Resumé

According to the classical parliamentary doctrine of majoritarian rule, governments should be large, united and accountable to the voters. Since the introduction of proportional representation in the beginning of this century, these requirements have seldom been fulfilled in Continental,politics. In this article the Swedish experience of minority parliamentarism is analyzed. The conclusion is drawn that the consensus model of democracy that has been practiced in this country comes closer to the ideal of the Conservatives who a hundred years ago opposed the parliamentary system.

Introduction

Parliamentarism was introduced into Sweden in 1917, when the King and the Conservatives were forced to agree that in the future the composition of the government would no longer be the King's personal choice but dependent on the party composition of Parliament. According to the leading ideologue of the parliamentary system, the Liberal Party leader Karl Staaff who also suggested a political science research program on parliamentarism, the rule of consensus characteristic of predemocratic society should in a parliamentary democracy be replaced by the rule of the majority. A government should be large enough to control Parliament and united enough to pursue consistent policies. To form a minimum winning coalition was the goal of the politicians. The opposition should be excluded from government. The function of the minority was to criticize the government with the aim of forming a majority government itself after the next election. This was the way to make government democratic, that is accountable to the voters (Staaff 1917).

Figure 1 depicts out the development that has followed. It is a picture that
would have shocked the victorious Left a hundred years ago, if they could
have seen it. For history took quite a different turn from what they expected.1

---

* Leif Lewin, Department of Government, Uppsala University, Skytteanum Box 514. S-751 20 Uppsala, Sweden.

Only one government fulfills the classical criteria of being "large" and "united": between 1968 and 1970 the Social Democrats held a majority of the seats in Parliament but in the election of 1970 they returned to the minority they had held earlier. The other governments above the fifty percent mark, indicated by a dotted line in Figure 1, were either coalition governments that were ideologically divided and split up before the subsequent election or the grand coalition during Second World War. Minority governments have been by far the most common. The average parliamentary support enjoyed by Swedish governments between 1920 and 1994 has been 41.5 percent.

The Swedish situation reflects an international pattern. At the turn of the century a large number of countries had, like Sweden, changed their electoral system from majority to proportional representation, thus making it easier for parties to be represented in Parliament. In a multi-party system it is of course more difficult to achieve a majority than in a two-party system. Thus, while majority rule prevailed as the dominant theory of parliamentarism, in reality minority governments have been common in Continental Europe (Laver & Schoefield 1990). The minimum winning coalition did not turn out to be a realistic picture of politics:

If participating in winning governments is the overriding goal of politicians, someone should tell the politicians. It is patently not the case that participation in winning government is the determinative goal. It is even more doubtful that participation in any government is preferable to non-participation (Luebbert 1983, 240).

Studies on Italy (Spotts & Wiser 1986; LaPalombara 1987), Germany (Pridham 1982), Switzerland (Linder 1994), The Netherlands (Lijphart [1968] 1975), Austria (Luther & Muller 1992), and the Nordic countries (Damgaard et al. 1992) all report that the minimum winning model of democracy is not a very forceful explanation for these countries.

Purpose of the Study

Against this background, how should a "minority parliamentary government" be understood? How have Swedish politicians argued for the governments they have put together? How should the form of government that in reality has been practiced in Sweden be characterized?

The Most Easily Tolerated Government

After the introduction of the parliamentary system there followed, contrary to expectations, a decade of small governments. The government exercised power despite its minority position by playing the role of a pivot, putting together voting majorities from its position in the political middle with support sometimes from the left and sometimes from the right.

From a normative point of view this was a deviation. "According to the basic rules of the parliamentary system, every minority cabinet is an unwanted crisis symptom" and may in the end lead to "decisional paralysis" (von Beyme 1970, 570-71). But if one would like to understand what actually happened, it seems more fruitful to regard the party leaders as rational actors and the formation of a government as a deliberate process that can stop "before majority status has been reached. This happens because core members of the coalition want no further expansion, because marginal parties do not want to join, or both" (Strom 1990, 52).

