‘The Democratic Deficit’

Joseph Raz

1) Introduction
‘In 2017 … autocracy was on the rise. The “democratic recession” continues.’
Martin Wolff, The new world disorder and the fracturing of the west (FT 2/1/18)

It is not surprising that many of us feel somewhat confused and conflicted

about democracy these days. Conviction that it is the best and the only
legitimate form of government struggles to come to terms with
disappointment over a series of electoral results in several countries, and over
a reported decline, mostly among young people, in a sense of the importance
of living under a democratic government.
We may well feel that the constitutional practices and laws in many
democracies are deficient, and vulnerable to exploitation and corruption. What
is needed is reform which will make our practices conform with our ideals.
And there can be no doubt about that. It’s not only that nothing is perfect, but
that continuing and accelerating changes in the demographic and economic
character of our societies, in the means of communication, and in the
character of interpersonal interactions require that democratic rules and
procedures should be continuously modified and adapted. This is a hard task
but not a dispiriting one. What is dispiriting is the creeping doubt that there
are fundamental obstacles hindering any political system from living up to the
aspirations of the democratic prophets.

1
Among political scientists it has always been easy to find a more sober
assessment of what can and cannot be expected of democracy. But their
studies have only local validity, and are often marred by theoretical
shortcomings.
It is difficult to judge whether the failings of existing democracies are
remediable shortcomings or evidence that whatever the attractiveness of
democratic ideals, they are incapable of being adequately implemented. Many
political philosophers are undaunted in upholding democracy as the ideal,
indeed the only morally acceptable form of government. That view and the
arguments on which it rests are the rock. The hard place are all international
and transnational organisations, be they the United Nations and its agencies
like the WHO and UNESCO, or the World Trade Organisation, and most
clearly territorial organisations like the EU, which resemble most closely
states, the primary locus of democratic government. Not only do all such
organisations suffer from a democratic deficit, but they seem doomed to suffer
from it. It is, to put it mildly, hard to envisage how any of them can be
genuinely democratic, without itself turning into a state.
Is this our choice? Either retreat from the existing transnational organisations,
insisting on the state as the only legitimate large scale political organisation, or
make do with imperfect transnational organisations? Perhaps we should revisit
and revise our understanding of the pros and cons of democratic
governments?
But perhaps these doubts are unfounded. I do not mean merely the possibility
that, regardless of their many failures, their democratic character makes life in
the existing democracies better (for most?) than life in the non-democratic
ones. That may be true, and is hard either to establish or to refute, which
makes it a belief that, once embraced, can easily persist. A more interesting
claim is that living in a democracy has a virtue or an advantage of great moral

2
and personal importance that has nothing to do with how well the government
performs. It depends on nothing more than its democratic character.
Something along these lines would establish an intrinsic value in democratic
governments, and presumably also in being governed by such governments.
That opens various possibilities: maybe transnational bodies can acquire that
value, whether by becoming democratic – which as I said is not likely – or in
some other way. Or, maybe so long as the transnational bodies supplement
states, and do not replace them, people may enjoy that value by being citizens
of states, and do not require that it should also be present in transnational
bodies. For example, if the value is in having a fair chance of wielding political
power, or of the experience of wielding political power, then people have that
chance in their states, and it may not matter that they have much less of a
chance of wielding political power in, for example, the European Union.
All this is very abstract, and I will come back to it later in my talk.

But this is not the only possible explanation that makes continued belief in
democracy rational. Suppose that there is no intrinsic value in democracy. It is
possible that, though mistaken, holding this view is not irrational. That could
also help explain people’s continued belief that legitimacy depends on being
democratic. It is a familiar fact that when knowledge depends on specialist
inquiry, most people, quite rightly, derive their beliefs from the common view
of those taken to be expert on the matter, and as some of those who are
reasonably taken to be expects are not in fact experts, or are experts no
longer, false beliefs which people have good reason to hold may persist or
spread.
Furthermore, so common a mistake can actually make it the case, or at least
contribute to making it the case, that the governments of democratic states in
which these mistakes are widely shared are legitimate. It does not matter why

3
you think that the government is legitimate, and you may not have any
explanation of why it is legitimate: so long as you believe that it is legitimate
independently of this belief of yours, so long as you believe that your belief
merely reflects how things are independently of it; you are mistaken, yet your
mistake contributes to making the government legitimate, brings it closer to
being legitimate and may make the difference between being and not being
legitimate.
This self-correcting belief is more than interesting. It is a major part of the
explanation of how governments can be legitimate at all; why legitimacy, which
looks like an almost impossible condition, is actually not that difficult to
achieve. It also explains why Hobbes was closer to the truth than most of his
detractors. To see that, we need to make a fresh start. And I will start with
some semi-technical clarifications.

