Content uploaded by André Blais
Author content
All content in this area was uploaded by André Blais on May 04, 2017
Content may be subject to copyright.
European Journal of Political Research 32: 107–129, 1997. 107
c1997 Kluwer Academic Publishers. Printed in the Netherlands.
Electoral formulas: A macroscopic perspective
ANDR´
E BLAIS & LOUIS MASSICOTTE
Department of Political Science, Universit´
edeMontr
´
eal, Canada
Abstract. The paper offers a macroscopic view of the legislative electoral formulas in force
in 166 sovereign countries, whether democratic or not. The purpose is to determine which
formulas are the most widespread, and whether the prevalence of a formula is correlated with
geographical, historical, economic, and political factors. While plurality and PR systems are
equally frequent, the former tend to be found in more populated countries. PR prevails in
Europe and South America, and among the most democratic countries, while the plurality
rule is more typical of Asia, Africa, and North America, and is especially popular among
former British colonies. Large countries are more likely to opt for single-member districts. No
significant correlation emerges between the level of economic development or French colonial
background and any electoral formula. The study suggests that the selection of an electoral
system is not merely the outcome of party interest, but is also strongly influenced by ideas
about what is good, just, or efficient.
Introduction
Asthe process of democratizationcontinues,countriesarefacedwiththe issue
of how to run elections and more particularly of which electoral formula they
should use. Much of the debate over electoral formulas pertains to the party
systems and kinds of government (one-party versus coalition) they tend to
produce (Blais 1991; Blais & Massicotte 1996). Here we look at the question
from a different perspective. We wish to understand what kinds of electoral
formulas are more widespread and which are not, and to see whether there
is any pattern in the choices that are, or have been, made. Do countries
with the same electoral formula share some common political or economic
characteristics?
Our approach is macroscopic.We look at the electoral formulas thatprevail
in the world and try to discern general patterns in what we see. This has
some important disadvantages.We cannot pay close attention to the richness
of individual cases. We cannot appreciate the specific factors that led to the
adoption of the electoral law in a given country(see, in particular, Bawn1993;
Brady & Mo 1992; Elklit 1992; Sisk 1994; Christensen 1994; Mainwaring
1994;Lehoucq1994; Lamare &Vowles1994;Juberias1994;Carstairs1980).
In depth case studies are needed to capture the gamut of considerations and
constraints that come to play in the choice of electoral systems.
108 ANDR´
E BLAIS & LOUIS MASSICOTTE
At the same time, it is important, we believe, to take some distance from
individual cases and to adopt a broad, comparative perspective. We intend
to show that there is much to learn from a macroscopic perspective: it is
useful to know which electoral formulas are most and least popular and
to sort out the main characteristics that appear to be associated with each.
Such knowledge may not provide a full-fledged explanation of why a certain
country has adopted PR. In our view, however, it constitutes a crucial first
step in that direction; by outlining general patterns, we will be in a position
to establish the broad factors that define the legitimate set of alternatives in
different groups of countries.
The study
Ouranalysiscoversall the sovereigncountriesoftheworldthathadaworking,
directly elected parliament, as of December 1995. One hundred and sixty-
six sovereign countries are included.1Our criterion for inclusion may look
exceedingly broad, in contrast with the dominant stream of research (Lijphart
1994; Powell, 1982; Rae 1967) which is typically confined to mature democ-
racies. The latter choice is appropriate to explore the relationship between
electoral rules and party systems.Our goal is different. We wish to present a
broad picture of the kinds of electoral rules that prevail in the world and to
identify clusters ofcountries with similar systems.
We do not assumethat eachcountry covered in ouranalysis hasfree and fair
elections. On the contrary, we explicitly examine whether more democratic
countries tend to adopt rules that are different from those existing in less
democratic ones. Furthermore, we analyzeseparately the patterns prevailing
in ‘democratic’ countries.
We deal with the electoral formula for the election of parliament or, in
the case of bicameral legislatures, the first chamber. We confine ourselves
to elections at the national level. We distinguish four basic electoral formu-
las: plurality, majority, proportional representation (PR), and mixed systems
which combine PR and plurality or majority (Blais 1988; Blais & Massicotte
1996).2The electoral formula that, to the best of our knowledge,3prevailed
in December 1995 in each of the 166 countries is indicated in Appendix A.
We seek to determine whether electoral formulas are associated with the
following characteristics:
1. The continent is which the country is located. We know some electoral
formulasare more widespreadinsome continents than in others. Themeaning
of such geographical differences is, of course, ambiguous. Geography may
reflect similar levels of economic development, a similar culture or history, or
patterns of diffusion. We will check whether geographicaldifferences remain
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 109
after controlling for demographic, economic, historical and political factors.
To the extent that they remain, they would point to culture and history as
important determinants of electoral Systems. We have five dummy variables:
North America, South America, Africa, Asia and Oceania. Europe is the
reference point.
2. Size. Rokkan (1970) was the first to suggest a link between the size of a
country and the choice of an electoral formula. He noted that in Europe ‘the
smaller democracies have been much more prone to accept the principle of
proportionality’ (1970: 76), arguing that ‘PR is tolerable in the smaller units
because they face lesser loads of decision-making’ (1970: 77).