The pivot was the ideal of C. G. Ekman, the leader of the Liberal party and twice prime minister with the support of only 14 percent and 13 percent respectively of the seats in Parliament. His arguments were quite curious. The art of government pursued by a pivotal party was, in his view, morally superior to that of a party with a majority because the decisive factor was the nature of substantive issues, not the tactics of bloc politics. In virtue of its small size, such a government could only count on voluntary support from other parties. He described the classical form of parliamentary government as "blocs of MPs who, whipped together by force, lack any independent vote". In relation to this ideal of democracy, the best government was a small government, a government that was not based on a majority of the people but rather one whose success was the result of the sincere support of many. The

pivot had an excellent capacity to capture the will of the people. It was the
best manifestation of democratic rule (Ekman 1928a; 1928b).

Long after the first era of pivotal rule was past in Sweden (the twenties), the Liberals continued to espouse their governmental ideal. One later leader of the Liberal Party found it strange that over the years the label "pivot" had come to be something of a pejorative. Rather, the pivot produced the policy that the people preferred in comparison with all other available alternatives. "Parliamentarism according to the classical theory" with two competing teams "no longer exists even in theory. In practice, something has happened that the architects of the system had not counted on: more than two parties." That one of these architects was Staaff was of course a matter of some embarrassment to a leader of the Liberal Party. His polemics was therefore blunted. Staaff had also indicated which types of government he put in second place, which in third and so on. "Even if Staaff thus put minority government in next to last place, he obviously considered such a government a possibility" (Andersson 1959, 15-25). Staaff's criticism of minority government was thus reinterpreted as pragmatism: since Staaff disliked minority government, one could see that he was enough of a realist to expect that it could occur!

In the 1970s the pivot surprised everyone by returning to Swedish politics. Once again the prime minister was a Liberal, now with only 11 percent of the MPs behind him. Since then the pivot has remained the most common form of government in Sweden. The prime minister declared that the Liberals had always been opposed to "irreconcilable bloc politics" (Proceedings in Parliament 1978/79 21:5). He showed discretion in not mentioning what Staaff had to say on the matter. What was more surprising, however, was that he later denied that he had led a pivotal government or that the small Liberal government that had succeeded his had been a pivotal government either.

The interpretation of [these as] a "pivot" not only displayed ignorance of the facts, it did not reflect the opinion of the Liberals about how governmental power should be exercised. It was not the government's numerical basis in Parliament that was decisive but rather its political position. It was considered that a purely Liberal government could more easily shape policies in line with the political direction the Liberals wished to promote than a coalition of Liberals and Moderates (Ullsten 1984, 460).

This statement is a combination of truism and self-contradiction: it goes without saying that the policies of a Liberal government would lie closer to the party-line of the Liberals than those of a coalition government of Liberals and Moderates; and the argument that it is position, not numbers that is the basis of policy is actually an excellent definition of pivotal government.

The pivot has continued to dominate the political scenery during the last 25
years: Social Democratic governments balancing between the non-socialists
and the Left Party-Communists during the seventies, eighties and again today,

as well as the non-socialist government during the early nineties balancing
between the Social Democrats and the right-wing party New Democracy.

Reality differed from theory. It was hard to form large, united and accountable governments. Political scientists began to reconsider the doctrine of parliamentarism. When one of Staaff s most devoted pupils got the chair in political science at Uppsala with the ambition to implement Staaff s research program, he therefore suggested that parliamentary theory be reformulated to account for the experience of minority parliamentarism. Unlike its British counterpart, the Swedish parliamentary system lacked an active expression of will on the part of Parliament with respect to government. It was therefore better to say that parliamentarism was a form of government in which the government "was tolerated" by the majority in Parliament. The pivot fitted well into this picture: it was the most easily tolerated government. To capture this idea, he coined the concept of "negative parliamentarism" (Brusewitz 1929, 323-34).

