Академический Документы
Профессиональный Документы
Культура Документы
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=cpsa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Canadian Political Science Association and Société québécoise de science politique are collaborating with
JSTOR to digitize, preserve and extend access to Canadian Journal of Political Science / Revue canadienne de
science politique.
http://www.jstor.org
PartySystemsandPatternsof Government
in WesternDemocracies
*Paper read at the Seventh World Congress of the International Political Science Associa-
tion at Brussels,September 1967.
1Although the literature on parties is vast, the number of works specifically devoted to the
study of party systems is relatively small, at least in relation to systems of more than one
party. In this context, practically all the discussionshave been related to the contrast between
two-party systems and other systems, as is for instance the case in M. Duverger's Political
Parties (Paris, 1951). The same applies to S. Neumann's Modern Political Parties (Chicago,
1956). More recently, distinctions have been made between stable and less stable multi-
party systems, usually on the basis of contrasts between Scandinavian countries and the
Netherlands, on the one hand, and France on the other. R. A. Dahl's Political Opposition
in Western Democracies (New Haven, 1966) considers this problem in relation to each
country and in a general fashion, but the precise characteristicsof the party systems of the
various countries of the group are not specifically mentioned. The forthcoming study of the
Smaller European Democracies will also help to fill the gap in this field. Interest in the
question was shown at the 1967 International Political Science Association Congress,
where a number of the papers (including the present study) were related to the close
examination of types of party systems, in particular those of A. Lipjhart, "Typologies of
Democratic Systems," and G. Sartori, "Parties and Party Systems: A Theoretical Frame-
work."
The present article would not have been written but for the suggestion of H. Daalder
whose interest in the subject has been long-standing and whom the author wishes to thank
particularly. Finally, the most comprehensive general review of types of party systems in
both Western and developing areas is that of S. M. Lipset, "Political Cleavages in
'Developed' and 'Emerging'Polities," in E. Allardt & Y. Littunen, eds., Cleavages, Ideologies
and Party Systems (Helsinki, 1964).
Canadian Journal of Political Science/Revue canadienne de Science politique, I, no. 2
(June/juin 1968). Printed in Canada.
Systemes de partis et types de structures gouvernementales dans les democraties
occidentales
PARTY SYSTEMS
If undertaken on a world basis, the analysis of party systems would require a
consideration of the number of parties, of their strength, of their place on the
ideological spectrum, of the nature of their support, and of their organization
and type of leadership. In the context of Western liberal democracies, it is
possible to limit the analysis to the first three of these characteristics. With
very few exceptions, Western parties can be deemed to be of the "legitimate"
mass type and to have a regularized system of leadership selection. Charismatic
leadership is exceptional and, at the other extreme, Communist parties are
becoming increasingly similar to other parties. Western parties appeal nationally
to the electorate by trying to put across a general image; differences are more to
184 J. BLONDEL
Second, five countries give between 75 and 80 per cent of their votes to the two
major parties (though Belgium fell markedly below this level at the last election).
In six countries, the two major parties obtain about two-thirds of the votes
(though Holland is somewhat different, as we shall see). Finally, three countries
give about half their votes to the two major parties, though, for the last two
elections, France would have to be moved from the fourth to the third category:
but as the French situation may well alter again after the end of the de Gaulle
era, it is probably more realistic to draw conclusions from the average of the
whole post-war period than merely from developments in the 1960s.
Countries belonging to the first group can be defined as two-party systems,
though there is some ambiguity in the cases of Australia and Austria; in both
these countries some governments depended for their constitution or maintenance
on the support of more than one party. The five countries of the second group
constitute the three-party systems, Germany having arrived at this status as a
result of the operation of the electoral system as well as of the Adenauer tactics.
The nine countries of the last two groups are the genuine multi-party systems, in
which four, five, or even six parties play a significant part in the political process.
2. Strengthof parties
Little has to be said about countries of the first group, in which 90 per cent or
more of the electors vote for, and 90 per cent or more of the seats are distributed
between, the two major parties. It is interesting to note that discrepancies in
strength between the two major parties are remarkably small, at least if averaged
over the post-war period. While it could in theory be the case that one party
might be permanently much larger than the other, no two-party system (i.e., no
system in which the two parties obtain 89 per cent of the votes) gives to the
larger of the two parties a permanent premium of over 10 per cent of its own
electorate. Despite differences in social structure, for instance, between the
Western Democracies 185
United States, the United Kingdom, and Austria, the electorate distributes its
preferences fairly evenly between the two parties. Although it cannot be stated
with assurance that such a situation cannot exist in any type of social system
(there are indeed examples of uneven distribution of party support among two-
party systems outside Western democracies), it seems possible to hypothesize, in
the absence of contrary evidence, that, in Western democracies, two-party
systems show a tendency towards a relative equilibrium between the two parties.
