Consensus or Conflict? Legislative Behaviour of Opposition Parties during Minority Government in Denmark

Consensus or Conflict? Legislative Behaviour of Opposition Parties during Minority Government in Denmark Very first draft. Not for quote. Flemming Juul Christiansen Department of Political Science Aarhus University Denmark fjc@ps.au.dk Paper prepared for the Annual Conference of the Italian Political Science Association 1

Palermo, 8 10 September 2011 Abstract This paper looks at consensus and dissent between governments and opposition parties in Denmark. The country is characterised by minority governments and cooperation at the same time. One of the major strategies for minority governments to pass legislation is to negotiate legislative agreements with opposition parties. With legislative roll call data, this paper finds support for the argument that such settlements are ways for political parties to handle cooperation and conflict simultaneously. 2

Introduction In a democracy consent and dissent are integrated parts of the political process. Political parties disagree over preferred aims, and they compete for vote and office. But unless there is a majority party in parliament, coalitions are also needed between parties to govern. So, party competition and cooperation go together, and coalition parties need to solve problems of how to do both at the same time. According to the literature of coalition governance (Strøm, Müller and Bergman 2008; Moury 2010), political parties try to solve such problems through institutional mechanisms such as inner cabinets and coalition agreements. The literature usually deals with majority coalition governments but this paper will take minority governments into consideration because here opposition parties are necessary for the majority building. Opposition parties do not have cabinet seats, and one of their aims may be to replace the current government. Nevertheless, the attitude of non-cabinet parties towards the government can take various forms from supporting the government over responsible opposition towards antagonism. Especially under minority parliamentarism a common feature in the Scandinavian countries (Christiansen & Damgaard 2008) the involvement of opposition parties will determine the ability of the political system to make political decisions (Strøm 1990; Green-Pedersen 2001). If a minority government is a result of no viable alternatives and faces hostile or reluctant opposition parties, the political decision making process could become gridlocked, and political crisis may evolve. However, the literature also points out that other forms of minority governments either quasi (or in the terms of Strøm (1990) formal) minority governments with permanent external support parties securing a majority - or flexible minority government, with the minority governments having a real possibility of finding ad hoc majorities, may function quite well. Even majority governments may be internally blocked for various reasons. If there are not enough opposition parties that suffice to a majority for a bill proposed by a minority government, the latter 3

will have to bargain to get an agreement. The aim of this paper is to show that such agreements provide a consensual manner to solve problems of dissent. Denmark is the country in Western Europe with the highest frequency of minority government, but it is also often characterised as a consensus democracy (Arter 2006; Lijphart 1984). Settlements or political accommodations between Danish minority governments and opposition parties follow Century old and highly institutionalised informal rules recognised by all parties of the Danish Parliament, the Folketing (Christiansen 2008; Christiansen & Damgaard 2008; Pedersen 2010). Political accommodations provide a manner for Danish minority governments and opposition parties to trade policy influence for procedural influence to opposition parties which become involved in the decision making process by a policy area, not only when the decision is decided but also in the longer run. Thereby, the accommodations become coalition governance institutions functionally equivalents to similar institution within government under majority rule. This paper will show that for laws passed as a consequence of political accommodations - in spite of these securing majorities and thereby broadening the support for the laws there will less consensual majorities, more political attention to the topic and more bloc politics involved. The paper tests these arguments with legislative data from the Folketing covering 2463 passed laws Theory Some issues are more consensual than others. Parliaments pass a number of laws that are either technical or regulate issues on which there is only very limited dissent, or where opposition parties do not choose to challenge the government. Some matters would be difficult to argue against in public for almost any party, and in most parliaments a certain share of laws is passed unanimously. 4

Therefore, we should expect that opposition parties out of its own free will support a number of the laws proposed by a government. There are other laws where we should expect an opposition party to be against a law proposed by the government, either it sincerely disagrees or because it finds tactical reasons to oppose. Strøm (1990) assumes that political parties compete over office, policy and votes. This means that, ultimately, all parties ultimately want office, even if they do not prioritise this goal in the short run. As a consequence, they have incentives to counteract the government and opposite its proposals to wear it out and perhaps topple it. In between mere accept and opposition, there is a room for bargaining and perhaps also agreements, especially relevant for a minority government, because it needs opposition support to pass its laws. If the government gives enough concessions, either through a compromise or through a log roll (also called a horse trade cf. Weingast & Marshall 1988), an opposition party may be tempted to enter into an agreement and give up its original intention not to support the law. From this follows the main hypothesis of this paper: H 1 : During minority government, laws following from bargained settlements are passed with less consensual support in parliament than other laws. Those issues that political parties prioritise and devote much attention to are also the one where they should be expected to compete with other parties. As a consequence we should expect laws passed on such issues to be less consensual than other issues: 5