The pivot produces the policy that the people prefer to all available alternatives, this had been the message from a Liberal party leader. Is that just political rhetoric or is there some substance to such a claim? When we talk of preferring various alternatives to each other or tolerating one government more easily than another, we assume that citizens feel differently about the various parties. Some are liked very much, some are accepted, some are unsympathetic, some are abominable. The intensity problem is a well-known shortcoming of the majority rule, for it cannot express this variation in feelings (Dahl 1956; Kendall & Carey 1968). But since Borda (1781) we know that this problem can be overcome (for a modern approach, see Riker 1982). One way of doing this is by giving the voters several votes so that they can grade their likes and dislikes. In actual elections this hardly ever occurs. But this situation can be simulated in research. In Swedish electoral surveys in 1991, those interviewed were asked to rank their party preferences, data that we have transformed into a Borda analysis of the election of that year (Figure 2).²

The main difference between this outcome and the official election result according to the majority rule is that the two large parties - the Social Democrats and the Moderates - are weakened and the other parties become larger. In other words, the Borda analysis tends to even out the size of the parties. Even if it is not apparent from the table, the underlying data informs us that the voters place Social Democrats and Moderates as their first and last preferences, whereas the middle parties are to a much greater extent secondorder preferences. The middle parties, and especially the Liberals, arouse greater sympathy among the electorate than the simple majority rule can express. "In Sweden, people like the Liberals but they don't vote for them". The parties in the middle are most easily tolerated as governors: a prime minister from this camp represents the minimum common politics.

Bargaining

Bargaining is a technique for the pivot to form voting majorities with the help from left or right. But this method is also used by other archetypical roleholders in the political game, perhaps most significantly by the coalition builder. The coalition builder tries to widen the parliamentary basis for the government by bargaining with another party to get an agreement that stands for a longer period of time.

Let us imagine the following preferences:

This captures the situation in a parliamentary committee before one of the most important coalitions in modern Swedish history, namely that between the Social Democrats (L) and the Agrarian Party (M), was formed in 1933, a step that led Sweden out of the minority parliamentarism of the twenties to the majority coalition in the thirties. In the model R depicts the non-socialist opposition. Of the issues, b symbolizes the agricultural policy of the Agrarian party, c represents the unemployment policy of the Social Democrats, and a stands for the defeat of these proposals. In the committee, the two policies became inseparably interconnected. None of the alternatives received a majority of first preferences. Instead we have an instance of Condorcet's voting paradox a > b > c > a (Condorcet 1785; Arrow 1963), something L exploited to undermine the expected success of the non-socialists. L

commenced bargaining and succeeded in reaching an agreement with M
through log-rolling. The adjusted preferences after this had occurred were as
follows (the preferences that have changed place are underlined):

The coalition of L + M, built on the adjusted preference orders, voted
assent to its first-order and second-order preferences while R's a was
thwarted.

The value of broad solutions and consensus was the main theme taken up by the leader of the Social Democratic Party, prime minister Per Albin Hansson, when he defended the log-rolling before Parliament. He took exception to a simplified procedure of the kind that only satisfies the firstorder preferences of the parties. A rule like that favored the no-votes; it was a simple matter to defeat M's b with votes from L and R, and there was no difficulty in defeating L's c with votes from R and M. But when the country was in the grips of the Great Depression, the stalemate facing the minority government in Parliament could be accepted no longer. The prime minister had therefore entered negotiations with all the parties and an agreement had also been reached. With arguments remarkably similar to those later to be put forward by political scientists Buchanan and Tullock, Hansson made an earnest plea for log-rolling and consensus. The intensity of the interest people took in different questions varied. Contrary to its reputation, log-rolling was a morally acceptable method by which most people could have their preferences satisfied depending on how intensely they felt about them. This was a method for "the politics of a good society" (Lewin 1988; cf. Buchanan & Tullock 1962).

The ultimate step in broadening the parliamentary basis of the government is to form a grand coalition. What was one to say about the fact that in 1933, one of the parties, R, was left out of the agreement? This had actually worried the prime minister greatly, and he emphasized time and time again in his address to Parliament how desirable it would have been to extend the bargain "in traditional Swedish fashion" so that R could also get some satisfaction. There was room for everyone in "the good society". Six years later, under the extraordinary circumstances of the Second World War, Hansson assembled such a coalition. In the bargain struck between the parties, partisan differences were put aside and agreement was reached on that which in the seriousness of the moment was essential: foreign and defense policy. However, what is more interesting from our perspective is the fact that at the end of the war Hansson reasserted his conviction that a grand coalition would be valuable in the future, in peacetime. "In classical parliamentary theory" it was assumed that there was a minority that could become a majority. This did not seem to hold now, declared the self-confident Social Democrat who could

look back on his party's electoral victories of the thirties and forties. The nonsocialist parties seemed doomed to perpetual opposition. This was not healthy for the country's political life. In "the good society" everyone should have an influence. A permanent grand coalition was one way of approximating the preferences of all (Proceedings of the Swedish Social Democratic Party Congress 1940,51-68). Developments did not turn out as Hansson wished but the idea of a permanent grand coalition has since then remained an alternative in Swedish politics, one that the parties return to from time to time (Ruin 1968).