Countries in the three-party system group also share several characteristics.
TABLE II
AVERAGE STRENGTH OF THE TWO MAJOR PARTIES IN TWO- AND TWO-AND-A-HALF PARTY SYSTEMS
(percentagevotes cast)
Difference Difference
United States 49 50 1 Germany 45 35 10
New Zealand 48 47 1 Canada 36 43 7
Australia 47 46 1 Belgium 43 35 8
United Kingdom 45 47 2 Eire 46 29 17
Austria 46 43 3
Averagedisparitybetween
the two majorparties 1.6 10.5
First, disparities in strength between the two larger parties are generally much
greater than among countries in the first group. They seem structural in that the
swing of the pendulum does not, as among countries of the first group, diminish
the gap, even if one considers a fairly long period: the average percentage point
difference between the two major parties among the five countries of the first
group is only 1.6 if all post-war elections are taken into account; it is 10.5 in
countries of the second group (excluding Luxemburg where data for all elections
were not available). The often stressed phenomenon of the German SPD, which
has not been able to achieve equality of strength with the CDU, is thus a general
phenomenon among countries of this group. It can therefore be stated that in
countries in which about 80 per cent of the votes go to two parties the distribu-
tion of the support tends to be uneven.
Three-party systems also have another, and converse, characteristic. They
all have two major parties and a much smaller third party. While it would seem
theoretically possible for three-party systems to exist in which all three significant
parties were of about equal size, there are in fact no three-party systems of this
kind among Western democracies: all the three-party systems are in the second
group; all systems of the third group have more than three significant parties. It
can thus be stated that, while theoretically possible, a genuine three-party system
is not a likely type of system among Western democracies: countries of the
second group should therefore be more strictly labelled as "two-and-a-half-party
systems." It may indeed be permissible to add, after considering the evolution of
party systems, particularly in the early part of the twentieth century, that
genuine three-party systems do not normally occur because they are essentially
transitional, thus unstable, forms of party systems.
The situation arising out of the 1965 Belgian election was thus of particular
186 J. BLONDEL
interest, since the upsurge of the Liberal party and the decline of the two major
parties was such that Belgium appeared to have moved into the exceptional and
seemingly transitional position of a "genuine" three-party system. If the proposi-
tion advanced earlier is correct, it would seem that in the next few years Belgium
will move to one of three types of further changes. The Liberals could return to
their "normal" position of small party; they could displace one of the two major
parties (an unprecedented development in Western Europe since all the move-
ments happened in the other direction); a split could occur among the supporters
of one or both of the major parties and Belgium might move from the second
group of countries (two-and-a-half-party systems) to the third or fourth (multi-
party systems). From the experience of other Western democracies the first and
third of these three outcomes appears more probable than the maintenance of a
genuine three-party system.
We included Holland among the countries which constituted the third group,
composed of those countries in which the two largest parties obtained about
two-thirds of the votes of the electorate. But if we consider the relative strength
of the two major parties, the five other countries of the group differ markedly
from Holland, in that they have one very large party, which might be defined as
dominant as it obtains about 40 per cent of the electorate and generally gains
about twice as many votes as the second party. In fact, in both Norway and
Sweden, the mechanics of the electoral system have sometimes enabled this
dominant party to obtain the absolute majority of seats in the Chamber. But
these countries are at the same time multi-party systems as they all have four or
five significant (and often well-structured) parties playing a crucial role in the
operation of the political system. Four of the five Scandinavian countries are
included in this group (the fifth being Italy), though in Iceland the dominant
party is not the Socialist but the Conservative (Independence) party. The Fifth
Republic, at least under de Gaulle, appears to be moving towards this pattern of
party system, though the characteristics of the Gaullist party are such that it
seems unreasonable to predict that the present party configuration is likely to
survive de Gaulle. Thus, alongside two-party systems and two-and-a-half-party
systems, the third group should be defined, not so much as a multi-party system
in which two parties obtain about two-thirds of the votes of the electorate, but
as a multi-party system with a dominant party which obtains about two-fifths and
less than half of the votes.
