|
Strasbourg, 4 February 2004
Study no. 250/2003
|
CDL-AD(2004)003
Or. Fr.
|
|
|
|
EUROPEAN COMMISSION FOR DEMOCRACY THROUGH LAW
(VENICE COMMISSION)
REPORT
ON ELECTORAL SYSTEMS
OVERVIEW OF AVAILABLE SOLUTIONS
AND SELECTION CRITERIA
Adopted
by the Venice Commission
at
its 57th Plenary Session
(Venice, 12-13 December 2003)
on
the basis of comments by
Mr Christophe BROQUET
(expert, France)
Mr Alain LANCELOT
(substitute member, France)
TABLE OF
CONTENTS
INTRODUCTION............................................................................................................................ 4
PART ONE–ELECTORAL SYSTEMS ON OFFER..................................................................... 6
1. Category
1: Guiding principles and procedures
governing
the casting of votes............................................................................ 7
1.1
Direct and indirect universal suffrage,
multi-tiered elections......................................... 7
1.2 Plurality/majority systems,
proportional representation and “hybrid” systems....................... 7
a.
The plurality and majority systems............................................................................................ 7
b.
Proportional representation...................................................................................................... 8
c.
Hybrid systems........................................................................................................................ 8
1.3
Constituencies/Electoral
districts........................................................................................ 9
a.
Personal
circles........................................................................................................................ 9
b.
Territorial
constituencies........................................................................................................... 9
c.
The existence of several levels of
constituencies/electoral districts............................................ 10
1.4
Methods of voting........................................................................................................... 11
a.
“Categorical” voting methods................................................................................................. 11
b.
“Ordinal” voting methods....................................................................................................... 12
1.5
Number of rounds of voting............................................................................................. 14
a.
Single-round ballots............................................................................................................... 14
b.
Two-round ballots................................................................................................................. 14
2. Counting votes and distributing the seats................................................. 15
2.1
Distributing the seats among constituencies....................................................................... 15
2.2
Distributing the seats between political
formations............................................................ 16
a.
Distribution of seats in plurality and
majority systems............................................................... 17
b.
Distribution of seats in proportional
representation systems...................................................... 18
-
First stage of the distribution– the electoral quota or quotient................................................. 18
-
Second stage of the distribution – Distribution of the remaining seats and
votes
cast............................................................................................................................... 19
-
Allocating the seats in one operation: the divisor methods...................................................... 20
c.
Some specific methods for counting votes and
distributing seats............................................... 21
d.
Thresholds and bonuses......................................................................................................... 22
2.3............. Allocation of seats within lists.......................................................................................... 23
a.
Modes of expression and distribution...................................................................................... 24
b.
Subdivided constituencies....................................................................................................... 24
2.4............. Examples of past or present electoral methods................................................................. 26
3. Recall.................................................................................................................................. 33
PART TWO – CRITERIA FOR
SELECTING A PARTICULAR ELECTION
SYSTEM AND THE IMPLICATIONS
OF THAT CHOICE...................................................... 34
1. The three major functions of an electoral system............................. 34
1.1
Representation.................................................................................................................. 34
1.2
Selection........................................................................................................................... 35
1.3
Investiture......................................................................................................................... 37
2. The
three historical models for the FUNCTIONALITY
of
electoral systems................................................................................................. 38
2.1
The elitist model................................................................................................................ 38
2.2
The mass democracy model............................................................................................... 39
2.3
The consumerist individuation
model.................................................................................. 40
3. Advantages and drawbacks of hybrid systems........................................ 41
3.1
The search for the happy medium....................................................................................... 41
3.2
Drawbacks and difficulties................................................................................................. 42
4. Some considerations on the specific situation of
the emerging democracies.................................................................................... 43
4.1
Initial identification of the
nature and degree of segmentation............................................... 43
4.2
Choice of a model of
citizenship......................................................................................... 44
4.3
Choice of a system of democratic
governance.................................................................... 44
CONCLUSION............................................................................................................................... 45
Table
summarising the three historical models.......................................... 46
INTRODUCTION
1. In any political society whose size
necessitates a certain “division of labour” between those who are governed and
those who govern, the system of representation lies at the very heart of the
democratic system. The electoral system is at the very core because its task is
to translate the will of the sovereign people in designating their legitimate
representatives who are responsible,
on their behalf, for supervising executive and legislative acts and, while not
having a binding mandate, make themselves accountable in the next set of
elections.
2.
Stated in such broad terms,
these basic principles of democratic representation enjoy unanimous
accord. But that accord will always show
cracks or even completely disintegrate once it is attempted to define
procedures for implementation. Is there
just one “sovereign people” or several?
Does the people’s ethno-cultural make-up (national minorities) or at
least its variable presence on the territory (division into constituencies)
require special consideration? Must
representation be purely political in nature (party- or ideology-based) or must
it take cultural and social factors (religion, sex, social category etc) into
account? Must the need to arrive at a
governing majority be catered for prior to the election, by tolerating
“distortions” of plurality or majority voting, or afterwards, by allowing free
representatives to form coalitions in the light of the challenges and the
balance of power? Since parliamentary
elections do not have the same scope from one political system to another
(parliamentary, presidential or semi-presidential) and are not the sole
elections held in most democracies (there being infra-national and sometimes
supra-national elections), should they not be regarded as part of a broader
“electoral system”, in which several patterns of representation may coexist
and, accordingly, several voting methods, with each offsetting the effects of
the others? All these issues should be
focal points for specialists and politicians.
3. However, somewhat paradoxically, the
issue of electoral systems, which should attract the attention of all election
analysts, generally interests only a handful of specialists, albeit very
intensely, if not obsessively, so that they rapidly become fervent advocates or
relentless adversaries of proportional representation, leaving the rest totally
indifferent or perhaps somewhat contemptuous.
4. This paradox is doubtless largely due
to the necessarily amoral, and for some people immoral, side of electoral
systems which smacks of “political jiggery-pokery” and unfortunately detracts
from the purity of the immortal principles of the sovereignty of the people,
the expression of the general will and the legitimacy of the elective
authorities. But politics, even
democratic politics, is not just about principles. It is also the arena of
combat between all those who dream of conquering legitimate power. Law and
philosophy weigh in less heavily here than history and sociology, which prompts
us to agree with Mosca that when we say that the electorate elects its
representatives, we have got things out of perspective: “the truth is that the
elected representatives do all they can to be elected by their electorate”. And
the choice of a particular electoral system is clearly part of the armoury of
any political undertaking wishing to “have its representatives elected” in
order that the elected assembly should correspond most closely to its own
expectations and interests.
5. It must be acknowledged that the
variety of systems which specialists offer practitioners is such as to satisfy
their wildest dreams. The variety of systems on offer is not only so varied as
to be bewildering, it enables almost any result to be obtained, as if the
electors’ vote was ultimately less important than the sophistication of those
responsible for drawing up electoral legislation. Alain Lancelot has shown this
(Commentary, No 73, 1996) using simulations carried out on the basis of the
results of a French region and applying 41 differing electoral systems.
Depending on the system chosen, with 57 seats to be filled, the moderate right
obtained between 18 and 57 seats and the socialists between 4 and 30!
6. The fact that there is such a
“predictive” factor for election results means that attention ought to be paid
to it. This is the aim of this introductory report. The first part, written by
Christophe Broquet, seeks to bring some order into the plethora of electoral
systems available, and the second part, drawn up by Alain Lancelot, sets out to
identify the main criteria for choosing between the various systems in the
light of the few fundamental interests which each system takes into account in
its own particular way.
PART ONE – ELECTORAL SYSTEMS ON
OFFER
7. There is no predetermined uniform
classification of electoral systems. In general, electoral systems are divided
into three main types: majority or plurality/first-past-the-post, proportional
representation and hybrid. Yet within these major types, there is a virtually
unlimited number of voting methods. Just to take plurality and majority systems,
Frédéric Bon[1]
calculated that there were no fewer than eighty on the basis of general
criteria alone (number of rounds, counting methods, types of
constituency/electoral district). Consequently in order to give the best
possible account of the various different forms which electoral systems may take,
this report will concentrate on identifying their main features.
8. The definition of electoral systems
as “the set of procedural rules governing the expression of votes cast in a
given election and their conversion into seats”,[2] enables us in the first
place to classify those features in two main categories: the first consists of
all those factors relating to the organisation, the conduct and the process of
the election; the second consists of the rules relating to the counting of
votes and the distribution of seats.
9. The first category covers the rules
governing how the elector exercises his or her right to vote. As far as this
area is concerned, electoral systems have to provide a response to five questions,
and several responses are possible:
-
What electors do you want to call to the ballot box and hence what sort
of suffrage is to be applied? Direct universal suffrage, indirect universal
suffrage or multi-tiered elections?
-
What guiding principle should the vote be based upon? Is the intention
to give preference to the majority principle, proportional representation or
“hybrid” systems?
-
How is the electorate to be divided up between constituencies/electoral
districts?
-
What method of voting is to be available to electors? Here we can
distinguish between “categorical” voting methods (electors are asked to make an
absolute choice by indicating their preference for one party or one political
movement to the exclusion of all others) and “ordinal” ones (under which
electors may qualify their choice).
-
How many times must electors vote? This is a question of deciding how
many rounds the vote should be spread over.
10. The second category of rules deals
with how votes are counted and how seats are distributed. In this context, four
factors have to be taken into consideration:
-
the distribution of seats among constituencies/electoral districts,
-
the choice of method for allocating seats between the various lists,
-
the determination of thresholds for election and of bonuses designed to
ensure that assemblies elected by proportional representation obtain government
majorities and
-
the distribution of seats within lists.
1. CatEgorY 1: GUIDING PrincipLes AND
procEdures GOVERNING THE CASTING OF voTES
1.1 Direct
and indirect universal suffrage, multi-tiered elections
11. Before considering the different
features of the various electoral systems, it is necessary to define what type
of suffrage it is sought to adopt. Three types of suffrage are available in
democratic systems: direct universal suffrage, indirect universal suffrage and
multi-tiered elections.
-
direct universal suffrage: namely “where every citizen, except those
who are excluded by law, can directly participate in the vote”
and elect his or her representatives without the involvment of
intermediaires. This type of vote is
employed in practically all democracies for the purposes of electing
representatives to lower chambers.
-
indirect universal suffrage: elected representatives are chosen by
leading citizens, who may themselves be elected representatives. However,
unlike electoral colleges, the leading citizens in question are not chosen for
that purpose.
-
Multi-tiered elections: representatives are chosen by electoral
colleges chosen for that purpose. Nowadays this type of suffrage is no longer
used to elect representatives to lower chambers, but the President of the United States is elected by a system of
electoral colleges.
1.2 Plurality/majority, proportional
representation and “hybrid” systems
12. Every electoral system works according to a guiding principle. It varies
between two extremes: the plurality/majority systems and proportional
representation. Hybrid or intermediate electoral systems fall between the two.
a. The plurality
and majority systems
13. In the plurality and majority systems,
the candidate or list of candidates that obtained the most votes in the
conclusive round of voting is declared elected. This system is the one most
consonant with European cultural traditions. It is also the only system
possible where there is only one vacant seat per constituency. One of the
effects of the first-past-the-post system is that it enables clear government
majorities to be formed.
14. The plurality and majority systems
may take many forms according to how one defines “the most votes”. There are
three possibilities:
¨ Relative (or simple)
majority (plurality system/first-past-the-post in English): the candidate
obtaining the most votes in the conclusive round (the one on which the election
depends) is declared to have been elected.
¨
Absolute majority: defined as half plus one of the votes cast.
¨
Qualified (or reinforced) majority: this corresponds to a greater
majority than half plus one of the votes cast (majority equivalent to, for
example, two-thirds or three-fifths of the votes cast). This type of majority
is little used in elections based on direct universal suffrage.
15. Systems requiring a relative majority are more likely to result in a
two-party race or at least a concentration of seats in the hands of the leading
two parties, whereas systems requiring an absolute majority (in the first
round) are more open to multi-party coalitions.
b. Proportional
representation
16. Proportional representation may be
defined in terms of its objective, which is to seek to achieve “a
proportional transposition of votes into mandates”.
To date, three countries have opted for full proportional representation: Israel,the Netherlands
and Paraguay.
