This assignment is a take-home essay assignment of 2 questions, 2 pages each

malld614

The_Democracy_Sourcebook_----_6_-_REPRESENTATION.pdf

Home >Government homework help >This assignment is a take-home essay assignment of 2 questions, 2 pages each

6 REPRESENTATION

Representative Government John Stuart Mill

On Elections Marquis de Condorcet

Liberalism against Populism William H. Riker

Saving Democracy from Political Science Gerry Mackie

Unlikelihood of Condorcet’s Paradox in a Large Society A. S. Tangian

Congruence between Citizens and Policymakers in Two Visions of Liberal Democracy

John D. Huber and G. Bingham Powell, Jr.

The Political Consequences of Electoral Laws Douglas W. Rae

South Africa’s Negotiated Transition: Democracy, Opposition, and the New Constitutional Order

Courtney Jung and Ian Shapiro

The Representation of Women Anne Phillips

The Democracy Sourcebook, edited by Robert A. Dahl, et al., MIT Press, 2003. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=3338872. Created from apus on 2020-03-18 08:38:59.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Representative Government

John Stuart Mill

That the Ideally Best Form of Government Is

Representative Government

It has long (perhaps throughout the entire dura-

tion of British freedom) been a common saying,

that if a good despot could be ensured, despotic

monarchy would be the best form of govern-

ment. I look upon this as a radical and most

pernicious misconception of what good govern-

ment is; which, until it can be got rid of, will fa-

tally vitiate all our speculations on government.

The supposition is, that absolute power, in the

hands of an eminent individual, would ensure a

virtuous and intelligent performance of all the

duties of government. . . .

A good despotism means a government in

which, so far as depends on the despot, there is

no positive oppression by o‰cers of state, but in

which all the collective interests of the people are

managed for them, all the thinking that has re-

lation to collective interests done for them, and

in which their minds are formed by, and con-

senting to, this abdication of their own energies.

Leaving things to the Government, like leaving

them to Providence, is synonymous with caring

nothing about them, and accepting their results,

when disagreeable, as visitations of Nature. . . .

Such are not merely the natural tendencies,

but the inherent necessities of despotic govern-

ment; from which there is no outlet, unless in

so far as the despotism consents not to be des-

potism; in so far as the supposed good despot

abstains from exercising his power, and, though

holding it in reserve, allows the general business

of government to go on as if the people really

governed themselves. . . .

There is no di‰culty in showing that the ide-

ally best form of government is that in which

the sovereignty, or supreme controlling power

in the last resort, is vested in the entire aggregate

of the community; every citizen not only having

a voice in the exercise of that ultimate sover-

eignty, but being, at least occasionally, called on

to take an actual part in the government, by the

personal discharge of some public function, local

or general. . . .

The ideally best form of government, it is

scarcely necessary to say, does not mean one

which is practicable or eligible in all states of

civilisation, but the one which, in the circum-

stances in which it is practicable and eligible, is

attended with the greatest amount of beneficial

consequences, immediate and prospective. A

completely popular government is the only polity

which can make out any claim to this character.

It is pre-eminent in both the departments be-

tween which the excellence of a political consti-

tution is divided. It is both more favourable to

present good government, and promotes a better

and higher form of national character, than any

other polity whatsoever.

Its superiority in reference to present well-

being rests upon two principles, of as universal

truth and applicability as any general proposi-

tions which can be laid down respecting human

a¤airs. The first is, that the rights and interests of

every or any person are only secure from being

disregarded when the person interested is himself

able, and habitually disposed, to stand up for

them. The second is, that the general prosperity

attains a greater height, and is more widely dif-

fused, in proportion to the amount and variety

of the personal energies enlisted in promoting it.

Putting these two propositions into a shape

more special to their present application; hu-

man beings are only secure from evil at the

hands of others in proportion as they have the

power of being, and are, self-protecting; and they

only achieve a high degree of success in their

struggle with Nature in proportion as they are

Excerpted from: John Stuart Mill, Representative Gov-

ernment. New York: E. P. Dutton, 1972.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

self-dependent, relying on what they themselves

can do, either separately or in concert, rather

than on what others do for them. . . .

It is an adherent condition of human a¤airs

that no intention, however sincere, of protecting

the interests of others can make it safe or salu-

tary to tie up their own hands. Still more obvi-

ously true is it, that by their own hands only can

any positive and durable improvement of their

circumstances in life be worked out. Through

the joint influence of these two principles, all

free communities have both been more exempt

from social injustice and crime, and have at-

tained more brilliant prosperity, than any others,

or than they themselves after they lost their

freedom. . . .

It must be acknowledged that the benefits of

freedom, so far as they have hitherto been

enjoyed, were obtained by the extension of its

privileges to a part only of the community; and

that a government in which they are extended

impartially to all is a desideratum still unreal-

ised. But though every approach to this has an

independent value, and in many cases more than

an approach could not, in the existing state of

general improvement, be made, the participation

of all in these benefits is the ideally perfect con-

ception of free government. In proportion as

any, no matter who, are excluded from it, the

interests of the excluded are left without the

guarantee accorded to the rest, and they them-

selves have less scope and encouragement than

they might otherwise have to that exertion of

their energies for the good of themselves and of

the community, to which the general prosperity

is always proportioned.

Thus stands the case as regards present well-

being; the good management of the a¤airs of the

existing generation. If we now pass to the influ-

ence of the form of government upon character,

we shall find the superiority of popular govern-

ment over every other to be, if possible, still

more decided and indisputable. . . .

. . . The maximum of the invigorating e¤ect of

freedom upon the character is only obtained

when the person acted on either is, or is looking

forward to becoming, a citizen as fully privileged

as any other. What is still more important than

even this matter of feeling is the practical disci-

pline which the character obtains from the oc-

casional demand made upon the citizens to

exercise, for a time and in their turn, some social

function. It is not su‰ciently considered how

little there is in most men’s ordinary life to give

any largeness either to their conceptions or to

their sentiments. Their work is a routine; not a

labour of love, but of self-interest in the most

elementary form, the satisfaction of daily wants;

neither the thing done, nor the process of doing

it, introduces the mind to thoughts or feelings

extending beyond individuals; if instructive

books are within their reach, there is no stimulus

to read them; and in most cases the individual

has no access to any person of cultivation much

superior to his own. Giving him something to do

for the public, supplies, in a measure, all these

deficiencies. If circumstances allow the amount

of public duty assigned him to be considerable, it

makes him an educated man. . . .

. . . Still more salutary is the moral part of the

instruction a¤orded by the participation of the

private citizen, if even rarely, in public functions.

He is called upon, while so engaged, to weigh

interests not his own; to be guided, in case of

conflicting claims, by another rule than his pri-

vate partialities; to apply, at every turn, princi-

ples and maxims which have for their reason of

existence the common good: and he usually finds

associated with him in the same work minds

more familiarised than his own with these ideas

and operations, whose study it will be to supply

reasons to his understanding, and stimulation to

his feeling for the general interest. He is made to

feel himself one of the public, and whatever is for

their benefit to be for his benefit. . . .

From these accumulated considerations it is

evident that the only government which can fully

satisfy all the exigencies of the social state is

one in which the whole people participate; that

any participation, even in the smallest public

Representation 313

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

function, is useful; that the participation should

everywhere be as great as the general degree of

improvement of the community will allow; and

that nothing less can be ultimately desirable than

the admission of all to a share in the sovereign

power of the state. But since all cannot, in a

community exceeding a single small town, par-

ticipate personally in any but some very minor

portions of the public business, it follows that

the ideal type of a perfect government must be

representative.

Chapter 6 314

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

On Elections

Marquis de Condorcet

In a free society based on equality, accurate

popular choices are necessary not only for public

prosperity and the safety of the state, but also to

ensure the preservation of the basic principles

upon which that society is grounded. . . .

The Theory of Elections

For a man who chooses alone and who wants to

adhere to a strict procedure, an election is the

result of a series of judgements comparing all the

candidates two by two. The candidates to whom

he restricts his choice are, in this case, the men he

considers worthy of the place.

Similarly, an election vote is the result of those

judgements which obtain a majority. In each

judgement, an individual weighs up the reasons

for preference between two candidates. In an

election, each person’s votes for or against a

candidate represent these reasons, and are

counted rather than weighed up. When a man

compares two individuals and prefers the second

to the first, and then, on comparing the second

with a third, prefers the latter, it would be self-

contradictory if he did not also prefer the third

to the first. If, however, on making a direct

comparison of the first and the third, he found

reasons for preferring the first, he would then

have to examine this judgement, balance the

reasons behind it with those behind his other

judgements (which cannot exist alongside this

new one) and sacrifice the one he considers least

probable.

In an election between three candidates, the

three judgements which obtain majority support

when the candidates are compared two by two

may sometimes be inconsistent, even when the

individual judgements of each voter involve no

contradiction. This can easily be shown through

examples and can be explained by the simple fact

that the majority in favour of each of these ac-

cepted propositions is not made up of the same

individuals. We must therefore reject the propo-

sition with the smallest majority and retain the

other two. Thus, whenever we are able to obtain

only a fairly large probability, we must reject a

proposition which is probable in itself if it ex-

cludes another which is more probable.

An election result incorporating a proposi-

tion which runs contrary to one of the majority

judgements may still be very probable. In fact,

this only occurs in situations where it is certain

that the majority has been mistaken at least

once. The probability of the result being accurate

is then equal to the probability that the majority

has made only one mistake regarding a certain

proposition.

There is no need to point out that the con-

tradiction which can occur between majority

judgements when there are three candidates must

be even more likely when there are a greater

number of candidates; or that, in such cases,

several of these judgements may contradict one

another, meaning that the majority must neces-

sarily have made more than one mistake. The

consequences of this are the same.

The list of candidates who are put forward for

election must be determined if all the members

of an assembly are to give a complete vote by

comparing the same candidates two by two, or

to list them in order of merit, which comes to

the same thing. That is, everyone must know

the names of the candidates between whom the

voters will be choosing.

However, it is not necessary for everyone to

compare all the candidates or to form a complete

list. A voter may for various reasons regard a

certain number of candidates as equal to one

Excerpted from: Marquis de Condorcet, ‘‘On Elec-

tions.’’ In Condorcet: Foundations of Social Choice and

Political Theory. Brookfield, Vermont: Edward Elgar

Publishing, 1994. Reprinted by permission.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

another, either after considering their attributes

or because he does not know the candidates

and is therefore unable, or unwilling, to judge

them. . . .

This condition in no way restricts the voters’

freedom, since it simply requires everyone to de-

cide which candidates he wishes to choose be-

tween. The list of all those put forward in this

way would then present each voter with the

names of all the candidates between whom the

other voters wanted the election to be conducted,

and he would then have complete freedom to

decide how he could share in this judgement:

which candidates he wanted to rank in order of

merit and which he wanted to reject entirely by

placing them after all the others.

Any election method in which the votes given

are incomplete will produce results which con-

tradict the will which the majority would have

had if complete votes had been collected.

The results of these incomplete votes will of

course have some degree of probability of being

correct, but it would be similar to that of a

proposition which has been only half examined.

In fact, we should support a probable proposi-

tion only when we have discovered the impossi-

bility of incorporating new information, and as

long as this impossibility lasts.

However, we would be just as far from fulfill-

ing our aim if we forced each voter to express,

not the complete vote which he actually forms,

but a complete vote in an absolute sense; that is,

if we forced him to establish an order of prefer-

ence between all the candidates, including those

he does not know. Clearly, he would then rank

the latter at random and his vote could result in

the election of a candidate who would not oth-

erwise have had su‰cient support. In the first

case, we are neglecting judgements which should

have been assessed, and in the second, we are

assessing judgements which have not been given.

In the first case, we are acting as if we had ran-

domly excluded a certain number of voters, and

in the second as if we were randomly giving

some of them double the number of votes.

In theory, therefore an election procedure

should be as follows: after having determined the

list of acceptable candidates, each voter should

express his complete will, whether of preference

or indi¤erence.

A table of majority judgements between the

candidates taken two by two would then be

formed and the result—the order of merit in

which they are placed by the majority—ex-

tracted from it. If these judgements could not

all exist together, then those with the smallest

majority would be rejected.

This is exactly the same procedure as that fol-

lowed by any individual who wants to make a

considered choice by using a general, regular

method which applies to all situations.

Since there is only one way of obtaining a true

decision, the procedures used by a deliberative

assembly should generally be as close as possible

to those used when an individual examines a

question for himself.

This principle can have other important ap-

plications. In this case, it allows us to develop an

election method which is reasonably natural and

as perfect as the nature of things permits.

Any substitute method put forward should be

compared to this one in terms of the time re-

quired by an election, the degree of enlighten-

ment of those to whom it is entrusted, the

problems of outlawing the influence of intrigue,

and the need both to preserve true equality be-

tween the voters and to interest them in fulfill-

ing this function with scrupulous care and

enthusiasm. . . .

Chapter 6 316

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Liberalism against Populism

William H. Riker

5.A. Arrow’s Theorem

. . . Kenneth Arrow published Social Choice and

Individual Values in 1951. Although his theo-

rem initially provoked some controversy among

economists, its profound political significance

was not immediately recognized by political

scientists.1 In the late 1960s, however, a wide

variety of philosophers, economists, and politi-

cal scientists began to appreciate how pro-

foundly unsettling the theorem was and how

deeply it called into question some convention-

ally accepted notions—not only about voting,

the subject of this work, but also about the

ontological validity of the concept of social wel-

fare, a subject that, fortunately, we can leave to

metaphysicians.

The essence of Arrow’s theorem is that no

method of amalgamating individual judgments

can simultaneously satisfy some reasonable con-

ditions of fairness on the method and a condition

of logicality on the result. In a sense this theorem

is a generalization of the paradox of voting (see

section 1.H), for the theorem is the proposition

that something like the paradox is possible in

any fair system of amalgamating values. Thus

the theorem is called the General Possibility

Theorem.

To make the full meaning of Arrow’s theorem

clear, I will outline the situation and the con-

ditions of fairness and of logicality that cannot

simultaneously be satisfied.2 The situation for

amalgamation is:

1. There are n persons, n b 2, and n is finite.

Di‰culties comparable to the paradox of voting

can arise in individuals who use several stan-

dards of judgment for choice. Our concern is,

however, social choice, so we can ignore the

Robinson Crusoe case.

2. There are three or more alternatives—that is,

for the set X ¼ (x1; . . . ; xm), m b3. Since tran- sitivity or other conditions for logical choice are

meaningless for fewer than three alternatives and

since, indeed, simple majority decision produces

a logical result on two alternatives, the conflict

between fairness and logicality can only arise

when m b3.

3. Individuals are able to order the alternatives

transitively: If x Ri y and y Ri z, then x Ri z. If it

is not assumed that individuals are able to be

logical, then surely it is pointless to expect a

group to produce logical results.

The conditions of fairness are:

1. Universal admissibility of individual orderings

(Condition U). This is the requirement that

the set, D, includes all possible profiles, D, of

individual orders, Di. If each Di is some per-

mutation of possible orderings of X by prefer-

ence and indi¤erence, then this requirement is

that individuals can choose any of the possible

Reprinted by permission of Waveland Press, Inc. from

William H. Riker, Liberalism against Populism. Pros-

pect Heights, Ill.: Waveland Press, 1982 (reissued

1. William H. Riker, ‘‘Voting and the Summation of

Preferences,’’ American Political Science Review, Vol.

55 (December 1961), pp. 900–911.

2. Since many proofs are easily available to those who

wish to follow up the subject, I will not reiterate the

proof here. For Arrow’s proof, as revised by Julian

Blau, see Kenneth Arrow, Social Choice and Individual

Values, 2nd ed. (New Haven: Yale University Press,

1963), pp. 96–100. A refined form of Arrow’s proof is

to be found in Amartya K. Sen, Collective Choice and

Social Welfare (San Francisco: Holden-Day, 1970),

chap. 3. See also Peter C. Fishburn, The Theory of

Social Choice (Princeton: Princeton University Press,

1973), p. 206. Bengt Hansson, ‘‘The Existence of

Group Preference Functions,’’ Public Choice, Vol. 28

(Winter 1976), pp. 89–98, contains a topological proof.

An informal sketch of Arrow’s proof is contained in

William H. Riker and Peter C. Ordeshook, An Intro-

duction to Positive Political Theory (Englewood Cli¤s,

N.J.: Prentice-Hall, 1973), pp. 92–94.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

permutations. For example, if X ¼ (x; y; z), the individual may choose any of the following 13

orderings:

1. x y z 7. x (y z) 10. (x y) z 13. (x y z)

2. y z x 8. y (z x) 11. (y z) x

3. z x y 9. z (x y) 12. (z x) y

4. x z y

5. z y x

6. y x z (5-1)

The justification for this requirement is straight-

forward. If social outcomes are to be based

exclusively on individual judgments—as seems

implicit in any interpretation of democratic

methods—then to restrict individual persons’

judgments in any way means that the social out-

come is based as much on the restriction as it

is on individual judgments. Any rule or com-

mand that prohibits a person from choosing

some preference order is morally unacceptable

(or at least unfair) from the point of view of

democracy.

2. Monotonicity. According to this condition,

if a person raises the valuation of a winning al-

ternative, it cannot become a loser; or, if a per-

son lowers the valuation of a losing alternative,

it cannot become a winner. . . . Given the demo-

cratic intention that outcomes be based in some

way on participation, it would be the utmost in

perversity if the method of choice were to count

individual judgments negatively, although, as I

have shown, some real-world methods actually

do so.

3. Citizens’ sovereignty or nonimposition. De-

fine a social choice as imposed if some alter-

native, x, is a winner for any set, D, of individual

preferences. If x is always chosen, then what

individuals want does not have anything to do

with social choice. It might, for example, happen

that x was everyone’s least-liked alternative, yet

an imposed choice of x would still select x. In

such a situation, voters’ judgments have nothing

to do with the outcome and democratic partici-

pation is meaningless.

4. Unanimity or Pareto optimality (Condition

P). This is the requirement that, if everyone

prefers x to y, then the social choice function, F,

does not choose y. . . . This is the form in which

monotonicity and citizens’ sovereignty enter all

proofs of Arrow’s theorem. There are only two

ways that a result contrary to unanimity could

occur. One is that the system of amalgamation is

not monotonic. Suppose in D0 everybody but i prefers x to y and y P0i x. Then in D, i changes to x Pi y so everybody has x preferred to y; but, if

F is not monotonic, it may be that x does not

belong to F({x; y}; D). The other way a violation of unanimity could occur is for F to impose y

even though everybody prefers x to y. Thus the

juncture of monotonicity and citizens’ sover-

eignty implies Pareto optimality.

