peter mair

RULING THE VOID?

The Hollowing of Western Democracy

‘A semi-sovereign people’ was the term coined nearly half

a century ago to suggest that control over political decision-
making might lie beyond the reach of the ordinary citizen.1
Schattschneider’s thesis was a familiar theme in the sixties,
discussed by a variety of critical scholars in the so-called pluralist-elitist
debate. It seems to me to remain highly relevant—albeit now in a
stronger and less equivocal form. For today even semi-sovereignty
appears to be slipping away, and the citizenry are becoming effectively
non-sovereign. What we see emerging is a notion of democracy that is
being steadily stripped of its popular component—democracy without
a demos. In what follows I examine the twin processes of popular and
elite withdrawal from mass electoral politics with particular focus on the
transformation of political parties. I conclude with a discussion of the
implications of this process for Western liberal democracies.

When I first began to consider the notion of non-sovereignty, I asso-

ciated it primarily with indifference—towards politics and, indeed,
towards democracy. This had been one of the more neglected elements
in the literature on political trust and mistrust that emerged in the late
1990s.2 Arguably, however, the sense of hostility which some citizens
clearly felt towards their political leaders was less important than the
indifference with which many others viewed the political world more
generally. Of course, the dividing line between indifference and hostility
is not always very pronounced, and, as de Tocqueville once observed,
the loss of function can easily breed contempt for those who continue
to base their privileges on its exercise. But it seemed worth recognizing

new left review 42 nov dec 2006 25

 26 nlr 42

that politics and politicians might simply be deemed irrelevant by

many ordinary citizens.

By the late 1990s, however, popular indifference was being compounded

by a new rhetoric from the politicians themselves. A salient case was
Tony Blair, who claimed during his first term as Prime Minister that
‘I was never really in politics . . . I don’t feel myself a politician even
now’.3 For Blair, the role of ‘progressive’ politics was not to provide solu-
tions from above, by exercising the ‘directive hand’ of government, but to
bring together ‘dynamic markets’ and strong communities so as ‘to offer
synergy and opportunity’.4 In Blair’s ideal world, politics would eventu-
ally become redundant. As one of his close cabinet colleagues was later
to remark, ‘depoliticizing of key decision-making is a vital element in
bringing power closer to the people’.5 At one level, this was a simple pop-
ulist strategy—employing the rhetoric of ‘the people’ in order to suggest
that there had been a radical break with past styles of government. At
another, however, it gelled perfectly with the tenets of what were then
seen as newly emerging schools of ‘governance’—and with the idea
that ‘society is now sufficiently well organized through self-organizing
networks that any attempts on the part of government to intervene will
be ineffective and perhaps counterproductive’.6 In this perspective, gov-
ernment no longer seeks to wield power or even exercise authority. Its
relevance declines, while that of non-governmental institutions and prac-
tices increases. In Ulrich Beck’s terms, the dynamic moves from Politics,
with a capital ‘P’, to politics with a lower-case one, or to what he has
called ‘subpolitics’.7

Anti-political sentiments were also becoming more evident in the

policy-making literature of the late 1990s. In 1997, an influential article

1
E. E. Schattschneider, The Semi-Sovereign People: A Realist’s View of Democracy
in America, Chicago 1960. An earlier version of this argument was rehearsed in
‘Democracy Beyond Parties’, Center for the Study of Democracy, uc Irvine 2005,
available online: repositories.cdlib.org/csd.
2
See, for example, Susan Pharr and Robert Putnam, eds, Disaffected Democracies,
Princeton 2000; Pippa Norris, ed., Critical Citizens, Oxford 1999.
3
Blair’s Thousand Days, bbc 2, 30 January 2000. For a discussion of New Labour’s
approach see my ‘Partyless Democracy’, nlr 2, March–April 2000.
4
Tony Blair, ‘Third Way, Phase Two’, Prospect, March 2001.
5
Lord Falconer, quoted by Matthew Flinders and Jim Buller, ‘Depoliticisation,
Democracy and Arena-Shifting’, unpublished paper 2004.
6
Guy Peters, ‘Governance: a Garbage-Can Perspective’, isa, Vienna 2002.
7
For example, Ulrich Beck, Risk Society, London 1992, pp. 183–236.
 mair: Depoliticized Democracy 27

appeared in Foreign Affairs expressing the concern that government in

the us was becoming ‘too political’. Its author, Alan Blinder, a leading
economist and deputy head of the Federal Reserve, suggested extend-
ing the model of independent Central Banks to other key policy areas,
so that decisions on health, the welfare state and so on would be taken
by non-partisan experts.8 The role of politicians in policy-making would
be confined to those areas in which the judgement of experts would not
suffice to legitimize outcomes. Similar arguments were emerging in the
European context. In 1996, for example, Giandomenico Majone argued
that the role of expert decision-making in the policy-making process was
superior to that of political decision-making in that it could take better
account of long-term interests. Politicians, by definition, worked only in
the short-term; to allow decisions to be dominated by considerations of
the electoral cycle was to risk less optimal outcomes: ‘the segmentation
of the democratic process into relatively short time periods has seri-
ous negative consequences when the problems faced by society require
long-term solutions’. The solution, once again, was to delegate powers to
what Majone defined as non-majoritarian institutions, ‘which, by design,
are not directly accountable to voters or to their elected representa-
tives’.9 Experts were better able to deal with the technical complexities of
modern law-making, which often confused elected politicians. As tradi-
tional forms of state control were replaced by more complex regulatory
frameworks, specialist knowledge was likely to prove more effective than
political judgement.10 Here too, then, politics was becoming devalued.

By the late 1990s, in short, it seemed that neither citizens nor policy-makers
placed much value on the role of political or partisan decision-making. But
while the evidence pointed to a widespread indifference to politics and
politicians, it was less clear that it indicated indifference towards demo-
cracy as such. Indeed, if one looked at the debates about constitutional
reform during the late 1990s, as well as at the more theoretical literature,
the impression was of a large and burgeoning interest in democracy, with
more attention being paid to how democratic systems worked, and to what
they meant in reality, than probably at any stage in the previous twenty
or thirty years. Far from being treated with indifference, democracy had

8
‘Is Government too Political?’, Foreign Affairs, no. 6, vol. 76, 1997.
9
Giandomenico Majone, ‘Temporal Consistency and Policy Credibility’, European
University Institute, Working Paper 96/57, 1996.
10
Majone, ‘The Politics of Regulation and European Regulatory Institutions’, in
Jack Hayward and Anand Menon, eds, Governing Europe, Oxford 2003, p. 299.
 28 nlr 42

become a research priority within both empirical political science and

political theory. The catalogues of academic publishers brimmed with
new titles on the subject. Oxford University Press, for example, posted
as the lead publication in the 2002 political theory catalogue Robert
Goodin’s Reflective Democracy, closely followed by Iris Young’s Inclusion
and Democracy, John Dryzek’s Deliberative Democracy and Beyond, and
Henry Richardson’s Democratic Autonomy. Democracy was also becom-
ing more of an issue on the daily political agenda: debates on institutional
reform took shape in many Western polities; emphases on ‘participatory
governance’ began to emanate from the World Bank and other interna-
tional organizations. Discussions of the reform of the European Union
polity achieved a degree of salience that would have been almost unim-
aginable ten years before. By the end of the 1990s, democracy—whether
associative, deliberative or reflective; global, transnational or inclusive;
electoral, illiberal or even just Christian—had become a hot topic.

