and grossly disproportionate to the industry’s impact on the national economy. Basic information regarding total national legal compliance and net litigation costs, for example, is not available. “The fund of basic information does not exist,” law professor Marc Galanter said in 1994. “It is as if we had a medical establishment,” he adds, “consisting entirely of practicing physicians with no research institutes like the National Institutes of Health and no public health monitoring facilities like the Centers for Disease Control.”3 Fourteen years later, little had changed, as law professor Theodore Eisenberg observes: “Policymakers and interest groups regularly debate and assess whether civil problems are best resolved by legislative action, agency action, litigation, alternative dispute resolution, other methods, or some combinations of actions. Yet we lack systematic quantitative knowledge about the primary events in daily life that generate civil justice issues.”4 Although the nation cannot track key data regarding the civil justice system and the legal services industry, its expenditures on legal services increase at rates much higher than the nation’s inflation and GDP growth rates. Between 2000 and 2005, total revenue for the legal services industry increased from $161 billion to $222 billion, and average profits per partner in the nation’s 100 top-grossing firms increased from $800,000 to $1,060,000.5 To place these figures in perspective, in 2005 the total amount spent on legal services exceeded the total amount spent by all American businesses on research and development and was more than twice the total amount the federal government spent on research and development.6 Americans thus paid more money for legal services than their businesses invested in securing a competitive advantage in the future. Another perspective is that, although 1,400,000 Americans are diagnosed with cancer every year and cancer kills about 560,000 Americans every year, total annual expenditures on legal services are about 50 times the amount the National Institutes of Health spend on cancer research.7

3

Galanter, Marc, et al. (1994, January–February). How to improve civil justice policy. Judicature, 77(4), 185, 230. 4 Eisenberg, Theodore. (2008, November 12). The need for a national civil justice survey of incidence and claiming behavior (p. 1). Available at SSRN: http://ssrn.com/abstract¼1305385. 5 United States Bureau of the Census. (2008). Table 1249, professional, scientific, and technical services — estimated revenue: 2000 to 2005. Statistical abstract of the United States 2008. Washington, DC. McCoy, Blythe. (2006). Trends & business strategies in the legal industry . . . the law firm perspective. Thomson West. Willing, Richard. (2006, May 1). Top law firms rake in bigger bucks. USA Today. (2006, May). Prime movers. The American Lawyer. (2005, June 28). The American Lawyer reports law firm revenues top $46 billion in 2005 am law 100 rankings. Business Wire. 6 National Science Board. Chapter 4 – Research and Development: National Trends and International Linkages. Science and Engineering Indicators 2008 (pp. 4–5, 4–13). Arlington, Virginia: National Science Foundation. 7 American Cancer Society. (2007). Estimated new cancer cases and deaths by sex for all sites, US, 2007. Cancer Facts & Figures 2007. Atlanta: American Cancer Society. (2006, May 19). Cancer Research Funding. National Cancer Institute Fact Sheet.

2.1 The Paradox of Copious Lawyers and Scant Data

13

The nation’s expenditures on legal services reflect its relatively large population of attorneys. Nearly one in every 262 Americans is an attorney, and every weekday morning 1,143,358 chairs throughout the nation await indentation behind a lawyer’s desk.8 By way of contrast, the total number of civil engineers in the United States is 256,000, and the number of physical scientists is 199,600.9 There are 173,000 law offices around the country, compared with 55,000 engineering services firms and 16,000 scientific research and development establishments.10 While more than 1.1 million active attorneys provide legal services, medical care is available from only 800,000 active physicians.11 Indigent citizens can avail themselves of more civil legal aid programs than federally funded community and migrant health care centers.12 Despite the large number of attorneys, the high demand for legal services and the intensive legal regulation of modern life, scant data exist regarding the economic benefits of legal services expenditures, the accuracy of attorneys’ advice and judgment and the efficacy of their representation in lawsuits. As law professor Douglas Rosenthal observes, “we ourselves have no reliable information about how competent, in the aggregate, lawyers actually are.”13 This paradox is evident to economists and psychologists: The appropriateness of lawyers’ probability judgments has important implications for the quality of their service – decisions about whether to sue, settle out of court, or plead guilty to a lesser charge, all depend on a lawyer’s judgment of the probability of success.

8

American Bar Association. (2007). National lawyer population by state. Compiled by ABA Market Research Department, Chicago, Illinois. United States Census Bureau, Population Estimates Branch. (2006). Estimated Population by State: 2000–2006. 9 U.S. Department of Labor, Bureau of Labor Statistics. Engineers. Occupational outlook handbook, 2008–2009 Edition (p. 6). National Science Foundation, Division of Science Resources Statistics. (2001). Scientists, Engineers, and Technicians in the United States: 2001 (NSF 05-313). 10 United States Bureau of the Census supra note 5, Table 1247, Selected service-related industries – establishments, employees, and payroll by industry: 2003 and 2004. 11 U.S. Department of Health and Human Services, Health Resources and Services Administration, Bureau of Health Professions. Table 201, total and active physicians (mds) and physician-topopulation ratios, selected years: 1950–2000. National Center for Health Workforce Analysis: U.S. Health Workforce Personnel Factbook. United States Bureau of the Census supra note 5, Table 1249, professional, scientific, and technical services — estimated revenue: 2000 to 2005. Ratios, selected years: 1950–2000. Cauchon, Dennis. (2005, March 2). Medical miscalculation creates a doctor shortage. USA Today. 12 Houseman, Alan. (2007). Civil legal aid in the United States: An update for 2007. Center for Law and Social Policy. Houseman, Alan W. (2005). Civil legal aid in the United States: An overview of the program and developments in 2005. Center for Law and Social Policy. United States General Accounting Office. (2000). Community health centers: Adapting to changing health care environment key to continued success (p. 43). Rosenbaum, Sara, Shin, Peter, and Darnell, Julie. (2004). Economic stress and the safety net: a health center update. The Henry J. Kaiser Family Foundation. 13 Rosenthal, Douglas E. (1976). Evaluating the competence of lawyers. Law & Society Review, 11, 257.

14

2 Prior Research on Attorney-Litigant Decision Making Surprisingly, then, there is relatively little research assessing the calibration of lawyers’ probability judgments in their day-to-day practice.14

Like medical practice before the advent of national databases recording patient outcomes, the legal services industry generally has eluded quantitative accountability and comparability in outcome assessment and peer benchmark performance standards. Consistent with the profession’s lack of objective performance standards and the absence of comparative quality measurements, the typical fee agreement in a litigation matter contains a clarion disclaimer of responsibility for the primary purpose of the retention: results. The lack of data about legal services in general and litigation outcomes in particular contrasts sharply with businesses’ rapidly expanding reliance on analytics. For more than two decades, businesses have evolved from subjective evaluations to quantitative analysis, from making decisions based on intuition and hunches to relying on data and algorithms. As Ian Ayres, a law professor and author of Super Crunchers notes, “We are in a historic moment of horse-versuslocomotive competition where intuitive and experiential expertise is losing out time and time again to number crunching.”15 This shift to quantitative analysis has occurred in virtually every major business sector except law. Even sports teams, as shown in the popular book Moneyball, are more likely to employ analytics than law firms.16 Companies that are data driven, “from automobiles, to textiles, to computer software, to baseball,” explain Stanford University professors Jeffrey Pfeffer and Robert Sutton in their book Hard Facts: Dangerous Half-Truths and Total Nonsense, consistently outperform their competitors: “Organizations can gain competitive advantage if they take the trouble to substitute facts for common lore and to test conventional wisdom against the data.”17 The competitive advantage of analytics-driven companies is reiterated by management professor Thomas Davenport in his Harvard Business Review article, “Competing on Analytics:” “Virtually all the organizations we identified as aggressive analytics competitors are clear leaders in their fields, and they attribute much of their success to masterful exploitation of data.”18 Apart from comparing gross revenue and profits per partner with competitors, law firms are not competing on analytics. Despite the widespread use of analytics in the business world, law firms are not running horse-versus-locomotive races, as depicted by Ian Ayres, but are still pitting their thoroughbreds against another 14

Koehler, Derek, Brenner, Lyle, & Griffin, Dale. The calibration of expert judgment: heuristics and biases beyond the laboratory. In Gilovich, Thomas, Griffin, Dale, and Kahneman, Daniel (Eds.). (2002). Heuristics and biases: the psychology of intuitive judgment (p. 705). Cambridge: The Press Syndicate of the University of Cambridge. 15 Leonhardt, David. (2007, September 16). Let’s go to the stats. The New York Times. 16 Lewis, Michael. (2005). Moneyball. New York: W.W. Norton & Co. 17 Pfeffer, Jeffrey and Sutton, Robert. (2006). Hard facts: Dangerous half-truths & total nonsense (p. 14). Boston, Massachusetts: Harvard Business School Press. 18 Davenport, Thomas. (2006, February). Competing on analytics. Harvard Business Review, pp. 99, 106.

2.2 Empirical Legal Research on Judge, Jury and Attorney Decision Making

15

firm’s thoroughbreds. Law firm clients, consequently, have become accustomed to working in two different realms. The first realm, outside their lawyer’s office, is increasingly directed by data and requires quantitative justification for decisions and objective evidence of accomplishment within budget. The second realm, inside their lawyer’s office, is characterized by “on the one hand this, on the other hand that” legal advice, lengthy narratives and memoranda, caveats that loom larger than general propositions, and attorneys who submit a total fee projection and then “blow through it in half the time.”19 Law firm clients who would be aghast if their financial institution relied on a personal interview instead of a credit score in making a $100,000 loan decision nevertheless are comfortable turning a $50,000,000 case over to an attorney who will not employ a quantitative analysis to assess case outcome probabilities and whose own record of trial losses, case management, cost control and decision errors is unknown. Clients reluctant to invest $5,000 in a mutual fund without checking its Morningstar rating retain attorneys to purchase buildings, sell businesses, and license patents based simply on a colleague’s recommendation or another lawyer’s referral. This anomaly arises not by client choice but rather by necessity; comparative performance data simply are not available.20

2.2

Empirical Legal Research on Judge, Jury and Attorney Decision Making

Although many aspects of legal practice remain in a pre-reformation mode – their language cryptic, their rituals opaque and their prefects autonomous – one aspect of the legal system sparked early and earnest quantitative research: decision making by judges, juries and litigation attorneys. One prong of this research focused on judge-jury agreement, i.e., whether judges and juries make similar determinations of guilt, liability and damages. A second prong concentrated on attorney and litigant predictions about case outcomes, i.e., whether attorneys and clients make accurate or over-optimistic assessments of what a judge or jury will decide at trial. As explained below, this research generally shows that judges and juries have similar opinions about how cases should be decided, but attorneys and clients are 19

(2008, November 7). Interview with Craig Nordlund, General Counsel of Agilent Technologies, reported in Silicon Valley: In the beginning. San Francisco Daily Journal, p. 1. 20 For the businesses and insurance companies that claim to employ quantitative methods for selecting and evaluating law firms, one may question their usefulness when 54% of corporate counsel report that they fired their primary law firms in the last 18 months and only 31% would recommend their primary law firm to another company. Source: BTI Consulting Group. (2006, March 3). Client Satisfaction with Law Firms Plummets [Press Release]. See also BTI Consulting Group. (2008). The survey of client service: Performance for law firms: The BTI client service A-team (reporting only 34.6% of corporate counsel surveyed in 2007 would recommend their primary firm).

16


not particularly accurate forecasters of trial results. The neutral roles of judge and jury are associated with relatively consistent and predictable case evaluations; the roles of advocate and litigant, however, are marked by conflicting and inaccurate case assessments. The research regarding decision making in civil cases suggests that judgments about risks and consequences are altered when attorneys and clients assume partisan roles. People whose judgment is otherwise sound and whose predictions are otherwise accurate lose their acuity when they adopt the roles of advocate and litigant. Because judges once acted as attorneys and jurors have been or may become individual plaintiffs and defendants, it appears that attorneys and their clients are not permanently misaligned decision makers but may act that way when they become legal representatives or parties in actual cases. The fact that jurors’ opinions are consistent with experienced judges’ opinions also indicates that jurors’ verdicts are not wildcards but rather predictably reflect the values, rules and decision-making processes judges employ. Judges and juries, in short, seem to agree on what is the “right” result, but attorneys and clients in litigated cases have seriously disparate views of how cases should and will be resolved. David Donoghue, an intellectual property attorney and partner at Holland & Knight in Chicago, reflects on attorneys’ difficulties in predicting case outcomes and opines that law school education itself may contribute to the gap between attorneys’ predictions and jurors’ verdicts: As a child, my dad (a criminal defense attorney) routinely asked my family and me to predict the outcomes of his trials. We were usually correct. My dad was not. At some point during law school, I stopped being able to predict his case outcomes.The law changes how you think. Perhaps lawyers become too clouded with burdens of proof and rules of evidence to appreciate how a jury sees a trial. As a federal district court law clerk, I had a similar experience. I saw a number of trials and as we waited for the jury, we would often try to predict the results in chambers. The only people who reliably predicted the results were those without law degrees.21

Noting that “legal training hinders your ability to understand, persuade and communicate with juries” and attorneys usually have little in common, socially or economically, with jurors, Patricia Steele, a jury consultant, voices a similar sentiment: “Lawyers are skilled at many things, but understanding and connecting to jurors is generally not one of them.”22

21

Donoghue, David R. (2007, July 6). Juries get it right – 80% of the time. Chicago IP Litigation Blog. Available at http://www.chicagoiplitigation.com/2007/07/articles/legal-news/juries-get-itright-80-of-the-time/ 22 Steele, Patricia. (2006, Summer). To deal better with juries, stop thinking like a lawyer! Defense Comment, 21(2).


2.2.1

17

Judge-Jury Agreement

The conventional wisdom is that attorneys and their clients cannot accurately predict case outcomes because juries are unpredictable. Media coverage of celebrity trials bolsters this perception and invariably includes references to the O.J. Simpson murder trial and the multi-million dollar verdict in the McDonald’s spilt coffee case. Empirical research, however, does not support the conventional wisdom but rather demonstrates that jurors usually make deliberate, thoughtful, and intelligent decisions that comport with a judge’s opinion of what the verdict should be. “Lawyers entertain longstanding perceptions of the jury as biased and incompetent, relative to the judge,” writes law professor Kevin Clermont. But, after reviewing the extensive research regarding jury decision making, Professor Clermont concludes that lawyers’ perceptions of jurors’ ineptitude are groundless: “There is, however, no actual evidence that juries are relatively biased or incompetent.”23 Although the jury system may be a convenient scapegoat when trial strategies and forecasts go awry, it does not deserve the invectives thrown at it. The fact that some attorneys and clients do not comprehend how a jury will decide their cases does not mean that a jury’s methods and decisions are incomprehensible any more than a medieval craftsman’s inability to understand how the Romans built the largest unsupported concrete dome over the Pantheon proves that the Romans, too, had no idea of what they were doing. Although one may conclude that attorneys and clients are often poor forecasters it does not follow that jurors are unpredictable adjudicators. During the last 40 years, studies consistently demonstrate that jurors understand trial evidence and the applicable law, and judges agree with their verdicts in the large majority of cases. The first major study of judge-jury agreement, The American Jury, was published in 1966 by law professors Harry Kalven and Hans Zeisel. The study was based on questionnaires completed by more than 500 judges in 4,000 civil trials throughout the United States. The judges recorded their opinions of the difficulty of the case and how they thought it should be decided before the jury rendered its verdict, avoiding the possible effects of hindsight bias. In 78% of the cases, the judge agreed with the jury’s verdict. The cases on which they disagreed did not reflect any bias in favor of or against plaintiffs; in 10% of the cases the judge would have issued an award for the plaintiff when the jury’s verdict favored the defendant, and in 12% of the cases the judge would have ruled in favor of the defendant when the jury rendered a verdict for the plaintiff.24 “The central

23

Clermont, Kevin M. (2008, March 22). Litigation realities redux (Legal Studies Research Paper Series, Research Paper No. 08-006, p. 32). Cornell Law School. 24 Kalven, Harry, and Zeisel, Hans. (1996). The American jury. Vidmar, Neil. (1998). The performance of the american civil jury: An empirical perspective. Arizona Law Review, 40, 849, 853. See Kalven, Harry. (1964, October). The dignity of the civil jury. Virginia Law Review, 50(6), 1055, 1065–1066.

18


findings from Kalven and Zeisel’s research,” state law professors Neil Vidmar and Valerie Hans, “are that agreement between judge and jury was substantial and that most instances of disagreement could not be ascribed to jury incompetence or unwillingness to follow the law.”25 The 78% agreement rate between judges and juries is especially impressive when compared with decision making in other fields, explains Kevin Clermont: “this 78% agreement rate proves better than the rate of agreement on dichotomous decisions between scientists doing peer review, employment interviewers ranking applicants, and psychiatrists and physicians diagnosing patients, and almost as good as the 79% or 80% rate of agreement between judges themselves making sentencing decisions on custody or no custody in an experimental setting.”26 Kalven and Zeisel’s findings have been replicated in many other studies conducted after 1966; and in one study conducted by the University of Chicago Jury Project, the judges disagreed with the jury’s finding of liability in only 1% of the cases.27 Those other studies also show that a case’s complexity and the amount of expert witness testimony do not affect significantly the extent of judge-jury agreement, indicating that juror comprehension and jury verdict concordance with judges’ opinions do not diminish with case difficultness.28 Even in the specialized area of child support awards, which are determined by judges, experimental research shows that potential jurors reporting for jury duty “follow a predictable and rational course in their intuitive lawmaking” when presented with hypothetical child support cases. In 89% of the hypothetical child support cases, the jurors’ intuitive opinions about an appropriate amount of child support did not vary from the state guidelines by more than 20% although the jurors were unaware of the guidelines; and their average award in one hypothetical child support case “was coincidentally a perfect match” to the amount established by the state guidelines.29

25

Vidmar, Neil, and Hans, Valerie. (2007). American juries: The verdict (p. 149). Amherst, New York: Prometheus Books. 26 Clermont supra note 23 at 31. 27 Vidmar and Hans supra note 25 at 149–151. Galanter, Marc. The regulatory function of the civil jury. In Litan, Robert E., Ed. (1993). Verdict: Assessing the civil jury system (p. 70). Washington, D.C.: Brookings Institution Press. 28 Vidmar and Hans supra note 25 at 150. See Wissler, Roselle L., Hart, Allen J., and Saks, Michael J. (1999). Decision making about general damages: a comparison of jurors, judges and lawyers. Michigan Law Review, 98, 751. Heuer, Larry, and Penrod, Steven. (1994, February). Trial complexity: A field investigation of its meaning and its effects. Law and Human Behavior, 18(1), 29–52 (evidence complexity, legal complexity, and quantity of information were not significantly related to judge-jury verdict agreement). Robbennolt, Jennifer K. (2005). Evaluating juries by comparison to judges: A benchmark for judging. Florida State University Law Review, 32, 469, 477. 29 Ellman, Mark, Braver, Sanford L., and MacCoun, Robert. (2006). Intuitive lawmaking: The example of child support. Papers presented at the Second Annual Conference on Empirical Legal Studies, New York University, November 9–10, 2007. The potential jurors’ congruent opinions about child support awards contrast sharply with attorneys’ perceptions of child support awards.


