269
© 2016 The Authors. Juncture © 2016 IPPR Juncture \ Volume 22 \ ISSUE 4
In the conventional view, democracy begins with the voters.1
Ordinary
people have preferences about what their government should do. They
choose parties and leaders who will do those things, or they enact their
preferences directly in referendums. In either case, in this view, what
the majority wants becomes government policy – a highly attractive
prospect in light of the dreary historical experience that human beings
have generally endured with respect to governments. Democracy makes
the people the rulers, and legitimacy derives from their consent. In
Abraham Lincoln’s stirring words in his Gettysburg address, democratic
government is ‘of the people, by the people, and for the people’. Robert
Dahl emphasised ‘the continued responsiveness of the government to
the preferences of its citizens, considered as political equals’.2
That way
of thinking about democracy has passed into everyday wisdom in a
great many countries around the globe, including places such as Britain
and America whose political systems were originally constructed along
very different lines. It constitutes a folk theory of democracy, a set of
accessible, appealing ideas that assure people that they live under an
ethically defensible form of government that has their interests at heart.
Unfortunately, while the folk theory of democracy has flourished as an
ideal, its credibility has been severely undercut by a growing body of
scientific evidence that presents a different and considerably darker
view of contemporary democratic politics. The old frameworks will no
longer do.
What are those conventional notions of democracy that, we argue, have
outlived their time? There are many kinds of democratic theory, including
participatory and deliberative versions that underlie institutions like New
England town meetings. But among those that are applied to mass
democracies, two main theories predominate: one popular with broad
swatches of democratic society, and a second whose appeal is largely
confined to scholars specialising in the study of elections.
1	 This article is adapted from Achen CH and Bartels LM (2016) Democracy for Realists, Princeton University
Press, which will be published in April 2016.
2	 Dahl Robert A (1971) Polyarchy: Participation and Opposition, Yale University Press: 1.
Democracy for realists
Holding up a mirror to the electorate
Conventional accounts of how democracy works are flawed on a fundamental level,
argue Christopher Achen and Larry Bartels. By accounting for the ways social
identities shape voting behaviour, they present a new model that not only offers
greater intellectual clarity but could make genuine political change possible.
Essay
J22-4_text_160222.indd 269 22/02/2016 18:48:05
Juncture \ Volume 22 \ ISSUE 4 © 2016 The Authors. Juncture © 2016 IPPR
270
‘The People, Yes’: populism, and its undoing
The first model – the populist ideal of democracy – emphasises the role of
ordinary citizens in ‘determining the policies’ of democratic communities.3
This notion of popular sovereignty has inspired a good deal of sophisticated
academic thinking derived from Enlightenment concepts of human nature
and the political views of 19th-century British liberalism. It also lies behind
the Labour party’s emphasis on its annual conference, at which ordinary
citizens are meant to set policy for the party in parliament. In its less rarified
forms, populism has undergirded the folk theory of democracy celebrated
in much patriotic democratic rhetoric. As the homespun poet of democracy
Carl Sandburg proclaimed, ‘The People, Yes’.4
But how precisely should the people govern, according to the logic of the
populist theory? In one version, the public elect leaders who represent
their views. Joseph Schumpeter was a forceful critic of this doctrine, but
he summarised it well: the citizenry ‘decide[s] issues through the election
of individuals who are to assemble in order to carry out its will’.5
In the
other version of populism, the people rule through ‘direct democracy’,
choosing policies themselves via initiative and referendum procedures.
Both representative democracy and direct democracy loom large in
popular understanding of democratic self-government.
Alas, the scholarship of the last half-century has demonstrated that the
assumptions propping up both versions of populist democracy are highly
unrealistic. That evidence shows that the great majority of citizens pay
little attention to politics. At election time, they are swayed by how they
feel about ‘the nature of the times’, especially the current state of the
economy, and by political loyalties typically acquired in childhood. Those
feelings and loyalties, not the facts of political life and government policy,
are the primary drivers of political behavior. In consequence, in country
after country, people’s policy preferences often match up poorly with the
parties that they vote for.
When preferences and parties do match up, the cause is usually not the
voters’ careful choice of the ideologically correct party for themselves.
More often, the voters adopt the opinions of their parties and of other
relevant social groups – ethnic, racial, religious and occupational. Thus
election outcomes are well determined by powerful forces, but those
forces are not primarily the ones that current theories of democracy
believe should determine election results.
Much the same criticism applies to direct democracy. In America, states
like California make extensive use of referendums. Its voters have often
adopted naïve tax-cutting schemes that have eroded vital public services;
wholly unworkable business regulations intended to repeal the law of
supply and demand; and bald expressions of prejudice toward immigrants
and minority groups. The populist spirit also animates America’s heavy
reliance on primaries to nominate political candidates – a reliance that has
3	 Dahl RA (1998) On Democracy, Yale University Press: 37–38.
4	 Sandburg C (1936) The People, Yes, Harcourt Brace & Company.
5	 Schumpeter JA (1942) Capitalism, Socialism and Democracy, Harper & Brothers: 250.
Essay
“The scholarship
of the last halfcentury
has
demonstrated that
the assumptions
propping up both
versions of populist
democracy are
highly unrealistic.”