Beyond mass participation?

Which leads to a puzzle: as we shall see, there is now quite consistent

evidence of popular indifference to conventional politics and, more argu-
ably, to democracy; and yet, at an intellectual level, and sometimes at the
level of practical institutional reforms, there has been a distinct renewal
of interest in democracy (if not necessarily in politics as such). How do
we square these developments?

There are two possibilities. The first is that they are in fact related, and
that the growing intellectual and institutional interest in democracy, its
meanings and its renewal, is in part a response aimed at combating the
expanding scale of popular indifference. Making democracy relevant, in
other words, comes on to the agenda at the time when it otherwise risks
becoming irrelevant. But while the timing suggests that this may be the
case, the actual content of the discussion points to a different story. For,
far from seeking to encourage greater participation, or trying to make
democracy more meaningful for the ordinary citizen, many of the contri-
butions on institutional reforms or democratic theory seem to concur in
favouring options that actually discourage mass engagement. This can be
seen in the emphasis on stake-holder involvement rather than electoral
participation that is found in both ‘associative democracy’ and ‘participa-
tory governance’, and in the emphasis on the sort of exclusive debate that
is to be found in ‘deliberative’ and ‘reflective’ democracy. In neither case
 mair: Depoliticized Democracy 29

is there real scope afforded to conventional modalities of mass democracy.

The new stress on ‘output-oriented legitimacy’ in discussions of the
European Union polity, and the related idea that democracy in the eu
requires ‘solutions that are “beyond the state” and, perhaps, also beyond
the conventions of Western-style representative liberal democracy’, are
equally geared away from mass involvement.11 In other words, while
there may be concern with the problem of popular indifference, making
democracy more mass-user friendly does not seem to be the favoured
answer. For Philip Pettit, for example, who discusses the issue of demo-
cratic renewal in the context of deliberation and depoliticization, the issue
comes on to the agenda because ‘democracy is too important to be left to
the politicians, or even to the people voting in referendums.’ For Fareed
Zakaria, in his more popular account, renewal is necessary because ‘what
we need in politics today is not more democracy but less.’12

Hence the second possibility: the renewal of intellectual and institu-

tional interest in democracy is not intended to open up or reinvigorate
the practice as such, but rather to redefine democracy in such a way that
does not require any substantial emphasis on popular sovereignty, so
that it can cope more easily with the decline of popular involvement.
At the extreme, it is an attempt to redefine democracy in the absence
of the demos. Part of this process of redefinition lies in highlighting
the distinction between what has been called ‘constitutional democracy’
and what we might call ‘popular democracy’, a division that overlaps
with and echoes Robert Dahl’s earlier distinction between ‘Madisonian
democracy’ and ‘populistic democracy’.13 The constitutional compo-
nent emphasizes the need for checks and balances across institutions,
and entails government for the people; the popular component empha-
sizes the role of the ordinary citizen and mass participation, and entails
government by the people; the two elements co-exist and complement
one another within a ‘unified’ sense of democracy. Today, however, we

11
Jo Shaw, ‘Constitutional Settlements and the Citizen’, in Neunreither and Wiener,
eds, European Integration after Amsterdam, Oxford 2000, p. 291.
12
Philip Pettit, ‘Deliberative Democracy and the case for Depoliticising Government’,
University of nsw Law Journal, no. 58, 2001, § 46; Fareed Zakaria, The Future of
Freedom, New York 2003, p. 248.
13
Robert Dahl, A Preface to Democratic Theory, New Haven 1956. See also Yves Mény
and Yves Surel, eds, Democracies and the Populist Challenge, Basingstoke 2002;
Dahl, ‘The Past and Future of Democracy’, Occasional Paper Number 5, cspc,
Siena 1999; Shmuel Eisenstadt, Paradoxes of Democracy, Washington, dc 1999.
 30 nlr 42

see them being disaggregated and then contrasted, both in theory and
practice. Hence the recent emergence of notions of ‘illiberal’ or ‘elec-
toral’ democracy, and the attempt to distinguish those democracies that
combine free elections—popular democracy—with restrictions on rights
and the potential abuse of executive power.14 As many studies of ‘Third
Wave’ democracies in particular seem to indicate, popular and constitu-
tional elements are no longer necessarily bound together.

Not only is there a growing conceptual distinction between the two com-
ponents, but also a widening disparity in practice—one in which the
popular element is downgraded with respect to the constitutional. For
Zakaria, for example, it is the presence of the constitutional rather than
the popular component that is essential for the survival and well-being
of democracy. As he puts it: ‘For much of modern history, what charac-
terized governments in Europe and North America, and differentiated
them from those around the world, was not democracy but constitu-
tional liberalism. The “Western model” is best symbolized not by the
mass plebiscite but the impartial judge.’15 In this view it is not elections
as such that make for democracy, but rather the courts, in combina-
tion with other modes of non-electoral participation. With respect to
the developing countries, as much of the ‘good governance’ literature
implies, the formula is very clear: ngos + judges = democracy. That is,
while an emphasis on ‘civil society’ is acceptable, and while a reliance on
legal procedures is essential, elections as such need not be.16

A similar reasoning can be seen in many of the debates around consti-

tutional reform, where democracy is again often redefined in ways that
downgrade the importance of the popular pillar. As Michelle Everson
has noted in her discussion of Majone’s work, for example,

non-majoritarian thought . . . forcefully claims that its isolation of market

governance from political forces serves the goal of democracy by safeguard-
ing the democratically set goals of the polity from the predatory inclinations
of a transitory political elite.17

14
Larry Diamond, ‘Is the Third Wave Over?’, Journal of Democracy, no. 3, vol. 7,
1996.
15
Fareed Zakaria, ‘The Rise of Illiberal Democracy’, Foreign Affairs, vol. 76, no. 6,
1997, p. 27.
16
See also Amy Chua, World on Fire, New York 2003.
17
Michelle Everson, ‘Beyond the Bundesverfassungsgericht’, in Bankowski and
Scott, eds, The European Union and its Order, Oxford 2000, p. 106.
 mair: Depoliticized Democracy 31

