Вы находитесь на странице: 1из 25

Société québécoise de science politique

Party Systems and Patterns of Government in Western Democracies


Author(s): J. Blondel
Source: Canadian Journal of Political Science / Revue canadienne de science politique, Vol. 1,
No. 2 (Jun., 1968), pp. 180-203
Published by: Canadian Political Science Association and the Société québécoise de science
politique
Stable URL: http://www.jstor.org/stable/3231605
Accessed: 01/02/2009 08:42

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

J. BLONDEL University of Essex

The study of patterns of government in Western democracies has been greatly


improved by the work devoted recently to the smaller democracies. Although it
may be premature at this stage to expect to obtain firm conclusions and although
one should probably only try to draw a sketch of the main lines of a comparative
analysis, it is at least possible to go beyond some of the classic typologies which
tended to contrast, for instance, the British to the American system of govern-
ment or the Anglo-Saxon to the French.'
An analysis of patterns of government in Western democracies must begin by
an attempt to circumscribe and operationalize the characteristic features of the
"population" under consideration. This population has many common institu-
tional and behavioural traits; its socio-economic structures are also very similar,
at least by comparison with the rest of the world. We are therefore able to control
for certain variations and, consequently, to come to more definite conclusions
about the influence of others. But at least three types of difficulties remain.
First, the characteristic features of the political systems of these countries are
almost certainly among the most awkward to summarize or grasp because of
their complexity. Second, the most advanced method which we have at our
disposal in comparative government, structural-functionalism, is only of limited

*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

11 existe quatre types de systemes de partis dans les democraties occidentales, le


systeme de deux partis, le systeme de deux partis et demi, le systeme multi-partiavec
parti dominant, et le systeme multi-partisans parti dominant. Presque toutes les
democratiesoccidentales ont conserve le meme type de parti tout au long de l'apres-
guerre. Chaque type a ses caracteristiquespropres qui constituent autant de meca-
nismes automatiquesd'equilibre.Les types de structures gouvernementalespeuvent
etre definiespar l'intermediairedes forces politiques qui se trouvent representeesau
gouvernement, soit conjointement, soit successivement; ils peuvent egalement etre
definis en termes de stabilite. Au cours de l'apres-guerre,le systeme multi-partiavec
parti dominantsemble avoir donne naissance a la forme de representationla moins
equitable;la stabilite gouvernementaleparait dependre de la presence d'un seul, et
non de deux partis importants.

usefulness when we try to apply it to the detailed analysis of patterns of govern-


ment. Functions are broad and consequently somewhat vague, whether on the
input or the output side. We shall therefore consider here certain concrete
variables and remain close to the complex realities of political structures and
processes in Western democracies, although these are of less general application
than the variables used in structural-functionalism. Third, no over-all yardstick
with which to compare the Western countries has been clearly defined. Cross-
national analysis has typically been developed around the ideas of stability and
development. By concentrating on nineteen developed countries which are also
essentially stable, although a few exceptions can be found, we are precluding
ourselves from using either of the accepted thermometers. Admittedly, if our
instruments were more refined, we could consider stability and/or development
as continuums and we might be able to measure, for instance, the relative stability
of Iceland and Belgium. But, except in a few cases, such as France, stability
cannot be measured directly in the context of Western democracies.
We shall therefore concentrate in this paper on manifestations of political life
which appear to be characteristic of Western democracies and which, surprisingly
enough, have not received up to now the attention they might have justified.
Among these manifestations, the two most important are probably the types of
party systems and the types of governmental structures. Their examination in
itself should reveal the existence (or absence) of patterns; these in turn should
indicate whether certain forms of governmental or political behaviour are more
likely or less likely to occur than others. The consideration of the relationship
between party systems and governmental structures should also reveal whether
certain regularities can be expected or whether the distribution of each of the
variables is random in relation to the others. While such an analysis is clearly
not of direct relevance to a general study of comparative government extending
beyond Western democracies, there may be some indirect pay-off to the extent
that it might be possible to assess the probability of certain types of developments
in countries whose political system is similar to, though none the less distinct
from, that of Western democracies.
182 J. BLONDEL