2.2.2

19

Punitive Damages

Empirical research also challenges the popular conception that juries are more likely to award punitive damages or a higher amount of punitive damages than judges. Comparing judges’ decisions in 101 cases with jury verdicts in 438 cases, law professor Theodore Eisenberg and his colleagues concluded that “juries and judges award punitive damages in approximately the same ratio to compensatory damages.”30 Although he found some variability in the incidence of punitive damage awards, Eisenberg concluded that “the differences in punitive award rates more likely are a function of case selection than of juror’s relative harshness in bodily injury cases.”31 Eisenberg’s conclusions are consistent with studies by Thomas Eaton and Jennifer Robbennolt, who found that juries did not award punitive damages more often than judges and made similar decisions about the appropriate amount of damages.32 The empirical studies on juries’ decisions, in sum, “provide evidence of massive stability and consistency in jury decision making.”33

2.2.3

Judges’ Assessments of Juries

Judges not only agree with jurors’ verdicts in the vast majority of cases but they support their deliberative processes as well. Multiple surveys of more than 1,400 state and federal court judges demonstrate that judges respect both jurors’ capacity for objective evaluation and their sound judgment in rendering verdicts. The surveyed judges, report professors Vidmar and Hans, “gave very positive evaluations of the jury for its competence and its fairness” and “generally reported that the juries had made the correct decision and had had no difficulties applying the In interviews with attorneys representing parties in contested divorce cases, where child support and property division were disputed issues, the attorneys “report that they have difficulty discerning court standards and that they cannot predict the outcomes of court processes . . . Even the lawyers in our sample who do think there are set standards and who do say they can predict outcomes differ in their opinion of the content of those court standards; obviously, they cannot all be correct.” Erlanger, Howard S., Chamblis, Elizabeth, and Melli, Marygold S. (1987). Participation and flexibility in informal processses: Cautions from the divorce context. Law & Society Review, 21, 585, 599, cited in Galanter, Marc, and Cahill, Mia. (1994, July). Most cases settle: Judicial promotion and regulation of settlements. Stanford Law Review, 46, 1339, 1385. 30 Eisenberg, Theodore. (2006, July). Juries, judges and punitive damages: Empirical analysis using the civil justice survey of state courts 1992, 1996, and 2001 data. Journal of Empirical Legal Studies, 3(3), 293. Cf. Hersch, Joni, and Viscusi, W. Kip. (2004). Punitive damages: How judges and juries perform. Journal of Legal Studies, 33(1), 1–36. 31 Eisenberg supra note 30 at 293. 32 Vidmar and Hans supra note 25 at 311. 33 Galanter, Marc. The regulatory function of the civil jury. In Litan, Robert E., Ed. (1993). Verdict: Assessing the civil jury system, p. 83. Washington, D.C.: Brookings Institution Press.

20


appropriate standards to the case.”34 The curious conclusion is that, although attorneys and the public may perceive jurors as impressionable if not wayward, the most authoritative sources – the judges who actually weigh the evidence alongside them – consider their deliberations to be commendable, their verdicts fair.

2.2.4

Attorney-Jury and Attorney-Attorney Agreement

Shifting from the study of judge-jury agreement to attorney-jury and litigant-jury agreement, one finds large disparities between what attorneys and their clients expect to occur in a case and what actually happens at trial. These disparities are evident in experimental studies of hypothetical cases as well as data compiled from actual cases and are directly proportional to attorneys’ confidence levels. Attorneys with the highest level of confidence in their assessments tend to be the most poorly calibrated, i.e., most likely to be wrong in forecasting case outcomes. For cases that are settled rather than tried to verdict, the studies also demonstrate that attorneys have widely divergent views among themselves of what a case is worth and what is the appropriate amount of initial settlement demands and offers. These discordant case evaluations often reflect unrealistic settlement positions and ultimately break the strategic link between skillful bargaining and probable case outcomes. When negotiations collapse because settlement positions have no relation to likely outcomes, clients bear the cost of testing their attorneys’ case assessments at trial. Hence the adage, “Attorneys learn by trial and error – the client’s trial, the attorney’s error.”

2.2.5

Attorney-Litigant Negotiation Positions, Assessments and Outcomes

Two of the earliest studies of pre-trial negotiations and case evaluations were completed in the 1960s. In the first study, entitled “Predicting Verdicts in Personal Injury Cases,” Philip Hermann analyzed cases where the parties exchanged settlement offers and demands, failed to settle the cases, and proceeded to trial. Comparing the plaintiff’s last settlement demand and the defendant’s last offer with the actual trial verdict in 443 personal injury cases, they discovered that the attorney’s settlement posture bore little relation to the actual trial value of the cases. Only onesixth of the demands and offers were within 25% of the verdict. The attorneys and the insurance companies, Hermann observed, were “equally wild” in guessing the value of their cases.35 34

Vidmar and Hans supra note 25 at 151. Galanter supra note 33 at 83.

35


21

In the second experiment, conducted by Douglas Rosenthal, the settlement amount negotiated by attorneys in pre-trial settlements was compared with the settlement valuation prepared by an independent panel of experts. The panel was comprised of two lawyers who usually represented plaintiffs, two claims agents for insurance companies, and one lawyer who was experienced in representing plaintiffs and insurance companies. Rosenthal thought that “a comparison of the actual recoveries with the mean panel evaluation would provide one relatively ‘objective’ empirical measure of the competence of professional service received by personal injury claimants.”36 He found that 40% of the cases were settled for less than twothirds of the case’s settlement value, as determined by the expert panel, and overall the settlement amounts varied from one-sixth of the panel’s valuation to twice their valuation.37 Rosenthal concludes, “In 77% of the cases (44 of 57) clients did worse than they should have according to the arithmetic means of the values assigned to their claims by each of the five panelists.”

2.2.6

Disparities In “Same Case” Evaluations and Outcomes

Following Hermann and Rosenthal’s studies in the 1960s, empirical research continued to demonstrate high variability among attorneys in evaluating cases and negotiating settlements. Professor Gerald Williams’ experiment with practicing attorneys, published in 1983, is regarded as one of the early and noteworthy attempts to use the “same case” method to simulate actual settlement negotiations among attorneys. In that experiment, designed to replicate pre-trial negotiations between practicing attorneys in a personal injury case, Williams assigned 40 practicing lawyers to 20 teams and randomly designated the attorneys on each team as the attorney for the plaintiff or the defendant. All attorneys read the same case facts, were informed that the case would be tried to a jury in Des Moines, Iowa, and were notified that the results of their negotiations, along with their names, would be published. Each attorney, moreover, received copies of jury awards in comparable cases tried to verdict in the Des Moines area. Despite the fact that all attorneys received identical case information and could have learned the outcomes in similar cases, their negotiation positions and settlements were astonishingly dissimilar. Attorneys assigned the plaintiff’s attorney role initiated settlement negotiations with demands ranging from $32,000 to $675,000, and attorneys in the defense role made opening offers ranging from $3,000 to $50,000. The amount of the ultimate settlement negotiated for their hypothetical clients varied from a low of $15,000 to a high of $95,000 – all for the same injuries in the same case in the same jurisdiction.38 36

Rosenthal, Douglas E. (1974). Lawyer and client: Who’s in charge (p. 59). New York: Russell Sage Foundation. 37 Galanter supra note 33 at 83. 38 Galanter supra note 33 at 81–83.

22

2.2.7


Comparisons of Predictions and Outcomes

Five years after the publication of Williams’ results, psychology professors Elizabeth Loftus and Willem Wagenaar studied actual predictions attorneys made regarding civil cases they expected to proceed to trial. They asked attorneys to record what they thought would be a good result for their client and the probability of obtaining the desired result. After the cases were resolved – and in many cases after considerable prodding – the attorneys reported the actual results to Loftus and Wagenaar. Comparing the attorneys’ goals and levels of confidence with the actual outcomes, Loftus and Wagenaar concluded that attorneys’ forecasts were poorly calibrated and “in general lawyers were overconfident in their chances of winning, especially so in cases in which they had been highly confident to begin with.”39 Loftus and Wagenaar observed that forecasting accuracy is particularly important for attorneys and clients because “we are concerned here with the possibility that erroneous predictions about an uncertain future might lead a lawyer to make the wrong decision about whether to proceed with further litigation or to settle.”40 They posit five possible explanations why attorneys are confident yet inaccurate forecasters: (1) few attorneys keep a record of their actual forecasting accuracy or their trial win rates; (2) the few attorneys who do keep track of their performance “might record how often they lose, but fail to fully analyze what went wrong;” (3) attorneys may systematically neglect important predictors, relevant law precedents and the personality of the judge; (4) lawyers may “need to feel and display overconfidence in order to attract clients, and, later, to keep those clients convinced that their interests are well served;” and (5) lawyers may tend to recall “a similar case in which a favorable verdict was achieved and ignore “similar cases in which unfavorable verdicts were achieved.”41 Overconfidence, Loftus and Wagenaar hypothesize, may help lawyers maximize their courtroom performance and persuasiveness but adversely affects their settlement evaluations and negotiations. A later study, also conducted by Loftus and her colleagues, tested the accuracy of lawyers’ predictions about whether their client would prevail at trial. She found that the lawyers’ judgments “showed no predictive validity” and were “hardly above chance.”42 The attorneys generally exhibited a marked “overextremity bias (underprediction of success for low probabilities and overprediction of success for

39

Loftus, Elizabeth F., and Wagenaar, Willem A. (1988, Summer). Lawyers’ predictions of success. Jurimetrics, 28, 437. 40 Id. at 441. 41 Id. at 450. 42 Goodman-Delahunty, J., Granhag, P.A. & Loftus, E.F. (1998). How well can lawyers predict their chances of success? Unpublished manuscript. University of Washington. Cited in Koehler, Derek J., Brenner, Lyle, & Griffin, Dale. (2002). The calibration of expert judgment: Heuristics and biases beyond the laboratory. In Gilovich, Thomas, Griffin, Dale, & Kahneman, Daniel (Eds.). (2002). Heuristics and biases: The psychology of intuitive judgment (pp. 705, 706). Cambridge: The Press Syndicate of the University of Cambridge.


23

high probabilities).”43 When she specifically analyzed the predictions of lawyers retained on a contingency basis, she observed that their predictions of success were especially biased by over-optimism. Although the contingency fee lawyers exhibited the same level of confidence about case outcomes as other lawyers, the contingency fee attorneys in her study won only 42% of their cases compared with an overall 56% win rate.

2.2.8

Damages Award Predictions

When researchers focus on projected damages awards instead of win/lose predictions, they discover that jurors’ opinions about damages are a close match with judges’ opinions and, in some instances, jurors’ collective judgment is less variable than the opinions of individual judges and attorneys. In two experiments conducted by Neil Vidmar the same hypothetical case facts were presented to lawyers, judges and citizens who had reported for jury duty, and they were asked to estimate the appropriate amount of damages to award to the plaintiff. (Liability was admitted in the hypothetical case). In both experiments, the jurors’ average estimated award was less variable than the individual judges’ estimated awards. Overall, Vidmar reports, “the findings show variability among legal professionals and hint that juries may produce more stable estimates of the community’s evaluation of the injury than a single judge acting alone.”44 About four years after the publication of Vidmar’s research regarding damages awards, Roselle Wissler and her colleagues completed a large-scale study designed to determine the degree of variability among judges, jurors and lawyers in assessing liability and awarding damages. In Wissler’s study, 1,060 judges, jurors and lawyers in two different states were presented with 62 case summaries based on actual personal injury cases. After hearing the case summary, the survey respondents were asked to state the amount of money they would award to the plaintiff, the amount of the award they thought an average juror would award, and their rating of the plaintiff’s injury in five aspects, e.g., overall severity. Reviewing the data, Wissler states, the “dominant theme of these findings is one of considerable similarity across the various groups of decisionmakers in the structure of thinking about injury severity and awards. Most importantly, an impressive similarity exists in the injury attributes that drive their decisions, the weight given to those attributes, and the shared sense of vertical equity held by jurors, judges, plaintiffs’ lawyers, and defense lawyers alike.”45

43

Id. Vidmar and Hans supra note 25 at 301. 45 Wissler, Roselle L., Hart, Allen J., and Saks, Michael J. (1999). Decisionmaking about general damages: A comparison of jurors, judges and lawyers. Michigan Law Review, 98, 751, 812. 44

24


Wissler’s study found that the most significant deviation in opinions about the severity of plaintiffs’ injuries was not between juries and judges but between defendants’ lawyers and the other respondents: Indeed, if any one group emerges as being out of step with all of the others, it is defense lawyers. This has somewhat paradoxical implications. When members of the defense bar evaluate the performance of jurors, and gauge them to be off the mark, these lawyers no doubt reach that assessment by comparing the jurors’ conclusions to their own. But their own impressions of injuries are the ones that depart most from the pattern shown by the other decisionmaking groups, at least in regards to judgments of injury severity.46

She notes that the severity of injury is consistently one of the strongest predictors of monetary awards, and defense lawyers’ opinions about the amount of awards may be “less responsive to case details” and “more mechanical.”47

2.2.9

Overview of Judge, Jury and Attorney Decision Making

The research regarding the decisions of judges, jurors and lawyers does not substantiate the strong criticisms frequently leveled at jurors – that, in Jerome Frank’s accusation, jurors are “uncertain, capricious, and unpredictable, ignorant and prejudiced, poor factfinders, gullible and incapable of following complex legal rules, thus making ‘the orderly administration of justice virtually impossible.’”48 To the extent that studies compare the decisions of jurors with judges and attorneys, the studies show “considerable similarity” in overall case assessments or, in some instances, higher variability among attorneys than judges and jurors. When studies compare attorneys’ case outcome predictions with actual trial results, attorneys appear to be over-confident, inaccurate forecasters; and when attorneys’ negotiation positions and settlement amounts are compared with those of other attorneys, the variances are broad in scope and deep in implications. “What,” asks law professor Marc Galanter, “are we to make of this persistent and sizable variability and error in lawyers’ readings of potential outcomes?”49

2.2.10 Attorney-Litigant Decision Making in Actual Cases Professor Galanter’s question is both answered and amplified by three studies comparing trial awards with rejected settlement proposals to determine whether litigants recovered more money at trial than they were offered in settlement 46

Id. at 805. Id. at 758, 794. 48 Id. at 753. 49 Galanter supra note 33 at 83. 47


25

negotiations. These studies assess whether, in deciding to try cases instead of settling them, attorneys and their clients are making financially advantageous decisions, and if not, what factors influence their decision to forego settlement. The three studies are reported in law professors Samuel Gross and Kent Syverud’s 1991 article, “Getting to No: A Study of Settlement Negotiations and the Selection of Cases for Trial,” their 1996 study, “Don’t Try: Civil Jury Verdicts in a System Geared to Settlement;” and law professor Jeffrey Rachlinski’s 1996 study, “Gains, Losses and the Psychology of Litigation.”50 In these three studies the authors analyzed settlement behavior in actual civil cases and concluded that the conventional model of rational decisions leading to optimal economic outcomes is inapplicable, misleading, or inaccurate. Noting that “the absence of data on pretrial negotiations has handicapped development of this topic,” law professors Gross and Syverud first studied a nonrandom sample of 529 cases between June 1985 and June 1986. Their data showed that “the main systemic determinants of success at trial and in pretrial bargaining are contextual and relational [e.g., litigants’ resources, reputations, insurance, fee arrangements, repeat litigants]” and that prior theoretical models of attorney/litigant settlement behavior were “quite alien to actual litigation.”51 Attorneys and clients, Gross and Syverud found, were not rational, utility-maximizing actors and “win-win” trials were rare. Only 15% of the trials they studied resulted in an award for a plaintiff that was greater than defendant’s settlement offer but less than plaintiff’s demand, “and in some of these cases the entire gain for one side, or both, will have been consumed by the trial costs.”52 Gross and Syverud’s study directly challenged a prior theoretical model of litigation posited by law professors George Priest and Benjamin Klein: “the fifty percent implication.”53 According to Priest and Klein’s theory, trials occur primarily in “close cases,” plaintiffs and defendants are equally adept in predicting trial outcomes, plaintiffs will win about 50% of the cases that proceed to trial, and “mistakes” about outcomes will be evenly distributed between plaintiffs and defendants. Priest and Klein’s hypothesis, however, turns out to be inconsistent with the data compiled by Gross and Seyverud: Economic theories of trial and pretrial bargaining call to mind the standard image of a competitive market: numerous individuals intelligently pursuing independent self-interests.

50

Gross, Samuel, & Syverud, Kent. (1991). Getting to no: A study of settlement negotiations and the selection of cases for trial. Michigan Law Review, 90, 319. Gross, Samuel, & Syverud, Kent. (1996). Don’t try: Civil jury verdicts in a system geared to settlement. UCLA Law Review, 44, 1, 51. Rachlinski, Jeffrey. (1996). Gains, losses and the psychology of litigation. Southern California Law Review, 70, 113. 51 Gross & Syverud (1991), supra note 50 at 330, 379. 52 Gross & Syverud (1991), supra note 50 at 379. 53 Priest, George L. & Klein, Benjamin. (1984). The selection of disputes for litigation. Journal of Legal Studies, Vol. 13:1. Priest, George L. (1985). Reexamining the Selection Hypothesis. Journal of Legal Studies, Vol. 14:215.