Katzenstein (1985)and Rogowski (1987) expanded this reasoning. Accord-
ing to Katzenstein, the choice of proportional representation reflects a search
for compromise, and such a search is more likely to be found in small coun-
tries, which strive for unity, than in large ones, where the winner-take-all
mentality prevails (Katzenstein: 170). For Rogowski, small countries are
more likely to adopt PR and large districts because they are more dependent
on trade: ‘trade-dependent, advanced economies, which are likely in the first
place to experience strong pressures for democratic participation, will find
it advantageous to develop democratic institutions that maximize the state’s
insulation, autonomy, and stability’ (Rogowski: 212), and PR, Rogowski
argues, facilitates the achievement of those goals.
There are also practical reasons that could make PR less appealing in larger
countries. PR entails the creation of multi-member districts which tend to be
larger than the single-member districts usually found in plurality and majority
systems. This createsa particular problem in a large country: many may feel
that the large multi-member districts that would have to be created would
seriously jeopardize representatives’ capacity to reach constituents during
campaigns and to keepclose contact with them afterwards.
All this, however, is mere speculation. With one exception, there has been
no systematic analysis of the relationship between size and the choice of an
electoral system. That exception is the study by Rogowski. He does show a
negativecorrelation between size of population and PR in 24 democracies. It
remains to be seen, however, whether that relationship holds when a greater
number of countries are considered and/or other factors are controlled for.
We also need to determine whether it is the size of the population or the
size of the territory that matters.4Rogowski, for instance, looked at the
size of population. But if the concern is the closeness of the link between
representatives and their constituents, we would expect territory to matter
more than population.
110 ANDR´
E BLAIS & LOUIS MASSICOTTE
We have two indicators of size: the size of the population, and that of
the territory. We also consider population density, the number of people per
square kilometer of territory.
3. The level of economic development. There is a strong relationship
between the level of economic development and democratization (Lipset
1959, 1994; Helliwell 1994; Burkhart & Lewis-Beck 1994). As far as we can
tell, however, no study has examined whether the level of economic devel-
opment is related to the existence of specific electoral rules. Our hypothesis
is that there is no direct link between economic development and the choice
of an electoral formula. We use two indicators: per capita GDP and average
life expectancy.
4.Coloniallegacy. A number of studies have established that former British
colonies have been more likely to democratize (Weiner 1983; Dominguez
1993), to adopt parliamentary government(Stepan & Skac 1993) and plural-
ity elections (Blais & Dion 1990). Colonial background may influence the
selection of a particular system in many ways. The setting by the foreign
power of representative institutions in a colony prior to emancipation creates
a pattern that may well survive in the future. The training of local elites in
the ‘mother country’ brings them into contact with a set of tried electoral
institutions they may later be inclined to emulate in their own country.
Among colonial powers we selected Britain and France for two reasons.
First, the number of nations formerly controlled for a long period of time by
either country is high (54 and 22 respectively).Second, both colonial powers
tended to create representative institutions prior to independence.
We created two dummy variables, corresponding to former British and
French colonies.5
5. The level of democracy. Elections matter much more when they are
free and fair than when they are rigged. We would expect less democratic
countries to be less concerned about the representation of minority groups
and thus to be less willing to adopt PR. Our measure of levelof democracy is
based on Freedom House’s ratings ofpolitical rights in l994.6The ratings are
on a scale from 1 to 7, 1 representingthe highest level of rights. These ratings
were transformed into an index ranging from 0, for a country with no political
rights, to 1, for one with full rights. Freedom House also provides ratings of
civil liberties, but we use only those of political rights, which ascertain the
degree to which elections are free and fair.
We also constructed two dummy variables, old and new democracy.Tobe
considered a democracy, a country had to get a score of 1 or 2 on ratings of
political rights in 1994. A country was considered an old democracy only if
it had obtained a score of 1 or 2 ten years in a row, since 1985; it was coded
as a new democracy if it had a score of 1 or 2 in 1994 but had failed to get
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 111
such scores in any of the previous nine years. These dummy variables will
allow us to detect the existence of potential patterns among the most recent
democracies. Sources for all variables are listed in Appendix B.
Findings
Table 1 shows that among our 166 countries 59 resort to the plurality rule,7
56 have adopted PR,8and 25 the majority rule;925 countries have mixed
systems,andone country(Chile)couldnot be fit intoanyof these categories.10
Plurality and PR are thus equally popular, while majority and mixed systems
are less frequent. Larger countries are more likely, however,to have plurality
elections, and, as a consequence, 50 percent of the total population covered
by our study elects its legislature througha plurality procedure, twice as much
as the population using proportional representation (22 percent).
Table 1. Electoral formulas (all countries)
Number of countries % of countries % of population
Plurality 59 36 50
Majority 25 15 11
Proportional representation 56 34 22
Mixed 25 15 16
Other 1 1 0
Total 166
The picture is slightly different if we confine ourselves to democratic coun-
tries (see Table 2), that is countries that had a score of 1 or 2 on ratings
of political rights. We observe in that group more PR than plurality systems.
