THE RATIONALITY OF MINORITY
GOVERNMENTS
A STUDY INTO THE FUNCTIONING OF SUPPORTED MINORITY CABINETS
MS
CT
HESISD
AANH
ENDRIKSON,
S2265427
11/6/2019
SUMMARY
This study seeks to understand parliamentary behaviour from the perspectives of Rational Choice Institutionalism (RCI) and Normative Institutionalism (NI) on the workings of supported minority cabinets. These theories both have explanative value in coalition/opposition cooperation of supported minority cabinets. However, the explanations are conflicting. Normative Institutionalism assumes that parties will act based on the existing norms and that in a system that is used to majority cabinets, even a supported minority cabinet will act as a majority cabinet. Rational Choice Institutionalism suggests that parties act to achieve their preferences, in this case, policy-change. According to this theory, parties would decide to support based on their preferences, and this would create a situation where parties could act counter to the norms of a majority cabinet during a supported minority cabinet.
In the Netherlands, a supported minority cabinet was recently created, and this was the first non-caretaker (supported) minority cabinet since 1922. This cabinet consisted of VVD and CDA. The support party was PVV. Primarily because of the Euroscepticism of the PVV, the transposition of European Union-related legislation was used as a possible dividing issue between the VVD and the CDA and the PVV. If the PVV were to vote differently on proposals that are EU-related than the cabinet parties, that would provide support for the perspective of RCI above the perspective of NI. Consequently, if this is confirmed, the opposition parties would then be needed to create legislative majorities for EU-related legislation. To test the explanative value of the theories of RCI and NI this study sought to answer two hypotheses; the first being that the support party supports more often other legislation than EU-related
legislation and the second being that the opposition parties support less often other legislation than EU-related legislation.
This study found that the supporting party in this case study, the PVV, seems to be able to vote on issues rationally, thereby, creating a divide with the minority cabinet on issues as European Union-related legislation. Furthermore, there seems to be no strong relationship between the support agreement and the voting behaviour of the supporting party, further confirming that parties can act rationally. This finding would create the need for the minority cabinet to rely on ad hoc majorities to pass legislation. However, the results do not show a significant difference between the opposition voting on EU-related legislation and other legislation. These results do indicate that there is a need for ad hoc support; however, further research is needed in order to clarify the unclearness that is present around the voting behaviour of the opposition parties.
INTRODUCTION
In recent history, there has been an increase in supported minority cabinets (Otjes & Louwerse, 2014). However, there are widespread arguments over the extent that parties cooperate across coalition-opposition lines during supported minority cabinets. There is no consensus whether minority cabinets work as if they are majority cabinets in disguise or as minority cabinets, which seek ad hoc majorities on the issues that fall outside of the support agreement party (Christiansen & Pedersen, 2012; Strøm, 1990). It could be argued that it would have a positive effect if coalitions worked in varying ad hoc coalitions. This could create a more distant relationship between the government and the Parliament, which could make it better in controlling the government, as was recently advocated by the Dutch parliamentary committee Remkes, which looked at the functioning of the parliament (Staatscommissie Parlementair Stelsel, 2018).
In 2010, a supported minority cabinet was created in the Netherlands for the first time since 1922. This minority cabinet can be useful in the understanding of how political parties operate during a supported minority government when a majority cabinet has been the norm (Andeweg, 2011). The supported minority cabinet existed out of right-wing parties CDA and VVD with the radical right-wing party PVV as the support party. PVV chose to be a support party instead of a cabinet party, due to the differences of opinion on issues, as for example the European Union and the Islam. The issues of support were expressed in a support agreement (Gedoogakkoord VVD-PVV-CDA, 2010). A study of Otjes and Louwerse suggests that this supported minority government has worked as a majority cabinet in disguise (Otjes & Louwerse, 2014).
The claim that supported minority cabinets essentially work as conventional majority cabinets is based on Normative Institutionalism. This theory suggests that the parties act based on norms and that a new cabinet form does, therefore not influence legislative behaviour. Otjes and Louwerse distinguished three broad groups of legislation in their study:
included in support agreement, a grey area and excluded from the support agreement. Based
on Normative institutionalism, there would likely be no significant variation of coalition-opposition cooperation within these groups, and therefore, the results would indicate that this cabinet worked similarly to a majority cabinet. However, if parties act based on their preferences and act rationally as Rational Choice Institutionalism suggests, differences could be present per issue, which would go against the norm of coalition-opposition voting during majority cabinets.