2) Some suppositions
My aim is modest: to say something instructive and helpful, by way of an
introduction, about how to establish what conditions make authorities
legitimate, the focus being on the so-called problem of the democratic deficit,
an expression most commonly used about international authorities, but which
can be applied to any authority when the suspicion arises that its legitimacy is
in doubt because it is not democratic enough, or not democratic in the right
way.
This may seem to be the place to define ‘democratic government’ or
‘democratic constitution’. I will avoid doing so, and would recommend avoiding
the issue altogether. The reasons are many. One important cluster shows how
disputes about the definitions tend to become verbal disputes. Some people
are so deeply convinced that democracy is the only good form of government,
and that it incorporates the main virtues of government, that they proceed to

4
include in their definition of democracy more or less any good property a
government should have. ‘Being compassionate is not part of democracy? How
can you say that? Surely a government that is not compassionate is at least
incompletely democratic.’ Etc. Besides, some believe that there is one and only
one form of democratic government. Others doubt it, and what could be a
verbal dispute about defining democracy disjunctively (it is any government
that is like this or like that etc., etc.) disguises disagreement about the ideal or
best forms of governments, which one need not indulge in discussing the
merits of democracy. For me the decisive reason for avoiding a definition is
that in all countries, however democratic they are, some organs and processes
of government are undemocratic, or less democratic than others. The most
helpful arguments are arguments about the suitability of including this or that
institution or governing process in the government of this or that country in
current, or likely conditions.
Should we then define ‘democratic institution’ allowing that it takes many of
them to make a country democratic? That route has its pitfalls, one being that
many institutions are democratic and contribute to the democratic character
of a government when they exist alongside some other institutions, but not
otherwise.
I would therefore confine myself to the observation that at the core of a
democratic government lies the right to vote for the main institutions of
government, possessed by all adults who live in the country, a right which is
accompanied by devices securing access to information relevant to its use, and
whose exercise is free from undue influence. The most cursory reflection on
this cumbersome sentence will incline some to add to it and other to delete
elements from it. Indeed, as it stands there is no country whose government is
democratic. But I am not offering the sentence as a definition, or part of one. It
is merely meant to indicate some of the least controversial elements in the

5
traditions of thinking about democracy. It is meant to locate the subject, not to
define it.1
The existence and activities of international authorities can have symbolic
significance, expressive meaning and normative force. The European Union may
symbolise the common identity of the people of its member states, and may be
a focus for sentiments of pride, hope, despair and others. Its existence or
activities may express striving towards solidarity, and its treaties and rules may
be normative reasons for action for governments and individuals. Legitimacy
relates primarily to the normative force of institutional action: if the
institutions are legitimate then their actions, which are meant to set reasons
for the guidance of themselves or others, do so because they are so intended.
In other words, we are bound by the rulings and decisions of legitimate
authorities that are meant to bind us, but not by those of illegitimate
authorities.
The symbolic, expressive and normative forms of significance of an institution
and its activities while distinct are inter-dependent. But the way they are is
commonly not dictated by the constitution or nature of the institution alone. It
is in large measure the result of the way the institution functions, and is seen
or believed to be functioning. Legitimacy relates to the normative force of its
rules and decisions because institutions, by their constitution, presume to have
the power to set reasons. That makes their legitimacy a focus of special

1
Another tempting thought is that an institution or process that is likely to make the
government or some part of it more responsive (than it would be without it) to
expressed preferences of the people, or some sections of the people, is a
democratic element in the government. There are many difficulties in any attempt to
clarify this obscure thought, but its main problem from our point of view is that
there are many such institutions in non-democratic governments, and they are often
perceived to be a substitute for democracy, rather than a way of implementing it.
Besides, the market can be said to be such an institution, but is it really constitutive
of democracy? We are again drawn to a dispute whose place is in developing a
complete theory, not in identifying the subject it deals with.

6
attention, even where the ceremonial and symbolic role of those institutions is
in some ways more important or more prominent.
You will notice that I freely use weasel words (like ‘primarily’, ‘commonly’).
That is meant to allow that the concepts used in the discussion are the ones
we use in discourse relating to the practices surrounding these institutions.
Such concepts are only as sharp as their common use requires; meaning that
their use allows for margins of vagueness when encountering circumstances
that are not normally envisaged by those involved or interested in the
functioning of the institutions, or when dealing with circumstances those
involved would prefer to be vague about.
I have started referring to institutions, rather than to international authorities
or state governments etc. This does not indicate a change of topic. It results
from a desire to avoid the need to mark boundaries that are not needed for
theoretical purposes. First, in today’s world the once firm distinction between
public and private powers has been eroded. I will therefore not limit the
discussion to public authorities. Second, the framework for determining the
legitimacy of international institutions is none other than the framework for
determining the legitimacy of any institution. So, to avoid unnecessary
distinctions we can consider institutions generally.
Institutions are agents, but unlike humans, whose existence and life necessarily
have or may have intrinsic value, the existence and life of institutions do not
necessarily have intrinsic value. Institutions have interests. They have an
interest in achieving the goals set by their constitutions, and they have an
interest, let us assume, in continued existence. But given that their existence
and life do not necessarily have intrinsic value there may be no reason to serve
or protect their interests, no value in their continued existence or in their
achieving their goals.

7
Things are different with people: if, or when, their life, or some aspects of it,
has intrinsic value it necessarily has that value. Their actions are guided by
those values, they aim to protect the valuable aspects of their life, or to
develop others, and they recognise and respect (have reason to recognise and
respect) the valuable aspects of people and other beings whose existence or
life has intrinsic value. None of this gives them reason to respect or serve the
interests of institutions that do not have such value. Hence the special
character of the question of legitimacy: under what conditions do the
existence and actions of institutions constitute or provide reasons for people?