Many writers have interpreted the unanimity

condition as purely technical—as, for example,

in the discussion of the Schwartz method of

completing the Condorcet rule. . . . But Pareto

optimality takes on more force when it is rec-

ognized as the carrier of monotonicity and

nonimposition, both of which have deep and

obvious qualities of fairness.

5. Independence from irrelevant alternatives

(Condition I). According to this requirement

. . . , a method of amalgamation, F, picks the

same alternative as the social choice every time F

is applied to the same profile, D. Although some

writers have regarded this condition simply as a

requirement of technical e‰ciency, it actually

has as much moral content as the other fairness

conditions. . . . From the democratic point of

view, one wants to base the outcome on the

voters’ judgments, but doing so is clearly impos-

sible if the method of amalgamation gives dif-

ferent results from identical profiles. This might

occur, for example, if choices among alternatives

were made by some chance device. Then it is the

device, not voters’ judgments in D, that deter-

mines outcomes. Even if one constructs the de-

vice so that the chance of selecting an alternative

is proportional in some way to the number of

people desiring it (if, for example, two-thirds of

Chapter 6 318

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

the voters prefer x to y, then the device selects x

with p ¼ 2 3 ), still the expectation is that, of sev-

eral chance selections, the device will choose x

on p selections and y on 1 � p selections from the same profile, in clear violation of Condition

I. In ancient Greece, election by lot was a useful

method for anonymity; today it would be simply

a way to by-pass voters’ preferences. Another

kind of arbitrariness prohibited by the indepen-

dence condition is utilitarian voting. Based on

interpersonal comparisons of distances on scales

of unknown length, utilitarian voting gives

advantages to persons with finer perception and

broader horizons. Furthermore, independence

prohibits the arbitrariness of the Borda count. . . .

6. Nondictatorship (Condition D). This is the

requirement that there be no person, i, such that,

whenever x Pi y, the social choice is x, regardless

of the opinions of other persons. Since the whole

idea of democracy is to avoid such situations, the

moral significance of this condition is obvious.

Finally, the condition of logicality is that the

social choice is a weak order, by which is meant

that the set, X, is connected and its members can

be socially ordered by the relation, R, which is

the transitive social analogue of preference and

indi¤erence combined. (This relation, as in x R

y, means that x is chosen over or at least tied

with y.) In contrast to the previous discussion, in

which the method of amalgamation or choice, F,

simply selected an element from X, it is now

assumed that F selects repeatedly from pairs in X

to produce, by means of successive selections, a

social order analogous to the individual orders,

Di. And it is the failure to produce such an order

that constitutes a violation of the condition of

logicality.2

Since an individual weak order or the relation

Ri is often spoken of as individual rationality,

social transitivity, or R, is sometimes spoken of

as collective rationality—Arrow himself so de-

scribed it. And failure to produce social tran-

sitivity can also be regarded as a kind of social

irrationality.

Arrow’s theorem, then, is that every possible

method of amalgamation or choice that satisfies

the fairness conditions fails to ensure a social

ordering. And if society cannot, with fair meth-

ods, be certain to order its outcome, then it is

not clear that we can know what the outcomes

of a fair method mean. This conclusion appears

to be devastating, for it consigns democratic

outcomes—and hence the democratic method—

to the world of arbitrary nonsense, at least some

of the time.

5.B. The Practical Relevance of Arrow’s

Theorem: The Frequency of Cycles

. . . One meaning of Arrow’s theorem is that,

under any system of voting or amalgamation,

instances of intransitive or cyclical outcomes

can occur. . . . Since, by definition, no one of the

alternatives in a cycle can beat all the others,

there is no Condorcet winner among cycled

alternatives. All cycled alternatives tie with re-

spect to their position in a social arrangement in

the sense that x y z x, y z x y, and z x y z have

equal claims to being the social arrangement.

Borda voting similarly produces a direct tie

among cycled alternatives. Hence a social ar-

rangement is indeterminate when a cycle exists.

When the arrangement is indeterminate, the

actual choice is arbitrarily made. The selection is

not determined by the preference of the voters.

Rather it is determined by the power of some

chooser to dominate the choice or to manipu-

late the process to his or her advantage. Every

cycle thus represents the failure of the voting

process. . . .

5.G. The Absence of Meaning

The main thrust of Arrow’s theorem and all the

associated literature is that there is an unresolv-

able tension between logicality and fairness. To

guarantee an ordering or a consistent path, in-

Representation 319

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

dependent choice requires that there be some

sort of concentration of power (dictators, oligar-

chies, or collegia of vetoers) in sharp conflict

with democratic ideals. Even the weakest sort of

consistency (bþ or �PI ) involves a conflict with unanimity, which is also an elementary condition

of fairness.

These conflicts have been investigated in great

detail, especially in the last decade; but no ade-

quate resolution of the tension has been dis-

covered, and it appears quite unlikely that any

will be. The unavoidable inference is, therefore,

that, so long as a society preserves democratic

institutions, its members can expect that some of

their social choices will be unordered or incon-

sistent. And when this is true, no meaningful

choice can be made. If y is in fact chosen—given

the mechanism of choice and the profile of indi-

vidual valuations—then to say that x is best or

right or more desired is probably false. But it

would also be equally false to say that y is best

or right or most desired. And in that sense, the

choice lacks meaning. . . .

Chapter 6 320

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Saving Democracy from Political Science

Gerry Mackie

Although democratization is the main trend in

the world today, the main intellectual trend in

American political science is the view that de-

mocracy is chaotic, arbitrary, meaningless, and

impossible. This trend originated with economist

Kenneth Arrow’s (1963) impossibility theorem,

which was applied to politics by the late Wil-

liam Riker, political scientist at the University

of Rochester. Riker’s famed Liberalism against

Populism (1982) claims that elections serve only

to randomly remove o‰cials. Define populism

(ignore its pejorative usages) as the doctrine that

political outcomes are or should be related in

some way to the collective opinion or will of the

free and equal citizenry. Riker, and his followers

in the so-called Rochester School of political

thought, are, in this special sense, antipopulists.

Because di¤erent voting systems may yield

di¤erent outcomes from the same profile of indi-

vidual voters’ preferences, says Riker, ‘‘populist

democracy’’ is inaccurate. The claim is made by

way of examples that establish the logical possi-

bility of such divergence. It turns out, however,

that with real-world distributions of individual

preference orders—individual orderings that are

at least mildly correlated with one another rather

than unrelated and random—the reasonable

voting rules do tend to converge on the same

outcomes. Next, Riker continues, given a fixed

voting system, then democracy is meaningless:

the outcome of voting is manipulable, and it

is not possible to distinguish manipulated

from unmanipulated outcomes because of the

unknowability of private intentions underlying

public actions. Riker’s claim that democracy is

irremediably manipulable is based on the as-

sumption that cycles are ubiquitous on impor-

tant issues.

Suppose that the Duckburg Troop of the

Junior Woodchucks are deciding by majority

voting how to spend their treasury over three

alternatives, as in table 6.1. Huebert and Louis

favor A over B, Huebert and Deuteronomy

favor B over C, and Deuteronomy and Louis

favor C over A. The collective choice cycles over

A > B > C > A. It seems that there is no collec- tive choice. Moreover, if only Huebert controlled

the order of voting, he could manipulate the

outcome to his first-ranked A, by declaring first a

majority vote between B and C, and then the

winner of that against A. Arrow’s possibility

theorem can be understood as a generalization of

this Condorcet paradox of voting. Cycles are

logically possible, but are they empirically prob-

able? Data on real individual preference rank-

ings show that they are distributed in a manner

such that cycles are rare, and theoretical consid-

erations predict this finding. Unfairly manipu-

lated outcomes are possible only if cycles are a

problem, and only if some political actors are

unfairly granted powers not granted to others.

That a mathematical model shows that those

with greater formal power may have greater in-

fluence over the outcome should come as no

surprise. Contrary to Riker, it does not establish

that there is a problem in principle with democ-

racy, which requires equal formal powers.

It is Riker’s dramatic empirical illustrations

of political disequilibrium, more than his theo-

retical arguments, that are responsible for the

wide popularity of his antipopulist views. I have

closely examined all of the well-developed em-

pirical claims of cycling and democratic irratio-

nality published by Riker and his followers, and

I claim to show that each is mistaken, either

conceptually or empirically (Mackie, forthcom-

ing). These case studies include a cycle alleged

in Congressional agricultural appropriations,

agenda control in a flying club, a cycle arising

from Powell’s desegregation amendment in the

1956 Congress, a cycle in votes in the Senate

over proposing the 17th amendment to the U.S.

Constitution, a cycle in consideration of the

Wilmot Proviso in 1846, a cycle over candidates

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

for the 1860 presidential election, introduction of

the slavery issue in order to arbitrarily manipu-

late political outcomes, a cycle in consideration

of the method for selecting the U.S. president

at the Constitutional Convention, Lincoln’s al-

leged ‘‘heresthetic’’ manipulation of the Lincoln-

Douglas debates, Magnuson’s alleged heresthetic

manipulation of the nerve-gas issue, and further

cases not from Riker’s hand. If these illustrations

fail, then so does the antipopulist creed.

It’s not possible in the space at hand to justify

my claimed findings. I present an abridged sam-

ple, however, having to do with Riker’s most

startling claim, that there was a cycle in the

presidential election of 1860, and that choice

among the candidates in that election was, in

principle, arbitrary and meaningless. Riker’s

overarching hypothesis is that the slavery di-

mension of concern in antebellum American

politics was suppressed by the Democratic Party

manipulative elite with the Missouri Compro-

mise of 1820. The competing Whig and Repub-

lican manipulative elite sought to revive that

dimension by contriving a cycle, and succeeded

spectacularly with the election of Lincoln in

1860, the story goes.

There were four major candidates in the

American presidential election of 1860. They

were Lincoln of the recently formed Republican

Party, Douglas of the northern Democrats, Bell

from a new third party, the Constitutional

Unionists, and Breckinridge of the renegade

southern Democrats. Lincoln received a plurality

of 40 percent of the popular vote, and Riker

suspects a cycle. From election data we have

only voters’ expression of first-ranked prefer-

ence, not what their second-, third-, and fourth-

ranked preferences were. In order to conduct his

analysis, Riker must infer voters’ full rankings

from historical considerations. He provides esti-

mates of voters’ full rankings over 15 likely of

the 24 rankings possible for strong preferences

over four candidates. His estimations then per-

mit the calculation of hypothetical outcomes by

alternative voting rules. The results of his exer-

cise seem to confirm Riker’s thesis that democ-

racy is arbitrary and meaningless. Di¤erent

voting rules lead to di¤erent outcomes, and by

pairwise majority voting there is a cycle. Here

are Riker’s results:

Plurality: Lincoln > Douglas > Breckinridge > Bell

Pairwise Majority: (Douglas > Lincoln > Bell > Douglas) > Breckinridge

Borda Count: Douglas > Bell > Lincoln > Breckinridge

Approval Voting (two votes): Bell > Lincoln > Douglas > Breckinridge

Approval Voting (three votes): Douglas > Bell > Lincoln > Breckinridge

Riker concludes that with five methods of voting

Douglas wins twice, Bell once, Lincoln once, and

they are in a cycle and hence tie once. ‘‘Clearly,

if my guesses are even roughly right, there was

complete disequilibrium in 1860’’ (1982, 229).

Riker’s first two rankings have to do with all

those voters whom we know ranked Lincoln

first. The two rankings contain 40 percent of the

voting population of the entire country, so if

there is a major error among them then we need

go no further. Reconstructing Riker’s data, he

has it that about 1,794,000 voters in the free

north voted for Lincoln, and he is correct

enough on that. His further claim is that of

those, one-fourth or about 450,000 ranked

Douglas second, and three-fourths or about

1,346,000 ranked Bell second. I claim that it is

the reverse, that most Lincoln voters ranked

Douglas second, and few ranked Bell second.

Table 6.1

Condorcet paradox of voting

Huebert Deuteronomy Louis

1st A B C

2nd B C A

3rd C A B

Chapter 6 322

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

The major controversy in the country was the

extent to which slavery would be allowed in the

vast territories, the so-called ‘‘free soil’’ question.

Lincoln was most free-soil, then Douglas, then

Bell, then Breckinridge (who was most pro-

slavery). If the slavery dimension was of greatest

concern to voters, then one would expect that

Douglas would be the second choice of most

Lincoln voters, not Bell. Riker does not justify

his estimates, but presumably his inference that

Bell was the second choice of most Lincoln

voters was based on the thought that the Con-

stitutional Unionists contained many southern

members of the defunct Whig Party, and that

northern former Whigs whose first choice was

Lincoln would second-rank Bell from the Whig-

gish Constitutional Unionists rather than Doug-

las from their ancient enemy the Democratic

Party. That did not happen, and to understand

why we have to look at what the parties stood

for, with special attention to the Constitutional

Unionists.

The parties’ platforms reflect the positions I

suggest, with the exception of the Constitutional

Unionists. Their brief platform is a platitudinous

appeal to Constitution, Union, and Enforcement

of the Laws. Because of this, and their poor

showing in the election, some commentators ne-

glect to locate them ideologically. A close read-

ing of their platform shows, however, that their

pledge of continued union is conditioned on

the reestablishment of states’ rights and equal-

ity between the sections as found in the origi-

nal constitutional compromise. It is the original

constitutional compact that commands their

loyalty, not present-day misinterpretations or

usurpations by the north. Outside the platform,

the Constitutional Unionists believed that Con-

gress must protect the property rights of slave-

owners in the territories, the southern position

on free soil. They denounced the Breckinridge

Democrats for threatening secession; but did not

deny the right to secede should compromise with

the north fail. The southern former Whigs who

were the largest faction in the organization of

the Constitutional Unionists aspired to attract

northern former Whigs, but they largely failed,

because, although more moderate than Breck-

inridge, they did not appeal to northern senti-

ments on free soil. The southern orientation of

the Constitutional Unionists was no secret to

northern voters: the party received four percent

of their vote.

A look at a map of the county-level winners in

the 1860 presidential elections (McPherson 1993,

128) shows that latitude is attitude. The upper

north voted for Lincoln (more free soil), the

lower north voted for Douglas (less free soil), the

upper south voted for Bell (less slave soil), and

the lower south voted for Breckinridge (more

slave soil). Why would a voter in the upper north

choose Lincoln? High among the reasons must

be Lincoln’s appeal to northern interests includ-

ing his position on free soil; otherwise one of the

other candidates would do. If such a voter were

denied the choice of Lincoln, who would his

second choice be? Why not Douglas, who is the

next most adjacent on issues of northern interest?

And if Douglas were not available where next

would the typical upper northern voter turn?

He would turn to Bell, the candidate of the

upper south, of course, not to Breckinridge the

candidate of the lower south. Similar reasoning

would predict the ranking by the typical voter in

the lower north, the upper south, and the lower

south.

If voter preferences were ideologically oriented

to the free soil question, then we would expect

upper northern voters to rank L > D > Bl > Br. Aggregate data are suspect because of the mis-

named ‘‘ecological fallacy.’’ Nevertheless, such

data are instructive, especially if they support a

position otherwise supported by independent

considerations. Aggregates at the state level

show that in the upper north Lincoln was first

everywhere and Douglas generally finished sec-

ond. The ideological approach would predict

that Bell would finish third in the northern states,

but he finishes third in three and fourth in eight

states. This is the candidate that Riker would

Representation 323

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

have us believe is ranked second by northern

Lincoln voters. In the upper north, 55 percent of

voters favored Lincoln, 35 percent Douglas,

eight percent Breckinridge, and two percent Bell,

L > D > Br > Bl. Aggregates at the state level for the lower north, the upper south, and the

lower south support the view that slavery was

the primary dimension of concern.

After my first analysis of the issue, Tabarrok

and Spector (1999) published an intriguing and

more technically sophisticated analysis of the

1860 election. Among other things, they queried

historians of the Civil War as to their views on

the rankings, and 15 responded. For voters who

ranked Lincoln first, the historians’ median

opinion was that 60% ranked Douglas second

and 40% ranked Bell second. Tabarrok and

Spector constructed a median historian’s pro-

file, and compared it to Riker’s. The plane of

possible outcomes in Tabarrok and Spector’s

positional-vote space, using Riker’s profile, is

large, and permits a large variety of rankings by

way of varying the weights of a positional voting

rule—from plurality to Borda count to anti-

plurality and everything in between. The posi-

tional vote plane using the median historian’s

profile is small, and ‘‘Douglas wins . . . under any

positional voting system which gives significant

weight to second- or second- and third-ranked

preferences’’ (278); but Lincoln wins by plurality

count, which gives no weight to any preferences

except for the first-ranked. Pairwise majority

voting is cyclic with Riker’s profile, but with the

median historian’s profile Douglas beats Bell

by 15.1%, Bell beats Lincoln by 0.033% (three-

hundredths of one percent), and Lincoln beats

Breckinridge by 25.4%.

The platforms, aggregate voting data, and the

opinion survey of historians support the hypoth-

esis that Douglas and not Bell was second-

ranked among most Lincoln voters. What does

this revision do to Riker’s demonstration that

democracy is arbitrary because di¤erent voting

methods result in di¤erent outcomes and that

democracy is meaningless because cycles will be

contrived on major issues such as the future of

slavery? In the upper and lower north, Lincoln

won 53% of the vote, Douglas 36%, and Bell 4%.

Riker has 25% of Lincoln voters ranking Doug-

las second, and 75% ranking Bell second. The

median historians’ estimate is 60% Douglas sec-

ond, and 40% Bell second. Suppose that at least

37% ranked Douglas second, and at most 63%

ranked Bell second. What happens then to the

arbitrarily di¤erent outcomes and the meaning-

less cycle? They disappear! Here are the new

outcomes:

Pairwise Majority: Douglas > Lincoln > Bell > Breckinridge

Borda Count: Douglas > Lincoln > Bell > Breckinridge

Approval Voting (two votes): Douglas > Lincoln > Bell > Breckinridge

Approval Voting (three votes): Douglas > Bell > Lincoln > Breckinridge

Moreover, with plurality runo¤, Lincoln and

Douglas would go to the runo¤ and Douglas

would win (a runo¤ between first-round losers

Bell and Breckinridge has Bell as the winner,

yielding the overall ranking D > L > Bl > Bk). All five of these voting methods select Douglas

as the winner. Except for approval voting with

three votes, the rules identify the same ranking:

D > L > Bl > Bk. A centrist such as Douglas should have won

the 1860 election. That one didn’t is due to the

antidemocratic design of the electoral college,

and due to the spoiler strategy of the Breck-

inridge faction, which perversely sought a Lin-

coln victory in order to justify secession. The

election of Lincoln was not due to a hidden flaw

in the ideal of democracy.