H2: Those laws passed that political parties devote mostly attention to are passed with less consensual support in parliament than other laws. Case Selection and data To test these propositions, a case is needed with frequent minority governments and a possibility to identify legislative bargains. Denmark provides such a case. Since 1945, 26 of 30 governments in this established parliamentary democracy in Northern Europe has been of the minority type (Damgaard 2000, 234-235). Furthermore, in Denmark, a certain form of political agreement exists between government and opposition parties (Christiansen 2008). In these political accommodations, parties do not only decide to pass specific policies but they also commit themselves over time to maintain the agreement as long as it is valid, that is until an agreed date when it ends or until it is cancelled before a general election. The partners in an agreement are consulted about policies, and they also hold veto rights concerning changes to the agreement. Strøm (1990) characterised legislative accommodations as narrow in policy scope but durable over time. Upholding political agreements is considered the most important part of Danish parliamentary culture when members of the Folketing were asked (Jensen 1993). These agreements evolved from the end of the 19 th Century and most of the informal rules referred above were fully established around 1920 (Pedersen 2011). These days, all parties from time to time participate in legislative accommodations. So, we know that in Denmark there exists a highly complex and institutionalised form of bargained policy agreements between governments and oppositions parties. Furthermore, Christiansen (2008) has identified those laws passed after 1953, which were agreed to by a political accommodation. Finally, Denmark has not experienced constitutional changes or other major upheavals of its political system. Therefore, Denmark is suitable case to test the propositions. 6

Denmark is a parliamentary democracy and a member of the European Union since 1973. Like the other Scandinavian countries, Sweden and Norway, it has been characterised as a consensus democracy but often combined with a bloc politic dynamic in the sense that two groups of parties compete for public office, although, especially in Sweden and Norway, the Social Democratic parties held a predominant position within the party systems (Arter 2006; Christiansen & Damgaard 2008). Furthermore, minority governments are common. Party cohesion is very high (Skjæveland 2001). To test the proposition, a data set has been collected with roll call data from four recent governmental periods: 1993-94, 1998-2001, 2001-2005 and 2005-2007. In all of them, Denmark had coalition government, 1993-2001 headed by the Social Democrats and after 2001 by the Liberals (Venstre), but only from 1993 to 1994 the government has a majority of its own, and therefore this case is singled out. The composition of the Folketing for each period is found in Table 1. 2463 passed laws have been coded. Only passed laws are considered because political accommodations usually manifest themselves in this manner. For the same periods I have coded all written political agreements, remarks to and parliamentary readings of the laws. Furthermore, the length of the debated in parliament has been coded as a proxy for the combined political saliency of a law to the political parties. 1 The idea is that the more they say in parliament, the more important they find the law. TABLE 1 about here. 1 I would like to thank Christoffer Green Pedersen for access to his data set on debate lengths. 7

To be able to control for other possible explanations, the EU status, majority status, ideological colour, policy area and place in the legislative cycle are also tested. First, EU is known to be an issue that divides the political parties differently than is usually the case for Danish legislative coalitions (Christiansen and Pedersen 2011). That is especially case when it comes to laws that directly implement EU decisions. Second, it could have an impact whether there is a majority government, because it has majority without opposition parties. Third, there are no theoretical reasons to expect the ideological colour of the government to have a systematic effect but it is worthwhile controlling for it. Fourth, we know from other studies that there may be cyclical factors over an electoral period so that especially when a government enters office it will try to change a number of matters and care less about broad majorities. To deal with this problem, I have divided the material into which legislative session that a law is passed in. Fifth, policy areas may make a difference with regard to consensus or dissent. We should expect some areas have more consensus than other because of varying competitive goals associated with for political parties (cf. Strøm 1990). The division follows Christiansen (2008), who in turn follows quite common standards in the Danish case. The dependant variable, consensus, is operationalised in three different manners. First, following the tradition of Pedersen (1967) by looking at which laws are passed with the accept of all of the four old parties, the ones with roots before 1920: Liberals, Social Democrats, Conservatives and Social Liberals. Second, by looking at pure bloc majorities, that is when laws have been passed with the sole support of the internal or external support parties of the government. Third, by looking at unanimously passed legislation. During the research period, the first of these variables were on 72 percent, the second one 88 percent, and the third on 33 percent. These levels seem to be pretty high. 8