Party systems and patterns of government:the selection of variables for analysis of


relationship
Western party systems have attracted attention for several decades, but in only
one respect, that of the possible influence of electoral systems on the number of
parties, has there been any systematic examination of a possible relationship.2
Little consideration has been given to the possible effect of party systems on the
type and characteristics of the governmental structure and of governmental
decision-making. It is generally assumed that two-party systems are concomitant
with relatively stable governments: yet even this proposition has not been directly
confronted with the available evidence. No other proposition of this or any
similar type has been subjected to an empirical analysis, nor have hypotheses even
been formulated.
Endeavours to generalize in this field are admittedly hampered by a number
of difficulties. One of these is structural: even when the smaller democracies are
included the number of countries is relatively small and the time span short. To
the fifteen democracies of northern and western Europe3 must be added four
Anglo-Saxon countries outside Europe, namely the United States, Canada,
Australia, and New Zealand. The homogeneity of the universe diminishes if one
moves outside this group of countries.
A second type of difficulty arises from the unavailability of certain types of
data directly relevant to the analysis of governmental patterns. Party systems are
generally well-known: the size of parties and their general ideology or broad
policy standpoints are usually easy to obtain. But much less is available about
governmental systems. We have practically no information, at least for most
countries, about any but the most formal arrangements. We do not know
whether governmental structures are really collective or oligarchical or monarchi-
cal. We have very little information on patterns of internal decision-making.
Nor is the "effectiveness" of the government easy to measure on the basis of
available data. Clearly the best test of governmental achievement-that of the
amount of "important" legislation passed by the legislature, with or without
amendment-would suppose a detailed knowledge of patterns of legislative
behaviour in relation to outputs, which are, at this stage, unfortunately only
available for a limited number of countries and, in many of these cases, only for
short periods. One has therefore to rely, as we shall do in this paper, on much
less satisfactory indicators, such as the stability of governments, a criterion which
cannot in any case be applied to the two countries in the group, the United States
and Switzerland, which do not have a parliamentary system.
Problems of operationalization have been the third major type of difficulty
encountered in comparative government. The notion of party system has
remained somewhat loose and the notion of pattern of government even more
2The literature on the relationship between electoral systems and party systems is vast, the
main works being those of F. A. Hermens, Democracy or Anarchy? (Notre Dame, 1941),
Duverger, Political Parties, and L'influence des systemes electoraux sur la vie politique
(Paris, 1950) and E. Lakeman & J. D. Lambert, Voting in Democracies (London, 1955).
See also C. Leys, "Models, Theories and the Theory of Political Parties," Political Studies,
VII, no. 2.
:'The United Kingdom, Ireland, France, Germany, Italy, Belgium, the Netherlands, Luxem-
burg, Iceland, Norway, Sweden, Finland, Denmark, Switzerland,and Austria.
Western Democracies 183
nebulous. Yet it is possible to acquire at least some information in relation to two
specific governmental characteristics, stability and extent of coalition-building.
Neither of these characteristics will reveal more than part of the aspects of
governmental life which are of importance. But, as long as other indicators are
not available, the study of the stability of governments and of coalition formation
and duration will give some information on various aspects of patterns of
government. In particular, it is usually assumed that stability leads to efficiency
and ensures support for the system, while governmental instability will foster
alienation and will be characterized by inefficiency. The study of types and
extent of coalition will provide information on the breadth of the support for
government, on its representativeness in party terms, at any given point in time.
Thus, although more work is required to ascertain whether governmental stability
or coalition government are indeed closely related to support for the regime,
executive efficiency, or representativeness of government, it seems reasonable to
assume that there is a relationship and it becomes significant to consider in some
detail the patterns which emerge from the analysis of the indicators which we
have at present at our disposal.
We propose to examine in this paper both whether concepts of party systems
and of governmental patterns can thus be operationalized and, if so, whether
certain tentative conclusions can be drawn about the relationship between the
variables. We shall thus begin by attempting to see whether the nineteen coun-
tries of western Europe and of the Anglo-Saxon world can be grouped, with
reference to their party system, in a small number of categories in which there
are common positive characteristics.
If patterns of party systems exist, it will also be possible to examine to what
extent these appear to be related with certain forms of government. Is it clear
that coalition government or unstable government or less party-representative
government is more likely to occur in countries which have certain types of party
systems? Are coalition governments less likely to be stable, while more likely to
be representative of the shades of opinion in the country? Obviously, our
knowledge of patterns of government will be increased when we can introduce
variables relating to prime-ministerial government or to the power of the govern-
ment in connection with legislation. But some of the variables which can be
operationalized may already give some idea of the possible influence of party
systems on patterns of government.