3) Legitimacy
As you see – legitimacy as I use the term (and there is nothing wrong in other
uses of it) relates to the power to set reasons (by expressing an intention to
do so), which is justified by the desirability of having that power. An institution,
including an international one, is legitimate if its rules and decisions have
binding force, namely if they set reasons for those for whom they are intended
to set such reasons. Under what conditions is it legitimate? The question can
be broken up into several, and in various ways. I will divide it into three,
though they are partly interlinked and inseparable:
First: every institution is constituted by a set of constitutive rules, which
determine its powers, and the standards governing their use. Among other
things, the rules may also define the purposes for which the powers should be
used. Some institutions are dynamic, in that their constitutions provide ways
for changing their purposes, and other aspects of their constitution.
I am not saying that all institutions, by their nature, have distinct or limited
purposes. Arguably sovereign states do not. But even they may set self-
imposed dynamically regulated limits to their jurisdiction partly defined by
setting specific purposes for themselves.

8
Naturally, legitimacy depends in part on there being reasons to conform to
putative reasons set by the institution. And that must depend on the value
served by its existence and activities. I can think of two ways in which such
dependence may be satisfied. First, there is sometimes value in those who have
the power to establish institutions, and to determine their constitutions, having
those powers. Second, the existence of an institution or its activities (its life)
has value. By the second test an institution is legitimate only if its existence and
purposes are worth pursuing, and can adequately be pursued by an institution
of that kind. The first test allows for the legitimacy of institutions whose
existence has no value, provided that there is value in those who determine
their existence or their demise having these powers. The value of the power is
not undermined by the fact that it is used in a mistaken or wrong-headed way.
The very fact that the one who has the power has it and uses it has value.
For those who doubt that this test makes sense let me add examples. It
is the only test that can apply to commercial corporations and to charitable
organisations: their existence depends on decisions of private individuals who
move to create them under laws that provide such powers. And these laws
and the powers they provide are justified if it is good that individuals should
have such powers. That value would then also justify the existence of the
institutions they created. That is also why promises are binding: there is value
in individuals having the power to bind themselves by expressing their
intention to do so. That gives them the power to make silly promises that are
nevertheless binding.
Naturally, this test can yield divided results: some aspects of an
institution may be legitimate, or partially legitimate, while others are not. It
may be legitimate at one time but not at a later time, because changes in the
world, or in the institution’s constitution, may render its existence unjustified.

9
Having spent some time introducing this way of establishing the legitimacy of
an institution I will now ignore it, because states, and super-state organisations
like the UN or the EU, cannot be justified that way. I do not mean that they
are never established by acts of founders, only that their justification does not
depend in the same way on the founders. They are justified by the results, by
the quality of their character and life.

While the first test for the legitimacy of an institution turns on the value of its
existence, given its origin and structure, the second test turns on the character
of those who govern it, on the reply to the question of who should hold the
powers of the institution. Strictly speaking, this question is but part of the
question of the constitutive norms of the institution, and the standards
governing its operation. But as the question ‘who should govern, who should
wield the powers that the institution confers on its organs?’ attracts much
attention as a separate issue in political debates, I am singling it out as the
second condition of legitimacy.
And then there is a third test, turning on the question whether the legitimacy
of any action by the institution depends on observing certain precepts, say of
justice or fairness. The constraints of the third test apply to individual
institutional acts, but if the rules or constitution of the institution violate them,
or even if not, but if its organs are habitually disposed to violate them, the
constraints may show that the institution itself is altogether, or in large degree
or in certain parts, illegitimate.
The three questions intermingle and overlap. Given our interest in democracy,
the second test is our focus. But it may be good to start with the first, for it
threatens to absorb the second altogether.
All three questions can be said to be answered by the value of the existence of
an institution, and of its likely activities. An argument about the first condition

10
for the legitimacy of an institution aims to show that the purposes for the
pursuit of which it exists are valuable, and that it is likely to secure them, or to
continue to secure them. Given that by and large the interests of institutions
have no intrinsic value, these conditions can only be met if the institution’s
existence and life serve the interests of people, which means that they are met
only if its interests are aligned with interests of people.
Needless to say, to give content to this abstract statement one needs to argue
about what interests people have, not a task for today. There are many other
difficulties in generating a definite and informative set of conditions for the
legitimacy of institutions. For example, it is not at all clear that legitimacy
depends on the institution doing on balance more good than harm. Some
institutions will be legitimate even if they do more harm than good, depending
on the relative importance of the harm and the good, and on the difficulty of
achieving that good in some other way. The cost of change, or of attempting
change, may be sufficient, in the circumstances, to show that the institution is
legitimate. It all points to the difficulty of saying much at a high level of
abstraction. Substantive normative conclusions require examination of more
limited types of institutions, and more specific circumstances in which they
exist.

