Resumé

The aim of this article is to analyse the political leaders' efforts to organize and manage relations between relevant party actors in a way that is suitable for the operation and preservation of coalition governments. Five coalition governments serve as illustrative cases showing how these relations have been managed in post-war Norway. The similarities between the different government coalitions are obvious. There are, however, interesting variations concerning the priority given to coordination and unity versus party differences and profilation.

Introduction

For a long time political scientists have been preoccupied with the study of coalition governments' formation. To some extent, they have also addressed the question of coalition duration and termination. Summing up several studies, Laver & Schofield (1990, 144ff.) describe four types of factors explaining coalition government stability:

(1) Attributes of the overall political system, i.e. the number of parties, the "size" of the party system, the presence of anti-system parties, the degree of ideological polarization, the level of policy influence open to the opposition, the salience of elections to the government formation process, and the presence of a formal investiture requirement.

(2) Coalition attributes, like the cabinet's majority or minority status,
minimal winning or surplus governments, and the ideological compatibility
of coalition members.

(3) The bargaining environment, with a key distinction between multipolar
systems on the one hand, and bipolar and unipolar (centre) systems on the
other, with unipolar (off-centre) systems in an intermediate category.
(4) The events approach, assuming that the fall of governments is caused
by random events.

As Laver & Schofield (1990, 158) point out, analyses based on factors 1-3

are "deterministic" in the sense that all the information used to predict a government's duration is known when the government is formed. The events approach, on the other hand, is "randomistic". The knowledge of why some events are "critical" or "terminal" is limited. A common feature of all four perspectives presented by Laver and Schofield, and of most of the literature I have read, is the lack of attention paid to the political leaders' efforts to organize and manage relations between relevant actors in a way that is suitable for the operation and preservation of the government.

The aim of this article is to fill in some of these blank spots. It is based on a study of the five coalition governments established in postwar Norway. As shown in Table 1, the coalition governments in Norway have always been composed of non-socialist parties (Rommetvedt 1992b, 58). In addition to using written sources like political memoirs and biographies, I have interviewed four of the five prime ministers, several ministers and state secretaries, Members of Parliament, party leaders and party secretaries.¹

The Party-Parliamentary System and the Coalition Apparatus

Rudy Andeweg (1988) suggests that cabinet decision-making is structured by two sources of disagreement: the departments headed by the cabinet ministers, and the parties or factions these ministers represent. The first source of disagreement is obviously relevant to both one-party and coalition governments. Even though the second source is relevant to one-party governments, its salience is most obvious to coalition governments.

The departments or ministries and the parties are certainly among the
most important organizations or organizational elements in this connection.

A comprehensive list of sources of disagreement needs, however, to take into account several other elements as well. The party-parliamentary system, as we may call it, is a complicated network consisting of inter- and intraorganizational relations between a number of elements. Figure 1 maps this system in periods with a three party non-socialist coalition government, i.e. 1983-86 and 1989-90. It is a great challenge to the political leaders to manage these network relations and to operate this complicated system in a manner that promotes coordinated decision-making and action. The maintenance of the coalition and the coalition government's ability to pursue its policy goals depend on the political leader's ability to take care

of this task of organization and coordination. There is no comprehensive theory describing how to fulfil this task. The aim of this article is not to present such a theory. It is limited to the presentation of some of the devices used to handle this challenge in the Norwegian cases. The "modes of executive-legislative relations" we shall concentrate on are the interparty and intra-party modes (King 1976).

Actually, a fruitful point of departure is still the "notes toward theories of coalition behavior" published by Sven Groennings more than twenty years ago. As Groennings (1970, 460f.) points out, "... one frequently encounters the assumption that, for analytical purposes, coalition maintenance is simply repeated coalition formation". However, this assumption is not adequate. Groennings emphasizes the need to identify the maintenance variables and to theorize about the impact and relationships of these variables. He presents a model of coalition maintenance based on four types of variables related to each party participating in the coalition: (a) motivation and communication variables, (b) situational variables, (c) compatibility variables, and (d) strategic variables. Some of these variables are clearly related to factors mentioned in the above references to Laver and Schofield's work.

