. 1
( 11)


This page intentionally left blank

This work offers a set of extended interpretations of Madison™s argument
in Federalist X of 1787, using ideas from social choice theory and from the
work of Douglass North, Mancur Olson, and William Riker. Its focus is not
on social choice theory itself, but on the use of this theory as a heuristic
device to better understand democratic institutions. The treatment adapts a
formal model of elections to consider rapid constitutional change at periods
when societies face social quandaries. The topics explored in the book include
Britain™s reorganization of its ¬scal system in the eighteenth century to pros-
ecute its wars with France; the Colonies™ decision to declare independence in
1776; Madison™s argument about the “probability of ¬t choice” during the
Rati¬cation period of 1787“8; the argument between Hamilton and Jefferson
in 1798“1800 over the long-run organization of the U.S. economy; the Dred
Scott decision of 1857 and the election of Lincoln in 1860; Lyndon Johnson
and the “critical realignment” of 1964; and Keynes™s rejection of the equilib-
rium thesis in 1937 and the creation of the Bretton Woods institutions after

Norman Scho¬eld is the William Taussig Professor of Political Economy at
Washington Univesity in St. Louis. He has served as Fulbright Distinguished
Professor of American Studies at Humboldt University Berlin in 2003“4,
a Fellow at the Center for Advanced Study in the Behavioral Sciences at
Stanford in 1988“9, and Sherman Fairchild Distinguished Scholar at the
California Institute of Technology in 1983“4. Professor Scho¬eld is the author
of Mathematical Methods in Economics and Social Choice (2003), Multiparty
Government (coauthored with Michael Laver, 1990), and Social Choice and
Democracy (1985). He received the William Riker Prize in 2002 for contribu-
tions to political theory and is co-receipient with Gary Miller of the Jack L.
Walker Prize for the best article on political organizations and parties in the
American Political Science Review for 2002“4.
political economy of institutions and decisions

Series Editor

Stephen Ansolabehere, Massachusetts Institute of Technology

Founding Editors

James E. Alt, Harvard University
Douglass C. North, Washington University, St. Louis

Other books in the series

Alberto Alesina and Howard Rosenthal, Partisan Politics, Divided Government,
and the Economy
Lee J. Alston, Thrainn Eggertsson, and Douglass C. North, eds., Empirical
Studies in Institutional Change
Lee J. Alston and Joseph P. Ferrie, Southern Paternalism and the Rise of the
American Welfare State: Economics, Politics, and Institutions, 1865“1965
James E. Alt and Kenneth Shepsle, eds., Perspectives on Positive
Political Economy
Josephine T. Andrews, When Majorities Fail: The Russian Parliament,
Jeffrey S. Banks and Eric A. Hanushek, eds., Modern Political Economy: Old
Topics, New Directions
Yoram Barzel, Economic Analysis of Property Rights, 2nd edition
Yoram Barzel, A Theory of the State: Economic Rights, Legal Rights, and the
Scope of the State
Robert Bates, Beyond the Miracle of the Market: The Political Economy
of Agrarian Development in Kenya, 2nd edition
Charles M. Cameron, Veto Bargaining: Presidents and the Politics
of Negative Power
Kelly H. Chang, Appointing Central Bankers: The Politics of Monetary Policy
in the United States and the European Monetary Union
Peter Cowhey and Mathew McCubbins, eds., Structure and Policy in Japan
and the United States: An Institutionalist Approach
Gary W. Cox, The Ef¬cient Secret: The Cabinet and the Development
of Political Parties in Victorian England

Continued on page following Index
Constitutional Quandaries and
Social Choice Theory

Washington University in Saint Louis
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press
The Edinburgh Building, Cambridge cb2 2ru, UK
Published in the United States of America by Cambridge University Press, New York
Information on this title: www.cambridge.org/9780521832021

© Norman Schofield 2006

This publication is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.

First published in print format 2006

isbn-13 978-0-511-21965-8 eBook (EBL)
isbn-10 0-511-21965-2 eBook (EBL)

isbn-13 978-0-521-83202-1 hardback
isbn-10 0-521-83202-0 hardback

isbn-13 978-0-521-53972-2 paperback
isbn-10 0-521-53972-2 paperback

Cambridge University Press has no responsibility for the persistence or accuracy of urls
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.

List of Tables and Figures page xi
Preface xiii

Constitutional Quandaries and Social Choice
1 1
1.1 Introduction 1
1.2 Balancing Risk and Chaos 3
1.3 Preferences and Judgments 11
1.4 The “Institutional Narrative” of the Book 17
Power and Social Choice
2 23
2.1 Introduction 23
2.2 The World Today 24
2.3 Democratic Dilemmas 30
2.4 The Logic of Empire 34
2.5 Social Choice Theory: Autocracy and Risk 44
2.6 Social Choice in Britain: 1625“1776 49
2.7 The Agrarian Empire in North America 58
2.8 The End of Empire in Britain 64
2.9 Concluding Remarks 68
Franklin and the War of Independence
3 71
3.1 Introduction 71
3.2 The Quandary of the Declaration of Independence 73
3.3 The Decision to Declare Independence 80
3.4 Appendixes 85
3.4.1 The Quebec Act, October 7, 1774 85
3.4.2 Declaration and Resolves of the First
Continental Congress, October 14, 1774 90
3.4.3 Declaration of Independence, July 4, 1776 94


Madison, Jefferson, and Condorcet
4 98
4.1 The Rati¬cation of the Constitution 98
4.2 The Con¬‚ict over Union and Confederation 100
4.3 Social Choice and Constitutional Theory 113
4.4 Land and Capital in North America, 1756“1800 117
4.5 The In¬‚uence of Condorcet on Madison and Jefferson 120
4.6 Origins of the Two-Party System in the 1790s 125
4.7 Concluding Remarks 128
4.8 Appendix 132
4.8.1 Speech by Benjamin Franklin to the
Constitutional Convention on
September 17, 1787 132
Lincoln and the Civil War
5 135
5.1 Introduction 135
5.2 The Intersectional Party Balance 136
5.3 Dred Scott and the Supreme Court, 1857 146
5.4 The Illinois Election of 1858 149
5.5 Lincoln in New York and New Haven 155
5.6 The Presidential Election of 1860 158
5.7 Concluding Remarks 163
Johnson and the Critical Realignment of 1964
6 166
6.1 Introduction 166
6.2 Partisan Realignments from 1896 to 2000 167
6.3 Party Competition in Two Dimensions 170
6.4 Equilibrium in Candidate Competition 174
6.5 Party Activist Equilibrium 176
6.6 A Joint Model of Activists and Candidates 178
6.7 Third Parties 182
6.8 Summary of the Model 183
6.9 Partisan Strategies 185
6.10 Choices, Credible Commitment, and Path Dependence 194
6.11 Concluding Remarks 195
Keynes and the Atlantic Constitution
7 200
7.1 Introduction 200
7.2 Ordering the Political Economy 201
7.3 Prophets of Chaos 207
7.4 Political and Economic Beliefs in the Constitution 212
7.5 The Collapse of Hegemony in the 1970s 217
7.6 Keynes and the Quandary of the 1930s 223


7.7 The Constitutional Quandary of 1944 229
7.8 Architects of Change, 1944“1948 233
7.9 Concluding Remarks 240
Preferences and Beliefs
8 243
8.1 Introduction 243
8.2 Hobbesian and Lockean Views of Society 249
8.3 Condorcet and Social Truth 252
8.4 Scienti¬c Truth 256
8.5 Core Beliefs and the Heart 260
8.6 Modelling Beliefs 263
8.7 Modelling Elections 266
8.8 A Formal Model of Elections 268
8.9 Appendix: Condorcet™s Jury Theorem 273
Political Change
9 276

Bibliography 281
Index 309

List of Tables and Figures

2.1a A Comparison of U.S. and Soviet Military Expenditure,
page 38
2.1b A Comparison of U.S. and Russian Military Expenditure,
1992“1999 38
2.2 Twelve Democratic Polities: 1980“1995 41
2.3 Twelve Democratic Polities: August 2005 42
5.1 U.S. Presidential Elections: 1836“1860 139
5.2 The Election of 1860 142
A6.1 The Election of 1964 197
A6.2 The Election of 1968 198
A6.3 The Election of 1972 199