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

be found in the type of image than in the structural characteristics of the


organization.
1. Number of parties in the system
Two-party systems are rarely defined operationally. Yet, among Western
democracies, four groups are clearly discernible (see Table I). Five countries
give on average more than 90 per cent of their votes to the two major parties.
TABLE I
AVERAGE TWO-PARTY VOTE, 1945-66
(percentage)
United States 99 Denmark 66
New Zealand 95 Sweden 66
Australia 93 Norway 64
United Kingdom 92 Italy 64
Austria 89 Iceland 62
Holland 62
West Germany 80
Luxemburg 80 Switzerland 50
Canada 79 France 50
Belgium 78 Finland 49
Eire 75

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)

Two-partysystems Two-and-a-halfparty systems

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.

3. Ideological spectrum and party strength


Western democracies can therefore be divided into four groups, if both numbers
and strengths are taken into account: five are in the two-party system group, five
Western Democracies 187

in the two-and-a-half-party system group, five in the group of multi-party systems


with dominant party, and four in the group of multi-party systems without
dominant party. But the party system can only wholly be defined if we take into
account the position of parties on the ideological spectrum, particularly when
the system does not have a "symmetrical"character.
Even in Western democracies, parties are difficult to categorize accurately.
The United States and Eire always had parties which do not fit in any easily
recognizable typology. The French Gaullists are not ordinary Conservatives,
though they may have to be lumped with Conservatives for comparative pur-
poses. The other multi-party system countries all have parties of a somewhat
peculiar type, such as the two Conservative parties in Holland, the Swedish party
in Finland, and the party of the Peasants, Artisans and Bourgeois in Switzerland.
In all these cases, certain rough approximations have had to be made.
Table III shows the panorama of party strengths within the ideological
spectrum. Group 1, except for the United States, is fairly homogeneous: the four
other countries in the group have a large Socialist party and a large Conservative

TABLE III
THE IDEOLOGICAL SPECTRUM OF PARTIES IN WESTERN DEMOCRACIES

Comm. Soc. Lib./Rad. Agr. Christ. Cons.

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.

4. The dynamics of party systems in the post-war period


Party systems are also very stable. Germany is the one clear case of shift from
one group to another during the post-war period. In 1949 Germany would have
had to be classified in the fourth group, its party system resembling in many
ways that of Holland; by 1957, the German party system was similar to that of
Belgium or Luxemburg. It is perhaps not altogether surprising that a country
which was so fundamentally reconstructed politically (and whose territory
strikingly diminished after the war) should have experienced such a transforma-
tion in its party system. All the other cases are doubtful. France, as we noted, is
apparently moving now from the fourth to the third group, but it seems difficult
to predict that the change will be permanent. In two other cases, Belgium and
Austria, future moves are possible, though they are still far from certain: while
Austria may move towards a two-and-a-half-party system if the third party is
more seriously taken into account as a result of the break-up of the Christian-
Socialist coalition, the electoral basis for the change is still not apparent. Belgium
may become a multi-party system if the growth of the Liberals encourages splits
among the major parties, but a more likely development still seems to be the
return to a normal two-and-a-half-party system. Party systems in Western
democracies therefore remain in existence for very long periods: they can be
categorized and grouped with a considerable degree of accuracy. That this should
be the case may be a sign of over-all political stability and a consequence of the
developed character of these political systems: cross-national studies extending
beyond Western democracies might indeed show that party systems elsewhere are
neither as "clear" nor as stable. There seems therefore a strong argument for
190 J. BLONDEL

examiningwhetherthese patternsof party systems are reflectedin the govern-


mentalstructuresof the countriesconcerned.