In our context the most interesting aspect of Groennings' model is the interaction of these variables through the coalition apparatus which is established when a governing coalition is formed. This coalition apparatus is based on some sort of initial agreement as to policies and positions, decision-making and contact structures, and rules of the game. The apparatus variables in the model are: (a) positions and the nature of leadership; (b) programmes; (c) committees; (d) decision-making model; and (e) rules of the game (Groennings 1970, 460). In addition to apparatus variables we will include some aspects of what Groennings calls communication variables.

Policy Statements and Programmes

As Groennings (1970, 461) points out, the coalition parties "normally issue
an extensive statement of their lowest common denominator of purposes....
This statement, perhaps synopsized from party programs and parliamentary
committee reports, will constitute the government's initial programmatic
framework". This applies to Norwegian reality as well. The coalition
formation process in Norway starts with negotiations concerning the
coalition's policies.

In fact, these negotiations may start long before the formation of a
coalition government. In the heyday of the Norwegian Labour Party, which
controlled a majority in the Norwegian Parliament from 1945 to 1961, the

non-socialist parties tried to show their ability to govern by establishing pre-election coalitions or declarations to the effect that the parties would collaborate to form a coalition government if warranted by the outcome of the election. During the 1950s and at the beginning of the 19605, they did not, however, succeed in working out any common platforms. This was due to non-socialist disagreement on policy issues in the Norwegian Parliament - the Storting. A prerequisite for the agreement on a pre-election coalition declaration seems to be a very low level of conflict or disunity between the parties in the Storting during the period before the election (Rommetvedt 1992b).

When the Labour Party lost its majority in the Storting at the election in 1961, it was due to the new Socialist People's Party which captured a pivotal position with their two seats in Parliament. In 1963 the Socialist People's Party voted in favour of a motion of no confidence against the Labour government during the debate on the so-called King's Bay issue, which had its origin in several mining accidents in Spitsbergen. Prime Minister Einar Gerhardsen had to resign, and the Conservatives, under their leader John Lyng, could form a coalition government together with the Centre Party, the Christian People's Party and the Liberal Party. Owing to the fact that there was a socialist majority in Parliament, it was quite obvious that the Lyng government would not last for long. During the inaugural speech debate, Labour and the Socialist People's Party proposed motions of no confidence, and Lyng resigned after four weeks in office.

In his memoirs, Prime Minister Lyng (1973, 200ff.) writes that, at first, he considered making a government declaration which only discussed the parliamentary principles debated in connection with the King's Bay issue. These principles concerned the government's relations with and obligations towards the Parliament. Lyng, however, reached another conclusion. He wanted to use the occasion to work out a comprehensive manifesto which to some extent, though not formally, could serve as a common programme for the coalition parties. This could be of great importance before the election in 1965.

The Lyng government's declaration was worked out by a committee with one representative from each party. The discussions concerning the declaration helped "cleaning up" disagreements between the non-socialist parties. The parties had to make compromises on some issues, but according to Lyng it was not difficult to reach such compromises. Quite a lot of disagreement, which had a historical bakground; was no longer relevant. Lyng's biographer, Lars Roar Langslet (1989, 173) maintains that this work went quite smoothly because Lyng was willing to give substantial concessions regarding issues close to the other parties' hearts. He wanted to kill old myths about the Conservative Party and to move the party towards the political centre.

Formally, the non-socialist parties have never prepared a common programme prior to elections. After the Lyng coalition they succeeded in establishing a pre-election coalition before the election in 1965. The declaration was limited to a statement on the non-socialist parties' ability and willingness to form a coalition government. In reality, however, extensive common remarks in the recommendation made by the Finance Committee in the Storting, concerning the government's long-term programme for the next four year election period, served more or less as a common election programme for the coalition parties. The coalition parties won the election, and Per Borten from the Centre Party could form a new coalition government.

