Jump to ContentJump to Main Navigation
Parties and DemocracyCoalition Formation and Government Functioning in Twenty States$

Ian Budge and Hans Keman

Print publication date: 1993

Print ISBN-13: 9780198279259

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198279256.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017

(p.208) Appendix A: Party Factions and Cabinet Reshuffles

(p.208) Appendix A: Party Factions and Cabinet Reshuffles

Source:
Parties and Democracy
Publisher:
Oxford University Press

This appendix reports a check carried out on Implications 4(i) and 4(ii) of Table 2.4, relating reshuffles and turnover of ministers to the Prime Minister's freedom from constraints, and therefore to the existence of single‐party governments as compared to coalitions.

Although these hypotheses follow from the main and auxiliary assumptions of Tables 2.1 and 2.2, just like other aspects of our reasoning, it proved difficult to combine this limited analysis, which is all we can undertake here, with the more extended analysis undertaken in the chapters. We therefore report it in this free‐standing appendix—which is not to say that positive results will not uphold the General Theory just as much (and negative findings damage it) as ones reported earlier. The analysis was in fact done contemporaneously with the last four chapters of the book.

To recap the theoretical argument briefly, parties are assumed to act in unity when there is a premium on doing so and external pressures are great—as in coalition formation and indeed during the lifetime of coalitions. Where both the premium on unity and external pressures are less, parties become more inward‐looking, and policy‐pursuing factions have greater scope. The Prime Minister, as a member of a faction, will use the powers given by the constitutional conventions to dismiss or transfer representatives of other factions in order to benefit his or her own faction. Policy‐pursuing motives can thus prompt the movement of individuals between offices (reshuffles). Such movement will be greater where the Prime Minister is less constrained by other constitutional powers, such as an executive or semi‐executive Head of State (found only in France and Finland of the countries we are dealing with, however), and above all by coalition partners. Hence the major contrast we shall be examining in the succeeding analysis is that between coalition and single‐party governments.

A published report has in fact already presented results from an earlier analysis of some of our data, for the period 1946–78.1 We shall include some findings from this along with our updated ones for 1946–84 below. The article defined ‘reshuffle’ as we shall use the term here: that is, as ‘the simultaneous movement or replacement of two or more Cabinet Ministers’. Such a definition deliberately avoids identifying the resignation and replacement of only one individual as a reshuffle, since very probably this would not have been precipitated by ‘political’ motives but by ill health or personal reasons. Certainly the transfer of two or more persons might also have been initiated by (p.209) these factors, but it constitutes more than a simple substitution and at least represents a decision by the Prime Minister to make more changes in the Cabinet than necessitated by a single appointment from outside.

There is another operational definition involved in the analysis. Governments as such consist of very diverse ministries, many of which have a very limited existence, and some of which are represented within Cabinets while others are not. To permit analysis over time, therefore (and also to allow for checks of the other hypotheses in the central theory), what has been examined is changes in the tenure of the central and most important ministries—those already used in the analyses of Chapter 4. In most governments these indeed

Table A.1. Incidence of Government Reshuffles in Eighteen Democracies, According to Type of Government, 1946–1984

Coalition governments

Single‐party governments

All governments

No. of reshuffles

No.

No. of reshuffles

No.

No. of reshuffles

No.

Australiaa

11

22

11

22

Austria

3

18

3

4

6

22

Belgium

4

23

4

4

27

Canada

24

15

24

15

Denmark

7

11

4

12

11

23

Finland

2

29

2

9

4

38

France 4

2

28

2

28

5

6b

21

6

21

West Germany

4

14

4

14

Ireland

1

4

8

10

9

14

Israel

2

29

2

29

Italy

0

33

14

47

Japan

4

5

24

24

28

29

Luxemburg

13

13

Netherlands

1

17

1

17

New Zealand

3

16

3

16

Norway

1

4

9

15

10

19

Sweden

3

6

7

11

10

17

United Kingdom

22

15

22

15

TOTAL

36

233

120

160

156

393

(a) The permanent and governmental alliance between the Australian Liberals and the Country party makes them a single party for the purposes of the analysis.

(b) At times the government majority in France 5 has seemed more like a single party, but it has never actually cohered in the end.

(p.210) represent the whole set of Cabinet Ministries, and in any case any important reshuffle would involve changes affecting them. So although the need to select only certain ministries in some countries is a limitation on the analysis, it should not seriously affect results.

In Table A.1 we present data for each country taken individually, distinguishing in each case between coalition and single‐party governments. Some countries, such as the United Kingdom and most Commonwealth democracies, have only single‐party governments, of course, while the classic multi‐party systems have only coalitions. A preliminary glance at these two extremes shows that the Commonwealth as a group has many reshuffles while in multi‐party systems these are very limited. In the intermediate cases where both coalitions and single‐party governments appear, the number of reshuffles is greater in the single‐party governments, with the exception of Denmark and Belgium (Italy never has reshuffles in any case). The high number of single‐party but decisively minority governments in Denmark perhaps affects behaviour there: parties, particularly the (relatively) predominant Social Democratic Party, seem to act in the same way regardless of what their present status as a government is. In Belgium single‐party government is exceptional and confined to the first half of the 1950s. The average duration of Italian governments is low; this, as we shall see, is an important influence on reshuffles, since for obvious reasons they are unlikely to occur in new governments. New Zealand is also exceptional, in that strictly single‐party government is associated with very few reshuffles. A special factor may be found in the small recruitment pool and great technical difficulty of finding replacements from outside government if a minister resigns rather than accepts relocation. More could be made of the contrast between the Fourth and Fifth Republics in France if we took the government majority in the latter case as essentially a single (highly factionalized) party. But the balance of considerations seems to point to governments as being essentially coalitions, though more stable and durable than those going before. Reshuffles are obviously limited under the Fifth Republic by another factor—the institutional position of the President as a separate and more powerful actor than the Prime Minister. Government reshuffles thus take the form of a change of Prime Minister along with the other ministers. Several outright reshuffles of this kind, which appear in our comparative definition as the termination of a government, have in fact taken place (involving, for instance, Pompidou in 1968, Chirac in 1976). The same holds true for Finland, where the President may often intervene to terminate a government as a preliminary to encouraging another more to his liking.