When we consider the total population voting undervarious systems in demo-
cratic countries, plurality still comes first, and we discover that 22 percent of
the population of democratic countries vote under mixed systems. This last
result may come as a surprise, as mixed systems are sometimes dismissed as
marginal. But over recent years New Zealand, Japan and Italy, in particular,
have switchedto mixed formulas, which seem to havegained in popularity.
Table 3 reports the results of logit regressionsthat indicate which character-
isticsare associated with plurality,majority,andproportional representation.11
Four sets of factors seem to affect the choice of an electoral system. First and
most important is colonial background. Former British colonies are much
more likely to have adopted the plurality rule.
The strongest predictorof the choice of an electoral formula is thus colonial
background. Former British colonies make up one-third of the countries
112 ANDR´
E BLAIS & LOUIS MASSICOTTE
Table 2. Electoral formulas (democratic countries)
Number of countries % of countries % of population
Plurality 26 34 38
Majority 6 8 6
Proportional representation 34 44 33
Mixed 10 13 22
Other 1 1 1
Total 77
Table 3. Factors associated with electoral formulas (all countries)
Variables Plurality Majority PR Single-member
North America 0.52 0.02 0.29 0.41
South America 0 41 0.18 0.60 0.38
Africa 0.52 0.01 0 27 0.09
Asia 0.55 0.05 0 31 0.05
Oceania 0.50 0.57 0.40 0.11
British Colony 0.44 015 026 0.41
Degree of democracy 0.03 0 15 0.43 0.24
Size of territory (log) 0.06 0.02 0.04 0.11
N 140 140 140 141
2log likelihood 118.6 107.5 116.2 141.9
Corrected pseudo R20.59 0.31 0.59 0.42
Significant at the 0.10 level; significant at the 0.05 level; significant at the 0.01 level.
All results are based on LOGIT regression. Table entries are changes in probabilities evaluated
at the sample mean of the dependent variable, estimated from LOGIT coefficients. See Petersen
(1985). All regressions exclude mixed systems. The first three regressions, but not the fourth,
exclude Chile. The corrected pseudo R2is the Aldrich–Nelson pseudo R2with the correction as
proposed by Hagle & Mitchell (1992).
covered in our study and they show a strong propensity to opt for the plurality
rule. Our data indicate that if there had been no British influence, 36 countries
instead of 59 would have plurality legislative elections.12 The strength of
the British influence may stem from the continuous existence, in the United
Kingdom, of a single electoral formula that has successfully resisted waves
of change and of a set of political institutions that is revered by the mass
public of the ‘mother’ country (Almond & Verba 1963: 102). This confirms
the crucial role of history in the development of electoral institutions.
Former French colonies are not, for their part, significantly more likely to
haveadopted themajorityrule, the rule thatFrancehas had formostits history.
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 113
French colonial legacy thus appears to be weak. This most likely results
from the fact that the French have often changed their electoral system (see,
especially,Campbell1958)and could noteasilyprovidea single unchallenged
model for their colonies. Furthermore, the French themselves have not been
particularly proud of their electoral systems and of their political institutions
in general (Aron 1965:217).
The second most important factor is the continental variable. We may
distinguish three groups of continents: first North America, Africa, Asia and
Oceania, where the plurality rule is more frequent;second Europe, where PR
tends to prevail; and third South America, where PR dominates completely.
Europe is the heartland of proportional representation. Belgium was the
first country to adopt PR in 1899. It was followed by Swedenin 1907 and by
most European countries in the late 1910s. As Carstairs (1980: 3) shows, PR
was perceived, from the late 19th century, as the only fair system, and only
tactical considerations delayed its adoption. There is ample evidence here of
the importance of ideas, ideas brought forward by people like Victor d’Hondt
and Thomas Hare, and of a process of diffusion.13 At present, 27 European
countries out of 33 have PR for legislative elections.14
PR is even more popular in South America. All South American countries
have either a PR or a mixed system.15 This can be imputed, we believe, to the
fact that South American constitutional lawyers were trained in continental
Europe and looked to Europe as a model for the choice of an electoral system.
In the other four continents, the plurality rule is the most frequent formula.
Its popularity does vary among these four continents, but our data show that
when we control for other factors (most importantly British colonial legacy)
differences among them vanish.
The third factor that seems to be associated with the choice of an electoral
system is degree of democracy. The more democratic a country the more
likely it is to adopt PR and the less likely it is to have a majority system.
Everything else being equal, the probability of having PR is 43 percentage
pointshigher in acompletely democraticcountry,with a score of1 onpolitical
rights, than in a completely undemocratic country.
This suggests that PR is particularly popular among the most democratic
countries, perhaps because the idea that each party should get its fair share
of seats in the legislature seems so compelling to those who believe that the
vote of all citizens should weigh equally.