26

2 Prior Research on Attorney-Litigant Decision Making Social reality, as usual, is inconsiderate of global theories. In this case it provides a competing image that is less susceptible to statistical prediction: stragglers picking their way in the dark, trying to avoid an occasional land mine.54

Presaging a broader application of behavioral economics theories (discussed in Chapter 4) to attorney-litigant settlement behavior, Gross and Syverud observed that plaintiffs usually are more risk averse than defendants; plaintiffs and defendants attach “separate values to each possible outcome;” and “their stakes may be unequal (or equal) with respect to victories, or defeats or both.”55 In their second study, Gross and Syverud added a sample of 359 cases reported between 1990 and 1991. Their results again conflicted with the Priest-Klein litigation model. Instead of a 50/50 distribution of “mistakes,” Gross and Syverud found that plaintiffs were more likely than defendants to make a decision-making mistake, that is, rejecting a settlement proposal which turned out to be the same as or more favorable than the actual trial award. Plaintiffs were “clear losers” in 61% of the cases in their first sample (1985–1986) and 65% of the cases in their second sample (1990–1991). The defendants, in contrast, made mistakes in only 25% and 26%, respectively, of the cases in the two samples. In the third major empirical study of attorney-litigant decision making in adjudicated cases, Rachlinski compared final settlement offers with jury awards in 656 cases. His data showed decision error by plaintiffs in 56.1% of the cases, contrasted with defendants’ decision error rate of 23%. Although plaintiffs’ decision error rate was markedly higher than defendants’ decision error rate, the average cost of plaintiffs’ decision error ($27,687) was dramatically lower than defendants’ mean cost of error ($354,900). Observing that litigants’ decisions are “suboptimal” and “may not comport with rational theories of behavior,” Rachlinski found that the “consistently divergent risk preferences between plaintiff and defendant” could be explained by behavioral economics’ framing theories.56 Litigants’ “risk preferences depend upon characterizing a decision as a gain or loss” and “vary systematically depending upon whether they are in the role of plaintiff or defendant.”57 Plaintiffs are consistently risk averse, while defendants are risk taking. Consequently, plaintiffs generally benefited from litigation and “defendants as a class paid heavily for their decision” to litigate: “When settlement negotiations failed, the plaintiffs were unwittingly forced to undertake a risk that, on average, benefited them and cost the defendants dearly.”58

54

Gross & Syverud (1991) supra note 50 at 385. Gross & Syverud (1991) supra note 50 at 381. 56 Rachlinski supra note 50 at 114, 118, 120, 142. 57 Rachlinski supra note 50 at 119. 58 Rachlinski supra note 50 at 160. 55

2.3 Chapter Capsule

27

2.2.11 Kiser, Asher and McShane Study of Attorney-Litigant Decision Making The continuing viability of the Gross and Syverud and Rachlinski studies was tested in 2008 by a large-scale analysis of attorney-litigant decision making. The results of that analysis appear in an article entitled, “Let’s Not Make A Deal: An Empirical Study Of Decision Making In Unsuccessful Settlement Negotiations,” co-authored by Randall Kiser of the decision services company DecisionSet1 and Martin Asher and Blakeley McShane of The Wharton School. The article, which describes the results of the largest multivariate analysis of attorney-litigant decision making, was published on behalf of Cornell Law School and The Society for Empirical Legal Studies in the Journal of Empirical Legal Studies, Vol. 5, No. 3 (September 2008). Six key findings of this study of 4,532 actual cases are: (1) 61% of plaintiffs and 24% of defendants obtained a result at trial that was the same as or worse than the result that could have been obtained through a pre-trial settlement; (2) the average cost of these decision errors was $43,100 for plaintiffs and $1,140,000 for defendants during the 2002–2005 period; (3) the incidence of decision errors increased and the cost of these errors soared between 1964 and 2004; (4)“Context” variables (systemic factors like case type) are more predictive of adverse trial outcomes than “Actor” variables (personal factors like attorney experience and law school ranking); (5) statutory cost-shifting procedures, intended to encourage settlement by financially penalizing parties whose trial result is worse than the settlement offer made by an adversary, may provoke rather than deter risk-taking behavior; and (6) parties who are represented by attorneys with mediation training experienced a lower incidence of decision error. The results of the Kiser, Asher and McShane study are remarkably consistent with the earlier results reported by Gross and Syervud and Rachlinski, evidencing consistent patterns of high plaintiff decision error rates and high costs of defendant decision error.

2.3

Chapter Capsule

Although the legal services industry assumes a large role in the American economy, scant data exists regarding the economic benefits of legal services expenditures, the accuracy of attorneys’ advice and the effectiveness of their representation in lawsuits. The legal industry has lagged behind other businesses and professions in establishing metrics to measure costs and assess performance. Limited research regarding attorney-litigant decision making indicates that attorneys and clients are over-optimistic in evaluating their cases and predicting trial outcomes. Four studies by three independent research teams found that, when cases proceed to trial, most plaintiffs obtain an award that is less than the defendant’s settlement offer. Defendants, for their part, obtain a worse result at trial than could have been achieved by accepting a plaintiff’s settlement demand in

28


23% – 26% of their cases. Although plaintiffs exhibit a higher incidence of adverse trial outcomes, the average cost of defendants’ adverse outcomes is dramatically higher than plaintiffs’ average cost. Adverse trial outcomes often are blamed on unpredictable and erratic jurors, but empirical research demonstrates that jurors generally discharge their duties faithfully, responsibly and intelligently. Although attorneys may hold widely divergent ideas about case settlement values and likely trial outcomes, judges and juries have similar opinions about how cases should be decided. The high degree of judge-jury agreement suggests that case evaluation may be clouded when clients and attorneys assume partisan roles.59

59

Some sentences in this chapter are excerpted with permission from the author’s article, “Let’s not make a deal: An empirical study of decision making in unsuccessful settlement negotiations” (co-authored with Martin Asher and Blakeley McShane), Journal of Empirical Legal Studies, 5(3), 551–591, published by Wiley Periodicals, Inc.

Chapter 3

A Current Assessment of Attorney-Litigant Decision Making In Adjudicated Cases

I often say that when you can measure what you are speaking about and express it in numbers you know something about it; but when you cannot express it in numbers your knowledge is a meager and unsatisfactory kind; it may be the beginning of knowledge but you have scarcely, in your thoughts, advanced to the stage of science, whatever the matter may be. Lord Kelvin (William Thomson 1824–1907), Lecture to the Institution of Civil Engineers (May 3, 1883)

This chapter tests the “crapshoot” theory of litigation, the popular belief that trial outcomes are inherently unpredictable and litigants proceed to trial at their own peril. In his article, “Forget Fair; It’s Litigation as Usual,” New York Times business columnist Joe Nocera epitomizes this widely held belief. Commenting on drug manufacturer Merck’s settlement of the mass tort litigation regarding the painkiller Vioxx, he writes, “And finally, when you get right down to it, litigation is a crapshoot, and it can be cruelly unfair.”1 The belief that trials are crapshoots is part of the American ethos. If judges and juries were perceived to be predictable, fair, and dutiful, novels like John Grisham’s The Runaway Jury would search desperately for a gullible reader instead of soaring to the top of the bestseller lists. Cynicism about jurors’ deliberative processes and the resultant suspicion that many verdicts are arbitrary, ill-considered, and tainted with prejudice is so engrained in the national belief system that even a selfdescribed victim of the jury system can wax philosophical about his murder conviction. Michael Quartararo, who claims he was wrongfully convicted of killing a 13-year-old boy in 1981 by shoving pebbles down his throat, wrote this review of The Runaway Jury from his jail cell: Most trial lawyers will tell you that juries are completely unpredictable, that selecting a jury is a roll of the dice, that you can’t predict what they’ll decide. And so, for completely irrational reasons, it is increasingly possible that a jury could run away with a verdict.

1

Nocera, Joe. (2007, November 17). Forget fair; It’s litigation as usual. The New York Times.


29

30

3 A Current Assessment of Attorney-Litigant Decision Making In Adjudicated Cases Perhaps this is why most civil cases are settled out of court, and most defendants plead guilty. They, like the characters in Grisham’s book, fear the Runaway Jury.2

Outside of his jail cell, Michael Quartararo would find an ideationally sympathetic audience with corporate counsel, 57% of whom gave negative responses to the question, “Overall, how would you describe the fairness and reasonableness of state court liability systems in America – excellent, pretty good, only fair or poor?”3 His concerns also seem to be shared by the American public, the majority of whom declared in an American Bar Association survey that “the justice system needs a complete overhaul.”4 If trials truly are crapshoots and it’s impossible to accurately predict case outcomes, as business columnists, convicted murderers, corporate counsel and the public seem to believe, trial outcomes would be randomly distributed. Chance would determine trials wins and losses – divided evenly between plaintiffs and defendants – and both plaintiffs and defendants would make a similar number of mistaken settlement decisions. If verdicts are inherently unpredictable, plaintiffs and defendants’ negotiation positions also would show equivalent deviations from the actual outcomes, and they would bear equally the financial losses resulting from mistaken settlement decisions. This follows from the fact that random adverse events, by definition, wreak havoc indiscriminately on their victims. If, for instance, aviation was inherently dangerous and plane crashes were wholly inevitable and unpredictable, crashes would occur randomly, unaffected by weather conditions, terrain, pilot experience, or air traffic controller training. Like litigants in a courtroom, airline passengers choosing to be seated in a dangerous and unpredictable environment would simply “pay their money and take their chances,” knowing they are powerless in the grip of fate. After decades of intensive investigations and statistical analysis, however, experts have learned that airplane crashes are not random and certain factors are predictive of the frequency of crashes and the extent of passenger injury and death. Because of this information and other advances in airline safety, the number of fatalities in aircraft accidents has declined from 16.7 deaths per billion miles flown in 1946–1950 to 0.14 in 1996–2000.5 Although judges and juries have been deciding cases for centuries longer than pilots have been flying airplanes, surprisingly little effort has been made to ascertain the determinants and reduce the

2

Quartararo, Michael. Review, The runaway jury. Bookreporter.com, retrieved at http://www. bookreporter.com/reviews/0440221471.asp. 3 Taylor, Humphrey. (2002, February 20). Survey of lawyers finds big differences in their perceptions of courts, judges and juries in 50 states. The Harris Poll1 #9. Available at http://www. harrisinteractive.com/harris_poll/index.asp?PID¼286. 4 Greenhouse, Linda. (1999, February 24). 47% in poll view legal system as unfair to poor and minorities. The New York Times, p. A12. 5 Federal Reserve Bank of Dallas. Exhibit 10. In 2001 Annual Report: Taking Stock in America. Between 1997 and 2007, accident rates dropped by 65%, from one fatal accident in about 2 million departures in 1997 to one fatal accident in about 4.5 million departures in 2007. Wald, Matthew. (2007, October 1). Fatal airplane crashes drop 65%. The New York Times, p. C1.

3.1 The Fifty Percent Implication

31

incidence of adverse litigation events. The belief persists that, while pilot and mechanical error may be susceptible of rigorous investigation and precise analysis, the decisions of judges, jurors, attorneys and clients are essentially idiosyncratic and impenetrable by scientific methods and quantitative analysis. Unlike other soft disciplines, law has never experienced physics envy.

3.1

The Fifty Percent Implication

The belief that trial outcomes are random and hence adverse verdicts and settlement mistakes are evenly balanced among plaintiffs and defendants is so consistent with intuition and conviction that it was embodied in an academic theory in 1984 and named the “fifty percent implication.”6 The “fifty percent implication,” posited by law professors George Priest and Benjamin Klein, is a logical outgrowth of the assumption that the cases that proceed to trial are “wildcards.” According to Priest and Klein’s theory, trials occur primarily in “close cases,” and plaintiffs and defendants are “equally successful at predicting the outcomes of the cases.”7 Therefore, plaintiffs will win and defendants will lose about 50% of the cases litigated to verdict, and “mistakes” about outcomes will be evenly distributed between them. Specific characteristics of the case, the parties and the attorneys, under the 50% implication, should not affect the random distribution of settlement decision errors and win rates. Priest and Klein note that “the most important assumption of the model is that potential litigants form rational estimates of the likely decision, whether it is based on applicable legal precedent or judicial or jury bias.”8 This assumption implies that litigants will consider not only how the case should be decided legally, as a matter of legal precedent, but also how it will be decided practically, as a result of any human biases. Their fifty percent implication further assumes that litigation costs are relatively high compared to settlement costs, the application of legal standards is predictable, both parties can evaluate outcomes with “equal precision,” and the stakes are “symmetrical” to the parties, i.e., gains and losses from litigation “are equal to both parties.”9 Possessing equivalent knowledge of likely outcomes and burdened with equal costs and risks, plaintiffs and defendants, Priest and Klein concluded, would experience the same number of trial wins and losses and make erroneous settlement decisions with the same frequency and equivalent costs.

6

Priest, George L. & Klein, Benjamin. (1984). The selection of disputes for litigation. Journal of Legal Studies, 13, 1. Priest, George L. (1985). Reexamining the selection hypothesis. Journal of Legal Studies, 14, 215. 7 Gross, Samuel, & Syverud, Kent. (1991). Getting to no: A study of settlement negotiations and the selection of cases for trial. Michigan Law Review, 90, 319, 325. 8 Priest & Klein supra note 6 at 4. 9 Priest & Klein supra note 6 at 5, 12, 14, 19, 20, 24.

32

3 A Current Assessment of Attorney-Litigant Decision Making In Adjudicated Cases

This chapter tests the validity of the fifty percent implication and attempts to answer four questions: l l l l

Do plaintiffs win 50% and do defendants lose 50% of the cases that proceed to trial? Are plaintiffs and defendants equally adept at predicting case outcomes? Do plaintiffs and defendants bear equally the costs of adverse settlement decisions? Do the settlement positions taken by plaintiffs and defendants reflect rational assessments of case risks and strategic bargaining framed by mutually acknowledged outcome ranges?

In answering these questions, this chapter explains whether certain characteristics of a case, the parties, and their attorneys are correlated with low win rates, suboptimal settlement decisions, disproportionate costs resulting from erroneous decisions, and settlement positions grossly disproportionate to the ultimate trial outcome. As will be shown below, the assumptions and predictive capacity of the Priest and Klein model are challenged by the study data showing that win rates, decision error rates and settlement negotiation positions vary widely with different types of cases, clients, and attorneys.

3.2

New Data

To answer the key questions regarding settlement decisions and case outcomes, this chapter analyzes 5,653 actual civil litigation cases in which approximately 14,250 attorneys represented the disputants. Since both a plaintiff and a defendant had to make a decision to accept or reject an adversary’s settlement proposal in each of the 5,653 cases, the total number of decisions dissected is 11,306. This chapter’s reliance on 11,306 actual case decisions distinguishes it from most other studies in nature and scope. Although scholars have produced hundreds of excellent research articles regarding negotiation theory and simulated settlement negotiations and trials, relatively few studies examine the characteristics and decisions of litigants and their attorneys in actual settlement negotiations, and none, to date, analyze as many decision-making variables and how those variables relate to case outcomes. Unlike most negotiation theories and studies, this study is empirical, not experimental, and it endeavors to advance prior decision-making research by examining settlement decisions made by individuals, governmental entities, and businesses facing real financial consequences. The study compares settlement offers and demands with the actual verdicts rendered in the litigants’ cases, revealing the frequency, correlates and costs of ineffective settlement decisions.

3.2.1

The Four Datasets

The 5,653 cases analyzed in this chapter are divided among four different datasets. The primary dataset consists of 2,754 civil cases adjudicated in California state

3.2 New Data

33

courts or decided by California arbitrators. These cases, reported in VerdictSearch California during the 58-month period between November 2002 and August 2007, include about 23% of all civil cases tried by juries during that period.10 The second dataset serves as a shadow study for the primary dataset; it contains 524 civil litigation cases filed in New York and reported in VerdictSearch New York during the 12-month period between January 1, 2004, and December 31, 2004. This second dataset allows the California litigants’ settlement positions and trial results to be compared with the positions taken and results experienced by their counterparts in New York during a shorter but comparable timeframe. The third dataset is a 40-year survey of settlement decisions and trial results in California cases adjudicated between 1964 and 2004. The 40-year survey provides an historical context to evaluate whether attorney-litigant decision error rates are constant and whether the frequency and costs of adverse outcomes in the 58-month California study and the 12-month New York study are typical. The historical survey also seeks to answer the question, “Are attorneys and their clients getting better or worse at predicting trial outcomes in cases that do not settle?” The fourth dataset differs from the other datasets by focusing on particular types of attorneys instead of time periods or states. It consists of California cases where attorneys who also serve as court-appointed mediators in other cases represented the parties in the dataset cases. In most cases, these “attorney-mediators” are parttime mediators and continue to represent clients in their regular litigation practices; in a small number of cases, the mediators now serve as full-time mediators but represented clients in litigated cases before they became full-time mediators. The attorney-mediators in this dataset met state-mandated mediator training requirements, including 30 minimum hours of classroom and experiential training in conflict resolution, and served on their local court’s panel of mediators. Because these attorney-mediators presumably are more objective in case evaluation and more skilled in conflict resolution than the average attorney, they may be expected to exhibit lower decision-making error rates. Alternatively, if the attorney-mediator decision error rates are similar to other practitioners, one may tentatively conclude that mediation training and experience do not appear to affect attorneys’ case evaluation skills or risk-taking propensities – or that clients, not their attorneys, call the shots and the attorneys themselves may be fungible.

3.2.2

VerdictSearch Publications

The cases in all four datasets were initially reported in the weekly trade publication VerdictSearch California, previously titled California Jury Verdicts Weekly, and 10

The total number of civil cases tried by juries in California state courts between Fiscal Year 2002– 03 and Fiscal Year 2006–07 is 11,208. Source: Administrative Office of the Courts’ 2008 court statistics report – statewide caseload trends. See Levey, Dhyana. (2009, February 13). For the vanishing civil trial, report shows another down year. Daily Journal Verdicts and Settlements, p. 1.

34


VerdictSearch New York. VerdictSearch California and VerdictSearch New York are published by New York Law Publishing Company, which is owned by American Lawyer Media (ALM). Similar jury verdict reports are published by ALM for cases filed in Texas, Florida, Illinois, New Jersey and Pennsylvania. (ALM also publishes major trade magazines and newspapers like The American Lawyer, Corporate Counsel, The National Law Journal, New York Law Journal, and Legal Times.) These reports are available by subscription (hard copy and electronic), and older paper volumes are bound and shelved in law libraries throughout the country. VerdictSearch California is the primary reporting source for judgments and settlements in California, and the reliability of its reports has been confirmed in law review articles and by research studies.11 Gross and Syverud, for instance, “concluded that the information contained in the journal is reliable and found no systematic bias among the errors by either plaintiff or defendant to misreport the winning party, the size of the award, or the settlement offers.”12 The RAND Corporation also utilized the data reported in VerdictSearch California to prepare its periodic reports on jury trials and verdicts in major metropolitan areas, including its 1996 report, “Trends in Civil Jury Verdicts Since 1985.”13 As recently as November 2007, Seth Seabury, an economist at RAND, was reporting on trends in civil litigation awards, relying on data from the RAND Jury Verdicts Database (JVDB). The data in RAND’s JVDB is derived from the jury verdict reports.14

3.2.3

Case Database Selection Criteria

VerdictSearch California does not report every verdict rendered in California but relies on voluntary submissions from attorneys and solicits reports based on court dockets and trade publications.15 The information VerdictSearch California obtains

11

Gross & Syverud supra note 7 at 319. Rachlinski, Jeffrey. (1996). Gains, losses and the psychology of litigation. Southern California Law Review, 70, 113. Peterson, M.A., & Priest, G.L. (1982). The civil jury: Trends in trial and verdicts, Cook County, Illinois, 1960–1979 (Publication No. R-2881). Santa Monica, California: RAND Institute for Civil Justice. Shanley, M.G., & Peterson, M.A. (1983). Comparative justice: Civil jury verdicts in San Francisco and Cook Counties, 1959–1980 (Publication No. R-3006). Santa Monica, California: RAND Institute for Civil Justice. 12 Rachlinski supra note 11 at 149, fn. 133. 13 Moller, Erik. (1996). Trends in civil jury verdicts since 1985. Santa Monica, California: RAND Institute for Civil Justice. 14 Seabury, Seth. (2007, July 5). Inferring beliefs from selected samples: Evidence from civil litigation. Papers presented at the Second Annual Conference on Empirical Legal Studies, November 9–10, 2007, New York University. 15 Contract cases appear to be under-reported, low monetary value cases presumably are underreported, and settlement demands and offers subject to confidentiality provisions or statutes are, of course, not reported.