In this case the opposition is unequivocal: in one corner, the goals of

the polity, objectively defined; in the other, the claims of a transitory—
because elected—and hence predatory elite. The one is sustained by the
networks of good governance, the other by the crude power and ambition
of electoral politics. Similarly, a recent review of new modes of delega-
tion underlines the growing importance of ‘procedural legitimacy’, which
‘relies on a process of decision making by nmis [non-majoritarian institu-
tions] being better than the insular, often secret, deliberations of cabinets
and executives.’ Here the benefits of transparency, legality and the provi-
sion of access to stakeholders are held up against the limits and distortions
induced by partisan politics, and are seen to lead to a process which can
offer ‘a fair and democratic substitute for electoral accountability.’ 18

Role of parties

What impact has this downgrading of the popular component of

democracy had upon political parties—and what role have the parties
themselves played in this process? Some twenty years before publish-
ing The Semi-Sovereign People, Schattschneider famously proposed that,
without parties, democracy was unthinkable. The phrase itself comes
from the opening paragraph of his Party Government, and is worth citing
in its context:

The rise of political parties is indubitably one of the principal distinguish-

ing marks of modern government. The parties, in fact, have played a major
role as makers of governments, more especially they have been the makers
of democratic government. It should be stated flatly at the outset that this
volume is devoted to the thesis that the political parties created democracy
and that modern democracy is unthinkable save in terms of parties. As
a matter of fact, the condition of the parties is the best possible evidence
of the nature of any regime. The most important distinction in modern
political philosophy, the distinction between democracy and dictatorship,
can be made best in terms of party politics. The parties are not therefore
merely appendages of modern government; they are in the centre of it and
play a determinative and creative role in it.19

As always in the writings of this period, democracy was both popular

and constitutional; it was the democracy of elections, mandates, popular

18
Mark Thatcher and Alec Sweet, ‘Theory and Practice of Delegation to Non-
Majoritarian Institutions’, West European Politics, no. 1, vol. 25, 2002, p. 19.
19
Elmer Schattschneider, Party Government, New York 1942, p. 1.
 32 nlr 42

accountability and representative government as well as of checks and

balances. This was the democracy that Schattschneider found unthink-
able except in terms of parties, and the sheer force of his conviction
has led to his proposition being cited by party scholars, especially in
their own defence, ever since. It is usually taken to mean that, since
the survival of democracy is guaranteed, this means that parties will
be, too. But we can also read it the other way around, to suggest that
the failure of parties might indeed imply the failure of democracy,
or at least of representative government.

Without parties (still following Schattschneider), we would then be left

with something that might still be called democracy, but which has been
redefined so as to downgrade or even exclude the popular component—
since it is this which depends so closely on party. Without parties, in
other words, we are left with a stripped-down version of constitutional
democracy, or with some system of modern governance that seeks to
combine ‘stakeholder participation’ with ‘problem-solving efficiency’.20
These are not unthinkable forms, but they are ones in which conven-
tional popular democracy plays little or no significant role, and in which
neither elections nor parties remain privileged. When democracy in
Schattschneider’s terms becomes unthinkable, in short, other modes of
democracy move to the fore. Hence the contemporary intellectual inter-
est in the theory of democratic renewal, and the more practical interest
in proposing new forms of institutional politics. All of these approaches
aim to find or define a notion of democracy that (a) works (b) is seen to
be legitimate and yet (c) no longer places at its centre the notion of popu-
lar control or electoral accountability.

Western trends

But in what sense are parties failing? First, as has been well attested,
parties are no longer managing to engage the ordinary citizen. Not only
are citizens voting in fewer numbers and with less sense of partisan
consistency, they are also increasingly reluctant to commit themselves
to parties, whether in terms of identification or membership. In this
sense, citizens are withdrawing from conventional political involvement.
Second, the party can no longer adequately serve as a base for the activi-
ties and status of its own leaders, who increasingly direct their ambitions

20
Beate Kohler-Koch, ‘European Government and System Integration’, European
Governance Papers, no. C-05-01, 2005.
 mair: Depoliticized Democracy 33

towards, and draw their resources from, external public institutions.

Parties may provide a necessary platform for political leaders, but this
increasingly serves as a sort of springboard from which to reach other
locations. In sum, parties are failing as a result of a mutual withdrawal,
whereby citizens retreat into private life or more specialized and often
ad hoc forms of representation, while party leaderships retreat into insti-
tutions, drawing their terms of reference ever more readily from their
roles as governors or public-office holders. The traditional world of party
democracy—as a zone of engagement in which citizens interacted with
their political leaders—is being evacuated.

On the question of citizens’ disengagement from conventional politics,

two qualifying remarks should be emphasized from the beginning.
First, this process of withdrawal is far from complete: indeed, in some
respects, but not all, it is not much more than a trickle, and I am there-
fore dealing with something that is ongoing rather than fully realized.
Second, although in some respects this is a familiar development, which
has already been dealt with in great detail in the scholarly literature as
well as in more popular commentary, the whole gamut of features of
this pervasive and wide-ranging process has not been brought together
in one overall and accessible assessment. Here I will attempt to do just
that, and to indicate the breadth and variety of modes of disengagement,
even if some of these are less substantial than others.

In fact, what we see here are two features that are not normally appli-
cable to cross-national changes at the level of mass politics. The first
of these is that virtually all of the separate trends that are treated here
point in the same direction. This in itself is very unusual. Analysts of
data relating to mass politics almost invariably expect to find mutually
opposing trends within the different streams of indicators, some point-
ing in one direction, some in another. Mass politics rarely moves en
bloc, as it were, but in this case it is precisely the uniformity of the
trends that is striking. Second, virtually all of these movements are
consistent across the advanced oecd democracies. This again is most
unusual. The normal expectation in comparative political research is
that, while particular trends may well be noted in some countries, they
are almost never universal. Some countries may shift together, but it is
only very rarely that all, or even most, shift in the same way and at the
same time. What we see now, however, is a much clearer indication of
cross-national convergence in the trends that matter. In other words,
 34 nlr 42

not only are these now pointing in the same direction, they are also
doing so almost everywhere. It is in this sense that the trends, though
often small, are very significant.

Electoral entropy

To begin with the most obvious and immediate indicator: the levels
of participation in national elections. Given what has been said about
citizen withdrawal, it is here that we might expect some of the most
striking trends to be identified; yet while expectations regarding the
possible decline in levels of electoral turnout have been current for
some years, they have often been found to have little backing in the
aggregate empirical data. Although long-term stability in levels of par-
ticipation has been followed by a slight decline, this is usually not seen
to be sharp enough to cause concern for the healthy functioning of
modern democratic life.