The last factor related to electoral rules is size, more precisely the size of
the territory. As Table 3 shows, however, contrary to what Rokkan (1970),
Katzenstein (1985), and Rogowski (1987) have argued, we do not find PRto
be more popular in smaller countries.16 We are thus led to conclude that the
114 ANDR´
E BLAIS & LOUIS MASSICOTTE
various arguments put forward by these authors,based on the assumption that
there is a greater affinity for PR in smaller countries, are not compelling.
Wedofind a relationship,however,betweensizeof territoryand the propen-
sity to use single-member districts. All PR systems have of course multi-
member districts, but there are a number of countries using the plurality or
the majority rule that have multi-member districts.17 Altogether, we have 53
countries with single-member districts, 38% of the total; 59% of the total
population votes in single-member districts.
The larger the country, the more likely it is to have single-member districts.
Whenwe move from our smallestcountry(Monaco)toour largest(Canada),18
the probability of having single-member districts increases by 74 percentage
points. The fact that the size of a country is associated with the type of
district rather than with the electoral formula suggeststhat in a country with
a large territory it is usually felt that single-member districts are required if
representatives are to keep close contact with their constituents.19
Two other factors prove not to be related to the choice of an electoral
formula. As expected,we found no correlation between the level of economic
development and the presence or a plurality, majority or PR system. We
also observed no distinct pattern among old and new democracies.Although
in general PR is more widespread among more democratic countries, that
tendency is neithermore nor less pronounced among new democracies.There
is thus no indication that a certain electoral formula is gaining or losing
ground.20
Theprecedinganalysisis based on aconsiderationof allcountries,whatever
their level of democracy. It could be objected that the choice of an electoral
formula is meaningful only when the election is truly democratic. Table 4
shows the results we obtain when the analysis is confined to ‘democratic’
countries, those that get a score of 1 or 2 on political rights. We have 77 such
countries: 26 have the plurality rule, 6 are majority systems, 34 have adopted
PR, 10 are mixed and Chile has its own peculiar rule. The following analysis
is based on 66 countries, excluding Chile21 and mixed systems.
The patterns uncovered in Table 4 are somewhat similar to those emerg-
ing in Table 3. First, in relation to colonial legacy, former British colonies
are still much more likely to have single-member districts; their propensity
to adopt the plurality rule does not appear as strong, however. As for con-
tinental patterns, Europe and South America still emerge as strongholds of
proportional representation. Finally, the patterns according to which more
democratic countries are more likely to choose PR and larger countries to
adopt single-member districts hold as well in our smaller sample.
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 115
Table 4. Factors associated with electoral formulas (democratic countries)
Variables Plurality Majority PR Single-member
North America 0.60 0.09 0.51 0.67
South America 0 39 0.09 0.48 0.31
Africa 0.46 0.03 0 32 0.27
Asia 0.54 0.09 0.39 0.13
Oceania 0.56 0.36 0.52 0.51
British Colony 0.15 0.04 0.10 0.46
Degree of democracy 0.39 0 09 0.48 0 31
Size of territory (log) 0.04 0.01 0.08 0.21
N66666667
2log likelihood 45.1 31.2 42.8 36.0
Corrected pseudo R20.70 0.32 0.72 0.74
Significant at the 0.10 level; significant at the 0.05 level; significant at the 0.01 level.
All results are based on LOGIT regression. Table entries are changes in probabilities evaluated
at the sample mean of the dependent variable, estimated from LOGIT coefficients. See Petersen
(1985). All regressions exclude mixed systems. The first three regressions, but not the fourth,
exclude Chile. The corrected pseudo R2is the Aldrich–Nelson pseudo R2with the correction as
proposed by Hagle & Mitchell (1992).
Conclusions
What do we learn from this broad overview of the electoral systems of the
world? First, the relative popularity of the four broad formulas we have
distinguished depends on the criterion one uses. In terms of total numbers,
plurality and PR are equally frequent; in terms of numbers of democratic
countries, PR is the most popular; in terms of population voting under the
various systems, plurality comes first.
We have also tried to establish whether some electoral formulas are more
likely to be found in certain types of countries and we have related the
presence or absence of electoral institutions to a number of social,economic,
and political-characteristics.
Starting with the negative findings, we have shown that the level of eco-
nomic development does not seemto affect the choice of electoral formulas.
This suggests that far from merely reflecting economic constraints, political
institutions have some logic of their own. We should keep in mind, however,
that even though economic development does not have a direct impact on
institutions, it is related to the level of democracy (Helliwell 1994), which
does have some impact. Economic developmentmay thus have some indirect
effect.
116 ANDR´
E BLAIS & LOUIS MASSICOTTE
Other variables have proved to be of greater substantial import. The most
important of all is British colonial legacy. Former British colonies are much
more likely to have single-member plurality elections, and they clearly form a
distinct set of countries in their electoral institutions. This raises of course the
issue of the durability of that colonial influence: Will it progressively wane
as time goes on? How would former colonies react should Britain switch to
PR?