J22-4_text_160222.indd 270 22/02/2016 18:48:05
271
© 2016 The Authors. Juncture © 2016 IPPR Juncture \ Volume 22 \ ISSUE 4
frequently (and not just in the current presidential election cycle) offered
avenues to prominence for neophytes, demagogues and extremists.
In the welter of political claims and counterclaims, most people simply
lack the time and relevant experience to sort out difficult truths from
appealing dreams. That is no less true for Ivy League and Oxbridge dons
than it is for average citizens. The folk theory doctrine of populist control
relies for its persuasiveness on our exaggerated self-importance and our
systematic self-deception about human capacities and motivations in
the political realm.
Selecting a leader through the rearview mirror
The second contemporary model in defense of democracy is less widely
popular, though more persuasive to most political scientists. This model
focuses on elections as mechanisms for leadership selection. Dispensing
with populist notions, Schumpeter insisted that ‘democracy does not
mean and cannot mean that the people actually rule in any obvious
sense of the terms “people” and “rule.” Democracy means only that the
people have the opportunity of accepting or refusing the men who are to
rule them’.6
Or as the distinguished postwar Tory member of parliament
LS Amery put it a decade later, government should be ‘of the people, for
the people, with, but not by, the people’.7
Schumpeter gave little attention to the criteria by which voters would – or
should – choose among potential rulers. However, subsequent scholars
have fleshed out his account. The most influential model of democratic
selection in contemporary political science is the retrospective theory of
voting, which portrays ‘the electorate in its great, and perhaps principal,
role as an appraiser of past events, past performance, and past actions’.8
In this view, election outcomes hinge not on ideas, but on public approval
or disapproval of the actual performance of incumbent political leaders.
This model of democratic accountability appeals to skeptical scholars
because it puts much less pressure on the voters to have elaborate, wellinformed
policy views. Ordinary citizens are allowed to drive the automobile
of state simply by looking in the rearview mirror. Alas, this turns out to work
about as well in government as it would on the highway.
In our book, we consider cases in which leaders are clearly not responsible
for good or bad outcomes – droughts and floods, for example. It turns
out that voters routinely punish incumbents for bad weather, as Benjamin
Disraeli observed in the 19th century and as voluminous statistical evidence
has confirmed since then. We estimate, for example, that American voters’
retribution against the incumbent Democratic party for their states being too
dry or too wet cost presidential candidate Al Gore seven states in 2000 – far
more than his losing margin to George W Bush. We even study the political
effects of the 1916 shark attacks in New Jersey, on which the film Jaws and
its sequels were based. We find that in that case, too, the voters punished
6	 Ibid: 284–285.
7	 Amery LS (1947) Thoughts on the Constitution. London: Oxford University Press: 20–21.
8	 Key VO (1966) The Responsible Electorate: Rationality in Presidential Voting 1936–1960,
Harvard University Press: 61.
Essay
“The folk theory
doctrine of
populist control
relies upon our
exaggerated
self-importance
and our systematic
self-deception
about human
capacities and
motivations in the
political realm.”
J22-4_text_160222.indd 271 22/02/2016 18:48:05
Juncture \ Volume 22 \ ISSUE 4 © 2016 The Authors. Juncture © 2016 IPPR
272
the incumbent president, Woodrow Wilson, for changes in their welfare
that were clearly acts of God or nature. Citizens’ ability (or inclination) to
make sensible judgments about credit and blame is highly circumscribed.
Thus retrospection is often blind, and political accountability will be
greatly attenuated.
Voters are not very good at assessing responsibility for their pain,
but they often also assess the pain itself in a peculiar way. The
most prominent, politically significant and best-studied example of
retrospective accountability is economic voting in US presidential
elections. Scholars have repeatedly found that voters do indeed reward
or punish presidential incumbents for real income growth or decline.
However, the voters are myopic, focusing almost entirely on income
growth in the months just before each election. The performance of the
economy over the course of a president’s entire term – which provides
a better measure of changes in voters’ welfare as well as a presumably
more reliable benchmark of the incumbent’s competence – is almost
entirely discounted by voters when they go the polls.
A less flattering portrait of the voting public
We also studied voting behavior in the midst of the most severe
economic crisis in American history, the Great Depression of the 1930s.
Here, one might think, was an emergency that would focus voters’ minds
on momentous policy choices, shaping the course of government and
public policy for decades to come. The stakes were indeed momentous.