Finally, the last four countries (if Holland is included and France maintained
in the group in the view of the past and possible future performance of that
country) constitute the "genuine" multi-party systems: they have no dominant
body and indeed seem to show a pattern of political behaviour in which three or
four parties are equally well placed to combine or form coalitions. Recent elec-
toral movements in Holland appear to bring that country gradually nearer
towards this model: the slow decline of the Labour party and the patterns of
governmental formation have tended to conform fairly closely to those which
have been customary of Finland.
TABLE III
THE IDEOLOGICAL SPECTRUM OF PARTIES IN WESTERN DEMOCRACIES
Group 1
Two-partysystems
United States L L
New Zealand L e L
Australia L e L
United Kingdom L e L
Austria L e? L
Group2
systems
Two-and-a-half-party
Germany L s L
Belgium L s L
Luxemburg L s L
Canada s L L
Eire s L L
Group 3
Multi-partysystemswith one dominantparty
Denmark L s/m s /m
Norway L s s s s/m
Sweden L s/m s /m
Iceland s/m s/m M L
Italy M s s L s
Group4
Multi-partysystemswithoutdominantparty
Netherlands M s M s
Switzerland M M M s
France M /s/ s M
Finland M M s M s
L = largeparty(about40 per cent)
M = medium-sizedparty(somewhatover 20 percent)
s/m = smallto medium-sizedparty(about 15 per cent)
s = smallparty(about 10 per cent or even less)
e = verysmallparty(only mentionedin relationto two-partysystems)
188 J. BLONDEL
or Christian party, together with a small other group, whose position in the
centre is perhaps sometimes problematic. Countries of group 2 have two types of
party systems: Belgium, Germany, and Luxemburg resemble countries of
group 1, except that the centre party is stronger and, as we noted earlier, this
increased strength is mainly at the expense of one only of the two major parties
(in fact the Socialist party). The other two countries of the group (Canada and
Eire) are different: the small party is not the centre, but the left-wing, party and,
probably not quite accidentally, the right-wing party is not Christian but Con-
servative. In the early part of the twentieth century, it would probably have been
argued that both these countries were still in a transitional stage: Canada has a
party system not unlike that of Britain in 1906. But it must by now be recognized
both that the two-and-a-half-party system is stable (Belgium and Germany have
had for long periods the British party system of the 1920s and do not appear to
move further towards a two-party system) and that the two-and-a-half-party
system is stable even if the left-wing party is the small party. Neither the Irish
Labour party nor the Canadian NDP appear to be in a position to overtake the
centre party in the near future: their position of small party seems stable. The
reasons for the stability of the Canadian model of two-and-a-half-party system
are probably to be found along lines similar to those which account for the
stability of the American parties, which have remained in existence despite
earlier predictions to the contrary.
Two types, and not more than two types, of multi-party systems with a
dominant party can be found. Three Scandinavian countries have a dominant
Socialist party; Iceland and Italy have a dominant Conservative or Christian
party. In the latter two cases, we encounter for the first time countries with a
really large Communist party. In Iceland and Italy (as in Finland and even
France) the Communist party remained fairly stable throughout the whole
period; the Socialist party is therefore correspondingly weak and the left is
divided; as is well known, the converse occurs in Norway, Denmark, and Sweden.
In these three countries, though the agrarian party is not placed symmetrically in
relation to the Communists on the political spectrum, it probably plays the same
divisive part. The absence of religious divisions might have led one to expect the
three Scandinavian countries which do not have a large Communist party to have
two-and-a-half-party systems, but the apparent feelings of identity of the agri-
cultural community have created permanent cleavages on the right which have
produced a party system in many ways "symmetrical" to that of Iceland and
Italy.
Countries of the fourth group have no dominant party: party strengths are
therefore fairly evenly spread across the ideological spectrum, though Holland
and Switzerland, with weak Communist parties, display somewhat less spread
than France and Finland. This latter country combines the splintering charac-
teristics of left and right of the Scandinavian countries: the divisions of Iceland
are superimposed on those of Sweden.
There are thus six types of party systems in Western democracies. At one
extreme are the broadly based parties of the two-party system countries: the
United States is the most perfect case of this type, but four other countries
closely approximate this model and they only diverge inasmuch as they have a
Western Democracies 189
small centre party and are divided ideologically between conservatives and
socialists. At the other extreme, the votes of the electors are spread fairly evenly,
in groups of not much more than 25 per cent and in many cases much less than
25 per cent over the whole ideological spectrum, as in Holland, Switzerland,
France, and Finland. Between these two poles, one finds four types of party
systems: five countries have two-and-a-half-party systems: among them, three
have a smaller centre party, while the other two have a smaller left-wing party.