Our study has also identified some clear continental patterns. Even control-
ling for other factors,PR is much more popular in Europe and South America,
while the plurality rule dominates in Asia and North America. It is not entirely
clear what these continental patterns reflect. They indicate, we believe, two
kinds of effects.First, they point to the role of history and culture as important
determinants of electoral institutions. Second, they suggest the presenceof a
diffusion processwhereby countries look at their neighbors when they adopt
electoral rules.22
We have also been able to examine systematically the relationship between
size and electoral systems. Many authors have argued that smaller countries
have a greater affinity for proportional representation. The evidence does
not support that hypothesis. We have rather shown that eventhough size does
matter,its impactis different from theone usuallyadvanced inthe comparative
politics literature. What really counts is not the size of the population but the
size of the territory and its effect is not on the electoral formula as such but
on the type of district. Larger countries find it more difficult to adopt multi-
member districts not because they nourish a less consensual type of culture
but for blunt geopolitical considerations. Multi-member districts are just too
big in a large country.
Perhaps the most interesting finding concerns the link between electoral
rules and the level of democracy. The fact that more democratic countries are
more likely to use a PR system tells us a lot about the powerful symbolic
appeal of the principle of proportional representation. The debate over the
vices and virtues of various electoral formulas is still very much alive, but
proportional representationmay be closely associated, in the minds of many,
with the ideal of democracy. Still, the absence of any dominant electoral
system among new democracies suggests that the latter are unlikely to settle
the ongoing debate in older democracies, and may indeed reflect the fact that
the jury is still out. It may also result from the unwillingness of major donor
countries to prescribe the adoption of a particular formula.
The patterns we have identified do not support a deterministic account of
how electoral systems are selected. We are able to explain only a fraction in
the total variation among countries. Yet, our findings show that the selection
of an electoral formula is not a random decision.
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 117
Finally, the overall thrust of our findings speaks to the crucial role of ideas
in the choice of electoral institutions. Most case studies of the adoption of
electoral rules stress the role of strategic considerations.The standard account
is to show how the main players seek to enact institutions that are most likely
to enhance their own power.
That picture does have some validity. We certainly do not dispute the fact
that parties generally tend to favor rules that they perceive to be in their
interest. That is, however, only part of the story. If it were the whole story
how do we explain the absence of the plurality system in South America?
Would it not be in the interest of some parties in some countries of that region
to have plurality elections? Is the configuration of party interests so different
in South America from that in other continents? Why is it that the great
majority of former British colonies have adopted single-member plurality
elections? And why is PR more frequent in more democratic countries?
The answer must be that ideas count as much as interests. As Mainwaring
(1994: 29) argues, politicians may choose electoral institutions on the basis
of their interests, but their perceptions of what institutions best suit them
are likely to be shaped by the range of options debated at a given moment.
Moreover, politicians sometimes make choices primarily on the basis of their
views about what is good, just or efficient.
Our study provides ample evidence to support that assertion. In South
America, for example, it would seem that the plurality rule is simply not on
the agenda though it would probably benefit some powerful interests. The
only debate is whether the system should be PR or mixed, and then on the
specific variants of the system. We have no doubt that strategic considerations
play an important role in choosing the specific details of electoral rules. We
suggest, however, that they do not tell us much about why there is no plurality
system in South America. To understand such a pattern, we need to adopt a
macroscopic perspective.
Acknowledgments
We thank the Social Science and Humanities Research Council of Canada
for financial support, Agnieszka Dobrzynska for research assistance, and the
referees for their comments on a previous draft of the paper. This is the
revised version of a paper presented at the Sixteenth World Congress of the
International Political Science Association, Berlin, 1994.
118 ANDR´
E BLAIS & LOUIS MASSICOTTE
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 119
120 ANDR´
E BLAIS & LOUIS MASSICOTTE
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 121
122 ANDR´
E BLAIS & LOUIS MASSICOTTE
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 123
124 ANDR´
E BLAIS & LOUIS MASSICOTTE
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 125
Appendix B: Data sources
B.1 The Dependent Variable: Electoral Formulas
We used three general sources:
Inter-Parliamentary Union (1993). Electoral systems: A world-wide compar-
ative study. Geneva: Inter-Parliamentary Union.
Inter-Parliamentary Union (various years). Chronigue des ´
elections et de
l’´
evolution parlementaire. Geneva: Inter-Parliamentary Union.
Keesing’s Record of World Events (various years). Cambridge: Longman.
We also consulted the following sources: Andrenkov & Andrenkova 1995;
Blaustein & Flanz;Christensen 1994; Cox 1995; Dimitras 1994; Diskin 1992;
Drake&Silva 1986; Espinola 1995;Gonzales1991; Grzybowski1994; Hsieh
1995; Jones 1995; Juberias 1994; Lamare & Vowles 1994; Lijphart 1994;
Lincoln 1993; Mackie & Rose 1992; McDonald & Ruhl 1989; Mathur 1991;
Morriss 1993; Mozaffar 1995; Nohlen 1993; Roberts 1994; Simon 1995; Sisk
1994; Solt´
esz 1994; Teh-fu 1994; Teague & Gould 1990.
We had access to the electoral laws of the following countries: Australia,
Austria, Azerbaidjan, Belgium, Benin, Burkina Faso, Burundi, Canada, Cape
Verde, Central African Republic, Congo, Czech Republic, France, Gabon,
Germany, Greece, Haiti, Hungary, Italy, Mexico, Nepal, Portugal, Romania,
South Africa, Spain, Switzerland, Taiwan, United Kingdom.