The voting pattern of the political parties in the Folketing during the research period is shown in table 2. The parties are listed in their approximate ideological order from left to right. It shows that government parties participate in the passing of almost all laws. The table also shows a regular ideological pattern in the votes. The further away a party is on the left-right scale from the government, the more likely it is not to vote in favour of the government. The moderate centrist, centre-right, and centre-left parties vote in favour of much more than half of the legislation even when in opposition. Strong opposition is only found on the fringes of the party system. One interesting change happens after 2001 when the Danish People s Party although a party on the outer wing of the party system become a permanent external support party for the Liberal- Conservative government and begin to vote in favour of most of the passed legislation. Table 2 about here. The final column in table 2 informs about the share of laws that had been agreed to in a legislative accommodation. The majority government from 1993 to 1994 clearly did not use this tool very much whereas the minority governments make agreements for more than a fourth of the laws. Analysis Table 3 presents the three measures for consensus and describes them statistically for each government period and for each of the theoretical explanatory variables introduced above. The level of consensus has been quite stable over the four governmental periods, regardless of how we measure it. However, it does seem to make a clear difference, whether a law has been agreed to in a 9

legislative accommodation. This indicates, in accordance with H 1 that these political settlements are associated with situations of dissent rather than consensus. Similarly, the length of debate for each bill indicates that there are longer debates for non-consensual issues. If debate length is used as a proxy for the saliency of issues, it means, in accordance with H 2, that the topics receiving most attention are less consensual. Table 3 about here. Looking at the control variables we do find a cyclical pattern in the material, especially in the first election session after a government takes over. Furthermore, EU implementation seems to promote consensus. Certain policy areas have high levels of consensus like foreign affairs, transport, business and justice and crime, whereas other policy areas take much lower values, especially taxation. The ideological leadership does not seem to make a difference, but the majority government 1993 to 1994 did to certain extend pass legislation with a pure block majority. Having established these relations in descriptive terms, the paper now proceeds with a test of the robustness of the findings. This is done with logistic regressions. Table 4 about here. 10

Table 4 test the explanations for consensus as measured by the four old parties. It does so in two models, one without the policy areas to reduce the complexity. In both models the two variables for party competition, namely legislative accommodations and lengths of debates are strongly significant and have the expected direction. But EU implementation and early session after a government comes to office also has an effect. Adding policy areas increases the explanatory power of the model but the other independent variables remain significant. The same basic results are found if consensus is measured as a law passed by a pure bloc majority like it is done in table 5. Table 5 about here. Conclusion During minority government, opposition parties become involved in legislation due to the mere necessity that they need to find a majority to pass laws and other decisions. In the parliamentary democracy with the most minority governments, Denmark, studied in this paper, one of the most important ways to involve opposition parties is through so-called legislative accommodations, which are highly institutionalised arrangements for exchange of support and procedural rights between Danish minority governments and opposition parties. Using roll-call data this paper seems to confirm that legislative accommodation are associated with dissent. It is a way to create consensus in spite of dissent. Legislative accommodations should be understood as informal institutional mechanisms of coalition governance relevant for minority government rule, because they support political parties in the interest of working together despite differences and competition 11

References Arter, D. (2006). Democracy in Scandinavia. Consensual, majoritarian or mixed? Manchester: Manchester University Press Christiansen, F. J. (2008) Politiske forlig i Folketinget. Århus: Forlaget Politica. Christiansen F. J. & Damgaard E. (2008) Parliamentary Opposition under Minority Parliamentarism: Scandinavia, Journal of legislative Studies 14(1): 46-76. Christiansen, F. J. and H. H. Pedersen (2011) Minority Coalition in Denmark, ECPR General Conference, August 25-27, 2011 Reykjavik, Iceland. Christiansen, F. J. og R. B. (2011) Europeanization and Coalition Management in Minority Parliamentary Systems in M. Wiberg and T. Persson (eds)., Government in the Nordic Countries at a Crossroads : Coping with Challenges from Europeanisation and Presidentialisation Stockholm : Santérus Förlag. Damgaard, E. (2000). Denmark: The Life and Death of Government Coalitions in W. C. Müller & K. Strøm (eds.), Coalition Governments in Western Europe, Oxford: Oxford University Press: 231-64. Green-Pedersen, C. (2001) Minority Governments and Party Politics: The Political and Institutional Background to the Danish Miracle, Journal of Public Policy, 21 (1), 53-70. Green-Pedersen, C. & L. H. Thomsen (2005) Bloc Politics vs. Broad Cooperation? The Functioning of Danish Minority Parliamentarism, The Journal of Legislative Studies, 11 (2): 153-169. Jensen, T. K. (1993) Politik i prâxis. Aspekter af danske folketingsmedlemmers kultur og livsverden (Frederiksberg: Samfundslitteratur). Lijphart, A. (1984) Democracies. Patterns of Majoritarian and Consensus Government in Twenty- One Countries (New Haven: Yale University Press). 12