Yet we found that voters in the 1930s behaved much as they do at other
times – punishing their leaders at the polls when economic conditions
worsened and rewarding them when economic conditions improved,
with short memories and little apparent regard for ideology or policy.
The primary implication of our analyses of retrospective voting is that
election outcomes are mostly just erratic reflections of the current balance
of partisan loyalties in a given political system. The voters usually follow their
parties, with various deviations primarily due to factors the government does
not control. In a system with two major parties and competitive elections,
that means that the choice between the candidates is essentially a coin
toss. Thus, the picture that emerges is not ‘a portrait of citizens moved to
considered decision as they play their solemn role of making and unmaking
governments’.9
Rather, elections are capricious collective decisions based
on considerations that ought, from the viewpoint of the folk theory, to be
largely irrelevant – and which will in any case soon be forgotten by the voters
themselves. We conclude that the retrospective model of democracy simply
will not bear the normative weight that its proponents want to place on it.
Hence, this second conventional model of democracy, like the first, crumbles
upon empirical inspection. A dramatic departure from the folk theory is
needed to make sense of how democracy actually works. In particular, we
believe that a democratic theory worthy of serious social influence must
engage with the findings of modern social science.
9	 Ibid: 4.
Essay
“Election outcomes
are mostly just
erratic reflections
of the current
balance of partisan
loyalties: the voters
usually follow
their parties.
In a system with
two major parties
and competitive
elections, that
means that the
choice between
the candidates
is essentially a
coin toss.”
J22-4_text_160222.indd 272 22/02/2016 18:48:05
273
© 2016 The Authors. Juncture © 2016 IPPR Juncture \ Volume 22 \ ISSUE 4
A new model: the group theory
In the final part of our book, we point toward a quite different way
of thinking about democracy, which we refer to as the group theory.
This model portrays citizens as, first and foremost, members of social
groups, with (frequently numerous and complex) social identities and
group attachments figuring crucially in their political loyalties and
behaviour. We argue that this model provides a surer foundation for
democratic theory than either populism or retrospective voting.
Voters, even the most informed voters, typically make choices not
on the basis of policy preferences or ideology, but on the basis of
who they are – their social identities. In turn, those social identities
shape how they think, what they think, and where they belong in the
party system. But if voting behavior primarily reflects and reinforces
voters’ social loyalties, it is a mistake to suppose that elections result
in popular control of public policy. Thus, we make a sharp break with
conventional political thought.
We present evidence in support of this third model. First, as one among
many such examples in American history, we take up John F Kennedy’s
presidential candidacy in 1960, demonstrating the powerful role of
religious identities in shaping responses to his Catholicism. Second,
we explore the partisan realignment of the South over the past half
century. The demise of the Democratic ‘Solid South’ has typically been
interpreted as a response to the momentous partisan policy conflicts
of the civil rights era.10
Instead, we interpret the change as primarily
a matter of social identity, as white Southerners – even those with
moderate racial views – increasingly came to feel that the Democratic
party no longer belonged to people like them. When voters are taking
on a new partisanship, identities usually matter and ideologies usually
do not.
Now it may be thought that this unfamiliar and somewhat shocking
argument has to be wrong somewhere. After all, no matter what
the well-known defects of the folk theory, when one listens to
ordinary citizens they often sound quite coherent. Tories generally
espouse judgments and policy views supporting their preferred
candidates; so do Labour supporters. Maybe, somehow, all is well.
Alas, we show that citizens’ perceptions of parties’ policy stands
and their own policy views are significantly coloured by their party
preferences. Voters sound as though they are in the right party
because the party and its supporters have taught them what to
say in order to rationalise those preferences. Even on purely factual
questions with clear right answers, citizens are sometimes willing
to believe the opposite if it makes them feel better about their
partisanship and vote choices.
10	 Carmines EG and Stimson JA (1989) Issue Evolution: Race and the Transformation of American Politics,
Princeton University Press.
Essay
“Voters sound
as though they
are in the right
party because
the party and its
supporters have
taught them what
to say in order to
rationalise those
preferences.”
J22-4_text_160222.indd 273 22/02/2016 18:48:05
Juncture \ Volume 22 \ ISSUE 4 © 2016 The Authors. Juncture © 2016 IPPR
274
A case study: Deficit-deniers
We illustrate this phenomenon by examining beliefs about a highly salient
and significant political fact – the size of the American federal budget
deficit. The deficit had decreased by more than half during Bill Clinton’s
first term as president; yet most Republicans in a 1996 survey managed
to convince themselves that it had increased. Even many Democrats
and independents had too little real information to get the facts right, but
for Republicans the lack of information was compounded by a partisan
desire to see a Democratic administration in a negative light. Indeed,
moderately well-informed Republicans had less accurate beliefs than the
least informed; a modicum of information was sufficient to discern what
they should want to be true, but not enough to discern what was in fact
true. They sounded as though they were thinking and following their beliefs
in making their partisan choices, but no one should be fooled. Democrats
behaved in much the same way in denying Ronald Reagan’s success in
taming inflation during his time in office.