Dominique Arel provided information on Belarus, Moldova, Russia and
Ukraine, Roger Ramigason Uzbekistan, Kai-chiang Lin on Taiwan, Mark P.
Jones on South America, and Shaheen Mozaffar on Africa.
B.2 The Independent Variables
– Size of population, size of territory, and density (1993): Encyclopaedia
Universalis (1994). Les chiffres du Monde. Paris: Encyclopaedia Universalis.
– Life expectancy(1992): World Bank (1994).World Tables 1994. Baltimore
and London: The John Hopkins University Press.
– Per capita GDP (1993): United Nations Conferenceon Trade and Develop-
ment (1994). Handbook of International Trade and Development Statistics.
New York and Geneva:United Nations.
– Continents: United Nations Educational, Scientific and Cultural Organiza-
tion (1994). Statistical Handbook 1994. Paris: UNESCO.
– Political rights (1985–1994): Freedom in the World, Political Rights and
Civil Liberties. New York: Freedom House.
126 ANDR´
E BLAIS & LOUIS MASSICOTTE
Notes
1. According to the Europa World Year Book (1995) there are 191 sovereign countries
in the world. Twenty countries were excluded because they did not have a working,
directly elected Parliament: Afghanistan, Algeria, Bahrein, Brunei, Chad, China, Ery-
threa, Kazakhstan, Liberia, Libya, Nigeria, Oman, Qatar, Rwanda, Saudi Arabia, Sierra
Leone, Somalia, Sudan, United Arab Emirates, Zaire. Five other countries could not be
included because of missing information: Armenia, Bosnia-Herzegovina, Guinea-Bissau,
Marshall Islands, Palau.
2. There are of course different types of mixed systems, some being closer than others to PR
(see Blais & Massicotte 1996). Some might argue that a mixed system, like the German
one, where PR seats are distributed in a corrective way, is, in the end, a PR system.
They are right with respect to the degree of disproportionality the system produces. But
a formula ought to be classified on the basis of its operation rather than its consequences.
We redid all the regressions classifying corrective mixed systems as PR; this does not
modify our conclusions.
3. We wish to thank one referee who pointed out one error in the previous version of our
Appendix.
4. In our data set, the correlation between these two variables is only 0.37.
5. A country is deemed to have a colonial background if it was formerly a colony, a
protectorate, or a trust territory held on behalf of the League of Nations or the United
Nations. We did not create a Spanish colony variable because almost all former Spanish
colonies separated from Spain in the early nineteenth century, when Spain did not offer
a set of political institutions to be emulated, and had created none in its colonies prior to
independence. Indeed Latin American constitution-makers seem to have been inspired
more by the USA than by Spain (Edelman 1965: 396; Lambert 1968: 344).
6. For a description of the methodology, see Ryan (1995).
7. This includes three countries which use the single non-transferable vote.
8. Including two countries with the single transferable vote.
9. Including two countries with the alternative vote.
10. Chile has two-member districts. If the leading party gets twice as many votes as the
second party, it wins the two seats; otherwise, the top two parties each win one seat.
11. We performed a similar analysis for mixed systems but found only two significant
correlations: mixed systems are less frequent among former British colonies but more
popular among economically developed countries (where life expectancy is higher).
These mixed systems are excluded in Tables3 and 4.
12. This result is obtained by estimating the predicted probability of having a plurality system
for each country, on the basis of the LOGIT coefficients, assuming that each country has
a score of 0 on British colony and keeps its observed score on all other variables. The
average probability for the 140 countries included in Table 3 is then 0.26, thus yielding
the prediction that if there wasno British colony, 36 (0.26 140) countries, instead of 59,
would have a plurality system. For an explanation of the methodology, see Rosenstone
& Hansen (1993: 73).
13. Carstairs (1980: 64) notes that the adoption of PR in the Netherlands was facilitated by
the fact that ‘Belgium, the immediate neighbour of Netherlands, had earlier been the
pioneer’.
14. The reader should keep in mind that mixed systems, as well as Chile, are excluded in this
part of the analysis. See supra, note 10.
15. The coefficient for South America does not come out significant because there are only
eight countries. The fact that the three countries with mixed systems have a strong PR
component makes us believe that there is a clear pattern in favor of PR in South America.
16 We get similar results when we use size of population. Density also proved to be non-
significant.
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 127
17. There are 21 countries with multi-member districts in the case of plurality systems and
10 in the case of majority systems.
18. Russia is the largest country in our data set but it is excluded in the regression analyses
because it has a mixed electoral formula.
19. The fact that it is the size of the territory rather than the size of population that matters
also supports that interpretation.
20. It is true that our model predicts that if the process of democratization continues, the
propensity to adopt PR should increase. On the other hand democratization should occur
mainly in Africa and Asia, and our data show countries in these two continents find the
plurality rule particularly attractive.
21. Chile is included in the regression concerning the choice between single and multi-
member districts.