Moury, C. (2010) Coalition agreement and party mandate: How coalition agreements constrain the ministers, Party Politics, accessed at: http://ppg.sagepub.com/content/early /2010/07/28/1354068810372099 Pedersen, H. H. (2011). Etableringen af politiske forlig som parlamentarisk praksis, Politica, 43 (1): 48-67. Pedersen, M. N. (1967) Partiernes holdning ved vedtagelsen af regeringens lovforslag 1945-66, Historie Jyske Samlinger, Ny række (VII), 404-435. Skjæveland, A. (2001) Party Cohesion in the Danish Parliament, The Journal of Legislative Studies, 7 (2), 35-56. Strøm, K. (1990). Minority Government and Majority Rule. Cambridge: Cambridge University Press. Strøm, K., W. C. Müller & T. Bergman (eds.) (2008) Cabinets and Coalition Bargaining. The Democratic Life Cycle in Western Europe. Oxford: Oxford University Press. Weingast, B. R. & W. I. Marshall (1988) The Industrial Organization of Congress, or, Why Legislatures, Like Firms, Are not Organized as Markets, Journal of Political Economy, 96 (1), 132-163. 13

Table 1: Composition of the Folketing for four government periods 1993-2007 1993-94 1998-2001 2001-2005 2005-2007 Red Green Alliance Na 5 4 6 Socialist People s Party 15 13 12 11 Social Democrats 71 63 52 47 Social Liberals 7 7 9 17 Centre Democrats 8 8 Na Na Christian People s Party 4 4 4 Na Conservative People s Party 31 16 16 18 Liberals (Venstre) 31 43 57 52 Danish People s Party Na 13 22 24 Progress Party 12 4 Na Na North Atlantic seats 0 3 (2 supports 3 3 unaffiliated to Danish parties the gov.) Total 179 179 179 179 Government parties. Parliamentary bloc parties. 14

Table 2: Yes votes to passed laws for four government periods in the Folketing 1993-2007. Percentages. RGA SPP SD SL CD CPP CON LIB DPP PP L.A. 1993-1994 na 83.8 100.0 100.0 99.8 99.5 69.5 68.8 Na 48.5 8.0 1998-2001 64.0 83.5 100.0 100.0 85.3 78.7 76.8 76.6 60.8 60.8 30.3 2001-2005 38.8 56.9 73.9 74.6 Na 80.4 99.9 99.9 94.3 Na 27.7 2005-2007 41.1 63.9 81.0 78.9 Na Na 99.7 99.7 95.0 Na 39.6 Sources: Christiansen & R. B. Pedersen (2011); Christiansen & H. H. Pedersen (2011). L. A.: Laws agreed to in legislative accommodations. Italics: Government parties. 15