These differences are unlikely to be found when distinguishing all legislation into these large three groups that are based on agreements. The study of Otjes & Louwerse chose to identify the issues based on the support agreement, thereby, creating a self-fulfilling prophecy, by neglecting to look at specific issues that could divide the minority cabinet and the support party and find differences in coalition-opposition voting on these issues. This study will seek to fill this gap in order to reassess the current understanding of the workings of supported minority cabinets and to further understand the debate between the theories of Rational Choice Institutionalism and Normative Institutionalism by answering the research question: To what extent do supported minority cabinets rely on ad hoc support to implement
legislation?
The case of the Netherlands is an extreme case to research this clash between the two theories because there has not been any minority cabinet since 1922, except for caretaker cabinets. If parties would act based on the norms, as normative institutionalism suggests, it
would arguably be most visible immediately after that change, as with time, norms tend to fade. In contrast, RCI predicts that preferences are leading in voting, so differences are present where policy present are divergent. This would indicate that parties can act rationally, which would mean that there would at least be a need for ad hoc support on some issues to reach majority support.
THEORETICAL FRAMEWORK
In recent history, there has been a rise of (supported) minority cabinets throughout the world (Otjes & Louwerse, 2014, p. 2). However, it is unclear among scholars how supported minority cabinets function. Some studies argue that supported minority cabinets tend to work as they are majority cabinets, by finding a stable support party that supports the minority cabinet as though they are part of the coalition (Bale & Bergman, 2006; Daalder, 1971).
However, other studies suggest that supported minority cabinets can work with varying ad hoc coalitions to create legislative majorities, as shown in Denmark and Spain (Christiansen & Pedersen, 2012; Green-Pedersen, 2001; Field, 2009). What is clear, is that supported minority cabinets cannot be as quickly defined as majority cabinets and minority cabinets, where the former has a coalition agreement on all issues, and the latter has no formal support partners (Field, 2009, p. 418). In understanding supported minority cabinets, many terms have been used: majority party in disguise (Strøm, 1990), but also contract parliamentarism (Bale & Bergman, 2006, p.422). The first proposes that supported minority cabinets will work as majority cabinets and the second meaning a relationship between coalition-opposition that is less rigid and where continuous negotiation between the cabinet and the parliament is needed, but where in the end the same legislative behaviour occurs (Otjes & Louwerse, 2014, p. 5; Bale & Bergman, 2006). The third classification by Christiansen and Pedersen does see supported minority cabinets as capable of acting different from a majority cabinet, by seeking ad hoc minorities on the issues that fall outside of the support agreement (2012).
In the Netherlands, the first minority cabinet since 1922 was created in 2010. This supported minority cabinet was studied by Otjes and Louwerse, and they stated that “The
data presented show that patterns of conflict and cooperation between coalition and opposition during the Rutte I cabinet are quite similar to the cooperation during a majority coalition.” (2014, p. 17) This indicates support for the idea that supported minority cabinets work as if they are majority cabinets. However, the study does point out that the focus on general patterns of legislative behaviour might exclude specific cooperation on issues between the opposition and coalition parties (Otjes & Louwerse, 2014, p. 14). This can be best shown in the decision that was made to divide the legislation into included in the support agreement, a grey area, and excluded from the support agreement.
In this manner, the study excludes studying differences in specific policy preferences, which could show the existence of varying ad hoc coalitions. The choice to have a more general focus is based on the theory of Normative Institutionalism. According to this theory, parties will act based on the established norms and will act based on those norms even after a change has occurred, in this instance a change in cabinet formation (Weber, 1922; Best & Vogel, 2014). It is understandable when basing a study on the theory of Normative Institutionalism , that it is decided to divided legislation based on the mentioned three groups, as they are structured around the general agreements that are used by majority cabinets. In this way, Otjes and Louwerse studied whether each of these categories showed if the support party worked as a member of a coalition, and they found that it mainly did. However, it could be argued that these categories are too broad as they are based on the full support agreement. This could lead to policies that the minority cabinet and the support party disagree on being pushed together. In this way, the differences that can occur between those policies would be squeezed out of the results. In this way, it seems to increase the chance of the research design creating the results, which the theory predicted. This could, therefore, lead to a self-fulfilling prophecy.