22. Asimilarfinding isreported byPrzeworskiet al. (1996: 43) who indicate that a democracy
is more likely to survive if it is situated in a region with many democratic countries.
References
Almond, G. & Verba, S. (1963). The civic culture: Political attitudes and democracy in five
nations. Princeton: Princeton University Press.
Andrenkov, V. & Andrenkova, A. (1995). Different ways to democracy: A comparative analysis
of the electoral systems of Russia, Ukraine, Lithuania, Latvia, and Estonia. Paper presented
at the Conference on Comparative DemocraticElections, Cambridge (Mass.).
Aron, R. (1965). D´
emocratie et totalitarisme. Paris: Gallimard.
Bawn, K. (1993). The logic of institutional preferences: German electoral law as a social choice
outcome, American Journal of Political Science 37: 965–989.
Blais,A.(1988).Theclassificationofelectoralsystems,EuropeanJournalof PoliticalResearch
16: 99–110.
Blais, A. (1991). The debate over electoral systems, International Political Science Review 12:
239–260.
Blais, A. & Dion, S. (1990). Electoral systems and the consolidation of new democracies, in D.
Ethier (ed.), Democratic transition and consolidation in Southern Europe, Latin America
and Southeast Asia. London: Macmillan.
Blais, A. & Massicotte, L. (1996). Electoral systems, in L. LeDuc, R.G. Niemi & P. Norris
(eds.), Comparing democracies. Beverly Hills: Sage.
Blaustein, A.P.& Flanz, G.P.(various years). Constitutions of the countries of the world.Dobbs
Ferry: Oceana Publications.
Brady, D. & Mo, J. (1992). Strategy and choice in the 1988 national assembly election of
Korea, Comparative Political Studies 24: 405–429.
Burkhart, R.E. & Lewis-Beck, M.S. (1994). Comparative democracy: The economic develop-
ment thesis, American Political Science Review 88: 903–910.
Campbell, P. (1958). French electoral systems and elections, 1789–1957. London: Faber.
Carstairs, A. McClaren (1980). A short history of electoral systems in Western Europe. London:
Allen and Unwin.
Christensen, R.V. (1994). Electoral reform: An analysis of individual and party level incentives
based on Japanese examples. Paper presented at the annual meeting of the American
Political Science Association, New York.
Cox, G. (1995). The politics of electoral coalitions. University of San Diego. Typescript.
Dimitras, P.E. (1994). The Greek parliamentary election of October 1993, Electoral Studies
13: 235–239.
Diskin, A. (1992). The Israeli general election of 1992, Electoral Studies 11: 356–361.
128 ANDR´
E BLAIS & LOUIS MASSICOTTE
Dominguez, J. (1993). The Caribbean question: Why has liberal democracy (surprisingly)
flourished?, in J. Dominguez (ed.), Democracy in the Caribbean: Political, economic, and
social perspectives. Baltimore: John Hopkins University Press.
Drake, P.W. & Silva, E. eds. (1986). Elections and democratization in Latin America, 1980–85.
San Diego: University of California Press.
Earle, V. (1968). Federalism. Infinite variety in theory and practice. Ithaca, NY: F.E. Peacock.
Edelman, A.T. (1965). Latin American government and politics. Homewood: Dorsey Press.
Elklit, J. (1992). The best of both worlds? The Danish electoral system 1915–1920 in a
comparative perspective, Electoral Studies 11: 189–205.
Espinola, R. (1995). Democracy and elections in Latin America. Paper presented at the Con-
ference on Comparative Democratic Elections, Cambridge (Mass.).
Gonzales, L.E. (1991). Political structures and democracy in Uruguay. Notre Dame, IN:
University of Notre Dame Press.
Grzybowski, M.T. (1994). Parliamentary elections in Czech and Slovak Republics and in
Poland, 1991–1993. Paper presented at the 16th International Political Science Association
World Congress, Berlin.
Hagle, T. & Mitchell, G.E., II (1992). Goodness-of-fit measures for probit and logit, American
Journal of Political Science 36: 762–84.
Helliwell, J.F. (1994). Empirical linkage between democracy and economic growth, British
Journal of Political Science 24: 225–248.
Hsieh,J.Fuh-sheng(1995). Electoral politics in newdemocracies in the Asia-Pacific area. Paper
presented at the Conference on Comparative DemocraticElections, Cambridge (Mass.).
Jones, M.P. (1995). A guide to the electoral systems of the Americas, Electoral Studies 14:
5–21.
Juber´
ıas, C.F. (1994). Electoral systems in Eastern Europe: How are they changing? Why are
they changing?. Paper presented at the 16th International Political Science Association
World Congress, Berlin.
Katzenstein, P.J. (1985). Small states in world markets: Industrial policy in Europe. Ithaca,
NY: Cornell University Press.
Lamare, J.W. & Vowles, J. (1994). Changing an electoral system: The case of New Zealand.
Paper presented at the annual meeting of the American Political Science Association, New
York.
Lambert,J.(1968).Am´
eriquelatine: Structuressociales et institutionspolitiques. Paris:Presses
Universitaires de France.