Japan is an interesting case where the frequency of reshuffles is very great. This seems clearly related to factional efforts to establish dominance in government. Japan comes closest, perhaps, to Italy in the way in which factions have become institutionalized and overtly recognized within the predominant party. In Italy, however, owing to the weaker position of (p.211) the Christian Democrats compared with the Japanese Liberal Democrats, internal factional conflict cannot be confined to the party itself. Thus it provokes the downfall of governments and often the Prime Minister, rather than an internal change.

Looking at the table as a whole, the contrasts seem to uphold the general line of argument advanced above, particularly in regard to the constraints on Prime Ministerial power imposed by coalition arrangements. Evidence on institutional constraints (relationships with other ministries) is limited to the French Fifth Republic and Finland, where different factors obviously operate. The influence of constraints imposed by other factions is limited to the observation that possession of the Premiership in a single‐party government, whatever the power of other factions, seems enough to enable the Prime Minister to initiate reshuffles (on this, Japan and Ireland can be contrasted with Italy and Israel). The obvious exception here is New Zealand where, whatever the reasons, the Prime Minister does not reshuffle his Cabinet.

One possible confounding factor is the effect which the duration of governments has on reshuffles. For the most part, a government has to get beyond the initial stage of organization before rearrangements can be thought of. Thus the length of time governments exist will affect the potentiality for reshuffles. This is doubly confounding because coalition governments and minority governments are generally shorter‐lived than single‐party majority governments. Is the apparent contrast between these types, attributed here to greater and lesser constraints on the Prime Minister, in fact due to a simple time factor?

A crucial comparison here is between coalition and single‐party governments within each country, where differences in duration are not so marked. The contrast does persist in six of the nine possible comparisons of this kind, so the hypothesized finding is still found even under the control for national factors.

We can control in another way, more directly, for the effects of duration as against those of single‐party versus coalition government, by incorporating both factors explicitly as variables in a multiple regression equation. Table A.2 presents several such equations, numbers 1 and 2 being based on countries as the basic case, using number of governments as an indicator of their greater or lesser duration (the more governments, the less long they last on average), and comparing the effects of this with those of the number of single‐party governments on number of reshuffles.

Equation 1 is already reported for a different time period (1946–78) and with India and Sri Lanka included as well as the countries shown in Table A.1.2 It produces a resounding confirmation for our hypotheses, as the effect of number of single‐party governments is still strong even under the control for numbers of governments—that is, for duration (which, as expected also, has a smallish negative effect on the number of reshuffles). The results (equation 2) for the countries and data given in Table A.1 are less resounding, (p.212)

Table A.2. The Relationship Between Single‐Party Government, Government Duration, and Number of Reshuffles, According to Various Multiple‐Regression Equations

Equation

Constant

Single‐party government

Duration

b 1

r

b 2

r

1

23 democracies (1946–78)

4.118

0.673

0.688

−0.117

−0.207

2

19 democracies (1946–84)

8.423

0.305

0.350

−0.152

−0.185

3

74 governments (1978–84)

0.475

0.522

0.532

0.030

0.401

Note: Number of reshuffles = Constant + b 1 Single government + b 2 Duration. Duration is measured by number of governments over the relevant period in each country (in Equations 1 and 2) and by number of months each government lasted (in Equation 3).

but support these observed before, with a moderately strong effect for number of single‐party governments and a small negative one for number of governments.

Using countries as cases when we want to examine government‐based phenomena is a slightly indirect way of proceeding. We can also look at the relationship between single‐party control and reshuffles using governments as cases, the duration of each directly measured in months, and single‐party government versus coalitions as a dichotomized variable scored 0 for coalitions and 1 for single‐party government. Unfortunately our project ran out of money and time before we could go very far with the recoding involved, so we have data with governments as cases only from 1978 to 1984. These six years suffice to give another test to our hypothesis, however, the results of which are again in the expected directions (except that duration appears as exerting insignificant rather than negative effects).

The advantage of looking at the relationship in several different ways is to establish that it holds under a range of different circumstances, which it clearly does. How can we interpret the findings? It is certainly true that, in itself, the ability of Prime Ministers to demote or move ministers around to suit their own purposes might equally be interpreted as office‐seeking (or office‐preserving) rather than policy‐pursuing behaviour. This is one of several cases where office‐seeking and policy‐pursuing assumptions give rise to similar predictions. The latter are better supported in the other tests made previously, so taken in context the result of the analysis made here is to validate further the reasoning of the General Theory set out in Chapter 2.

Notes:

(1.) I. Budge, ‘Party Factions and Government Reshuffles’, European Journal of Political Research, 13 (1985).

(2.) Ibid.