Chaos or autocracy in a polity
1.1 4
Estimates of GDP per capita (in 1985 dollars) for six OECD
countries, 1950“1992 39
Chaos or autocracy in a polity
2.2 46
Whig and Tory bargains in the British polity in the
eighteenth century 53
A schematic representation of land and capital in the United
States, 1800“1860 62
A schematic representation of the election of 1860 in a
two-dimensional policy space 63
A schematic representation of land and capital in Britain,
1720“1846 65

List of Tables and Figures

Estimated voter distribution and party positions in Britain,
1997 67
Chaos or autocracy in a polity
4.1 114
A schematic representation of the election of 1860 in a
two-dimensional policy space 144
Illustration of Republican and Democrat economic policy
positions in a two-dimensional policy space 171
Illustration of ¬‚anking moves by Republican and Democrat
candidates in a two-dimensional policy space 179
Policy shifts by the Republicans and Democrats, 1860“1896
6.3 187
Policy shifts by Republicans and Democrats, 1896“1932
6.4 188
Estimated positions of presidential candidates, 1976“2000
6.5 193
A hierarchy of models
8.1 256


Four decades ago, William H. Riker published Federalism: Origin, Oper-
ation, Maintenance (1964). Riker™s motivation in writing this book came
from a question that he had raised in his earlier book, Democracy in the
United States (1953) about the origins of Federalism in the United States.
His argument was that only an outside threat could provide the motivation
to politicians to give up power by joining the Federal apparatus. His later
book, The Theory of Political Coalitions (1962), also attempted to answer
the question why plurality rule in the U.S. electoral system seemed to be
the reason for both minimal winning coalitions and the two-party system.
A further book, Positive Political Theory (with Peter Ordeshook, 1973),
attempted to develop the theory, available at that time, on two-party elec-
tions. The convergence result presented in that volume was later shown
to depend on unrealistic assumptions about the dimension of the space
of political decisions. Later, using the so-called “chaos theorems,” Riker
returned to the historical questions that had earlier intrigued him and sug-
gested that manipulability and contingency were features of democratic
systems (Riker, 1982, 1986, 1996).
Riker™s work provides the motivation for this book and for a compan-
ion volume (Scho¬eld and Sened, 2006). The formal theory of elections
and coalitions, together with empirical analyses of elections in Britain, the
United States, Israel, the Netherlands, and Italy, makes up that coauthored
volume. This present volume addresses many of the historical questions
raised by Riker, using as a conceptual basis the formal electoral model
presented in the companion book. This model is only brie¬‚y described in
the Introduction, and somewhat more extensively in Chapter 8. However,
the focus here is not on “social choice theory” itself, but rather on the
use of this theory as a heuristic device to better understand democratic


The essays included in this book were written over a number of years.
Obviously, I owe a great debt to William Riker. I considered it a great
honor to be the recipient of the Riker prize from Rochester University,
in acknowledgment perhaps of some of the earlier versions of this work.
Douglass North pressed me to apply the formal reasoning to more general
topics than elections. I hope he ¬nds the result of interest.
I received very helpful comments on the versions of these essays pre-
sented at the Hoover Institution, Harvard, Yale, MIT, and ICER, Turin.
The notions of quandary, of the Atlantic Constitution, and of a “factor
coalition” came about from discussion with Andy Rutten. The idea of “dy-
namic stability” developed out of long conversations with Gary Miller.
Chapter 6 is adapted from work (Miller and Scho¬eld, 2003) coauthored
with Gary. Chapter 5 is partly based on work with Kim Dixon, and ma-
terial used in that chapter was collected by Alexander Fak. Imke Kohler
kindly made available her research on Truman and McCloy, and this I
found helpful in the discussion in Chapter 7, on the founding of the World
Bank and of the Marshall plan. I am indebted to Alexander, Andy, Gary,
Imke, and Kim, and to my colleagues at Washington University in St.
Louis, particularly Andrew Martin, John Nachbar, Douglass North, John
Nye, Robert Parks, and Andrew Rehfeld. Iain McLean, who has written
extensively on Condorcet and on applying “rational choice theory” to
British politics, kindly listened to earlier versions of aspects of the argu-
ment. I thank James Alt, Keith Dowding, Robert Goodin, Manfred Holler,
Margaret Levi, Carole Pateman, Maurice Salles, and Albert Weale, who
were editors associated with earlier versions of these essays. The original
versions of the chapters were typed by Alexandra Shankster, and many
of the diagrams were drawn by her and by Diana Ivanov. Cherie Moore,
Robert Holahan, Ekaterina Rashkova, and Tsvetan Tsvetkov provided
further assistance.
I appreciate the support of the National Science Foundation (under
Grants SBR-98-18582 in 1999 and SES-0241732 in 2003) and of Wash-
ington University. The Weidenbaum Center at Washington University pro-
vided support for the completion of the manuscript. A year spent at Hum-
boldt University, Berlin, under the auspices of the Fulbright Foundation,
as distinguished professor of American Studies, gave me the opportunity
to develop the formal model that provides the theoretical foundation of
the current volume. Finally, Scott Parris, chief editor in economics at Cam-
bridge University Press, exercised great patience during the period of more
than a decade that it has taken to complete this work.
”Norman Scho¬eld, December 27, 2005, Saint Louis, Missouri

Constitutional Quandaries and Social Choice

1.1 introduction —
[I]t may be concluded that a pure democracy, by which I mean a society, consisting
of a small number of citizens, who assemble and administer the government in
person, can admit of no cure for the mischiefs of faction. A common passion or in-
terest will . . . be felt by a majority of the whole . . . and there is nothing to check
the inducements to sacri¬ce the weaker party. . . . Hence it is that such democ-
racies have ever been spectacles of turbulence and contention; have ever been
found incompatible with personal security, or the rights of property; and have in
general been as short in their lives, as they have been violent in their deaths.
A republic, by which I mean a government in which the scheme of representa-
tion takes place, opens a different prospect.
The two great points of difference between a democracy and republic, are
¬rst, the delegation of the government, in the latter, to a small number of citizens
elected by the rest; secondly, the greater number of citizens and the greater sphere
of country, over which the latter may be extended.
It may well happen that the public voice pronounced by the representatives of
the people, will be more consonant to the public good, than if pronounced by the
people themselves . . .
If the proportion of ¬t characters be not less in the large than in the small
republic, the former will present a greater option, and consequently a greater
probability of a ¬t choice.
As each representative will be chosen by a greater number of citizens in the
large than in the small republic, . . . the suffrages of the people . . . will be more
likely to centre on men who possess the most attractive merit. . . .
The other point of difference is, the greater number of citizens and extent
of territory which may be brought within the compass of republican, than of
democratic government; and it is this . . . which renders factious combinations
less to be dreaded in the former, than in the latter. . . . Extend the sphere, and you

This chapter is partly based on a talk presented at the Conference on Constitutional
and Scienti¬c Quandaries, at the International Center for Economic Research, Turin,
June, 2005.