Forming a majority by coalition building seldom results in a government that is sufficiently "united," though, to meet the standards of parliamentary theory. As already mentioned, most majoritarian coalition governments have split up before the election. There have been five of them (Figure 3).³

Not even the large reform government of 1917-20, which introduced parliamentary democracy, lived up to Staaff's ideal; after carrying out the reform, the Liberals and the Social Democrats split on economic issues - "socialism" - and the government resigned. Interesting from the point of view of majoritarian or consensus democracy are the words that the out-going prime minister, Staaff's successor as the leader of the Liberal Party, wrote after the dissolution of the government.

So much should be clear, that one cannot and should not attempt to transfer the English Cabinet's position of power to our country without modification ... A Swedish parliamentary government cooperates with Parliament and its party to a far greater extent that its English counterpart, and as long as there is mutual trust, the leadership of the government need not be the least impaired by it (Eden 1920, 436-37).

The red-green coalition of 1936-39 was a successful one - only the outbreak of the Second World War made it necessary to replace it with a grand coalition. From the point of view of evaluation there is reason to regard it as a "united" government and consequently acceptable to the doctrine of parliamentarism (like the already mentioned one-party Social Democratic majority government of 1968-70).

The subsequent coalition formed between the Agrarians and the Social
Democrats in 1951-57 split on the supplementary pension issue.
The 1976 coalition of the three non-socialist parties - Agrarians (now

Center Party), Moderates and Liberals - could not reach agreement on nuclear
energy policy.

In 1981 a second attempt of the three non-socialist parties to govern together ended in disunity over taxes when the two middle parties took the remarkable step of bargaining with the Social Democrats instead of trying to reconcile their differences with the Moderates.

And the largest government of all according to Figure 1, the grand coalition during the war (not included in Figure 3), was fundamentally disunited, especially on economic issues or "socialism" like the reform government in 1920, and resigned as soon as the war was over - in spite of what the prime minister had wished. By definition, a consensus government including all parties with no division of roles between a majority in government and a minority in opposition is contrary to the original meaning of parliamentarism.

Representativeness

In majoritarian democracy the opposition is excluded from power not because it is unimportant. Staaff underlined the importance of an opposition that could criticize the government, aiming at forming the government itself after the next election. For the adherents of the British parliamentary doctrine, making the government accountable to the voters was the way in which the opposition fulfilled the vital function of strengthening the legitimacy of democracy.

In the Swedish political tradition that we have tried to describe in this article, "accountability" is seldom mentioned as a value. Instead, legitimacy is promoted by another strategy. By sharing power with the parties in opposition and including them in the rule of the country, the government is supposed to be regarded as representative for the people as a whole and consequently one that all can feel loyal to. To "reach consensus," to "find a common policy," to "capture the will of the people" have been the declared motives of Swedish politicians. Representativeness is the central norm in Swedish political culture. The government should represent the people's opinion. Or to quote a leading authority:

To create "representativeness of opinion'1 has been interpreted as the central as well as the
operationalizable goal that the proportional electoral method is intended to provide for. . . .
This also means that it is deemed unsatisfactory to reduce democracy to the view that the
influence of the citizenry is essentially restricted to the possibility of ex post facto being able
to dismiss those elected when it is felt that they have misbehaved (Westerståhl 1994. 99-100).

The analysis of Swedish parliamentarism then leads us to a well-known

confrontation: the one between majoritarian and consensus democracy, or to put it in other words, between the Downsian and Lijphartian models of democracy (Downs 1957; Lijphart 1984). Stating this, we immediately must add that Lijphart's democracy is characterized by large government, even grand coalitions, whereas the most common Swedish government is a minority government. How can this be reconciled? The answer is that there often is, as we have seen, "a hidden majority" behind the Swedish minority governments. Their aim has been - to use a typical Swedish expression - to "anchor" its politics with the majority. Even small governments have had the ambition to pursue policies that are liked by a plurality. That was what prime minister Hansson referred to when he spoke of the "traditional Swedish fashion". "Consensus" is the basic value.