Is this a reasonable conclusion? On the face of it, and especially with

regard to the European data, the interpretation seems plausible.21 Thus
through each of the four decades from the 1950s to the 1980s, average
turnout levels in Western Europe scarcely altered, increasing margin-
ally from 84.3 per cent in the 1950s to 84.9 per cent in the 1960s, and
then falling slightly to 83.9 per cent in the 1970s and to 81.7 per cent in
the 1980s. This was essentially the steady-state period.22 That said, the
decline from the 1970s to the 1980s, while small, was remarkably con-
sistent across the long-established European democracies, with just three
(Belgium, Norway, the Netherlands) of the fifteen countries countering
an otherwise general trend. The decline may have been marginal when
looked at cross-nationally, but it was almost universal, and therefore
might well have justified a sense of concern.

But what is even more important to note is that this very marginal shift
accelerated in the 1990s, with average turnout across Western Europe
falling from 81.7 to 77.6 per cent in the last decade of the century. To be

21
For details see Mair, ‘In the Aggregate: Mass Electoral Behaviour in Western Europe,
1950–2000’, in Keman, ed., Comparative Democratic Politics, London 2002.
22
Pippa Norris, Democratic Phoenix, Cambridge 2002, pp. 54–5; Mark Franklin,
‘The Dynamics of Electoral Participation’, in LeDuc, Niemi and Norris, eds,
Comparing Democracies 2, London 2002.
 mair: Depoliticized Democracy 35

Figure 1. Turnout levels in Western Europe, 1950s–1990s (per cent)

100

90

80

70

60

50
1950s 1960s 1970s 1980s 1990s

sure, even at this level, which is the lowest recorded in any of the post-
war decades, turnout remains relatively high, with an average of slightly
more than three-quarters of national electorates casting a ballot in the
elections held during the 1990s, a figure that remains substantially
higher than that recorded in nationwide elections in the United States,
for example. Nevertheless, even allowing for the fact that this drop from
the 1980s to the 1990s is less than 5 per cent, it is striking to see the
overall European figure now dipping below the 80 per cent mark for
the first time in five decades. Moreover, there is a striking consistency
across countries, in that eleven of the fifteen democracies involved also
recorded their lowest ever decade averages in the 1990s. The exceptions
to this pattern again include Belgium, where the lowest turnout came in
the 1960s, and Denmark and Sweden, where it was in the 1950s. Even
in these three cases, however, it should be noted that the average level
of turnout in the 1990s was lower than in the 1980s. The fourth excep-
tion is the United Kingdom, which was unusual in recording its trough
in participation in the 1980s. Indeed, Britain is the only one of these
fifteen countries that had a marginally higher level of turnout in the
1990s than in the 1980s, although it plunged to an all-time low of just
59 per cent in 2001.
 36 nlr 42

This trend has continued into the 21st century. In addition to the uk, the
2001 elections in Italy and Norway, and the 2002 elections in Portugal,
France and Ireland were also marked by all-time low turnouts, as was
the 2000 election in Spain. Levels close to historic lows were recorded
in Greece in 2000, Austria in 2002, and Finland and Switzerland in
2003 (the last year included in this survey). In short, the trend towards
ever lower levels of participation has persisted. Both unidirectional and
pervasive, it offers a striking indicator of the growing enfeeblement of
the electoral process.

Before leaving these crude turnout figures, it is worth noting one other
telling feature. Indicators here are somewhat like those of climate
change: the shifts do not necessarily occur in great leaps or bounds, and
are not always linear. For these reasons, the importance of what is often
just a slight or uneven trend may be underestimated. Climatologists
have responded to this problem by laying less stress on the trends as
such, but instead noting patterns in the timing and frequency of the
peak values in their indicators. Thus, for example, evidence of global
warming is derived by noting that the warmest decade on record was
the most recent, the 1990s, while 1998 emerges as the warmest single
year, followed by 2001. Further evidence is adduced from the fact that
the eight warmest years on record have all occurred since 1990, even
though in that same period air temperatures were also recorded (e.g. in
1992, 1993 and 1994) which were little more than those reached in the
late 1970s.23 In other words, the pattern is evident, even if the trend is
not wholly uniform. This is also more or less true of turnout levels, and
indeed of many other indicators of mass political behaviour, and for this
reason the extent of change at this level is also often underestimated.
Although there is no undisturbed downward trend in levels of participa-
tion, for example, record lows now come with greater frequency, and in
a greater number of polities.

As can be seen from Table 1, which lists the three elections with the
lowest levels of turnout in each of the 15 long-established European
democracies, more than three-quarters of these 45 elections have taken
place since 1990. Not only do the 1990s hold the record for the low-
est turnout of any postwar decade in Western Europe, but within the

23
P. D. Jones and A. Moberg, ‘Hemispheric and large-scale surface air temperature
variations’, Journal of Climate, no. 16, 2003.
 mair: Depoliticized Democracy 37

Table 1: Low Turnout Elections

(a) Record low levels of turnout, 1950–2003 (b) Frequency of record low turnouts

Years of lowest turnout Period No. Percentage

Austria 1994, 1999, 2002 1950–59 6 13.3
Belgium 1968, 1974, 1999 1960–69 1 2.2
Denmark 1950, 1953 (i), 1953 (ii) 1970–79 2 4.4
Finland 1991, 1995, 1999 1980–89 2 4.4
France 1988, 1997, 2002 1990–2003 34 75.6
Germany 1990, 1994, 2002 All 45 100.0
Iceland 1995, 1999, 2003
Ireland 1992, 1997, 2002
Italy 1994, 1996, 2001
Luxembourg 1989, 1994, 1999
Netherlands 1994, 1998, 2002
Norway 1993, 1997, 2001
Sweden 1952, 1956, 1958
Switzerland 1995, 1999, 2003
uk 1970, 1997, 2001

great majority of West European democracies, most, and sometimes

even all of the individual national elections that are marked by record
low turnout have occurred since 1990. The two clearest exceptions are
Denmark and Sweden, where, seemingly for contingent reasons, the
lowest turnouts came in the 1950s. Beyond these cases, the only excep-
tions are one low-turnout election in the 1960s (in Belgium), two in
the 1970s (in Belgium and the uk) and two in the 1980s (in France
and Luxembourg). The remaining 34 cases all date from 1990 or later.
However small the overall shifts might be, they are nevertheless cluster-
ing together in a remarkable fashion. Indeed, this pattern also extends
to the newer southern European democracies: the three lowest levels
of turnout recorded in post-authoritarian Greece were in 1974, which
was the first free election, 1996 and 2000; in Portugal, the lowest levels
were recorded in 1995, 1999 and 2002; and in Spain in 1979, 1989 and
2000. Here, as in the longer-established democracies, the more recent
 38 nlr 42

the election, the higher the odds that it will record a trough in participa-
tion. There is no certainty; like the pattern evinced by climate change,
turnout also sometimes bucks the overall trend, even today. In the long
term, however, the overall direction of the change is unmistakable,
and offers the first strong indicator of increasing popular withdrawal
from conventional politics.24

Voter volatility

A second key aggregate indicator relates to those citizens who do partici-

pate, and measures the consistency of partisan preferences. Those who
continue to vote are clearly still engaged with conventional politics, even
if at the most minimal level.25 As popular involvement fades, however,
we may anticipate that even those who do take part will prove more vola-
tile in their preferences; not only the readiness to vote, but the sense of
partisan commitment will start to fade; choices are likely to prove more
susceptible to short-term factors. In practice, this means that election
outcomes will become less predictable; new parties and candidates may
prove more successful and traditional alignments come under pressure.
Inconsistency goes hand in hand with indifference.