The five remaining countries are multi-party systems with a dominant party,
three of them having a dominant socialist party opposed by a divided right,
largely because of the presence of an agrarian sentiment in the countries con-
cerned, while the other two have a strong right-wing party opposed by a
divided left, largely because of the presence of a substantial Communist party.
A number of types of party systems, which are theoretically possible, do not
appear to exist. There are no three-party systems, as we saw; there are no
"unbalanced" two-party systems. There are, moreover, only two existing ideo-
logical models of two-and-a-half-party systems out of the three which might have
existed. Dominant parties in multi-party systems tend to be of the right or left,
not of the centre. Patterns of party systems in Western democracies are thus
limited in number. There are discrete points at which party systems are to be
found: not only social structures but balance and equilibrium have surely to be
taken into account in an analysis of the real world distribution of Western party
systems.
1. Duration of governments
The durationof governmentscan, of course,be analysedonly for the seventeen
countriesin which the systemof governmentis parliamentary(it does not seem
necessaryto exclude France from the group, but Switzerlandand the United
Statescertainlyhave to be). For these seventeencountries,an index of average
durationof governmentswas thereforeconstructed:as such an index has to rest
on an operationaldefinitionof government,it is not possible to avoid drawing
some arbitraryboundaries. For the purposes of the present analysis, any
administration was consideredas one governmentwhichfulfilledtwo conditions:
that of being headed by the same prime minister and that of relying on the
support of the same party or parties in the Chamber.No other operational
definitionappearedacceptable.A more restrictivedefinitionwhich would have
includedreshuffleswould have had in order to be truly operational,to include
many minor changesand would have thereforemade us consideras ministerial
criseschangesof personnelwhichappearroutineto observersand are in no sense
seen as crises.Thereare majorreshuffles,admittedly;in some cases, as in Britain
in 1962, the scale of the changes is such that it can be, and might elsewhere
have been deemed to be, a ministerialcrisis, but it did not seem possible to
define operationally,at this stage, a "reshufflebar,"4above which reshuffles
could have been equatedto ministerialcrises. Conversely,while a changein the
basis of partysupportfor the governmentsurelyconstitutesa change of govern-
ment, a change of prime minister may be due to accidentalcauses, such as
ill-healthor death:the numberof crisesmay thus be somewhatoverestimatedin
some countries(as mightindeedbe the case for Denmark).But changesof prime
ministers,even with the same party support,are normallyan indicationthat a
governmentalcrisis has taken place: the French and Italian governmentalsys-
tems would become surprisinglystable if only changes in party supportwere
takeninto account.
The averagefor the seventeencountriesranges from five to seven years in
Swedenand Australiato less than a year in Italy, Finland,and France (despite
the new stabilityof the Fifth Republic) (Table IV). Countriescan be divided
into four well-definedgroups:in four countries,governmentslasted five yearsor
more (Australia,Sweden,New Zealand,and Canada);in a second groupof four
countries,governmentslastedan averageof little morethan threeyears (Britain,
Eire, Germany,and Norway); in the third group of five, governmentslasted
somewhatover two years (Austria, Luxemburg,Iceland, Denmark, and Hol-
land); and in four countries,governmentswere in office on averagefor about a
year (Belgium,Finland,Italy, and France). Thus governmentalstabilityis much
lower everywherethan is intuitivelyfelt: a Britishprime ministeris unlikelyto
remainmore than three or four years in office;Dutch governmentsare not only
4Such a reshufflebar could be defined by taking into account the number and importanceof
ministerial posts which change hands at a given point in time or over a very short period.