3.2 New Data

35

from attorneys, including the factual contentions, damages, results and settlement offers, is compiled in a draft case report. To confirm the contents of the draft case report, VerdictSearch California then attempts to contact counsel for all parties by facsimile and telephone. All information received from the parties’ attorneys, VerdictSearch California affirms, is incorporated in the case report. Cases reported in VerdictSearch California during the 58-month core study period were included in the study database if they met five basic requirements: (1) a jury’s verdict, judge’s decision, or arbitrator’s award was entered in a specific monetary amount; (2) the plaintiff submitted a settlement demand in a specific monetary amount; (3) the defendant made a settlement offer in a specific monetary amount or its settlement offer was described as “none;” (4) there was no reported disagreement among the parties regarding the amount of the ultimate result and the parties’ prior settlement positions; and (5) the parties were represented by counsel. The core study database thus is limited to documented cases in which the parties conducted unsuccessful settlement negotiations and the parties’ liability, if any, was ultimately decided by a judge, jury or arbitrator. The database excludes a few cases that otherwise might satisfy the five requirements above. Cases terminated on technical or procedural grounds, e.g., motions for nonsuit, directed verdict, summary judgment, and judgment notwithstanding the verdict, were excluded.16 The outcome in those cases is a matter of law, as opposed to an attorney-client decision about mixed and disputed issues of both fact and law. Class actions also are excluded from the database because the relationship between attorneys and clients in those cases is too attenuated to assess attorney-client decision making. Cases in which typographical or reporting mistakes appeared on the face of the report or the parties’ settlement positions were not adequately allocated among multiple parties were eliminated.

3.2.4

Attorneys in Dataset

What percentage of all California litigation attorneys is represented in the primary dataset? The primary dataset includes 2,754 cases and 6,945 attorneys. The attorneys are evenly divided between those representing plaintiffs (3,492) and those representing defendants (3,453). These attorneys represent an estimated 22% – 27.5% of all California litigation attorneys, although a precise count is elusive because the exact number of litigation attorneys in California has never been ascertained. This task is complicated by intra-profession disagreements as to whether attorneys who dabble 16

About 9% of all cases reported in VerdictSearch California during the subject period included a reference to “summary judgment,” “directed verdict,” “judgment n.o.v.,” “judgment notwithstanding the verdict,” or “nonsuit,” although the actual ruling in the case, if any, is undetermined. In a different study of federal cases, Eisenberg found that “summary judgment, judgment on pleadings, motion before trial” comprised 12.14% of all dispositions. Eisenberg, Theodore, & Lanvers, Charlotte. (2008, November 21). What Is the Settlement Rate and Why Should We Care? Cornell Legal Studies Research Paper No. 08–30. Available at SSRN: http://ssrn.com/abstract¼1276383.

36


in litigation should be considered real “litigators” and whether real litigators are something different from trial lawyers. Although the State Bar of California does not maintain records regarding the precise number of civil litigation attorneys in California, 16% of the attorneys responding to its February 2006 survey identified civil litigation as their primary area or field of practice.17 Forty-five percent of all surveyed attorneys indicated they had a “secondary area of legal practice,” and among that group 14% designated “civil litigation” as the secondary area. When asked what State Bar section the members belonged to, however, only 7% of all surveyed members designated “litigation.”18 Because the State Bar information is ambiguous, the data reported by MartindaleHubbell, the largest directory of attorneys, is helpful. In the Martindale-Hubbell database, about 23% of all California attorneys designate “litigation” as their practice area, but this figure overrepresents the number of attorneys who actually represent clients at trials or practice civil litigation exclusively. Attorneys in the Martindale-Hubbell directory are permitted to list more than one practice area, and Martindale-Hubbell loosely defines “practice area” as an area to which an attorney “devotes a significant portion of professional time.” Since “practice area” is an expansive concept intended more for marketing purposes than demographic research, it is not surprising that attorneys who list “litigation” as a practice area also list an average of 2.78 practice areas in addition to litigation. (Similarly, 20% of New York attorneys list “litigation” as a practice area, along with an average of 2.42 other practice areas). Hence, the “litigation” attorneys listed in Martindale-Hubbell include attorneys who practice litigation exclusively and attorneys for whom litigation may be a secondary or tertiary practice area. For the purposes of this study, acknowledging the limited data available and the possibility that attorneys for whom litigation is a peripheral practice area do not often try cases to verdict, the estimated percentage of California litigation attorneys is 16%–20% of the total 159,807 active members as of March 13, 2008. Thus, the total estimated number of California litigation attorneys is 25,569–31,961. Because the total number of attorneys included in the primary database is 6,945, the study attorneys represent an estimated 22%–27% of all California civil litigation attorneys.

3.3

Concepts and Definitions

Although civil litigation may be expensive, protracted and seemingly imponderable, only three basic problems occur when trying to resolve a case. These three problems direct the data analysis in the balance of this chapter and hence must be described and understood before proceeding to the data analysis. 17

Hertz Research. (2006). Final report of results, member services survey, The State Bar of California – February 2006. 18 Id. at 17.

3.3 Concepts and Definitions

37

First, a litigant can sabotage settlement negotiations by demanding too much money or offering too little money relative to the actual value of the case. In this study, the actual value of a case is the amount of the judgment or award made by the judge, jury, or arbitrator in that case. All of the cases in this study have a benchmark award or judgment – against which the parties’ settlement positions can be measured – because the parties declined their adversaries’ pre-trial settlement proposals and proceeded to trial or arbitration. In litigation cases, however, neither the plaintiff nor the defendant knows what the actual monetary value of the case is unless their settlement negotiations fail and a judgment or award is entered. Consequently, their settlement positions reflect, at least in theory, an estimate of the case’s value, with some allowances being made for attorney’s fees, court costs, the possibility of an appeal, the difficulty of enforcing a judgment, and other negotiation concessions. The gap between the parties’ settlement positions and the actual value of the case is called “negotiation disparity,” and the parties’ settlement positions are either “underpriced” or “overpriced” relative to the actual case value. Like a seller’s determination of the listing price for a house or a prospective buyer’s offer to purchase the house, underpricing or overpricing of a case may reflect strategic bargaining – or they may reflect overconfidence and unfamiliarity with market values. Asking too much may turn away viable buyers ready to pay fair market value, while asking too little may result in a sale below fair market value. For rational negotiators, the objective is not to sell the house or settle the case at any price but to negotiate a price close to fair market value and avoid the mistake of rejecting a proposal that turns out to be better than the price ultimately paid or received. The second basic problem a litigant can encounter is simply losing the case in pretrial motions or at trial or arbitration. For plaintiffs, this means they recover nothing and fail to obtain any relief sought from the judge, jury or arbitrator – the proverbial goose egg – and may be required to pay the defendant’s court costs or attorneys fees. For defendants, losing at trial or arbitration means being ordered to pay some amount of money to the plaintiff or performing some act previously deemed unacceptable or unwarranted. In this study, the term “win rate” describes the frequency of verdicts and awards in favor of plaintiffs and against defendants and serves as one indicator of the likelihood that a plaintiff or defendant will prevail in a particular type of case. The third source of consternation for litigants is that they can simultaneously win and lose at trial or arbitration. A litigant can prevail at trial but actually sustain a net financial loss because the award at trial is the same as or less than the settlement proposal made by an adversary. Many trial “victories” actually are financial defeats because the nominal winner could have obtained the same result or a better result by accepting an adversary’s settlement proposal. As law professors Gross and Syverud state, “Any plaintiff who was offered as much as the verdict or more, and any defendant who could have settled for as much as the verdict or less, has lost.”19

19

Gross, Samuel, & Syverud, Kent. (1996). Don’t try: Civil jury verdicts in a system geared to settlement. UCLA Law Review, 44, 41–42.

38


These Pyrrhic victories are called “decision errors” in this study. A decision error occurs whenever a party rejects a settlement proposal and the ultimate result at trial or arbitration is the same as or worse than the demand or offer it declined. In absolute terms, the attorney and/or client made a decision error and the client sustained an unequivocal, quantifiable financial loss.20 Decision error is strictly a mathematical calculation and does not mean that an attorney was negligent or improperly advised a client. Many decision errors, moreover, occur because clients do not follow their attorneys’ advice. Although negotiation disparity, underpricing, overpricing, win rates, and decision errors are distinct concepts, they are often confused and at an intuitive level become intermingled. Because an understanding of these distinct concepts is a threshold requirement for analyzing the data in the rest of this chapter, these concepts are further explained and illustrated below.

3.3.1

Negotiation Disparities and Decision Error

Negotiation disparities represent the mathematical differences between litigants’ settlement positions and the ultimate judgment or award in the case. Negotiation disparities occur when a party’s settlement proposal is either (a) less than the ultimate award at trial or arbitration; or (b) greater than the ultimate award at trial or arbitration. In casual conversation, one would say a plaintiff’s settlement demand was “too high” (overpriced) or a defendant’s settlement offer was “too low” (underpriced). Decision errors and negotiation disparities are conceptually and mathematically different. Decision error is determined by comparing the adverse party’s settlement offer with the actual trial result to test whether an opportunity to obtain a better financial result was lost by proceeding to trial, while negotiation disparity is calculated by comparing a party’s own settlement offer with the actual trial result. Decision error thus represents an economic loss caused by declining an adversary’s settlement offer, while negotiation disparity represents the mathematical difference between a party’s own settlement position and the award at trial. Although negotiation disparities often reflect mistaken assessments of trial outcomes, they do not necessarily cause financial harm to the mistaken client. For that reason, they are regarded simply as negotiation disparities and not decision errors. Because a party obviously cannot decline or accept its own settlement proposal, whether overpriced or underpriced, negotiation disparities alone do not cause financial harm. For a financial loss to be sustained a party has to commit a decision error, i.e., declining a settlement proposal that is equal to or better than the result at trial. 20

Parties, of course, may be motivated to litigate for reasons other than obtaining an optimal economic outcome. Gross and Syverud (1996), however, interviewed 735 attorneys in their dataset and reported that “only three attorneys mentioned a desire for vindication as an explanation for why their case went to trial,” and a “non-economic motive” was highly infrequent.” Id. at 57.


3.3.2

39

Underpricing

When a party’s pretrial settlement offer turns out to be less than the amount ultimately awarded at trial, it has underpriced its case. Underpricing by the plaintiff may be a strategic or inadvertent discounting of its gain at trial; for the defendant underpricing may be a crafty negotiating tactic or a careless discounting of its liability at trial. For both the plaintiff and the defendant, though, underpricing simply means that the amount of the demand or offer is less than the ultimate award. If, for example, the plaintiff demanded $8 to settle the case, the defendant offered $6, and the trial verdict was $10, both parties underpriced their cases – the plaintiff’s demand underpriced its gain by $2, and the defendant’s offer underpriced its loss by $4. Although both parties underpriced their demands and offers, only the defendant committed a decision error by declining the plaintiff’s demand to settle for the lesser sum of $8 and later being held liable for a $10 award. The plaintiff did not commit decision error because it was in a financially superior position as a result of rejecting the defendant’s offer. The following case reflects this model of mutual underpricing and a defendant’s decision error. In 2004, a consultant sued her former employer, alleging sexual harassment, sex discrimination, fraud, breach of contract and retaliatory termination. She made a settlement demand of $2,900,000, and the defendant employer offered to settle by paying her $58,000. Both parties rejected the other party’s settlement proposal and took their chances at trial. The jury awarded the plaintiff employee $4,200,000. Although the parties’ settlement proposals displayed wide negotiation disparities and both parties underpriced their settlement positions, only the defendant committed a decision error by declining what turned out to be an economically advantageous settlement demand.21

3.3.3

Overpricing

Conversely, when a party proposes a settlement sum more than the ultimate award, it is considered to have overpriced its case. If the plaintiff submitted a settlement demand of $14, the defendant offered to pay $12, and the award again is $10, both parties overpriced their settlement demands and offers. The plaintiff overpriced its demand by $4 and the defendant overpriced its loss by $2. Although both parties made overpriced demands and offers, only the plaintiff committed a decision error by rejecting an offer of $12 and recovering only $10 at trial. An actual case illustrates this model of overvaluation and plaintiff’s decision error. In 1998, Dennis Berkla sued Corel Corporation for breach of a non-disclosure 21

Adapted from facts in Mayer v. CSC Consulting, Inc., as reported in VerdictSearch California, December 13, 2004. The outcome of subsequent appeals, motions, and settlement negotiations, if any, and the existence and importance of non-economic factors are unknown.

40


agreement regarding an alleged misappropriation of an image file database.22 Both parties were represented by attorneys from highly regarded law firms with extensive intellectual property experience. Berkla initially demanded $1.6 million to settle, and Corel offered $200,000. After a court-facilitated settlement conference, Corel raised its offer to $400,000. Berkla declined Corel’s settlement offer and later countered at $900,000. Corel rejected Berkla’s counter-offer and the case proceeded to trial. Berkla prevailed at trial, convincing a jury that Corel indeed had breached the non-disclosure agreement. The jury’s verdict was in favor of Berkla and against Corel in the amount of $23,502. Although Berkla prevailed at trial, he overpriced his case by $876,498, about 37 times the ultimate award, and made a decision error by forgoing a $400,000 settlement offer. Corel, for its part, overpriced its liability by $376,498, a comparatively conservative figure only 16 times the ultimate award. Despite its generous, overpriced settlement offer, Corel did not commit a decision error. It was far better off declining Berkla’s overpriced demand and benefited from Berkla’s consequent decision to reject Corel’s $400,000 offer. The verdict of $23,502 demonstrates that both parties had vastly overpriced their settlement positions, but only Berkla committed a decision error.

3.3.4

Negotiation Disparities Without Decision Error

As previously explained, decision error is defined and measured differently from negotiation disparity. Consequently, a negotiation disparity can exist in the absence of a decision error. Another example illustrates this point. If the verdict was $10, the plaintiff submitted a pre-trial demand of $12, and the defendant’s pre-trial offer was $8, both parties’ positions displayed negotiation disparities – plaintiff overpricing and defendant underpricing – but neither party committed a decision error. Plaintiff overpriced its demand by $2, defendant underpriced its offer by $2, and both parties achieved a better financial result at trial than could have been obtained by accepting the adversary’s settlement proposal. Thus, despite their respective overpricing and underpricing, neither the plaintiff nor the defendant committed a decision error. This paradigm of mutual negotiation disparities without decision error is shown in the following negligence case.23 The plaintiffs, a father, mother, and their children, were standing near an entrance/exit door at a Costco Wholesale store when two gunmen started shooting AK-47 assault rifles at an armored truck messenger exiting the Costco store from the same door. A stray bullet hit the plaintiff husband in his abdomen and hip; his spleen and part of his pancreas had 22 Berkla v. Corel Corp., 302 F. 3d 909 (9th Cir. 2002). The outcome of subsequent appeals, motions, and settlement negotiations, if any, and the existence and importance of non-economic factors are unknown. 23 Adapted from facts in Chau v. Sectran Security, Inc., as reported in VerdictSearch California, June 7, 2004. The outcome of subsequent appeals, motions, and settlement negotiations, if any, and the existence and importance of non-economic factors are unknown.


41

to be surgically removed. His wife was shot in the abdomen, leg and hip, requiring partial removal of her stomach and intestines. Alleging that the armored truck messenger used an improper entrance/exit and violated his employer’s security procedures, the plaintiffs sued the messenger’s employer for negligence and made a settlement demand of $19 million. The defendant security company offered $3 million – $16 million less than the plaintiffs’ demand. The plaintiffs declined the offer and proceeded to trial; the jury then awarded $3,372,887 to the plaintiffs. In this case, the plaintiffs’ settlement demand was overpriced, while the defendant’s offer was underpriced. Despite these mutual negotiation disparities, neither party committed a decision error, as they were both in financially superior positions after rejecting the other’s settlement proposal.24

3.3.5

Effect of Negotiation Disparity on Decision Error

Under two conditions, it is important to note, one party’s underpricing or overpricing always results in the other party’s decision error. First, whenever a plaintiff underprices its demand a defendant commits a decision error, as the plaintiff’s underpriced demand, by definition, is less than the award the plaintiff obtains at trial. A defendant invariably would have been in a financially superior position had it accepted an underpriced plaintiff demand. Second, a defendant’s overpricing always results in decision error by the plaintiff, as the defendant’s overpriced offer, by definition, is more than the award plaintiff obtains at trial. A plaintiff invariably commits decision error by declining a defendant’s overpriced settlement offer. These concepts of decision error, negotiation disparity, overpricing and underpricing are used throughout this chapter to explain negotiation behavior and trial outcomes in the actual cases described below. All possible sets of decision errors and negotiation disparities are presented in Table 3.1. By identifying decision errors, measuring the range of negotiation disparities, and discovering the case factors associated with those errors and disparities, the study results, described Table 3.1 Examples of decision errors and negotiation disparities Case Plaintiff Defendant Award Decision Plaintiff Negotiation Demand Offer Error Disparity 1 $8 $6 $10 Defendant Underpriced 2 $10 $8 $10 Defendant None 3 $10 $10 $10 Both None 4 $12 $8 $10 Neither Overpriced 5 $12 $10 $10 Plaintiff Overpriced 6 $14 $12 $10 Plaintiff Ovepriced

24

Defendant Negotiation Disparity Underpriced Underpriced None Underpriced None Overpriced

The parties’ legal costs, the results of appeals and post-judgment discounts of awards generally are unknown and hence usually are not included in the calculation of decision error. See Appendix.

42


below, may serve as statistical guideposts to assist litigants and attorneys making tough choices under uncertainty.