17.
Proportional representation is generally regarded as being the “fairest
system” on the grounds that it tends towards a more faithful representation of
the various political forces. However, the drawback with this system is that it
favours a fragmentation of those seeking to be elected and consequently makes
it more difficult to form stable majorities in the assemblies. This was shown
by the Polish legislative elections of June 1993: 29 parties were represented
in the lower house (460 members in all), but no party obtained more than 12.3%
of the seats.
18. Lastly, it should be noted that a
proportional representation system is not automatically synonymous with a list
system. The proportional system may be effectively used, albeit very rarely, in
combination with individual candidatures.
c. Hybrid systems
19.
Hybrid or combined systems is the term generally used to describe systems
presenting a combination of the first-past-the-post/majority and proportional
systems (Italy, Japan), and also systems in which proportional representation
is applied but the actual distribution of seats does not exactly match the
proportions identified (Greece, Portugal, Spain). Their aim is to arrive at
governmental majorities, secure optimum representation for the different
political tendencies and preserve the link between the electorate and elected
representatives, which is not always possible with the rigid application of
first-past-the-post/majority or proportional systems. While such electoral
systems were long regarded as “institutional anomalies symptomatic of an
inferior political civilisation”, they
are now being taken on board by more and more countries, particularly those
having undergone sweeping political changes.
20. The main combinations proposed are the
following:
¨
Application of plurality/majority system in some constituencies and
proportional representation in the others.
¨
Use of the plurality/majority systems for candidates or lists having
obtained the absolute (or qualified) majority and the distribution of the
remaining seats on the basis of proportional representation.
¨
In a given constituency, application of the plurality/majority system
for a predetermined number of seats and use of proportional representation for
the remainder.
¨
Application either of the plurality/majority system or the proportional
representation system depending on the size of the constituency.
¨
Use of proportional representation in order to offset the effects
brought about by plurality/majority voting. Amongst those systems, mention may
be made of “personalised representation” in Germany and the 1994 Electoral Act
in Italy.
1.3 Constituencies/electoral
districts
21.
A third variable characterising electoral systems is “the units within
which voting returns are translated into distributions of parliamentary seats”,[7] namely
the constituency or electoral district. The constituency may have a social
(“personal circles”)[8] or a
territorial (territorial constituencies) basis.
a. Personal circles
22. Whereas
elections organised on the basis of personal circles were frequently
encountered in the past, few countries still apply such a breakdown of
electors. Several criteria have been employed, such as social class, income,
occupation or membership of an ethnic group. Some commentators have also
proposed other distinctions: formation of opinion groups or age groups or the
possibility for electors themselves to determine the personal circle to which
they wish to belong.
b. Territorial constituencies
23.
Territorial constituencies constitute the principal way of dividing up
electors. They may be based on pre-existing administrative divisions (member
states of federal states, provinces, departments) or have been created ex
nihilo by the legislature for a given election (legislative and interregional
constituencies in France,for example).
24. Territorial constituencies may be
broken down into two types:
-
Single-member constituencies (one seat in contention per constituency)
are the most widespread today, in particular in the common-law countries. It is
possible to make use of such constituencies for elections based on the plurality/majority
principle. In contrast, they are inconceivable for elections based on pure
proportional representation.
-
Multi-member constituencies (several seats in contention per
constituency) are compatible both with plurality/majority systems and
proportional representation. As far as plurality/majority voting is concerned,
they are used in practice in Europe virtually only for local elections. Outside
continental Europe, multi-member constituencies are frequently used for the election of
national parliaments.
25. Determining the constituency
boundaries is an important step in the electoral process, since they can give
rise to marked “structural”[9]
inequalities in representation. Three factors may be behind such disparities:
demographic change, gerrymandering and “natural gerrymandering”.
-
The varying demographic trends experienced by different territories
make regular redrawing of constituency boundaries necessary. A failure to
redraw constituencies has an adverse effect on the most demographically dynamic
constituencies (urban areas) as compared with those constituencies with a low
population growth (rural areas).
-
Gerrymandering consists in skilfully redrawing constituencies on the
basis of previous electoral results with a view to boosting the representation
of the party in power. In such case, two specific examples may arise: either
the minority is voluntarily brought together in a few constituencies where it
has a strong majority, whilst the majority has only a slight edge over it in a
higher number of constituencies; or the redrawing is carried out in such a way
that the minority cannot obtain a majority in any constituency. In order to
avoid this, it is absolutely necessary that the redrawing of constituency
boundaries be carried out by an independent and politically neutral body.
-
Lastly, there is what Maurice Duverger terms “natural gerrymandering”.
Unlike gerrymandering proper, it does not result from a deliberately biased
redrawing of boundaries. It may be that one category of the population is
strongly concentrated in a small number of electoral constituencies, whilst
another has a very small majority in a large number of constituencies.
c. The existence of several levels of
constituencies/electoral districts
26. During an election, there may be
several levels of constituencies coexisting alongside one another. The method of casting votes or the guiding
principle may therefore vary depending on constituency size. For
elections to the US House of Representatives, use has sometimes been made of
single-member majority voting in small constituencies and a multi-member
majority system in bigger constituencies.
When the number of seats allocated to each State increased, these different
constituency levels were used to designate additional elected representatives
and no fresh divisions of single-member constituencies were made. A
number of constituencies may also be superimposed over each other. In hybrid systems, representatives are
sometimes elected at two levels of constituencies. In Mexico, for example, of the 500
representatives making up the Chamber of Deputies, 300 are elected by a
single-member, one-round first-past-the-post system and 200 by means of a
system of regional proportional representation using closed lists in
multi-member constituencies. In South Korea, half the deputies are
appointed by single-member constituencies, with the other half being elected by
proportional representation in a national constituency. The joint use of personal circles and
territorial constituencies is also possible.
It enables specific majorities to be represented, as in the case of the
election of the House of Representatives in New Zealand (61 territorial
constituencies, 6 constituencies reserved for the Maori minority and 53 seats
for list MPs).
1.4 Methods
of voting
27. When voting, each elector is called
on to make a choice and one of the functions of the electoral system adopted is
to govern how that choice is expressed.
On the basis of the choice which the elector is asked to make, a
distinction can be made between “categorical” voting methods which require the
voter to make an absolute choice and “ordinal” voting methods under which the
voter may qualify his or her choice. It
should be made clear immediately that this analysis variable has no direct impact
on the fragmentation of the party system and elected assemblies.[11]
a. “Categorical”
voting methods
28. “Categorical” voting methods oblige
the voter to indicate a preference for a party, a list or a candidate to the
exception of all others. They are used
both in plurality/majority systems and in elections run under proportional
representation. The following types
exist.
29. Single-member ballot: there
is where there is just one seat to be filled per constituency and each elector
has only one vote. It is therefore applicable only in plurality/majority or
hybrid systems.
30. The single non-transferable vote:[12]
irrespective of the number of seats to be filled, each elector has only one
vote. Candidates receiving the most
votes are elected in proportion to the number of seats to be filled. This system is regarded as being the
“fairest” of the non-proportional methods, since it enables a high degree of
proportionality to be achieved between the votes cast and the number of seats.
31. The closed list: Under this
system the elector has to choose a list in its entirety. It is generally used in proportional
representation systems; but it can also be used in plurality/majority and
hybrid systems. Under proportional representation, candidates on closed lists
are elected in the order in which they appear on the list.
b. “Ordinal”
voting methods
32. “Ordinal” voting methods enable
electors to express a more complex choice. For example, they can class parties
by order of preference or choose between the lists or candidates presented to
them. Whereas they are chiefly used in proportional representation systems,
“ordinal” voting methods are sometimes used in plurality/majority systems. In
all, there are nine types of “ordinal” voting systems:
33.
The preferential vote: this type of vote is possible only
for list-based proportional systems. It enables the elector to mark a
preference for one or more candidates appearing on the competing lists.
However, it is not a question of classifying candidates within lists but only a
way of expressing a preference for a specific candidate or candidates.[13] The
number of preferential votes which an elector may cast is limited.
34. “Latoisage”[14] or “negative vote”:[15] this
is the converse of the preferential vote in that the elector can strike one or
more candidates from the list, so as to show which candidate he or she does not
wish to vote for. The “negative vote” differs from panachage (or
vote-splitting) as the elector does not indicate the name(s) of other
candidate(s) to replace the candidate(s) struck out.
35. Cumulative vote: electors may
give two or more votes to a given candidate. In practice therefore, each
elector has a stock of votes equal to the number of seats to be filled which he
or she can use to mark his or her preference. The leading candidates are
elected in proportion to the number of seats to be filled. When used in the plurality/majority
systems,[16] the
cumulative vote tends to accentuate artificially the representation of
significant minorities. The drawback with this system is that it may result in
a defeat for the majority if there is an excessively large number of votes for
one of its candidates. Lastly, it should be noted that the cumulative vote may
be combined with the limited vote.
36. Graduated or rank-ordered vote:
this allows electors to classify candidates in order of preference. To this
end, electors have a number of votes which they distribute between the various
candidates for whom they wish to vote. Each position corresponds to a number of
votes determined by the legislature. It is in this respect that this method
differs from the cumulative vote. It should be added that the number of
candidates for whom an elector may express a preference may be limited.
37. The aforementioned ordinal voting methods do not cater for the
possibility of voting for candidates from several lists. That is possible,
however, with the following variants, which are designed to enable voters to
vote simultaneously for candidates from different lists.
38. Panachage or vote-splitting:
the elector can modify a party list and include on it candidates from other
lists. Panachage is rarely authorised in proportional representation systems.
39. The blank list: the elector
draws up his or her list of candidates that he or she would like to see
elected, but the candidates’ names must already appear on other lists.
40. The totally open list:
electors draw up their lists themselves. They can include persons of their
choice, whether or not they are standing for election. The persons obtaining
the most votes are declared elected. Multi-member plurality/majority ballots
belong in this category. The candidates stand independently and electors
indicate the names of those they would like to see elected.
41. The limited vote: the elector
has fewer votes than there are seats to be filled and he or she is not allowed
to cast more than one vote for a given candidate. The candidates with the most
votes are elected in proportion to the number of seats to be filled.
42. The use of the limited vote in a
list-based proportional representation system is extremely rare. There is only
recorded occurrence: Finlandbetween 1906 and 1935. In contrast it is used somewhat more frequently in plurality/majority
systems. For example, it is used to elect the Spanish Senate.
43. This type of ballot has the
advantage of securing a measure of representation for minorities in plurality/majority
systems. The reason for this is that the party likely to obtain the majority
has no interest in putting up a number of candidates equal to the number of
seats to be filled because, if it did so, it would no longer be sure of
obtaining a majority as a result of the dispersion of votes. Strategically,
therefore, it is compelled to present a number of candidates equal to the
number of votes available to each elector in order to optimise its chances of
obtaining a majority. This therefore enables minority parties to be
represented.
44. Contingency vote systems: the
elector has only one vote under this voting system. On his ballot paper, he is
asked to classify all the candidates standing in his constituency in order of
preference. In the count, his vote will initially go to his or her
first-preference candidate. Subsequently, that vote may be transferred to the
second, third, etc preferred candidate. The aim of this voting method is to
enable electors to qualify their choice, while avoiding the wastage of votes
granted to candidates who would not be elected or who already have a sufficient
number of votes to be declared elected. In practice, this type of voting method
has invariably been used where there are individual candidatures. However, it
could theoretically be applied to list-based ballots provided that the lists
are closed. It should also be noted that contingency vote systems are only used
for single-round ballots.
45. In general, a distinction can be
made between two types of contingency votes which depend on the guiding principle
used for the election. Where the majority system is used, the contingency vote
is known as the alternative vote. In the case of proportional representation,
it is known as the single transferable vote. The difference between the two systems
applies solely to the distribution of seats and has no effect on the manner in
which the elector may vote.[17]
1.5 Number
of rounds of voting
46. A further variable affecting the
make-up of electoral systems is the number of rounds of voting. Most ballots
consist of one or two rounds but there are also systems with n rounds.
a. Single-round
ballots: these can be used not only in proportional representation systems
but also in first-past-the-post systems. This type of ballot generally leads to
a bipolarisation of political life (Duverger’s Law) or at least to the
conclusion of pre-electoral agreements between large and small parties.