Architects of Political Change

take in a greater variety of parties and interests; you make it less probable that a
majority of the whole will have a common motive to invade the rights of other
citizens . . .
Hence it clearly appears, that the same advantage, which a republic has over
a democracy . . . is enjoyed by a large over a small republic “ is enjoyed by the
union over the states composing it.
”James Madison, Federalist X, 1787

This book may be thought of as an extended interpretation of Madi-
son™s argument, using ideas from the work of Douglass North, William
Riker, and Mancur Olson, and aimed at developing “social choice” ap-
proaches to the evolution of society. As I suggest at the end of the book,
I see this research program as continuing the work of Madison™s contem-
poraries, Condorcet and Laplace.
North™s early work with Thomas (North and Thomas, 1970, 1973,
1977) attempted an economic explanation of the transition from
hunter/gatherer societies to agriculture. Later, he proposed a “neo-
classical theory of the state,” wherein contracts with “Leviathan” set
up a system of property rights and taxes (North, 1981). His later work
has focused on institutions and how they change as a result of incentives,
knowledge, and beliefs (North, 1990, 1994, 2005). One of his most per-
suasive pieces is his work with Weingast (North and Weingast, 1989) on
Britain™s Glorious Revolution in 1688 and how this transformed Britain™s
ability to manage debt, ¬ght wars (particularly with France), and develop
an empire.
Riker™s earliest work was on American Federalism, particularly the
logic underlying the need for Union in 1787 (Riker, 1953, 1964) and the
stability of parties as coalitions (Riker, 1962). After working for a number
of years on rational choice theory (Riker and Ordeshook, 1973), Riker
returned to American political history, to interpret key events in terms
of “heresthetic” (1982, 1984, 1986, 1996). Riker coined the word her-
esthetic from the greek ±ιρµ„ ικoζ , meaning “able to choose.” His book,
Liberalism against Populism (1982), argued that social choice theory im-
plied that populism, in the sense of existence of a “general will” was
vacuous. At best, all democracy could hope for was the liberal capacity
to remove autocrats.
Much of Olson™s work attempted to grapple with understanding how
some societies are successful and others much less so. In his early book,
Olson (1965) used the idea of the prisoner™s dilemma to suggest that coop-
eration may fail as individuals pursue their sel¬sh ends (through strikes,
revolutions, etc.) and indirectly constrain economic growth. Later, Olson

Constitutional Quandaries and Social Choice

(1982a, b) used this argument to provide a “declinist” explanation of why
stable democracies such as Britain and the United States appeared less vi-
tal (in the 1980s) than the newer democracies of the post“World War II
era (such as France, Germany, and Japan).
In this book I attempt to construct the beginnings of a theory of
democratic choice that I believe can be used as a heuristic device to
tie together these differing historical accounts. The basic underlying
framework is adapted from social choice theory, as I understand it, on
which I graft a “stochastic” model of elections. This model is an at-
tempt to extend the Condorcetian theme of electoral judgment. I shall
argue that its logic was the formal principle underlying Madison™s jus-
ti¬cation for the Republican scheme of representation that he made in
Federalist X. While this logic does not imply a general will in the sense
of Rousseau, it does suggest that Riker was overly pessimistic about the
nature of democracy. On the other hand, the social choice framework
suggests that democracy, indeed any polity, must face dif¬cult choices
over what I call chaos and autocracy. These dif¬cult choices are the
constitutional quandaries of the subtitle of this book. The historical
choices that I discuss often involve a leader or theorist”an architect of
change, either in the realm of politics or economics”who interprets or
frames the quandary troubling the society in a way that leads to its

1.2 balancing risk and chaos
Figure 1.1 is intended as a schematic representation of the formal results
of social choice theory. This ¬gure is replicated in Chapter 2, where a more
detailed discussion is provided of its interpretation. This ¬gure is intended
as a theoretical construct whose purpose is to suggest the relationship
between the many differing results of the theory. The vertical axis denotes
the “axis of chaos.” The theorems of social choice, from the earliest result
by Arrow (1951) to the later work on spatial voting theory (McKelvey
and Scho¬eld, 1986, 1987) imply that as factionalism increases, then utter
disorder can ensue. The term chaos was introduced to describe the possible
degree of disorder by analogy to mathematical chaos, which was used to
characterize a deterministic dynamical system, f, with the feature that
for almost any pair of outcomes x, y in the state space, X, there exists a
trajectory (see Li and Yorke, 1975)

x ’ f (x) ’ f 2 (x) ’ . . . f t (x) = y. (1.1)

Architects of Political Change

Lebanon, etc.


D=Democratic Stability
Weak Autocrat

Risk axis

B=Blocking veto Single veto group Oligarchy. A=Autocracy
“extreme risk avoidance” or collegium: The “extreme risk-taking”
The European Union The President Praesidium of N. Korea
and Congress of the USSR
the US

Figure 1.1. Chaos or autocracy in a polity.

For a voting rule, with speci¬ed voter preferences and an initial point
x, let f (x) be the set of alternatives that beat x. More generally, we can
think of the set, f (x), as the set of alternatives that can come about
from x, as determined by the social rule. The idea of social chaos is
that there are conditions under which, starting from almost any x, it
is possible to reach almost any possible outcome y = f t (x) by reiterating
the social rule. When the set Y that can be reached is large, in a for-
mal sense, then we can call Y the chaotic domain, Chaos(f ). In contrast,
we can identify the core or social equilibrium, Core(f ), as a singularity
of f, where y is in Core( f ) if and only if f (y) is empty. An element
y of Core( f ) may be an attractor of f , that is, a single outcome with
y = f t (x), which results from any x, after some number of iterations of the

Constitutional Quandaries and Social Choice

The social chaos theorem sets out the conditions for existence or oth-
erwise of the social equilibrium and for the situation where the chaotic
domain becomes almost the whole of X. For example, for any voting pro-
cedure, f , without a dictator, oligarchy, or collegium1 able to control, or
at least restrict, social choice, then, as the dimension of X increases, so
does the extent of voting chaos.2 For a general social rule, f, Scho¬eld
(1985a) formally de¬nes Chaos( f ) in terms of local cycles of the rule and
then shows that the union of Chaos( f ) and Core( f ) is non-empty. Thus,
if the rule has the property that Chaos( f ) is empty, then Core( f ) must
be non-empty. The theoretical problem for democratic theory is that if
Chaos( f ) for the social rule, f, is non-empty, then there may be no social
equilibrium. However, as discussed at length below, it may be the case that
democratic power resides in veto groups. Since a veto group is a collegium
in some limited domain of policy (namely a subset of X), then Chaos( f )
will be empty, and the social chaos theorem will not apply.
Note however that chaos, as I interpret it, is not just a property of
voting procedures. For a society where the social rule, f, is war rather
than voting, then I suggest that the chaotic domain, Chaos( f ), is likely
to be a large subset of X. For less violent methods, the chaotic domain
will typically depend on the heterogeneity of preferences in the society.
These results do not imply that democracies are necessarily chaotic, but
they do suggest that they can be.3 Throughout this book I shall use the
term chaos somewhat loosely, to refer to a social situation where there is

Chapter 2 gives more detail on this assertion. Roughly speaking, a voting rule is

characterized by a family of winning coalitions, D, say. A dictator is a single agent
who belongs to every winning coalition and is also winning. An oligarchy is a group
that belongs to every winning coalition and is itself winning, while a collegium is
a group of voters that belongs to every winning coalition in D, but need not be
Chapter 8 discusses the similar results on chaos in different domains. For social

choice, the chaos theorem is presented for a voting rule D, with speci¬ed voter
preferences. If D is collegial, in the sense that there is a collegium, then the core,
Cor e( f ), of the social rule, f, will generally exist. If D is non-collegial, then there
is an integer, w(D), called the “chaos dimension,” which characterizes D in the
following sense: If the dimension of the space, X, exceeds w(D), then the chaotic
domain, Chaos( f ), of the social rule, f, will be almost the whole of X. What I call
the social chaos theorem is the result of a long sequence of results by Plott (1967),
Kramer (1973), McKelvey (1976, 1979), Scho¬eld (1978, 1980, 1983), McKelvey
and Scho¬eld (1986, 1987), Banks (1995), Saari (1996), and Austen-Smith and Banks
(1998, 1999).
There has been much debate about the applicability of the social chaos result to

democratic theory. See, for example, Riker (1980, 1982, 1984, 1986), Hammond
and Miller (1987), and the essays in Ordeshook and Shepsle (1982).