3.4

Overall California Results

Two cardinal findings emerge from the primary dataset: plaintiffs make decision errors more often than defendants, but the cost of defendants’ decision errors is dramatically higher than the losses plaintiffs sustain. As shown in Table 3.2, the decision error rate for plaintiffs is 60%, compared to defendants’ decision error rate of 25%. In other words, plaintiffs would have achieved better financial results if they had flipped a coin to decide whether to settle or try a case, and defendants made a decision error in one out of every four cases. In only 15% of the cases did both parties obtain a superior economic result by rejecting each other’s settlement proposal and proceeding to trial; out of every 100 trials in the dataset, only 15 trials resulted in a nominal “win-win” award after the parties walked away from the negotiating table. This distribution of decision error persists in cases where the amount at issue is relatively high and the parties and their attorneys may be expected to exhibit a superior level of experience and sophistication. When the dataset is limited to cases in which a plaintiff’s demand is between $1,000,000 and $50,000,000, the incidence of plaintiff error is 58%, compared with 28% for defendants. The “no error” cases constitute 14% of all cases in the high-end dataset.

3.4.1

Costs of Decision Error

For the plaintiffs, the average cost of decision error – the difference between what they received at trial and the amount they could have received through settlement – is $73,400. The defendants’ average cost of error, in comparison, is $1,403,654, about 19 times the loss sustained by plaintiffs. When the dataset is narrowed to relatively large cases in which plaintiff’s demand is between $1,000,000 and $50,000,000, this pattern persists; the average cost of plaintiff error in the high-end cases is $327,158, compared with defendants’ average cost of error of $5,325,785. Table 3.2 Overall results for primary dataset (2,754 cases)

Total Attorneys Decision Error Rate Mean Decision Error Cost Mean Decision Error Cost/Mean Award Mean Demands/Offers as % of Mean Award Overpriced Demands/Offers Underpriced Demands/Offers

Plaintiffs 3,492 60% $73,400 11.93%

Defendants 3,453 25% $1,403,654 228.07%

121% 75% 25%

21% 42% 40%

3.4 Overall California Results

43

The total financial loss sustained by all plaintiffs making a decision error is $120,890,536, compared to an aggregate loss of $981,154,097, nearly one billion dollars, for all defendants committing a decision error. Because the attorneys fees and court costs are unknown in most of these cases, those costs would have to be added to these aggregate financial losses to calculate the actual total loss.25 If the total cost of attorneys fees and costs were ascertainable, the percentage of “no decision error” cases would decline, and both the incidence and magnitude of decision error would increase. (In the extreme circumstance where the parties incur attorneys fees and costs for both a trial and an appeal, one prominent mediator says he has never seen a case where either party realized a net gain after factoring in the attorneys fees). For every party that commits a decision error, an adverse party necessarily benefits from that error. When an erring plaintiff foregoes a $10,000 settlement and recovers only $6,000 at trial, a defendant “saves” the additional $4,000 he was willing to pay to settle the case; conversely, when an erring defendant declines a plaintiff’s settlement demand of $13,000 and is hit with a $20,000 judgment at trial, the plaintiff “gains” an additional $7,000. In a sense, some attorneys argue, there is no net financial loss from decision error in litigation cases because the money that might have gone into one party’s pocket simply ends up in the other party’s pocket. From some advocates’ perspective, a foregone financial benefit is not a decision error but simply an episodic, unintended distribution of assets among morally neutral parties with competing claims. Because the permanent players in the litigation system earn the same fees regardless of the clients’ results and attorneys cannot be expected to foresee trial outcomes that are inherently unpredictable, some attorneys and academicians assert, the whole idea of decision error is wrong and irrelevant. Steve Brill, the founder of Court TV and the American Lawyer magazine, noted this attitude while covering trials earlier in his career; he calls it the “I’m getting $500 an hour no matter what these goober jurors say and we’ll win on appeal anyway smugness.”26 The “who cares what happens to the gamblers as long as the card shufflers get paid” attitude overlooks three basic problems. First, for most litigants the objective is to make a financially sound decision, assisted by the advice of counsel. Although parties also may attempt to vindicate a principle in litigation, they invariably desire to vindicate the principle in addition to making a financially sound decision. When that party commits a decision error, its litigation objectives are defeated; and it takes no more solace in the fact its adversary has profited from its decision error than it would when informed that its stockbroker is enjoying the upscale house purchased with the commissions obtained from churning its account. Second, the 25

Other tangible factors that would affect the incidence and cost of decision error are costs of postjudgment collection, bankruptcy of the judgment debtor, inability to collect the entire amount of the judgment from an evasive or impecunious defendant, and the time value of money. Intangible factors that nevertheless hit the bottom line include the diversion of an organization’s key resources and damage to reputation and business relationships. 26 Brill, Steven. (2008, March 30). Uncivil action. The New York Times.

44


argument ignores the attorneys fees, court costs, public resources, and intangible assets like time and attention diverted from the litigant’s core business and personal relationships and unnecessarily expended by all parties for months, sometimes years, after a settlement would have yielded a superior economic result. The party “saving” $100,000 when an adversary declines a $200,000 pre-trial offer and recovers only $100,000 at trial, may well have spent the $100,000 “savings” on its attorneys fees between the dates of the offer and the rendition of a verdict. Third, the argument perpetuates the “luck of the draw” philosophy of litigation, absolving attorneys of any responsibility for realistically determining a trial’s likely outcome and learning how to use the decision making tools employed by other professionals and managers with equivalent private and public responsibilities.

3.4.2

Negotiation Disparities

The settlement negotiation positions that preceded the decision errors in the primary dataset are archetypal, evidencing strong patterns of risk aversion by plaintiffs and risk taking by defendants. The average plaintiff’s demand is 121% of the ultimate award, while the mean defendant’s offer is 21% of the ultimate award. Plaintiffs’ mean demand rests on a narrow margin above the actual adjudicated outcomes, arguably reflecting strategic bargaining and a relatively modest degree of overconfidence. Defendants’ mean offer, in contrast, falls far below the mean award and appears to be more consistent with behavioral economics theories of risk taking in the losses frame than strategic negotiation behavior. Offering 21 cents on the dollar seems, on its face, to be more aggressive than deliberative. Comparing the mean cost of error with the mean award also reveals the degree of defendants’ risk taking. For defendants, the mean cost of error ($1,403,654) is more than twice the amount of the mean award ($615,436); for plaintiffs, meanwhile, the mean cost of error is only 11.9% of the mean award amount. The parties’ patterns of underpricing and overpricing settlement demands and offers also are consistent with behavioral economics theories. Plaintiffs overprice their demands (relative to the ultimate award) in 75% of the cases, and the mean overpriced demand is $649,069 above the actual award. Plaintiffs underprice their demands in the other 25% of the cases, but the magnitude of their underpricing is more severe than the magnitude of their overpricing. When plaintiffs underprice, i.e., discount their own cases by demanding an amount less than the ultimate award, the average underpriced demand is $1,447,130 below the actual award. The magnitude of plaintiffs’ underpricing thus is more than double the magnitude of their overpricing. For defendants, the pattern of overpricing and underpricing is reversed. Although defendants’ offers are roughly split between underpriced and overpriced offers (40% and 42%, respectively), the mean amount of underpricing is $1,326,684 below the mean award, while the mean amount of overpricing is a relatively meager $103,858 above the mean award. Thus, the apparent balance between defendants’

3.5 New York Results

45

underpricing and overpricing of offers masks the enormity of defendants’ discounting and the relatively small amount of defendants’ overpricing. Ironically, when plaintiffs underprice their own demands, the magnitude of their underpricing roughly matches the discounting by defendants. Underpricing plaintiffs discount their demands $1,447,130 below the mean award, while their underpricing counterparts on the defense side discount their offers $1,326,684 below the mean award. The risk averse, underpricing plaintiffs give their own cases the same haircut offered by their adversaries.

3.5

New York Results

The New York dataset replicates the overall results observed in the primary California dataset. As shown in Table 3.3, the plaintiff win rate in New York is 48%, compared to 49% in the California dataset. Plaintiffs’ decision error rate in New York is slightly lower than in California (56% vs. 60%), but the New York defendants’ error rate is slightly higher than the California defendants’ error rate (29% vs. 25%). Due to these offsetting decreases and increases in error rates, the incidence of “no decision error” is 15%, exactly the same percentage in both the New York and California databases. The disparities between the cost of plaintiff decision error and the cost of defendant decision error are similar in New York and California. The New York database shows that plaintiffs’ mean cost of error is $52,183, compared to defendants’ mean cost of error of $920,874, roughly 18 times the cost of plaintiffs’ error. In California, defendants’ mean cost of error is 19 times the mean amount of plaintiffs’ decision error. The New York parties’ settlement negotiations reflect the same patterns of plaintiff risk aversion and defendant risk taking as the California dataset. Plaintiffs’ mean demand, as a percent of the mean award, is 118% (New York) and 121%

Table 3.3 Comparison of New York and California Data Total Cases Mean Award Plaintiff Win Rate Plaintiff Decision Error Rate (%) Defendant Decision Error Rate (%) No Decision Error Rate (%) Plaintifff Mean Cost of Error Defendant Mean Cost of Error Plaintiff Mean Cost of Error As % of Mean Award Defendant Mean Cost of Error as % of Mean Award Plaintiff Demand as % of Mean Award Defendant Offer as % of Mean Award

New York 525 $596,282 48% 56% 29% 15% $52,183 $920,874 8.75% 154.44% 118% 23%

California 2,754 $615,436 49% 60% 25% 15% $73,400 $1,403,654 11.93% 228.07% 121% 21%

46


(California), while defendants’ mean offer, as a percentage of the mean award, is 23% (New York) and 21% (California). These patterns of relatively modest demands by plaintiffs and de minimis offers by defendants result in major financial losses for the defendants in both states. Plaintiffs’ mean cost of error, as a percentage of the mean award, is 8.75% in New York and 11.93% in California. In contrast, defendants’ mean cost of error, relative to the mean award is 154.44% in New York and 228.07% in California. Overall, plaintiffs in both states are risk averse in their settlement negotiations and sustain relatively minor financial hits when they make a decision error, while defendants in both states display aggressive, risk-taking positions in settlement negotiations and sustain major losses as a result of decision errors in roughly onequarter of their cases. Defendants’ settlement positions appear to be more autonomic than economic, and the costs of their decision errors, when compared to mean awards, more closely reflect framing biases (risk aversion in the “gains” mode and risk taking in the “losses” mode) than strategic bargaining in the shadow of the law.

3.6

40-Year Historical Study

To provide an historical context for the California and New York results, a third dataset was compiled, representing cases negotiated and adjudicated during the 40–year period between 1964 and 2004. The results of this 40-year study are remarkably similar to the results derived from the primary California dataset and the New York dataset in all categories – win rates, negotiation disparities, and decision error rates. The plaintiff win rate in the historical dataset is 50%, compared with 49% in the primary California dataset and 48% in the New York dataset. Plaintiffs’ settlement demands exceed the actual award in 78% of the historical dataset cases, compared with overpricing rates of 75% and 71% for the primary California and New York datasets, respectively. Defendants, on the other hand, made underpriced settlement offers in 39% of the historical dataset cases, contrasted with defendant underpricing rates of 40% and 39% in the primary California dataset and the New York dataset, respectively.

3.6.1

Historical Decision Error

The decision error rates for plaintiffs during that 40-year period ranged from a low of 49.7% in 1969 to a high of 67.9% in 1994, with an overall decision error rate of 61%. The historical plaintiff decision error of 61%, as shown in Table 3.4 is similar to the plaintiff decision error rate in the primary California dataset (60%) and the New York dataset (56%). For defendants, the historical dataset shows decision error rates ranging from a low of 19.0% in 1964 to a high of 26.2% in 1984; and the overall defendant error rate for the entire 40-year period is 22%. This overall

3.6 40-Year Historical Study

47

Table 3.4 Decision error and cost of error – historical samples Year Error Type % of Cases 1964 1964 1964 1969 1969 1969 1974 1974 1974 1979 1979 1979 1984 1984 1984 1989 1989 1989 1994 1994 1994 1999 1999 1999 2004 2004 2004

No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error No Error Plaintiff Error Defendant Error

27.20 53.80 19.00 25.20 49.70 25.20 14.70 65.40 19.90 19.30 57.90 22.80 11.30 62.40 26.20 15.00 63.00 22.00 10.20 67.90 21.90 17.50 60.10 22.40 14.00 65.70 20.20

Mean Cost of Error ($1,000’s) NA 1.2 5.9 NA 1.8 27.3 NA 5.7 42.6 NA 6.6 67.7 NA 18.4 628.4 NA 38.4 546.5 NA 22.4 1,120.60 NA 45.7 2,259.80 NA 40.8 649.1

decision error rate of 22% is lower than the error rate in the primary California dataset (25%) and the New York dataset (29%), showing an increased incidence of defendant decision error over time. Interestingly, the incidence of plaintiff decision error over the 40-year period is never less than the incidence of defendant decision error, varying from two to three times the incidence of defendant decision error. Despite some volatility over time, the frequency of decision error is greater at the end of the 40-year period than at the beginning. The number of cases in which no decision error occurs drops from 27.2% and 25.2% in 1964 and 1969, respectively, to 17.5% and 14.0% for the years 1999 and 2004, respectively, revealing a rise in aggregate plaintiff and defendant decision error. If lawyers and their clients were physicians and patients, one would conclude that diagnostic error fluctuated but ultimately increased during the last 40 years.

3.6.2

Historical Cost of Decision Error

The financial consequences of decision error in the historical dataset are shown in Table 3.5. As that table indicates, the cost of decision error for both plaintiffs and

48


Table 3.5 Mean cost of error in constant 1964 dollars

Period

1964, 1969, 1974 1979, 1984, 1989 1994, 1999, 2004

Plaintiff Mean Cost of Error ($1,000’s) 2.6 5.9 7

Defendant Mean Cost of Error ($1,000’s) 20.5 122.5 300.6

defendants is substantially greater at the end of the 40-year historical period. When the cost of error for this period is adjusted for inflation (the nominal values in Table 3.4 are converted to real values (in 1964 dollars)), one observes a dramatic rise in the magnitude of the parties’ decision errors. From the earliest period (1964–1974) to the latest period (1994–2004), plaintiffs experienced nearly a threefold real (inflation-adjusted) increase in mean cost of error. During that same period, defendants experienced in excess of a 14-fold real increase in mean cost of error. The mean cost of plaintiff decision error rose from $2,600 to $7,000 (in constant 1964 dollars), while the cost of defendant decision error soared from $20,500 to $300,600 (in constant 1964 dollars). In the medical context, similar outcome data would be interpreted as showing an increase in diagnostic error accompanied by startling increases in morbidity and mortality rates of 3–14 times previous levels, depending on the type of patient undergoing elective surgery. In hospitals, however, a mere twofold increase in mortality rates could lead to suspending a program, closing a department or terminating staff privileges. This surge in the cost of error defies legislative and judicial efforts to expand pretrial discovery, broaden the scope of information exchanged by litigants, and compel pre-trial disclosure of documents and contentions regarding damages. Civil discovery in California changed significantly during this period due to liberal interpretations of the Civil Discovery Act of 1957 and the enactment of the Civil Discovery Act of 1986. These changes were intended to encourage settlements, reveal the strengths and weaknesses of an adversary’s case, eliminate surprise, and generally end the “trial by ambush” era.27 Although the California cases that settle may well have achieved those objectives, the historical sample indicates that, for non-settling parties, the surprises are neither less frequent nor less costly.

3.7

Attorney-Mediator Results

Although the primary dataset includes 6,945 attorneys – roughly a quarter of all California litigation attorneys – and the decision error rates in that dataset, the New York dataset, and the historical dataset are remarkably consistent, one may question 27

See Fairmont Insurance Co. v. Superior Court, 22 Cal. 4th 253 n. 2, 92 Cal. Rptr. 2d 70 (2000). Greyhound Corp. v. Superior Court, 56 Cal. 2d 355, 15 Cal. Rptr. 90 (1961).

3.7 Attorney-Mediator Results

49

whether the attorneys in the dataset are trial junkies, determined to take cases to trial regardless of the odds. Skeptics may ask whether the attorneys in these datasets had singular risk-taking propensities that impeded a negotiated settlement and ultimately resulted in significant decision errors. The issues of selection bias or dispositional bias, unfortunately, can never be resolved conclusively because researchers cannot compare the decision error rates in the study datasets with decision error rates in negotiated settlements. Settlements effectively eliminate any financial markers by which one can compare the negotiated settlement sum with the amount that would have been obtained at trial. Like responsible campers, settling parties leave no traces when they move on. But researchers can very roughly probe for selection bias by learning whether the arguably over-confident attorneys in the datasets exhibit higher decision error rates than attorneys who possess strong conflict resolution skills and are trained to be objective, settlement-seeking professionals. To identify attorneys with strong dispute resolution skills and substantial experience in negotiating pre-trial settlements, lists of California attorneys serving on Superior Court mediator panels, affiliated with private dispute resolution companies, or designated as members of the Southern California Mediation Association were reviewed. From these sources a total of 939 “attorney-mediators” were identified. Each attorney-mediator’s name then was entered in the VerdictSearch electronic database, limiting the search to California cases reported between 1985 and 2006. The search was further limited to cases in which the attorney-mediator had represented a plaintiff or defendant in a case tried through verdict or arbitration award. (Not all of the 939 mediators were necessarily litigation attorneys at any time during that period, since the courts’ lists include some non-attorneys, former judges, and non-litigation attorneys). The search initially yielded 672 cases reported during the 1985–2006 period, of which 369 met the case selection criteria used for the primary study dataset. Of the remaining 303 cases, 150 were settled and 153 did not meet the selection criteria for other reasons.28

3.7.1

Attorney-Mediator Decision Error

As indicated in Table 3.6, the presence of an attorney-mediator generally is associated with a reduced decision error rate.29 Table 3.6, Panel 3.6a summarizes 28

The attorney-mediator dataset spans a 21-year period (1985–2006), while the primary study dataset covers a 58-month period (November 2002-August 2007). Whether a party is represented by an attorney who also serves as a mediator is not a fact separately reported in VerdictSearch California and hence was not a variable coded in the primary study dataset. 29 Rachlinski suggested that framing effects might be mitigated by the intervention of attorneys familiar with framing biases: “The framing theory suggests another positive influence attorneys may have in reducing the costs of litigation. An attorney may have some power to reframe a settlement offer, sparing the client the most costly aspects of framing . . .. Thus, the framing model of litigation poses a powerful role for the attorney. The attorney can control the client’s frame,

50


Table 3.6 Decision error and cost of error (attorney-mediator sample) Error Type % of Cases Mean Cost of Error ($1,000’s) Panel 3.6a - All Cases No Error 21.1 NA Plaintiff Error 52.6 48.4 Defendant Error 26.3 900 Panel 3.6b - Attorney-Mediator Represents Plaintiff No Error 19.5 Plaintiff Error 48.5 Defendant Error 32.0

NA 68.4 908.6

Panel 3.6c - Attorney-Mediator Represents Defendant No Error 22.5 Plaintiff Error 56.0 Defendant Error 21.5

NA 33.8 889.2

the experience for 369 cases in which one of the parties was represented by an attorney-mediator. Total decision error in this sample is less relative to the primary dataset presented in Table 3.2; “no error” in attorney-mediator cases is 21% compared to 15% in the primary dataset. Ironically, the percentage of “no error” cases in the attorney-mediator dataset is similar to the percentage of “no error” cases for all litigation attorneys in the 1960s, a period when litigation attorneys’ judgment may have been sharpened by spending more time in actual trials than laboring over pre-trial motions.30 In cases where attorney-mediators represented plaintiffs, plaintiffs’ win rate is higher (62% vs. 49%) and plaintiffs’ decision error rate is lower (48% vs. 60%) than the win rates and decision error rates in the primary dataset. The overall incidence of decision error is lower for the plaintiff attorney-mediator set than the primary dataset; “no error” cases comprise 19.5% of the plaintiff attorney-mediator set relative to 15% of the cases in the primary dataset. In cases where defendants were represented by an attorney-mediator, summarized in Panel 3.6c, the incidence of defendants’ decision error is similarly reduced. The decision error rate for the defendant attorney-mediator cases is 21.5%, compared to 25% in the primary dataset. The percentage of cases with “no decision error” is again higher in the attorney-mediator set (22.5%) than the primary dataset (15%). Specific case types were examined to assess the incidence of decision error in the attorney-mediator cases. Because the sample of attorney-mediator cases, when classified by case type and whether the mediator represented a plaintiff or a

thereby influencing settlement decisions in either direction.” Rachlinski supra note 11 at 171–2. See Korobkin, Russell, and Guthrie, Chris. (1997). Psychology, economics and settlement: A new look at the role of the lawyer. Texas Law Review, 76(1), 77. 30 See Galanter, Marc. (2004). The vanishing trial: An examination of trials and related matters in federal and state courts. Journal of Empirical Legal Studies, 1(3), 459–570.