The Arrow theorem disclosed the logical pos-

sibility of a majority cycle, of perpetual political

instability. But we observe stability rather than

instability in democratic politics. Riker origi-

nally responded that cycles are common but

rarely detected. Simulations show and empirical

Chapter 6 324

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

studies corroborate, however, that cycling is an

empirical improbability. Riker later conceded

that uncontrived cycles are quite rare, but that

on major issues actors will contrive cycles by in-

troduction of new alternatives. In order to be

persuaded to abandon the concept of the public

good and the idea of democracy as in some sense

the expression of the people’s will, most people

would require that it be robustly demonstrated

that the manipulation of political outcomes is

frequent, harmful, and irremediable. Riker’s po-

sition is that it is either theoretically impossible

or empirically di‰cult to detect such manipula-

tion. He is able, however, to produce spectacular

anecdotes that purport to show harmful manip-

ulation on major issues, including a demonstra-

tion that the biggest event in American history

was the consequence of a contrived cycle. I have

worked through each of his examples only to

find that each is mistaken. Theoretical consid-

erations about the distribution of preference

orders suggest that cycles are most unlikely; em-

pirical studies show that cycles are of no practi-

cal importance; finally, every developed and

published example of a political cycle has now

been challenged. Thus, I argue that after fifty

years of scholarship, from the first publication of

Arrow’s theorem, no one has satisfactorily dem-

onstrated the existence of a normatively trou-

bling cycle in the real world.

The antipopulist interpretations of social

choice results inspired by Riker, widely endorsed

in the discipline of political science, are unsup-

ported by evidence and must be abandoned.

References

Arrow, Kenneth J. 1963/1951. Social Choice and Indi-

vidual Values, second edition. New Haven: Yale Uni-

versity Press.

Mackie, Gerry. Forthcoming. Democracy Defended.

Cambridge: Cambridge University Press.

McPherson, James M. 1993. Ordeal by Fire, vol. 1, The

Coming of War. New York: McGraw-Hill.

Riker, William H. 1982. Liberalism Against Populism:

A Confrontation between the Theory of Democracy and

the Theory of Social Choice. Prospect Heights: Wave-

land Press.

Tabarrok, Alexander and Lee Spector. 1999. ‘‘Would

the Borda Count Have Avoided the Civil War?’’ Jour-

nal of Theoretical Politics 11: 261–288.

Representation 325

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Unlikelihood of Condorcet’s Paradox in a Large Society

A. S. Tangian

1 Introduction

1.1 Social Choice with Respect to Cardinal and

Ordinal Utility

As early as in the last quarter of the 18th cen-

tury, Borda, Condorcet, and Laplace mathe-

matically investigated election by vote (see Black

1958; Young 1988; McLean and Urken 1994).

Borda described cases when voting implies

intuitively unacceptable results. Condorcet dis-

covered the intransitivity of majoritarian prefer-

ence, the fact known as Condorcet’s paradox.

Both authors developed new methods to over-

come this problem with voting, which posed

further problems. Laplace provided some math-

ematical arguments to back up Borda’s model.

These studies, having questioned the faithfulness

of election systems, have had numerous succes-

sors and thus have given birth to a new scientific

branch, the social choice theory.

Borda and Condorcet founded two di¤erent

standpoints, which continue to be the two prin-

cipal approaches to the problem. Borda’s count

is based on summing ranks of candidates in in-

dividual schedules, which are cardinal utilities.

Laplace considered a model with real utilities as

a logical development of Borda’s model. Con-

dorcet’s count uses pairwise votes, which are or-

dinal data. . . .

The ordinal approach, including a majority

rule, is usually considered to be ‘more objective’

and better substantiated than the cardinal one.

On the other hand, the intransitivity of prefer-

ence discovered by Condorcet is generally inher-

ent in all ordinal models. Thus having a rigorous

methodological basis, the ordinal approach lacks

the universal solvability inherent in the cardinal

approach. In a general formulation, the unsolv-

ability of the ordinal approach to social choice

has been proved by Arrow (1951).

However, contrary to Condorcet’s predictions,

the intransitivity of social preference is seldom

observed in real voting. Moreover, all demo-

cratic systems are based on di¤erent forms of

voting, ignoring the unsolvability of the ordinal

approach. So we have an apparent inconsistency

between theory and practice.

In this paper, we analyze the cause of this

inconsistency. It turns out that Condorcet’s pre-

dictions are true for a quite particular combina-

torial case, when an ‘average voter’ votes for or

against certain candidates with equal probabil-

ity. The related analytical studies have been done

by DeMeyer and Plott (1970), May (1971),

Merril and Tideman (1991), Niemi and Weisberg

(1968), Garman and Kamien (1968), Gehrlein

and Fishburn (1976), and Gehrlein (1983). The

results of computer simulation have been re-

ported by Campbell and Tullock (1965), Klahr

(1966), Jones et al. (1995), and Van Deemen

(1996).

The assumption of equal probability of voting

for or against candidates means that these can-

didates have equal merits, a circumstance which

is possible theoretically, but can hardly ever oc-

cur. Much more frequently, candidates have dif-

ferent merits, reflected by the vote asymmetry.

We show that in such a case the probability of

intransitive social preference vanishes as the

number of voters increases. From this we con-

clude that it is very unlikely to obtain Con-

dorcet’s paradox in a real vote. . . .

. . . We prove with a probability model that

group decisions based on a pairwise majority

Editor’s note: Tangian develops some lengthy and

technical proofs to support his findings, which are not

included here. For our purposes, it su‰ces to provide

an explication of his methods and a summary of his

results.

Excerpted from: A. S. Tangian, ‘‘Unlikelihood of

Condorcet’s Paradox in a Large Society.’’ Social

Choice and Welfare 17 (2000): 337–365. Reprinted by

permission.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

vote or on summing individual cardinal utilities

(in the form of either reals, or integers, or ranks)

tend to be the same as the number of individ-

uals increases. The only exception is the case

when the vote ratio for pairs of alternatives is

50%:50%. Since such an equilibrium is rather

unlikely, we conclude that in most practical sit-

uations pairwise vote count and utility count are

almost equivalent.

Let us explain the main idea at an intuitive

level, in terms of tournaments. Imagine a chess

competition with three participants A, B, and C.

Due to chance, a weaker player can gain a vic-

tory over a stronger one, and the preference

order on the set of players can have cycles like

A � B � C � A. However, in multiple rounds, a stronger player wins more often, getting a higher

total score of victories over weaker players. After

a certain number of rounds, the preference order

on the set of participants based on paired games

will correspond to their real strength rating.

(Recall that each chess professional is assigned a

rating calculated with respect to his results in

several tournaments.)

A pairwise vote for candidates A, B, C in a

small group of voters is like a single tournament

round with a risk of misrepresentation of real

merits of the participants. This can result in in-

transitive cycles in the preference order. A vote

in a large society, due to the error-vanishing ef-

fect, is similar to a multiple-round tournament

where the merits of participants are revealed

more accurately. If social merits of candidates

are adequately reflected by the vote in an elec-

tion, then the (latent) merit scale linearizes

the resulting preference, as the rating scale lin-

earizes the order of players in a multiple-round

tournament.

The idea of the underlying ‘‘objective’’ prefer-

ence being distorted in individual preferences

goes back to Condorcet who has assumed prob-

abilistic deviations from some ‘‘correct’’ vote.

From this standpoint, the intransitivity of the

majoritarian preference is no longer a paradox

but rather an error to be corrected—the argu-

ment exploited by Condorcet and some other

scholars.

Our approach can be explained with a refer-

ence to Black’s (1958) model with single-peaked

preferences. He proved the transitivity of a ma-

jority preference, having assumed an alignment

of candidates such that each elector has a single-

peaked preference with respect to this alignment.

In our model, such an alignment is the latent

social merit scale. In a large model, the devia-

tions from this scale in individual preferences can

be considered as ‘‘errors’’ which become negligi-

ble as the number of individuals increases.

Note that, unlike Black, we do not restrict

the domain of preference profiles (combinations

of individual preferences). We consider all pro-

files under some probabilistic assumptions (in-

dependence of voters, inequality of the mean

preference probability from 0.5, that is no

equilibrium of votes). Then we show that the

probability of the profiles which result in an

intransitive social preference vanishes as the

number of voters increases. We conclude that to

avoid Condorcet’s paradox in a large society,

there is no need to introduce any domain re-

striction, since the domain ‘restricts itself’ almost

for sure in the probabilistic sense under rather

weak assumptions. That is, the profiles resulting

in an intransitive social preference constitute an

event with vanishing probability. . . .

In intermediate reasoning we make use of in-

dividual cardinal utilities. We assume that they

do exist and have some probability distribution,

but do not require their expressibility. This as-

sumption is not crucial, since the question of ex-

istence of probability measures is trivial for finite

models, which is our case (the number of indi-

viduals is finite, and individual utilities can be

restricted to some finite grid).

Since the limit theorems of the paper are valid

for arbitrary distributions, the results hold even

if these distributions are not known. Here we

follow the same line of reasoning as (Tanguiane

1991, 1994) while revising Arrow’s paradox: We

prove a property for every measure and then

Representation 327

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

conclude that knowing the measure is not neces-

sary to be sure that this property holds.

. . . [W]e consider a model with two alter-

natives. We prove that the decision obtained by

voting and that obtained by summing individual

cardinal utilities become the same as the number

of individuals increases. The two decisions con-

tradict each other with a significant probability

only if the probability of vote for one candidate

against another is equal to 0.5. This case corre-

sponds to precisely equal social merits of the

candidates which is a rare coincidence.

Then, we extend this result from a pair of

alternatives to several alternatives. This gives the

main statement of the paper: The social prefer-

ences based on ordinal or cardinal counts tend to

become the same as the number of individuals

increases. As a corollary, in a model with a large

number of voters, the transitivity of the Borda

preference is inherent in the preference based on

pairwise vote.

. . . [W]e provide numerical evidence of the

coincidence of results from voting and from

summing individual utilities. The speed of con-

vergence of the two methods of social decision

making depends on several parameters. We clas-

sify distributions of individual cardinal utilities

with respect to these parameters and interpret

these classes with regard to the way of social

behavior. . . .

. . . [I]n most practical cases (not simulta-

neously vote symmetry and indi¤erent social

response) decisions based on voting and on

summing individual utilities provide close re-

sults. For a large society, reservations should be

made only for the case of symmetric vote in an

indi¤erent society.

It should be emphasized that Condorcet’s

paradox remains relevant for small groups. Since

our model assumes the probabilistic indepen-

dence of individuals, a large society means a

large number of independent individuals. A few

large coalitions with members sharing the same

preference do not meet this condition. Such a

model is a model with a few participants with the

weight coe‰cients, corresponding to the size of

coalitions.

Independence may be regarded as the most

consistent realization of the idea of probability.

Therefore, our results relate with reservations to

a nonindependent probabilistic model, as dem-

onstrating some kind of a general trend.

Besides, our results illustrate the opinion that

cardinal measurements are not very important in

processing ordinal data. In fact, we show that

voting leads to the same results as summing car-

dinal utilities, implying that most essential infor-

mation is already inherent in ordinal preferences.

Finally, it follows that the inaccuracy of cer-

tain election procedures should not play a deci-

sive role in case of a large society. In particular,

a majority vote will reveal the same preference

as more sophisticated and ‘‘precise’’ methods of

election that are based on cardinal utilities.

For instance, consider the presidential elec-

tion in Russia for Jelzin against Sjuganow on

June 16, 1996. In spite of a small vote asymme-

try (34% and 32% of votes, respectively) we can

expect that the social response was not indi¤er-

ent (since the political programs of the candi-

dates di¤ered significantly). From our standpoint

we can be quite sure that even a hypothetical

voting system which takes into account the de-

gree of individual preferences would give the

same election result.

In conclusion we summarize the main state-

ments of the paper.

1. Our model is not based on any domain re-

striction on the individual preference profiles

implying the transitivity of a majority prefer-

ence. We consider all profiles but show under

rather weak assumptions that the probability of

an intransitive collective preference is vanishing

as the number of individuals increases (Con-

dorcet has considered equiprobable collective

preferences which, as one can see, is not relevant

to real social choice).

Chapter 6 328

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

2. For this purpose we assume the following

. Independence of individuals (following Con- dorcet, Laplace),

. Independence of ordinal and cardinal parts in individual preferences (following Borda,

Laplace),

. Asymmetry of pairwise votes (Tangian). 3. In a large society, a di¤erent choice with re-

spect to voting or summing individual utilities is

statistically not significant. The solvability of a

large ordinal model of social choice (transitivity

of a majoritarian preference) follows from the

existence of the dual cardinal model, not requir-

ing its explicit knowledge.

4. In a small group, a di¤erent choice with re-

spect to voting and summing individual utilities

is statistically significant. Testing the di¤erence

requires additional information about the distri-

bution of preference probabilities.

5. A di¤erent choice with respect to voting and

summing individual utilities in caused by the

following:

. a high deviation coe‰cient s=m (since s < 1, the expectation m must be small, that is, individ-

uals are indi¤erent);

. almost symmetrical votes ( p is close to 1/2). 6. One should not worry about the imperfection

of election systems based on voting:

. The risk that an intransitive preference emerges is negligible.

. Election based on voting is almost as accurate as complex systems with cardinal utilities. . . .

References

Arrow KJ (1951) Social choice and individual values.

Wiley, New York.

Campbell CD, Tullock G (1965) A Measure of the

Importance of Cyclical Majorities. Economic Journal

75: 853–856.

DeMeyer F, Plott Ch (1970) The Probability of a

Cyclic Majority. Econometrica 38: 345–354.

Garman M, Kamien M (1965) The Paradox of Voting:

Probability Calculations. Behavioral Science 10: 306–

317.

Gehrlein WV (1983) Condorcet’s Paradox. Theory and

Decision 15: 161–197.

Gehrlein WV, Fishburn PC (1976) The Probability of

the Paradox of Voting: A Computable Solution. Jour-

nal of Economic Theory 13: 14–25.

Jones B, Radcli¤ B, Taber Ch, Timpone R (1995)

Condorcet Winners and the Paradox of Voting: Prob-

ability Calculations for Weak Preference Orders.

American Political Science Review 89: 113–144.

Klahr D (1966) A Computer Simulation of the Para-

dox of Voting. American Political Science Review 60:

384–390.

May RM (1971) Some Mathematical Remarks on the

Paradox of Voting. Behavioral Science 16: 143–151.

McLean I, Urken AD (eds.) (1994) Classics of Social

Choice. University of Michigan Press, Ann Arbor.

Merril, Tideman (1991) The Relative E‰ciency of Ap-

proval and Condorcet Voting Procedures. Rationality

and Society 3: 65–77.

Niemi RG, Weisberg H (1968) A Mathematical Solu-

tion for the Probability of the Paradox of Voting. Be-

havioral Science 13: 317–323.

Tanguiane AS (1991) Aggregation and Representation

of Preferences. Introduction to Mathematical Theory of

Democracy. Springer, Berlin Heidelberg New York.

Tanguiane AS (1994) Arrow’s Paradox and Mathe-

matical Theory of Democracy. Social Choice and Wel-

fare 11(1): 1–82.

Young HP (1988) Condorcet’s theory of voting.

American Political Science Review 82: 1231–1244.

Representation 329

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Congruence between Citizens and Policymakers in Two Visions of Liberal Democracy

John D. Huber and G. Bingham Powell, Jr.

Liberal democracy claims to establish connec-

tions between citizens and policymakers. The re-

peated processes of electoral competition and

legislative bargaining are supposed to ensure that

policymakers do what citizens want them to do.

There are, however, at least two quite di¤erent

visions of the democratic processes that can cre-

ate congruence between citizen preferences and

public policies.

In what we call the Majority Control vision,

democratic elections are designed to create

strong, single-party majority governments that

are essentially unconstrained by other parties in

the policy-making process. Policymakers are

likely to do what citizens want them to do be-

cause the party that controls the government has

won majority support in the election. Its an-

nounced policy comments, previous record, or

both were preferred to the partisan alternative by

a majority of the citizens. In the other vision,

which we call the Proportionate Influence vision,

elections are designed to produce legislatures

that reflect the preferences of all citizens. After

the election legislative bargaining between

parties is necessary for policymaking, and the

influences of the various parties in post-election

bargaining processes determine the extent to

which policymakers do what citizens want them

to do.

In this paper, we o¤er a simple way to con-

ceptualize the degree of congruence between

citizens and their governments, comparing citi-

zen self-placements on the left-right scale with

the placement of the governing political parties

on the same scale by expert observers. We then

attempt to give explicit theoretical form to the

Majority Control and Proportionate Influence

visions, to link them empirically to specific types

of modern democracies, and to measure their

successes and failures at creating congruence. We

want to know in particular how such theoreti-

cally critical features as responsible incumbent

governments at the time of the election, identifi-

able future governments in electoral competition,

proportional representation in electoral out-

comes, and the formation of majority govern-

ments after the election are related to levels of

congruence.

Congruence, of course, is not the only demo-

cratic virtue: some of the processes treated here

as intervening may be highly valued in their

own right.1 Voters oriented to control may

wish to see government formations that change

in response to even small vote shifts. Voters

may prefer to have very distinctive choices.

Voters may prefer that policy-making be highly

e‰cient. Permanent minorities may prefer pro-

portionate representation and consultative legis-

lative bargaining, especially if other processes

directly impose the preferences of the majority.

We therefore do not propose that congruence

between citizen preferences and public policy

should be the only grounds for choosing or sup-

porting one vision over the other. We do think,

however, that congruence between the prefer-

ences of citizens and the actions of policymakers

constitutes a major claim and goal of liberal de-

mocracy. Thus, Dahl’s ‘‘reasonable justification

for democracy’’ posits that ‘‘a majority of citi-

zens can induce the government to do what they

most want it to do and to avoid doing what they

Excerpted from: John D. Huber and G. Bingham

Powell, Jr., ‘‘Congruence between Citizens and Policy-

makers in Two Visions of Liberal Democracy.’’ World

Politics, 46 (3) (Apr., 1994), 291–326. 6 1994 The Johns Hopkins University Press. Reprinted with the

permission of The Johns Hopkins University Press.

1. Some of these other virtues are more fully described

and elaborated in G. Bingham Powell, Jr., ‘‘Elections

as Instruments of Democracy’’ (Manuscript, Univer-

sity of Rochester, 1993).

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

most want it not to do.’’2 This is not a unique

position but rather articulates more clearly than

most a common assumption of those who theo-

rize about liberal democracy. Hence, although

congruence is only a part of our general inter-

est in democratic processes, it is an important

part. . . .

Alternative Visions of Democracy and

Congruence

In each of the two visions of democracy exam-

ined here, there is a clear path by which both

electoral and legislative processes can create

congruence between the citizen median and the

behavior of governments and policymakers. But

the path in each vision is di¤erent and so are the

areas where one may expect problems.