Architects of Political Change

reason to believe that it is impossible to determine, even in general terms,
where the social trajectory will go.
When war, or intense and unrestrained con¬‚ict, dominates, then we
can expect chaos, as in Kosova, in Lebanon during the civil war, and
in Iraq at the present time. For a pessimist like Hobbes, it was obvious
that any society could fall into chaos, unless mitigating institutional de-
vices were constructed. The quote from Madison™s Federalist X suggests
that Madison certainly viewed direct democracy as subject to chaos. In-
deed, in his other writings, he used the phrase “the mutability of the law”
in commenting on the possible choices of the legislature. I take his com-
ments to mean that he considered that legislative bodies such as the House
and Senate were subject to a degree of disorder”possibly not the com-
plete disorder of chaos. It should be noted that the chaos theorem refers
to situations where individuals with speci¬c and heterogeneous prefer-
ences come together in either war or assembly and are in con¬‚ict over
an outcome. Thus a legislative assembly can be understood as a direct
democracy, and consequently can exhibit chaos, as suggested by the so-
cial choice results. Madison was very clear that representative democracy
involves the choice of a person, and he obviously believed that the voters
in the Republic could make a sound choice for the Chief Magistrate if
their judgments were not contaminated by preferences. One purpose of
this book is to explore the nature of social choice when it depends on
judgment rather than simply individual preferences.
The rationalizability of social choice may hold when an electorate
makes a speci¬c and limited choice, particularly in a binary situation
of yes or no. For example, the negative referenda votes in May and early
June 2005 in France and the Netherlands over the European Union (EU)
Constitution, while unexpected, cannot be seen as truly chaotic, because
they were one-off events. However, the frantic responses by the political
leaders of the EU may have elements of considerable disorder. At the same
time, there are many institutional devices within the EU that are designed
to control disorder.
The effect of these institutional “equilibrium” devices are well un-
derstood from the point of view of social choice theory. They all force
“rationality” by concentrating power in various ways. This is shown in
Figure 1.1 by the power characteristics of the decision rule, f , along the
risk axis. The work on social choice by Arrow (1951) considered a very
strong rationality axiom. Using this he showed that if this rationality prop-
erty is to be satis¬ed then the most extreme form of power concentration,
namely “dictatorship,” is a necessary condition in the case that individual

Constitutional Quandaries and Social Choice

preferences are unconstrained. Less extreme forms of power concentra-
tion include existence of an “oligarchy,” or “collegium,” or multiple veto
groups. Because a “dictator” can make any choice “he” deems ¬t, and
such a degree of power concentration almost never occurs in a polity, I
shall use the term autocrat for one who controls the levers of power of
the polity and has at least the ability to declare war without being con-
strained by some form of political veto. Clearly, Saddam Hussein was not
a dictator in the formal sense, but he certainly was an autocrat. Similarly,
I use the term oligarchy for a group who, if they agree, have “autocratic”
powers. A collegium is a group without full autocratic powers, but who
must all agree before the exercise of such power to pursue war or other
endeavors. A veto group is one with collegial power within a speci¬c re-
stricted domain of policy. Obviously there can be many veto groups in
any complex society.
Figure 1.1 presents my hypothesis that autocrats are likely to be ex-
treme risk takers. To some degree, this is an empirical assertion. One
only need make a list: Genghis Khan, Attila, Philip II of Spain, Napoleon,
Hitler, Stalin. Kennedy™s book, The Rise and Fall of the Great Powers
(Kennedy, 1987) argued that great nations tend to over-exert themselves
in the military realm, and through lack of ¬scal caution, bring about their
own demise. If we translate this argument by regarding the lack of ¬scal
caution as an element of risk taking more generally, then Kennedy™s logic
certainly seems valid for Philip II and Napoleon, and possibly for the lead-
ers of the USSR during the cold war. Kennedy also argued that it applied
to the United States in the post“World War situation. Table 2.1 in the next
chapter gives the relevant data on military spending for the United States
and USSR up until 1991 and suggests that there was little indication of
this risk-preferring military incaution by the United States until that date.
Whether the same inference is valid today is another question entirely.
On the risk axis, an autocrat is likely to be much more risk taking than
an oligarchy. I also suggest that an oligarchy will tend to be more risk
taking than a collegium. It is dif¬cult to precisely differentiate between an
oligarchy and a collegium. An example of an oligarchy is the Praesidium of
the Soviet Union. All members of the Praesidium must agree, in principle,
for a choice to be made, but if they do, then no decision-making body
can override them. A possible example of a collegium is the U.S. President
together with his cabinet, in a situation where the majority parties of the
House and Senate are in line with the president, and agree with his policy
initiatives. The more general situation, of course, is where the President
may veto Congress, and Congress may, in turn, counter his veto, with a

Architects of Political Change

supermajority. Thus the U.S. executive and Congress, regarded as a unit,
can be interpreted as having collegial power. Because the Congressional
counter-veto requires a supermajority, only very extreme situations can
lead to chaos as a result of presidential/congressional interaction. Note,
however, that the President and Congress together do not comprise an
oligarchy, since there are obvious policy domains in which Congress and
the President may concur but are blocked by state legislatures.
Because Congress may be factionalized, it can, as Madison expected,
exhibit what he called mutability”a degree of disorder or incoherence
in the laws that are passed. My understanding of the U.S. Constitution
is that it has a precise design to allow the presidential veto to overcome
congressional mutability. Of course, if there is a well-disciplined majority
party in Congress, then it can act as a collegium, thus ensuring stability
of some kind. However, it is certainly possible for Congress to become
factionalized, leading to the collapse of the collegium. One instance of
this was the presidential election of 1844 and its aftermath, as discussed
in Chapter 5. Because of the actions of Southern Democrats in block-
ing the candidacy of the New York Democratic, Martin Van Buren, the
Northern and Southern wings of the Democratic party split, and Northern
Democrats voted with Northern Whigs to suspend the gag rule, forbid-
ding discussion of the issue of slavery in the House. This factionalization
led eventually to a realignment of the party structure in the election of
Madison, of course, was concerned that the President would gain au-
tocratic power, and to avoid this, the Congressional counter-veto was
devised. However, even with the counter-veto, the President does have
some autocratic power, and I shall use the term weak autocrat to charac-
terize his power. It is evident that there is a tendency for U.S. presidents to
display the degree of risk preference that characterizes autocrats. I judge
that Congress will generally be risk-averse, which is why, I believe, power
to declare war resides in Congress. Even when Congress and the President
are aligned, then one would still expect the Presidential risk preference to
be muted by Congressional risk avoidance.
On the other hand, Congressional risk avoidance has the effect of de-
laying the resolution of fundamental constitutional quandaries. Typically,
a quandary can only be faced if there is a risk-taking leader capable of
forcing resolution. Without such a leader, the result can be the opposite
of chaos, namely “gridlock.”An illustration of this is given in Chapter 6,
in the discussion of the passage of Civil Rights legislation in 1957, while
Johnson was leader of the Senate (Caro, 2002). Decisions in the Senate

Constitutional Quandaries and Social Choice

could be blocked by the ¬libuster, and this could only be overcome by
“cloture.” This rule required “support from two-thirds of those present
and voting to impose cloture. This meant that a minority coalition of one-
third plus one of those present and voting could prevent a vote” (Rohde
and Shepsle, 2005). First, as leader of the Senate, and later as president
in 1964, Johnson was a risk taker able to persuade the collegium (of
one-third plus one) of Southern Democrats to lift its block.
Rohde and Shepsle (2005) go on to observe that “as a consequence
of a huge upsurge in ¬libusters in the decade following the civil rights
revolution, Rule 22 was amended in 1975, changing the requirement to
an absolute standard”sixty votes”to close debate [in the Senate].” Ob-
viously a group of forty-one senators has blocking power, and the change
in the rule has reduced the collegial veto power of such a minority.
As I discuss in Chapter 4, and as indicated by the aforementioned quo-
tation, Madison developed an argument in Federalist X that derived from
Condorcet. This led him to expect that the election of the president could
be assumed to be characterized by a high “probability of a ¬t choice.”
In constrained situations where we may assume that judgments predomi-
nate and voters evaluate the options in a clear-sighted fashion, then their
choice of Chief Magistrate may indeed be well formed in this way. For
this reason I locate the weak autocrat in Figure 1.1 at a position where the
risk taking of the autocrat is balanced by the risk avoidance of Congress,
as well as by judgment of the electorate. It would be natural to assume
that electoral judgment will generally be risk avoiding. However, there are
situations where a society feels threatened in some fashion and may ex-
hibit a degree of risk preference. It seems to me that the current situation
with regard to the United States and Iraq is unusual, precisely because the
electoral judgment has seemed to be much more risk preferring than is
common. As the true risks of the current siuation become apparent, this
risk posture may change.
It is important for my interpretation of electoral judgment that when
the “preferences” of the electorate are muted by judgments, then their
choice of the Chief Magistrate need not be subject to the chaos results.
Whether this is an entirely valid argument is a somewhat delicate matter.
Madison hoped that, because the election of the Chief Magistrate involved
the selection of a person, rather than an option (as in the passage of a law),
judgment rather than preference or interest would predominate. To argue
this formally requires analysis of an electoral model where judgment and
preference are both incorporated. In this book, I present the tentative
outline of such a model. It is of course entirely possible that beliefs or