Western Democracies 191
TABLE IV
AVERAGE DURATION OF GOVERNMENTS IN SEVENTEEN WESTERN
DEMOCRACIES DURING THE POST-WAR PERIOD
(years per government)
Australia 7.0 Austria 2.3
Canada 5.2 Luxemburg 2.3
Sweden 5.0 Iceland 2.3
New Zealand 4.4 Netherlands 2.2
Denmark 2.1
United Kingdom 3.3
Germany 3.3 Belgium 1.2
Norway 3.3 Finland 1.0
Eire 3.0 Italy 0.9
France 0.7
difficult to form-they are not likely to remain in being for very long. Admittedly,
both these cases can be "explained" by a series of accidents. Dutch governments
were stable in the first decade of the post-war period when Mr. Beel and
Mr. Drees dominated the political scene; it may be that the present period of
instability will be followed by a return to greater governmental duration, though
the results of recent general elections do not seem to give much support for such
a conclusion. In Britain, Churchill's age, Eden's Suez policy and subsequent bad
health, and Macmillan's illness appear to account on the superficial level for
fairly frequent governmental changes; but it is interesting to note that "accidents"
of this kind should take place relatively often; their repetition does throw some
doubt on the accidental theory itself. In fact, prime ministers, like ministers, are
frequently replaced. Parliamentary regimes are in this sense more unstable than
a presidential system like that of the United States. Few countries keep govern-
ments with the same political base and the same head for as many as four years
(one US presidential term) and only the Australian, Canadian, and Swedish
governments show the same stability as that of the United States executive. As a
result, it is perhaps ironical to note that the difference between a stable and an
unstable governmental pattern remains very narrow: the borderline appears to
be somewhere between one and two years. Expectations of public opinion and
politicians may be different, and governmental efficiency may be greater on the
upper side of the line: but it remains true that parliamentary governments are
rarely very stable and that stable parliamentary regimes change their government
fairly often.
much in accordance with their own stated policies but on the basis of circum-
stances and in order to win over opponents without whose support the govern-
ment may not be able to remain in office. But it can at least be assumed that,
over a period, the relative amount of political astuteness is reasonably well
spread and the relative degree of policy implementation is broadly similar. While
such an assumption might not be valid if Western and non-Western countries
were compared, it appears acceptable, at least as a working hypothesis, among
Western democracies, where the conditions of the political game are broadly
similar.
The relative participation of parties in government can fruitfully be viewed
from two perspectives. First, it is interesting to relate the participation of parties
in government to their share of the national vote. Are certain significant groups
of electors constantly under-represented in the executive circles? Second, the
analysis of the relative participation of parties in the government makes it pos-
sible to compare the effect of two types of governmental arrangements practised
in Western democracies, the one-party government system and the coalition
system.
There are, in fact, at least two types of coalitions: the small and the grand
coalition, the latter having sometimes been known as the "Austrian solution."
The grand coalition is rarely imposed upon politicians by the circumstances of
party representation in parliament: it is more often than not the product of a
determined and conscious political philosophy as a result of which a high
premium is given to the sharing of power and a rather low value to clear-cut
decisions.
The Austrian solution (now abandoned by Austria, of course) has been
widely practised among Western democracies. In Table V, countries have been
ranked on the basis of one point per year where there was no coalition, two
points per year of small coalition, and four points per year of grand coalition.5
TABLE V
COALITIONS IN WESTERN DEMOCRACIES, 1947-66
A majority of Western countries receive at least forty points: they have therefore
at least occasionally practised the grand coalition and only Anglo-Saxon and
some Scandinavian countries have not done so.
Do coalitions enable the various political parties to be represented more fairly
than single-party government? In principle, the idea of coalition government is
5For the purpose of the allocation of points, grand coalition is defined as one in which
parties representingjointly over 75 per cent of the electorate take part.
Western Democracies 193
that of simultaneous representation of the various forces (or some of the political
forces). Single-party government has to be based on a different unwritten rule,
that of "successive" representation. But such a successive representation can be
fair representation only if a "divine hand" makes the pendulum swing regularly.
Conversely, coalition government, unless it takes the form of a permanent grand
coalition, can equally bring about distortions, though these are less likely to be
caused by accidents of electoral success than by conscious (or at least half-
conscious) patterns of behaviour of the politicians.
The operation of the divine hand can only be examined if countries which
practise single-party government are considered over a long period. It seems
equally reasonable to consider the operation of coalitions over a longish period,
as the prejudices and biases of politicians are apt to change after a decade and
certain ostracisms may tend to be less marked. An index was therefore con-
structed for Western democracies based on the twenty-year period starting in
1946. It was assumed that if serious distortions did still appear when considering
such a long time span, it could be claimed with assurance that, on the whole,
fair representation was not normally achieved.