4) Subsidiarity
We can illustrate the difficulty in finding universal principles of legitimacy by
examining briefly one apparently promising candidate. A familiar and powerful
limitation of legitimacy is the subsidiarity principle. I will use ‘subsidiarity
principles’ in the plural to refer to any principle that says that an institution is
legitimate only if the good that establishes its legitimacy cannot be secured by
individuals acting without the mediation of institutions, or by institutions
whose decisions are taken at a less centralised level (towns rather than

11
regions, regions rather than countries, etc.). It is not easy to formulate a test
for being a subsidiarity principle. But the thought may be clear enough.
The striking feature of subsidiarity principles is that they are strong enough to
be tie breakers: they reject the legitimacy of an institution even if it is as good,
in its aims and in its life, as the aims and results of individuals or a less
centralised institution acting without its mediation. That means that subsidiarity
principles cannot be justified on grounds of efficiency alone. They express a
preference for actions by individuals (or by institutions that are closer to
individuals) over actions by institutions (or more remote institutions). How
could that be justified?
Perhaps because, other things being equal, it is better that a person does
something worth doing him- or herself than that they have it done to or for
them.2 Obviously, other things are often not equal.
Indeed, subsidiarity principles presuppose that where they apply there is
nothing distinctively valuable in the activities of the institutions concerned that
is lost when they do not exist. And that includes the presupposition that no
involvement of individuals in institutions and their activities is distinctively
valuable, a value that is lost when the institutions do not exist. If there is
something distinctively valuable in the existence or functioning of an institution
then even if its aims can be secured by individual actions, it may be better to
secure them through the activities of that institution and have the extra
advantage of whatever is distinctively good in its existence.
That possibility cannot be easily brushed aside.3 For one thing, sometimes it
may be better for individuals if certain tasks are not in their hands, if they are

2
It is important to understand that subsidiarity principles do not express an
individualistic preference, when that is understood as a preference for people being
active. There are plenty of opportunities for people to co-operate and act in
combination with others, without the intervention of institutions.
3
My observations here illustrate what can be called the anti-consequentialist
presupposition underlying this paper. The presupposition has two elements: The first

12
looked after rather than having to look after themselves, and that can be so
even when they can look after themselves. And it is plausible that there are
other ways in which the existence of institutions can itself be valuable:
Institutions work through the activities (and omissions) of individuals who are
organs of the institutions or are assigned certain roles in their governance.
Plausibly, being so involved with an institution, and having the opportunities to
be so involved, are (or can be, given the right conditions) valuable in a
distinctive way, for the individuals concerned. So even subsidiarity principles
are not universally valid. Principles of legitimacy demand a nuanced and
contextualised examination.
That is the point where the value of democracy returns to the discussion.
******
I have been arguing that if there is an independent, distinctive, value in being
engaged in the government of an institution then the existence of that
institution has value that may override its deficiencies in protecting and
securing the goods that it is meant to secure. The first question about
legitimacy made success in securing these goods the foundation of legitimacy.
Subsidiarity principles introduced a comparative element in the legitimacy test:
to be legitimate an institution has to fulfil its aims and functions better than any
other arrangements that keep power closer to its subjects. The fact that there
may be value in being among the governors of an institution shows that
subsidiarity is not always valid.

can be called value pluralism, i.e. the view that not all valuable things, activities,
experiences etc. are valuable in the same way. They may manifest different values,
and the difference should be noted in our appreciation of their value. Second, that
given value pluralism there may be cases in which choices among different values are
forced on us where there is no “right answer” to the question which combination of
values should be preferred. The text above displays belief in value pluralism, and is
neutral on the question of the commensurability of value combinations and the
existence or otherwise of a right answer to choices among them.

13
The possible limitation we noted on the generality of subsidiarity principles
gives the question: ‘who should govern the institutions?’ independent standing.
But for that limitation the answer would have been totally subservient to the
answer to the first question: those should govern who would best secure the
legitimate purposes of the institution, that is, those who would best secure
those purposes that if secured to a sufficient degree would make the
institution legitimate.
A final clarification before turning to democracy: It is possible for an institution
to count among its aims the creation of opportunities for people to engage in
activities that only the institution provides. To the extent that it does, the
appropriate subsidiarity principle that applies to it will reflect that. It will allow
for the legitimacy of an institution resting on its ability to provide such
opportunities. In suggesting that the value of engaging with an institution sets
limits or exceptions to subsidiarity principles I did not deny that. I was merely
pointing out that not all these opportunities need be part of the institution’s
purposes. Some people will be inclined to argue that by implication they always
are. But such arguments force artificial unity on the questions, at the cost of
ignoring workable and natural distinctions that serve us well in thinking about
these issues.
There is another example of this kind: arguably, institutions are legitimate only
if their structures allow them to adapt, develop and self-correct past faults.
This is another matter which belongs with the second of my three questions:
how should institutions be governed? It would be a pity to force it into the
first, namely whether the institution achieves worthwhile goals. The two
address different time dimensions, and raise different issues which are best
kept apart.