Thus, making use of the media and the Internet, our parties and our social
groups tell us what to think. The best-informed people often mimic the party
line better than the less politically engaged. We infer that group and partisan
loyalties, not policy preferences or ideologies, are fundamental in democratic
politics. Thus, a realistic theory of democracy must be built, not on the
French Enlightenment, on British liberalism, or on American progressivism,
with their devotion to human rationality and monadic individualism, but
instead on the insights of the critics of these traditions, who recognised that
human life is group life.
Get real: Beyond folk theories of democracy
In politics, powerful interest groups and identity groups abound, while many
citizens have only weak organisations speaking for them. Thus every modern
democracy is a long way from representing ‘the preferences of its citizens,
considered as political equals’. In particular, America is a democracy, but it is
not very democratic.
In our book, we try to face without flinching the logical consequences of
what democracy’s most thoughtful observers have long seen, and what
political scientists over the past several decades have demonstrated
in meticulous detail: democratic citizens – all of us – have to think
differently. All too often when we look at democracy we do so through
rose-tinted classes – glasses handed to us from the dead hands of
Enlightenment thinkers. ‘One person, one vote’ acquires a glossy sheen,
and we can see little else.
In consequence, we not only propose bad solutions – often enough, we
cannot even see the problems. The gross inequalities of political power in
contemporary democracies, on both the left and the right – are the most
obvious instances of this. The daunting challenge of altering a deeply
entrenched and powerfully defended status quo that embodies those
inequalities is often obscured by simplistic folk-theoretic faith that the
current system is directly responsive to its citizens, or that it could be
Essay
“Our parties and
our social groups
tell us what to
think. The bestinformed
people
often mimic
the party line
better than the
less politically
engaged. Group
and partisan
loyalties, not
policy preferences
or ideologies,
are fundamental
in democratic
politics.”
J22-4_text_160222.indd 274 22/02/2016 18:48:05
275
© 2016 The Authors. Juncture © 2016 IPPR Juncture \ Volume 22 \ ISSUE 4
made so if only a few bad apples were removed from office, or populist
ideology were applied even more rigorously. Too many democratic
reformers, even the successful ones, have squandered their energy on
misguided or quixotic ideas of direct popular control that have left the
powers that be comfortably undisturbed. Weakening the role of elected
officials and party leaders in nominations is an important example, as is
expanding the plebiscitary features of modern democracies.
What changes in society and government would be required to make
democracy more effective? We have no blueprint. We are certain only
that the old ways of thinking about democracy will not get us there. If the
formal procedures of electoral democracy were the main tools ordinary
citizens needed to produce responsive, accountable government, most
contemporary democracies would already be working pretty well. But
robust democracy is not primarily a matter of counting ballots. It is a
power struggle in which ballots provide one rather limited source of clout,
and ordinary citizens’ interests are likely to matter only insofar as the
organised groups representing those interests – including labour unions,
consumer groups, and ethnic, racial, professional, fraternal, religious,
civic, and neighborhood associations – are themselves politically engaged,
well-resourced, and internally accountable. As a practical matter, that
sort of healthy democratic struggle would require a degree of social and
economic equality than no contemporary democracy even approximates.
Just as a critical step towards democracy occurred when people lost faith
in the notion that the king had been anointed by God, we believe that
abandoning the folk theory of democracy is a prerequisite to both greater
intellectual clarity and genuine political change. A great deal of hard thinking
and organising will be required to develop more sophisticated notions of real
democracy, and to implement the deep transformations that will be required.
None of us, as individual citizens, can do it alone. But one vital implication of
our account is that more realistic party doctrines of democracy are central
to progress. Developing them will require robust debate in journals like this
one. The conventional wisdom about democracy too often leads us toward
policies that fail to improve people’s lives. A progressive agenda that can curb
illegitimate inequality and serve the interests of ordinary citizens will require us
to give up some cherished but badly outdated science, and to substitute for it
a vision with honest 21st-century intellectual credentials.
Christopher H Achen is the Roger Williams Straus professor of social
sciences and professor of politics at Princeton University. His books
include (as co-editor with Robert Thomson, Frans N Stokman and
Thomas König) The European Union Decides (Cambridge University
Press, 2006). Larry M Bartels holds the May Werthan Shayne chair of
public policy and social science at Vanderbilt University. His books
include Unequal Democracy: The Political Economy of the New
Gilded Age (Princeton University Press, 2008).
Essay
“A great deal of
hard thinking
and organising
will be required
to develop more
sophisticated
notions of real
democracy, and
to implement
the deep
transformations
that will be
required.”
J22-4_text_160222.indd 275 22/02/2016 18:48:05