Moreover, opting for a single round ballot magnifies, virtually systematically,
the dominance of the majority party in terms of votes and under-representation
of other parties in the assemblies. This trend is even more marked where there
is a multi-member ballot.
Moreover, where there is a single-member
first-past-the-post system in which two parties are virtually of equal strength
(two-party system), the distribution of seats approximates to a cubic
relationship (Law of Cubic Proportions):[18]
003-e_files/image002.gif)
in which S1 = number of seats obtained by party 1,
S2 = number of seats obtained by party 2
V1 = number of votes obtained by party 1, and V2 =
number of votes obtained by party 2.
It should be noted however that there are numerous
exceptions to this rule. Two conditions have to be fulfilled in order for this
law to apply: all the constituencies must be substantially equal and the
difference between the votes cast for the majority and the minority has to be
virtually identical in all constituencies. As a result, the Law of Cubic
Proportions can provide only an imperfect indication as to the possible make-up
of assemblies.
b. Two-round
ballots: a second round of voting is habitually used where the leading
candidate or list failed to obtain an absolute majority[19] in the first round and this type of majority is
necessary in order to be declared elected in the first round. It is therefore
closely linked to the majority system.
47. In order to guarantee a measure of
representativeness of elected persons, a quorum is often necessary in order to
validate an election in the first or, indeed, the second round. The quorum can
take two forms: a minimum score in terms of the registered electorate for the
leading candidate or the attainment of a certain participation rate. In France, for the election of members
of the National Assembly, for example, a candidate cannot be declared elected
in the first round unless he or she obtained the votes of at least 25% of the
electors on the register. In Lithuania, a 40% participation rate
is necessary in order to validate a parliamentary election. On top of this
there are frequently conditions designed to restrict the number of candidates
standing in the second round. Accordingly, thresholds may be introduced so as
to authorise only candidates or lists which have obtained a minimum score to
participate. In the case of parliamentary and regional elections in France, the respective thresholds
are 12.5% of electors on the register and 10% of the votes cast. Participation
in the second round may also be restricted to the two candidates or the two
lists which came out on top in the first.
This restriction to two candidates for the second round is still used in
France, Portugal, Austria, Poland, Brazil and Peru for the election of the
President of the Republic and in Ukraine for the election of members
of parliament. It was formerly used for the election of the Reichstag in
Imperial Germany and in Austria and Italy before 1918.
48. In the second round of voting, only
a relative majority is needed for a given candidate or list to be declared
elected.
49. In single-member ballots, opting for
a two-round election has the main effect of encouraging coalitions between
parties: “The influence of the two-round, single-member ballot on the
strategies of political parties may depend as much on alliances as on the
number of votes obtained”.[20] The logic behind this type of ballot is that
in the first round, each party should quantify the strength of the other
parties. In the second round, these parties support those candidates best
placed to win and promote their ideas. More
often than not, this support is expressed by withdrawal. In multi-member
systems, support may take tangible form in the merging of lists (if authorised
and subject to the reaching of certain applicable thresholds). It should
be noted that where alliances are possible only in a single camp – owing in
particular to the influence of extreme parties or parties excluded from
alliances with government parties –, the inequalities in terms of
representation may prove to be particularly marked.
50. Ballots consisting of n
rounds: little used nowadays, this type of ballot can be contemplated only
where there is a requirement for an absolute majority. The principle underlying
this system is simple: since an
absolute majority is required in each round, as many rounds of voting
are held as are necessary in order to be able to declare one of the candidates
or one of the lists elected; on the
other hand, in a two-round system, an absolute majority is required only in the
first round. Sometimes, the candidates who do the least well are
excluded between the various rounds but this does not happen systematically.
Votes consisting of n rounds appear to be largely incompatible with
elections by secret ballot. For practical reasons, it is hard to contemplate
calling all electors to the ballot box more than two or three times.
Consequently, this system is used only for votes carried out by assemblies of
limited size.
2. COUNTING VOTES AND
DISTRIBUTING THE SEATS
2.1 Distributing the seats among constituencies
51. Before considering the allocation of
seats among the various political formations, it is worth dwelling briefly on
the distribution of seats among constituencies. The number of seats must be
approximately proportional to the population of the constituency, either the
number of nationals – including minors –, the number of electors, or the number
of actual voters. In the last case, the seats can only be allocated on the
evening of the election. There are, however, examples of “non-proportional”
distribution. Examples are the US Senate and the Swiss Council of States, where
all the states or cantons (except the former half-cantons) have two
representatives, irrespective of their population size. In these cases,
preference is given to equality between states or cantons over equality between
citizens.
52. There are often other rules in
addition to this requirement for proportionality. For example, there are five
rules governing the election of the House of Representatives in the United
States: 1) the number of representatives of a given state cannot decrease where
the total number of representatives increases (Alabama paradox), 2) only the
quota determines the number of seats to be allocated (rounded up or down to the
nearest whole number), 3) all the States are subject to the same method of
allocation, 4) the method used must not excessively favour small States over large
States, and vice versa, 5) each State must receive at least one
representative. Similarly when the 1958 electoral map was drawn up in France, the principle of
proportionality was supplemented by the obligation to give each département
at least two deputies.
53. It is for this reason that a number
of methods have been drawn up in order to take account of all of these rules.
The most frequently used ones are as follows:
-
The Jefferson method:it is used today for the purposes of distributing seats under the name of the
“highest average method” although it was originally invented for the
purposes of allocating seats among constituencies.
-
The Webster method: corresponds to the
Saint-Laguë method.
-
The Hamiltonmethod: this is method of the largest remainders. The snag with this mode
of distribution is that it does not preclude the effects of the Alabamaparadox.
-
The Huntingtonmethod: the votes cast for each list are divided by the following series of
numbers: 003-e_files/image004.gif)
, 003-e_files/image006.gif)
, 003-e_files/image008.gif)
,…, 003-e_files/image010.gif)
. The seats are allocated in the same way as in the divisor
method.
This system yields arbitrarily at least one seat per constituency, as the first
divisor is 0. This system has been adopted in the United States for the House of Representatives. It
may be used for the purposes of allocating seats between parties, provided that
the first division is not carried out. It is not possible to grant one seat to
all the competing lists without exception.
2.2 Distributing the seats between political formations
54. Counting votes and distributing
seats amongst the various candidates (or amongst political tendencies) depends
essentially on the guiding principle adopted. We shall therefore distinguish
between the plurality/majority systems and proportional representation.[23]
a. Distribution of seats in plurality
and majority systems
55. As mentioned earlier, where the
guiding principle is majoritarian, the candidate or list which obtains the
requisite majority[24] in the
conclusive round of voting obtains all the seats in a constituency. In
single-round ballots, a relative majority is generally required. In systems
where an absolute or qualified majority is required, there are generally two
rounds of voting.
56. However, there is a single-round
voting method which enables representatives to be elected by an absolute
majority. This is the alternative vote, a system whose method for counting
votes warrants our full attention, since it differs significantly from that of
the other majority (or plurality) systems. In order to explain this method of
counting votes, we shall draw a distinction between its application in
single-member constituencies (only one seat to be filled) and multi-member
constituencies (several seats to be filled).
-
The
single-member alternative vote: the
number of first preferences obtained by each of the candidates is counted.
If one of them obtains an absolute majority, he or she is elected. If this is
not the case, the candidate who obtained the fewest first preferences is
eliminated. The second preferences recorded on the ballot papers of the
candidate who has been eliminated are then considered. Those second preferences
then become first preferences. The total number of first preferences is
recalculated to see if one of the candidates obtains an absolute majority. If
that is not the case, the candidate with the fewest votes is eliminated and his
or her votes are carried forward as described above. The operation is repeated
as many times as is necessary for one candidate to obtain the absolute majority
of the votes cast.
The
advantage of this manner of counting votes is that it avoids any candidate
being elected by chance owing to the dispersion of his or her competitors. It
also affords a higher degree of representativeness of the person elected in his
constituency, since votes cast for small candidates will be taken into account
by the carry-over system. This method of voting is used in Australia for elections to the House of Representatives
and Assemblies of States, with the exception of Tasmania.
-
Multi-member
alternative vote: the first seat is
allocated using the same method as in single-member constituencies. For the
purposes of allocating the second seat, the second preferences on the ballot
papers of the candidate elected become first preferences. The first preferences
are then recalculated. The second elected representative is then appointed
using the same procedure as in the case of the single-member alternative vote.
All these operations are repeated until all the seats are filled.
57. The votes of electors who gave their first
preference to an elected candidate are taken into consideration several times.
This has the effect of enabling the majority party to obtain almost all the
seats. Thus in 1925 the Labour Party obtained 45% of the votes cast in the
election for the Australian Senate
yet obtained no seats.
58. With
alternative votes, a vacancy in a seat usually results in the holding of a
by-election.
b. Distribution
of seats in proportional representation systems
59. The allocation of seats in
proportional representation systems generally requires two operations to be
carried out. The first allocation is effected using an electoral quota.
However, use of such a quota does not enable all the seats to be allocated and
all the votes cast to be dealt with. It is therefore necessary to have a second
stage designed to convert the remaining votes into seats.
FIRST STAGE of
the distribution: the electoral quota or quotient
60. An electoral quota or electoral
number is a divisor which helps determine the number of seats to be allocated
to each of the lists. During this first stage of the distribution, a list
therefore obtains a number of seats (S) equal to the quota – rounded up to the
next whole number – of the total number of votes obtained by the list (V)
divided by the number of votes needed in order to obtain a seat (the electoral
quota, EQ).
003-e_files/image012.gif)
61. There are two major classes of
electoral quotas: “fixed” electoral quotas and
“variable” electoral quotas.
1. The “fixed” electoral quota or uniform number:
this is a number of votes predetermined by the legislature and identical for
all constituencies. Use of this quota means that the number of seats in the
assembly will not be determined until election night. The number of seats will,
moreover, depend on the participation rate. Moreover, adoption of a “fixed”
electoral quota tends to preclude the representation of a substantial number of
votes, particularly those cast for small parties. Consequently, only the choice
of a relatively low electoral quota, coupled with its application in large
constituencies, is capable of curbing this tendency. This type of electoral
quota has only been used in the Weimar Republic in Germany.
2. The “variable” electoral quota is determined on
election night. There are various forms:
-
The simple or Hare quota: this quota is obtained by dividing, in each
constituency, the total number of votes cast (V) by the number of seats to be filled (S).
Hare quota 003-e_files/image014.gif)
In practice, the Hare quota corresponds to the
maximum number of votes which a candidate requires in order to be elected. The
drawback with this system therefore is that it means that only a small number
of seats can be allocated during the first distribution. It is, moreover, for
this reason that other quotas have been formulated so as to allow the maximum
possible number of seats to be allocated during the first stage of distribution.
-
The Hagenbach-Bischoff quota: this quota is obtained
by dividing, in each constituency, the total number of votes cast (V) by the number of seats (S) plus one.
Hagenbach-Bischoff quota 003-e_files/image016.gif)
-
The Droop quota: this quota is calculated in the
same way as the Hagenbach-Bischoff quota, but one is added.
Droop Quota 003-e_files/image018.gif)
-
The Imperiali quota: this quota is calculated in
the same way as the Hagenbach-Bischoff quota but two is added to the divisor
instead of one.
Imperiali Quota 003-e_files/image020.gif)
-
The double quota: this is, in a manner of
speaking, a combination of the “fixed” electoral quota and the Hare quota. In
the first instance, an initial electoral quota is fixed corresponding to the
minimum number of votes needed to participate in the distribution of seats. A
second electoral quota is then calculated which takes account only of the votes
obtained by the lists which were successful in the first stage.
The double quota must not
be confused with a qualifying threshold for participating in the distribution
of seats. First, the quota is expressed as a number of votes. Second, it is
identical regardless of the constituency. The double quota is used chiefly in Latin America (Bolivia,
Costa Rica, Ecuador).
Second staGE: distribution of
the remaining seats and votes cast
62. It is rarely possible to distribute
all the seats using quota methods. At the end of the first distribution, there
very often remain unrepresented votes and seats to be allocated. Accordingly, a
second distribution has to be carried out using one of the following methods:
-
The largest remainder method: using this system,
the list with the highest number of unrepresented votes at the end of the first
distribution obtains one seat. The operation is repeated until all the seats
unfilled at the end of the first distribution have been allocated.