14
5) The Advantages of Democracy – Part One
Democracy is in some circles the most popular answer to the question who
should govern, at least when the institutions concerned are political ones.
In considering this claim I will focus on state governments, though I assume
that similar considerations apply to other political institutions. There may be
fairly general agreement on three advantages of democratic governments: First,
given contemporary circumstances, those whose life is governed by democratic
regimes are more likely to enjoy civil and political rights than those who live
under different regimes. Some civil and political rights are an integral part of
democratic governments, meaning that the democracy is at best imperfect if
these rights are not enjoyed by the inhabitants of the state concerned. Other
civil and political rights may not be integral to democracy, but are more likely
to be respected in democratic countries. Second, democracies provide more
and more extensive opportunities for people to engage in public affairs, and to
express publicly their attitudes to public matters. These opportunities are
protected by the civil rights, but extend beyond them. Third, given
circumstances prevailing in post-industrial societies, democratic governments
are more likely to be responsive to the expressed preferences of their citizens
than other kinds of regime.
I am aware that many will think that my qualifications ‘given contemporary
circumstances’ and ‘in post-industrial societies’ are unnecessary. I think that
they are essential, but I will avoid that issue. In combination, these three
factors suggest that some forms of democratic government in some
institutions and under some circumstances are more likely to secure the value
of the institutions, with minimum bad consequences.
This qualified conclusion, ‘some forms of democratic government’ fits
well with actual practices in contemporary democracies. All of them discharge
some functions of government through non-democratic organs, or through

15
bodies which are only indirectly and to a limited degree controlled by
democratic organs of government. Famously, in many countries the existence
of an unelected and independent judiciary is not only preferred but is believed
to be vital to protecting civil and political liberties. So that even democracy
itself depends on some important functions of government being discharged by
institutions that are not themselves democratic. In recent times some
countries placed their central banks beyond “political”, namely democratic,
intervention. Others look for ways of guaranteeing that the main media outlets
be protected from political, i.e. democratic, interference, and so on.
It would be interesting to know how big is the advantage of enjoying civil and
political rights in democracies relative to non-democratic regimes. There are
some helpful surveys, but they do not allow us to draw any general
comparisons. Both democracies and non-democracies vary greatly in their
respect for such rights. It is also very difficult to collect the data needed for the
comparison. It is relatively easy to scan the law of different countries and
identify the rights it recognises. It is much more difficult to assess the degree
to which the condition the right is meant to secure (the right to freedom of
expression aiming to secure for people a condition in which they are free to
express themselves, etc.) obtain in different countries. It can obtain without
the protection of a legal right, or fail to be realised in spite of legal protection.
We have to be content with partial surveys and personal knowledge. But they
are a sufficient ground for confidence in the three advantages of democracies
mentioned above.

6) The Advantages of Democracy – Part Two

There is a second group of advantages of democratic government, or at any
rate advantages that some of them achieve in contemporary conditions better
than many other forms of government. I will mention four:

16
1) Stability: the relative stability of one’s political and economic conditions
is essential for being able to feel that one is master of one’s life, that one
can have plans and hopes for the future. It is also essential for the
possibility of successful medium term political planning. In some parts of
the world democratic governments have been more successful than
most other contemporary governments in securing stability.
2) Peaceful transition of power. I tend to think of English history during
the 15th to the mid-18th centuries, but it’s not unique in illustrating how
absence of a secure and generally accepted way of effecting transition of
governmental power from one person or group to another encourages
repression, conflict and war. Again, in the last 200 years or so, more or
less democratic governments in various countries (though not in all
countries toying with democracy) were more successful in providing
reliable processes for transfer of power than most contemporary
alternatives.
3) Loyalty: Simplified and speculative as my last remarks were, the
remaining remarks in this section are even more so. All I will venture to
say is that in recent times in many more or less democratic countries
their democratic character, real or believed, enhanced their inhabitants’
loyalty to their countries. It is plausible that in this regard as well,
democratic governments enjoyed a greater degree of loyalty than many
contemporary non-democratic governments. And loyalty to the
government does facilitate governing and inclines to make the
government more successful.
4) Solidarity: closely related is the tendency of the democratic character
of a government to engender or cement solidarity among the governed,
solidarity that can be valuable in itself, and is vital for the willingness of
sections of the population to support others, even at considerable cost

17
 to themselves, again making it easier for government to attend to the
diverse needs of different sections of the population. Arguably, both
loyalty and solidarity are manifestation of a more general feature
essential for the successful functioning of governments, namely trust.
Again, it is likely that in recent times democratic governments enjoyed a
greater degree of trust by those subject to them than other
governments.
There is a common theme to these four absolutely vital advantages of
democratic governments: they all derive from and depend on belief that
democratic governments are legitimate because they are democratic. Many
who believe that, further believe that only democratic governments are
legitimate.
It is arguable that acceptance of the legitimacy of the government by those
subject to it is, generally speaking, a condition of its legitimacy. Here
‘acceptance’ means putting up with the government either without reflection
on the matter, or without much reluctance. But it also incorporates belief that
one is accepting the legitimacy of the government because one should, and that
one should because it is legitimate.
There can be various arguments for acceptance being a condition of legitimacy.
I will rely on a single simple argument, namely that without acceptance by the
bulk of the population the effectiveness of government action is greatly
reduced, and the undesirable effects of any government measure are
multiplied, as people take action to evade its consequences that are unpleasing
to them.
The result is that when people accept the legitimacy of the government
because of a belief that it is legitimate, that very belief, whether independently
true or false, tends to self-verify. Of course, there are other conditions, but in
many fairly standard circumstances, acceptance due to belief that the

18
government is legitimate tilts the balance, and in this day and age that belief
often derives from the view that democratic governments are legitimate
because they are democratic.

