Social Choice with Partial Knowledge of Treatment Response

132 Pages·2005·6.357 MB·English

by  Charles F. Manski

Social Choice with
Partial Knowledge of
Treatment Response
Social Choice with Partial Knowledge of
Treatment Response
IS A PART OF THE SERIES
ECONOMETRIC INSTITUTE LECTURES
Series Editors
Herman K. van Dijk and Philip Hans Franses
The Econometric Institute, Erasmus University Rotterdam
The Econometric Institute Lectures series is a joint project of
Princeton University Press and the Econometric Institute at
Erasmus University Rotterdam.
This series collects the lectures of leading researchers which
they have given at the Econometric Institute for an audience
of academics and students.
The lectures are at a high academic level and deal with topics
that have important policy implications. The series covers a
wide range of topics in econometrics. It is not confined to any
one area or sub-discipline.
The Econometric Institute is the leading research center in
econometrics and management science in the Netherlands.
The Institute was founded in 1956 by Jan Tinbergen and
Henri Theil, with Theil being its first director. The Institute
has received worldwide recognition with an advanced training
program for various degrees in econometrics.
The first volume in this series is Social Choice with Partial
Knowledge of Treatment Response by Charles F. Manski.
Social Choice with
Partial Knowledge of
Treatment Response
Charles F. Manski
Princeton University Press
Princeton and Oxford
Copyright © 2005 by Princeton University Press
Published by Princeton University Press,
41 William Street, Princeton, New Jersey 08540
In the United Kingdom: Princeton University Press,
3 Market Place, Woodstock, Oxfordshire OX20 1SY
All rights reserved
Library of Congress G'ataloguing-in-Publication Data
Manski, Charles F.
Social choice with partial knowledge of treatment response /
Charles F. Manski
p. cm. - (The Econometric Institute Lectures)
Includes bibliographical references.
ISBN 0-691-12153-2 (cl: alk. paper)
1. Social sciences—Statistical methods. 2. Social choice.
3. Estimation theory. I. Title. II. Series.
HA29.M2466 2006
300'.l'5195—dc22 2005048691
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available
from the British Library
This book has been composed in Computer Modern and
typeset by TfcT Productions Ltd, London
Printed on acid-free paper 0
www.pup.princeton.edu
Printed in the United States of America
10 987654321
Contents
Preface vii
1 Utilitarian Treatment of
Heterogeneous Populations 1
1.1 Studying Treatment Response to Inform
Treatment Choice 1
1.2 The Planning Problem 3
1.3 Practices that Limit the Usefulness of
Research on Treatment Response 11
2 The Selection Problem 19
2.1 Treatment Choice Using the Empirical
Evidence Alone 20
2.2 Monotone Treatment Response 37
2.3 Exclusion Restrictions 49
3 Treatment Using Experimental Data 56
3.1 The Expected Welfare (Risk) of a Statistical
Treatment Rule 56
3.2 Using a Randomized Experiment to Evaluate
an Innovation 62
3.3 Using Covariate Information with Data from
a Randomized Experiment 74
vi Contents
4 The Selection Problem with Sample Data 108
4.1 Sample-Analog Rules Using the Empirical
Evidence Alone 109
References 115
Preface
This monograph codifies and elaborates the Economet-
ric Institute and Princeton University Press (PUP) Lec-
tures that I presented at Erasmus University Rotterdam
in June 2004. I am grateful to the Econometric Insti-
tute, to PUP, and especially to Herman van Dijk for
this opportunity to unify and further develop a program
of research that I began in the late 1990s and continue
today. I am grateful to Adam Rosen for his careful read-
ing of the manuscript and to several anonymous review-
ers for their comments. The National Science Foundation
supported this work through grant no. SES-0314312.
Chapter 1 is a largely verbal, and I hope broadly
accessible, development of basic themes. Chapters 2-4
are necessarily more technical, but empirical illustra-
tions and numerical examples aim to help one inter-
pret the math. I feel that the analysis in these chapters
answers some important questions. Nevertheless, I antic-
ipate that readers will conclude, as I do, that the study
of social choice with partial knowledge of treatment
response is barely beyond its infancy. I shall be pleased
if publication of this book stimulates new research that
broadens and deepens the present analysis.