This method is by far one
of the most favourable for small lists, thereby tending to encourage a
proliferation of small parties. Moreover, the lower the number of available
seats, the more this system favours small political groupings. The largest
remainder method also suffers from the disadvantage that it does not take
account of the relative strengths of the parties, that is to say the number of
seats obtained during the first distribution. On top of this, a number of
paradoxes, such as the “Alabama paradox”,
are associated with this method.
-
The strongest lists method: this system provides
for those seats which remained unfilled at the end of the first distribution to
be allocated to the lists or list obtaining the greatest number of votes cast.
This system markedly favours the large parties. This method, which was only
infrequently used in the past, is not used any more today.
-
The highest average method: this is the system
in most widespread use internationally. Under this method, the number of votes
cast for each of the lists (V1) is
divided by the number of seats which the list in question obtained during the
first distribution (S1), to which a
fictitious seat is added. The list which has the highest average per seat is
then allocated the seat at stake. This operation is repeated as many times as
necessary until all the vacant seats have been allocated. During the first distribution
of seats, either the simple quota or the Hagenbach-Bischoff quota may be used.
Average votes per seat 003-e_files/image022.gif)
This system tends to put large parties at an
advantage and to exclude small parties from the distribution of seats. It also
guarantees that coalitions will obtain at least as many seats as they would
have obtained if the parties belonging to them had stood as single parties.
Lastly, mention should be made of a variant of the
highest average method, the Balinski-Young method. First, an initial
distribution of seats is carried out using the Hare quota. Next the highest
average method is used as described above with one variant. The division is
carried out once only and the seats go to the lists with the highest average. Consequently,
no party obtains more than one of the remaining seats; this has the effect of
reducing the over-representation of large parties.
63. Lastly, it should be noted that the
distribution of seats remaining vacant may be carried out at the level of the
basic constituency, at the level of groups of constituencies or even at the
national level. The choice of a second level of constituencies for the purposes
of allocating the remaining seats makes for greater proportionality between
seats and votes cast, since as D.W. Rae[29] has observed, that factor
depends on constituency size.
allocating the seats in one
OPERATION: the divisor methods
64. Some allocation methods have,
however, been introduced so as to enable all the seats to be allocated in a
single operation. These methods are all based on the following general
principle: the number of votes cast in favour of each of the lists is divided
by sequential numbers. The seats are allocated to the lists which obtain the
highest quotas during this operation. We have set out below the main divisor
methods which have been applied in practice.
-
The D’Hondt method: the votes obtained by each
list are divided by a sequence of whole numbers: 1, 2, 3, 4, 5… The seats are
allocated to the lists obtaining the highest quotas. This method tends to
favour the majority party.
-
The Sainte-Laguë method: the votes obtained
by each list are divided by a sequence of odd numbers: 1, 3, 5, 7… The seats
are distributed among the lists which obtain the highest averages. The
Sainte-Laguë method is distinctly more favourable to small parties than the
D’Hondt method.
-
The modified Sainte-Laguë method: it
differs from the Saint-Laguë method only in that the first divisor is replaced
by 1.4. This method is more favourable to small parties than the D’Hondt
method, but does not overly advantage them. It also affords fairer
representation for medium-sized parties. It is used today in Sweden,in Norway and
also in Denmarkfor small constituencies.
-
The so-called “Danish” method: the number of votes
obtained by each list is divided by the following numbers: 1, 4, 7, 10…This
system is extremely favourable for small parties. In Denmark,this method is used to distribute, among small constituencies, seats attributed
to a party at the level of a group of constituencies.
c. Some specific methods for counting votes and distributing
seats
65. In the following section, we propose
to examining two specific methods for accounting for votes (the single
transferable vote and the “apparentement” – or alliance – system) before
discussing the distribution of seats in “personalised proportional
representation systems”.
The single transferable
vote
66. The contingency vote applied in the
context of a proportional representation ballot is known as the single
transferable vote. It is used principally in Ireland and in Australia for elections to the upper
House. As in the case of the alternative vote, vote counting is a relatively
complex operation. This is how it works. First, a count is made of the number
of first preferences obtained by the competing candidates. Any candidate
obtaining the electoral quota[30] is
declared elected. If no candidate satisfies this condition, the one who
obtained the fewest first preferences is eliminated. The second preferences on
the ballot papers are then taken into consideration. A new total is worked out
to see whether one or more candidates has obtained the electoral quota. Once a
candidate has been elected, his or her name is deleted from all the ballot
papers.
67. The question which then arises is
how to carry forward votes which have not been used directly to elect a
candidate. There are two possible solutions. The first is to transfer the votes
appearing on the ballot papers which were sorted after the electoral quota was
reached. The final outcome of the election will therefore depend on the order
in which the ballot papers were sorted. The second solution, adopted in Ireland, eliminates this random
factor. All the preferences marked on the ballot papers on which the elected
candidate received first preference are counted. These are then weighted by a
quota equal to the surplus votes of the elected candidate divided by the number
of transferable votes. Unlike the alternative vote in multi-member
constituencies, the votes are therefore taken into account only once. The
preferences weighted as described above are added to the first preferences of
the respective candidates. If one of the candidates attains the electoral
quota, he or she is elected. If not, the least well-placed candidate is
eliminated and the preferences indicated on the ballot papers where that
candidate had a first preference vote go up one place in the hierarchy. A
candidate is elected when he or she attains the electoral quota. When at least
one of the competing candidates is declared elected, the operation is repeated
from the beginning by striking out the name of the candidate elected and then
taking account of all the preferences on the ballot papers on which the elected
candidate had a first preference vote and so on.
68. Should a seat fall vacant, the procedure for designating a new elected
representative varies according to the country or assembly concerned. In the
Irish House of Representatives for example, a by-election is held, as in the
Indian Council of States and the Maltese House of Representatives. In the
Australian Senate, on the other hand, the incumbent is replaced by a member of
the same party elected by the Chambers of the State or territory concerned. For
vacant seats in the Irish Senate, the Prime Minister may appoint a replacement
in some cases.
The “apparentement” or
alliance system
69. The “apparentement” or “alliance”
system can be summarised as follows: “Several lists of different parties which
are entirely separate, each having its own programmes may declare by mutual
agreement that they intend to link up, in other words this means that when the
seats are allocated, the votes which they have obtained separately must be
added together”.[31] It is
used today in Finland, the Netherlands, Spain and above all Switzerland.
70. “Apparentement” declarations generally
have to be made before the election, but it is possible to conceivable to have
a system in which such declarations are not disclosed until after the election.
In systems involving two rounds of voting, “apparentements” may be authorised
only for the second round so as to allow electors to express their choices.
“Apparentement” is often subject to conditions. For example, in France, under the 1951 Electoral
Act for the election of the National Assembly “apparentement” was permitted
only between lists of national parties or groupings, that is to say parties or
groupings putting forward candidates in at least thirty départements. It
should be added that “apparentements” are sometimes operative only for the
allocation of remainders in the basic constituency or in groups of
constituencies.
71. In comparison with union lists, the
“apparentement” system enables a more proportional distribution to be achieved
between the various linked parties. It is more favourable than the higher average
system to small parties. This system benefits only parties which are able to
form alliances.
“Personalised
proportional representation”
72. The idea behind “personalised
proportional representation” is to offset the effects of first-past-the-post (or
majority) voting during the distribution of seats. Its best-known application
is that of the election of the Bundestag in Germany, which we shall use as a
guide in explaining this system. It has however been used in Estonia since 1992 and in New Zealand since the 1993 referendum.
73.
Under this system, each elector has two votes. The first vote is designed to
elect half the representatives of the Bundestag in single-member
constituencies. The second enables the total number of mandates attributed to
the various parties to be determined on a proportional basis at national level.
Those mandates are then distributed between the various Länder in proportion to
the votes obtained by the lists in those Länder. The number of seats attributed
as a result of the first votes is subtracted from the total number of seats
allocated to each party in the Länder. If a party obtains more seats with the
first votes than it obtains in the proportional vote, the seats gained in this
way are kept. As can be seen, such a system may resemble a preferential vote
within a party. The advantage of this method of distribution is that it enables
members to be designated directly while ensuring proportional representation in
the assemblies. However, there are risks of manipulation as Pierre
Martin points out: “in this two-vote
system, there is nothing to prevent a party from not putting forward official
candidates in the single member ballot and allowing them to stand as
independents (…) In that case, it could, in practice, obtain a substantially
higher number of direct seats while benefiting to the maximum from the
offsetting mechanism”.
d. Thresholds and bonuses
74. In
order to ensure that there are stable majorities in assemblies elected by
proportional representation, the legislature very often has recourse to
thresholds entitling access to the distribution of seats and to bonuses for the
most successful list.
75.
Thresholds are “fixed or variable limits, established using the election
result, which determine whether a given list or a given candidate takes part in
the distribution of seats”.[33] They
are in some way the equivalent of the thresholds used in the first round
election in the majority system. This restriction on access to distribution
generally applies to the award of seats in basic constituencies, but it may
also be applied solely to groups of constituencies or to the distribution of
remainders. For example, in Austria, only those lists obtaining the simple quota in the
basic constituencies are admitted to the distribution of seats at the level of
groups of constituencies. In Greece, there are three rounds
of distribution of seats. During the first two rounds the seats are distributed
between all parties according to the votes they have acquired, excluding those
parties, which have not reached the threshold of the 3% of the votes. In the
third round the remaining seats are usually allocated to the party having
acquired the majority of the votes. In any case it has to be made certain that
the representation of all parties is equal to at least 70% of their
proportional share of votes.
76. The
thresholds, which are generally expressed as a percentage of registered
electors or of voters, are a matter for the discretion of the legislature.
However, the role played by the thresholds differs depending on how high they
have been set and on the party system existing in each country. The choice of a
low threshold eliminates only very small parties, which makes it more difficult
to build stable majorities in assemblies. Where there is strong fragmentation
of the party system, a high threshold results in the exclusion from
representation of a substantial proportion of votes.
Examples of thresholds adopted for the
election of lower houses
-
Netherlands:obtaining 0.67% of the votes cast at national level.
-
Israel:obtaining 1% of the votes cast at national level.
-
Bulgaria:obtaining 4% of the votes cast at national level.
-
Liechtenstein:obtaining 8% of the votes cast at national level.
-
Denmark:obtaining 2% of the votes cast at national level or obtaining a specific number
of votes in two of the three geographical areas of the country.
-
Germany:obtaining 5% of the votes cast at national level or obtaining three direct
mandates.
-
Sweden:obtaining 4% of the votes cast at national level or 12% of the votes cast in
the basic constituency in which the seat is awarded.
77. Bonuses are mandates granted to the most
successful list before the distribution of seats strictly speaking is carried
out. They are principally used for local elections. In the French regional
elections, a bonus equivalent to one quarter of the seats is given to the list
which ends up in the lead in the conclusive round. For elections to the
Corsican Assembly, three bonus seats are granted. Likewise, half the seats are
granted directly to the leading list in the case of municipal elections in
municipalities of more than 3,500 inhabitants. Bonuses can sometimes exist in
other forms. For instance, the 1953 Italian Act provided that linked lists
obtaining more than 50% of the votes were to receive 64.5% of the seats.
Consequently, these bonuses enable government majorities to be conferred on
assemblies and hence avoid the necessity of bargaining with extremely minority
parties.
2.3 Allocation of seats within lists
78. Once
the seats have been distributed among the political parties, the question
arises as to how to distribute them within the lists in the case of elections
under a system of proportional representation. That distribution can be done in
many ways, but we will confine ourselves here essentially to presenting
procedures which have already been used.
a. Modes of expression and distribution
79. In
the simplest case, that of closed lists, candidates are elected in the order in
which they appear on the lists. In this case, the political parties have very
substantial power, since they determine the order in which candidates appear.
In the case of preferential, cumulated and rank-ordered voting on the other
side, the electors determine the position of each candidate on the list which
they compose.
80.