Coalition government formation in Norway starts with negotiations concerning policies. The distribution of ministries is negotiated after the parties have reached an agreement on policies. During negotiations concerning the coalition's policy documents, it is the goal of each party to influence the coalition's policies in accordance with their own preferences. To some extent, we may say that the distribution of "policy rewards" is decided by these documents. The parties sign a protocol and issue a statement on the government's policies. To a large extent this protocol and statement are based on and refer to the remarks in the Finance Committee's recommendation concerning the government's long-term programme. Together these three documents form the programme of the coalition government.

A number of the formulations used in these programmatic documents seem rather vague and indistinct. The public documents do not present concise directions for action. Nevertheless, the importance of the goals presented in the documents was underlined in interviews with members of the Willoch government, especially by the Prime Minister. An important function of these documents was related to the management of party relations. The documents served as a coordinating device for the coalition parties. With reference to this "contract" the Conservative Prime Minister Kåre Willoch could avoid or at least reduce the problems connected with the middle parties' claims for public expenditures. It should be added that when the Willoch coalition was formed in 1983, the parties made a secret deal in addition to the official declaration and negotiation protocol.

Policy compromises are necessary to facilitate the formation of coalitions. The participating parties need, however, to maintain their profiles concerning vital issues. Thus, the Willoch government's declaration contained a statement saying that the parties were free to pursue their differing views on the abortion issue. Without this statement, the Christian People's Party would not have participated in the coalition. A similar formulation was also included in the Syse government's declaration in 1989. The Syse declaration also contained a statement on differences in the parties' views on alcohol. Most important, however, was the so-called "suicidal paragraph" in

the declaration from the Syse government. The paragraph ensured the parties' freedom of action concerning Norway's connection with the EC. This freedom was vital to the Centre Party's participation. Later, disagreement concerning the EC/EFTA negotiations on the European Economic Area led to the dissolution of the Syse government in 1990.

Decision-Making Model and Rules of the Game

Relations between coalition parties are regulated by explicit and implicit rules. As pointed out by Groennings (1970, 461f.), dissent is bound to occur in any coalition. The question for coalition maintenance is how it should be handled. There are basically two decision-making models, the unanimity model and the dissent model. In the dissent model, disagreement can be handled "either by presenting coalition proposals with the dissent appended or announced or by allowing the dissenting party to propose, on the floor of the parliament, certain bills which it will make clear would not be a basis for cabinet proposals". According to Groennings, it seems that "if the dissent model were adopted, conflict would be minimized and both durability and legislative output would be increased. However, apparently every coalition opts for the unanimity model, presumably to minimize the opposition's opportunities to exploit differences within the coalition and to promote a spirit of consensus" (1970, 462).

This view applies to Norway as well. The basic decision-making model of Norwegian coalition governments seems to be the unanimity model. Each party has a veto assuring that the coalition will not put forward proposals to which the party is deeply opposed. There have been, however, different views on how consistently this model should be followed in practice. We have already seen examples deviating from the unanimity principle in the pre-election and government declarations of the coalition parties.

During the Borten government's time in office, the ministers were allowed to dissent. Prime Minister Borten regarded this as a safety valve necessary to prevent the frustration which would occur if the ministers were forced to support a proposal against their will. According to Borten's opinion, the members patience will corrode in a "streamlined" coalition.²

Prime Minister Willoch, on the other hand, was strongly opposing ministers' dissent within his government. Willoch, who was Minister of Trade in Borten's government, feared that if one party was allowed to present alternative proposals and was thus released from the common responsibility, then other parties would be tempted to make similar claims. Consequently, the coalition would risk too much disagreement to govern efficiently, and might even fall apart as the Borten government did in 1971.

It is easier to avoid policy conflicts within the coalition when ministers' dissent is ruled out of accepted practice.³ Prime Minister Willoch also underlined the importance of showing agreement between the coalition parties outside the cabinet. This view was not, however, shared by all the coalition leaders. In interviews with the author of this article, some of the centrally positioned persons during Willoch's time in office maintained that Willoch's "strict regime" caused frustration.