Therefore, it is vital to not only look at one explanation but look at other perspectives as well, as for example, Rational Choice Institutionalism. This theory sees parties as unitary actors that focus on achieving their preferences (Strøm, 1990, p. 566). These preferences can, in short, be expressed by the terms, office seeking, policy seeking, and vote-seeking or a combination of these (Strøm & Müller, 2010, p. 5; Strøm, 1990, p. 566). Starting with the work of Riker (1962), there was a focus within coalition theory on office-seeking behaviour of political parties. However, this focus seemed unable to explain the rise of minority cabinets throughout the world because if office seeking is seen as rational it would be logical that all parties would like to be part of the cabinet and that would make a majority cabinet likely (Herman & Pope, 1973; Otjes & Louwerse, 2014, p. 2). However, that does not happen in an increasing number of cases (Strøm, 1990, p. 199). To explain why this does not happen, it would be logical to assume that other preferences are predominant. Budge and Herman, argue therefore that there is a predominance of policy concerns above office-seeking (Budge & Herman, 1978; Budge & Laver, 1986, p. 486). This would mean that political parties value enacting their policy preferences above office-seeking.
It can be argued that grouping varying policies together as one group can lead to missing differences between those groups if political parties would act rationally and therefore seek policies according to their preferences. Each party will seek to achieve its policy preferences per issue. Therefore, it would be logical that in the case of a supported minority cabinet there are policy proposals where the supporting party chooses not to support measures that go against their preferences and vice versa for opposition parties. They likely choose to support some measures that support their policy preferences. This creates the possibility of varying ad hoc coalitions when a supported minority cabinet is divided into one or more issues. When there is a division on an issue, the minority cabinet will seek
opposition support, and if it is within their preferences, the opposition will likely support the measures. This would go against the norms of coalition-opposition voting.
CASE STUDY
To further study how supported minority cabinets work; this study chose to use the Netherlands as a case study. In October 2010 a supported minority cabinet was formed there. It was the first (supported) minority cabinet since 1922, except for caretaker cabinets. In comparison with other Western European Countries, the Netherlands waited for the longest to form a minority cabinet after the Second World War, as shown in figure 1, Moreover, the Netherlands is also the only country, from the countries which had minority cabinets, which only has had one minority cabinet. This distinctiveness makes the Netherlands an extreme case study in the typology of Seawright and Gerring (2008, p. 301).
Countries Duration before
minority cabinet in years
Start First Minority cabinet
Austria 24 04/12/1945 21/04/1970 Belgium 8 13/03/1946 22/05/1954 Denmark 0 30/10/1945 30/10/1945 Finland 3 18/03/1945 29/07/1948 France 9 22/01/1947 31/01/1956 Germany NA 20/09/1949 NA Iceland 3 30/06/1946 06/12/1949 Ireland 0 18/02/1948 18/02/1948
Italy 0 02/02/1947 02/02/1947 Luxembourg 0 01/03/1947 01/03/1947 Malta 3 27/10/1947 04/09/1950 Netherlands 64 03/07/1946 14/10/2010 Norway 16 05/11/1945 11/09/1961 Sweden 0 19/11/1948 19/11/1948 United Kingdom 29 26/07/1945 04/03/1974
FIGURE 1– MINORITY CABINETS IN WESTERN EUROPEAN COUNTRIES
The supported minority cabinet of Rutte I existed of the parties VVD and CDA and the PVV, as support party. All of these parties are on the right of the ideological left-right dimension, as can be seen in figure 2 (Polk, et al., 2017). However, not on all issues the supported minority cabinet parties have similar views when looking at the data of the Chapel Hill Expert Survey, there is a clear difference in the position of the parties on the European Union. The PVV has a negative view of the EU, whereas the VVD and CDA score much higher.
As shown in figure 2, the opposition parties, PVDA, D66, and GL seem to have more similar positions on the EU compared with minority cabinet parties than the PVV.