In practice, preferential votes do
not always influence the distribution of mandates within lists. Although
the electors can indicate their preferences, a number of conditions have to be
fulfilled almost systematically before they can be taken into account. In Belgium, for example, in the case
of both Chambers, the votes of electors who have not cast a preferential vote
are automatically considered to be preferential votes in favour of the
candidates at the top of the lists. Once a sufficient number of votes has been
obtained in order to declare the candidate at the head of the list elected,
non-preferential votes are counted as being votes in favour of the candidate in
second place and so on. For the election of the Austrian National Council,
there is a threshold for taking preferential votes into account. The threshold
is fixed at a level equivalent to the number of voters in the basic constituency
divided by the number of seats to be filled. This generally results in a de
facto closed list situation.
81.
Lastly, the single transferable vote, panachage and open or blank lists, when
applied to the letter, inherently incorporate a system for allocating seats
within lists.
b. Subdivided constituencies
82. Some
electoral systems provide for subdivided constituencies. These are subdivisions
of the constituencies used as the basis for the election and in which the
candidates stand. When the seats are distributed, the mandates are attributed
between the various lists at the level of the basic constituencies.[34] Once
this operation has been carried out, the seats obtained by each list are then
distributed among the various subdivided constituencies, generally in
proportion to the votes obtained in each of those constituencies. Within those
constituencies themselves, candidates are then elected in the order in which
they appear on the lists.
83. The
total number of representatives in subdivided constituencies is not known in
advance. It depends on the level of participation. The higher the turnout in a
subdivided constituency, the higher its number of seats will be. In addition,
the most populated subdivided constituencies are placed at an advantage. They
generally obtain more seats than less populated constituencies simply as a
result of the pro rata distribution of the votes obtained.
84. Today, the Netherlands and Denmark apply this distribution
system for elections to the lower chamber. In Germany, the Länder act as
subdivided constituencies for the election of the Bundestag. Recently,
subdivided constituencies have been introduced for French regional elections
(departmental sections).
2.4 Examples of past or present electoral systems
ESTONIA (Electoral Act of 7 June 1994, last amended:
December 1994)
Chamber (monocameral system): Riigikogu
Number of seats: 101
seats
Constituencies: 11
multi-member constituencies
(8 to 11 deputies per
constituency)
Guiding Principle: Proportional and
proportional by compensation
Distribution of seats: Triple proportional
allocation of seats using the Hare quota in each of the 11 multi-member
constituencies. The seats remaining to be allocated (“compensatory” seats) are
distributed at national level between the parties and coalitions which obtained
over 5% of the votes cast.
France (1951
Electoral Act, in force for the 1951 and 1956 elections)
Lower Chamber: Assemblée nationale
Deputies
of metropolitan France
Constituencies: Each
département constitutes a constituency with the exception of the départements
of Bouches-du-Rhône, Pas-de-Calais, Rhône, Seine, Seine-et-Oise, Seine
Inférieure and Gironde, which are divided into several constituencies
Guiding
Principle: Hybrid: use
of the majority system for candidates or lists that have attained a majority
and distribution of the remaining seats on the basis of proportional representation
Manner of
voting: Single-round
list-based ballot with “apparentement” of lists, panachage and preferential
votes
Counting of votes and distribution of seats:
-
“apparentements” authorised at constituency level between lists of
national parties and groupings or between lists composed solely of candidates
belonging to national parties or groupings provided this “apparentement” is
accepted by all candidates subject thereto.
-
The list or the group of linked lists which obtains the absolute
majority obtains all the seats. Seats are distributed between linked lists
according to the rule of the highest average, account being taken of
preferential votes and panachage. If no list satisfies these conditions, the
seats are distributed on the basis of proportional representation in accordance
with the rule of the highest average (largest remainders in Seine and Seine-et-Oise).
Distribution within lists is carried out on the basis of the highest average.
No seat is awarded to lists obtaining less than 5% of the votes cast.
FRANCE (1985 Electoral
Act)
Election: Municipal
elections
Constituencies: Municipalities with
more than 3,500 inhabitants (except Paris, Lyon, Marseille)
Guiding principle: Hybrid,
the ultimate aim being to archieve a majority
Manner of
voting: Two-round
ballot. Closed lists. Only lists which obtained 10% of the votes cast may go
through to the second round. Lists obtaining 5% may merge and can go through to
the second round.
Distribution
of seats: Bonus equivalent
to half the seats for the leading list in the conclusive round. The remaining
seats are distributed on the basis of proportional representation between lists
which obtained 5% of the votes cast including the leading list.
Germany (1993 Electoral Act)
Lower Chamber: Bundestag
Number of seats: 656
members (500 as from 2002)
Constituency: 328 constituencies +
1 subdivided national constituency (the Länder)
Manner of voting: Each
elector has two votes
-
one vote for the single-member one-round ballot
-
one vote for a list under a proportional representation system
Distribution
of seats - 328 members
designated by the single-member one-round ballot by the first vote
- 328 seats
allocated by proportional representation at national level. In order to
participate in the distribution of seats, a list must obtain 5% of the votes
cast or have obtained 3 seats in the single-member ballot. This threshold does
not apply to parties representing national minorities. The seats are then
distributed between the Länder in proportion to the votes obtained in the
Länder themselves. The number of seats obtained on the basis of the first vote
is subtracted from the total number of those seats. If a party obtains more
seats in the single-member ballot than it is allowed according to the proportional
representation vote, it keeps those seats.
IRELAND (Electoral Act of 5
November 1992, last amendment: March 1998)
Lower Chamber: Dáil
Eireann - House of Representatives
Number of seats: 166
seats
Constituencies: 42
multi-member constituencies (3 to 5 seats).
Guiding principle: Proportional
representation
Manner of voting: Single
transferable vote
Distribution of seats: Included in the
single transferable vote
ITALY (Electoral
Act of 6 February 1948, last amendment: August 1993)
Upper Chamber: Senato
della Repubblica
Number of seats: 315
seats
Constituencies: -
232 single-member constituencies
- 20
multi-member constituencies
Guiding
principle: Hybrid
combining a single-member ballot and proportional representation by a system of
offsetting
Method of
voting: - plurality
vote in the 232 single-member constituencies
- distribution of the 83
remaining seats by proportional representation on the basis of regional results
and in accordance with the d’Hondt method. These are “compensatory” seats and
accordingly, before a distribution on the basis of proportional representation
is carried out, the votes given to the candidates elected are subtracted from
the total votes for the list to which they belong.
ITALY (Electoral
Act of 6 February 1948, last amendment: 4 August 1993)
Lower Chamber: Camera
dei deputati
Number of seats: 630
seats
Constituencies 475
single-member constituencies
155 seats elected by
proportional representation in 26 constituencies
Guiding
principle: Hybrid
combining a single-member ballot and proportional representation by a system of
offsetting
Method of voting: Each
elector has two votes
Distribution
of seats: -
First-past-the-post in the single-member constituencies, with candidates
belonging to lists
- The
lists have to pay a “levy” (scorporo) for each candidate in the quota elected
under the first-past-the-post system. Distribution takes place at national
level on the basis of the total obtained in the constituency. Initially, only
lists which which obtained 4% of the votes cast take part in this stage. The
155 seats are allocated among the lists by means of the method of whole quotas
and largest remainders. Distribution
at the level of the lists must fulfil two conditions. Firstly, each list is
allocated seats in the constituencies where, proportionally, it gained the
highest number of votes. Then, the total number of deputies elected in a
constituency (the sum for all lists) may not exceed the number of seats
allocated to that constituency for the proportional part of the vote.
JAPAN (Electoral Act of 1
January 1900, last amendment: May 1998)
Upper Chamber: Sangiin
Constituencies: -
47 multi-member constituencies, metropolitan or prefectoral
- One
national constituency for the remainder of the seats
Guiding
principle: Hybrid:
first-past-the-post for a predetermined number of seats and proportional
representation for the remainder.
Method
of voting: - 146
representatives elected by first-past-the post in geographical constituencies.
Candidates obtaining a number of votes equal to or greater than one-sixth of
the electoral quota (total number of votes divided by the number of seats to be
filled in the constituency) are declared elected in an order reflecting the
amount of valid votes received.
- 96
representatives elected at national level by a proportional list system in
accordance with the d’Hondt method.
LITHUANIA (Electoral Act of 9 July
1992, last amendment: April 2003)
Chamber (monocameral
system): Seimas
Number of seats: 141 seats
Constituencies: -
71 single-member constituencies
- 1 multi-member constituency
(70 seats)
Guiding Principle: Hybrid: parallel application
of first-past-the-post (in single-member constituencies) and proportional (in
one multi-member constituency) systems.
Distribution
of seats: - first-past-the-post
system in 71 single-member constituencies. A candidate shall be considered
elected when that candidate receives the majority of the votes cast. A turn-out
of 40% is required for the election to be valid.
- proportional voting for
one nation-wide multi-member constituency. The list of candidates of the party
may receive mandates only if not less than 5% of the voters participating in
the elections voted for it. The threshold for joint lists of candidates is 7%. If less than 60% of the voters have voted for the
lists that have received more than 5% of the votes cast, the lists that have
not taken part in the distribution up till then may acquire the right to take
part in the distribution of mandates. The lists that have not taken part in the
distribution are added one at a time, starting with the one that received the
highest percentage of votes cast until the lists taking part in the
distribution of mandates make up a total of at least 60% of the votes cast. The
distribution of mandates is done on the basis of Hare’s quota. The mandates
still to be distributed after the first distribution are distributed according
to the highest remainders. A 25% turn-out is required for the election to be
valid.
LUXEMBOURG (Electoral
Act of 31 July 1924, last amendment: February 2003)
Chamber (monocameral system): Chambre
des Députés
Number of
seats: 60
seats
Constituencies:
4 electoral constituencies: South (23 deputies),
Centre (21 deputies), North (9 deputies) and East (7 deputies)
Guiding principle: Proportional
representation
Method of voting: Preferential
vote or panachage
Distribution of seats: Hagenbach-Bishoff method
Remaining
seats distributed in accordance with the highest average
MALTA (Electoral Act of 27
September 1991, last amended 2002)
Chamber (monocameral system): Il-Kamra Tad-Deputati
Number of seats: 65 seats
Constituencies: 13
multi-member constituencies (5 seats per constituency)
Distribution of seats: Application
of the single transferable vote, the quotient used is the Hagenbach-Bischoff
quotient. If a party obtains the majority of first preferences, it is granted a
“bonus” supplement of seats if needed, to ensure that it has a majority in the
Chamber.
NETHERLANDS (Electoral
Act of 28 September 1989)
Lower Chamber:
Tweede Kamer der
Staten-Generaal
Number of seats: 150
seats
Constituencies: 18
subdivided constituencies
Guiding principle: Full
proportional representation
Method of voting: Preferential
vote
Distribution
of seats: The seats are
distributed at national level in accordance with the d’Hondt method. A
threshold of 0.67% of the votes cast at national level is required in order to
take part in the distribution.
NORWAY (Electoral Act of 7
September 1984)
Chamber (monocameral system): Stortinget
Constituencies:
19
multi-member electoral constituencies (from 4 to 15 deputies) corresponding to
the 19 provinces
Guiding principle: Proportional
representation
Method of voting: Lists
Distribution of seats: Modified
Saint-Laguë method
POLAND (Electoral Act of
12 April 2001)
Lower Chamber: Sejm
Number of seats: 460 seats
Constituencies: 41
provincial multi-member constituencies (between 7 and 19 seats per
constituency)
Guiding Principle: Proportional
Distribution of seats: -
391 deputies are elected from local lists in multiple-seat constituencies.
Seats are divided between the parties using the d’Hondt method and then
allocated to the candidates who individually obtained the highest number of
votes. In order to be eligible for seat allocation, a party must obtain 5% of
votes cast at national level and a coalition 8% of votes. These thresholds do
not apply to the national minority lists.
-
69 deputies are elected from national lists. Seats are divided between the
lists using the d’Hondt method and then allocated in the order in which the
candidates appear on those lists. Seats are allocated only to those parties and
coalitions which obtained 7% of votes cast at national level.
SPAIN (1985 Electoral
Act, last amendment: March 1995)
Lower Chamber: Congreso
de los Diputados
Constituencies: -
50 multi-member constituencies (2 seats at least per province, the remainder
being distributed on the basis of population size) corresponding to the
provinces
- 2 single-member
constituencies (the North African enclaves of Ceuta and Melilla)
Guiding
principle: Proportional, except for the constituencies
of Ceuta
and Melilla
where the first-past-the-post principle is applied. However, the small number
of seats per constituency does not allow the guiding principle to come fully
into play. The voting method is closer de facto to the
first-past-the-post method.