As with turnout patterns, predictions of volatility growth have been

current for a number of years. Here too, however, the empirical record
at the aggregate level usually failed to meet expectations. While some
countries experienced a substantial increase in electoral flux through the
1970s and 1980s, others appeared to become more stable, resulting in
a relatively subdued level of aggregate change across Western Europe as
a whole.26 But here again we see the picture changing in the 1990s, the
peak decade for electoral volatility, with a score of 12.6 per cent, almost 4
points higher than that recorded in the 1970s and 1980s. Not too much
should be made of this; on a scale with a theoretical range from 0 to 100,
and decade averages that run in practice from 2.5 (1950s Switzerland) to
22.9 (1990s Italy), a mean value of 12.6 still reflects more (short-term)

24
This is also the conclusion drawn by Thomas Paterson in his valuable study of the
American case, The Vanishing Voter, New York 2002; see also Mair, ‘Voting Alone’,
European Political Science, no. 4, vol. 4, 2005, pp. 421–9, which incorporates parts
of the present discussion.
25
For example, Geraint Parry, George Moyser and Neil Day, Political Participation
and Democracy in Britain, Cambridge 1992.
26
Stefano Bartolini and Peter Mair, Identity, Competition and Electoral Availability,
Cambridge 1990.
 mair: Depoliticized Democracy 39

stability than change. On the other hand, the 1990s is the first of the five
postwar decades in which the overall mean of instability breaches the 10
per cent threshold; it is also the first to record such a major shift from
the previous mean.

The significance of the 1990s is underscored by individual national

experience. Thus, in all but four of the countries (the exceptions are
Denmark, France, Germany and Luxembourg), the 1990s constitute a
peak in volatility levels, which, in the majority of cases, easily exceeds
10 per cent. This confluence is also unprecedented, and again signals
that end-of-century patterns are markedly different from those of the
earlier postwar years.27

As with turnout data, there is no sign that these peaks are abating in the
21st century. In 2002, both Austria and the Netherlands experienced
record levels of aggregate instability, as did Italy in 2001. France, Norway
and Sweden also saw remarkably high levels of volatility in these years,
although no absolute records were broken. More generally, as can be
seen in Table 2 (overleaf), a clear majority of the most unstable national
elections since 1950 have occurred after 1990. In this case the pattern is
not so one-sided: volatility data inevitably prove more erratic than turn-
out data, being responsive to political crises as well as to institutional
and social-structural change. Nevertheless, the period since 1990 seems
exceptional: not only do more than half the record highs fall in this
period, but no other decade comes close to this clustering. With the mar-
ginal exceptions of Denmark and Luxembourg, it seems that the more
recent the election, the less likely it is to yield a predictable outcome.

Since 1990, then, it seems that ever fewer voters are ready to participate
in elections, although turnout levels still remain reasonably high; while
among those who do participate, there is a greater likelihood that they will
switch preferences from one election to the next.28 The exceptions have
been Luxembourg, which had very low turnout but only moderate volatility;

27
The 1990s rise in volatility outside Western Europe—Japan, Mexico, India, for
example—though noteworthy, lies outside the scope of this paper.
28
This counters an earlier observation based on us data by Lance Bennett, who
suggested that even though conventional political participation may be in decline,
‘those who continue to participate in traditional politics exhibit stability and sub-
stance in electoral choice, opinion formation and policy deliberation.’ Bennett, ‘The
Uncivic Culture’, ps: Political Science and Politics, December 1998, p. 745.
 40 nlr 42

Table 2: High Volatility Elections

(a) Record Levels of Volatility, 1950–2003 (b) Frequency of record high volatility

Years of Highest Volatility Period No. Percentage

Austria 1990, 1994, 2002 1950–59 5 11.1
Belgium 1965, 1981, 2003 1960–69 2 4.4
Denmark 1973, 1975, 1977 1970–79 7 15.6
Finland 1970, 1991, 1995 1980–89 6 13.3
France 1955, 1958, 2002 1990–03 25 55.6
Germany 1953, 1961, 1990 All 45 100.0
Iceland 1978, 1991, 1999
Ireland 1951, 1987, 1992
Italy 1992, 1994, 2001
Luxembourg 1954, 1984, 1989
Netherlands 1994, 1998, 2002
Norway 1989, 1997, 2001
Sweden 1991, 1998, 2002
Switzerland 1987, 1991, 1999
uk 1974 (i), 1978, 1997

Sweden, which recorded high volatility but not exceptionally low turnout;
and Denmark, which proved extreme on neither indicator. Beyond these
cases, the evidence of unusual patterns since 1990 is both striking and
consistent. Across Western Europe, electorates are not only voting less,
but they are also slackening in terms of partisan commitment.

Partisan attachment

This is also the message from survey data, the evidence of individual
experiences collected by election studies and commercial polling projects,
which now corresponds closely to the aggregates on turnout and volatil-
ity. Many of the former have been collated by Dalton and Wattenberg in
their comprehensive Parties without Partisans, and again both the con-
sistency and the ubiquity are striking. One key indicator is the degree
to which individual voters feel a sense of belonging or commitment to
 mair: Depoliticized Democracy 41

particular political parties. In seventeen of the nineteen countries for

which relevant data are available (the two exceptions are Belgium and
Denmark) the percentage of voters claiming a sense of identification with
parties has fallen over the past two decades or so. Even more strikingly,
the smaller numbers of voters who report a strong sense of belonging or
identification has also decidedly fallen, and this time in every single one
of the countries concerned. As Dalton notes, it is not just the scale of the
decline that is important here, but more the fact that it occurs in each of
the cases for which figures are available. ‘The similarity in trends for so
many nations forces us to look beyond specific and idiosyncratic expla-
nations . . . For public opinion trends to be so consistent across so many
nations, something broader and deeper must be occurring’.29