GOVERNMENTAL PATTERNS: STABILITY AND STRUCTURE OF GOVERNMENT

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.

2. The political structure of governments


The duration of governments is an indicator of the stability of the system; it can
also be taken to constitute an indicator of the potential efficiency of governments.
By studying the political complexion of governments over a relatively long period,
one can measure, on the other hand, the representativeness of the executive. If
different political groups have a share in the government over a given period,
the system would appear to be representative. It would also seem reasonable to
assume that, the more a party shares in government, the more it is able to carry
out its policies. Admittedly, parties may be frustrated by the action of others
which act consistently more cleverly; some parties may also tend to act not so
192 J. BLONDEL

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

Austria 78 points Germany 42 points


Switzerland 70 Denmark 30
Netherlands 66 Eire 26
Luxemburg 62 Sweden 26
Iceland 54 Norway 22
France 52 United States 20
Belgium 49 Canada 20
Finland 47 United Kingdom 20
Italy 46 Australia 20
New Zealand 20
_!

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)

Comm. Soc. Rad./Lib. Agr. Christ. Cons.

United States Rep. 40


Dem. 60
Canada 70 30
United Kingdom 35 65
Eire 9 1.5 F. Fail 70
F. Gael 18
Rep. 1.5
France 1.5 11 18 14.5 21.5 Gaullists28.5
Germany 3 17 68 9+2
Italy 1 4+6.5 9.5 74.5 4.5
Belgium 1.5 31 13.5 54
Netherlands 29.5 9 39.5 12+10
Luxemburg .5 24 16 59.5
Norway 95 1.2 1.2 1.2 1.4
Denmark 69.5 10.5 12.5 7.5
Sweden 91 9
Finland 3 22 10 49 6+10
Iceland 4.5 29.5 12 49
Austria 50 50
Australia 20 80
New Zealand 30 70

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.

Patterns of government, institutions and party systems


THE PARTICIPATION OF POLITICAL PARTIES IN GOVERNMENT
We saw that governments in Western democracies have experimented with three
main types of arrangements: one-party, small coalition, grand coalition. But the
three types of governmental arrangements are not equally embedded in the
political culture of a particular country. While party government is an unwritten
rule of government in Anglo-Saxon countries, there is less stability and "agree-
ment" about patterns of governmental arrangements on the Continent.
By using the ranking and the "points year" system, however, one can examine
whether the governments of countries which have tended to practise the coalition
idea have in fact represented the various forces in the nation better than the
governments of countries which have normally been constituted by single parties.
Table VII shows that the countries which have practised the grand coalition
Western Democracies 195
tend most regularly to have the best party representation, that one-part govern-
ments tend to have ensured a moderate level of representation, and that the
greatest distortions are to be found among those countries which have practised
the small coalition idea normally or only intermittently. It can therefore be
stated, at least in the first instance, that grand coalitions are more likely to ensure
fair representation of parties in government that one-party government, but that
one-party government ensures normally a fairer representation than the small or
intermittent coalition.
TABLE VII
COALITION GOVERNMENT AND THE REPRESENTATION OF PARTIES IN THE GOVERNMENT

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

This finding needs to be qualified, however, by relating the degree of repre-


sentation to the type of party system. Distortions in representation vary con-
siderably within each type of party system and, in three cases out of four, the
average distortion is about the same. But the average distortion is much higher
in the fourth type of party system, that of the multi-party system with a dominant
party (see Table VIII). Thus two-party systems are not more likely to ensure
representation than two-and-a-half- or multi-party systems.
Perhaps because twp-and-a-half-party systems are more likely to provide a
ground for grand coalitions than two-party systems, these systems tend to ensure
as good a representation as two-party systems. In both groups of countries, there