The Majority Control vision assumes that po-

litical power will be concentrated in the hands of

identifiable governments chosen by the electorate

and responsible to it. Elections involve competi-

tion between incumbent governments and chal-

lengers. Voters evaluate the past performance

and future promises of each and choose the con-

tender whose policies they expect will be closest

to their preferences. That contender wins elec-

toral and legislative majorities and comes to of-

fice committed to a set of policies favored by a

citizen majority. When in o‰ce, the new gov-

ernment carries out those policies under the eye

of the electorate, which can evict it in the next

election if it fails to keep its promises.

The key stage in the Majority Control vision is

clearly electoral competition: party alternatives,

voter choices, and the aggregation of the two.

Elections must provide voters with identifiable

alternative governments; they must also produce

clear control over policy-making for the party

preferred by the citizens. If these defining char-

acteristics of majority control are achieved, then

whether there is a close correspondence between

voters and policymakers will depend on another

feature of the election: the presence of a party or

candidate located at or very near the median

voter. If neither identifiable alternative govern-

ment is close to the median voter, then by our

definition, the majoritarian democratic process

will not result in a government that is committed

to ‘‘what the voters want.’’ Responsible incum-

bents in o‰ce at the time of the election should

be helpful: single-party majority governments

that bear clear responsibility for their actions will

be pressed to anticipate the citizen majority as

they look to the election; voters will find it easier

to evaluate the credibility of promises and to

choose the party whose true position is closest to

their preferences.

Scholars have o¤ered a variety of specific

models to explain how Majority Control systems

can deliver policies that the citizens want. In the

well-known two-party competition model pro-

posed by Anthony Downs, the desire to win

elections drives both parties toward the position

of the median voter. With a single dimension of

party competition, a party that fails to converge

nearly to the median can always be defeated by a

party that does move to the median. The strate-

gic incentives for the parties and the rational

choices of voters act together to provide victories

for the party that is closest to the median.18 If

the theory of center-driven party competition

were empirically true, it would provide a power-

ful underpinning for the claim of the Majority

2. Robert A. Dahl, Democracy and Its Critics (New

Haven, Conn.: Yale University Press, 1989), 95. In a

similar vein, see Hanna Pitkin, The Concept of Repre-

sentation (Berkeley: University of California Press,

1967), 234.

18. Downs, An Economic Theory of Democracy (New

York: Harper and Row, 1957); see also Heinz Eulau

and Kenneth Prewitt, Labyrinths of Democracy: Adap-

tations, Linkages, Representation and Policies in Urban

Politics (New York: Bobbs-Merril, 1973); Joseph

Schlesinger, Ambition and Politics: Political Careers in

the United States (Chicago: Rand McNally, 1966); and

Joseph Schumpeter, Capitalism, Socialism and Democ-

racy (New York: Harper and Row, 1942).

Representation 331

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Control vision to create congruence. However,

there is much controversy about the correspon-

dence between Downs’s theory and the empirical

facts of party competition.

Since only the winning party needs to be near

the citizen median to create congruence, the

Majority Control vision need not depend on

Downs’s strategic parties. It can also encompass

‘‘mandate’’ versions of democracy19 or other

models in which incumbents face challengers

who over time o¤er a large array of possible al-

ternatives.20 We cannot here explicate the vary-

ing assumptions of these models, but we merely

note that various specific models in the broad

Majority Control vision can lead to the predic-

tion that the winner of the election should usually

be at or near the median voter. All of these

models then tend to assume that the election

winner will subsequently dominate the policy-

making process and implement the promised

policies.

If some combination of these models proves

empirically accurate, then the other (presumed)

virtues of majoritarianism will be buttressed by

good congruence between the preferences of the

electorate and the commitments of the policy-

makers. The potential problem, of course, is that

various empirical studies, and also some theo-

retical work, show failure of competition to pro-

duce consistently a party at the median.21

The Proportionate Influence vision gets to a

similar prediction of congruence in a very di¤er-

ent way. The models and research associated

with this vision are not directly oriented to

majorities or to control, but rather are oriented

to representation and bargaining. This vision is

less clearly articulated in its multiple stages. At

the electoral level the large literature on pro-

portional representation stresses the fairness of

having all voters’ voices count in getting o‰cials

into o‰ce.22 At the policy-making level, various

analysts of accommodative or consociational

democracy, most influentially Arend Lijphart,

argue that minorities in deeply divided systems

will want ‘‘grand coalition’’ arrangements that

guarantee them a voice in policy-making.23 In

his model of ‘‘consensus democracy,’’ Lijphart

19. See Anthony Birch, Representation (London:

Macmillan, 1972); and Austin Ranney, The Doctrine of

Responsible Party Government (Urbana: University of

Illinois Press, 1962).

20. See Kollman, Miller, and Page (fn. 5); and Richard

D. McKelvey and Peter C. Ordeshook, ‘‘Elections with

Limited Information: A Fulfilled Expectations Model

Using Contemporaneous Poll and Endorsement Data

as Sources,’’ Journal of Economic Theory 36 (June

1985).

21. For theoretical results, see, e.g., Alberto Alesina,

‘‘Credibility and Policy Convergence in a Two-Party

System with Rational Voters,’’ American Economic

Review 78 (September 1988); Peter J. Coughlin, ‘‘Can-

didate Uncertainty and Electoral Equilibria,’’ in James

M. Enelow and Melvin J. Hinich, eds., Advances in the

Spatial Theory of Voting (Cambridge: Cambridge

University Press, 1990); Melvin J. Hinich, ‘‘Equilib-

rium in Spatial Voting: The Median Voter Result Is an

Artifact,’’ Journal of Economic Theory 16 (December

1977); Donald A. Wittman, ‘‘Candidates with Policy

Preferences: A Dynamic Model,’’ Journal of Economic

Theory 14 (February 1977); and idem, ‘‘Spatial Strat-

egies When Candidates Have Policy Preferences,’’ in

Enelow and Hinich. For empirical results, see, e.g.,

David Robertson, A Theory of Party Competition

(London: Wiley, 1976); Ian Budge and Dennis Fairlie,

Voting and Party Competition (London: Wiley, 1983);

and Bernard Grofman, Robert Gri¤en, and Amihai

Glazer, ‘‘Identical Geography, Di¤erent Party: A Nat-

ural Experiment on the Magnitude of Party Di¤erences

in the U.S. Senate, 1960–84,’’ in R. J. Johnston, F. M.

Shelley, and P. J. Taylor, eds. Developments in Elec-

toral Geography (London: Routledge, 1990).

22. See Douglas Rae, The Political Consequences of

Electoral Laws (New Haven: Yale University Press,

1967).

23. Lijphart, Democracy in Plural Societies (New

Haven: Yale University Press, 1977); see also G. Lem-

bruch, ‘‘A Non-Competitive Pattern of Conflict Man-

agement in Liberal Democracies,’’ in Kenneth McRae,

ed., Consociational Democracy (Toronto: McClelland

and Stewart, 1974); and Jürg Steiner, ‘‘The Principles

of Majority and Proportionality,’’ British Journal of

Political Science 1 (January 1971).

Chapter 6 332

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

draws attention to various institutional devices

less inclusive than a grand coalition that induce

majorities to bargain with minorities.24 Most of

this work assumes that multiparty elections and

proportional representation are highly desirable

prerequisites for such negotiation.

To convert the Proportionate Influence vision

into a more clearly identified model of elections

connecting citizens and policymakers, we must

spell out the assumptions at each of the two

important stages in the process of government

formation. At the election stage, the vision as-

sumes multiple parties o¤ering a variety of al-

ternatives, so that all groups of citizens can find

compatible parties. The parties do not—must

not—converge to the center unless virtually all

the voters are located very close to it.25 At the

time of the election, then, the choices of voters

and the working of proportionate election laws

result in a legislature with parties representing all

these groups in their proportionate strength. A

critical implication of this fact is that the posi-

tion of the median legislator (or median party, if

parties are in fact the relevant units) should be

very close to that of the median voter.

The second stage of the Proportionate Influ-

ence vision concerns coalition bargaining. Since

an election often creates a legislature with no

single-party majority, various coalitions could

form among the many parties represented. Nat-

urally, the more diverse the electorate and, con-

sequently, the legislature, the more possible in

the abstract to build a coalition that strays from

the position of the median citizen. But as in

electoral competition, coalition theory predicts

that in one-dimensional situations the median

party will play a dominant role in government

formation,26 that is, all coalitions should include

the median party, although any coalition may

incorporate other parties that fall to one side or

the other (or both).

Existing research does not, however, provide

a clear prediction about whether the median

party will dominate policy-making, even in the

one-dimensional situation. Laver and Schofield

and de Swann argue that the policy position of

the median legislator will prevail, but Austen-

Smith and Banks, in a model that integrates

electoral competition and government forma-

tion, find that in equilibrium, final policy out-

comes never correspond to the preferences of the

median legislator.27 More generally, in situa-

tions where a single party or coalition of parties

forms a government and must maintain tight

party discipline—which empirically is the case in

almost all parliamentary systems—the govern-

ment might be expected to make policies that

correspond to its own internal median, not to the

legislative median.

Thus, we have here a potential for connec-

tions—through inclusion of the median party in

the coalition—without very close congruence.

As in the concerns about the failure of party

competition to produce at least one party at the

median in the majoritarian vision, the processes

that connect legislative bargaining to govern-

ment policy may also lead to consistent policies

o¤ the median.

Up to this point, we have considered only

the congruence between governments and citi-

zen preferences. That is, we have assumed that

24. Arend Lijphart, Democracies: Patterns of Ma-

joritarian and Consensus Government in Twenty-one

Countries (New Haven: Yale University Press, 1984);

see also Dahl (fn. 2), chap. 11.

25. In Lijphart’s empirical analysis of ‘‘consensus’’

systems, the number of e¤ective parties virtually de-

fines one of his dimensions (fn. 24), 214.

26. For an excellent review of the coalition formation

literature, see Michael Laver and Norman Schofield,

Multiparty Government: The Politics of Coalition in

Europe (Oxford: Oxford University Press, 1990),

chap. 5.

27. Ibid., 111; Abram de Swaan, Coalition Theory and

Cabinet Government (Amsterdam: Elsevier, 1973); and

David Austen-Smith and Je¤rey Banks, ‘‘Elections,

Coalitions, and Legislative Outcomes,’’ American Po-

litical Science Review 82 (June 1988).

Representation 333

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

the representation process ends with the forma-

tion of a government coalition. In practice all

governments will probably be somewhat influ-

enced by the issues raised by other parties in the

legislature. . . .

Under the Majority Control vision of congru-

ence (especially in the ‘‘mandate’’ formulations),

one would expect that the greater the opposition

influence in policy-making, the less the congru-

ence between policymaker and citizens. In this

vision, the selection of a governing party close

to the citizen median should already have re-

sulted in good congruence between final policies

and the citizen majority. If all other parties have

not converged to the median position, however,

giving weight to the opposition after a govern-

ment forms can only move the policymakers

away from the median. Moreover, such influence

will make it more di‰cult for voters to make

clear retrospective judgments about government

responsibility—a fact that decreases the incen-

tives for parties to converge to the median in the

first place.

In the Proportionate Influence vision, giving

some weight to the opposition may pull policy-

makers back toward the median citizen if the

government includes the median party but ex-

tends from it to the right or left. Hence, in con-

trast to what we would expect in ideal Majority

Control systems, giving opposition parties sig-

nificant weight in policymaking may improve

congruence between what citizens want and what

policies result in Proportionate Influence sys-

tems; but it also may not, depending on the spe-

cific positions of the government and the other

parties. . . .

Characteristics of Liberal Democratic Control

Systems: Theory and Practice

The two visions of democracy are founded in

experience and custom as well as in theory.

[Table 6.2] suggests that we can identify features

of electoral competition, electoral outcomes, and

legislative bargaining that will enable us to cate-

gorize empirically the di¤erent systems accord-

ing to the extent to which they follow one vision

or the other. Table 6.3 presents the data neces-

sary to accomplish this task; the data are used

to categorize the systems in twelve industrial

democracies for the period 1968–87. Subse-

Table 6.2

Visions of democracy and processes that create congruence between voters and policymakers

Process stages Majority control vision Proportionate influence vision

Electoral competition identifiable alternative governments, one a

responsible incumbent, one or both close

to the median voter

wide range of party choice; absence of

explicit coalition commitments

Election outcomes party close to median voter wins majority proportionate legislative representation of

all parties and voters

Government formation election winner forms majority government bargaining: government coalition

includes the median legislator

Policy-making between

elections

government dominates all policy-making coalitions may change but still include

median; negotiation with opposition

parties may help balance government

parties right or left of median party

Congruence prediction government is the policymaker and is close

to the median voter

government includes median legislator,

but average weight of all policymakers

will be closer to the median voter

Chapter 6 334

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

T a b le

6 .3

S y st e m

c h a ra ct e ri st ic s in

tw e lv e d e m o c ra ti c c o u n tr ie sa

M .C .

P .I .

M .C .

P .I .

M .C . &

P .I .

(1 )

(2 )

(3 )

(4 )

(5 )

(6 )

C o u n tr y

Id e n ti fi a b il it y o f

F u tu re

G o v t.

P a st G o v t.

S ta tu s

E ¤ e c ti v e N o .

o f P a rt ie s

S in g le

P a rt y

o r P E C

W in s

M a jo ri ty

P ro p o rt io n a li ty

O p p o si ti o n

C o m m it te e

In fl u e n c e

A u st ra li a

1 0 0

.4 4

2 .5

1 0 0

8 7

N o

B e lg iu m

0 0

6 .1

1 4

9 1

Y e s

D e n m a rk

3 8

0 5 .4

0 9 7

Y e s

F ra n ce

3 8

.1 2 5

3 .5

7 5

7 9

N o

W e st G e rm

a n y

1 0 0

0 2 .7

1 0 0

9 8

Y e s

Ir e la n d

3 3

.3 3

2 .6

5 0

9 6

N o

It a ly

0 0

3 .6

0 9 5

Y e s

N e th e rl a n d s

1 7

0 5 .1

1 7

9 6

Y e s

N e w

Z ea la n d

1 0 0

1 .0

2 .0

1 0 0

8 0

N o

S p a in

1 0 0

.3 3

2 .6

4 3

8 3

Y e s

S w e d en

2 9

.1 2 5

3 .3

8 3

9 8

Y e s

U n it e d K in g d o m

1 0 0

.6 7

2 .2

6 7

8 5

N o

a . S e e te x t fo r d e sc ri p ti o n s o f th e m e a su re s. D a ta

fo r c o lu m n s 1 – 4 a re

a v e ra g e s fr o m

th e p e ri o d 1 9 6 8 – 8 7 . D a ta

fo r c o lu m n 5 a re

fr o m

T h o m a s T .

M a c k ie

a n d R ic h a rd

R o se , T h e In te rn a ti o n a l A lm

a n a c o f E le c to ra l H is to ry , 3 d e d . (W

a sh in g to n , D .C .: C Q

P re ss , 1 9 9 1 ), 5 0 9 – 1 0 . C o lu m n 6 is

b a se d o n th e a n a ly si s b y S tr o m

(f n . 2 9 , 1 9 9 0 ), c h a p . 3 ; a n d P o w e ll a n d W h it te n (f n . 3 3 ).

Representation 335

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

quently, we will use data from 1978 to 1985 in

order to test congruence.30

At the stage of electoral competition, the Ma-

jority Control vision stresses that the voters be

able to identify future alternative governments

and that responsibility for past policy-making

by the incumbent government should be clear.

The Proportionate Influence vision emphasizes a

large number of parties o¤ering a wide range of

choices. In table 6.3 we see the countries in our

study followed, in columns 1–3, by features of

their electoral competition.31 . . .

Citizens, Governments, and Ideological

Congruence in Majoritarian and Proportionate

Influence Systems

We develop two measures, called Government

Distance I and Government Distance II, of the

congruence between the position of the govern-

ment and the estimated position of the median

voter.38 . . . Government Distance I takes the av-

erage position of all the parties in the government

weighted by the size of the respective parties. . . .

Since the number of portfolios a party receives

may not be a good measure of its influence in the

coalition, we developed an alternative measure,

Government Distance II. This second measure

assumes that the left-right position of the gov-

ernment coalition is dominated by the placement

of the median party within it. Hence, Govern-

ment Distance II is simply the left-right posi-

tion of the median party within the government.

Which of these two measures is more appropri-

ate depends, of course, on whatever theory we

might have about how policy-making goes on

within the government. As we shall see, however,

the results for both measures are quite similar.40

30. It would be ideal to be able to analyze congruence

for this entire twenty-year span, but since the Castles

and Mair expert survey measured party positions in

1982, we have used only the 1978–85 period so that we

can rely on the assumption that party positions have

not changed much. There are thirty-eight governments

in these twelve countries in this time period. Various

readers have suggested that we extend our time period

and bring in more cases, but we simply cannot locate a

comparable survey of experts at another time period

that asks the appropriate left-right question.

31. Some readers may be troubled by the absence of

electoral laws from the analysis. Clearly the electoral

law of a given political system shapes many features of

electoral competition and government formation that

are important to this study, including the e¤ective

number of parties, proportionality, identifiability, and

the election of single-party majorities. Nonetheless,

for the purposes of this study, it would be di‰cult—if

not wrongheaded—to categorize a country as either a

Majority Control or Proportionate Influence system on

the basis of its electoral laws. One problem is the di‰-

culty of developing an appropriate measure of electoral

laws because each one has unique features, with im-

portant di¤erences in aggregation rules and in district-

ing. (For example, Spain and the Netherlands both

have proportional representation [pr], but the propor-

tionality of electoral outcomes in Spain is much lower

than in the Netherlands, as shown in Table 6.3.) More

important for analysis, the nature of electoral compe-

tition varies over time within systems having the same

election law. In systems with single-member district

pluralities, for example, if there is a minority govern-

ment at the time of an election, clarity of responsibility

for past policy-making will be low. In systems with pr,

to take another example, there are often cases in which

identifiability is high because of the formation of pre-

election coalitions. As our analysis focuses on election-

specific characteristics of party competition, we do not

use the election laws directly to classify the various

political systems. However, analysis of the indirect

impact of election laws on congruence under various

conditions, through the features here examined, is an

interesting topic for future research.

38. Since the left-right scales have discrete boundaries

between the di¤erent cells, we approximate the loca-

tion of the median voter using a technique described in

Thomas H. Wonnacott and Ronald J. Wonnacott, In-

troductory Statistics for Business and Economics, 3d ed.

(New York: John Wiley, 1984), 671.