Split-ticket voting, whereby voters opt for one party in one electoral
arena, and a different one in another, is also on the rise across all
those cases where it has been measured over time (Australia, Canada,
Germany, Sweden and the United States). A committed and engaged
voter, with a strong partisan loyalty, will undoubtedly vote for the same
party regardless of the arena involved—for example, voting Democrat in
us Presidential and Congressional elections, as well as local state and
county ones. Lesser partisan commitment is more likely to be associated
with a greater willingness to split the ticket. Voters are also less ready
or less able to tell pollsters how they will vote. Here too, with a single
Danish exception, almost every election study has reported an increase
in the proportion of voters who decide how to vote during the campaign
or shortly before polling day. Again, ‘the trend is clear: contemporary
voters are less likely to enter elections with standing partisan predis-
positions’. Hardly surprising then, that these voters are also unlikely
to engage in more demanding campaign activities, such as attending
political meetings, working for a party, persuading others to vote for a
particular candidate or donating money. On almost all of these meas-
ures, and in almost all countries for which data are available, the survey
evidence once again points to decline: voters are less willing to partici-
pate; for many, at least as far as conventional politics is concerned, it is
enough to be simply spectators.30

29
Russell Dalton, ‘The Decline of Party Identification’, in Dalton and Wattenberg,
eds, Parties without Partisans, Oxford 2000, p. 29.
30
Dalton, McAllister and Wattenberg, ‘The Consequences of Partisan Dealignment’,
in Dalton and Wattenberg, Parties without Partisans, pp. 49, 58.
 42 nlr 42

Voters are also less willing to take on the obligations associated with
party membership. Again, it is striking to note not only the sheer decline
in the number of party members, but also the pervasiveness of the phe-
nomenon across all long-established democracies. Although the pattern
here is more pronounced than in turnout or electoral instability, until
the 1980s the evidence of decline tended to be somewhat equivocal. The
first major study based on aggregate—often official party—data in 1992
found that, although the party membership ratio had fallen in most of
the relevant European polities (the only exceptions were Belgium and
West Germany), absolute levels of membership had often held up.31 It
offered little support for the idea that these countries were experiencing
‘a spreading disillusionment with partisan politics’.32

By the end of the 1990s, however, the patterns in the aggregate data had
become unequivocal. As Table 3 reveals, the ratio of party membership
to the electorate across a range of West European democracies had fallen
markedly between 1980 and the end of the 1990s.33 In 1980, an aver-
age of 9.8 per cent of the electorates were party members; by the end of
the 1990s, this had fallen to just 5.7 per cent. Still more strikingly, for
the ten European democracies in which it is possible to trace reliable
membership figures from 1960, the average membership ratio was 14
per cent; in a majority—six of the ten—more than one in every ten elig-
ible voters were members of political parties. At the end of the 1990s,
by contrast, there were twenty democracies for which it was possible
to find reliable membership data. Across all twenty, the average mem-
bership ratio was just 5 per cent, and only one—Austria—recorded a
ratio that exceeded 10 per cent.34

31
See Katz, Mair et al., ‘The Membership of Political Parties in European
Democracies, 1960–90’, European Journal of Political Research, vol. 22, no. 3, 1992,
pp. 329–45.
32
Norris, Democratic Phoenix, pp. 134, 135.
33
The table is adapted from data in Mair and Ingrid van Biezen, ‘Party Membership
in Twenty European Democracies, 1980–2000’, Party Politics, no. 1, vol. 7, 2001,
where they are discussed in more detail; see also Susan Scarrow, ‘Parties without
Members’ in Dalton and Wattenberg, Parties without Partisans, pp. 86–95.
34
The pattern is comparable in the advanced democracies outside Europe. In
Australia in 1967 there were 251,000 members, 4.1 per cent of the electorate; by
1997, the number had fallen to 231,000, just 1.9 per cent of a much larger electorate.
Canada went from 462,000 members in 1987 to 372,000 in 1994: 2.6 to 1.9 per
cent. In New Zealand, the decline was from 272,000 in 1981, or 12.5 per cent, then
the peak of a growing wave, to 133,000 in 1999, or 4.8 per cent. Webb et al., Political
Parties in Advanced Industrial Democracies, Oxford 2002, pp. 355, 389–90, 416–9.
 mair: Depoliticized Democracy 43

Table 3: Change in party membership, 1980–2000

Party membership as
Change in Change as %
% of electorate
numbers of party of original
members membership
Country Period Start of period End of period

France 1978–99 5.05 1.57 –1,122,128 –64.59

Italy 1980–98 9.66 4.05 –2,091,887 –51.54
uk 1980–98 4.12 1.92 –853,156 –50.39
Norway 1980–97 15.35 7.31 –218,891 –47.49
Finland 1980–98 15.74 9.65 –206,646 –34.03
Netherlands 1980–2000 4.29 2.51 –136,459 –31.67
Austria 1980–99 28.48 17.66 –446,209 –30.21
Switzerland 1977–97 10.66 6.38 –118,800 –28.85
Sweden 1980–98 8.41 5.54 –142,533 –28.05
Denmark 1980–98 7.30 5.14 –70,385 –25.52
Ireland 1980–98 5.00 3.14 –27,856 –24.47
Belgium 1980–99 8.97 6.55 –136,382 –22.10
Germany 1980–99 4.52 2.93 –174,967 –8.95
Portugal 1980–2000 4.28 3.99 50,381 17.01
Greece 1980–98 3.19 6.77 375,000 166.67
Spain 1980–2000 1.20 3.42 808,705 250.73

This was reinforced by the fall-off in the absolute numbers of party mem-
bers, in marked contrast to the pattern noted in 1992. In every one of the
long-established democracies the number of party members had fallen,
sometimes by as much as 50 per cent of 1980 levels. In no country had
there been an increase. This was exit on a grand scale—both in terms
of reach and direction. Throughout the old democracies, as the analysis
concluded, parties were simply haemorrhaging members.35

What conclusions can be drawn from this brief review of the evidence?
Clearly, it supports the contention that citizens are disengaging from
the arena of conventional politics. Even when they vote—and this is
less often than before, or in smaller proportions—their preferences

35
Mair and Biezen, ‘Party Membership in Twenty European Democracies’.
 44 nlr 42

are determined closer to polling day and are less guided by partisan
attachments. Electorates in this sense are becoming progressively de-
structured, affording more scope to the media to set the agenda, and
requiring a much greater campaign effort from parties and candidates.
What we see here, in short, is a form of voting behaviour that is increas-
ingly contingent. Much of this change has only become really apparent
since the end of the 1980s.

To be sure, we are dealing with some quite small pieces of evidence,

and the changes noted are sometimes relatively marginal: this is in
some instances a trickle rather than a flood. But when all the disparate
pieces of evidence are summed together, they offer a clear indication
of a marked shift in the prevailing patterns of mass politics, consistent
not only in terms of its focus—all of these indicators now point in a
common direction—but across the different European polities. The con-
clusion is unambiguous: all over Western Europe, and in all likelihood
all over the advanced democracies, citizens are heading for the exits of
the national political arena.