It should be emphasized that majoritarian and consensus democracy are just two models and models seldom correspond to reality one hundred percent. As a matter of fact Swedish politics has also majoritarian characteristics. During the first two decades of this century, before parliamentarism was accepted, and again during the seventies and eighties, we witnessed a two bloc system with shifting governments, even if during the latter period there were also many agreements across this bloc cleavage; during the nineties cooperation and consensus between parties is again the dominant trend. The best description is perhaps therefore to say that Sweden has been somewhere between the two models, "between the grand coalition and a two-party system," to quote the title of a well-known study (Ruin 1968). However, there is no doubt about the fact that of the two models Sweden is closer to consensus democracy - with the qualifications just made.

In order to understand this tradition of consensus, it is useful to go back to pre-democratic times where we started. There can be found a governmental ideal that comes close to Lijphart's power-sharing democracy. In the predemocratic form of government, against which Staaff and the left-wing parties revolted, the opposition was included. The King wished all opinions to be represented - no one was to be left out and everyone should have a place. Policies were to be formed so that they could be accepted by all. "British parliamentarism" - with its bipartite structure, the majority party holding power and the minority party being excluded from it in opposition - was considered by the leading ideologists to be foreign to the Swedish political style. Swedish governments based their policy instead "on a compromise, consisting in the fact that two opposed parties and interests, each for its own part, yield somewhat and in this manner carry out the intended reform or measure" - thus functioned "the true compass" of Swedish politics, which "levels out differences and thereby brings about a decision that is for the benefit of the country as a whole and not simply for one party" (Fahlbeck 1916, 45, 84). To adopt parliamentarism in Sweden would entail "extremely grave difficulties" in a country that was accustomed to "assessing issues

objectively and in unison" (Hallendorff 1911, 395-400). "The parliamentarism that the Left is striving for is by its nature partisan rule, something which is greatly foreign both to our constitution, to the intentions of our constitutional fathers, and to our political tradition" (Rexius 1917, 181-93).

Sweden's consensus tradition is often portrayed as the result of the Social Democrats' famous and successful log-rolling with the Agrarians in 1933. In this article we have searched for deeper roots. We wish to draw attention to the similarity between the ideal that was embraced by Swedish Conservatives at the turn of the century and Lijphart's power-sharing model. Perhaps it is too much to expect that the victors of the Left would admit that in the long run it is the ideal of the losers, the Conservatives, that has come to shape the development of parliamentary government in Sweden. That is nevertheless what has happened.

---

NOTES

1. Adopting a historicistic argument Staaff maintained that parliamentary government was the ultimate outcome of a long process of constitutional development. To oppose this end was to defy a historical law. To be critical of parliamentary government, as the King was, was tantamount to acting unconstitutionally. In thus anticipating a practice that was established only later, the Left succeeded in further provoking the Right. See my article Lewin 1987.

2. A net sample of 3,588 people was selected. Most of these were interviewed at home; a smaller number by telephone. Non-response was 24.2 percent. The respondents were asked to place each party on a scale from —5 to +5. The value —5 represents "strongly disapprove", the value 0 stands for "neither approve nor disapprove", and value +5 means that the person "strongly approves" of the party in question. Borda points are calculated in such a way as to allow respondents to have "weak" preference orders, i.e. that they are indifferent to two or three parties. Borda points for each party are calculated and the sum of these yields the grand total of all Borda points. This makes it possible to calculate the number of seats in Parliament each party would obtain: a party's share of the total number of seats is equal to its share of the grand total of Borda points. For example, if a party wins 6.000 Borda points of a grand total of 60.000, it receives 10 percent of the seats. Should it instead win 12.000 points, it would receive 20 percent of the seats etc. Points of criticism to this approached are that the voters are "forced" to take all parties into consideration and that the distances between all "ranks" are equally large. Acknowledgment: Department of Political Science, University of Gothenburg and Sverker Härd at my department, who has carried out the technical adaptation for this study.