Certain features of an index of this kind have to remain somewhat arbitrary,
as it is based on the attribution of a share of power or at least of representation
to parties in a coalition. In the formulation used in this paper, 10 points per year
are attributed to each country. When the government is wholly composed of
members of the same party, the party is allocated all 10 points: thus, as the
British Conservative party was in office alone for thirteen years of the period
1946-65, it receives 130 points, while the Labour party, in office for the other
seven, received 70 points. It therefore follows that Labour had a 35 per cent
share in the government during the whole of the period while the Conservatives
had a share of 65 per cent. When, on the other hand, the government consists
of a coalition, one basis for allocation might be the ratio of ministers of each
party to the total number of ministers (though perhaps an extra point should be
given to the party which has the prime minister). As data on all governments
could not easily be obtained, the calculations used in this paper are based not so
strictly on the percentage of ministers of each party as on a rougher division
corresponding broadly to the relative strength of the various parties in Parlia-
ment. Thus, in the Austrian case, 100 points were allocated to each of the two
parties for the whole of the period, since they were in constant partnership up
to 1966. Admittedly, the People's party should probably have been allocated a
few more points than the Socialists, since they were more often than not the
senior partners of the coalition; but the difference would be small-probably not
more than 105 or at the most 110 for the People's party. Thus it seems that,
while the British Labour party obtained, during the period, about one-third of
the total British government's share, the Austrian Socialists, who obtained
generally about the same percentage of votes cast as the British Labour party,
obtained a percentage of participation of about 50 per cent.
Calculations for most countries were much less simple than for Austria or
Britain and therefore somewhat more arbitrary: they are reproduced in Table VI
on a percentage basis (the details for the "points year" calculation for each of
the countries concerned are given in the Appendix). Even if exact percentages
of ministers had been taken into account, it would have been difficult to know
194 J. BLONDEL
TABLE VI
PARTICIPATION,1946-66
SHAREOF GOVERNMENTAL
(in percentageof total share for each country)
precisely how much each of them had participated in the government. Points
give orders of magnitude: the Dutch Liberals may perhaps have had a share of
about 8-12 per cent in government: it is unlikely that their participation was
as high as 20 or as low as 2 per cent.
The main and not altogether surprising general conclusion is that left-wing
participation has been markedly low, except in three Scandinavian countries and
in Austria. What is also clear, however, is that differences in electoral strength
appear to explain very little: the type of governmental arrangement (single-party
government or coalition) and the type of party system need to be considered at
least as carefully as the electoral strength of parties.
Grand
coalition
Practised from time Some small One-party
often to time coalitions government
Representation:
Fair Austria Belgium
Netherlands
Moderate Luxemburg Iceland Eire New Zealand
United Kingdom
Canada
Low France Denmark Australia
Germany
Verylow Italy Norway
Finland Sweden
TABLE VIII
DISCREPANCY BETWEEN VOTES AND REPRESENTATION IN GOVERNMENT IN
WESTERN DEMOCRACIES*
is one instance of a high distortion rate. In Australia, the divine hand did not
operate to ensure a swing of the pendulum; in Germany, Adenauer's efforts to
work a party government have prevented for a time (but only for a time) the
government of Bonn from moving towards a grand coalition and from joining
Belgium and Luxemburg in the direction of a low distortion rate.
This finding can be generalized by introducing the notion of balance and
particularly of balance between right and left. We noticed that in two-party
systems the two parties tended to be evenly balanced and that in two-and-a-half-
party systems there was some unbalance between the two major parties, but that
it was small. At the other extreme, multi-party systems without a dominant party
are also balanced, though the distinction between right and left is less clear in
these countries: the balance which prevails in this type of party system appears
to be that of the political oligopoly. Hence a large number of combinations;
hence also, particularly where communists are not strong, the impression that
representation can be effective, obviously on a coalition basis, but between all
parties. There is a premium given to the centre parties, around which most
coalitions tend to be formed: this premium is largest in Finland (Agrarians), but
it applies to the Dutch Catholic party, the French Radicals and MRP, and to the
Swiss Radicals.
The multi-party system with a dominant party is, on the contrary, structurally
unbalanced. The dominant party is in the position of a large firm faced with a
number of much smaller competitors: it can manipulate the situation much more
easily than any party in any other of the party systems. Indeed, one can list in
broad terms the conditions which will maximize the dominance of the large party.
First, it must "appear" to be dominant in terms both of votes (40 per cent of
the electorate seems to constitute a minimum) and of strength in relation to the
second party (about twice as many votes). The nearer the dominant party comes
to receiving 50 per cent of the votes, the more secure its position is, not only
because it comes nearer to obtaining the absolute majority by itself, but also
because the second party is likely, by the same token, to be relatively smaller.