Party EU Position Left-Right Dimension CDA 5.55 6.78 PvdA 5.45 3.67 VVD 5.18 7.89 D66 6.82 5.56 GL 6.55 2.33 SGP 2.56 8.11 SP 2.10 1.00 CU 3.44 5.44 PVV 1.09 9.25 PvdD 3.71 2.89 50PLUS 4.80 5.25
FIGURE 2–DUTCH POLITICAL PARTIES POSITIONS ON THE LEFT-RIGHT DIMENSION AND THE EU (POLK, ET AL.,
2017)
Because of the difference of position of the minority cabinet parties and the support party on the issue of the European Union, this issue has the potential to divide the supported minority cabinet. It would be rational for the PVV not to support EU-related legislation that is proposed by the VVD and the CDA as they have different views on the issue. Because of this divide in opinion on the European Union and the possibility that it may divide the supported minority cabinet EU-related legislation is chosen as a test variable to see if the support party can act rationally and choose to vote based on their preferences or the norms of a majority cabinet. If the support party can act unbounded by the structure of the support agreement and the norms of a majority cabinet, it would create a need for the opposition parties to help
create majority support for EU-related legislation. In this situation, where the support party does not support EU-legislation, it would be rational for these opposition parties to support this sort of legislation as they have similar positions on this topic to the minority cabinet. If this happened even though they disagree on other issues as shown in the left-right dimension, this would indicate that opposition parties can also act rationally. Moreover, the European Union is not only a dividing issue in the Netherlands; it is throughout Europe, which is another reason to look at EU-related legislation in this study (de Vries & Marks, 2012; Grande, Kriesi, Lachat, Dolezal, & Timotheos, 2008).
In order to analyse the voting behaviour of the Dutch political parties, this study will define and quantify European Union-related legislation based on the methods of measuring Europeanisation from earlier studies (Bovens & Yesilkagit, 2010; Page, 1997; Töller A. E., 2008; Töller A. E., 2010). In this way, two separate groups of legislation will be created, EU-related legislation and other legislation. This variable will then be used to assess the voting behaviour of both the supporting party and the opposition parties by contrasting EU-related legislation and other legislation. Based on the theoretical framework, this study expects the following outcomes of the analysis:
The Support party hypothesis: The support party is less likely to support bills that implement European Union-related bills than other bills
If this hypothesis is confirmed, it would show support for the Rational Choice Institutionalism perspective. As this would show that the support party can follow its preferences and vote outside of the norms of coalition agreements. However, if the hypothesis is rejected, as the NI perspective suggests, it indicates that the support party behaves similarly as a coalition party even though their preferences might differ from the other parties. If the RCI perspective is confirmed, it would increase the possibility of
opposition parties also being able to vote based on their preferences. Furthermore, it would be needed for opposition parties to support legislation of the minority cabinet as the support party does not on certain issue. Therefore, this research expects that:
The Opposition parties hypothesis: The opposition(s) parties are more likely to support bills that implement European Union-related bills than other bills.
If the hypothesis is supported, the RCI perspective is further confirmed. If the opposition parties will support more EU-related bills than other bills, this supports the RCI perspective as they vote in contrast with the expectation of the NI perspective. If this hypothesis is rejected, this would show support for the NI perspective, when the opposition parties choose to oppose, no matter their preferences.
To further study if the theory of Rational Choice Institutionalism has more explanative value than Normative Institutionalism, it is not enough to only test if parties can act rationally as RCI would expect, but also that parties do not act based on norms, which would contrast with NI. Therefore, the support agreement variable will be used as a control variable. In this way, this study seeks to check the explanative value of the NI theory. Finally, The influence of ‘Eerste Kamer’/Senate committees and if the legislation is a vote on a budget will be used as control variables as proposed by Christiansen and Pedersen (2012, p. 194). In this way, this research seeks to establish if committees and budget legislation influence the voting behaviour of the support party and the opposition parties.
RESEARCH DESIGN
In order to test the stated hypotheses, this research will seek to analyse voting data from the Parliamentary Behaviour Dataset from the Netherlands during the Rutte I cabinets. The dataset consists of all adopted bills, motions, and amendments during Rutte I (Otjes & Louwerse, 2014). The data will be further divided into EU and non-EU related legislation. As EU-related legislation is expected to be an issue that will create the need for ad hoc coalitions to pass legislation.