TABLE VIII
DISCREPANCY BETWEEN VOTES AND REPRESENTATION IN GOVERNMENT IN
WESTERN DEMOCRACIES*

United Kingdom 39 Denmark 53


Australia 59 Norway 83
New Zealand 38 Sweden 88 66.7
Austria 22 Iceland 43
Italy 47
Canada 44
Eire 38 Netherlands 13
Germany 61 38.4 France 49 43.5
Belgium 15 Finland 68
Luxemburg 34
*The distortion in representationin governmentis calculated on the
basis of the differencebetween party support in the country and party
representationin government
196 J. BLONDEL

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*

5 years About About About


or more 3 years 2 years 1 year

Two-partysystems Australia United Austria


New Zealand Kingdom
Two-and-a-half party Canada Germany Luxemburg Belgium
systems Eire
Multi-partysystemswith Sweden Norway Denmark Italy
dominantparty Iceland
Multi-party systems without Netherlands Finland
dominantparty France
The name of countrieswhichnormallypractisethe coalitionsystemare in italics.
Western Democracies 199
Austria, though in the end (at least up to 1966) the same parties became
partnersagainin the coalition.But, in psychologicalterms, coalitionbreak-ups,
even if followed by a rejoining,appear to be ministerialcrises; they usually
mean that the machineryof the state comes to a halt for a period, sometimes
lasting several weeks; they involve the machineryof the state to resolve the
conflict (in particularthe action of the head of state or a formateurappointed
by him). They do not take place in the privacyof the prime minister'smind or
among a few colleagues. They are public events and, as such, contributeto
mould attitudesaboutthe politicalsystem and to have an effecton the efficiency
of government.
Two factorsthus appearto contributeto the relativedurationof governments
in parliamentary systems:the fact that the systemis a straightclash betweentwo
parties and the fact that one partygovernmentprevails.The two factorsare quite
clearly connected: one-partygovernmentsare not possible without at least one
large party,except on a temporarybasis. But, as we noted, the ideology of one-
party governmentis accepted in some countries, particularlyin Anglo-Saxon
countries,while the notion of coalitioncan prevail even where it is not strictly
necessary,as in Austria or duringat least one period of Adenauer'sGermany.
On balance,althoughthe link betweenthe two factorsleads to some uncertainty,
it does seem that one-partygovernment,rather than the two-partysystem, is
the factor contributingmost decisivelyto the stabilityof government.None of
the countrieswhich have had party governmentduringmost of the period has
a low recordof governmentalstability.
Conversely,no countryin the coalition group had governmentswhich lasted
more thanaboutthreeyears.Germanyis indeedthe only coalitioncountrywhich
belongsto this groupand its place there appearsdirectlyconsequentialupon the
enterpriseof Adenauer:eventssince 1963 show that post-Adenauergovernments
may not last for even as long as three years. While practicallyall the one-party
governmentsare to be found in the top two categories (more than five years
and aboutthree years), these categoriesmay soon no longerinclude any of the
countrieswhichare ruledby coalitions.
The durationof governmentscan thereforebe said to be essentiallydependent
on the presencein power of a "team,"and, almost certainly,on the presence
of a team leaderwho does not merelyarbitrate,or spendmuch time in arbitrat-
ing, divergencesbetweencoalitionpartners.The existenceof a two-partysystem,
in the strict sense in which we definedit here, does apparentlyhelp in that a
two-partysystemincreasesthe probabilitythat the ideologyof one-partygovern-
ment will prevailin the political culture.It also constitutesa sufficientthough
not a necessaryconditionfor the presenceof a largepartywhichcan take alone-
or at least is in a position to claim that it can take alone-all the posts in the
government.
But where (and when) the two-and-a-half-party system, or even the multi-
party system with a dominantparty, is so configuredto allow one party to
take power alone at least fairly frequently,governmentalstability is-or may
be-equally assured.Coalition,whethersmall or large, appearsdirectlyantago-
nistic to stable government,though differencescan also be large, from Austria
to Franceand fromthe Netherlandsto Finland.The way to ensureboth stability
and coalitionthereforeseems to be the Swiss way. When, as in Switzerlandbut
200 J. BLONDEL