In early 2002, Anthony Giddens pointed to the watershed that had been
passed in mass media entertainment through the growing popularity of
reality tv. ‘Previously television was something that reflected an external
world which people watched. Now television is much more a medium
in which you can participate.’36 In conventional politics, by contrast, the
shift has been the other way around. Previously, and probably through to
at least the 1970s, conventional politics was seen to belong to the citizen,
and something in which the citizen could, and often did, participate.
Now, it has become part of an external world which people watch from
outside: a world of political leaders, separate from that of the citizenry.
It is the transformation of party democracy into ‘audience democracy’.37
Whether the increasing disengagement of voters is responsible for the
emergence of this new mode of politics, or whether it is an emerging
form of politics that is encouraging voter withdrawal is, at least for now,
a moot point. What is beyond dispute is that each feeds the other. As citi-
zens exit the national political arena, they inevitably weaken the major
actors who survive there—the parties. And this, in turn, is part of, and

36
Interview with Henk Jansen in Facta, no. 1, vol. 11, February 2003, p. 4
(my translation).
37
Bernard Manin, The Principles of Representative Government, Cambridge 1997,
pp. 218–35.
 mair: Depoliticized Democracy 45

promotes, audience democracy, or, in another formulation, ‘video poli-

tics’; for these grow stronger when parties are weak. Strong parties are
difficult to sustain when politics turns into a spectator sport.

From civil society to state

Given how difficult it has become to engage citizens in the conven-

tional political arena, it might be expected that political leaders would
devote considerable effort to keeping politics alive and meaningful.
As noted above, there has rarely been such widespread discussion of
institutional reform. But beneath the beating of official breasts and the
apparent distress at the hollowing out of mass politics, in practice there
exists a clear tendency for political elites to match citizen disengage-
ment with a withdrawal of their own. Just as voters retreat to their own
particularized spheres of interest, so too have political and party lead-
ers withdrawn into the closed world of the governing institutions. Both
sides are cutting loose.

Contemporary changes in the form of party politics may be speci-

fied under two broad headings: the location of the parties, and their
political identity. As far as location is concerned, the past few decades
have witnessed a gradual but inexorable withdrawal of the party leader-
ships from the realm of civil society into that of government and the state.
The same period has also seen the steady erosion of the parties’ political
identities, and the blurring of inter-party boundaries. Together, these
parallel developments have led to a situation in which each party tends to
become more distant from the voters that it purports to represent, while
at the same time becoming more closely associated with the various
protagonists against whom it purports to compete. Party–voter distances
have been stretched, while party–party differences have lessened; both
processes combine to reinforce a growing popular indifference and dis-
trust of parties, and of political institutions more generally.

If we conceive of the role and location of parties within a democratic

polity as standing somewhere on a spectrum between society and the
state, then we can suggest that they have shifted along this contin-
uum from a position in which they were primarily defined as social
actors—as in the classic mass-party model—to one where they might
now be reasonably defined as state actors. As we have seen, the strength
of electoral identification with parties is now almost universally in
 46 nlr 42

decline. At the same time, the former privileges of membership have

also tended to disappear, as party leaders look beyond their shrinking
membership to the electorate at large. The voice of the ordinary voter
is seen to be at least as relevant to the party organization as that of the
active party member, and the views of focus groups often count more
than those of conference delegates.38

In addition, a sense of dispersal and atomization marks the broader organ-

izational environment within which the traditional parties used to nest.
As workers’ parties, or as religious parties, the mass parties in Europe
rarely stood on their own, but constituted the core element within a wider
and more complex network of trade unions, churches, business asso-
ciations, mutual societies and social clubs. These helped to root the old
mass parties in society and to stabilize and distinguish their electorates.
Over the past thirty years, however, these broader networks have largely
disintegrated. In part, this is because of a weakening of the sister organi-
zations themselves, with churches, trade unions and other traditional
forms of association losing both members and a sense of engagement.
With the increasing individualization of society, traditional collective
identities and organizational affiliations have become enfeebled.

Party leaderships have also sought to reduce the weight of their ties to
associated groups, and to downgrade the privileged access once accorded
to affiliated organizations.39 Increasingly, parties tend to think of them-
selves as self-sufficient and specialized organizations, willing to listen
to particular social actors but avoiding any close formalized links with
them. Leaders have distanced themselves from civil society and its social
institutions, while at the same time becoming more firmly entrenched
in the world of government and the state. We may summarize the key
developments of this process as follows.

38
As, for example, when British Labour leaders shrugged off their defeat when the
Labour Annual Conference voted to restore the link between pensions and average
earnings. The vote had gone 60/40 against the leadership, and Gordon Brown
responded: ‘I’m not going to give in to the proposal that came from the union lead-
ers today . . . It is for the country to judge, it is not for a few composite motions to
decide the policy of this government and this country. It is for the whole commu-
nity, and I’m listening to the whole community.’ Guardian, 28 September 2000.
39
A trend already noted in nuce by Otto Kirchheimer in his then highly prescient
analysis, ‘The Transformation of West European Party Systems’, in LaPalombara
and Weiner, eds, Political Parties and Political Development, Princeton 1966,
pp. 177–200.
 mair: Depoliticized Democracy 47

Firstly, as is now widely recognized, parties in most Western democracies

have moved from being principally dependent for their organizational
survival on the resources provided by members, donors and affiliated
organizations to being increasingly reliant on public funds and state
support. In most countries today, and in particular in almost all newly-
established democracies, the preferred source of party funding has
become the public purse.40

Second, parties are now increasingly subject to new laws and regula-
tions, which sometimes even determine their internal organizational
functioning. Many of these regulations were introduced in the wake
of public funding for parties, with the distribution of state subventions
inevitably demanding a more codified system of party registration and
control. Controlling party access to the public broadcasting media has
also required a new system of regulations, which again acts to codify the
status of parties and their activities. From having been largely ‘private’
and voluntary associations, parties have increasingly become subject to
a regulatory framework which has the effect of according them a (quasi-)
official status. As the internal life and the external activities of parties
become regulated by public law, the parties themselves become trans-
formed into public-service agencies, with a corresponding weakening of
their own organizational autonomy.