40. The mean scores by country for Government Dis-

tance I (II) are Australia 1.35 (1.35), Belgium .74 (.74),

Denmark 1.36 (1.46), France 1.96 (2.15), West Ger-

many 1.55 (1.81), Ireland .47 (.84), Italy .92 (1.24),

Netherlands .90 (.50), New Zealand .95 (.95), Sweden

1.28 (1.17), Great Britain 2.39 (2.39), and Spain 1.94

(1.94).

Chapter 6 336

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Comparing Congruence between Citizens and

Governments in the Three Types of Systems

Table 6.5 shows the average distance scores for

the three types of systems using our two di¤erent

measures of the position of the government. The

data show that the two measures of distance

work quite similarly. It is also clear that the

Majority Control and Mixed systems have

governments that are on average substantially

farther from the median voter than are govern-

ments in the Proportionate Influence systems: the

average government in the Majority Control and

the Mixed system is over 1.5 points from the

median; the average government in the Propor-

tional Influence system is about 1 point away.

Even with so few cases, the di¤erence between

the mean of the Proportionate Influence systems

and the mean of the Majority Control systems is

statistically significant at .05 (one-tailed test).

In parentheses in table 6.5 we show the per-

centage of voters between the government and

the median citizen. This figure depends on both

the absolute distance and the distribution of

Table 6.4

Characteristics for majority control, mixed, and proportionate influence systems 1968–87 (1978–85)a

System type

Majority control:

Australia, Great

Britain, New Zealand

Mixed: France,

Germany, Ireland,

Spain, Sweden

Proportionate influence:

Belgium, Denmark,

Italy, Netherlands

Electoral competition

Identifiability 100

(100)

(75)

(45)

Past government status .67

(.80)

.17

(.06)

(0)

E¤ective no. of parties 2.2

(2.2)

3.1

(3.0)

5.1

(5.4)

Election outcomes

Percentage of elections won

by a single party or a

preelection coalition

(100)

(58)

(10)

Proportionality 85

(83)

(93)

(96)

Legislative bargaining

Percentage of committee

systems that permit

opposition influence

(0)

(60)

(80)

Number of elections 29

(5)

(12)

(10)

a. The top number in each cell is for the period 1968–87. The figure in parentheses is for the period 1978–85

(the period for which we analyze congruence). The top number in the ‘‘Proportionality’’ row is calculated using the

figures in Mackie and Rose, The International Almanac of Electoral History, 3d ed. (Washington, D.C.: CQ Press,

1991), 510, which are calculated using only the last election reported in their study.

Representation 337

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

voters on the left-right scale. If the voters were

more dispersed in the Majority Control systems,

for example, a larger distance might a¤ect the

same number of voters as does a smaller distance

in the Proportionate Influence systems. How-

ever, we see the same pattern as in the absolute

distances. The Majority Control systems find, on

average, 28 percent of the electorate between the

government and the median, whereas the figures

are 23–25 percent in Mixed systems and 20 per-

cent in the Proportionate Influence systems.

The advantage of the Proportionate Influence

systems in o¤ering greater congruence between

governments and voters is somewhat theoreti-

cally unexpected. We expected that governments

in the Majority Control systems would be close

to the median as the direct result of party com-

petition and voter choices (under either Down-

sian theory or some of the nonstrategic or

partially strategic alternatives). We also expected

coalition bargaining in the Proportional Influ-

ence systems might result in governments that

are often farther away from the median voter.

But the converse is true. The reason for the

poorer performance of the Majority Control

systems is basically that the two main parties in

Britain and Australia are far from the median

(over 2 points) during the period of our study.

The closer of the two large parties does come to

power, but it is still rather extreme.41 In New

Zealand the Majority Control vision seems to

work better; in fact both major parties are fairly

close to the median (about 1 point).

A similar problem is evident in the Mixed

systems, although it is less theoretically surpris-

ing in the multiparty situations. A common pat-

tern here is the formation of formal or informal

preelection coalitions that pit right against left.

These coalitions frequently fail to converge, but

the one that gets a majority forms a govern-

ment without bargaining with the opposition. In

France, especially, both major alternative gov-

ernments are very far from the median voter. In

1978 the winning conservative coalition was 2.75

from the median—the farthest in our sample. In

Germany, Spain, and Sweden, too, the alter-

natives are rather far apart, each around 1.5

points from the median. Only in Ireland are both

of the two alternatives quite close to the median.

A Regression Analysis

While the results in table 6.5 are interesting, it is

troubling that we have only five cases of pure

Majority Control elections. A regression analysis

can help us go beyond the typology to illuminate

the contribution of various properties of political

systems to the degree of congruence. . . .

Table 6.5

Congruence between government and citizen left-right

orientationsa

System type

Majority

control Mixed

Proportional

influence

Government

Distance I

1.61

(28%)

1.43

(23%)

.96

(20%)

Government

Distance II

1.61

(28%)

1.55

(25%)

1.03

(20%)

N 5 16 17

a. Government Distance I measures the di¤erence

between the weighted mean left-right position of the

government and the left-right position of the median

voter. Government Distance II measures the di¤erence

between the left-right position of the median party in

the government coalition and the left-right position of

the median citizen. The numbers in parentheses give

the percentage of voters between the government and

the median citizen. Positions of the parties are taken

from Castles and Mair (fn. 10).

41. In Britain the closest parties to the median voter

were the Liberals in 1979 and the Alliance in 1983, but

neither of these parties won as much as a quarter of the

votes, and both were heavily penalized by the election

laws. The Conservatives were somewhat closer to the

median than was Labour, but both large parties were

rather far away.

Chapter 6 338

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Table 6.6 gives the results of six ols regres-

sions where the formation of a new government

is the unit of analysis and the electoral and leg-

islative characteristics described in table 6.4 are

the independent variables.42 . . .

The general conclusions from the first two

regressions are clear: Majority Control charac-

teristics have mixed e¤ects, which have the net

result of moving the government away from the

median voter; Proportionate Influence charac-

teristics move the government closer to the me-

dian voter. . . .

. . . [W]hen all the variables are entered into

the model, the main variables from the Majority

Control vision have a mixed e¤ect with identifi-

ability of future governments harmful to con-

gruence; the variables from the Proportionate

Influence vision are helpful to congruence.43 The

Table 6.6

Predicting distance between the median voter and the left-right position of the government (OLS model of Govern-

ment Distance I and II)

Dependent variable:

Government Distance I

Dependent variable:

Government Distance II

Independent Variablesa (1) (2) (3) (4) (5) (6)

Identifiability of future government .010

(.003)

— .009

(.003)

.009

(.003)

— .0073

(.0036)

Past government status �.52 (.40)

— �.54 (.38)

�.58 (.42)

— �.61 (.42)

Majority or PEC wins election �.17 (.30)

— �.29 (.29)

�.22 (.31)

— �.34 (.32)

E¤ective number of parties — �.131 (.070)

�.05 (.07)

— �.14 (.07)

�.08 (.08)

Proportionality of electoral outcome — �.047 (.021)

�.041 (.020)

— �.038 (.021)

�.036 (.021)

Opposition influence in committees �.27 (.33)

.44

(.33)

.07

(.37)

�.46 (.36)

.28

(.35)

�.12 (.40)

Intercept .95

(.36)

5.79

(1.80)

4.83

(1.88)

1.27

(.38)

5.19

(1.87)

4.78

(2.03)

N 38 38 38 38 38 38

Adjusted R2 .20 .14 .26 .15 .11 .18

Standard error of the regression .65 .68 .63 .69 .71 .68

a. The independent variables are described in the text.

42. Our theoretical discussion does not suggest what

the appropriate functional form should be, so we

examined a wide variety of functional forms and a

simple linear relationship turned out to be the most

appropriate for each variable.

43. Our findings regarding the Proportionate Influence

variables are interesting when compared with empirical

studies of budget deficits by political economists; for a

recent review of this literature, see Alberto Alesina and

Guido Tabellini, ‘‘Positive and Normative Theories of

Public Debt and Inflation in Historical Perspectives,’’

European Economic Review 36 (April 1992). Roubini

Representation 339

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

regression analysis therefore supports and clari-

fies the simple comparison of system types. De-

spite the plausibility of Downsian theory and

some of the other formulations of majoritarian

democracy that predict congruence, when there

exist clearly identifiable future governments at

election time, the elected governments tend to be

far from the median voter. And despite concerns

about government formation processes, as the

e¤ective number of parties and proportionality

of electoral outcomes increase, congruence in-

creases. In fact, this congruence is best when

parties do not undertake preelection commit-

ments that may lead to the formation of minor-

ity governments. To put the comparison between

the two types of systems another way, on aver-

age the failures of electoral competition in the

Majority Control (and Mixed) systems seem

more serious for congruence than does the failure

of government formation in the Proportionate

Influence systems. . . .

Citizens, Policymakers, and Ideological

Congruence

Governments in parliamentary systems are not

totally uninfluenced by the opposition in making

policy, and the Majority Control and Propor-

tionate Influence visions make di¤erent predic-

tions about how the role of opposition parties

should a¤ect congruence. Consequently, an ex-

ploration of the e¤ectiveness of the two visions

in linking policy-making and voters must look

beyond the parties that formally share govern-

ment responsibility. . . .

Comparing Congruence between Citizens and

Policymakers in the Three Types of Systems

Table 6.7 compares the distance between policy-

makers and the citizen median in the three types

of systems. . . .

. . . Most strikingly, congruence improves in all

three types of systems.48 Across all thirty-eight

governments, the average distance between the

median voter and the government was about 1.3;

the average distance between the median voter

Table 6.7

Congruence between policymakers and citizen left-

right orientationsa

System type

Majority

control Mixed

Proportionate

influence

Policymaker

Distance I

1.17

(22%)

1.03

(18%)

.50

(10%)

Policymaker

Distance II

1.17

(22%)

1.15

(20%)

.59

(12%)

a. For Policymaker Distance I and Policymaker Dis-

tance II, a party’s weight is determined by whether it is

a government party, a support party, or an opposition

party. Policymaker Distance I uses the same measure

as Government Distance I to calculate the position of

the government. Policymaker Distance II uses the same

measure as Government Distance II to calculate the

position of the government. Further details are in the

text and the appendix. The numbers in parentheses give

the percentage of citizens between the Policymakers

and the median citizens.

and Sachs, for example, find that systems with a high

incidence of coalition and minority governments have

relatively large levels of public debt; Roubini and

Sachs, ‘‘Political and Economic Determinants of Bud-

get Deficits in the Industrial Democracies,’’ European

Economic Review 33 (May 1989). Since the central

characteristics of Proportionate Influence systems lead

to coalition and minority governments, it appears that

system characteristics which improve congruence be-

tween governments and citizens may also be associated

with large budget deficits. We are grateful to Bill Keech

for pointing this out to us.

48. Congruence of policymakers was not greater than

that of governments in every case, however. It is not a

tautology. In eight of the thirty-eight cases the congru-

ence was less for policymakers on at least one of the

two measures, although the di¤erences are usually not

very large.

Chapter 6 340

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

and the weighted policymakers was about .85.

This decrease was of approximately the same

magnitude for all three types of systems. This

does not mean, however, that oppositions are

closer to the voters than are the governments.

They are not. Rather, it means that as long as we

continue to weight the governments more heav-

ily than the oppositions in our estimate of policy-

making, more congruence is created by giving

the oppositions some weight than by leaving

them out of the process. . . .

Regression Analysis

Since the results from table 6.7 are interesting

and somewhat unexpected, it is useful to analyze

further the relationship between the system

characteristics and the policymaker variables

using multivariate regressions. . . .

The regression results help interpret and sup-

port the results of the simple comparison of

system types. The policymaker measures show

reduced distances in all types of systems. How-

ever, the Proportionate Influence systems still

show a substantial advantage. Identifiability in

electoral competition remains costly, although

the cost is not as great in absolute terms after we

upweight opposition influence on policy-making.

Poor proportionality in representation remains

highly costly. Single-party majority incumbents

before the election seem to be helpful. Yet mul-

tiparty systems also remain somewhat helpful,

as does opposition influence. The net congru-

ence advantage is to the Proportionate Influence

vision.

Concluding Comments

We have attempted to being into more precise

focus two general visions of the processes that

link citizens and policymakers in contemporary

democracies. We should stress that the gener-

ality of our results is constrained by our research

design: it may be that a di¤erent slice of time

would reveal majoritarian electoral competition

in which the parties are not so extreme and pro-

portionate influence bargaining is less centrist.

Moreover, we are well aware that the ‘‘commit-

ments’’ of governments and their actual policy

outcomes are not necessarily the same. This dif-

ference would be especially troubling for our re-

sults if policies diverged further from promises in

the Proportionate Influence systems than in the

Majoritarian ones. The di‰culty of identifying

clear responsibility for policy in the former cre-

ates prima facie grounds for concern.

With these caveats said, the results of our

analysis seem clear and consistent. In the simple

comparison, contrary to our expectations from

the theoretical arguments about creating con-

gruence connections in each approach, the gov-

ernments in the Proportionate Influence systems

are on average significantly closer to their me-

dian voter than are governments in the Majority

Control and Mixed systems. The regression

analysis reassures us on this point. If voters are

presented with two clear alternatives (parties or

preelection coalitions), these alternatives—and

resulting governments—tend to be rather distant

from the median voter. If voters are presented

with a wide range of choices and electoral out-

comes are proportional, governments tend to be

closer to the median. It is reassuring that the

regression results are supportive because they

allow us to take advantage of the mixture of

properties in the Mixed systems, rather than

relying solely on the number of pure Majority

Control cases.

Our analysis of policymakers, although neces-

sarily more speculative because of the weighting

problem, is also illuminating because the results

in part run counter to our initial theoretical ex-

pectations. Taking some account of opposition

influence helped congruence with voters in virtu-

ally all the systems, with surprisingly large e¤ects

in the Majority Control and Mixed systems.

However, the net advantage remained with the

Proportionate Influence systems, which gained

congruence, especially because of the strong

Representation 341

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

weighting of the opposition parties during mi-

nority government and because these systems

usually permit more opposition influence in

committees.

The results with respect to Majority Control,

and more generally with respect to high identifi-

ability, raise an important additional question

about policy-making. We examine congruence

on a government-by-government basis, and do

not have a long enough time span to take aver-

ages of the governments over several decades.

Hence, although each government in Britain and

Australia may be quite distant from the median

voter, the average position over time might be

much closer to the center. Of course, the long

predominance of such governments as the Con-

servatives in Britain from 1979 to the present

(or the conservative coalition in France from

1958 to 1981) may imply that this oscillation

does not redress the balance very quickly (or

at all).

The appropriate time frame for congruence is

an important issue for future research. The rela-

tionships between congruence and other features

of democratic government frequently proposed

as desirable (stability, e‰ciency, responsibility)

remain another rich area for exploration. The

consideration of these questions reminds us

again of the challenging empirical, theoretical,

and normative issues associated with the study of

congruence. For this reason, we see the current

results as a contribution, not a conclusion, to

our understanding of the fascinating problem of

the electoral connection between citizens and

policymakers. . . .

Chapter 6 342

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

The Political Consequences of Electoral Laws

Douglas W. Rae

Variations of Degree

The degree to which seat allocations diverge

from the condition of perfect proportionality is a

function of two electoral law variables: (1) elec-

toral formulae, and (2) district magnitudes. As a

rule, P.R. formulae and high district magnitudes

produce more nearly proportional results, while

‘‘first-past-the-post’’ formulae and low district

magnitudes produce the greatest disproportion-

ality (i.e. the greatest advantage for large parties

over small ones). Let me consider the two elec-

toral law variables individually, beginning with

the electoral formulae.

Electoral formulae diverge from proportion-

ality along two institutional dimensions. First,

proportional representation formulae produce

smaller deviations from proportionality than

‘‘first-past-the-post’’ formulae, whether the latter

are based on pluralities or majorities.9 Second,

highest averages P.R. formulae diverge farther

from proportionality than largest remainder P.R.

formulae.10 The first di¤erence—between P.R.

and ‘‘first-past-the-post’’ formulae—is by far the

greater of the two.

It follows that P.R. formulae are likely to

minimize (but not to eradicate) the general bias

of electoral formulae in favor of strong parties.

P.R. formulae give a smaller advantage to the

strong elective parties and exact a smaller price

from weak ones.11 They give a smaller bonus to

the strongest single elective party.12 And, natu-

rally enough, they are less apt to deny represen-

tation to elective parties.13

The proportionality of seat allocation also

varies with the number of seats assigned to elec-

toral districts—district magnitudes. Where many

seats are allocated in each electoral district, the

outcome is likely to approximate proportion-

ality.14 But where fewer seats are allocated in

each district, outcomes are likely to diverge more

sharply from proportionality.15 This relation-

ship is, however, curvilinear: as district magni-

tudes increase, disproportionality decreases at a

decreasing rate.16 Another way to put the same

thing is: as district magnitude increases, the pro-

portionality of outcome increases at a decreasing

rate.17 One need not be surprised that the bonus

obtained by the strongest party declines as dis-

trict magnitude rises.18

Since plurality formulae are always associated

with a single-member district, it is hard to dis-

tinguish the disproportionality of these formulae

from that of the single-member district’s low

magnitude. Indeed the distinction has no em-

pirical meaning, except in the Australian case,

where a majority formula is associated with a

single-member district.

Given that the defractionalization pattern,

which works in favor of fewer, stronger legisla-

tive parties, is a general fact of electoral life, it

is also evident that specific institutions pro-

duce variations of degree within the pattern. Po-

litical ingenuity can render the defractionalizing

pattern stronger or weaker through the ma-

nipulation of the institutional variables: electoral

formulae and district magnitudes.

Suppose one wanted to design an electoral

law that would maximize the defractionalizing

Excerpted from: Douglas W. Rae, The Political Con-

sequences of Electoral Laws, revised edition. New

Haven: Yale University Press, 1971. Reprinted by

permission.

9. Di¤erential Proposition Four.

10. Di¤erential Proposition Nine.

11. Di¤erential Proposition One.

12. Di¤erential Proposition Two.

13. Di¤erential Proposition Five.

14. Di¤erential Proposition Ten.

15. Di¤erential Proposition Ten.

16. Di¤erential Proposition Eleven.

17. Di¤erential Proposition Eleven.

18. Di¤erential Proposition Thirteen.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

pattern, producing strong advantages for strong

parties, strong penalties for weak ones, and

often ‘‘manufacturing’’ legislative majorities.

The findings of this study suggest that the single-

member district is the only necessary instrument.