Architects of Political Change

judgments in the electorate can be transformed in a chaotic fashion. Many
of the illustrations of belief transformation presented in this book suggest
that while the transformations are highly contingent, they are associated
with changes in what I call a core belief. A core, in social choice theory,
is an unbeaten alternative. By analogy, a core belief is a belief that has
general acceptance in the society.
As Figure 1.1 indicates, at the opposite end of the risk spectrum from
autocracy is the situation of extreme risk-avoiding blocking groups. Veto
groups are like collegia but with power in a limited domain. As indi-
cated earlier, social choice theory implies that veto groups induce sta-
bility, so the effect is the opposite of chaos. A good illustration is pro-
vided by the veto power that French farmers have over changes in the
EU Common Agricultural Policy (CAP). Obviously French farmers, to-
gether with their agrarian allies in Germany and the new members of
the EU, such as Poland, have a great deal to lose if the CAP is reorga-
nized. CAP is only one instance of a variety of protectionist, risk-averse
mechanisms that several veto groups have been allowed to deploy in the
expanding European polity. As Table 2.3 in Chapter 2 indicates, the con-
sequence seems to be that the core polities of France, Germany, and Italy in
Europe have stagnating economies. As of August 2005, the estimates of
growth by the Organisation for Economic Co-operation and Development
were less than 2 percent (1.8 percent in France, 1.1 percent in Germany,
and less than zero in Italy) with unemployment roughly 10 percent (about
8 percent in Italy, 10 percent in France and 11.6 percent in Germany). With
risk aversion comes high saving, low imports, high trade surplus, and an
appreciating euro. This will be increasingly exacerbated as the popula-
tion structure ages. These facts compare with growth and unemployment
of 3.6 percent and 5.0 percent respectively in the United States and 1.7
percent and 4.8 percent, respectively in Britain.
The “non” in France and “nee” in the Netherlands in May and June
2005 may have been induced by voter irritation at the apparent incom-
petence of the EU institutions, and it is reasonable to infer that these
referenda were based on electoral judgment. The problem is that, out-
side Britain, almost every group, except possibly teenagers and students,
has a veto over changes in crucial aspects of the social contract, partic-
ularly over unemployment and retirement bene¬ts. Without doubt, it is
much more comfortable to live in Europe rather than in the United States.
The degree of risk avoidance could be reduced, but only by institutional
mechanisms that are more risk preferring. The political institutions of
the EU (the Commission, Council of Ministers, European Parliament, the

Constitutional Quandaries and Social Choice

rotating President of the EU) all appear to be risk averse. The negative
referenda have induced some degree of disorder into the Council of Min-
isters, because the policy arena is now much more like a zero-sum game
than before, with ministers arguing over “rebates,”and agricultural subsi-
dies. Although the CAP budget has fallen over the last few years, from 70
percent of the EU budget to 40 percent, its effect is to distort agricultural
trade, harming farmers in less-developed countries. It is unclear at present
whether or how this EU quandary will be resolved. What is interesting is
that the Labour Party in Britain, though recently chosen by a proportion
of only 35 percent of the British electorate (much reduced from its support
in 2001) still controls 55 percent of parliamentary seats. Unlike a party
leader in the same situation in a polity based on proportional represen-
tation, Blair, as leader of the party, has the power to engage in a fairly
risky strategy against the other party leaders in the EU polities, directed
at transforming the CAP. This is consistent with the view that leaders of
polities based on proportional representation tend to be risk averse, while
leaders chosen through plurality electoral methods are more likely to be
risk seeking.
Social choice theory suggests that the EU quandary could be resolved by
the selection of a weak autocrat, such as a popularly elected EU president.
However, to satisfy Madison™s fears of autocracy, it would be necessary
for the electoral choice to be based on the judgments of voters rather than
their preferences. It is dif¬cult to see how a Europe-wide election could
have an information base that would be suf¬cient to support such a social
choice based on judgment.
With this preamble in mind, I shall attempt to formulate a Madisonian
model of election of the Chief Magistrate, President, or political leader,
that is in principle applicable to any democratic polity. The model will
involve both judgment and preference. Variations of the basic model can
then be interpreted in terms of a pure Condorcetian model of judgment,
or belief aggregation, as well as a pure, potentially chaotic model of pref-
erence aggregation.

1.3 preferences and judgments
For the formal electoral model I shall assume that individuals have pref-
erences that can be represented as functions on some “policy” space X.
This space characterizes both voter interests and possible eventualities. In
many of the examples, I argue that X conceptually derives from the soci-
etal deployment of the three factors of land, labor, and capital. Because

Architects of Political Change

the factors are bounded at any time, we may more conveniently regard
X as two dimensional. In empirical applications, for example, surveys
nearly always indicate that voters conceive of a con¬‚ict between the re-
quirements of capital and labor. What I term the labor axis is often derived
from beliefs about civil rights or religion. A third non-factor dimension
may involve attitudes toward war. In some cases the social attitudes with
regard to war are attributable to the desire for territorial expansion. Ob-
viously this notion of factor dimensions is a heuristic device, but it does
allow me to represent fundamental constitutional problems in a diagram-
matical form.
The interests or beliefs of the population or “electorate,” N (of size
n), are described by a set {xi } of “ideal points,” one for each “voter,” i =
1, . . . , n. An individual™s ideal point in the space, X, is used to describe or
represent that voter™s interests. In electoral models the ideal point can be
obtained from a survey. Whether we view xi as representing preferences
or beliefs is immaterial.
The set of options, S, of size s, is a set {z j }, each one being a point
in X.
In the situation of an election, each element of S is a declaration of
intended or proposed policy. There is one for each candidate, j.
While it is usual to conceive of each z j as simply a point, we can
easily allow z j to involve various possibilities associated with differing
probabilities of occurrence.4
In the simplest model, the “latent utility,” ui j of voter i for candidate
j has the form
ui j (xi , z j ) = »i j (xi ) ’ Ai j (xi , z j ) + θ T ·i . (1.2)

Here θ T ·i models the effect of the sociodemographic characteristics ·i
of voter i in making a political choice. That is, θ j is a k-vector specify-
ing how the various sociodemographic variables appear to in¬‚uence the
choice for option j. Thus θ T ·i is simply the in¬‚uence of i™s sociodemo-
graphic characteristics on the propensity to choose j.
The term Ai j (xi , z j ) is a way of representing the “preference disagree-
ment” between the interests of voter i and the jth option. In particular
Ai j (xi , z j ) may be some function of the distance between xi , the preferred

In principle we can construct a more general model where beliefs are probabilities

of outcomes, so the possible states are lotteries. This provides no technical problem,
since we can put an appropriate topology on this extended state space. The topology
I have in mind is a ¬ne topology taking into account differentiability. See Scho¬eld
(1996, 1999a,b) and Scho¬eld and Sened (2006).