7) Legitimation
What is often discussed under the title ‘the legitimation problem’ is not the
question of legitimacy that I am discussing. It is close to the question of when
do people accept the legitimacy of governments? My observations above are
meant to show that when a government meets the condition of legitimation,
i.e. when it is accepted by the bulk of its subjects, it may well also be legitimate.
That assumes that it does not violate the conditions mentioned in the third
question about legitimacy, and more. Legitimation is by no means sufficient for
legitimacy. But it means that a major hurdle is overcome.
Some people would find this conclusion unsurprising because, they believe,
people would not believe that a government is legitimate if it did not meet the
other conditions of legitimacy. I find that regarding any large institutions,
groups or countries, this is a rather optimistic assumption: the conditions in
which it is likely to be true are not often met. That is not primarily because of
doubts about people’s ability to know what is good for them, but because of
doubts about their likely understanding of what political choices are conducive
to their good, and because the outcome of votes in democracies is often
anything but a simple aggregative function of the views of the voters, and
because of doubts about people’s willingness to act in the general interest and
for the common good.
What is important is that acceptance of a government, whether for good or
bad reasons, or for no reason at all, has a double effect: First it is a necessary
condition of its legitimacy. Second, by itself, whether or not it makes a
government legitimate, it secures the four advantages of my second list above,

19
and by doing so it contributes to the realisation of the three advantages of my
first list. For example, civil and political rights are likely to be better respected
in a country with a stable government.
Do not these observations overestimate the boot-strapping effect of
belief in legitimacy? You will not be surprised that I am willing to defend it,
relying on two broad observations:
First, by and large whenever sizeable human groups develop modes of co-
operation that enable them to survive over generations, their continued
existence rests on the prevalence within them of practices that regulate
economic activities, sex and procreation and provide outlets for innate drives,
for physical activity, imagination, aesthetic expression and more. As a result,
children born or brought up in such groupings are able to absorb their
practices, and find within them possibilities of survival, friendship, creativity and
self-expression that make rewarding and fulfilled life possible, though not
always attained.
The big and important caveat is that human groups of that kind often do not
extend access to all their members to the conditions that make rewarding life
within them possible. Often their practices rely on excluding some groups,
exploiting them or repressing them. Most of the successful, and many of the
failed, reform movements that grew within societies strove to mitigate or
abolish the exclusions and extend the conditions that enable fulfilling lives to
some of the exploited or repressed groups. True, the transformations that
success in such efforts bring about may well partly transform the meaning of
opportunities available in the society (think of the transformation of marriage
when no longer confined to couples of different gender). But these often are,
and should be, incidental results, not ends in themselves.
The second observation I rely on is that while there is hardly any (virtually, no)
group or society whose condition could not and should not be improved,

20
attempts to improve them are unlikely to secure their intended results while
avoiding undesirable consequences that undermine their value, unless they are
based on relatively small improvements in existing features of life in the group.
These observations should be fairly elementary. What do they say? That
humans are born into conditions not of their own making. Many of them
having worthwhile lives because they adjust to their conditions and find ways
of making a life for themselves in the circumstances in which they find
themselves. In doing so they also change their conditions, and much of this
change is not planned or intended by anyone. It is brought about, largely, by
the multitude of deliberate actions of those who live at the time. But the
collective outcome of their deliberate actions is generally not what they
intended; this is true even of the small-scale changes in their own personal
conditions. They decide on their professional education, on their job
preferences, on the friends and partners they share their life with, and often
succeed because they adjust to the consequences of their decisions, rather
than because they aimed at these consequences.
This is even more so regarding the conditions of societies at large. Here too
reforms often succeed not because they achieve their aims but because people
can adjust to their consequences. They are likely to achieve their aims to any
significant degree, only when those are not to transform the social and
economic condition of their societies but to build on what is there, and
ameliorate it, removing some of its shortcomings.

8) Justifying the boot-strapping effect

Why do these observations matter to consideration of the legitimacy of
institutions? Because, and again I will illustrate the point by reference to states
and other territorial institutions, people are born and bred in societies with
legally shaped institutions and cultures, which are not of their choosing and

21
which in the generality of cases were not made with them in mind, and yet in
many countries the majority of them find their way to build a rewarding life
that is created by turning the cultural practices and the legally supported
practices into the basis on which to build their lives. They succeed not because
those practices are so good, but because they make them good or tolerable by
adjusting to them, and by adapting them.
Seeking paradox, we may say that the law is not good because it meets some
moral or other desirability conditions. It is made good by people putting up
with it, just as they learn to put up with the weather, and making a life for
themselves that takes the law for granted, and is built around and within it.
Even in those modest requirements governments often fail, but if the
governments are accepted, they achieve the four virtues of stability, peaceful
transfer of power, loyalty and solidarity, and these are almost enough to make
them legitimate. I say almost enough for in all societies I know of there are
excluded and oppressed groups, whose condition should be improved, and
there are always defects, pre-existing or accruing through changes in nature,
technology or social practices, defects that should be remedied. And they may
sometimes affect the very legitimacy of the governments. There are more
reasons for that than can be explored here. They relate to the way the four
virtues that acceptance can be said to secure ramify and enable institutions to
have other virtues as well. They are better able to achieve their aims, they are
more likely to acquire symbolic importance for their members, and more.
*****
It may be said that I take the virtues of democracy to be due to a placebo
effect. There is some limited truth in that. Acceptance of government
strengthens the case for its legitimacy. When acceptance is due to a false belief,
including a false belief in the virtues of democracy, that too strengthens the
case for the legitimacy of the government. And this time that effect is similar to