Manner of
voting: Closed lists
or single candidate depending on the constituency
Distribution
of seats: - multi-member constituencies: closed lists with seats being
distributed proportionately in accordance with the d’Hondt method; each elector
chooses a list from among those made available in the constituency (province)
-
single-member constituencies: plurality system.
SWEDEN (Electoral Act of
1 June 1997)
Chamber (monocameral system): Riksdagen
Constituencies: -
29 multi-member constituencies (2 to 34 seats) for 310 members
- 1 other multi-member
constituency for 39 compensatory seats
Guiding principle: Hybrid:
offsetting system
Method of voting: List
system with preferential votes
Distribution of seats: - proportional distribution in accordance with the
modified Sainte- Laguë method in the 29 multi-member constituencies. In order
to obtain a seat, a party has to obtain either at least 4% of the votes cast at
national level or 12% of the votes cast in a basic constituency.
- the
39 remaining (“compensatory”) seats are awarded by full proportional
representation on the basis of the votes obtained at national level.
Nevertheless, the seats are distributed between the constituencies. Parties
which obtain seats only by virtue of the rule of 12% of the votes cast in a
basic constituency are disqualified from taking part in this distribution. The
minimum required to be elected on the basis of the preferential vote is 8% of
the total votes cast for the candidate’s party in the constituency concerned.
UKRAINE (Electoral
Act of 22 October 1997)
Chamber (monocameral system): Verkhovna Rada
Number of seats: 450 seats
Constituencies: - 225 multi-member
constituencies
-
1 national constituency governed by the proportional system (225 seats)
Guiding Principle: Hybrid:
parallel application of first-past-the-post and proportional systems
Method of voting: closed
lists for election by proportional representation
Distribution of seats: -
single-round first-past-the-post ballot in the 225 multi-member constituencies
-
proportional ballot of lists for the other 225 seats. Seats are allocated only
to those parties and coalitions exceeding the threshold of 4% of votes cast.
3. RECALL
85. Recall is a semi-direct democratic
procedure whereby a public office holder who no longer gives satisfaction to
the electorate may be dismissed.
This procedure is similar, therefore, to the instigation of a binding
mandate. Recall may apply to a single
elected representative or an entire assembly, since in certain Swiss cantons
the cantonal assembly may be dismissed in this manner (Abberufungsrecht).
86. There are generally two grounds for
recall. The first is to control the functioning of institutions. In order to limit the use of the presidential
right of dissolution, the Lithuanian Constitution of 1991 authorises the Seimas
to decide on the holding of early presidential elections, for example. However, that decision must be adopted by a
three-fifths majority within thirty days following the first sitting of
Parliament. The second reason is to give the electorate a means of controlling
its elected representatives. In some American States recall has been introduced
for that purpose. It takes the form of a
petition which, if signed by a sufficient number of voters, enables them to
decide on whether the elected representatives should remain in office or a new
candidate should be elected. If the number of signatures falls short, the
office holder remains in place. It should be noted that the procedure is a
rarity in this day and age, since the regular holding of elections ensures
greater effective control over elected representatives.
PART TWO
CRITERIA
FOR SELECTING A PARTICULAR ELECTORAL SYSTEM AND THE IMPLICATIONS OF THAT CHOICE
87.
Given the account set out in the first part of this paper of the incredible
diversity of electoral systems available, it is difficult not to succumb to a
degree of scepticism or at least admit to the fact that relativism is
inevitable. The only obvious thing emerging from this survey is that there is
no evidence leading one to make an absolute choice in favour of any particular
electoral system. Or, more precisely, no electoral system can be the best or
the worst, or perhaps even intrinsically good or bad. In order to assess its
quality, it is necessary to know precisely what is expected of it – what is first
expected of it – by assessing the way in which it fulfils its different,
sometimes conflicting functions, namely those of the electoral system in the
context of the democratic political system. We have used this functional
approach in order to illuminate the criteria in choosing an electoral system
and the implications of that choice. We will then examine in succession the
three major functions of an electoral system, the three major models which have
been successively applied in order to fulfil those functions in Western
democracies, the advantages and drawbacks of hybrid systems and a number of
problems specific to the emerging democracies.
1. THE
THREE MAJOR FUNCTIONS OF AN ELECTORAL SYSTEM
88.
Reducing matters to their most essential, there are three main functions of an
electoral system: representation, selection and investiture.
1.1 Representation
89.
Representation is the most obvious function. Did it not give its name to the
representative system? But it is as complex as it is vital because it covers
the various fundamental and conflicting issues at stake.
a.
The most traditional and doubtless the most investigated contradiction
is that between representation as mandate, which confines the role of the
person represented to designating his or her agent, and representation as a
reflection which further requires that the agent is as similar as possible
to the person represented in order to prevent the latter’s sovereignty from
being excessively altered, not to say betrayed or substantially alienated, by
the machinery of representation. Adherents of the theory – encouraged by law
and philosophy – that the electorate determines the office uphold or upheld the
first hypothesis, whereas those more sensitive to sociology – considering that
the electorate confers the right – uphold the second.
But this contradiction does
not say everything about the complexity of the function of representation. What
in fact does it mean to serve the function of representation in an election? At
least three answers can be given to this question.
b. The first
relates to the representation of political opinions and hence of the movements
and forces giving form to them. It is axiomatic that representation itself
raises numerous problems. Is it necessary to represent all tendencies,
including the most minority ones? Should
existing tendencies be favoured to the detriment of emerging parties or should
emerging tendencies be encouraged in order to make it easier to adapt what is
on offer politically to suit new issues? Should the spirit of compromise be
fostered by making the formation of coalitions possible through mechanisms such
as the second round or “apparentement”? Most of the answers given to each of
these questions dictate the choice of the electoral system or of one or more of
its refinements.
c. The
second meaning attached to representation relates to the representation of
territories. Here too the range of options available is to a large degree open.
The first choice concerns the very existence of such representation: should the
emphasis be given to the overall representation of the electoral population –
as is the case in certain democracies with a relatively small population but
also in larger states in the case of certain elections (the European elections
in France up until a recent reform) – or should representation be broken down
to the level of smaller entities? The second choice has to do with the nature
of those entities: should a territory be divided into ad hoc electoral
constituencies or should administrative divisions simply be adopted? The third
choice relates to the demographic and perhaps spatial dimension of the entities
decided upon: should the constituencies be large enough to guarantee a measure
of heterogeneity of the electorate or smaller in order to guarantee that the
electorate is homogenous?
d. This
consideration cuts across the third meaning of representation, which relates to
the representation of specific categories of the electorate. Some of those
categories may be represented through geographical representation if where they
are to be found coincides with one or more electoral constituencies. This may
be the case with ethnic and cultural categories, which are often referred to as
national minorities in multi-cultural states. But, because of their nature or
the fact that they are more widely dispersed, other categories do not coincide
with particular areas. Is it necessary to provide in the electoral system for
specific representation and, if so, what form should that representation take?
Some systems make use of quotas, occasionally coupled with financial penalties
in the event of non-compliance, as in the case of Franceand the application of equality between men and women in elections. This
example is very interesting since it marks a conclusive break with the
traditionally abstract and formally egalitarian conception of citizenship under
the French model. According to that tradition, the citizen (and hence the
elector) is a reasonable being who allows the public interest to transcend the
specific interests of his or her gender, age, occupation, ethnic and cultural
origin and religious belief. Insofar as it introduces positive discrimination
in favour of classes which are equal under the law but de facto disadvantaged
vis-à-vis the dominant classes, the electoral system fulfils a new function. It
does not merely distribute seats in accordance with votes, but seeks to
influence a change in social relationships and becomes an overt instrument for
redistributing political power between categories of the electorate.
1.2 Selection
90.
Something which is less obvious when considering electoral systems is that
selection of the people in the governing class is as important a function as
representation. Moreover, the two functions merge when it comes to the
representation of specific categories, one of the potential forms taken by the
”pre-selection” of representatives.
91.
Essentially, what is at stake in the democratic selection of persons to govern
through elections can be assessed in three respects: that of the independence
of candidates – and hence of future elected representatives – vis-à-vis the
political machine, social fairness in recruitment and the renewal of the elite
in power.
a. The
extent to which candidates and elected representatives are dependent on the
party machine relates directly to the electoral system. The latter can
influence matters in this area in two ways. The first is whether candidatures
are for single member or multi-member constituencies. A ballot for a single
member, which brings personalities into play, naturally favours the candidate
with the greatest “personal assets”. The qualities of the candidates,
which have to do with their character, talents, training, or experience are as
important as – or perhaps more important than – their membership of a party
when it comes to determining whether they are to be candidates and whether they
are elected. They owe the fact of their election first and foremost to
themselves. This is not the case in a multi-member constituency where the most
important thing is first of all to get a place on the list of candidates.
Indeed whether one appears on the list depends almost invariably on the wishes
of the local and/or national party authorities. Moreover, the more a candidate
wishes to be on an electable position on the list, the more closely he or she
is dependent on those authorities.
The second influence concerns whether the electoral
system is plurality/majority or proportional. This is because the dependence of
elected representatives on parties is greatest in a proportional representation
system. In most cases, the head of the list or the two or three candidates
heading the list of the large parties have every chance of being elected in a
large constituency, since it is only from the third or fourth seat to be filled
that the real competition begins. In such a case, the fact that the party
places a candidate in an electable position is tantamount to guaranteeing his
or her election. It is the party and not the elector which selects the elected
representative. How can the elected representative not be indebted to the
kingmaker?
b. The
problem as to the degree of social fairness in the selection of elective elites
is less directly dependent on the choice of electoral system, even though it
can reasonably be argued that a single-member ballot favours those who already
have certain cultural and social assets whereas a list ballot enables less
privileged persons to stand. This is naturally facilitated still further by the
introduction of “positive discrimination” procedures as shown above (see
paragraph 1.1). It tends to be through the related systems of the electoral
system and the representative system that social fairness in the recruitment of
candidates can be best assured. As regards the former, particular mention may
be made of public funding for electoral campaigns and as regards the latter, of
remuneration for elected office and provisions fostering the re-employment of
elected representatives at the end of their term of office. As far as the
latter aspect is concerned, public office holders, who are often guaranteed to
return to their job at the end of their term of office, are markedly more
advantaged than their opposite numbers from the private sector.
c. The
extent to which elective elites are replaced by new blood is another major
feature of selection. It depends first and foremost on the constitutional
provisions relating to assemblies (duration of terms of office, whether or not
there is a limit to the number of successive mandates allowed, whether local
and national mandates may be combined etc.). But it also depends on provisions
concerning who may be elected: the minimum age required to stand as a candidate,
whether foreign residents may stand (citizens of other member states of the
European Union for example) or the provisions on equality between men and women
referred to above. The latter provisions, adopted when the French constitution
was revised in 1999, are potentially the most important for the renewal of
elective elites. Their effect has so far been more potential than actual
because the parties – or rather the men leading them – have taken advantage of
the difficulties entailed by the single-member ballot so as to delay their
application even if it means paying the penalties provided for by the law.
1.3 Investiture
92. Elections are not merely the time when electors
are given the chance to express their party preferences officially and
secretly; they are also the means of putting into office, more or less
directly, the party or coalition of parties which will govern the country until
the next election (or the first crisis).