With these very low district magnitudes, the ad-

vantage of strong parties will be maximized, no

matter what formula is used. Plurality and ma-

jority formulae will behave in the same way,

defractionalizing legislative party systems by

favoring fewer stronger parties. For that matter,

even P.R. formulae would have the same e¤ect

in single-member districts. Because only one

party can win in each district, the strong parties

benefit at the expense of the weak ones, and leg-

islative party systems are composed of fewer,

stronger parties.19

Suppose, on the other hand, that one were

worried about proportionality, and wished

therefore to minimize defractionalization, giving

each party its electoral due—no more and no

less. Given these objectives, he should insist on

a P.R. formula, preferably based on the largest

remainder procedure, linked with high district

magnitudes (i.e. many seats per district). It would

probably not be worthwhile to expand the dis-

tricts beyond ten or twenty seats, since the added

proportionality seems to decline rapidly beyond

that level. But with very low district magnitudes

(i.e. less than six seats per district), even the

largest remainder P.R. formula would produce a

very substantial defractionalizing e¤ect.

Distal Consequences of Electoral Laws

Do the short-run e¤ects we have been discussing

have long-run consequences for party systems?

What are they? These questions can be answered

only with considerable caution, since party sys-

tems are influenced by many variables—social,

economic, legal, and political. Proximal e¤ect of

electoral law upon the legislative representation

of parties is to be counted only one of many

determining forces. And it is, secondly, impos-

sible to sort out all the contributing factors, or

to assign even approximate weights to them.

Worse yet, electoral laws are themselves shaped

by party systems.

In the face of these di‰culties, one can only

suggest limited connections. I have chosen to

formulate my commentary in response to the

question: ‘‘Where electoral laws do in fact make

a long-run di¤erence, which specific properties of

electoral laws are apt to produce what di¤er-

ences?’’ By choosing to work within the assump-

tion that electoral laws do exert long-run e¤ects,

yet without demonstrating this assumption’s va-

lidity, I have settled for a very limited level of

analysis. But to do more would require not one,

but twenty or more, developmental studies, each

devoted to a single country. The limited sugges-

tions o¤ered here may provide some guidelines

for research of that kind, and may have at least

tentative significance in their own right.

Party systems vary over a continuum, from

non-fractionalization in one-party systems to ex-

treme fractionalization in systems where a great

many parties compete on about equal terms.

Among the party systems analyzed in this study,

the actual range of variation lies between two

less distant points: U.S.-style, two-party compe-

tition and Israeli-type multi-partism. Every party

system, at any one point in time, may be as-

signed a place on this continuum, although (and

this is important) individual systems may move

along these scales, toward or away from the em-

pirical extreme of two-party competition.

How do electoral systems influence the move-

ment of systems on the fractionalization con-

tinuum? I wish to suggest that the pattern of

proximal defractionalization described above is

the source of whatever influence electoral laws

have on the fractionalization of party systems.

19. This assertion must be modified where parties

which are weak on a national scale enjoy pockets of

local support, enabling them to profit from the single-

member district in those areas.

Chapter 6 344

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

Where the pattern is strong—large parties are

greatly advantaged—the electoral system exerts

pressure on the system for two-party competi-

tion. But where the pattern is weak—large

parties are only slightly advantaged—a weaker,

often negligible, pressure is exerted in that direc-

tion. The defractionalizing pattern is a restraint

on the fractionalization of party systems, and the

e¤ective pressure exerted by electoral laws varies

with the intensity of the defractionalizing pattern

itself. Multi-partism is most likely where elec-

toral laws produce a weak defractionalizing

pattern, and two-party competition most likely

where the electoral laws produce a strong

defractionalizing e¤ect.

Now, according to my earlier comments, the

defractionalizing pattern is complex. It entails at

least five related subpatterns: (1) the advantage

of large parties over small ones in the division of

legislative seats; (2) the awarding of a ‘‘bonus’’ in

seats to the strongest party; (3) the exclusion of

small parties from the legislative arena; (4) the

overall defractionalization of legislative party

systems; and (5) the fairly frequent creation of

‘‘manufactured majorities’’ in legislative party

systems. These are the subpatterns present in

all electoral systems, but stronger in some than

others, which constitute the defractionalizing

process that seems so important. The question

thus becomes, ‘‘What electoral law variables

produce this syndrome?’’

The answer to this question was foretold in the

examination of proximal e¤ects. Here let me re-

capitulate the e¤ects of these institutions, with

attention to the contribution they make to the

shaping of party systems over time. Logically,

the sequence of inferences is: (a) electoral law

variables to intensity of the defractionalizing

pattern, and, with less confidence, (b) the inten-

sity of the defractionalizing pattern to the long-

run fractionalization of the party system.20

What electoral law provisions intensify the

defractionalizing pattern and therefore seem

likely to exert pressure toward two-party

competition? The answer is simple: the single-

member district, or, failing that, small multi-

member districts. In a single-member district,

almost any formula21—the plurality is most

common—is likely to advantage the strong

parties and, in general, to establish the defrac-

tionalizing pattern. This much is confirmed by

the analysis of proximal e¤ects. But what about

the long-range e¤ects? The findings of the study

show a fairly consistent association between

single-member plurality formulae and two-party

systems.22

A causal interpretation of this association falls

upon several exceptions, the clearest of which are

Canadian and Austrian. Nevertheless, the com-

bination of the proximal defracationalizing pat-

tern and the distal association with two-party

competition suggests that the single-member dis-

trict is likely to contribute to the development

and sustenance of two-party systems. Other fac-

tors, such as regional minorities, may reverse this

condition, as is the case in Canada. But, insofar

as the electoral law exerts a controlling pressure,

the single-member district is likely to press the

system toward two-party competition.

And what arrangements are most likely to

press party systems toward multi-partism, be-

cause they exert a very weak defractionalizing

e¤ect? These would be the institutions that

optimize the proportionality of outcomes: largest

remainder P.R. formulae, operating in high

magnitude electoral districts. Because the out-

comes are more nearly proportional under

these provisions, the defractionalizing process is

weakened. Is there an association between these

arrangements and multi-partism? The study’s

findings show that there is:

20. A third inference is from the degree to which an

electoral system is presently fractionalized to the kind

of electoral laws adopted for future elections.

21. The French double ballot may be an exception,

although limited evidence suggests that it is not.

22. Di¤erential Proposition Three and Similarity

Proposition Seven.

Representation 345

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

1. In general, P.R. formulae are associated with

more fractionalized elective and parliamentary

party systems.23

2. Among P.R. formula electoral laws, those

using the largest-remainder procedure are asso-

ciated with greater fractionalization, both elec-

tive and parliamentary, than are those using

highest-average procedures.24

3. High district magnitudes are associated with

greater fractionalization in both elective and

parliamentary systems.25

The distal association between these institu-

tions and high fractionalization, even when seen

beside the weakness of the proximal defraction-

alizing pattern which they produce, does not

suggest a simple causal relationship. It does,

however, imply that insofar as the electoral law

exerts a controlling pressure, these provisions are

apt to press systems toward multi-partism and

away from two-party competition.

These conclusions suggest that the statesman

who must choose between electoral laws con-

fronts a dilemma. On the one hand, he may opt

for highly proportional election outcomes, in

which case he is likely to encourage the frac-

tionalization of party systems over time. Or, on

the other hand, he may opt to encourage the de-

velopment and maintenance of two parties, or

less fractionalized multi-party competition, with

the price being less proportional outcomes.

These alternatives may not be inevitable, but the

findings reported here make them seem proba-

ble: if proportionality, then multi-partism; but if

two-party competition, then also a dispropor-

tional outcome.26 . . .

Election Outcomes and Government Stability

It is generally supposed that election outcomes—

allocations of seats among parties—quite di-

rectly determine which party or coalition will

control a parliamentary government. So, for ex-

ample, one of our most sophisticated accounts

holds out these conditions for representative de-

mocracy: ‘‘A single party (or coalition of parties)

is chosen by popular election to run the govern-

ing apparatus,’’ and ‘‘Any party (or coalition

of parties) receiving the support of a majority of

those voting is entitled to take over the powers

of government until the next election.’’16 It is

revealing that these requirements, most obvi-

ously the second, lead to logical contradictions

and must therefore be violated by any empirical

system of parliamentary politics. Since this logi-

cal di‰culty carries our attention toward an

important empirical question, let me begin by

examining it briefly.

During a campaign, we have ne parties. Before

the electorate decides, one can imagine that

any of 2n e � 1 parties or coalitions might com-

mand majorities and form a government. Thus,

23. Di¤erential Proposition Six.

24. Di¤erential Proposition Nine.

25. Di¤erential Proposition Twelve.

26. The time-sequence data for the twenty-year period

I have studied do not produce clinching evidence for

these speculations: except for West Germany, there is

no system which changes drastically in its degree of

fractionalization. But that is not altogether surprising,

since all of the systems, save Germany and Israel, had

been in operation for many decades before the period

covered by this study. Historical analyses of the indi-

vidual systems would be of interest. I must conclude

that association, but not sequential data, support my

speculations.

16. Anthony Downs, An Economic Theory of Democ-

racy (New York: Harper & Row, 1957) pp. 23–24. A

related but less clearcut di‰culty arises in Robert

Dahl’s notion of polyarchy, when it is required that

alternatives with more votes replace alternatives with

fewer. See A Preface to Democratic Theory (Chicago:

University of Chicago Press, 1956), Chap. 3. In both

cases, the problem arises, I think, because the authors

are thinking primarily about two-party systems and

only secondarily about multiparty systems. In Dahl’s

analysis the requirement is stated as an ideal, but in

Downs’s it is treated (incorrectly) as a feature which

‘‘in practice’’ distinguishes democracy from other

forms.

Chapter 6 346

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

with three electoral parties one can imagine

23 � 1 ¼ 7 conceivable governments (e.g. A, B, C, AB, AC, BC, ABC.) According to ‘‘common

sense’’ and Downs’s stipulations, the election

outcome should reduce this set of possibilities to

one, the party or coalition which commands a

majority. But if more than one party wins repre-

sentation, this cannot be literally true. If np

parties win representation, the election outcome

must leave open the possibility of exactly 2np�1

distinct majority governments. The e¤ect of an

election outcome is thus to reduce the set of po-

tential majority governments without ‘‘picking’’

a single government. If, for example, three

parties win representation, then exactly 23�1 ¼ 4 possible governments must remain no matter how

seats are divided among these parties. And this

implies that both of Downs’s requirements are

unrealizable: the election itself cannot pick a

party or coalition to govern, and some parties or

coalitions holding majorities must inevitably be

denied control of government (e.g. 2np�1 ¼ 1). This very formal di‰culty is not without its rel-

evance to experience.

If a single party wins a majority of its own, the

di‰culty is merely technical. Say that A holds a

majority over B and C. It is still true that the

election permits four possible governments (A,

AB, AC, ABC) but there are compelling reasons

to presume that one will be chosen (A). In this

event, we need only add the stipulation that po-

tential governments with inessential members

(AB, AC, ABC, in our example) may be ignored.

One might suppose that the problem could thus

be reduced to a linguistic quibble were it not for

the fact that about two-thirds of all parliamen-

tary elections fail to produce single-party ma-

jorities (of 107 elections analyzed in the original

study, only 8 produced natural majorities, 25

produced manufactured majorities, and the re-

maining 74 produced neither). Thus a more typ-

ical case occurs when no single party holds a

majority and one is left without an obvious

choice of govenment. Thus, if A, B, and C each

have less than half the seats in a parliament, then

the possible governments include AB, AC, BC,

and ABC. This is a more than technical violation

of Downs’s requirement, for it leaves an open set

of potential governments (and, significantly, a

choice of potential alternatives to any govern-

ment that forms).

A useful example is o¤ered by the 1963 elec-

tion of the Icelandic parliament (the Althingi).

The Independence Party obtained 24 seats, the

Progressives 19, the Communists 8, and the Social

Democrats 7. With four electoral parties, one

might have imagined 24 � 1 ¼ 15 possible gov- ernments, and (since no party was denied repre-

sentation) the outcome simply restricts this to a

subset of 24�1 ¼ 8 groupings that command majorities and could conceivably form govern-

ments. To each of these, there naturally corre-

sponds a potential set of opposition parties:

Possible Government

Possible

Opposition

1. Ind, Prog, Com, SD nil

2. Ind, Prog, Com SD

3. Ind, Prog, SD Com

4. Prog, Com, SD Ind

5. Ind, Com, SD Prog

6. Ind, SD Prog, Com

7. Ind, Prog SD, Com

8. Ind, Com Prog, SD

The election itself comes to a very partial ar-

bitration of the question ‘‘Who governs?’’ Only

seven of the fifteen otherwise conceivable gov-

ernments have been eliminated, leaving eight

possibilities. Each party is a member of at least

five such possible governments, and no party

belongs to all eight. The e¤ective choice must

therefore rest with party leaders as they are

influenced by conflicts and compatibilities of

program, ideology, or personality in the forma-

tion of coalitions. It is a choice bounded but left

undecided by the election outcome. As it hap-

pens, government 6 (Independence–Social Dem-

ocrat) was formed in 1963, but roughly similar

outcomes have since World War II, also led

Representation 347

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

to the formation of governments 3, 4, and 7.17

Unless the process of government formation is

understood by the electorate, this surely raises

important di‰culties for theories of electoral

representation, for it is impossible to see how the

citizen could draw useful associations between

his voting decision and the eventual choice of a

government.18

Parliaments that are fractionalized enough to

produce these uncertainties19 would seem espe-

cially prone to government instability, for at

least two reasons. First, their numerical struc-

tures o¤er wide opportunity for the destruction of

governments. With more than one party-elite in a

government, it is possible to reach a vote of no

confidence even without a breach of party disci-

pline. And, as the number of essential partners

increases, one might expect an actuarial increase

in the frequency with which a particular decision

alienates a member party, bringing a new elec-

tion or a change of governments without one.

Moreover, since only one of 2np�1 possible gov- ernments has been formed, it seems probable

that member parties could, at some point, iden-

tify alternative partners with whom a more at-

tractive government could be formed. These

reasons are very roughly analogous to the struc-

tural relations outlined for partisan insurgency

earlier, although their interconnections are less

clear. A second reason points toward the like-

lihood that highly fractionalized parliaments

are indices of intense political conflict. If this is

so, then there will be a convergence of strong

motives and wide opportunities for changes of

government in such systems.

Unhappily, none of these conjectures can be

mistaken for a theory. And, although coalition

behavior has been the object of intensive study in

recent years, research on the specific question of

stability remains inconclusive. We do, however,

have a useful exploration of several common-

sense hypotheses in a recent study by Michael

Taylor and V. M. Herman.20 These authors

considered 196 governments in 19 countries

(those studied here, minus the U.S. and Switzer-

land, plus Japan) from the end of World War II

up to 1969. They define ‘‘government stability’’

in calendar days elapsed without a change of

prime minister or party support.21 Their study is

an e¤ort to treat variation in stability as a func-

tion of the parliamentary party system variables

discussed here. The most straightforward of their

results are these.

Independent Variable

Product-Moment

Correlation with

Government

Stability

(1) Number of parliamentary

parties (np)

–.39

(2) Fractionalization of

parliamentary parties (Fp)

–.448

(3) Number of parties in

government

–.307

(4) Proportion of seats held

by anti-system parties

–.450

These results tend to confirm the intuitive

expectations sketched out above, but they do not

include the most plausible explanation turned

17. Nils Andrain.

18. See Downs, pp. 142–63 for a discussion of these

di‰culties. In Thomas Casstevens, ‘‘An Axiom About

Voting,’’ American Political Science Review, 62 (1968),

205–07, an attempt is made to show that (on a sim-

plistic assumption about ‘‘winning’’ and a vitiating as-

sumption about the probabilities of parties doing so)

the voter should always just vote for his most preferred

party.

19. If parliamentary fractionalization reaches .5, there

may be no single-party solution, and if it reaches .75,

there cannot possibly be such a solution.

20. ‘‘Party Systems and Government Stability,’’

American Political Science Review, 65 (1971), 28–37.

21. In this, they follow an earlier study, Jean Blondel,

‘‘Party Systems and Patterns of Government in West-

ern Democracies,’’ Canadian Journal of Political

Science, 1 (1968), 180–203.

Chapter 6 348

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

up by the Taylor-Herman study. It is inter-

esting that their e¤ort to include ideological

distances in the notion of fractionalization actu-

ally weakened its relationship to instability (to

�.417.) But the related distinction between anti- system and pro-system parties nevertheless

proved highly interesting (variable 4 above.)

Their best explanation turns on the joint e¤ect of

‘‘proportion anti-system’’ and the fractionaliza-

tion of seats among the remaining parties. Using

these two variables in a multiple regression esti-

mate they find a fit of R ¼ .496. Ironically, the estimate is improved slightly (R ¼ .506) if right- wing anti-system seats are dropped, leaving only

communist seats in the proportion anti-system.

The apparent explanation for this account is

that, if anti-system parties always refuse to sup-

port governments, the proportion of available

votes required is increased (being r/2(r � a) where r is the parliament’s size and a is the size

of the anti-system contingent). The e¤ective de-

cision rule for sustaining governments thus

grows more restrictive with increases in anti-

system representation (or communists) and is less

easily met as remaining seats grow more frac-

tionalized. This seems, at any rate, the most

plausible conjecture.

Although this section is less developed ana-

lytically than its predecessor, its tendence is

similar—to suggest complexity of interconnec-

tion. I will not attempt to draw all of these

connections together here, but finish with the

conjecture that this analysis concerns itself with

only a few fragments of the immeasurably more

complex pattern by which ruling groups, as they

are sustained and replaced over time, leave their

marks on publics and the policies by which they

are governed.

Representation 349

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

South Africa’s Negotiated Transition: Democracy, Opposition, and the New Constitutional Order

Courtney Jung and Ian Shapiro

In 1965, Robert Dahl remarked that ‘‘of the

three great milestones in the development of

democratic institutions—the right to participate

in governmental decisions by casting a vote, the

right to be represented, and the right of an

organized opposition to appeal for votes against

the government in elections and in parliament—

the last is, in a highly developed form, so wholly

modern that there are people now living who

were born before it had appeared in most of

Western Europe.’’7 He might perhaps have

added that opposition is also the milestone that

has been least studied by contemporary political

scientists. . . .

Although the notion of a loyal opposition

finds its origins in monarchical rather than dem-

ocratic politics, democratic systems rely on insti-

tutionalized oppositions, and it is doubtful that

any regime could long survive as minimally

democratic without them.10

This is true for at least three related reasons.

The first is functional, having to do with the

peaceful turnover of power by governments that

lose elections. . . . If democratic politics is seen

as requiring at a minimum that there be turnover

of power among elites, then there must be sites

for counterelites to form and campaign as po-

tential alternative governments. Such opposition

requires the permissive freedoms of speech and

association as well as the presence of institutions

and practices that make it possible for counter-

elites to organize and inform themselves so as to

be able to contest for power. . . . If the opposition

is not thus perceived, then the possibility of

turnover is diminished and crises for the govern-

ment are correspondingly more likely to become

crises for the democratic regime.