22
a placebo effect. The analogy is far from complete for various reasons. For one
thing, there are genuine benefits of democracy, like a greater chance that one’s
civil and political rights would be recognised and protected. Second, and as a
result of the preceding point, not all beliefs in the virtues of democracy are
false, and that is especially so since many people believe in democracy but do
not have a specific account of its virtues. They just believe that it has important
virtues.
The crucial point is not that false belief in democracy has beneficial results, but
that belief in it (or in any other form of government) contributes to its
legitimacy whether or not it is true. The placebo analogy misleads in another
respect as well. The legitimacy of a regime depends in part on the available
alternatives, and on the cost of transition. To be available a different form of
government, even if it is merely a modification of the current one, has to enjoy
support that could turn into acceptance should it become the regime in that
country, and it also depends on forms of interaction and structures of
institutions that could sustain it, that can enable it to function as the regime of
the country. Belief that only democracy can in principle be legitimate is false,
but if widespread it can make a democratic regime legitimate for it will remove
any possibility of a legitimate alternative.

9) What has been missed?

Have I not missed the main reasons for democracy? Are they not that in
democracies everyone enjoys equal political power, and comes as close as
possible to realising the ideal of self-government as is compatible with its equal
realisation by others? I will have to be brief and rather peremptory in
dismissing these suggestions, relying on common knowledge of political
realities.

23
In no democratic country today do all inhabitants, or even all adult citizens,
enjoy equal political power. Broad observation suggests that there is no
discernible tendency to approach that condition. If anything, the obstacles on
the road to it are increasing. It is an impossible dream, which is not to say that
it is impossible to limit disparities in political power. And sensible policies can
do so in a desirable way.
Ironically, democracy may perfectly realise the equal power of self-
government. After all no one has that power in a democracy or in any other
form of government. We are all equal in this regard. However valuable are the
powers to cast a vote and to campaign for political positions or candidates,
they do not amount to power to govern oneself.
Of course, belief that democracy gives all voters a power to govern
themselves, should it be widespread, contributes to the acceptance of the
system of government. So does another view, namely that having the vote and
the right to campaign establish that those who have it are taken seriously, are
recognised as people of moral standing, whose interests should be taken into
account. Those who hold this view often also believe that only in democracies
are people, the voters, recognised as having moral standing, whose interests
should be taken into account. This second view seems to me false. Arguably,
Christian states recognise the moral standing of all people, but not only they.
The first conviction, that democracies qua democracies recognise the moral
standing of voters, is not always true. For example, there can be racist
democracies, which while allowing universal suffrage, ignore in their laws and
policies the intrinsic value of the interests of some groups among their voters.
Interestingly, the claim that in democracies and only in them do people feel
that they are recognised as having moral standing and interests that count may
be true. All marks of recognition of status are conventional, and therefore
contingent and local. In some, both democratic and non-democratic, countries

24
it may be the common view, regarded as self-evident, that the vote expresses
such recognition. And due to its prevalence having the vote comes to have that
meaning. It lacks it where and when it is not taken.
Given another occasion, it would be good to consider what we should and
what we should not expect of political ideals. Let me briefly illustrate: suppose
someone affirms
A: Only a state in which everyone living in it has his heart’s desires can be a
good state.
People having their heart’s desires is a condition on its being a good state.
Some people would object that that condition cannot be realised. Assuming
that that is true A can still be an independent desirability condition for the
state to be good. The closer A gets to being true of a state the stronger the
reason to regard it as a good state. But that is an acceptable part of the theory
of the good state only if there is something about the constitution, norms and
political culture of the state that makes it approximate the realisation of A
more than other states. Now the problem emerges: the features of the
constitution that will make it conform better with A may include features that
make it conform less well to some other ideal. Perhaps only in A Brave New
World can one approach satisfying A, that is, it will only be satisfied by
abandoning ideals of freedom, autonomy, respect, and others. If so, then A is a
sound political principle only if conforming to it does not harm other more
important ideals (and of course this misleadingly simple formulation hides many
pitfalls). Whether or not it does will depend, as it does in the example of Brave
New World, on contingent facts. Various contingencies can be ignored as very
unlikely or as involving circumstances whose bearing on political ideals is easy
to gage, and is unlikely to include any undesirable surprises. But this leaves an
indefinite number of conditions whose likelihood should not be ignored, and
whose normative implications are unknowable. Hence, for most normative

25
purposes political ideals should be judged relative to current or likely
circumstances, and the resulting theories will not enjoy universal validity.