After the Willoch government the pendulum swung back. At the 1985 election the coalition government lost its parliamentary majority. Prime Minister Willoch resigned in 1986 when the Progress Party voted against the government together with Labour and the Socialist Left Party. In the spring of 1987 the Conservatives, the Christian People's Party and the Centre Party tried to establish a new coalition. The three parties, however, had conflicting views on several policy issues. When I discussed the situation with the parliamentary leader of the Conservatives, Jan P. Syse, in March 1987, he maintained that the potential coalition partners should be given some latitude concerning policies. The parties should be allowed to flag the policy preferences close to their hearts.⁴ When Syse formed a new coalition government after the election in 1989, he carried this view into effect. Prime Minister Syse allowed the coalition parties to emphasize their specialities to a larger extent than Prime Minister Willoch did.

The Syse practice leads one's mind in the direction of decentralization and self-governance as a means of managing conflict. Together with problemsolving and bargaining, self-governance has been a predominant policy style developed in Norway since 1945 (Olsen 1983, 188). As Olsen (1983, 89) points out, "The enlarged public agenda has been countered by a delegation of authority from the Cabinet to the individual minister, making him or her more independent". The opinion expressed by Syse indicates that this kind of delegation may be used to manage conflicts of interest and reduce tension between coalition partners as well. It is quite possible, however, that this strategy is a "quasi resolution" of conflict (Cyert & March 1963, 117f.).

Communication and Coordination

As we have already mentioned, committees are among the apparatus variables included in Groennings' model of coalition maintenance. The number of cabinet committees and the formal allocation of their membership and responsibilities are of interest as evidence of how a government organizes itself to coordinate its tasks and to ensure consultation and resolve conflicts (Mackie & Hogwood 1985, 2).

The most important "committee" is the full cabinet. The Willoch government may serve as an illustration in this respect. During its time in office, all the ministers met with each other three times a week. They held a government conference on Mondays, a preparatory State Council meeting on Thursdays, and the official State Council meeting at the Royal Castle on Fridays. Thus, the full government could coordinate the ministers' work quite often. In addition to the formal meetings, the ministers used to have informal discussions to coordinate their views during the joint lunches held at these meetings. Some of the ministers also used the opportunity to consult each other during the car ride from the government building to State Council meetings at the Royal Castle.

In particular, the government conferences on Mondays had an important coordinating function in the Willoch government. On these occasions the parliamentary party leaders of the coalition parties met with the ministers as well. The government used the opportunity to inform the parliamentary leaders about important issues coming up, and to receive their reactions to the ministers' proposals.

This coordination "mechanism" was also applied by the Borten government. The coordinating effect, however, was not as good as in the Willoch government, mainly due to Prime Minister Borten's style of leadership. The parliamentary leader of the Christian People's Party, Lars Korvald, felt that the government was suffering from lack of decision-making ability. Too often, discussions ended with the Prime Minister postponing the decisions on complicated matters. After a while Korvald asked to be excused from the "endless" meetings (Korvald & Heradstveit 1982, 116f.).

The full government's meetings are important to the coordination of policies. Nevertheless, the size of the full cabinet (15-19 members) made it difficult to resolve severe conflicts within the coalition. Hence, the Willoch government appointed a "government's subcommittee" to handle such problems. Officially, there is no inner cabinet in Norway. In practice, however, the subcommittee functioned as an inner cabinet or inner circle within the Willoch government. The subcommittee worked out compromises which the full cabinet normally accepted.

Members of this committee were Prime Minister Kåre Willoch and Minister of Finance Rolf Presthus, Minister of Church and Education Kjell Magne Bondevik and Minister of Transport and Communication Johan J. Jakobsen. The choice of the first two members, both Conservatives, was obvious due to their positions in the government. The other two became members not due to the importance of their ministries, but due to their positions as chairmen of the party organizations of the Christian People's Party and the Centre Party respectively. In other words, the main task of the subcommittee was to make compromises between parties, not to coordinate ministries.