This suggests a second reason why opposi-

tion institutions matter: for the legitimacy of the

democratic political order. Providing the institu-

tional space for opposition is essential for ensur-

ing that discontent and dissatisfaction can be

directed at particular governments rather than at

the democratic regime itself . . . Unless there are

such institutional outlets for dissent within the

regime’s institutions, those who are discontented

with the status quo may not even distinguish the

government from the regime, undermining the

possibility of the ebb and flow of competitive

party politics that democracies require. . . .

Last, institutional arrangements that facili-

tate loyal opposition perform important public-

interest functions in democracies. They are

necessary to ensure the presence of healthy po-

litical debate. They encourage competition over

ideas among elites and counterelites, and this

leads to demands for reason giving and coher-

ence in public debate. Moreover, opposition

institutions empower groups and individuals

who have an interest in asking awkward ques-

tions, shining light in dark places, and exposing

abuses of power. Without an organized political

opposition that has rights to information and

other resources, governing in secret becomes all

too tempting for the administration of the day.

Governments always have incentives to cam-

ouflage mistakes or controversial decisions that

might otherwise threaten their popularity and

to misuse the perquisites of o‰ce. Unless their

members know that they can be called to public

account for their actions, the temptation to act

Excerpted from: Courtney Jung and Ian Shapiro,

‘‘South Africa’s Negotiated Transition: Democracy,

Opposition, and the New Constitutional Order.’’ Poli-

tics and Society 23(3): 269–308. 6 1995 Sage Pub- lications, Inc. Reprinted by permission of Sage

Publications, Inc.

7. Robert A. Dahl, Political Opposition in Western

Democracies (New Haven: Yale University Press,

1965), p. xiii.

10. Barrington Moore, Jr., Liberal Prospects under

Soviet Socialism: A Comparative Historical Perspective

(New York: Averell Harriman Institute, 1989), pp. 8,

25. . . .

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

on these incentives will in many cases prove

irresistible.

II. The Opposition/Nonopposition Continuum

Although the principal alternative to democratic

systems that institutionalize opposition may be

authoritarian systems that repress it, many sys-

tems of government are designed to channel dis-

sent and opposition away from national political

institutions and otherwise to di¤use it. Such sys-

tems are usually defended on the grounds that

disagreements in the society in question are so

potentially explosive that anything else will re-

sult in the war of all against all. Many ortho-

dox prescriptions for ‘‘divided’’ societies—those

in which unbridgeable divisions among the

population are thought to preclude pluralist

politics—take this form. Perhaps the best known

of these is Lijphart’s consociational model of

minority vetoes and enforced coalitions, which

might be thought of as occupying a position at

the far end of a continuum from the opposition

model just sketched.13

Consociational models of democracy empha-

size participation and representation to the vir-

tual exclusion of opposition. Their organizing

principles are proportionality in the electoral

system, an institutional structure that forces

power sharing and a system of mutual vetoes

among ‘‘a cartel of elites,’’14 and a predisposi-

tion toward robust federalism to insulate ter-

ritorially based minorities from the power of

whoever controls national political institutions.

Advocates of consociational democracy contend

that particularly when societies are deeply di-

vided along ethnic lines, if democracy can be

realized at all it will be only if the e¤ects of ma-

jority rule are mitigated by institutional devices

of this sort. The assumption is that ethnic divi-

sions so completely overdetermine other con-

flicts, and that they are so intense and enduring

(if not primordial), that the only viable institu-

tional recipe is one that is designed to minimize

political competition and keep the groups from

getting at one anothers’ throats.

Consociational systems undermine the func-

tional, legitimacy-generating, and public-interest

roles of opposition discussed in section I. With

regard to the first, consociational systems are not

designed to foster alternation of major parties in

power. Instead, they permit the same combina-

tion of elites to entrench themselves at the peaks

of spoils and patronage hierarchies more or less

continuously. The democratic benefits that can

accrue from ‘‘tossing the rascals out’’ are there-

fore unavailable. On the legitimacy front, be-

cause consociational arrangements ensure that

every major political player is part of the gov-

ernment, there is little basis for the disa¤ected to

di¤erentiate the government from the regime. It

is hard for an ethos of loyal opposition to de-

velop when there is no institutional outlet.15

Those who are not in government are removed

from politics altogether, making it more likely

that they will turn to extra-institutional politics if

they can. With regard to the public-interest role

of opposition politics, consociational systems do

not give powerful parliamentary players incen-

tives to keep government honest by shining light

in dark corners. Because consociational systems

require high levels of consensus among governing

13. See Arend Lijphart, Democracy in Plural Societies

(New Haven: Yale University Press, 1977), and Power-

Sharing in South Africa (Berkeley: University of Cali-

fornia Institute of International Studies, 1985).

14. Arend Lijphart, ‘‘Consociational Democracy,’’

World Politics 4, no. 2 (January 1969), pp. 213–215,

222.

15. This insight is at the core of Horowitz’s critique of

consociational systems: that the very circumstances in

which they are most needed—when there are powerful

divisions in the society—are the circumstances when

they are also least likely to be e¤ective; see Donald L.

Horowitz, A Democratic South Africa? Constitutional

Engineering in a Divided Society (Berkeley: University

of California Press, 1991), pp. 142–143.

Representation 351

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

elites, the only way for anyone in the governing

elite to get her way on issues that matter in-

tensely to her is to give other powerful players

what they want on issues that matter intensely

to them. Therefore, mutual vetoes can be ex-

pected to lead to mutual logrolling, rather than

to political confrontations among elites, and to

promote insider clubism. By the same token,

logrolling minimizes the likelihood that govern-

ment elites will be called to account by members

of potential alternative governments. In short,

consociational systems, based on the politics of

elite coalitions, maximize both representation

and participation in government, but at the price

of almost complete abandonment of a viable

opposition politics.

The position of a regime along the opposition/

nonopposition continuum depends on the elec-

toral system and on the divisions of power within

the legislature and between the legislature and

the executive. First-past-the-post electoral sys-

tems based on single-member districts are likely

to produce two-party systems and, as a result,

strong parliamentary oppositions.16 These op-

positions may or may not have the potential to

become alternative governments (depending on

their potential grass roots support). Nonetheless,

they will have an interest in becoming magnets

for antigovernment sentiment, which should give

them incentives to perform at least some of the

conventional functions of loyal oppositions:

organizing and channeling dissent, disseminating

information, and exposing government corrup-

tion. Thus although two-party–dominated plu-

rality systems have high barriers to entry and

are comparatively unrepresentative, they pro-

duce significant institutionalized opposition.

Things will be less predictable in multiparty

systems dependent on coalition governments.

Given the well-documented propensity of pro-

portional representation to produce party frac-

tionalization, particularly when constituencies

are large and thresholds are low,17 uncon-

strained proportional representation makes par-

liamentary structure unpredictable. In certain

circumstances it may produce strong opposition

coalitions and alternation in government, as is

sometimes the case in Israel and The Nether-

lands. In other circumstances, such as Austria

and Switzerland, the major parties will often

govern in coalition. Given the unpredictability of

unconstrained proportional representation ver-

sus the predictability of first-past-the-post plu-

rality systems in generating two-party politics,

one would not choose the former over the latter

if the goal was to produce oppositional rather

than consociational parliamentary politics.

Whether a system is presidential or parlia-

mentary also a¤ects the prospects for institu-

tionalized opposition. Parliamentary systems

link the fortunes of the executive to those of the

majority party in parliament, so that institution-

alized opposition is apt to be centered in the

legislature. In presidential systems, significant

opposition may occur between the executive and

the legislature, depending on the powers of the

presidency (e.g., whether they include the power

to appoint a cabinet or to conduct some a¤airs

unilaterally) and the interactions between the

electoral rules for parliament and presidency.

An electoral system that produces a govern-

ment consisting of all the powerful players in the

legislature is likely to weaken the presidency,

reducing the president, in the limiting case, to a

figurehead. If the president’s party is a minority

16. Douglas W. Rae, The Political Consequences of

Electoral Laws (New Haven: Yale University Press,

1967).

17. Arend Lijphart, Electoral Systems and Party Sys-

tems (Oxford: Oxford University Press, 1994). Rae

suggests that with proportional representation, small

and medium-sized constituencies and higher thresholds

combine to limit fractionalization and give parties

incentives to form blocs, whether in government or

opposition. In e¤ect the constraints make multiparty

systems operate more like two-party systems: see

Douglas W. Rae ‘‘Using District Magnitude to Regu-

late Political Party Competition,’’ Journal of Economic

Perspectives 9, no. 1 (winter 1995), pp. 65–75.

Chapter 6 352

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

in the legislature, as is often the case in Brazil for

instance, he has to seek support from the major-

ity (opposition) party to avoid governmental

paralysis. This type of system tends closely to

approximate a power-sharing type of arrange-

ment. If, however, the constitution confers sub-

stantial autonomous power on the president, as

it does in the United States, the president may be

an institutional figure to be reckoned with even if

his party is a minority in the legislature. For

presidential systems to be relatively close to the

opposition end of the continuum in virtue of

their presidentialism, then, either presidents must

be institutionally powerful or there must be sig-

nificant minority parties in the legislature with

which they can ally themselves. As the latter is

less predictably so in proportional representa-

tion systems, it is less predictably the case in

presidential proportional representation sys-

tems. Perhaps for this reason, observers have

noticed a coincidence of presidential propor-

tional representation systems and power-sharing

governments.18 . . .

18. Scott Mainwaring, ‘‘Presidentialism, Multiparty

Systems, and Democracy: The Di‰cult Equation,’’

Comparative Political Studies 26, no. 2 (1993), pp.

198–230; Arend Lijphart, ‘‘Democratization and Con-

stitutional Choices in Czechoslovakia, Hungary, and

Poland, 1989–1991,’’ Journal of Theoretical Politics 4,

no. 2 (1993), p. 209.

Representation 353

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

The Representation of Women

Anne Phillips

In the countries that lay claim to the title of de-

mocracy, women have enjoyed many decades of

formal equality, sharing with men the right to

vote, to stand in elections, to compete for any

o‰ce (political—not yet religious) in the land.

Their participation in voting is now much the

same as men’s. Yet almost regardless of the date

at which women won their rights (ranging from

1902 in Australia, 1919 in West Germany, 1920

in the USA, 1928 in the UK, to the much

delayed 1971 in Switzerland), there has been a

marked consistency in the figures for female

participation in national and local politics. With

the major recent exception of the Nordic coun-

tries (to which I shall return), women figure in

national politics at something between 2 and 10

per cent; in Britain and the USA, women have

found it notoriously hard to break the 5 per

cent barrier.

Figures for local politics are only marginally

more promising. By 1983, women representa-

tives had captured 13 per cent of the seats on

West German local councils; 14 per cent on the

French conseils municipaux; 14.4 per cent on

county councils in England and Wales; 11.1 per

cent on regional councils in Scotland; and 7.9

per cent on district councils in Northern Ireland

(Lovenduski 1986). By 1985, women made up 14

per cent of the membership of municipal and

township governing boards in the United States,

but had been elected to mayor in only four of the

hundred largest cities (Randall 1987:105). The

percentages are nothing to write home about,

and women’s relatively higher profile in local

politics only confirms what is frequently ob-

served: that the numbers rise where the power of

the o‰ce is less. We all know there are more men

than women in politics, but the details still come

as a shock: only forty-three women out of 650

members of the British parliament? Only twenty-

eight women out of 435 members of the US

House of Representatives? What kind of democ-

racy is this?

Liberal democracy makes its neat equations

between democracy and representation, democ-

racy and universal su¤rage, but asks us to con-

sider as irrelevant the composition of our elected

assemblies. The resulting pattern has been firmly

skewed in the direction of white middle-class

men, with the under-representation of women

only the starkest (because they are half the pop-

ulation) among a range of excluded groups. The

campaign for women’s right to vote was always

linked to a parallel campaign for women’s right

to be elected. Success in the first has not brought

much joy in the second. . . .

The under-representation of women within

conventional politics is nonetheless crucial in

thinking about democracy and gender. The

general critique of liberal democracy leaves a

teasing vacuum on what could serve as alter-

natives, while the questions raised over the two

most common alternatives suggest that neither

can be simply adopted in its place. We can per-

haps move on to more substantial ground if we

examine more closely the weaknesses (and pos-

sibly strengths) of current liberal democratic

practice. What does the under-representation of

women add to the understanding of democracy?

It shows that there is a problem undoubtedly,

but is the problem then in the theory or ap-

plication? Setting aside for the moment what

may be more fundamental problems with liberal/

representative democracy, can we anticipate a

trend towards sexual parity? Is there a theo-

retical problem with the ‘‘representation of

women,’’ an incongruity between this and the

assumptions of liberal democracy? . . . [D]oes

feminism provide us with a novel angle on these

issues, a di¤erent way of conceiving either possi-

Excerpted from: Anne Phillips, Engendering Democ-

racy. Cambridge: Polity Press, 1991. Reprinted by

permission.

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

bilities or limits? There are two major aspects of

this, which I shall discuss in separate sections.

What are the theoretical issues implied in the

notion of representation? What are the chances

of electing more women?

‘‘Mirror’’ Representation

Confront people with the damning evidence on

the number of women elected and they tend to

divide into those who think this matters and

those who say it does not. Much of the dis-

agreement reflects the complacency, not to say

dishonesty, of those who enjoy a monopoly of

power, but there are more intriguing issues at

stake. As with many feminist demands, the case

for greater parity in politics has been made in

three ways (Hernes 1987). Part of it relies on a

notion of basic justice, and fits within a broad

sweep of arguments that challenges sexual segre-

gation wherever it occurs. Just as it is unjust

that women should be cooks but not engineers,

typists but not directors, so it is unjust that they

should be excluded from the central activities in

the political realm; indeed, given the overarching

significance of politics, it is even more unfair that

women should be kept out of this. But for the

hundred years and more that access to political

power has been an issue, women’s organizations

have combined the case for justice with at least

one additional point. Sometimes the argument is

that women would bring to politics a di¤erent set

of values, experiences and expertise: that women

would enrich our political life, usually in the di-

rection of a more caring, compassionate society.

A more radical version is that men and women

are in conflict and that it is nonsense to see

women as represented by men.

The case for justice says nothing about what

women will do if they get into politics, while the

two further arguments imply that the content of

politics will change. All unite in seeing a sexual

disproportion between electors and elected as

evidence that something is wrong. The striking

homogeneity of our existing representatives is

proof enough of this, since if there were no sub-

stantial di¤erences between men and women, or

between black people and white, then those

elected would undoubtedly be a more random

sample from those who elect. Consistent under-

representation of any social category already

establishes that there is a problem. Such a

marked variance from the population as a whole

could never be an accidental result. Leaving

aside as mere prejudice the notion that women

are ‘‘naturally’’ indi¤erent to politics, there must

be something that prevents their involvement.

The argument from justice then calls on us to

eliminate or moderate whatever obstacles we

find to women’s participation, while the argu-

ments from women’s di¤erent values or di¤erent

interests go one stage further. The sexual dif-

ferentiation in conditions and experience has

produced a specifically woman’s point of view,

which is either complementary or antagonistic

to the man’s. Any system of representation

which consistently excludes the voices of women

is not just unfair; it does not begin to count as

representation.

All three arguments are at odds with what

has become the orthodoxy, for while there are

a number of competing versions on o¤er, the

idea that representatives should in some way

‘‘mirror’’ those they represent is probably the

most contested. . . .

The dominant practice in most contemporary

democracies is a muddled combination of both

accountability and autonomy. Our representa-

tives are said to represent our views (political

parties present us with alternative policies, and

we make a choice between them), but only in the

vaguest of ways (election manifestos o¤er bland

generalities, and those elected then fill in the

details themselves). Those elected are seen as

carrying some responsibility for their area, but

are not permitted to take this too far, for they

are ultimately bound by party lines. On any of

the major social or demographic characteristics

(age, sex, race, class) they do not represent us

at all. Taking the example of British Members

of Parliament: lawyers make up the largest

Representation 355

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

single occupational group; women have only just

pushed beyond five percent of the total; and the

proportion of the population that is nonwhite

is currently ‘‘represented’’ by a mere handful of

MPs, whose election in 1987 marked the first

substantial breach in the white monopoly. . . .

Representative democracy cannot produce a

perfect reflection of society: the only guarantee

of that would be all the citizens meeting together

in national assembly. Within the limits of repre-

sentation, it is hard to see how to get agreement

on the categories to be covered. Even where such

agreement becomes possible, proportionality in-

evitably reduces local autonomy, for it must in-

volve some form of national party directive over

the kind of candidates each constituency should

choose. But arguments that rely on the impossi-

bility of one extreme in order to justify its op-

posite are always suspect, and as long as those

who speak for us are drawn from such an

unrepresentative sample, then democracy will

remain profoundly flawed. The obstacles that

deny certain people the chance of election are as

undemocratic in their way as the laws that once

excluded them from the right to vote. And

moving on to the more positive point, di¤erent

experiences do create di¤erent values, priorities,

interests; while we may all be capable of that

imaginative leap that takes us beyond our own

situation, history indicates that we do this very

partially, if at all. Those who regard the current

situation with complacency are not too far in

spirit from the nineteenth-century apologists of

male su¤rage, who claimed that a man spoke

for himself and ‘‘his’’ woman, and thus that the

woman had no need for a separate voice. Where

there are di¤erent interests and di¤erent experi-

ences, it is either naive or dishonest to say that

one group can speak for us all. . . .

It is easy enough to show that women are

under-represented in politics, and not too much

more di‰cult to make the case that women are

oppressed. On both counts something should

surely be done. The disproportion between those

elected and those who elect is too astounding to

be attributed to accident, while the fact that it

serves those who are already advantaged is too

striking for any democrat to ignore. The di‰-

culties arise in the next stage, for within the

framework of representative democracy, it is

political parties that have provided the vehicle

for representation, and in its more substantial

sense, the representation of women does not fit.

‘‘Women’’ are not homogeneous and do not

speak with a single voice. . . .

. . . Those who resist an increase in women’s

representation no doubt harbour similar fears,

anticipating that the women elected will alter the

game. I hope there will be changes, but in the

interests of democratic accountability these have

to take place in the open, through the decision-

making processes of each party and the publicity

of electoral campaigns. We cannot jump too

easily into the notion that there is an interest of

women; and short of women’s constituencies or

women’s elections, there is no clear mechanism

for their representation.