Lehoucq, F.E. (1994). Institutional change and political conflict: Evaluating alternative expla-
nations of electoral reform in Costa Rica. Paper presented at the annual meeting of the
American Political Science Association, New York.
Lijphart, A. (1984). Democratic patterns of majoritarian and consensus government in twenty-
one countries. New Haven, CT: Yale University Press.
Lijphart, A. (1994) Electoral systems and party systems: A study of twenty-seven democracies,
1945–1990. Oxford: Oxford University Press.
Lincoln, A. (1993). The Georgian election of 1992, Electoral Studies 12: 174–179.
Lipset, S.M. (1959). Some social requisites of democracy: economic development and political
legitimacy, American Political Science Review 53: 69–105.
Lipset, S.M. (1994). The social requisites of democracy revisited, American Sociological
Review 59: 1–22.
Mackie, T.T. & Rose, R. (1992). The international almanach of electoral history. Washington,
DC: Congressional Quarterly Press.
Mainwaring, S. (1994). Explaining choices of political institutions: Interests and ideas in
Brazil,1985–1988. Paper presented at the annual meeting of the American Political Science
Association, New York.
Mathur,H. (1991).ParliamentinMauritius.Stanley and Rose-Hill: ´
Editionsde l’Oc´
eanIndien.
McDonald, R.H. & Ruhl, J.M. (1989). Party politics and elections in Latin America.San
Francisco and London: Westview Press.
ELECTORAL FORMULAS: A MACROSCOPIC PERSPECTIVE 129
McGregor, J.P. (1993). Fragmentation and capacity in central and east European parliaments.
Paper presented at the annual meeting of the American Political Science Association, New
York.
Morriss, P. (1993). The South Korean national assembly elections 1992, Representation 31:
60–62.
Mozaffar, S. (1995). The political origins and consequences of electoral systems in Africa:
A preliminary analysis. Paper presented at the Conference on Comparative Democratic
Elections, Cambridge (Mass.).
Nohien, D. (1993). Enciclopedia electoral Latino-americana y del Caribe.SanJos
´
e: Instituto
Interamericano de Derechos Humanos.
Ordeshook, P.L. (1994). Ethnic heterogeneity, district magnitude, and the number of parties,
American Journal of Political Science 38: 100–123.
Petersen, T. (1985). A comment on presenting results from logit and probit models, American
Sociological Review 50: 130–131.
Powell, G.B. (1982). Contemporary democracies: Participation, stability, and violence.Cam-
bridge, MA: Harvard University Press.
Przeworski, A., Alvarez, M., Cheibub, J.A. & Limongi, F. (1996). What makes democracies
endure?, Journal of Democracy 7: 39–55.
Rae, D.W. (1971). The political consequences of electoral laws. New Haven, CT: Yale Uni-
versity Press.
Roberts, D. (1994). The Cambodian elections of 1993, Electoral Studies 13: 157–162.
Rogowski, R. (1987). Trade and democratic institutions, International Organization 41: 203–
225.
Rokkan, S. (1970). Citizens, elections, parties.NewYork:DavidMcKay.
Rosenstone, S.J. & Hansen, J.M. (1993). Mobilization, participation, and democracy in Amer-
ica. New York: Macmillan.
Ryan, J.E. (1995). The comparative survey of freedom – 1994–95 survey methodology, in J.
Finn (ed.), Freedom in the world. New York: Freedom House.
Simon, J. (1995). Electoral systems and democracy: comparative electoral systems in Central
Europe during the transition. Paper presented at the Conference on Comparative Demo-
cratic Elections, Cambridge (Mass.).
Sisk, T.D. (1994). Electoral system choice in South Africa: Implications for intergroup mod-
eration. Paper presented at the 16th International Political Science Association World
Congress, Berlin.
S´
oltesz, E. (1994). Democracies in Eastern Europe: Hungary, Poland and Slovakia. Paper
presented at the 16th International Political Science Association World Congress, Berlin.
Stepan, A. & Skach, C. (1993). Constitutional frameworks and democratic consolidation:
Parliamentarism versus presidentialism, World Politics 46: 1–22.
Taagepera, R. & Shugart, M.S. (1989). Seats and votes: The effects and determinants of
electoral systems. New Haven, CT: Yale University Press.
Teague, R.C. & Gould, R.A. (1990). Guyana. A pre-election technical assessment report.
Washington: International Foundation for Electoral Systems.
Teh-fu, H. (1994). Political consequences of the single non-transferable vote system: The
cases of Japan and Taiwan. Paper presented at the 16th International Political Science
Association World Congress, Berlin.
Weiner, M. (1987). Empirical democratic theory, in M. Weiner &E. Ozbudun (eds.), Compet-
itive elections in developing countries. Durham, NC: Duke University Press.
Address for correspondence: Andr´
e Blais, Department of Political Science, Universit´
ede
Montr´
eal, C.P. 6128, Succursale Centre-Ville, Montr´
eal, Canada H3C 3J7
Fax: (514) 343-2360; E-mail: blaisa@ere.umontreal.ca