3.7 Attorney-Mediator Results

51

defendant, was small compared to the primary dataset, this study focused on personal injury cases, the most common type of cases in the primary sample. Consistent with the overall findings of reduced decision error in attorney-mediator cases, the study found that personal injury cases in which the parties were represented by attorney-mediators showed a lower decision error rate than those in the primary sample. Plaintiffs’ decision error rate in personal injury cases was 45.2% in the attorney-mediator sample and 53% in the primary sample. Defendants’ decision error rate in personal injury cases showed a similar pattern – 16.8% in the attorneymediator sample and 26% in the primary sample.

3.7.2

Attorney-Mediator Negotiation Disparities and Settlement Rates

Although representation by an attorney-mediator is associated with lower decision error rates, their clients’ negotiation patterns are not markedly different from the patterns evident in the primary dataset. In the plaintiff attorney-mediator sample, plaintiffs’ mean demand is 122% of the mean award; in the primary dataset, the mean demand is 121% of the mean award. (Plaintiffs represented by attorneymediators, however, are less likely to submit overpriced demands; 68% of the demands in the plaintiff attorney-mediator set are overpriced, compared to 75% of the demands in the primary dataset). The negotiation positions taken by defendants in the attorney-mediator set also reflect the same negotiating styles displayed in the primary dataset. The average offer made by defendants represented by attorney mediators is 22% of the average award, compared to 21% in the primary dataset. In addition to analyzing decision error rates and negotiation strategies, the study compared the settlement rates of attorney-mediators with those of other attorneys reporting cases during the primary dataset period. (VerdictSearch California reports settlements as well as trial and arbitration awards). The purpose of this comparison was to determine indirectly whether attorney-mediators were more likely to settle than litigate their cases and to test whether attorneys reporting cases during the primary dataset period exhibited an anti-settlement bias. This analysis revealed that a settlement was reported in 22% of all cases in which an attorney-mediator represented a party; for the primary dataset period, a settlement was reported in 29% of all cases. Thus, the attorney-mediators were not more likely to settle their cases than the ordinary attorneys reporting cases during the primary dataset period.

3.7.3

Tentative Conclusions About Attorney-Mediators

What tentative conclusions can be drawn about the decision-making qualities of attorney-mediators? First, because decision making about case settlement is a joint

52


attorney-client effort, one cannot discern whether the reduction in decision error is caused by the attorney-mediators themselves or their clients. It is entirely possible that those clients who make financially better decisions choose to be represented by attorneys who have undertaken mediation training and endeavor to reach negotiated resolutions; representation by an attorney-mediator may be incidental to the client’s superior decision-making skills, not a cause of the lower decision error rates. Second, despite completing a minimum of 30 hours of mediation training and, in some instances, functioning as a neutral for many years, the attorney-mediators did not achieve a dramatic reduction in decision error rates; their superior performance is noteworthy but essentially incremental. For plaintiffs, representation by an attorney-mediator reduced decision error rates from a “worse than chance” incidence to a “same as chance” incidence. Third, the decision error rates evidenced by the ordinary attorneys in the primary, historical and New York datasets do not appear to be attributable to distinctly litigious dispositions, anti-settlement bias or aberrational risk-seeking behavior. Even attorney-mediators, committed to achieving pretrial settlements, find it difficult to settle cases and effectuate major reductions in decision error.

3.8

Predictor Variables

What case factors are predictive of decision errors or correlated with high win rates and extreme negotiation disparities? To answer this question, cases in the primary dataset were analyzed and coded to identify every publicly reported, verifiable factor that could affect decision making. These variables generally can be classified into two groups: “Actor” variables and “Context” variables. The Actor variables can be thought of as the personal characteristics of the attorneys and their clients, while the Context variables depict the external conditions that may affect or reflect the actors’ decisions. An attorney’s gender, for instance, is an Actor variable, while the type of case handled by the attorney is a Context variable. Some of the Actor variables describing attorneys include the number of years an attorney has practiced law after admission to the bar, the academic and diversity ranking of the law school from which the attorney graduated, the attorney’s gender, and the size of the law firm in which the attorney practices. Actor variables describing the parties themselves include the type of party, e.g., corporation, unincorporated business, insurance company, governmental entity, or individual, and, if a party is an individual, the individual’s gender. The Context variables include the type of case being litigated, the forum in which the case was adjudicated (jury trial, judge trial, or arbitration hearing), the types and amounts of settlement demands and offers exchanged between the parties, the degree of disparity between the plaintiff’s demand and the defendant’s offer, and the type of damages being sought (out-of-pocket damages already sustained, future damages expected to be sustained, and punitive damages). Context

3.8 Predictor Variables

53

variables also include whether a defendant has insurance coverage and whether the alleged wrong was an act of commission or an omission. (A full description of all variables and coding methods is contained in the Appendix). After all of these variables are ascertained from a variety of sources and then coded and entered in the dataset, different models of multivariate analysis are utilized to assess correlations between these variables and the three measures critical to litigants: win rates, decision error rates, and negotiation disparities. Because this is not a technical book and some statistical models were described previously in the Journal of Empirical Legal Studies article,31 it is sufficient to explain that a multivariate analysis attempts to determine the independent effect of a variable. If, for instance, graduation from an “elite” law school is correlated with relatively low decision error rates, is the low decision rate really affected by the elite law school variable, or is another variable actually affecting this correlation? If most litigation lawyers who graduate from elite law schools also are employed by large law firms, are predominantly male, handle only the types of cases correlated with low decision error, and have practiced for 20 years, which variable is significant – school, firm size, gender, case type or experience? That kind of question is answered by multivariate analysis, which can include regression analysis generally and logistic regression specifically. In light of the traditional antipathy between attorneys and statistics, probably too much already has been said on this subject. Let’s turn, then, to results.

3.8.1

Context Variables Trump Actor Variables

Despite the popular emphasis on “star” trial attorneys, the multivariate analysis indicates that Context variables have a stronger effect on decision errors, negotiation disparities and case outcomes than Actor variables. In real life, Perry Mason’s success would depend more on the type of case he accepted than his experience or law school. (Not surprisingly, the author who crafted the Perry Mason stories, Erle Stanley Gardner, attended law school for less than a month; he was ejected from Valparaiso University for cuffing a law professor and passed the bar about four years later after studying law under various practicing attorneys). The British bar seems to have recognized the importance of Context variables long before empirical research confirmed that star power is more often a product of background lighting than personal luminosity.32 Taking a somewhat sardonic view, the British bar appraises an individual attorney’s contribution to the case outcome with

31

Kiser, Randall, Asher, Martin, and McShane, Blakeley. (2008). Let’s not make a deal: An empirical study of decision making in unsuccessful settlement negotiations. Journal of Empirical Legal Studies, 5(3), 551–591. 32 Groysberg, Boris, Nanda, Ashish & Nohria, Nitin. (2004, May). The risky business of hiring stars. Harvard Business Review, p. 93.

54


mathematical precision: “of every hundred cases, ninety win themselves, three are won by advocacy, and seven are lost by advocacy.”33 The eminent American trial lawyer, Edward Bennett Williams – the “magic mouthpiece” with the “affidavit face”34 – expressed a somewhat similar opinion about whether lawyers make cases successful or cases make lawyers successful. After a lifetime of representing clients like Frank Sinatra, Jimmy Hoffa, and The Washington Post, Williams opined that an individual attorney made a difference in only 20% of the cases. Given 100 cases and assuming that 50 should be won and 50 should be lost, he mused, the country’s best attorney would win 60% of the cases. “Turn the same case over to the most incompetent trial man,” Williams continued, “and he will win forty and lose sixty.”35 When he reviewed every case he had lost and noted next to each case all possible reasons for the loss – bribed jurors, biased judges, unethical opposing counsel, racial prejudice – he could find only one common factor among the cases he lost: bad facts. Like the British bar, Williams concluded that facts trumped personalities.

3.8.2

The Five Major Context Variables

The five Context variables that appear to exert a relatively strong effect on decision errors, win rates, or negotiation disparities in the primary dataset are (1) the type of case; (2) the service of statutory settlement demands and offers which raise the monetary risks of not settling before trial; (3) the forum (whether the case is decided by a judge, jury, or arbitrator); (4) the type of damages claimed by the plaintiff; and (5) the degree of disparity between the plaintiff’s demand and the defendant’s offer. Playing a lesser, but nonetheless notable role are other Context variables like the existence of insurance coverage for the defendant and whether the alleged wrongful act is an act of omission (usually negligent behavior), commission (intentional or willful conduct), or both. The Actor variables generally exert a secondary effect on win rates, decision errors and negotiation disparities. On a spectrum of Actor variables, characteristics like the type of defendant (corporation, individual, or governmental entity), attorney experience and attorney gender usually would be ranked higher than the academic ranking of the law school from which a defense attorney graduated, the size of the defendant’s law firm, and the type of plaintiff.

33

Bradford, Glenn. (2002, July-August). Losing. Journal of the Missouri Bar, 58(4), quoting Mayer, Martin. (1968). The lawyers (p. 44). New York: Dell Publishing Company, Inc. and Shragger, David, & Frost, Elizabeth. (1986). The quotable lawyer, Chester, Connecticut: New England Publishing Associates. 34 Thomas, Evan. (1991). The man to see (p. 17). New York: Touchstone. 35 Bradford supra note 33 at 5.


55

To some extent, establishing which variables are less important is as valuable as ascertaining which variables appear to be most important. Some clients, for instance, may believe that an attorney’s graduation from an elite law school or her employment in an elite law firm is the critical element in trial outcomes, but this conclusion is not substantiated by the limited information in the primary dataset. Among the many variables found to be of lesser consequence in predicting win rates, decision errors and negotiation disparities are the type of plaintiff, the size of the defendant’s law firm, the academic ranking of the law school from which the defendant’s attorney graduated, the diversity ranking of the law school from which the plaintiff’s attorney graduated, and whether the parties participated in non-binding arbitration or mediation before trial. Although some attorneys may be more effective forecasters and counselors than other attorneys, the key to their effectiveness may be found beyond the traditional icons of law school rank and law firm size. The most important variables and their correlation with win rates, decision error rates, and negotiation disparities are discussed below. Although every variable may appear to be important, the discussion below is limited to the variables that have relatively strong predictive value in a multivariate analysis. Stated differently, an individual variable may show intriguing and statistically significant differences if studied separately from the other variables in the dataset, but it is excluded from the discussion below if it has relatively little predictive value in the context of all other variables. Client representation by a defense attorney who graduated from a “Top 20” law school, for example, might be associated with a higher decision error rate than representation by a defense attorney who graduated from a non-Top 20 law school, but the defense counsel law school variable is omitted from further discussion because it shows relatively little predictive value in a multivariate analysis.

3.8.2.1

Case Type

Under the Priest and Klein “fifty percent implication,” one expects win rates and decision error rates to be balanced between the parties and unaffected by the case type. Plaintiffs theoretically would win 50% of their cases, regardless of case type, and with respect to decision error, plaintiffs and defendants would be “equally successful at predicting the outcomes of the cases.”36 The assumptions and predictive capacity of the Priest and Klein model, however, are challenged by the study data showing that both win rates and error rates vary widely with different types of cases, as shown in Table 3.7. In general, high plaintiff error rates are associated with cases in which contingency fee arrangements are common, e.g., medical malpractice (80% error rate) and products liability (70% error rate), while low error rates are associated with cases in which contingency fee arrangements are less common, e.g., contracts (31% error rate) and eminent domain (33% error rate). The higher error rates attendant to 36

Gross and Syverud supra note 7 at 325.

56


Table 3.7 Win rates and decision error rates by case type Case Type Plaintiff Win Plaintiff Decision Rate (%) Error Rate (%) Contract 73 31 Contract/Tort 58 49 Eminent Domain 100 33 Employment 52 51 Fraud 52 48 Intentional Tort 31 72 Medical Malpractice 21 80 Negligence (Non-PI) 42 63 Personal Injury 52 53 Premises Liability 40 65 Products Liability 32 70

Defendant Decision Error Rate(%) 60 36 47 34 44 17 17 21 26 19 23

No Decision Error (%) 9 15 20 15 8 11 3 16 21 16 7

plaintiff contingency fee cases may reflect optimistic overconfidence, according to an earlier study. In that study, lawyers retained on a contingency basis showed the same level of confidence about case outcomes as other lawyers, although the contingency basis attorneys won only 42% of their cases compared with an overall 56% win rate. In general, that study found that lawyers’ predictions regarding whether they would win their case “showed no predictive validity” and were “hardly above chance.” The lawyers, the study concluded, exhibited a marked “overextremity bias (underprediction of success for low probabilities and overprediction of success for high probabilities).”37 For the clients themselves, the contingency fee agreement may present an economic incentive to forego settlement and litigate cases to a final verdict, as the client ultimately decides whether to settle the case and may incur little or no additional cost for trial.38 On the defense side, high decision error rates are noted in cases where insurance coverage is generally unavailable, e.g., contract cases (60%) and fraud cases (44%), while low error rates are associated with cases in which insurers are more likely to represent defendants, e.g., medical malpractice (17% error rate) and premises liability (19% error rate).

37

Goodman-Delahunty, J., Granhag, P.A. & Loftus, E.F. (1998). How well can lawyers predict their chances of success? Unpublished manuscript. University of Washington. Cited in Koehler, Derek J., Brenner, Lyle, & Griffin, Dale. (2002). The calibration of expert judgment: Heuristics and biases beyond the laboratory. In Gilovich, Thomas, Griffin, Dale, & Kahneman, Daniel (Eds.). (2002). Heuristics and biases: The psychology of intuitive judgment (pp. 705, 706). Cambridge: The Press Syndicate of the University of Cambridge. For other results regarding attorneys’ predictive capabilities, see Loftus, Elizabeth F., and Wagenaar, Willem A. (Summer 1988). Lawyers’ predictions of success. Jurimetrics, 28, 437. 38 Contingency fee agreements typically provide for an increase in the percentage of the net recovery paid to the attorney, as the case gets closer to a trial date. A typical agreement may increase the attorney’s payment from 33% to 40% of the recovery within 30–60 days before trial. If the parties’ settlement negotiations occur during the specified pre-trial period, the client’s share of any net recovery remains the same whether the case is tried or settled, although the client likely will incur some additional court costs (e.g., jury fees) for the trial.


57

In general, an inverse relationship exists between plaintiff decision error rates and plaintiff win rates. Plaintiff decision error is lowest in cases with high plaintiff win rates and highest in cases with low win rates. Contract cases, for instance, have a 31% plaintiff decision error rate and a 73% win rate, while medical malpractice cases have an 80% plaintiff decision error rate and a 21% win rate. For defendants, the pattern generally is reversed; high defendant decision error rates are evident in high plaintiff win rate cases. Defendants commit decision errors in 60% of the contract cases, where the plaintiff win rate is 73%, but they make decision errors in only 19% of the premises liability cases, where the plaintiff win rate is only 40%. For defendants, calibration deteriorates as the risk of monetary loss increases. The decision error rates, when classified by identical case types, appear to be roughly consistent with Gross and Syverud’s data for 1985–1986 and 1990–1991 cases. In Gross and Syverud’s study, for instance, plaintiffs in medical malpractice cases were “clear losers” in 71% (1985–1986) and 78% (1990–1991) of the cases, compared with a 80% plaintiff decision error rate for medical malpractice actions in the primary dataset. Defendants’ decision error rate for medical malpractice cases in Gross and Syverud’s study was 17% (1985–1986) and 16% (1990–1991), compared with 17% in the primary dataset. The results in products liability cases are more disparate, but reflect similar qualitative differences between plaintiff and defendant decision error. Gross and Syverud’s data show plaintiffs in products liability cases either recovered nothing or less than the defendants’ offer in 64% (1985–1986) and 61% (1990–1991) of the cases, compared to plaintiffs’ decision error rate of 70% in the primary dataset. Defendants, on the other hand, committed decision error in 25% (1985–1986) and 32% (1990–1991) of the Gross and Syverud product liability cases, contrasted with 23% in this study. The type of case appears to affect not only decision error rates but negotiation disparities as well. The parties’ negotiation positions, when segregated by case type, elude rational, utility-based explanation. Plaintiffs’ mean demand, as a percentage of the mean award, is highest in intentional tort cases, where plaintiffs have a 31% chance of prevailing, and lowest in fraud cases, where plaintiffs have a significantly higher chance of prevailing (52%). Similarly, the incidence of plaintiff overpricing is highest in intentional tort and medical malpractice cases (83% of demands in both case types are overpriced), where plaintiffs have a relatively small chance of prevailing, and lower in contract cases (40% of demands are overpriced) and mixed contract/tort cases (53% of demands are overpriced) where plaintiffs have relatively high win rates. Plaintiffs, in short, are inclined to discount their strong claims and inflate their weak claims. Defendants’ negotiating strategies display similar patterns of risk aversion when facing gains and risk taking when facing losses. Defendants’ offers, as a percent of average awards, are low in many cases that defendants are likely to lose and high in many cases which defendants are likely to win. In both fraud and employment cases, for example, the plaintiff win rate is 52%, and the average defendant offer is about 6% of the average award. For the two types of cases with the lowest plaintiff win rates (medical malpractice – 21% and intentional tort – 31%), the mean defendant offers are 15% and 34% of the mean awards, respectively. Defendants’

58


offers suggest that they are gracious winners and ignoble losers, offering outsized settlement offers to likely losers and serving their cheapest offers to likely victors. If “no good deed goes unpunished” in the larger world, no good claim appears to go undeterred in the litigation world, at least for the cases that proceed to trial.