Thus the first condition is in itself both absolute and relative. The second
condition is that the system should be a multi-party system: dominance is
exercised most adequately (as can be seen in the economic world) if there is a
multitude, and not simply a small number, of other firms. If, to take an extreme
position, the dominant party was faced with a large number of insignificant
forces, each comprising 2 or 3 per cent of the votes, its dominance would be
assured permanently. In practice permanent dominance may be obtained with a
minimum of four solidly established parties, though dominance is not ensured, as
can be seen from the example of Iceland in which the Independent party was out
of office for three years. Even five parties are not fool-proof, since (but
admittedly with great effort and after a split had taken place) the dominant
Socialist party in Norway was thrown out of office by a coalition of all the
others. But these are very exceptional occurrences. On balance, a dominant
party is fairly safe in a five-party system: for all five countries, the two cases of
Norway and Iceland were the only ones in which the dominant party was out
of power.
There is, moreover, a third and subsidiary condition. Dominant parties thrive
particularly if they can count on the potential support of satellite parties on at
Western Democracies 197
least one ideological side. The Norwegian and Icelandic dominant parties suf-
fered from not being able to generate such satellites; the Swedish, Danish, and
Italian dominant parties (to which must be added the French Gaullists-and
indeed the German CDU during the short period when it moved from its
original "Dutch" position to its present situation) have on the other hand tended
to benefit from the existence of satellite parties and, in the Italian case, to
generate them. When, as in Italy, the dominant party is relatively weak with
only the minimum electoral strength entitling it to be dominant, the condition of
satellite creation and development becomes particularly critical. But the condi-
tion exists in all cases of this type, as these "dominant" parties usually do not
reach even a bare absolute majority of seats and remain somewhat below the line.
The representation of the political parties in the government, either succes-
sively or simultaneously, is therefore dependent upon the balance of the political
parties. Unbalance creates or multiplies distortions. Yet, this type of unbalance
can remain permanent in Western democracies. The distorted representation of
Scandinavian countries, of Italy, and possibly of France has not created serious
political problems. Admittedly, in France and Italy, present changes in electoral
alignments or in the representation in the government seem to indicate that the
existing unbalance is felt as a problem which needs to be solved. But in the
three Scandinavian countries, the right appears to have accepted its traditional
fate: perhaps the recent victory of the bourgeois parties in Norway will help to
maintain the status quo, for it suggests that the dominant party may be put, at
least for a time, in opposition.
While distortions elsewhere are, on average, lower than they are in the coun-
tries with a dominant party, recent changes in coalition patterns are leading to
reduced distortions in the two countries in which they are highest, Finland and
Germany: the return of the Communist party to the Finnish government and the
establishment of the grand coalition in West Germany are all conducive to
a sizeable reduction of the distortion level. The previous conclusion about
unbalance should therefore perhaps be restated: behavioural patterns may lead
automatically to a modification of the rules about governmental formation when
the level of distortions is sizable-provided, however, that the structural condi-
tions (dominant party) or the political attitudes (party government, as in
Australia) do not limit the flexibility of the adjusting mechanisms.
Representation of the various political parties thus depends very markedly
upon the nature of the party system as well as upon the ideology relating to
party government which prevails in the country. One-party government will
ensure representation only if there is a fairly even balance, implying therefore a
two-party system and indeed a two-party system which does (automatically, by
some divine hand) produce a change of government at a regular interval:
one-party government thus depends, like the traditional capitalist system, on the
goodwill of unknown underlying forces. Multi-party government, if left to its own
devices as in Holland, Switzerland, and (though perhaps oddly) in the tradi-
tional French regime, does ensure good to moderate representation-but the con-
dition which is required is that a party should not be ostracized by the others (or
make itself ostracized by its behaviour). The two-and-a-half-party system, if
flexible about the rules of the game, is perhaps the system which ensures best (and
most meaningful) representation. It oscillates between one-party government
198 J. BLONDEL
and a grand coalition, and these oscillations give flexibility to the arrange-
ments and make it possibleto avoid permanentnon- or under-representation of
some groups.In fact, if looked at from the point of view of the left (which is
weakerin two-and-a-half-party systemsthanin two-partysystems), the two-and-
a-half-partysystem has made partyrepresentationin governmenteasier;it would
have madefor very good representation in Germanyhad not Adenauerattempted
a systematicbut unsuccessfulmove towardsa two-partysystem. The distortion
levels of systemswith dominantpartiesare serious,thoughthey appearto be, in
the context of stableScandinavia,toleratedby the electorate.The distortionsof
two-partysystemsare potentiallyserious,if the divinehand does not act to make
the pendulumswing.Two-and-a-half-party systemshave shownsome capacityto
operate flexibly and to generate some representation.Is it at the price of less
stable, and thereforeprobably less efficientgovernment?