As Eriksen (1988, 45) points out, the Willoch government organized its work somewhat differently from other Norwegian governments. To a larger extent than before, policy issues were prepared by permanent and ad hoc committees with members from all three coalition parties. In addition to the subcommittee, the following committees are assessed to be the most important ones by Berggrav (1985, 41): the economy committee, the security committee, the research committee, the labour committee, the polar committee, and the petroleum committee. The composition of these committees is presented in Table 4. Some of the committees were established by the Conservative one-party government led by Prime Minister Willoch from 1981 to 1983. In this period the committee members were all heads of ministries involved with their respective committee's policy issues and therefore they represented their ministries. In 1983, when the coalition government was formed, additional members were appointed to ensure the representation of all three parties participating in the coalition. In order to represent the Centre Party, the Minister of Transport and Communication became a member of the economy committee, and the Minister of Environment was appointed member of the polar committee. The Christian People's Party was represented by the Minister of Petroleum and Energy in the security and polar committees. When the petroleum committee was established in 1984, the Minister of Agriculture became a member in order to represent the Centre Party.

In earlier coalition governments the use of government committees was
rather limited compared with in the Willoch government. During the
Borten and Korvald governments the state secretaries played a more

important role as party coordinators. Issues which created problems between the coalition parties were left to state secretaries to deal with. They met every Friday to discuss important issues, and state secretary committees were also appointed to take care of coordination problems and conflicts (Eriksen 1988, s_lff.).

Relations Between Government, Parliament and Party

John Lyng, who was prime minister in the first Norwegian coalition government, was appointed Minister of Foreign Affairs in the second coalition, the Borten government. In his memoirs he takes up the relations between the government, the parliamentary party groups and the party organizations in a coalition (Lyng 1976, 326ff.). With reference to the British politician Leo Amery, Lyng describes this relationship as follows: "A government coalition is like a pyramid with a weak foundation. It is tightly connected at the top, i.e within the cabinet. As one moves downwards from the top to the foundation, the ties become looser".

As Lyng points out, this is explained by the fact that the ministers are a relatively small group of people having almost daily contact with each other. The cabinet members have to act as a unity within the political environment. They have common responsibilities towards Parliament. Hence, they develop a sense of common interests and solidarity. The cooperation within the coalition is more complicated on the parliamentary level. The number of politicians involved is larger, and they have less contact with each other. The temptation to operate on their own is greater, due to the competition between MPs and parties.

The competition is even keener between the foundations of the parties, i.e. the central and local organizations. Lyng maintains that these organizations take care of the daily contact with voters. They are responsible for attracting voters' support. Parties that are relatively close to each other, like members of a coalition, are competing for the same voters. What one party gains, another coalition member is likely to lose. Consequently, Lyng (1976, 328) says, coalitions are rather short-lived, while parties have a natural proclivity to survive as long as possible.

Earlier in this article attention is drawn to two sources of disagreement suggested by Andeweg (1988); the ministries and the parties. What Lyng does, in fact, is to point out a third source of disagreement, namely the functional specialization of the various parts of each party. The formation of a coalition government implies that the functional differentiation is extended. Consequently, the potential for intra-party conflict is enlarged, and the management of intra-party relations becomes more important.

The main functions of the party organization are to work out the party manifesto and to organize the election campaign. The party organization is expected to emphasize its peculiar policy profile to attract voters.⁶ Interparty relations are characterized by disagreement and competition. The party's government members, on the other hand, have to cooperate with their coalition partners. They have to build compromises and to emphasize agreement with their colleagues within the government. Difference and unity are the catchwords attached to the main functions of the party organization and the government respectively.

The parliamentary party's function is two-sided. The MPs constitute the parliamentary foundation of the coalition government. They have to ensure parliamentary support of the cabinet. Hence, they have to agree with their coalition partners on major issues. At the same time, however, the MPs have to show their parties' ability to fulfil their promises to the voters and to implement the policies advocated in their respective election programmes. The parliamentary party has to manage both cooperation and competition. Cooperation and compromise are mainly taken care of by the parties' members in the standing committees in Parliament. Plenary debates are the arena for competition where parties emphasize their peculiar abilities distinguishing them from their competitors.

The importance of the two major functions is illustrated in Fig. 2. The different parts of the party are ranked according to the degree of coordination with other parties. The potential for intra-party conflict connected with the functional specialization within parties is large. Hence, the aim of coalition leaders is to manage intra-party relations so as to reduce and avoid manifestation of such conflicts.