3.8.2.2

Statutory Offers of Compromise

The multivariate analysis indicates that the presence or absence of a statutory settlement offer, called a “998 offer” by California practitioners, is correlated with wide variations in win rates, decision error rates, and negotiation disparities. A 998 offer, which refers to settlement demands and offers made under California Code of Civil Procedure Section 998, is a statutory cost-shifting mechanism designed to encourage settlement and penalize unreasonable settlement positions. Any party can serve a written 998 offer to the other party while a case is pending, up to 10 days before trial commences.39 A party who does not accept an adverse party’s 998 offer and obtains a worse result at trial may be liable for the adverse party’s court costs, expert witness fees and, in personal injury cases, interest from the date of the offer. Although a party prevailing at trial usually is entitled to recover its costs from the losing party, the service of a 998 offer prevents a prevailing party from recovering costs if the award in its favor is less than its adversary’s 998 offer. “The purpose of section 998,” the court held in Taing v. Johnson Scaffolding Co.,40 “is to encourage the settlement of lawsuits before trial by penalizing a party who fails to accept a reasonable offer from the other party.” Decision Error. The multivariate analysis indicates that 998 offers are salient predictors of decision error for both plaintiffs and defendants. The results of the four possible 998 conditions (plaintiff alone serves a 998 offer on defendant, defendant alone serves a 998 offer on plaintiff, both parties serve 998 offers on each other, and neither party serves a 998 offer) are shown in Table 3.8. In general, serving a 998 offer reduces both decision error and mean cost of error for the serving party, but increases decision errors and the cost of errors for the recipient party. Total decision error also increases in the presence of a 998 offer, i.e., “no error” is always a lower percent when 998 offers are served. This increase in total decision error occurs because the reduction in the serving party’s decision error is more than offset by the rise in the recipient party’s decision error. 39

The “offer of compromise” under Section 998 must expressly refer to the statute or otherwise notify the offeree that costs otherwise allowed to a prevailing party may be reduced or augmented if the offer is not accepted. See Stell v. Jay Hales Development Co., 11 Cal. App. 4th 1214, 1231, 1232, 15 Cal. Rptr. 2d 220 (1992). An oral offer purportedly made under Section 998, even if placed on the record during a deposition, does not satisfy the statutory requirements. Saba v. Crater, 62 Cal. App. 4th 150, 153, 72 Cal. Rptr. 2d 401 (1998). Cf. Berg v. Darden, 120 Cal. App. 4th 721, 727, 15 Cal. Rptr. 3d 829 (2004). 40 9 Cal. App. 4th 579, 583, 11 Cal. Rptr. 2d 820 (1992). Taing was distinguished in Bihun v. AT&T Information Systems, 13 Cal. App. 4th 976, 6 Cal. Rptr. 2d 787 (1993).

3.8 Predictor Variables Table 3.8 Win rates and decision error rates by 998 offer of compromise Party Serving Plaintiff Win Plaintiff Decision Defendant Decision Compromise Offer Rate (%) Error Rate (%) Error Rate (%) Plaintiff 65 39 50 Defendant 36 83 8 Plaintiff & 58 56 30 Defendant None 46 59 23

59

No Decision Error (%) 11 9 14 18

A plaintiff 998 offer reduces both decision error and cost of error for plaintiffs, but it raises decision error and the cost of error for defendants. Plaintiff’s service of a 998 offer appears to provoke or at least is correlated with more risk-taking behavior by defendants. Plaintiff decision error is reduced from 59% in the “no 998 offer” condition to 39% in the “plaintiff serves 998” condition; defendant decision error rises from 23% in the “no 998 offer” condition to 50% in the “plaintiff serves 998” condition. Plaintiffs’ service of a 998 offer thus is correlated with a 34% decrease in plaintiff decision error and roughly a twofold increase in defendant decision error. When a defendant serves a 998 offer on a plaintiff, a similar effect is observed. A defendant 998 offer is correlated with a reduction in defendant decision error from 23% in the “no 998” condition to 8% in the “defendant serves 998” condition, while plaintiff error rises from 59% in the “no 998” condition to 83% in the “defendant serves 998” condition. Defendants serving a 998 offer reduce their decision error rate by 65% while raising plaintiffs’ decision error rate by 40%. Service of a 998 offer also is correlated with higher overall decision error. Eighteen percent of the cases showed “no decision error” in the “no 998 offer” condition, but the percentage of “no decision error” cases declined to 11% in the “plaintiff serves 998” condition and 9% in the “defendant serves 998” condition. Omitting the cases in which a 998 offer was submitted would roughly double the instances in which neither party made a decision error. Ironically, eliminating the cases where a party utilized a statutory procedure intended to deter unreasonable settlement positions would result in a lower overall incidence of decision errors in this dataset of cases with adjudicated outcomes. Cost of Error. Parties serving a 998 offer also display a dramatically lower mean cost of error compared to the mean cost of error sustained when they receive a 998 offer. Plaintiffs serving a 998 offer incur a mean cost of error of $21,671, which doubles to $43,153 when they are on the receiving end of a 998 offer. For defendants, the mean cost of error is $857,454 when they serve a 998 offer and $1,574,129 when they receive a 998 offer. For both plaintiffs and defendants, the mean cost of decision error is substantially higher when they are a recipient of a 998 offer. Dual 998 Offers. In the dual 998 offer condition – when both parties serve 998 offers – the results are mixed. Plaintiffs in the dual 998 offer condition show a slight reduction in decision error compared with the “no 998 offer” condition (56% v. 59%)

60


and a reduction in the mean cost of error ($56,525 v. $94,828). Defendants in the dual 998 offer condition exhibit an increase in decision error compared with the “no 998 offer” condition (30% v. 23%) and a more substantial decrease in the mean cost of error ($920,546 v. $1,528,599). Negotiation Disparities. Apart from its correlation with increased decision error and higher mean cost of error for parties receiving a 998 offer, the existence of a 998 offer is associated with extreme negotiation disparities. Service of a 998 offer is correlated with relatively conservative demands by plaintiffs and generous offers by defendants, while receipt of a 998 offer is associated with extremely high plaintiff demands and low defense offers. The mean amount of a 998 offer served by a plaintiff upon a defendant is 54% of the mean award, but the mean amount of a plaintiff’s demand soars to 331% of the mean award when a defendant serves a 998 offer upon a plaintiff. The average amount of a defendant’s offer is 47% of the average award when it serves a 998 offer upon a plaintiff, but when a defendant is on the receiving end of a 998 offer from a plaintiff, the average amount of a defendant’s offer is reduced to 19% of the average award. Plaintiffs overprice their demands in 50% of the cases where they serve a 998 offer, but plaintiffs receiving 998 offers from defendants overprice their demands in 92% of the cases. Defendants overprice their offers in 76% of the cases where they serve a 998 offer, but their incidence of overpricing falls to 17% when plaintiffs serve a 998 offer upon them. When defendants underprice their offer, moreover, the average amount of underpricing, relative to the average award, varies dramatically depending on who has served a 998 offer. The mean amount of underpricing by a defendant who receives a 998 offer from a plaintiff is $1,753,131, but the mean amount of defense underpricing drops to $505,675 when a defendant serves a 998 offer upon a plaintiff. Both plaintiffs and defendants, on average, are magnanimous when serving a 998 offer and penurious when receiving one. Limitations on 998 Data. The correlation between 998 offers and increased risk taking has at least two major limitations. First, this study is limited to adjudicated cases; 998 offers may be effective in promoting settlement and facilitating reasonable settlement positions in settled cases. As observed earlier, decision error and negotiation disparities cannot be tested in settled cases because they lack a benchmark award. Second, one may argue that a 998 offer does not cause the risk-taking behavior but rather is propounded to curb or penalize extreme settlement positions after an adverse party has manifested unreasonable settlement behavior. Under this argument, a 998 offer may be a reaction to, not a cause of, an adverse party’s risktaking behavior. The weakness in this second argument, though, is that it overlooks the underlying intent of the 998 statutory procedure: to promote reasonable settlement behavior by imposing a financial penalty on unreasonable settlement positions, whether the recipient party is a reckless or a rational decision maker. Although 998 offers may have a salutary effect on those cases that settle, in this study of adjudicated cases the service of a 998 offer is correlated with significantly higher risk taking and decision error by the recipient party. Four Corroborative Studies. The legislative intent underpinning the Section 998 offer of compromise procedure, as previously explained, is to encourage


61

settlements by imposing financial penalties on parties who take unreasonable settlement positions. This study of 998 offers of compromise, however, indicates that this cost-shifting statutory scheme may actually induce risk taking by litigants. The possibility that punitive measures in general and cost-shifting statutory schemes in particular are ineffective or may actually provoke the gambling mentality they are intended to curb is demonstrated by four other studies, two of which are empirical and two of which are experimental. These corroborative studies, which raise questions regarding the effect of “tort reform” propositions, are briefly described below. First, law professor Jeffrey Rachlinski studied the effect of “loser pays” legislation, enacted to deter meritless lawsuits and increase settlements. He analyzed the research conducted by Snyder and Hughes on medical malpractice claims filed before and after Florida adopted “reform” legislation requiring the losing party at trial to pay the prevailing party’s litigation costs. (“Loser-pays” is also known as the “English Rule”). Comparing the “before” and “after” results, Snyder and Hughes found that “a case was less likely to be settled under a loser-pays system” and “defendants spent much more on litigation under a loser-pays system.”41 These findings, Rachlinski observes, are consistent with behavioral economics models showing that the propensity to take risks increases as the stakes are heightened. “Increasing the risks associated with litigation,” Rachlinski observes, “increases the attractiveness of wasteful litigation to risk-seeking defendants.”42 He further notes “by raising the stakes at trial, the loser-pays system makes litigation itself more valuable and can discourage settlement.”43 Second, the instant study separately analyzed the effect of a different statutory cost-shifting procedure under California Code of Civil Procedure Section 1141.21. That section applies when the parties participate in a non-binding arbitration hearing before a trial and, dissatisfied with the arbitrator’s decision, one party exercises its right to a trial after the arbitration. Under Section 1141.21, a party who requests a trial after non-binding arbitration but fails to obtain a better result at trial may be required to pay specified expert witness fees, court costs, and the arbitrator’s fees. Like the 998 procedure, this statutory scheme is intended to discourage trials by financially punishing over-confident plaintiffs and defendants.44

41

Rachlinski supra note 11 at 163. Rachlinski supra note 11 at 168. 43 Rachlinski supra note 11 at 161. 44 The court explained the background and purpose of Section 1141.21 in Phelps v. Stostad 16 Cal. 4th 23, 65 Cal. Rptr. 2d 360, 939 P.2d 760 (1997): “To encourage parties to accept reasonable arbitration awards, the Legislature enacted Code of Civil Procedure section 1141.21, which provides that if a party elects a trial de novo following judicial arbitration, and fails to obtain a judgment that is more favorable than the arbitration award, that party shall pay the costs incurred by the opposing party following the election of the trial de novo and shall not recover his or her own costs incurred following the election of the trial de novo. Section 1141.21(a)(ii) creates an 42

62


Analysis of the Section 1141.21 variable, however, reveals no statistically significant difference in plaintiff or defendant error rates between cases where the parties requested a new trial after non-binding arbitration and those cases where the parties proceeded directly to trial without non-binding arbitration. Nor is there any statistically significant difference in error rates between parties who requested a new trial after arbitration, exposing themselves to a monetary sanction if they failed to obtain a superior award at trial, and parties who participated in pre-trial mediation and had no exposure to monetary sanctions if the trial award was inferior to what they could have obtained in mediation. Thus, as applied to the adjudicated cases in this study, Section 1141.21 and its cost-shifting sanctions do not appear to decrease risk taking or decision error by either plaintiffs or defendants. Third, an experimental study published in November 2007 by Kevin McCabe and Laura Inglis at George Mason University simulated litigation settlement negotiations and specifically evaluated the effects of 998 offers on negotiation outcomes. Overall, McCabe and Inglis found that “in 46% of the cases that went to court in our experiments, both parties had received settlement offers that would have made them better off than the expected court outcome.” Based on experiments controlling for 998 offers of compromise, they concluded that the Section 998 cost-shifting mechanism “provides no significant improvement in settlement rates.” They further note that the English Rule, requiring an unsuccessful litigant to pay both sides’ court costs, “is actually less effective at promoting settlement than the American Rule.”45 Fourth, writing in the journal Nature four months after publication of the McCabe-Inglis study, Harvard professor Martin Nowak and his colleagues noted that punitive measures do not promote cooperative behavior and are correlated with the least successful negotiation results. Their study, entitled “Winners Don’t Punish,” tested the premise that punitive measures can be successful in forcing parties to cooperate in one-time competitive interactions. Based on outcomes recorded in a modified version of “Prisoner’s Dilemma,” a game in which the players compete against anonymous opponents and choose among cooperation, betrayal or punishment, Nowak concluded that “winners are those who resist the temptation to escalate conflicts, while the losers punish and perish.” His co-author, Professor David Rand, specifically observes that the players who most frequently invoked the punishment option “earned the lowest playoffs in the game.” As Rand explains, “Punishment can lead to a downward spiral of retaliation, with destructive outcomes for everybody involved. The people with the highest total payoffs do not use costly punishment.”46

exception to the usual rule that the prevailing party in an action ‘is entitled . . . to recover costs.’” The purpose of this statute, the Phelps court stated bluntly, “is to discourage trials de novo.” 45 McCabe, Kevin, and Inglis, Laura. (2007, November 30). Using neuroeconomics experiments to study tort reform. Arlington, Virginia: Mercatus Center at George Mason University. 46 Dreber, Anna, et al. (2008, 20 March). Winners don’t punish. Nature, 452(7185), 348–351.


3.8.2.3

63

Forum

The three forum variables are jury trials, bench (judge only) trials, and arbitration hearings. Each of the forum variables is correlated with distinct decision error rates, negotiation disparities and win rates. Decision error rates and win rates for each type of forum are presented in Table 3.9. Defendants committed substantially fewer decision errors in jury trials relative to bench trials (23% v. 39%). By contrast, plaintiff decision error was considerably higher in jury trials relative to bench trials (62% v. 45%). In arbitration cases, decision error rates for both plaintiffs and defendants differed substantially from their rates in jury cases. Defendants’ decision error rate in arbitration cases (48%) was higher than their error rate in bench trials (39%) and roughly double their decision error rate in jury trials (23%). Plaintiffs’ decision error in arbitration cases (27%) was notably lower than in either bench trials (45%) or jury trials (62%). The total amount of decision error, moreover, is much lower in arbitration cases than in either bench or jury trials, with “no error” cases comprising 25% of all arbitration cases relative to 16% of bench trials and 15% of jury trials. The cost of error for both plaintiffs and defendants shows a similar progression from arbitration to bench and jury trials. Plaintiffs and defendants sustain the lowest mean cost of error in arbitration cases ($17,264 and $692,614, respectively), the next lowest mean cost of error in bench trials ($25,140 and $972,313, respectively) and the highest mean cost of error in jury trials ($61,875 and $1,495,184, respectively). Regardless of the forum, however, defendants’ mean cost of error is always higher than plaintiffs’ mean cost of error, ranging from a high of 40 times plaintiffs’ cost of error in arbitration cases to a low of 24 times plaintiffs’ cost of error in jury trials. Defendants also exhibit extreme negotiation disparities in arbitration cases. Defendant’s mean offer in the arbitration cases, as a percentage of the mean award, is 6%, compared with 21% in jury and bench trials. Defendants also are more likely to underprice their settlement offers in arbitration cases; 73% of defendants’ offers are underpriced in arbitration cases; but the incidence of defendant underpricing is reduced to 55% and 38% in bench and jury trials, respectively. The relatively low amount of defendants’ offers in arbitration cases suggests that defendants have a diminished incentive to avoid adjudicated outcomes in arbitration cases. Plaintiffs, in contradistinction, do not exhibit the same degree of negotiation disparity in the three different forums. Plaintiffs’ mean demand, as

Table 3.9 Win rates and decision error rates by forum Forum Plaintiff Win Plaintiff Decision Rate (%) Error Rate (%) Arbitration 74 27 Bench (Judge) 58 45 Jury 47 62

Defendant Decision Error Rate (%) 48 39 23

No Decision Error (%) 25 16 15

64


a percentage of the mean award, is 110% in arbitration cases, 96% in bench trials, and 113% in jury trials. Consistent with behavioral economics theories of loss aversion in the “gains” frame and risk taking in the “losses” frame, plaintiffs’ decision error rate decreases as the probability of a trial victory increases. Plaintiffs exhibit the highest decision error rates in jury trials (62%), where the win rate is lowest (47%), and the lowest decision error rates (27%) in arbitration cases, where the win rate is highest (74%). Defendants display a similar pattern, evidencing a high decision error rate (48%) in arbitration cases, where they win only 26% of the cases, and a low decision error rate (23%) in jury trials, where they win 53% of the cases. For both plaintiffs and defendants, the incidence of decision error increases as the likelihood of prevailing decreases; they are most likely to reject a settlement demand or offer that is proven to be superior to the ultimate award when they have a relatively lower chance of prevailing at trial. Like losing bettors at a racetrack, litigants are more likely to bet on long shots as their financial prospects dim. The pattern of risk taking in the losses mode and risk aversion in the gains mode is again reflected in the parties’ settlement posturing. As noted above, the average defendants’ offer is a miserly 6% of the average award in arbitration cases and a relatively generous 21% of the mean award in jury trials, indicating that defendants offer less money, relative to their actual liability, in cases they are likely to lose, and more money in cases they probably will win. Defendants underprice their offers in 73% of the arbitration cases and 38% of the jury cases, again reflecting the pattern of miserliness toward plaintiffs in cases defendants are likely to lose and relative profligacy toward plaintiffs in cases defendants are likely to win. Plaintiffs are similarly aggressive and risk taking in cases they are likely to lose and more conservative and risk averse in cases they are likely to win; plaintiffs underprice their demands in 23% of the jury trials (47% win rate) and 39% of the arbitration cases (74% win rate). The higher win rates for plaintiffs in arbitration cases are consistent with other studies documenting win rates ranging from 63% to 80%.47 This research suggests