3. I agree with Damgaard (1994) that the problem is not what we might mean by "termination" of a government but rather how we should define "government". Here the following definition is used: a new government is considered to have been formed if one of the following occurs. 1) a new prime minister is appointed (except when the prime minister dies or resigns as party leader and is replaced by the person who takes over the position of party leader); 2) a change takes place in the parties that constitute the government.

ACKNOWLEDGMENTS

The article, which was translated by Donald Lavery, summarizes my books Votera eller förhandla? Om den svenska parlamentarismen (Norstedts Juridik 1996), and "Bråka inte!" Om vår tids demokratisyn (SNS 1998). The research has been sponsored by the Bank of Sweden Tercentenary Fund.

REFERENCES

Andersson, G. 1959. "Regeringar utan riksdagsmajoritet," Ekonomi, politik, samhälle Folk och
samhälle.

Arrow, K. 1963. Social Choice and Individual Values. Yale University Press.

Brusewitz, A. 1929. "Vad menas med parlamentarism?," Statsvetenskaplig Tidskrift.

Damgaard, E. (ed.), et al. 1992. Parliamentary Change in the Nordic Countries. Scandinavian
University Press.

Downs, A. 1957. An Economic Theory of Democracy. Harper and Row.

Eden, N. 1920. "Regeringsbildning och regeringsarbete. Några anteckningar och reflektioner,"
Forum.

Ekman, C. G. 1928a. "Skrämma, locka eller övertyga?," Politiska dagsfrågor.

Ekman, C. G. 1928b. "Valmetoder, sociallagstiftning och regeringsansvar," Politiska
dagsfrågor.

Fahlbeck, P. 1916. Engelsk parlamentarism contra svensk. Ett stycke nutidshistoria. Gleerups.

Kendall, W. & Carey, G. 1968. "The Intensity Problem and Democratic Theory," The
American Political Science Review 62.

LaPalombara, J. 1987. Democracy, Italian Style. Yale University Press.

Laver, M. & Schofield, N. 1990. Multiparty Government. The Politics of Coalition in Europe.
Oxford University Press.

Lewin, L. 1987. "Historiens mening," in Lewin, L., ed., Festskrift till professor skytteanus Carl
Arvid Hessler. Acta Universitatis Upsaliensis.

Lewin, L. 1988. Ideology and Strategy. A Century of Swedish Politics. Cambridge University
Press.

Lewin, L. 1996. Votera eller förhandla? Om den svenska parlamentarismen. Norstedts Juridik.

Lijphart, A. [1968] (1975). The Politics of Accommodation: Pluralism and Democracy in the
Netherlands. University of California Press.

Lijphart, A. 1984. Democracies: Patterns of Majoritarian and Consensus Government in
Twenty-One Countries. Yale University Press.

Linder, W. 1994. Swiss Democracy. Possible Solutions to Conflict in Multicultural Societies.
St Martin's Press.

Luebbert, G. 1983. "Coalition Theory and Government Formation in Multiparty
Democracies," Comparative Politics 15.

Luther, K. & Muller, W. (eds.) 1992. "Special Issue on Politics in Austria: Still a Case of
Consociationalism?" West European Politics 15.

Pridham, G. 1982. "The Government/Opposition Dimension and the Development of the Party
System in the 1970'5: The Reappearance of Conflictual Politics," in Döring, H. & Smith, G.,
eds., Party Government and Political Culture in Western Germany. The MacMillan Press.

Proceedings in Parliament.

Proceedings of the Swedish Social Democratic Party Congress.

Rexius, G. 1917. "Parliamentarismen och svensk tradition," Svensk Tidskrift.

Riker, W. 1982. Liberalism Against Populism. A Confrontation Between the Theory of
Democracy and the Theory of Social Choice. Freeman and Company.

Ruin, O. 1968. Mellan samlingsregering och wåpartisystem. Bonniers.

Spotts, F. & Wiescr, T. 1986. Italy. A Difficult Democracy. Cambridge University Press.

Strom, K. 1990. Minority Government and Majority Ride. Cambridge University Press.

Westerståhl, J. 1994 in Statsvetenskaplig Tidskrift.