The importance of contact and communication between the various parts
of the party is obvious. We have already mentioned that parliamentary

leaders of the coalition parties normally attended the regular government meetings on Mondays. Thus, the parliamentary party leader plays the role of liaison officer between the party's government members and its Members of Parliament.

When the non-socialist parties have participated in pre-election and government coalitions, they have established a common coalition faction within each of the standing committees in Parliament. Ministers have communicated directly with MPs at regular meetings arranged every Wednesday between the coalition faction of each committee and the relevant minister. During these meetings, the minister has discussed with the MPs propositions under preparation in the ministry. Hence, the MPs have been given an opportunity to influence the government's and ministers' policies. In my interviews with MPs, some of them have underlined the fact that the information given at these meetings represents an advantage to the coalition MPs compared with their opponents. The integrating effect of this practice is relevant to both intra-party and inter-party relations.

It seems natural to expect that the parliamentary groups of the coalition parties would arrange common meetings with cabinet members. This, however, has not been the case in Norway. Common group meetings have been held on a few exceptional occasions only. Neither have individual ministers from one party met with the parliamentary groups of the other coalition parties.

Traditionally, the parliamentary group is the most influential part of the non-socialist parties in Norway. The party organization used to be a campaign organization rather than a policy-making body. This has changed over the years. Lars Korvald, who was Prime Minister in the third coalition, maintains that the 1970s became the era of the party organization.⁷ The party organization engaged political advisors, and the central board of the party became a policy-making body to a larger extent than before.

The politicization of the party organization influenced intra-party and coalition relations. Consequently, the position of the party chairman was influenced as well. When Minister of Finance Rolf Presthus was elected chairman of the Conservative party organization in 1986, it was considered whether he should leave the cabinet to concentrate on party work. However, the chairmen of the Christan People's Party and the Centre Party protested. They feared that Presthus would be free to emphasize interests specific to the Conservative Party in a way which they, as members of the cabinet, could not do. As a result of this discussion, Presthus remained in government. He was, however, appointed Minister of Defence in order to have an easier work load.

This question was relevant at the formation of the Syse government as
well. The chairman of the Centre Party, Johan J. Jakobsen, was uncertain
whether or not to become a cabinet member. After pressure from Prime

Minister Syse, he let himself be appointed Minister of Local Government.
On both occasion we see that the coalitions chose a strategy of co-optation.

The co-optation of party chairmen was met with some scepticism within the party organizations. A kind of compensation was established in order to reduce this scepticism and to give the party chairmen the opportunity to show their parties' ability to influence policies. As Minister of Finance during most of the Willoch government's term of office, Rolf Presthus had the task of limiting public expenditures. His job was to say no when other ministers asked for more money. However, the ministers of transport and of church and education who were also chairmen of the Centre Party and the Christian People's Party respectively were given a little extra, a "chairman's surplus", when they asked for money. According to Presthus, this practice caused some envy among other ministers.⁸

Conclusion

Coalition governments operate within a complicated network consisting of a great number of organizational elements. This network, or the partyparliamentary system as we have called it, represents a great challenge to the leaders of the coalition parties. They have to organize relations between elements with different functions, thus representing a considerable potential for conflict. Executive-legislative relations, as well as inter- and intraparty relations have to be managed by the leaders.

In this article five Norwegian coalition governments have served as illustrative cases showing how these relations have been managed in practice. The coalition leaders' task is two-sided. In order to ensure the coalition's ability to govern, the coalition parties have to coordinate decision-making and action. At the same time, they try to influence coalition policies in accordance with their own peculiar preferences. The coalition partners are also concerned with the preservation of party profiles in order to attract voters. In short, the coalition leaders and parties have to manage unity and difference at the same time.

Government coalitions establish a coalition apparatus with lines of communication intended to resolve this problem. Party leaders organizing Norwegian coalitions have chosen quite similar patterns of organization. Party chairmen are appointed as ministers, all parties participating in the coalition are represented in government committees, parliamentary party leaders meet regularly with the cabinet, etc. These are devices used to coordinate parties' and politicians' views and actions. On the other hand, ministries are allocated to parties with special preferences, and normally one party only is represented in the political leadership of each ministry. Thus, characteristic party profiles are taken care of.