This function is crucial under the parliamentary system or the
semi-presidential system of the French type, as experiences with “cohabitation”
have shown, and it remains very important under the presidential system where,
by definition, the presidential election fulfils this function, but where the
essential role played by the legislature in the political process invariably
results in the parliamentary elections having a clear impact in terms of
investiture or censure, especially when they take place between two
presidential elections, as in the case of the US mid-term elections.
a. This
function is naturally particularly significant and the most strongly felt in
countries relying on the plurality and majority systems for their elections,
especially all its most extreme form of the single-round plurality ballot,
which is notorious for fostering the development of a functional two-party
system involving two potentially majority organisations which assume in turn,
with swings in one or the other direction, the functions of government and counter-government
(which means the official opposition whose aim is to become the governing
party). In democracies of this type,
electors in voting have the direct, clear potential to put their government
into office. It is therefore possible to
speak in this connection of a “governing democracy”.
b. This
is not necessarily the case with democracies whose electoral culture is that
of proportional representation, unless there are sociological
(exceptionally clear social divisions reflecting religious and cultural splits,
for example) or historical (memories of a trauma associated with the absence of
a government majority) factors which come into play. This is because the logic of proportional
representation is naturally to give preference to the function of
representation over the function of investiture. The more this logic is adhered to
(constituencies having numerous seats, rule of the highest average, no
threshold and no second round or majority correcting factor), the more the
election becomes a mere poll of all the existing political tendencies. Pure proportional representation favours a
multi-party system and the jealously guarded independence of parties. It deprives the elector of the possibility to
have a direct influence in the formation of the government, which is invariably
the outcome of bargaining between the parties represented in parliament, more
often than not after rather than before the elections. Hence the expression “governed democracy”
that is used for these types of systems in which the conclusive choice of the
government majority is monopolised by the political class, thereby depriving
the sovereign people of the full exercise of their sovereignty: “democracy
without the people”, to borrow the title of a work by Maurice Duverger on the
French tradition of the Parliamentary Republic before 1958.
c. The
example of France also shows that proportional
representation is not the only system being questioned. Doubtless this form of electoral system may
be introduced to respond to exacerbated criticism to the effect that in the
long run a majority system causes small and medium-sized parties to be excluded
from the distribution of seats and that the smallest parties even find it
difficult to survive as independent parties in their own right. But it can also “record” a fundamental
political split that the majority system – especially in a less effective form
than the first-past-the-post – cannot eradicate owing to cultural factors or
historical events resulting in irreconcilable differences (war, revolution,
economic crisis, etc). This was the case
in France between the two world wars when the continuance of a revolutionary
movement in a working class awakened by the Soviet revolution had created an
irreconcilable cleft between the communists and the socialists and the
continuing cleft between conservative Catholicism and anti-clerical radicalism
had prevented the right from regrouping even before the extreme fascist right
burst on to the scene. In such a socio-political setting, two-round majority
voting was considered above all from the point of view of the results of the first
round – which enabled each political family to count its electors – rather
than the second, which, admittedly, eradicated “in functional terms” the
divisions within the two major tendencies by presenting a façade of a
right/left split, but the latter did not take long to shatter in the Assembly
when major problems facing the country had to be dealt with. The lack of interest in the second round is
so manifest and seems so natural that second-round results were virtually never
quoted before the 1970s in the historical and political-science literature
dealing with the period. So much so that
a man with a mind as keen as Maurice Duverger claimed to study the
“inequalities of representation” under the Third Republic by relating the
distribution of seats following the second round to the distribution of votes
in the first, without paying the slightest attention to the transfers of votes
between the first and second rounds and without this legerdemain seeming to
pose any problem to him.
93. The
latter considerations show how complex it is to undertake a functional
interpretation of electoral systems. The
truth is that for every major stage in the development of democracy there is a
corresponding “ideal/typical” electoral system which carries out in a
particular way the three functions described above. This specific combination may be appraised in
the three successive forms which they took in the course of the last one
hundred and fifty years.
2. THE THREE HISTORICAL MODELS FOR THE
FUNCTIONALITY OF ELECTORAL SYSTEMS
94. If we consider the development of elective
democracy in very broad terms, we can see three major stages, corresponding to
three socio-political models for the electoral system: the elitist model, the
mass organisations model and the consumerist individuation model.
2.1 The elitist model
95. a. This model corresponds to the running-in
period of universal suffrage. The three major functions of the electoral system
are fulfilled in such a way that it could almost be categorised as being
pre-democratic. Whether regarding
representation, selection or investiture, each is stamped by the moral,
cultural and political power of the elites vis-à-vis the great mass of
the electorate granted voting rights.
Accordingly, the logic
of representation is geared to obtaining votes. Prominent persons of all opinions canvassing for
electors’ votes appeal to their immediate clientele, namely the section of the
population that knows them directly and naturally has confidence in them in
order to defend their material and spiritual interests much better than they
could do themselves. The population
divides according to each individual’s relationship with the immediate
clientele of given prominent persons.
b. The
logic of selection is a reverse selection.
It is not so much the elector who chooses the elected representative
than the would-be elected person who gathers together his potential
electors. The personalities are
pre-selected in one camp or another and are candidates only because they have
all the tangible or intangible attributes of distinguished figures. This goes without saying in the conservative
camp because they defend the “natural right” of the traditional governing
classes to govern, although this may also be found in the liberal camp in the
form of what Maurice Agulhon described as “democratic patronage”, that is to
say, the political commitment of the new governing strata who feel that they
have been invested with a duty to speak in the name of the people in order to
defend their interests of which they may be unaware.
c. The
logic of investiture is carte blanche. Each of the clients mobilised by the
competing notables delegates to them all of his powers – without having any
idea of how it will affect him or her personally – in order that they should
pursue in parliament or, perhaps, in government the policy which seems to them
to be the best and, which therefore cannot but be the best.
d. The
most functional electoral system for such a model is the single-member,
one-round ballot. First-past-the-post with one round of voting because this
is the simplest system to understand and agree upon in an inegalitarian,
uneducated society. The single-member
system is adopted because it is most commensurate with the logic of the model,
based as it is on notables.
2.2 Mass
democracy model
96. This
model corresponds to the exploitation of the political potential of universal
suffrage through mass organisations having the aim of mobilising electors. It was established during the stage of the
emergence and development of employees’ trade unions and of mass parties
associated with them which gradually become generalised by a process of
emulation to the whole of the political field, as shaped by the major
socio-economic divisions of industrial society.
Three conditions have to be satisfied in order to move from the first to
the second model. The first is that the electors should be in some way masters
of their own votes and capable of joining a party while moving on from the
logic of clan-like affiliation of clients: it may then be considered that the
electorate has attained the threshold of individualisation. The former model
was vertical, whereas this one is horizontal.
Individuals come together because they resemble each other, not because
they depend on the same boss. The second
condition is that the electors back their party because they share a similar
ideology, not because of their personal short-term interest. It is possible to speak in this connection of
a threshold of politicisation. The third
condition is that electors of all constituencies vote more or less in
accordance with the same socio-political logic and not for reasons differing
radically from one region to another. It
is then possible to speak of the threshold of nationalisation.
a. Under
this system, the logic of representation is socio-cultural. Parties form in
line with the major social divisions (mainly the split between the
self-employed and employees) and the major cultural divisions, which are
generally based on religious affiliation (catholic/secular or between various
Christian churches). They appeal to
social groups – often consolidated by associations or trade unions – wanting
these groups to be represented as such in the assemblies.
b. The
logic of the selection is party-political. It is essentially the outcome of
militancy. Candidature as such, and then
the candidate’s position in the list, are stages in the militant’s
progress. The electors ratify this
choice in so far as they put in place those whom they regard as most
representative of their background and therefore, from this point of view, the
best of them.
c. The
logic of investiture is not directly present in electors’ minds, whether in
the form of the great myths, such as preparation for proletarian revolution, on
the one hand, or restoration of the natural order of society, on the
other. In the shorter term, the
important thing is to do whatever one can to prevent trouble-makers or
reactionary forces from unilaterally calling into question the implied
socio-political pact. As a result, the
logic of “two opposing fronts” often put forward by both sides in the 1930s is
redolent rather of trench warfare than warfare on a moving front. It is a question of prevention rather than of
investiture.
d. The
most functional electoral system for this model is clearly the list system with
proportional representation. But it
may be able to cope with a plurality or majority list system if each camp
exhibits the potential to gather electors together and it is therefore possible
to give some strength to a coalition government outside the usual compromises
of the “blocked society”.
2.3 The
consumerist individuation model
97. This model, which is tending to replace the
preceding one in the advanced democracies, is less perfected and less
wide-spread to date. But its most
significant features are beginning to emerge in most western countries which
have shared the same experience of prosperity, peace, openness to the world and
liberalisation of morals over recent decades.
Those transformations of living conditions have fostered the emergence
of a culture described by Ron Inglehart as post-materialist, a culture whose
features he has described and whose consequences on electors’ attitudes and
political behaviour he has assessed.
a. The
logic of representation has clearly abandoned the principles of loyalty
under the system of faithful patronage and affiliation to a particular
socio-cultural category. It has come
closer to the logic of the market. From
the supply constituted by the available parties and candidates, the elector
makes a choice which best – or least badly – matches his or her present
demands. Those demands are ranked by
each person in the light of the priorities which he or she gives to the issues
at stake (security, employment, environment protection, moral liberation,
taxation, globalisation, etc). As a
result, the system no longer represents social groups but the aggregation of
individual priorities assigned to the various public policies.
b. The
logic of the selection is therefore eminently political: the electors
choose the people who seem to them to be the most capable of understanding
their dominant demands and of supplying them with answers or, if you like, the
“political enterprises” best satisfying the “call for tenders” that elections
now represent for the electorate.
c. The
logic of the investiture is very much present: what is involved is a
contract concluded between the “consumer” of public policies and the
“entrepreneur” whom the consumer instructs to resolve his or her problems. But the contract is a precarious one. It clearly constitutes a fixed-term contract
for the achievement of a priority objective.
So much so that the vote is as much concerned with sanctions as it is
with investiture. The consumer may try
out a new product, but he or she may reject it with a clear conscience if he or
she does not find it entirely satisfactory.
Electoral behaviour was so stable in the past that people did not shrink
from likening it to geology. It is so
volatile nowadays that it may be more appropriate to liken it to the dramatic
changes more familiar to the field of meteorology.
d. The
most functional voting method in a model of this kind is certainly
first-past-the-post or majority, which clearly invests and sanctions and
favours alternating between political parties in power. But such a system may suffer from the
drawback that it does not leave enough room for emerging movements susceptible
of capturing new demands. As a result,
in this new culture there is a blossoming of proposals for, or attempts at,
implementing hybrid systems, predominantly first-past-the-post (or
majoritarian) for the purpose of having alternating parties in power, but
embodying a corrective feature of proportional representation in order to allow
for innovation.
98. The
table, at the end of this report, recapitulates the chief characteristics of
the three models described above. It
must not be forgotten that what we are dealing with here are “ideal/typical”
models. Instances of them in history are
often more complex and the developments less clear-cut. This is because the models described as being
successive are still in part concurrent and sometimes intermeshed. The elitist model of top-down democracy based
on patronage has not completely disappeared and here and there residual pockets
may still be observed and there may even be episodic reversions in the
behaviour of present-day electors. This is still more often the case for the
great dominant model of the twentieth century – mass democracy. A social movement inspired by active
minorities and orchestrated in street demonstrations is enough to make the
public – which had voted for an enterprise dedicated to resolving a major
problem – realign itself behind the spokesman of the categories to which it
belongs. What was formerly called the gross variables of electoral behaviour
(objective social class, subjective social class and religious affiliation, in
particular) are doubtless no longer as healthy as they used to be, but strong
traces of them linger on and re-emerge to interrupt the regular decline in
their explanatory power. The fact is that
voting has never been and doubtless will never be one-dimensional, a feature
which it shares with all human behaviour.
Moreover, it is this which adds to the interest taken by many players
and observers in hybrid systems.
3. ADVANTAGES AND DRAWBACKS OF HYBRID
SYSTEMS
99. Reformers are increasingly tending
to regard hybrid systems as a panacea. This is partly because they have
internalised the multi-functionality of electoral systems, which may lead to
conflicting provisions.
3.1
The search for the happy medium
100. a. Hybrid
systems have the main advantage of combining logics which would be
incompatible if an attempt were made to implement them without placing them
in a ranking order. This is true mainly of the majority principle – which is
essential for the investiture function – and of the proportional representation
system – which is the most favourable for the representation function. This may
also be true of single-member or multi-member constituencies – which favour
personal selection – and the closed list
– which favours discipline and hence the cohesion of parliamentary
groups and parties – or the territorial approach – which favours geographical
representation – and the national approach – which favours political/ideological
representation. A concrete example of seeking the happy medium by means of
combining and prioritising may be found in a proposal made by Alain Lancelot in
1993 to the Association Nationale des Élus Régionaux at the request of its
chair, Valéry Giscard d’Estaing, for a possible reform of the method of
electing French Regional Councils. The proposal, drawn up by Jean-Claude
Casanova and Alain Lancelot, was summarised in our report as follows: “So as
not to be exposed to accusations of manipulation, the choice of voting method
should be based on a number of strong principles, which should be stated and
classified by order of importance, since they each have the effect of limiting
the others.