Finally, and perhaps most obviously, parties have cemented their link-
age to the state by according increasing priority to their role as governing
(as opposed to representative) agencies. In political-science terms, they
have become more ‘office-seeking’, with a place in government not only
a standard expectation, but also an end in itself. Some forty years ago, a
now classic review of political developments in Western democracies was
organized around the theme of ‘oppositions’.41 Today opposition, when
structurally constituted, increasingly comes from outside conventional
party politics, whether in the form of social movements, street politics
or popular protests. The parties, on the other hand, are either governing
or waiting to govern. With this new status has come a downgrading of
the role of the ‘party on the ground’, and a shift in the party’s organi-
zational centre of gravity towards those elements that serve its needs

40
Ingrid van Biezen, Financing Political Parties and Election Campaigns, Strasburg
2003.
41
Robert Dahl, ed., Political Oppositions in Western Democracies, New Haven 1966.
 48 nlr 42

in parliament and government. This move might also be seen as a

final manifestation of the classic Downsian or Schumpeterian notion
of parties as ‘competing teams of leaders’, in which the party organiza-
tion outside the institutions of the polity gradually withers away. What
remains is a governing class.

Passive and privatized masses

All of this has had major implications for the functions that parties per-
form within the wider polity. Conventionally, parties are understood to
integrate and, if necessary, to mobilize the citizenry; to articulate and
aggregate interests, and translate these into public policy; to recruit and
promote political leaders; and to organize parliament, the government
and the key institutions of the state. That is, just as parties aimed to com-
bine government for the people with government by the people, so too
they combined key representative functions with key procedural ones—
all within the same agency. As parties have changed, however, and as
the mass-party model has passed away, the functions they perform in
contemporary polities have also shifted, and now focus much more on
procedural ones. This development goes hand in hand with their move
from society to the state, and is part of the process by which parties and
their leaders separate themselves from the arena of popular democracy.
Parties have become agencies that govern—in the widest sense of the
term—rather than represent; they bring order rather than give voice. It is
in this sense that we can also speak of the disengagement or withdrawal
of the elites, although while exiting citizens are often headed towards
more privatized worlds, the exiting political leadership is retreating into
an institutional one—a world of public offices.

The process, then, is mutually reinforcing.42 Citizens turn from being

participants into spectators, while the elites gain more space in which to
pursue their own shared interests. As one commentator put it:
Our governors have become a self-perpetuating elite that rules—or rather,
administers—passive or privatized masses of people. The representatives
act not as agents of the people but simply instead of them . . . They are
professionals, entrenched in office and in party structures. Immersed in a
distinct culture of their own, surrounded by other specialists and insulated

42
See also John Hibbing and Elizabeth Theiss-Morse, Stealth Democracy,
Cambridge 2002.
 mair: Depoliticized Democracy 49

from the ordinary realities of constituents’ lives, they live not just physically
but also mentally ‘inside the beltway’.43

Two outcomes of this mutual disengagement may briefly be noted. In

the first place, the resulting gap has sometimes helped to fuel a populist
mobilization usually, but not exclusively, on the right. In other words,
partly as a result of this withdrawal, the political class has itself become
an issue of contention in a large number of democratic polities. Second,
and as noted above, the growing distance between citizens and their
political leaders has also helped to fuel elite demands for more ‘non-
majoritarian’ decision-making, and a greater role for non-partisan and
non-political agencies—judges, regulatory bodies, central banks and
international organizations.

In addition, with the separation of representative and procedural party

functions, and the shift from society to state, the distinction between pop-
ular and constitutional democracy becomes more salient. Through the
party, one and the same institution within mass democracy gave voice to
the citizenry and governed on their behalf. In such a context, popular and
constitutional democracy were more or less inseparable. With the grow-
ing gulf between the citizenry and the political leaderships, it becomes
increasingly difficult to effect this sort of symbiosis. A space is created in
which the features of popular democracy, taken more or less on its own,
can be weighed against those of constitutional democracy; government
‘by the people’ comes to be judged against government ‘for the people’. In
this assessment, it is usually popular democracy that is found wanting.

The difficulty runs deeper. Elsewhere, I have argued that, with the sep-
aration of parties’ representative and procedural roles, the increasing
emphasis on the latter was part of a more or less necessary process of
adaptation: precisely because they no longer functioned so effectively
as representatives, parties sought to compensate by building up their
role within the institutions. These were not therefore parties in decline,
I then argued, but instead had adapted to a new set of circumstances,
seeking to survive in the context of a new organizational equilibrium.44

43
Hanna Pitkin, ‘Representation and Democracy’, Scandinavian Political Studies,
no. 3, vol. 27, 2004, p. 339.
44
Mair, ‘Political Parties and Democracy: What Sort of Future?’, Central European
Political Science Review, vol. 4, no. 13, 2003.
 50 nlr 42

This now seems far too sanguine an interpretation. Parties might well
seek to compensate for diminished capacities in one direction by enhanc-
ing those in another, but there is no guarantee that they will succeed. On
the contrary: parties may be able to fill public office, but having aban-
doned their representative role, they may no longer be able to justify
doing so. In other words, if parties as governors are to be trusted, and
if party government more generally is to be legitimate, it is likely that
the parties must also be seen to be representative. For an elected politi-
cian, it is not enough to be just a good governor; without some degree of
representative legitimacy neither the parties themselves, nor their lead-
ers, nor even the electoral process that allows them to be chosen, will be
seen to carry sufficient weight or authority. The result will be to encour-
age distrust and scepticism.

Scepticism towards elected politicians is nothing new, of course. Nearly

sixty years ago, Schumpeter warned against relying too heavily on those
who were emerging from the electoral process, and suggested that ‘the
qualities of intellect and character that make a good candidate are not
necessarily those that make a good administrator, and selection by means
of success at the polls may work against people who would be successes
at the head of affairs.’45 The argument has been reiterated many times
since. But while the skepticism may not be new, it does acquire a more
robust foundation when articulated within a context in which popular
democracy has become distanced from constitutional democracy.

In fact, what we see here is a largely self-reinforcing process. As politi-

cal and party competition are hollowed out even further, they offer even
more encouragement to the politics of the spectacle and the horse-race.
And this becomes more likely to produce the sort of candidates and
elected politicians whose qualities, following Schumpeter, are even less
likely to be those of the good administrator.

What are the implications of these processes for the future of Western
democracies? I have suggested that the transformation in the role of
parties, as they have shifted away from expressive and representative
functions and moved closer to becoming appendages of the state, has
played a central part in the disaggregation of democracy’s popular and
constitutional components. Any broader reckoning as to why this is

45
Joseph Schumpeter, Capitalism, Socialism, and Democracy, 2nd edn, New York
1947, p. 288.
 mair: Depoliticized Democracy 51

happening—and why now, barely one decade after the much heralded
‘triumph of democracy’, attempts are being made to downgrade its
popular pillar and limit its scope—must take into account a number
of themes that fall outside the purview of this essay: the impact of the
end of the Cold War, the decline of ‘embedded liberalism’, the declin-
ing purchase of party government, and the more general fallout from
processes of globalization and Europeanization. But the focus on parties
makes one further irony impossible to ignore: the victory of democracy,
in this form, poses stark problems of representative legitimacy for
the new governing class.