Constitutional Quandaries and Social Choice

position (or ideal point) of voter i, and z j , the declared policy of candidate
j, according to some appropriate metric. In the standard electoral model,
where X is a policy space, it is assumed that Ai j (xi , z j ) = β xi ’ z j 2 is
the Euclidean quadratic loss (with β > 0) associated with the difference
between the two positions. We can, however, conceive of Ai j (xi , z j ) much
more generally. In the general case z j will involve a lottery across dif-
ferent possibilities, and different individuals could evaluate these various
possibilities in heterogeneous ways.
The model is stochastic because of the implicit assumption that
»i j (xi ) = » j (xi ) + µ j for j = 1, . . . , s. (1.3)
Here {µ j } is a set of possibly correlated disturbances and » j (xi ) is the
perception by a voter, i, with beliefs or interests, xi , of the “valence”
of the option presented by the candidate j. (See e.g., Ansolabahere and
Snyder, 2000; Groseclose, 2001). This valence is a way of modelling the
non-policy judgment by voter i of the quality of candidate j.
In the general model, the probability, Pr, that voter i chooses option j
ρi j = Pr[ui j (xi , z j ) > ui j (xi , zk) for all k = j]. (1.4)
Previous versions of this model have assumed that the valence com-
ponents {» j (xi )} are all zero and have usually asserted that all candidates
would converge to an “electoral mean” when they attempt to maximize
their expected vote shares. In the discussion of this model given in Chap-
ter 8, it is argued that, in the situation where the candidate valences dif-
fer, this mean voter theorem will only hold when a particular necessary
condition is satis¬ed. The condition depends on the valence differences
between candidates, on the coef¬cient β that speci¬es the importance of
policy, and on the variation of the distribution of voter ideal points, de-
noted as v2 . Further, the greater the stochastic variance (or uncertainty) of
the disturbances, then the easier it is for this condition to be satis¬ed. In
contrast, high electoral variation will tend to produce divergence of can-
didate positions. The upshot of this analysis is that empirical situations
can be found where convergence in candidate positions is very unlikely to
occur. Scho¬eld and Sened (2006) give examples from a number of poli-
ties based on proportional electoral systems where extreme divergence of
party positions is explained by this model.
We can apply this model in various ways. First, consider the pure
preference-based “non-stochastic” or deterministic case, µ j ’ 0, where
valence is zero. As noted earlier, a very extensive literature has shown

Architects of Political Change

that if decision making is binary (pitting options one against another)
and based on majority rule, or more generally on a non-collegial voting
mechanism, then “chaos” or disorder can ensue as long as the dimension
of X is suf¬ciently large. The formal results show that chaos can be pre-
vented by requiring that there be a collegium or veto player. Chapter 2
discusses this possibility in the context of an analysis of decision mak-
ing in Britain in the seventeenth and eighteenth centuries. The outcome
in this situation of a collegium, oligarchy, or autocrat may be a core or
institutional equilibrium. In the absence of a core, and if the dimension of
X is suf¬ciently low, then the set of probable outcomes will be restricted,
and I shall use the term the heart of the institution to refer to this set of
possible outcomes.
In the stochastic situation, with µ j = 0, it is necessary to focus on the
“beliefs” or judgments of the participants.
In the case that β ’ 0, then this is a situation of pure “belief aggrega-
tion.” Individuals will choose among the various options with probability
determined by the valence judgment that they have made. I suggest that
the ¬nal decision is often the consequence of what I call a belief cascade.
As more individuals decide that option zs , say, is superior, then other vot-
ers will in turn be swayed to form a judgment in favor of zs . I use the term
architect of change for an agent, s, who is able to trigger this change in
the social situation by providing a plausible argument for the option zs .
In the more general case with β = 0, the valences {» j (xi )} and therefore
the choices will depend on {xi }. It may be the case that different and
opposed belief cascades are generated in the population. For example,
in Chapter 5, I suggest that Lincoln™s arguments about the signi¬cance of
the Dred Scott decision generated opposing belief cascades in the northern
and southern electorates.
More generally, suppose that there is information available to some
subset M of the electorate which is consistent with the judgment
»s > »s’1 > . . . > »1 (1.5)
by the members of M. Then it will be the case that, for every voter i in
M, the subjective probabilities will be ranked
ρis > ρis’1 > . . . > ρi1 . (1.6)
It follows that the majority rule preference within the set M will choose
candidate s with option zs with greater probability than candidates s ’ 1,
s ’ 2, . . . , 3, 2, 1. If M is itself a majority under the electoral rule (or is a
winning coalition of more than half the electorate) then candidate s will

Constitutional Quandaries and Social Choice

win. When an alternative such as zs wins in this fashion, then it will be
sustained by a belief (or set of related beliefs) held by a winning coalition.
By analogy with the idea of a core, or unbeaten alternative, I use the term
core belief to refer to this common belief held by such a set of voters.
Condorcet in his Essai of 1785 argued essentially that a core belief
would tend to be a correct belief.5 A statement and proof of this result,
known as Condorcet™s Jury Theorem, is given as a short appendix to
Chapter 8. The demonstration is given for the case where only judgments
are involved, but it is obvious that the result holds in some weaker sense
when both interests and judgments are relevant, as long as interests do
not predominate. I argue in Chapter 4 that Madison had a version of this
argument in mind when he wrote about the “probability of a ¬t choice” for
the President in Federalist X. Of course, because interests may intrude in
the calculation of a ¬t choice, we cannot assert, as did Condorcet, that the
choice is necessarily superior.6 Notice also that the electoral rule (such as
deployed in the Electoral College) may de¬ne a coalition as winning even
though it does not comprise a majority. Recent literature has considered
extensions of the Jury Theorem when individuals have private information
and the decision problem is one of common value, so that all individuals
would agree over the correct choice if they had full information.7 The
societal decisions considered in this book have the characteristic that both
preferences and beliefs in the society are heterogeneous. I do not attempt
to present a full theory of such situations. Instead I hope to combine
elements of social choice theory and the theory of elections to present
a set of concepts that I feel can be useful in understanding democratic

Roughly speaking, the theorem asserts that, in a binary choice situation, the proba-

bility that a majority selects the true outcome will be greater than the probability that
a typical individual will select the truth. Rae (1969) and Scho¬eld (1972a, b) used a
version of the theorem to argue that majority rule would be “rationally”chosen by
an uncertain society as a constitutional rule. The theorem depends on the condition
of voter (pairwise) “independence,” which is a very strong assumption, and unlikely
to be satis¬ed. Recent work by Ladha (1992) and Ladha and Miller (1996) has at-
tempted to extend the theorem to include correlated choice. Empirical techniques
also allow for modelling correlated choices (Scho¬eld et al., 1998; Quinn, Martin,
and Whitford, 1999).
Scho¬eld (1972a) noticed the connection between Madison™s argument and a ver-

sion of the Jury theorem. Recently various authors have developed the theme of the
“wisdom of crowds” and how they might be swayed (Gladwell, 2000; Ball, 2004;
Surowiecki, 2004).
Recent work includes Austen-Smith and Banks (1996, 2005), Ladha (1996), Fedder-

son and Pesendorfer (1997), McLennan (1998), and Martinelli (2003).

Architects of Political Change

Thus the core belief underpins the selection of the option s, with the
greatest valence. I also use the notion of the heart of the Constitution to
refer to the con¬guration of beliefs that form the foundation for social
choice at each point in time. A constitutional quandary is a situation of
great uncertainty in the electorate. In the formal model, this is associated
with signi¬cant stochastic variance and relatively insigni¬cant valences.
According to the standard electoral model all candidates should converge
to the electoral center. Another way of expressing this is that the can-
didates should be risk averse. However, this assertion only holds true if
the electoral variation is relatively small. If electoral preferences are very
heterogeneous then candidates should rationally adopt very different po-
sitions.8 We might say, for a situation with very great uncertainty, that
these candidates for the attention of the electorate are prophets of chaos.
Sometimes, out of this cacophony of voices, there is one who can overcome
the barriers to clear perception and present a sensible way to interpret the
quandary. Naturally, this does not always happen. I suggest that a polity
will prosper when it is both open to the arguments of such an architect
of change and able to evaluate the opposing arguments. The evolution of
the Constitution is due to this continuous process of argument, shifting
beliefs, and changing valences.
This model is applied in Chapter 6 to suggest that the changing
valences of parties in the United States is due to the in¬‚uence of
activists on candidate positions. This accounts for what I call a structurally
stable dynamic, involving a slow rotation of party positions in what I
consider to be a fundamental two-dimensional policy space based on eco-
nomic factors and civil rights (Miller and Scho¬eld, 2003). There is some
evidence that a two-dimensional policy space is also relevant for Britain
(Scho¬eld, 2005a), though I suggest that the second dimension may be de-
rived from, or sustained by, beliefs that were appropriate during the period
of the British Empire. While my discussion largely focuses on Britain
and the United States, it is the larger question of the evolution of
what I call the Atlantic Constitution9 that forms the narrative of this

Scho¬eld and Sened (2006) show for example why the combination of stochastic and

electoral variation leads to many small, radical parties in Israel. Chapter 6 applies a
version of this model to presidential elections in the United States.
Bailyn (2005) argues that all the polities on the Atlantic littoral are connected through

a common history It seems natural to refer to the Atlantic Constitution as the set of
political, economic, and social beliefs common to these polities.