Table 3: Consensus levels for passed laws during four government periods in the Folketing 1993-2007. Percentages. Four old parties No pure bloc majority Unanimity Number of laws 1993-1994 68.0 75.2 36.2 400 1998-2001 75.3 92.8 28.0 789 2001-2005 68.6 91.3 32.7 678 2005-2007 74.3 87.8 36.6 596 Bargained settlement laws 58.1 83.4 18.0 695 Non-bargained settled laws 77.5 90.2 38.5 1768 Average length of debate 18.9 20.3 14.2 Na with consensus (columns) Non-do. (columns) 27.3 28.1 24.5 Na First legislative session 65.1 77.9 27.6 602 Second legislative session 76.5 90.6 35.8 865 Third legislative Session 72.5 93.1 33.0 697 Fourth legislative Session 71.8 91.3 33.6 298 Majority government 68.0 75.2 36.2 400 Minority government 72.8 90.8 32.0 2063 Social Democratic govt 72.8 86.9 30.8 1189 Liberal government 71.3 89.6 34.5 1274 EU implementation 89.5 96.9 36.9 295 Laws of national origin 70.0 89.8 31.2 1768 Agriculture and Fisheries 76.7 87.2 23.3 86 Business 82.5 93.4 38.4 229 Church 85.7 100.0 35.7 14 Culture 68.8 84.4 25.0 64 Defence 88.2 94.1 11.8 17 Education 70.5 90.6 32.9 234 Energy 76.8 85.7 26.8 15 Environment 67.6 89.8 29.6 108 Foreign Affairs and EU 95.5 95.5 40.9 22 Housing 70.9 91.3 32.0 33 Immigration 72.8 83.7 15.2 92 Justice and Crime 86.7 96.7 44.5 211 Labour Market 61.6 88.6 33.2 227 Municipalities and Health 71.4 89.4 43.2 98 Public Finances 88.9 87.5 36.1 72 Social Affairs 66.7 88.9 35.6 180 Taxation 57.5 76.4 25.9 402 Transport 85.3 95.4 25.7 109 Other policies 92.0 96.0 44.0 25 Total 72.0 88.3 32.7 2463 Sources: Calculated from data described in: Christiansen & R. B. Pedersen (2011) 16

Table 4: Logistic regression of four old parties voting yes to passed laws for four government periods 1993-2007 in the Folketing Model 1 Model 2 B SD Sig B SD Sig Legislative -1.107 0.117 0.000-0.960 0.127 0.000 accommodation EU Implementation 1.084 0.228 0.000 1.047 0.236 0.000 Length of law -0.006 0.002 0.001-0.016 0.003 0.000 First legislative session -0.443 0.151 0.003-0.447 0.159 0.005 Second legislative 0.280 0.149 0.061 0.326 0.156 0.036 session Fourth legislative session -0.269 0.188 0.153-0.275 0.196 0.161 Agriculture and Fisheries -1.797 0.835 0.031 Business -1.564 0.805 0.052 Culture -1.997 0.839 0.017 Education -1.714 0.793 0.031 Energy -1.517 0.877 0.084 Environment -2.132 0.815 0.009 Foreign Affairs and EU 0.505 0.280 0.071 Housing -1.107 0.853 0.194 Immigration -1.457 0.830 0.079 Justice and Crime -1.219 0.804 0.129 Labour Market -2.241 0.793 0.005 Municipalities and -1.763 0.794 0.026 Health Social Affairs -1.749 0.802 0.029 Taxation -2.494 0.785 0.001 Transport -0.779 0.844 0.356 Other policies 0.172 0.263 0.513 Constant 1.528 0.125 0.000 3.401 0.790 0.000 Nagelkerke R Sq. 0.144 0.234 N 2463 2463 Source: Christiansen & Pedersen (2011). 17

Table 5: Logistic regression of laws not only passed in pure bloc vote for four government periods 1993-2007 in the Folketing Model 1 Model 2 B SD Sig B SD Sig Legislative -1.202 0.179 0.000-1.057 0.193 0.000 accommodation EU Implementation 1.037 0.404 0.010 1.097 0.420 0.009 Length of law -0.007 0.002 0.000-0.016 0.003 0.000 First legislative session -1.298 0.224 0.000-1.370 0.239 0.000 Second legislative 0.337 0.268 0.210 0.365 0.276 0.186 session Fourth legislative session -0.357 0.311 0.251-0.293 0.323 0.364 Agriculture and Fisheries -3.721 2.090 0.075 Business -3.003 2.095 0.152 Culture -2.920 2.115 0.167 Education -2.509 2.081 0.228 Energy -3.003 2.095 0.090 Environment -3.273 2.087 0.117 Foreign Affairs and EU 0.994 0.554 0.073 Housing -0.911 2.288 0.691 Immigration -2.316 2.090 0.268 Justice and Crime -1.561 2.134 0.464 Labour Market -3.215 2.074 0.040 Municipalities and -2.490 2.073 0.230 Health Social Affairs -2.728 2.083 0.190 Taxation -4.058 2.076 0.051 Transport -1.392 2.188 0.525 Other policies 0.364 0.464 0.433 Constant 3.314 0.219 0.000 6.419 2.087 0.002 Nagelkerke R Sq. 0.165 0.280 N 2463 2463 Source: Christiansen & Pedersen (2011). 18