1. The regional government
must be founded, within the Council, on a stable and clear majority.
2. The identity of the
regions must be affirmed through the choice of a regional constituency.
3. Representation of
minorities must be ensured.
4. Representation of the
various départements likewise.
The third principle corrects the first and the last
corrects the second. If these objectives are accepted, one is led to propose a
reform of the electoral system along the lines of a majority system and
adoption of the region as the electoral constituency in place of the départements, whilst limiting the
effects of the majority system through the representation of minorities and
ensuring that the départements are
fairly represented.”
b. Hybrid
systems also have the advantage of evening out changes in
representation, selection and the capacity for investiture or sanction. These
changes tend, in the long run, to be virtually paralysed by the application of
the systems which most favour representation, such as pure proportional
representation, and conversely run the risk of becoming somewhat chaotic when
first-past-the-post (or majority) systems are implemented in their most radical
form, owing to repeatedly alternating victors. Between “sehr langsam” and
“allegro vivace”, if not “furioso”, the happy-medium tempo “bien tempéré”
attracts those who believe that controlled change lies at the heart of good
democratic governance.
3.2 Drawbacks
and difficulties
101. a. Notwithstanding
this analysis, the main drawback of hybrid systems lies in their complexity.
Doubtless, some “simple” systems, such as proportional representation, are in
themselves very complicated when it comes to the way in which seats are
distributed, especially in the use of remainders, for an elector who is not
arithmetically minded. But at least their underlying principle is unambiguous
and everybody can grasp it straight away. That is not the case with hybrid
systems, which often involve procedures that distort the results of the ballot
boxes so as either to exclude certain votes from representation (through the
operation of thresholds) or, on the contrary, to increase the weight of other
votes (the majority bonus, for example, for the leading party). The elector who
has difficulty in understanding the complexity of the arithmetic will find it
even more difficult to accept the resulting discrepancy compared to the votes
cast. This may lead sometimes, or even often, to a feeling of alienation
vis-à-vis the operation of the electoral system, which is “manipulated by
politicians”.
b. To
this difficulty can be added that of the nature of the “mix” itself. A
hybrid system would be paralysed or simply random if the various logics it
combined did not involve an appropriate ranking. But only experience can help
in choosing the correct mix. If the secondary logic is instilled too weakly, it
will scarcely have any corrective effect on the result of the election. If,
conversely, it is too strong, the main logic system will be diluted to an
excessive extent. The problem is particularly acute when a quantitative
correction factor has to be introduced in order, for example, to set the level
of a majority bonus or the threshold required to be surpassed in order to be
represented.
c. In
order to overcome these kinds of difficulties, the Committee for the Reform
of the Voting Method, set up in France in 1992 by Prime Minister Pierre
Bérégovoy, on which Alain Lancelot sat under the presidency of Professor Vedel,
proposed maintaining the majority system with two rounds of voting for most of
the seats in the Assemblée nationale and adding, “for a non-negligible part” of
those seats, the election by proportional representation of national lists. The
representatives of the left-wing majority and the right-wing opposition sitting
alongside the experts on that committee finally agreed that 10% of the total
number should constitute the “non-negligible part” of national seats to be
filled by proportional representation, but only after bitterly debating the
very principle of a hybrid system, since they were concerned that it would look
like a return to “rule by political parties”.
The argument that finally won them over was our proposal that each
elector should be given two votes: the first to elect one of the candidates in
the single-member majority ballot, and the second so as to choose one of the
lists of candidates by proportional representation at the national level. The
originality of this hybrid system lay in its transparency, since having a
double vote would definitively preclude any accusation of inequality and any
suspicion of distorting the will of the electorate. But the idea ultimately ran
aground owing to the practical difficulty in finding fifty seats to be filled
by proportional representation, either by increasing the total number of seats,
which, in the eyes of the parliamentarians on our committee, ran the risk of
worsening anti-parliamentary feeling, or else by reducing by the same amount
the number of constituency seats, which would be very unpopular in the départements whose representation would
be reduced. And the reform was not taken on board by the Government. In France, we like to say that “the
devil is to be found hiding in the details” and hybrid systems, alas, do not
lack detail.
4. SOME
CONSIDERATIONS ON THE SPECIFIC SITUATION OF THE EMERGING DEMOCRACIES
102. The emerging democracies tend
legitimately to skip stages of political, economic and social development in
order to take short cuts to catch up with the developed states which took
decades and sometimes centuries to mature. It would be pointless and unjust to
preach patience to them. That said, it is wise to help them to make at the start
of their democratisation process the fundamental choices that are going to
shape how they develop, in full knowledge of the facts. From the point of view
of electoral systems which is our subject here, this preliminary phase, should,
at the “constitutive moment” cover three aspects: initial identification of the
nature and degree of segmentation, the choice of a model of citizenship and
finally, the choice of giving electoral democracy greater or lesser influence
over governance.
4.1
Initial identification of the nature and degree of segmentation
103. Identifying the nature and degree
of political stratification and its socio-cultural foundation is a necessary
first step. It is on this segmentation that the choice of models to be created
will depend for the purposes of defining citizenship and for the investiture of
those who are to govern.
104. In order to carry out this
identification, it seems to us that there needs to be a first general election
- for a constituent assembly for example – under proportional representation.
This is the formula best suited to measuring the initial degree of
fragmentation of the political forces and to pinpointing the nature of the
allegiances (political, religious, ethnic/cultural, etc.) which bind each of
them together.
105. Some might fear that such an overt
initial phase would not merely make an initial finding, but would firmly
establish the differences by making them visible; this would, consequently,
have a heavy influence on subsequent choices.
That objection is not a negligible one. But experience shows that this
is not necessarily the case. Setting forth the divisions may well trigger a
desire to reduce them and, in any event, all the effects – particularly the
harmful ones – of an electoral system are not really experienced until it is
applied for a second time, that is to say after two consecutive elections.
4.2
Choice of a model of citizenship
106. By using the results of an initial
consultation under proportional representation, it is possible to distinguish
the main fault lines in the electorate, to assess their independence or any
possible overlap and to evaluate their respective importance. If fragmentation
is marked and based essentially on qualitative factors (ethnic/cultural or
religious for example), the crucial choice lies between (i) a universalist
conception of citizenship which transcends de facto differences and boosts de
jure equality so as to define a general higher interest, and (ii) a pluralist
conception, that of a mosaic that recognises, protects and promotes the
specific rights of the various communities. These two conceptions are equally
valid, whatever followers of Jean-Jacques Rousseau may think. And to impose an
abstractly egalitarian conception of citizenship on a very heterogeneous
society constitutes a very weak protection against the risks of actual
discrimination or even genocide. If the divisions run deep, it is better to
recognise them officially and to try to compel the communities to recognise
their respective rights as constituting reciprocal duties.
107. The choice of one or other of these
models entails consequences for the lasting choice of electoral system after
the initial stage of proportional representation. The universalist model calls
for constituencies to be drawn up along politico-administrative lines and a
first-past-the-post or majority voting
system. The pluralist model calls for constituencies that follow the spatial
distribution of communities as much as possible and a hybrid electoral system
or proportional representation, depending on how deep the divisions go and
whether the intention is to limit their effects or to accept that they will
become more rigid.
4.3
Choice of a system of democratic governance
108. The choice extends and reinforces
that of the model of citizenship. It includes the choice between a unitary
system and a federal system and the choice between a parliamentary system, a
presidential system and, possibly, a semi-presidential system.
109. From the point of view of electoral
systems, the choice of the unitary or federal form of state organisation is not
very important. It may be noted, however, that the federal form could permit
different electoral systems to be used for elections at federal level and for
elections at the level of the federated states. One could readily imagine, for
example, a first-past-the-post or even a hybrid system with first past the post
dominating at the level of the federal parliament, whose attributions relate
more to supra-community (or “universal”) issues such as defence, foreign
affairs, human rights or economic and social bargaining for the whole of the
federation, and a system using proportional representation to a greater or
lesser extent for all or some of the parliaments of the federated states.
110. As regards “governing democracy” as
discussed above which is a characteristic feature of societies whose citizens
have a real power to elect their government, the type of political regime and
electoral system which may be associated with it introduces a very significant
difference. It is clear that the presidential system, which is intrinsically a
majority system, is the most capable of simplifying the political choice of the
nation and forcing the nation to come together. And this is all the more so if
it is supported by a first-past-the-post (majority) ballot in parliamentary
elections. Especially if parliamentary elections occur in the immediate wake of
a presidential election. The French
elections of 2002 clearly showed the beneficial effect of this, from the point
of view of the investiture of a majority, of inverting the electoral calendar
that had initially scheduled the presidential election to take place after the
parliamentary elections.
111. The presidential regime may,
however, impose too harsh a choice on a pluralist society and may run the risk
of turning into a movement towards personal power in a society whose democratic
culture is too recent. If so, it may then be appropriate to introduce a hybrid
system which gives a proportionalist outlet to minority forces and guarantees
them a role as a democratic counter-power.
112. The same arguments apply to the
parliamentary system, which, since it is intrinsically a majority system, must
rely on the electoral system to identify a governing majority on the basis of
the ballot box, whilst respecting the rights of minorities and the possibility
of alternating the parties in government. Even if they are difficult to
understand – although we have seen above that this difficulty may be reduced –
hybrid systems with plurality/majority dominating seem to be the most
functional for an emerging democracy, both from the point of view of governing
democracy and that of respecting community pluralism.
113. Be that as it may, it seems
dangerous to state, in the name of some sort of theory of legal evolution, that
proportional representation is the supreme stage or as far as one can go in
democratic elections. This is obviously not true of the old advanced
democracies and it is not at all certain that it has to be true of all emerging
democracies.
CONCLUSION
114. The closing remark in the last
section enables us to turn full circle from where we began in the general
introduction to this report: despite the preference for proportional
representation, often expressed by supporters of a “politically correct” view
of the “democratic kit” or the equally peremptory assertion that outside a pure
first-past-the-post system there can be no “governing democracy”, there is no
electoral system which is good from every angle. Each has its advantages and
its drawbacks, which vary in magnitude depending on what function fulfilled by
the electoral system is considered.
115. The stakes are high since it is a
question of identifying and implementing in practice the legitimacy of
democratic power and ensuring that it is effective. Doubtless, this should
bring forth some modesty on our part. It is not so much a question of choosing
between ideal types as identifying – from minute examination of the
socio-cultural realities, local legal traditions and the prevailing
circumstances – what constitutes the best possible mix of conflicting
solutions. There must be no hesitation in rectifying a system that is starting
to produce perverse effects, since it is as easy to get into bad habits as
good, and bad habits become difficult to eradicate when they turn into a
cultural tradition. This is a sadly relativist conclusion for a lawyer who
believes in the strength of principles. But, under cover of legal principles
and mathematics, the question of electoral systems has to do with the art of
politics, which, in order to reconcile conflicting interests peaceably,
requires everyone to compromise without compromising themselves.
TABLE SUMMARISING THE THREE
HISTORICAL MODELS
ELITIST MODEL
Period of dominance: beginnings of universal suffrage
to the start of the 20th century
Logic of
representation: patronage
Logic of selection: notability
and democratic patronage
Logic of investiture: carte-blanche for the representative
Ideal/typical system: one-round, one-member
first-past-the-post/majority
MASS DEMOCRACY MODEL
Period of dominance: first
three-quarters of the 20th century
Logic of representation: identification with a socio-cultural category
Logic of selection: militancy
Logic of investiture: indirect ex post
Ideal/typical system: proportional representation
CONSUMERIST INDIVIDUATION MODEL
Period of dominance: last
decades of the 20th century
Logic of representation: demand for priority public policies
Logic of selection: most
credible offer
Logic of investiture: contract with a limited objective and
for a specified duration
Ideal/typical
system: hybrid
system predominantly based on the first-past-the-post or majority system