Constitutional Quandaries and Social Choice

1.4 the “institutional narrative” of the book
Here I shall brie¬‚y sketch the narative scheme that I shall use, based on
the ideas of social choice and on the notion of factor coalitions forming
in the policy space. Rogowski (1989) earlier made use of the assumption
from economic theory that there can be assumed to be three factors of
production: land, labor, and capital. External and internal features may
grant advantages to particular coalitions of these factor “interests.” For
example, the United States in the late 1700s could be characterized as
abundant in land, with both labor and capital relatively scarce. Principal
imports were manufactures, intensive in capital and skilled labor. Thus
protection in the form of tariffs would necessarily bene¬t capital and
“industrial labor.” In contrast, since land was abundant, this economic
interest, together with “agricultural labor,” would bene¬t from free trade.
Consequentially, the political con¬‚ict between the commercial Federalist
Party and the agrarian Jeffersonian Republicans, at the election of 1800,
can be interpreted in factor terms. However, some of the elements of the
controversy of that time can only be understood with respect to earlier
factor con¬‚icts in Britain, in the period from 1688.
North and Weingast (1989) had argued that the creation of the Bank
of England in 1694 provided a method of imposing credible commitment
on Parliament. The dilemma facing any government of that time was that
war had become more expensive than government revenue could cover.
Consequently, governments, or monarchs, became increasingly indebted.
Risk-preferring, or war-loving, monarchs, such as Philip II of Spain or
Louis XIV of France, were obliged to borrow. As their debt increased,
they were forced into repudiation, thus making it more dif¬cult in the fu-
ture to borrow. Since the Bank of England “managed” the debt in Britain
after 1694, there was an incentive for Parliament to accept the necessary
taxation and to avoid repudiation. However, it was clear after 1688 that
William III would pursue the war with France with great vigor and cost.
Contrary to the argument of North and Weingast, this escalating debt
could, in fact, force Parliament to repudiation. Until 1720, it was not ob-
vious how Parliament could be obliged to commit to ¬scal responsibility.
How this was done was through the brilliant strategy of Robert Walpole,
¬rst “prime” minister.
The fundamental problem was that the majority of members of both
Commons and Lords were of the landed interest. The obvious method of
funding government debt (which had risen to 36 million pounds sterling

Architects of Political Change

by 1713) was by a land tax. Indeed the land tax raised approximately
50 percent of revenue. War weariness had brought in a Tory government
in 1710, and the obvious disinclination of the Tory landed gentry to pay
increasing land taxes forced up the interest rate on long-term government
debt from 6 percent to 10 percent (Stasavage, 2002). In some desperation
the government created the South Sea Company in 1711. After Queen
Anne died in 1714, and the Hanoverian, George I, became sovereign, in-
creasing speculation in South Sea Company stock, and then the collapse
of the “bubble” in September 1720 almost bankrupted the government.
Walpole stabilized con¬dence in the company by a swap arrangement
with the Bank of England. In April 1721, Walpole, then Chancellor of the
Exchequer and First Lord of the Treasury, began his scheme to stabilize
government debt by instituting a complex system of customs and excise.
By restricting imports, mostly foodstuffs and land-intensive commodities,
this system had the effect of supporting the price of the scarce commod-
ity, land. From 1721 to 1740, these excise taxes and customs raised an
increasing share of government revenue. As Brewer (1988) has described,
the system required a sophisticated and skilled bureaucracy. The Walpole
device had many effects. Firstly, it ushered in a long period of Whig dom-
inance (at least until the 1800s). Protection of land remained in place
until the repeal of the Corn Laws in May 1846. As McLean (2000) has
described, the repeal was effected by Robert Peel, leader of the Tories (or
conservatives), together with Wellington in the Lords, against the inter-
ests of the majority of their party. Famine in Ireland made it obvious to
Peel and Wellington that unless food prices were lowered social unrest
could lead to civil strife. The Walpole “bargain” of 1721 essentially cre-
ated a compact between the “commercial” Whig interests and both Whig
and Tory “landed” interests. By supporting land prices, the bargain led
to increased investment in agriculture and (possibly counter intuitively)
the decline of the agricultural labor force. Increased food prices may have
reduced the real wage of industrial labor (Floud and McCloskey, 1994).
Although agricultural output increased in Britain, the population grew
even more rapidly, and Britain became increasingly dependent on food
imports, particularly from the United States.
Jefferson was well aware of the implications of the Walpole bargain.
His reading of the works of Henry St. John, Viscount Bolingbroke, led him
to believe that the land-capital bargain led to corruption, as well as the
¬lling of Parliament by placemen. In fact, Bolingbroke™s arguments against
Walpole were, to some degree, invalid, since the compact did make it
possible for Britain to manage its debt, ¬ght its wars, and create an empire.

Constitutional Quandaries and Social Choice

Bolingbroke™s logic was, however, valid for the United States. Hamilton™s
attempt in 1793 to re-create Walpole™s system would have necessitated
both a land tax and tariff protection. Since U.S. imports were primarily
manufactures, a tariff would protect the scarce factor (capital) associated
with these imports. In Jefferson™s view, this would have disadvantaged
the landed interest. By creating an agrarian coalition, essentially of the
Southern slave-owning landed interest and western free farmers, Jefferson
created a long-lasting compact under which the United States became the
food supplier for Britain. Just as the Walpole compact persisted until 1846,
so did Jefferson™s agrarian coalition survive until 1860. At that point, the
southern demand for expansion to the Paci¬c destroyed the Jeffersonian-
Jacksonian democracy.
The aftermath of the Civil War created a new coalition of commercial
interests and industrial labor, as represented by the presidential victory of
the Republican, McKinley, over the populist Democrat, William Jennings
Bryan in 1896. From this perspective, U.S. politics in the period 1896“
1960 can be interpreted in terms of a single factor dimension”capital”
since we can regard the interest of land to be generally in opposition to
capital. Thus, for the period from 1896 until the 1930s, the inclination
of Republicans for the preservation of a hard money or gold standard
rule was in opposition to the need for available credit in the agricultural
In the 1960s, agitation for greater civil rights brought the labor axis
into prominence. L. B. Johnson™s positioning on this axis contributed
to his great electoral victory in 1964, but also opened the way for the
Republican Party to adopt an increasingly conservative position on the
social dimension and gain political control in the southern states (Miller
and Scho¬eld, 2003).
In Britain, since 1846, all these factors have played a role at various
times. For example, McLean (2002) has observed that the success of the
Reform Bill, under the Conservative, Disraeli, in 1867, depended on be-
liefs about Empire. For industrial labor, “Empire” meant the opportunities
for emigration and a better life in the Dominions of America, Canada, and
South Africa. By using the rhetoric of “Empire,” the conservatives could
hope to appeal to working-class voters. In fact, such rhetoric was an im-
portant aspect of Thatcher™s electoral success in the 1980s. Indeed, recent
empirical analysis of electoral beliefs in Britain (Scho¬eld, 2005a) make
it clear that in addition to the usual economic (or “capital”) axis, it is
necessary to employ a second “social” axis. This axis incorporates “civil
rights,” but is also characterized by attitudes to the European Union. The