The similarities between Norwegian government coalitions' patterns of organization are obvious. It is, however, interesting to note variations in the management of party relations. The priority given to coordination and unity versus party differences and profilation has varied, especially among the prime ministers of Norwegian coalition governments. In this respect, the pendulum seems to swing back and forth, indicating a sequential attention to inconsistent goals (Cyert & March 1963, 118). The kind of sequential attention paid to shifting goals which is involved here, seems to be connected mainly with the formation of new coalition governments learning from earlier experiences. What the movement of the pendulum indicates is that there is no final solution to the difference-unity dilemma.

---

NOTES

* This article is basedon a paper presented to the workshop "The Core Executive and the Management of Party Relations", ECPR Joint Sessions of Workshops, Leiden, 2-8 April, 1993.

1. See Rommetvedt 1991. The study was funded by the Research Council of Norway.

2. Interview by author 14 March 1987.

3. Interview by author 1 April 1987.

4. Interviews by author 30 March 1987 and by the newspaper Verdens Gang 4 April 1987.

5. Interviews by author 2 April 1987 and 3 April 1987.

6. This view ties in with the "directional theory" proposed by Rabinowitz & Macdonald (1989). It deviates from Downs's "economic theory" (Downs 1957).

7. Interview by author 23 April 1987.

8. Interview by author 2 April 1987.

REFERENCES

Andeweg, R. B. 1988. "Centrifugal forces and collective decision-making: The case of the
Dutch Cabinet", European Journal of Political Research 16, 125-151.

Budge, I. & Keman, H. 1990. Parties and Democracy. Coalition Formation and Government
Functioning in Twenty States. Oxford: Oxford University Press.

Berggrav, D. 1985. "Regjeringen", in Nordby, T., ed.. Storting og regjering 1945-1985:
Institusjoner - rekruttering, Oslo: Kunnskapsforlaget.

Cyert, R. M. & March, J.G. 1963. A Behavioral Theory of the Firm. Englewood Cliffs, New
Jersey: Prentice-Hall.

Downs, A. 1957. An Economic Theory of Democracy. New York: Harper & Row.

Groennings, S. 1970. "Notes Toward Theories of Coalition Behavior in Multiparty Systems:
Formation and Maintenance", in Groennings, S., Kelley, E. W. & Leiserson, M., eds.,The
Study of Coalition Behavior, New York: Holt, Rinehart and Winston, pp. 445-465.

King, A. 1976. "Modes of Executive-Legislative Relations: Great Britain, France, and West
Legislative Studies Quarterly 1, 11-36.

Korvald, L. & Heradstveit, P. 0. 1982. Politikk og kall: Lars Kor valds memoarer i samarbeid
med Per Øyvind Heradstveit. Stavanger: Rocon.

Langslet, L. R. 1989. John Lyng: Samarbeidets arkitekt. Oslo: Cappelen.

Laver, M. & Schofield, N. 1990. Multiparty Government. Oxford: Oxford University Press.

Mackie, T. T. & Hogwood, B. W. 1985. "Decision-making in Cabinet Government", in
Mackie, T. T. & Hogwood, B. W., Unlocking the Cabinet. London: Sage, pp. 1-15.

Olsen, J. P., 1983. Organized Democracy. Oslo: Universitetsforlaget.

Rabinowitz, G. & Macdonald, S. E. 1989. "A Directional Theory of Issue Voting", American
Political Science Review 83, 93-121.

Rommetvedt, H. 1991. Partiavstand og partikoalisjoner. Stavanger: Rogaland Research.

Rommetvedt, H. 1992a. "The Norwegian Storting: The Central Assembly of the Periphery",
Scandinavian Political Studies 15, 19-91.

Rommetvedt, H. 1992b. "Norway: From Consensual Majority Parliamentarism to Dissensual Minority Parliamentarism", in Damgaard, E., ed., Parliamentary Change in the Nordic Countries. Oslo: Scandinavian University Press/Oxford: Oxford University Press, pp. 51-97.

Solstad, A. 1969. "The Norwegian Coalition System", Scandinavian Political Studies, Yearbook
4, 160-167.