nowhereelse, the coalitionsystem is institutionalizedand when, as a result,the


parties have little leeway to try differenttypes of combinationsrangingfrom
grand to small coalitions and includingvarious types of small coalitions, the
governmentalcoalitioncan remainstable.Perhapsthe Swissmodelis inapplicable
to largecountries;indeed the alternatepresenceand absenceof Socialistsin the
SwissFederalCouncilshowsthat, even wherecoalitionis institutionalized,some
instabilitydoes take place. But the stability of governmentwhich one-party
governmentcan give could not, it seems, be achievedas long as the rules of the
game remain uncertain in relation to coalitions, their size, and their type.
Whetherthe efficiencyof governmentis directlyaffectedneeds to be considered
in the light of other indicatorsas well, such as, for instance,legislativeanalysis.
But to the extentthatthe durationof governmentdoes affect,as it clearlyappears
to do, the efficiencyof its work, the coalition system and the absence of one
(but not necessarilymore than one) largeparty appearinstrumentalin increas-
ing inefficiency.
Stabilityandfair representationof partiesin governmentare difficultto achieve
jointlyin Westernsystemsof government.Clearly,if the divine hand is at work,
the two-party system can provide, over time, successive representationand
reasonableefficiency.But severalexamplesindicatethat the divine hand can be
at rest. In fact, only one of the five countrieswhich belong to the group, the
United States,appearsable to achievea relativelyhigh level of successiverepre-
sentation;and this conclusion can be drawn only provided little attentionis
paid to the relativeweight of the subgroupswhich divide Americanpartiesin
Congress: if these were to be taken into account, the over-representation of
the right might appear as large in America as in Britain, if not perhapsas in
Australia.Coalitiongovernmenthas providedmeans of achievingbetter repre-
sentation of the various political forces in some Continentalcountries and,
possibly as a result, obliged these forces to compromiseon their policies. But
this is clearly at the expense of possibilitiesof long-termaction: deadlockin
governmentand lengthygovernmentalcrises are the price paid for such "true"
representation.
Yet, as was noted earlier,the four types of party systems,with theirrelative
efficiencyand representativeness,have remainedsurprisinglystable. Even the
system of the dominantparty,thoughclearly unfairto those who do not move
into the government,showsvery little sign of being seriouslychallenged.Govern-
mental patterns are clearly shaped. They provide solutions not only to the
specificproblemsof the countriesconcerned (whether the presence of a large
Communistparty, or the ideology of one-party government,or the idea of
equilibriumof political forces) but to the general problemsof efficiency(and
apparent efficiency) and representation(including apparent representation).
No Western governmentappears to have solved both except by luck; and
reflectionson how to create or maintainthe luck have been generallyabsent.
Perhaps,however,behind the apparentlack of representationin governmentor
the apparentlack of efficiencyresultingfrom instability,other forces, through
parliamentsandthe interestgroups,succeedin promotingpolicies whichcounter-
balance, to a large extent at least, some of the most apparentdefects of the
systems. Studies of governmentalpatternsshould consider, throughsystematic
studiesof outputs,this efficiencyand this representativeness.
Appendfix
THE INDEX OF PARTY REPRESENTATION IN GOVIERNMENT

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

Germany CDU (Christ.) 7 7 7 7 6 6 6 6 8 8 8 8 7


FDP (Lib.) 2 2 2 2 2 2 2 2 3
DP (German P.) 1 1 1 1 1 1 1 1 2 2 2 2
GDB (Germ. P.) 1 1 1 1
SPD (Soc.)

Italy PCI (Comm.) 2


PSI (Soc.) 2
PSDI (Soc. Dem.) 1 1 1 1 1 1 1 1
PRI (Rep.) 1 1 1 1 1 2 2 1 1 1 2 1
DLI (Lib.) 1 1 1 1 1 1 1 1
DC (Christ.) 5 7 7 7 7 8 8 10 7 7 7 10 8 10 10 7

Belgium PCB (Comm.) 3


PSB (Soc.) 4 5 5 7 7 7 7
PLP (Lib.) 33 3 3 3 3 3 3 3
PSC (Christ.) 5 5 7 10 10 10 10 10 7 7 7

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

Вам также может понравиться