10) Concluding reflections

It is time to point to some lessons as we move towards a conclusion.
It is as certain as anything about human societies that what we currently
understand as democracy will not remain for ever the best form of
government for the states for which it is so at the moment. Human relations
and the institutions that make them possible, change. But it seems that at the
moment it is the best form of government for many countries, meaning by this
that it is best for them if their form of government includes a good dose of
democratic elements, among others.
My reflections were not designed to cast doubt on that. Their aim was to
explain why this is so, in spite of the fact that so many conventional beliefs
about democracy are false.
You may think that my sketch reflects a quiescent attitude, an acceptance of
any form of government whatever its demerits. It does not.
True, my sketch manifests a conservative spirit in two ways. First, in its
appreciation of how valuable the four virtues are both in themselves and in
serving other aims of government, and how fragile they are. Recent history
provides ample evidence for these points. Second, in its belief that far-reaching
reforms whether ending well or badly, do not end with the intended results.
To the extent that we care about deliberate political actions that enable us,
those who take them, to achieve intended results, we should engage in gradual
ameliorative reform. Though, admittedly, even when a political action does not
achieve its intended ends it may achieve some of them, and anyway, in some

26
respects some existing conditions may be so bad as to justify terminating them
even with no alternative clearly in our sights.
****
Judging a government or some other institution legitimate is different from
judging an action of theirs justified or good. But the two assessments are not
independent. An institution is legitimate if its existence is justified, is good or
worthwhile. And that depends on the quality of its aims, the likelihood that it
will achieve those of them that are worth achieving, and the benefits that its
own existence brings. All these and other incidental advantages are to be
judged comparatively, that is comparing the feasible alternatives, taking account
of the cost and benefits of transition. Broadly speaking the same criteria
determine the legitimacy of state governments and of international institutions.
But subsidiarity principles, however qualified, make their legitimacy
interdependent; their application depends on what options are available in the
circumstances. If there is a legitimate and adequate state, or other more local
way, of dealing with a task, it may be illegitimate for an international authority
to claim jurisdiction over it.
This interdependence can be illustrated with regard to the place of democratic
elements in the case for legitimacy. It may be thought, for example, that
regarding one and the same population the importance of democracy for
securing acceptance will be the same for state institutions and for international
ones with jurisdiction over that same population. But that would be rash. It
depends, at least in part, on the aims of the relative institutions. I have noted
that in all states some state organs are accepted even though, and sometimes
in part because, they are not democratically governed. Views about which
organs should and which should not be democratically run vary from country
to country and from time to time. So do views about which functions should
be run by public rather than by private institutions, where most commonly

27
there is not much demand that they, the private institutions, should be
democratic. We can think of TV and the press, mail and communications
services, and education and health services as examples of functions which in
some countries are deemed best excluded from direct public management. In
others they are believed to require public management, but of a professional
quality, relatively independent of democratic controls, whereas in others still
they are thought to be best run by public and democratic managements. And
all of them may be justified in light of their local traditions and current
conditions.
It will be clear by now why I avoided defining democracy, or even democratic
institutions. There is no single democratic ideal. Not only is democracy, like
any other form of government, a historical creation, valid for some conditions
and not for others, and one that will eventually be superseded. Even where it is
the right form of government, there is no single ideal of which it is the right
form. Rather, as is indeed the practice, however imperfect, of all existing
democracies, various functions of government are subject to various forms of
democratic control, and some are and should not be democratically controlled
either by excluding them from the public domain or by insulating their public
governance from democratic influences. When considering reforms we should
not think of how to make our country more democratic, but which form of
governance would be adequate for this or that purpose, given our traditions
and conditions.
The same is true of international organisations. It may be the case that the
acceptance of the authority of international organisations that lack democratic
control will depend on their jurisdiction. All communications, by mail,
telephone, internet, etc. and all transportation, by land, sea and air, among
countries depend on treaties and organisations that oversee the existence and
adaption of common standards. Much of this is done by bureaucracies without

28
attracting the attention of domestic politicians or public opinion. In these
matters, governance by professionals and experts may be working well enough.
Democratisation, were it possible, may cause more harm than good. If states
were conscientious in securing civil rights, which they are not, we would not
need international bodies to pressurise states to secure them. In that happy
circumstance, the case for democratising the governance of international
institutions that oversee the implementation of civil rights may evaporate. And
we can expand on the interdependence of the case for democratic governance
domestically and internationally.
In other words, there is no such thing as a ‘democratic deficit’. We need
specific examination of the functions and structures of specific institutions,
conducted in light of the conditions in which they operate, to assess what, if
any, democratic elements are required or would improve their government.
Are complaints about the democratic deficit due entirely to a
misconceived understanding of democracy? Yes and No. No, because while
there are many flaws with the structures and functioning of international
bodies which are remote from our topic today, some of them result from the
failures of bodies like the UN or the EU to attract trust and loyalty, and to
profit from solidarity and mutual concern among the people subject to them,
attitudes that in many countries have come to be fostered domestically
through bonds encouraged by their democratic character. It is therefore
tempting to those of us formed in democratic cultures to think that
democratization of international organisations will remedy these absences. But
also Yes, not only because in principle there are other solutions, but because
democratisation is not one of them. Given their scale, the UN, the EU and
other transnational bodies cannot be improved by democratisation, as a
proper understanding of the success of democracy, where it does succeed,
does make clear.

29