The relative autonomy of politics

Let me now turn the argument the other way. I

have said that the case for more women repre-

sentatives rests partly on a notion of equality in

participation and partly on a notion of reflection,

and that neither in principle nor in practice can

this guarantee that women are then represented

as women. We can anticipate that more women

elected will shift the context and priorities of

public policy, but cannot say this is necessarily

so, nor indeed presume that their choices would

win general acclaim. The problem may be even

more acute. If feminists are right in the critiques

they have developed of liberal democracy, then

the nature of the public/private divide may make

even equal numbers an impossible dream. If the

constraints on women’s political activity are set

by their economic and social position, then the

idea that you could have a parliament made up

equally of men and women, without also having

substantial changes in social relations, is non-

sensical. The absurdity lies not in thinking that

Chapter 6 356

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

women would prove no di¤erent from men, but

in imagining that they could ever get into such

positions of power until after the changes were

made. A democracy may require the policies on

equality before the equality of representation—

and yet without the second what hope is there of

the first? . . .

Explanations for women’s relative invisibility

in politics tend to be multi-causal—which is

probably why I used to find them tedious, for

they lack the drama of the singular cause.

Women have been socialized into regarding pol-

itics as an alien a¤air; women are constrained by

their responsibilities as mothers, and indeed their

general role as carers for the young, sick and old,

all of which make the commitments of full-time

politics an almost impossible choice; women are

under-represented in the jobs that most favour

political careers; women are discouraged by the

attentions of a hostile media; women are actively

excluded by the male selectorates who guard the

gateways to political life (Randall 1987). Within

the general tendency to give due weight to each

of these factors, there are still important varia-

tions in emphasis. Some writers have focused

more on what stops women coming forward, and

others on the male barriers when they none-

theless do (Lovenduski and Hills 1981 is an ex-

ample of the latter). And some writers focus

attention on political conditions, while others

identify the social and economic obstacles to

change.

In the more ‘‘political’’ category are the argu-

ments about proportional representation, which

have been most associated in the British context

(they are less of an issue elsewhere) with the

work of Elizabeth Vallance (1979) and more

recently Pippa Norris (1985). Single-member

constituencies with a first-past-the-post system

for selecting the winner are said to stack the

cards in favour of the male, for in our sex-

divided patriarchal societies, the man is the

norm, the woman peculiar. . . .

. . . [I]n a multi-member constituency the

selectorate might take the risk—and indeed,

when they are choosing not one but three or five

candidates, it begins to look odd if they are all of

them the same. In particular, where elections are

organized through party lists there is consider-

ably more scope for dictating sexual composi-

tion. It becomes possible to set a minimum quota

for each sex in the winnable positions, or, even

more ambitiously, to put men and women in al-

ternate order on the list and produce parity in

the final results. The evidence from Europe is not

decisive but certainly suggestive. Those countries

which rely on the single-member, first-past-the-

post system have found it hardest to advance on

the 5 per cent token representation. Most of those

with a system of multi-member constituencies

and party lists edge towards a 10 per cent repre-

sentation of women. All those that stand out as

spectacularly progressive employ proportional

representation with party lists (Norris 1985;

Lovenduski 1986; Haavio-Mannila et al. 1985).

General arguments for and against propor-

tional representation establish the di‰culties in

saying what is most democratic. The single

transferable vote system allows voters to indi-

cate a hierarchy of preferences, and protects us

against a situation in which party A wins with a

minority of votes, when parties B and C are be-

tween them the first and second preferences of

the majority. The implication, however, is that

the party that ends up with most seats could be

one that has fewer first preference votes—and

if people really wanted their first choice, but

only marginally preferred their second to their

third, this could be a compromise that leaves the

majority unhappy. The multi-member constitu-

ency can iron out some of the inequalities in first-

past-the-post systems, but the smaller parties

within each constituency may still end up with

no members in parliament even if their support

through the nation turned out quite impressive.

The additional or mixed member system tries to

deal with this by leaving some (usually half )

seats selected on the first-past-the-post system,

but distributing the remainder in proportion

to the parties’ share of the national vote. The

Representation 357

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

problem here is that if the way we vote in our

constituencies is still premised on our perception

of which parties have the most chance in our

area (we don’t simply vote for what we most

want, but for what we prefer given what we

think others will do) then the figures for national

voting cannot be presumed to reflect genuine

preference. And, as opponents of any of these

systems will argue, proportional representation

can give undue influence to the small party that

then holds the balance of power. Getting a ‘‘fair’’

reflection of parties in parliament does not guar-

antee fair reflection in government.

In terms of the ‘‘mirror’’ e¤ect of reflecting the

population, however, some form of proportional

representation wins hands down over first-past-

the-post systems (and when the New Zealand

Commission on Electoral Reform recommended

the abolition of Maori seats, it was with the

proviso that the country should switch to a

mixed member proportional system, thereby

increasing Maori chances of election). In her

study of ‘‘Women’s legislative participation in

Western Europe’’ (1985) Pippa Norris compares

twenty-four liberal democracies, including eigh-

teen in Western Europe, to assess the correlation

between the numbers of women elected and in-

stitutional, cultural and socioeconomic condi-

tions. The institutional di¤erences refer to the

electoral system: whether the country operates a

party list proportional system or some version of

majoritarianism, including first-past-the-post, al-

ternative vote and single transferable vote sys-

tems. Cultural di¤erences were defined by the

strength of Catholicism versus Protestantism (the

former seen as more traditionalist for women),

and with the help of data contributed by nine

countries in the European Community in a sur-

vey on attitudes to sexual equality in politics.

Socio-economic di¤erences were roughly identi-

fied by the proportions of women in the labour

force, their proportions in higher education and

their proportions in the professions.

As in any cross-national survey, the measures

are a bit rough and ready, but the results proved

striking enough. Di¤erences in electoral systems

emerged as by far the most significant in rela-

tion to the election of women, followed at some

distance by positions on the index of political

egalitarianism. The dominant religion proved

insignificant, as did socio-economic conditions.

Among the notable individual examples, Swit-

zerland is particularly convincing. Women were

not granted full equal rights till 1971, and the

society is so traditional in its views that as late as

1982 male citizens were still voting in one canton

to deny women the vote in local elections. Swiss

women nonetheless weigh in at over 10 per cent

in the Nationalrat—under a system of propor-

tional representation (Norris 1985:99).

Ten per cent is no great cause for rejoicing, and

the fact that a particular electoral system can

marginally enhance women’s participation does

not take us very far down the road. Looking be-

yond the clear evidence that proportional repre-

sentation o¤ers more favourable conditions for

women to be elected, what dictates when real

parity will be won? Is it a matter of women’s

political mobilization, both inside and outside of

party politics, which can make sexual equality a

question of public concern? Does it depend in

any way on shifts in economic and social ar-

rangements: an equalization in male and female

rates of participation in the labour market? a

desegregation in the jobs that men and women

typically do? Is the heart of the matter the con-

ditions under which children are cared for and

reared, so that as long as this remains the private

responsibility of invisible women, there is limited

time for them to invade men’s political sphere?

Coming as I do from a tradition that has viewed

political equalities as a reflex, however compli-

cated and over-determined, of social and eco-

nomic equalities, I thought for many years that a

social revolution was the only real answer. As

long as women are positioned as the dependants

of men—a positioning which seems to rest on

such an interlocking of social, economic and

cultural forces that it is hard to know where to

begin—it had seemed inconceivable to me that

Chapter 6 358

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

men and women could take equal parts in the

political realm. Add on to this the more spe-

cifically feminist analysis of the relationship

between public and private, which sees the de-

termined amnesia over the domestic domain as

central to liberal democracy, and the facts re-

quired no further explanation. What else could

one possibly expect? What has shifted me to-

wards the more political end of the spectrum—

to the belief, that is, that greater political

equality is possible even pending fundamental

social transformation—is the recent experience

of the Nordic countries.

Women in Nordic Countries

Up till the 1960s there was not much to choose

between any of the countries in Western or

Northern Europe: a somewhat more damning

obstructionism in first-past-the-post systems; a

more conservative tendency where the church

held its sway; generally just the deafening domi-

nance of men. By 1984, however, women had

taken 15 per cent of the parliamentary seats in

Iceland, 26 per cent in Norway and Denmark, 28

per cent in Sweden and 31 per cent in Finland

(Lovenduski 1986:152). In 1985, Norway took

the world record. Women made up 34.4 per cent

of the Storting (the national assembly), held

eight out of eighteen cabinet posts, contributed

40.5 per cent of the membership of county coun-

cils and contributed 31.1 per cent of the mem-

bership of municipal councils. The relatively

poor performance of women in Iceland further

confirms the role played by electoral systems, for

while Iceland shares with other Nordic countries

the practice of proportional representation, the

small number of constituencies combines with a

large number of parties to mean that few parties

can anticipate winning more than two seats in

any one constituency. As in so many countries,

women are typically nominated to the ‘‘orna-

mental’’ positions on a party’s list, with the first

two places secured to men. Their position is thus

more comparable to that in first-past-the-post

systems (Skard and Haavio-Mannila 1985b).

Taking these countries as a whole, what is

striking is that, despite significant economic dif-

ferences with countries in Western Europe, the

pattern of women’s employment and the pre-

sumption of women’s primary responsibilities for

children are not so many miles away from the

European norm. . . .

. . . The social position of women is still in-

formed by the statistically anachronistic but

nonetheless powerful assumption that men oc-

cupy the world of work and women the world of

the home.

The contrasts seem puzzling indeed. In 1988 I

attended a forum that brought together feminists

from Western and Northern Europe, centring

much of its discussion around the impact of

quota systems in increasing the political profile

of women. The women from West Germany

were still digesting the causes and consequences

of the recent conference decision of the German

Social Democratic Party (SPD), which just that

year had committed itself to a minimum of 25

per cent women among its Bundestag members

at the next election, progressing to a minimum of

40 per cent in the course of the next ten years.

Feminists from Britain could hardly begin to

conceive the conditions under which comparable

decisions might be reached in Britain. By con-

trast, those from Norway or Sweden had experi-

enced quota systems for so long that they could

barely remember when they won this battle—or

indeed if it was a battle at all!

The pattern, in fact, was as follows. In the

1970s a number of Scandinavian parties adopted

the principle of at least 40 percent female repre-

sentation at all levels of elected delegation within

the party itself. This was adopted by the Swedish

Liberal Party and the Swedish Communist Party

in 1972, by the Norwegian Liberal Party in 1974

and the Norwegian Socialist Left Party in 1975,

and by the Danish Socialist People’s Party in

1977 (Skard and Haavio-Mannila 1985a). In the

1980s there was a push to extend this to the level

of female representation in national parliaments.

In 1980, parties in both Norway and Sweden

Representation 359

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

proposed legislation that would commit all po-

litical parties to a minimum of 40 per cent women

on their electoral lists; failing the success of this

bid, various parties introduced the practice uni-

laterally (Skard and Haavio-Mannila 1985b). In

1983, for example, the Norwegian Labour Party

introduced a 40 per cent quota for candidates in

local and national elections and, as the largest

party after the 1985 elections, contributed sig-

nificantly to the lead that Norway then assumed.

Among the parties that remain ideologically

opposed to the principle of a quota system, there

has also been substantial movement: the Norwe-

gian Conservative Party espouses ‘‘competence’’

in opposition to formal quotas, but women

nonetheless make up 30 per cent of its current

national representation (Skjeie 1988). Again this

seems to confirm the importance of specifically

political factors. Once a few parties have put

their toes in the water, the exigencies of compe-

tition require that the others do not lag too far

behind.

The initial moves towards increasing women’s

representation pre-dated—or at the latest coin-

cided with—the emergence of the contemporary

women’s movement; indeed, with the exception

of Denmark and to a lesser extent Iceland, the

Nordic countries were rather slow to develop the

characteristically counter-cultural politics of sec-

ond wave feminism. This itself is rather daunting

in its implications. Through most of Europe, the

women’s movement showed little enthusiasm for

the conventionalities of representative democ-

racy, and it was not till quite late in the 1970s

that feminists began to look to existing political

parties as an important arena for change. Many

parties had their own women’s organizations,

often dating from the late nineteenth century

when women members were mobilized to per-

form various supportive or fund-raising activ-

ities. At key periods, these women’s sections had

worked hard to raise feminist issues, but this

early history was lost to most of the activists in

the new women’s movement; outside of the

Nordic countries, connections were rarely made.

Yet in Sweden, Norway and Finland, it was

the women’s sections in the social democratic

parties that began to make the running, and in a

chronology that has no parallel in the rest of

Europe, it is the 1950s that are described as the

‘‘decade of the build-up’’ and the 1960s as the

‘‘explosive decade’’ (Eduards et al. 1985:136). . . .

The remarkable levels of female participation

in these countries can I think be explained by a

combination of three factors. The first is the en-

abling condition of their systems of proportional

representation, which opens up but does not

guarantee more space. The second is the strength

of women’s organizations within the traditional

social democratic parties, and the political choices

made by feminists who have attached greater

significance to conventional power. The third is

the di¤erences between liberal and social democ-

racy, which hinge around a di¤erent relationship

between public and private spheres; women’s

position has been made a more explicit public

concern. Each of these begs questions about

what explains them in turn, but they combine to

highlight the importance of politics, and the

scope for specifically political change.

This does not mean the road is now clear. We

can draw some reassurance from the e¤ects of

example and competition and anticipate that the

process will speed up once the first stages are

won. But repeating the experience of one country

in another is not easy, and when Pippa Norris

identified the importance of proportional repre-

sentation she found herself correspondingly de-

pressed. ‘‘Given the many institutional barriers

to political equality, including a resistance to the

implementation of electoral reform in major-

itarian systems, it seems unlikely that cross-

national di¤erences in the political position of

women will diminish in the near future’’

(1985:100). All the trends seem to suggest a con-

tinuing growth in female participation in wage

employment, higher education and the profes-

sions. By early 1989, even the Conservative

Government in Britain had woken up to the

Chapter 6 360

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

anticipated shortages of skilled labour and burst

forth in a series of unlikely pronouncements

on the desirability of increasing the proportion

of women in higher education, of establishing

work-place nurseries (not of course financed by

the state), and encouraging more women to

work. But if the entry of women into higher ed-

ucation and paid employment has limited impact

on their numbers in elected o‰ce (as the example

of the United States reveals all too clearly), none

of these changes can be expected to make much

di¤erence. It is crass materialism to say that

what is ‘‘only’’ political is therefore more open to

change. On the other hand, if it is a matter of

politics we do at least know where to begin.

With all its limitations, representative democ-

racy is not necessarily inimical to the election of

women, and indeed gender may now prove less

intransigent than class. I do not hold out great

hopes for my own country, but find it entirely

plausible that liberal and, more specifically, so-

cial democracies will witness a growing propor-

tion of women in politics, even pending that

social revolution I once thought a necessary

condition. But those elected will be peculiarly

skewed to a certain kind of woman who, like

the generations of men who went before her, will

be a well-educated professional, and devoted to

politics full time. Even in the Nordic countries,

the marked shifts of recent years do not guaran-

tee access for all women, any more than previous

patterns of elected and corporate representa-

tion guaranteed access for all men. The women

elected to parliament may be thoroughly ‘‘un-

representative’’ in terms of their class, their in-

come, the number and age of their children, or

whether they previously worked full or parttime.

The fact that gender quotas are increasingly

accepted is no doubt a reflection of this: with

women distributed (however unevenly) across

the range of occupations and professions, they

can be incorporated into our representative

assemblies without disturbing the conventions

of competence and leadership, and without dis-

rupting the dominance of class. However distant

the prospect of a gender quota may seem to

feminists in Britain or the United States, it is at

least conceivable. No one even talks of a formal

quota for class.

Representative democracy may prove itself

more amenable than I once thought to the elec-

tion of women, but it has trouble with their

‘‘representation.’’ Grounded as it is in a tradi-

tion that rests on the abstract individual, liberal/

representative democracy has to define politics as

the realm of public rationality in which we con-

test opposing ideas. The least democratic ver-

sions of this will leave the bulk of issues to be

settled by those elected, and the electorate as a

whole will be permitted to make its occasional

foray into the voting booth only to indicate a

preference over who these people should be.

More radical versions will try to commit the

representatives to an explicitly delineated set of

policies and ideas, so that we will not only be

able to ‘‘punish’’ those who disappoint us by

not voting for them the next time round, but

can more positively influence the decisions they

make. Here the emphasis will be on party con-

gresses, definite commitments, explicit proposals,

a heightened profile for ideas.

The representation of women as women po-

tentially founders on both the di‰culties of

defining the shared interests of women and the

di‰culties of establishing mechanisms through

which these interests are voiced. It has been

noted that women politicians are often reluctant

to see themselves as representing women. While

we may regret this refusal of feminist concerns,

we cannot jump straight to the opposite camp.

Feminism should not give unwitting support to

a version of democracy that rests too exclu-

sively on trust, as if merely by virtue of their sex

women can presume a mandate to speak for us

all. The representation of women as women does

not fit within the framework of representative

democracy, and while this may count as ammu-

nition in the battle for democracy of a di¤erent

kind, it should not be glossed over in discussions

of change.

Representation 361

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .

References

Bashevkin, Sylvia (ed.) 1985: Women and Politics in

Western Europe. London: Frank Cass.

Eduards, Maud, Halsaa, Beatrice and Skjeie, Hege

1985: Equality: how equal? In Haavio-Mannila et al.

1985.

Graubard, Stephen R. (ed.) 1986: Norden: The Passion

For Equality. Norwegian University Press.

Haavio-Mannila, Elina, Dahlerup, Drude, Eduards,

Maud et al. 1985: Unfinished Democracy: Women in

Nordic Politics. Pergamon.

Hernes, Helga Maria 1987: Welfare State and Woman

Power: Essays in State Feminism. Norwegian Univer-

sity Press.

Lovenduski, Joni 1986: Women and European

Politics: Contemporary Feminism and Public Policy.

Wheatsheaf.

Lovenduski, Joni and Hills, Jill (eds.) 1981: The Poli-

tics of the Second Electorate: Women and Public Par-

ticipation. Routledge and Kegan Paul.

Norris, Pippa 1985: Women’s legislative participation

in Western Europe. In Bashevkin 1985.

Randall, Vicky 1987: Women and Politics, 2nd edn.

Macmillan.

Skard, Torild and Haavio-Mannila, Elina 1985a: Mo-

bilization of women at elections. In Haavio-Mannila

et al. 1985.

——— 1985b: Women in parliament. In Haavio-

Mannila et al. 1985.

——— 1986: Equality between the sexes: myth or re-

ality. In Graubard 1986.

Skjeie, Hege 1988: The Feminization of Power: Nor-

way’s Political Experiment (1986– ). Institute for

Social Research, Norway.

Vallance, Elizabeth 1979: Women in the House.

Athlone.

Chapter 6 362

C op

yr ig

ht ©

2 00

3. M

IT P

re ss

. A ll

rig ht

s re

se rv

ed .