DURATION OF GOVERNMENTS
We noted earlierthat governmentscould be dividedinto four groups according
to whetherthey lasted five years or more, three years or more, two years or
slightlymore, or about one year or even less. This durationis unquestionably
influencedby the type of party system prevailingin the country. The longest
duration can be found in two-party systems, even if the United States is
excluded,the shortestamongmulti-partysystemswithouta dominantparty.The
other two groups stand in the middle (see Table IX). There is no overlap
betweenthe first and fourthgroups: Austriawith the most short-livedgovern-
ments amongthe two-partysystemshad a slightlyhigheraveragethan Holland,
the country with the highest average among the multi-partysystem countries
withoutdominantparties.
There is little difference,on the other hand, between countrieswhich have
a two-and-a-half-party system and countrieswith a multi-partysystem of the
dominant-partyvariety: variationsare considerable,from Canada to Belgium
and from Swedento Italy. In both groups (and indeed even in the two-party
group if one compares Austria to the other countries), countries which are
normallygovernedby a coalitionare more likely to have unstablegovernments
thancountrieswhichare governedby one party.This resultis, of course,partlya
functionof the index used to measureministerialcrises: what is a reshufflein
Britain (e.g., sacking ministers as in 1962) becomes a ministerialcrisis in
TABLE IX
DURATION OF GOVERNMENTS BY GROUP OF PARTY SYSTEMS*
1946' 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
United Rep. 10 10 10 10 10 10 10 10
States Dhem. 10 10 10 10 10 10 10 10
Canada Cons. 10 10 10 10
Lib. 10 10 10 10 10 10 10 10 10 10 10 10
United Cons. 10 10 10 10 10 10 10 10 10 10
Kingdom Lab. 10 10 10 10 10 10
New Zealand Cons. 10 10 10 10 10 10 10 10 10 10
Lab. 10 10 10 10 10 10
Australia Cons. 10 10 10 10 10 10 10 10 10 10 10 10
Lab. 10 10 10 10
10 10 10 10 10 10 10 10 10 10
F. Gael 6 6 6 6 6 6
Lab. 3 3 3 3 3 3
Rep. .5 .5 .5 .5 .5 .5
Farmi. .5 .5 .5 .5 .5 .5
France Comm. 3
Soc. 4 2.5 2.5 2. 5 2.5 2. 5 2.5 2.5
MRP 3 2.5 2.5 2.5 4 2.5 3 3 2.5 2.5 2.5 2.5 2.5 1 1 1
Rad. 2.5 2.5 2.5 3 2.5 3 3 2.5 2.5 2.5 2.5 2.5 1 1 1
Cons. 2.5 2.5 2.5 3 2.5 4 4 2.5 2.5 2.5 2.5 2.5 1 1 1
Gaull. 2.5 2.5 7 7 7
- ------
APPENDIX-continued
1946 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
Netherlands Lab. 5 5 5 4 4 4 4 4 4 4 4 4 4
Lib. 1.5 1.5 1.5 1.5 2 2 2
Cath. 5 5 5 3 3 3 3 3 3 3 3 3 3 5 5 5
Chu. 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
AR 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
Luxemburg Comm.
Soc. 3 5 5 5 5 5 5 5 5
Lib. 2 3 3 3 3 3 3 3
Cath. 4 7 7 7 7 5 5 5 5 5 5 5 5 7 7 7
APPENDIX-concluded
1946 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
Norway Lab. 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
Cons.
Christ.
Agr.
Lib.
Denmark Agr. 2.5 2.5 2.5 2.5 5 5 5
Soc. 7.5 7.5 7.5 7.5 10 10 10 10 7 7 7 7 7
Cons. 5 5 5
Rad. 3 3 3 3 3
Sweden Soc. 10 10 10 10 10 10 7 7 7 7 7 7 10 10 10 10
Centre 3 3 3 3 3 3
Finland Comm. 3 3
Soc. 3 3 10 10 3 3 3 3 3 3
Agr. 4 4 7 5 5 8 3 3 3 7 3 10 10 10
Lib. 2 1 1 2 2 2 1.5
Cons. 3 1.5
Swed. P. 1 1 1 2 1 2 2 1 1
Iceland Comm. 3 3 3
Soc. 3 4 4 4 3 3 10 4 4 4
Progr. 4 4 4 4 4 4 4
Ind. 4 6 6 6 10 6 6 6 6 6 6 6 6
Austria Soc. 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
OVP 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5