Architects of Political Change

responses of Conservative Party members of Parliament to a questionnaire
on this topic suggest that they are strongly opposed to the incorporation
of Britain within the European Union. In other words, political beliefs
that were founded on an economic rationale dating back more than one
hundred years are still relevant, in a somewhat different form, today.
This narrative suggests that preferences, or interests, on economic fac-
tors, or dimensions, play an important role in political decisions. However,
the manner in which these interests are transformed into beliefs is, to a
considerable degree, still a matter of conjecture. Indeed, how these be-
liefs take political expression seemingly depends on the perceptions and
strategies of political leaders such as Walpole, Peel, Disraeli, Franklin,
Washington, Madison, Jefferson, Lincoln, or Johnson.
It has been a long-standing controversy whether political economy is
best described by the concepts of “equilibrium” or “chaos” (Austen-Smith
and Banks, 1998, 1999). In his later work, after 1980, Riker saw chaos
as a fundamental property and focused on key “contingent” events in
United States political history, like the rati¬cation of the U.S. Constitution
in 1787“8, or the onset of the Civil War in 1860“1.
The brief description of British and U.S. political history offered here
suggests that neither equilibrium nor chaos is an accurate description of
social choice. Instead, there can be long periods during which the political
economic equilibrium is quite stable. However, equilibria can be destroyed
and dramatically transformed at key historical periods, as previously de-
scribed. Denzau and North (1994) have adopted ideas from evolutionary
theory in biology (Eldredge and Gould, 1972) and from the notion of
“informational cascades” (Bikhchandani, Hirschleifer, and Welch, 1992)
and proposed the concept of “punctuated social equilibrium.” As they
suggest, this idea is an analogue in the social realm of Kuhn™s notion
of scienti¬c revolution (Kuhn, 1962). At least intuitively, the notion of
punctuated social equilibrium would seem relevant to the puzzle of the
collapse of the Soviet Union that so intrigued Olson. Indeed, it is entirely
possible that the apparent relative decline of the United States and Britain
(which seemed so obvious to Olson in 1982 and Kennedy in 1987) has
been reversed, as the underlying political economic equilibrium has been
transformed in these two countries since 1980.
The chapters in this volume address these central questions, raised by
North, Olson, and Riker. Chapter 2 provides a more detailed overview
of the differing political economic equilibria in Britain and the United
States. In particular, the chapter discusses how Walpole™s “equilibrium”
or balancing of Whig and Tory interests set the scene for British imperial

Constitutional Quandaries and Social Choice

expansion, and gave ammunition to Jefferson in his campaign against
Hamilton. This chapter introduces some of the basic social choice ideas
used later. The “institutional” narratives of Chapters 3, 4, and 5 con-
sider the constitutional transformations in the United States in the key
periods of the Revolutionary War, 1776“1783; the formation of the two-
party system in 1787“1808; and the period 1857“1861 leading up to
the Civil War. More speci¬cally, Chapter 3 focuses on the nature of the
British threat to the Colonies, as expressed in the Quebec Act of 1774, and
the need for French aid before the Declaration of Independence. Chap-
ter 4 discusses the period from the rati¬cation of the Constitution in
1787 to the election of Jefferson in 1800. The chapter emphasizes the
in¬‚uence of Condorcet on both Madison and Jefferson. Chapter 5 de-
velops the idea of a constitutional quandary, prior to an institutional
transformation, and examines the election of Lincoln in 1860 in the con-
text of the Dred Scott decision by the Supreme Court in 1857. Chapter 6
argues that the theoretical accounts posing chaos against centrist equi-
librium miss the underlying feature of dynamic stability, in the United
States in particular. This chapter, based on Miller and Scho¬eld (2003),
suggests that political parties in the United States slowly cycle in the two-
dimensional policy space that was created in the period just prior to the
Civil War. In certain periods (such as 1896“1920) the principal axis is one
of land/capital. However, in the more general situation, which has held
from 1964 to the present, a second dimension”the social axis (a mirror
of the free labor/slave axis)”is also necessary for understanding political
The electoral model suggests that the kind of analysis performed by
political leaders such as Franklin, Madison, Jefferson, Lincoln, and John-
son transforms social uncertainty into the much more amenable aspect of
risk. Thus plurality, or majority decision making, allows such risk-taking
political agents to create solutions to dangerous political quandaries. It
is for this reason that I use the term architect of change to refer to such
agents of political transformation.
Chapter 7 addresses the problem facing Keynes prior to the creation
of the Bretton Woods Scheme after World War II. Economic uncertainty
could induce a retreat to totalitarianism, as in the 1930s. The chapter
suggests that Truman was the “architect of change” who facilitated the
creation of the “uncertainty-reducing” institutions that helped promote
world economic growth after 1945. In the 1970s, new quandaries over
the balance of economic logic and political expediency again became par-
ticularly relevant. The declinist arguments of many “prophets of chaos,”

Architects of Political Change

including Olson, Kennedy, and others, proved, however, to be unwar-
One of the themes of this book is the contrast between equilibrium
(whether economic or political) and chaos. Chapter 8 attempts to survey
the theoretical arguments for one side or the other. As Chapter 2 suggests,
equilibrium can always be attained by the imposition of a dictator or au-
tocrat. Indeed, as Arrow™s Theorem (Arrow, 1951) implies, a strong form
of equilibrium-inducing rationality can only be induced by dictatorship.
As I have indicated, social choice theory suggests that factional chaos may
occur if the underlying dimension of the political-economic world is suf¬-
ciently high and there are no institutional controls. Chapter 8 argues that
beliefs need to be incorporated into formal models of societal decision
making and provides an outline of how the stochastic electoral model
introduced earlier can be generalized. The chapter suggests that there is
a parallel between science, regarded as a truth-seeking device, and social
choice and gives examples of how the resolution of a quandary, whether
scienti¬c or social, may reside in the overcoming of a core belief that acts
as a barrier to resolution. An appendix to Chapter 8 gives a short proof
of the Condorcet Jury Theorem.
The narrative presented here suggests that when beliefs rather than sim-
ply preferences or interests are relevant, then representative democracy
can maintain a kind of structural stability balanced between chaos and
the rigidity of permanent equilibrium. While the evolution of the devel-
oped Atlantic polities may be structurally stable, it is also true that social
change in many parts of the world appears to be fundamentally chaotic.
A very brief Chapter 9 suggests that in a world of interconnection and
technological change, it is this likelihood which forms the constitutional
quandary that these Atlantic polities currently face.10

To put this formally, I suggest that typically the chaotic domain, Chaos( fdev ) is

relatively limited for the social choice rule in developed polities, though it may
change over time. In many less-developed polities the set Chaos( fnondev ) is large
with respect to possible outcomes, and it is the linkage between these two sets that
creates the quandary.

Power and Social Choice

2.1 introduction —
Mancur Olson™s book, The Rise and Decline of Nations (1982a), used
ideas from his earlier Logic of Collective Action (1965) to argue that en-
trenched interest groups in a polity could induce economic sclerosis, or
slow growth. These ideas seemed relevant to the perceived relative decline
of the United States and Britain during the 1970s. Five years later, Paul
Kennedy™s The Rise and Fall of the Great Powers (1987) proposed a more
general “declinist” argument, that a great power such as the United States
would engage in ¬scal irrationality through increasing military expendi-
ture, thus hastening its own decline. Neither of these two declinist argu-
ments seem applicable to the situation of the new millennium. Olson™s last
book, Power and Prosperity, published posthumously in 2000, attempted
a more general theoretical analysis of the necessary and suf¬cient causes
of prosperity and growth. For Olson, only “securely democratic societies”
could be conducive to long-lived individual rights to property and con-
tract, but democracy itself need not be suf¬cient for the protection of
rights. This chapter attempts to further develop Olson™s logic on the con-
nection between prosperity and liberty, by exploring insights derived from
Riker™s interpretation of U.S. Federalism (Riker, 1964), from the contribu-

. 1
( 11)