47

See Maltby, Lewis. (1999, Fall). Employment arbitration: Is it really second class justice? Dispute Resolution Magazine. Maltby, Lewis L. (1998). Private justice: Employment arbitration and civil rights. Columbia Human Rights Law Review, 30, 29, 46–48. Bingham, Lisa B. (1995). Is there a bias in arbitration of nonunion employment disputes: An analysis of active cases and outcomes. International Journal of Conflict Management, 6, 369, 378. Howard, William. (1995, October-December). Arbitrating claims of employment discrimination. Dispute Resolution Journal, 40–43. Ernst and Young. (2004, December 2004). Outcomes of arbitration: An empirical study of consumer lending cases. National Arbitration Forum. (2000, January 7). Millennial issues regarding arbitration fairness: An administrator’s view. Baxter, George. (1993–94). Arbitration in litigation for employment civil rights? 2 Vol. of Individual Employee Rights 19. Delikat, Michael, and Kleiner, Morris. (2003, Winter). Comparing litigation and arbitration of employment disputes: Do plaintiffs better vindicate their rights in litigation? Conflict Management, 6(3). United States General Accounting Office. (1992, May 11). Securities arbitration: How investors fare (Rep. No. GAO/ GGD-92–74). American Arbitration Association. Analysis of the American Arbitration Association’s consumer arbitration caseload: Based on consumer cases awarded between January and


65

that plaintiffs may benefit from reconsidering the conventional wisdom that jury trials favor them (although many scholars argue that the probability of a plaintiff award is vastly outweighed by the relatively small amount of arbitration awards). The high win rate in bench trials (58%) vs. jury trials (47%) also augurs for a reconsideration of the conventional plaintiff preference for jury trials. (In the criminal law context, recent research also raises questions about the advantages of jury trials for defendants. This research shows that judges in 2006 convicted defendants in only 64% of the cases, compared to a jury conviction rate of 89%. As legal affairs writer Jason Krause notes, “The figures contradicted one of the oldest pieces of conventional wisdom in the legal profession: that defendants get more sympathy from a jury than a judge”).48

3.8.2.4

Damages Claim

Damages are coded in the datasets as (a) “Past” damages, representing injuries and damages already sustained; (b) “Future” damages, representing prospective, anticipated losses not yet paid or sustained, and (c) punitive or exemplary damages. A plaintiff in a personal injury suit against an intoxicated driver, for example, may seek compensation for medical expenses already incurred and pain and suffering previously suffered (Past damages); the cost of future surgery anticipated by her physician and prospective pain and suffering (Future damages); and punitive damages based on the defendant’s reckless behavior while driving intoxicated. The coding in the datasets is based on plaintiffs’ damages allegations, not the type of damages ultimately recovered by plaintiffs. Awards generally are not sufficiently allocated by VerdictSearch California and the adjudicator to consistently determine the type of damages ultimately awarded. Distinctions Among Damages Claims. Behavioral economics theory posits that a party is more likely to recover losses already incurred (“Past” damages) than presently unrealized future profits or other relatively remote, prospective damages (“Future” damages), even when a party is entitled to recover both types of damages.49 In a breach of contract action against a contractor who abandons a house construction project, for example, the plaintiff is more likely to recover its advance payment to the contractor than the rental income that would have been realized between the original contract completion date and the actual completion date.50 Although a non-breaching party is entitled to “the amount which will August 2007. Consumer Arbitration Task Force - Searle Civil Justice Institute. (2009, March). Consumer arbitration before the American Arbitration Association preliminary report. 48 Krause, Jason. (2007, June). Judge v. jury. ABA Journal, p. 46. 49 See Baron, Jonathan. (2000). Thinking and deciding (pp. 409–431). Cambridge: Cambridge University Press. 50 Facts based on Henderson v. Oakes-Waterman Builder, 44 Cal. App. 2d 615, 12 P. 2d 662 (1941), reversing trial court’s determination of damages and holding owner was entitled to recover advance payment, cost of demolition and reconstruction, and loss of rental value.

66


Table 3.10 Win rates and decision error rates by damages claim Damages Claim Plaintiff Win Plaintiff Decision Defendant Decision Rate (%) Error Rate (%) Error Rate (%) Past Losses Only 46 64 21 Future Losses Only 35 65 26 Past/Future Losses 51 57 28 Past/Punitive 56 49 37 Past/Future/Punitive 75 32 49

No Decision Error (%) 15 9 15 14 19

compensate the party aggrieved for all the detriment proximately caused thereby,” i.e., the equivalent of the benefits of contract performance,51 studies show that jurors and judges are reluctant to award both damages actually incurred and damages yet to be sustained.52 This bias purportedly manifests because previously incurred damages are perceived as tangible and proximate, while future losses may be perceived as speculative and remote.53 Effects of Damages Claims. The primary dataset presents mixed messages about the existence of a cognitive distinction between “Past” damages awards and “Future” damages awards. As indicated in Table 3.10, plaintiffs seeking only Future damages fared poorly, recovering a net award in only 35% of the cases. Plaintiffs alleging only Past damages, by contrast, prevailed in 46% of their cases. Plaintiffs seeking both Past and Future damages, however, recovered a net reward in 51% of the cases, a win rate slightly higher than the cases alleging only Past damages. Table 3.10 demonstrates the relation between the damages claim and the parties’ decision errors. Cases alleging only Past damages are correlated with a low defendant decision error rate (21%), but defendants’ error rate increases markedly as the damages claim becomes more complex. Faced with a claim alleging Past, Future, and punitive damages, defendants’ decision error rate more than doubles to 49%. Plaintiffs’ decision error rate, however, moves in the opposite direction, from a high of 64% in cases alleging Past damages only to a low of 32% in cases asserting Past, Future, and punitive damages. Litigants’ mean cost of error also is skewed differently for defendants and plaintiffs. Defendant’s mean cost of error is highest in cases alleging Past, Future, and punitive damages ($6,995,791) and lowest in cases alleging only Past damages ($333,938). Plaintiffs’ mean cost of

51

California Civil Code Section 3300. Cohen, David, & Knetsch, Jack L. (1992). Judicial choice and disparities between measures of economic value. In Kahneman, Daniel, & Tversky, Amos (Eds.). (2000). Choices, values, and frames (pp. 436–439). Cambridge: The Press Syndicate of the University of Cambridge. 53 An employer’s wage cuts, to cite another example, are more likely to be considered unacceptable than the employer’s failure to increase wages. Cohen, David, & Knetsch, Jack L. (1991). Loss aversion in riskless choice. In Kahneman, Daniel, & Tversky, Amos (Eds.). (2000). Choices, values, and frames (p. 157). Cambridge: The Press Syndicate of the University of Cambridge. 52


67

error, by comparison, is $78,404 in Past, Future, and punitive damages cases and $30,021 in Past damages cases; plaintiffs’ highest mean cost of error is sustained in cases alleging Past and Future damages ($121,700). But when mean cost of error is replaced with median cost of error, both plaintiffs and defendants exhibit their lowest median costs of error in Past damages only cases and their highest median cost of error in Past, Future, and punitive damages cases. Punitive Damages Claims. The presence of a punitive damages claim is correlated with high defendant decision error rates. Defendant decision error rose from 21% in Past damages only claims to 37% in Past and punitive damages claims, and from 28% in Past and Future damages claims to 49% in Past, Future, and punitive damages claims. By contrast, plaintiffs’ decision error was consistently lower in cases alleging punitive damages. When a punitive damages claim was joined with a Past damages claim, plaintiff decision error decreased from 64% (Past damages only) to 49% (Past and punitive damages). In cases where a punitive damages claim was joined with a Past and Future damages claim, plaintiff decision error decreased from 57% (Past and Future damages only) to 32% (Past, Future and punitive damages). The substantially higher defendant decision error rates in punitive damage claims may be attributable to the difficulty of predicting the amount of punitive damage awards and defendants’ inadequate evaluative adjustments for non-paradigmatic claims. Experimental studies show that individual differences in punitive damage awards “produce severe unpredictability and highly erratic outcomes;” study participants show strong agreement in finding punitive intent, but “there is no consensus about how much in the way of dollars is necessary to produce appropriate suffering in a defendant.”54 The proposition that punitive damages awards are unpredictable, however, is challenged by Theodore Eisenberg’s empirical study published in 2006. His study found, inter alia, “minimal, though observable, variation in the dispersion of the punitive and compensatory damage ratio over the years [1992–2001] and between trial modes.”55 Negative Problem-Solving Transfer in Punitive Damages Claims. Whether the amount of punitive damages is predictable or unpredictable, the defendants in the primary dataset displayed seriously diminished predictive capacity in punitive damages claims. The defendants’ relatively poor outcomes suggest that they either ignore the non-paradigmatic variable (punitive damages claim) or erroneously draw problem-solving analogies between the unexceptional cases (no punitive damages claim) and the exceptional cases (punitive damages claims). The risk of this type of decision-making error (“negative problem-solving transfer”) is high when cases 54

Sunstein, Cass, et al. Assessing punitive damages (with notes on cognition and valuation in law). In Sunstein, Cass (Ed.). Behavioral law & economics (pp. 232, 240). Cambridge: Cambridge University Press. 55 Eisenberg, Theodore, et al. (2006). Juries, judges, and punitive damages: Empirical analysis using the civil justice survey of state courts 1992, 1996, and 2001 data. Journal of Empirical Legal Studies, 3(2), 276. Cf. Hersch, Joni, and Viscusi, W. Kip, (2004). Punitive damages: How judges and juries perform. Journal of Legal Studies, 33(1), 1–36.

68


appear superficially similar: surface similarity in story line, causes, context, and phrasing, explains Professor Miriam Bassok, frequently leads decision makers to “retrieve and apply a solution to a nonanalogous problem (negative transfer) and thereby waste their cognitive resources or arrive at an erroneous solution.”56 This inability to meet decision-making challenges by changing mental templates is well documented and explained in the 2008 Harvard Business Review article, “The Experience Trap:” We conclude that managers find it difficult to move beyond the mental models that they have developed from their experiences in relatively simple environments or that have been passed on to them by others. When complications are introduced, they either ignore them or try to apply simple rules of thumb that work only in noncomplex situations. What they don’t do is materially improve the quality of their mental models to take into account the realities of complex projects.57

Like the recalcitrant managers described in “The Experience Trap,” many defendants seem to impose tried and true mental models on new and uncertain conditions, often leading to disastrous results. The cognitive template that is so successful in defendants’ assessment of simple claims appears to be ill-suited for defendants’ evaluation of complex claims. As Peter Drucker observed in his classic essay, “The Effective Decision,” one of the most common decision-making errors “is the mistake of treating a new event as if it were just another example of the old problem to which, therefore, the old rules should be applied.”58 Although defendants may not lump all cases onto the “routine” pile, the danger is that they misclassify cases when sorting the routine from the exceptional cases or treat cases in both categories with the same evaluative models. When decision makers have “seen it all before,” their time-tested evaluative tools may superimpose comparisons and analogies over incongruent facts and encourage decision makers to minimize nuances later proven to be determinative. Effect of Aggregated Claims on Win Rates and Negotiation Disparities. Apart from its impact on decision error, a plaintiff’s damages claim may affect win rates and the parties’ negotiation strategies. As plaintiffs aggregate different types of damages in their claims, their win rates increase. Cases alleging both Past and Future damages, for example, have a higher win rate (51%) than cases alleging Past damages alone (46%). When punitive damage claims are added to cases alleging Past and/or Future damages, the win rate soars to 75%. Plaintiffs overprice their demands in 79% of the cases alleging Past damages only, but plaintiffs’ incidence of overpricing decreases to 51% in cases alleging Past, Future, and punitive

56

Bassok, M. (2003). Analogical transfer in problem solving. In Davidson, J.E. and Sternberg, R. J. (Eds.) The psychology of problem solving (p. 343). Cambridge: Cambridge University Press. 57 Sengupta, Kishore, Abdel-Hamid, Tarek K. and Van Wassenhove, Luk N. (2008, February). The experience trap. Harvard Business Review, p. 94. 58 Drucker, Peter. (1967). The effective decision. In Harvard Business Review on Decision Making (p. 6). Boston, Massachusetts: Harvard Business School Press.


69

damages. Conversely, defendants underprice their offers in 36% of the cases alleging Past damages only and 68% of the cases alleging Past, Future and punitive damages. In alleging multiple types of damages, plaintiffs may increase their win rates by providing story and remedy alternatives to a jury and affording an opportunity for the jury to compromise the parties’ competing narratives, motivations, and expectations. Cognitive psychology studies demonstrate that choices are affected significantly by the number of alternatives,59 “different frames, reference points, contexts and elicitation procedures,”60 the addition of options, and the ability to compromise between two extremes.61 People avoid making any choices when faced with two unattractive or unpersuasive alternatives, but the addition of a third, less attractive or less persuasive alternative sometimes increases the desirability of an option about which one was previously ambivalent.62 Faced with a stark choice between awarding speculative, uncertain Future damages to a plaintiff or rendering a defense verdict, the jury is more likely to render a defense verdict. But given a choice between a defense verdict and a plaintiff’s aggregated claim for Future damages, Past damages, and punitive damages, two cognitive changes appear to occur: (1) the bias against Future damages ebbs when Future damages are combined with other types of damages clams; and (2) Past damages claims are more appealing and hence successful when combined with other damages claims. The bundling of damages claims may functionally obscure the blemishes evident when a lone damages claim is asserted. The opportunity to choose among different damages claims and to select among the alternative narratives that underpin those claims also enables the jury to negotiate mentally with the plaintiff and defendant and compromise their competing claims. A plaintiff alleging both contract damages and punitive damages (e.g., “the defendant breached the contract by failing to perform and defrauded my client because he never intended to perform”) permits the jury to choose a credible, middle ground story (“the defendant breached the contract but did not defraud the plaintiff”) and avoid the dissonance and unease generated by making stark, zero sum decisions. Although arbitrators often are accused of avoiding tough decisions by “splitting it down the middle,” jurors also may prefer to effect a compromise among competing claims. Asking for more damages instead of different types of damages also can result in higher verdicts. In mock trials, jurors presented with the same evidence awarded 59

Thaler, Richard H. (1999). Mental accounting matters. In Kahneman, Daniel, & Tversky, Amos (Eds.). (2000). Choices, values, and frames (p. 265). Cambridge: The Press Syndicate of the University of Cambridge. 60 Shafir, Eldar, Simonson, Itmar, & Tversky, Amos. (1993). Reason-based choice. In Kahneman, Daniel, & Tversky, Amos (Eds.). (2000). Choices, values, and frames (p. 618). Cambridge: The Press Syndicate of the University of Cambridge. 61 Baron supra note 49 at 288. 62 Hastie, Reid, & Dawes, Robyn M. (2001). Rational choice in an uncertain world (pp. 237–244). Thousand Oaks, California: Sage Publications, Inc.

70


higher amounts to the plaintiff when the only variable was a higher amount requested by the plaintiff’s attorney.63 As law professor Chris Guthrie observes, the lesson from the studies seems to be “Ask and ye shall receive.”64

3.8.2.5

Offer/Demand Ratio

Negotiation theorists contend that a negotiated resolution is less likely to occur when the parties’ demands and offers are far apart. All things being equal, they assert, parties whose negotiation positions are extremely disparate are less likely to settle than parties whose negotiation positions are relatively close. This section does not challenge this theory (its dataset would be useless in testing the theory because it includes only non-settled cases) but rather examines the relationship between decision error and the magnitude of the difference in the parties’ settlement positions. It specifically tests the conventional wisdom that, if the parties are far apart in their settlement positions, the “truth” lies somewhere in between, i.e., plaintiff’s demand is too high relative to the probable outcome, and defendant’s offer is too low. Applied to the concept of decision error, the conventional wisdom dictates that neither party is likely to make a decision error when their settlement positions are far apart. If, for example, plaintiff demands $1,000,000 and defendant offers $100,000, the conventional wisdom holds that the likely verdict is somewhere in between and neither party will commit a decision error by rejecting the other party’s “extreme” settlement proposal. As shown in Fig. 3.1, all cases in the dataset can be broken down into categories based on the range of quotients derived from dividing the amount of a defendant’s offer by the amount of a plaintiff’s demand. When defendant’s offer is $10 and plaintiff’s demand is $100, for example, the quotient is .10; when defendant’s offer is $90 and plaintiff’s demand is $100, the quotient is .90. If the conventional wisdom is correct, a lower quotient will be correlated with lower overall decision error, as the margin of error is broad, and a higher quotient will be correlated with higher overall decision error, as the margin of error is narrowed for both parties. The data, as shown in Fig. 3.1, do not support the conventional wisdom. The “truth” in settlement negotiations does not lie somewhere between two extreme positions but is nestled closer to defendants’ settlement positions when the parties’ demands and offers are far apart. Plaintiffs are more likely to commit decision errors when the offer/demand quotient is low, and their incidence of decision error generally decreases as the parties’ settlement positions move closer to each other. Defendants’ decision error rate, however, generally increases as the parties’ settlement positions converge. (Caveat to defendants: although the incidence of 63

Chapman, Gretchen B., & Bornstein, Brian H. (1996). The more you ask for the more you get: Anchoring in personal injury verdicts. Applied Cognitive Psychology, 10, 519, 525. 64 Guthrie, Chris, Rachlinski , Jeffrey J. and Wistrich, Andrew J. (May 2001). Inside the judicial mind. Cornell Law Review, 86(4), 789–790.


71

80% 70% 60% 50% Plaintiff Defendant

40% 30% 20% 10%

9%

9% %

–7 80

%

–8

9% –6 70

9% –5

%

60

9% % 50

9% 40

%

–4

9% %

–3

–2 30

%

–1

20

% 10

Beyond Right and Wrong: The Power of Effective Decision Making for Attorneys and Clients

Beyond Right and Wrong: The Power of Effective Decision Making for Attorneys and Clients

Right and Wrong

Tools and Techniques for Effective Data-Driven Decision Making

Wrong for the Right Reasons

Wrong for the Right Reasons

Ethics: Inventing Right and Wrong

Decision making and programming

Right and Wrong Thinking and Their Results

Assessment and Decision Making for Sustainable Transport

Environmental Management and Decision Making for Business

Voting and Collective Decision-Making: Bargaining and Power

Decisions: Making the Right Decisions, Righting the Wrong Ones

Organizational Decision Making (Cambridge Series on Judgment and Decision Making)

Experiment, Right or Wrong

Wrong Groom, Right Bride

Naturalistic Decision Making and Macrocognition

Distributed Decision Making and Control

Pharmacy Ethics and Decision Making

Voting and collective decision-making

Measurement, Judgment, and Decision Making

Measurement Judgment and Decision Making

Moral Education: Beyond the Teaching of Right and Wrong (Philosophy and Education)

Handbook of Decision Making

The Art of Decision Making

The Art of Decision Making

Experiment, Right or Wrong

Right Place, Wrong Time

Right Place, Wrong Time

Wrong Groom, Right Bride

Beyond Right and Wrong: The Power of Effective Decision Making for Attorneys and Clients