To analyse the influence of EU-legislation, this first must be operationalised into quantitative data. This study will seek to do so by going further than the studies of Bovens & Yesilkagit and Christensen that solely focus on the implementation of European directives as EU-related (2010; 2010). Expanding this search is important as not all legal acts are transposed in that manner (Steunenberg & Toshkov, 2009). Therefore, this study will adopt the broad definition coined by Töller of European Impulses, which also contains EU-regulations and decisions. The method of creating this EU-variable will be based on the research of Töller and the Pilot Monitor of the Asser Instituut, where keywords are used to establish what is EU-related and what is not (Töller, 2010; Douma, Pieters, Feenstra, & Koch-Hartmanová, n.d.). The following keywords were used to identify EU-related legislation1:
EG (European Community) EU Europees/europees/Europa/Europa/Europese/europese Verordeningen/verordeningen (Regulation) Richtlijnen/richtlijnen (Directives)
A keyword search was done in the full proposal as can be found on the ‘Eerste Kamer’ website of all legislation during the cabinet of Rutte I (Eerste Kamer der Staten-Generaal, n.d.). The legislation that included one or more of the keywords in the proposal will henceforward be seen as EU-legislation, and consequently, the other legislation will be seen as other legislation within this study. This distinction between EU-legislation and non-EU legislation will then be used to study the influence of this variable on the voting behaviour during a supported minority cabinet, as this is an issue that could divide the minority cabinet and the support party.
The voting behaviour of all parties2 within the Parliament during the cabinet of Rutte I will be used to analyse the divide between EU Legislation and other legislation. This will be done by using the Parliamentary Behaviour Dataset (Louwerse, Otjes, & van Vonno, 2018). The Dataset gives votes in favour a 1 and votes against a 0. The distinction of EU-related and other legislation will be used to understand the difference of voting behaviour between EU legislation and other legislation by using the average of these votes per party. The statistical significance of the differences between EU-related and other legislation will then be tested by a T-test in SPSS.
Furthermore, there will be multivariate tests. These will assess the influence and significance of the Support, Opposition, and EU variables. Furthermore, the control variables budget, support agreement, and the Senate commission shall also be taken into account in this analysis. Moreover, intervariate variables of the EU and support/opposition and the support agreement and support opposition variables will be included in a sequent analysis, to check for interaction effects between these variables. Then there will be a comparison
between the voting behaviour of the support and opposition variables when legislation is EU-related and when it is not, in order to test the hypotheses.
Furthermore, there will be a similar comparison between the voting behaviour of the support and opposition variables when legislation is part of the support agreement and when it is not. These comparisons are done to test for the two theories. The first seeks to understand the explanative value of the RCI perspective and if there is support for the idea that political parties can act rationally. The second seeks to understand the explanative value of the NI perspective and if there is support for the idea that political parties act based on norms, as quantified in the support agreement variable. In this way, this study seeks to contribute to the ongoing debate on the functioning of supported minority cabinets.
RESULTS
In order to test the Support and Opposition hypotheses multiple tests were run to determine the differences between voting behaviour on EU-legislation and other legislation, intervariate variables and control variables and the significance of those differences. First, a T-Test was run to determine the significance of the differences between EU-legislation and other legislative votes per party. After that, multivariate tests were run to further understand the differences in legislative behaviour between EU-related legislation and other legislation per party when controlled for other factors; these were two binominal logistic regressions. The first logistic regression included the variables of EU, Support party, Opposition parties, Budget, Support Agreement and Commissions. The second logistic regression checked for intervariation effects by creating interaction variables of EU and Support party and EU and Opposition parties, and Support Agreement and Support party and Support Agreement and Opposition parties, while also considering the variables of the EU, Support party, Opposition parties, Budget, Support Agreement and Commissions. The dataset consists of 585 legislative votes in the Second Chamber of Parliament of proposals that were adopted, from these 97 and 16,6% of the 585 proposals were found to be European Union-related based on a keyword search.
The voting records of all parties were evaluated based on their voting record of EU-related and other legislation as can be seen in figure 3. This figure shows the level of support for legislation per party in percentages. EU-related legislation is on the left, and other forms of legislation are on the right, and the margin of error is displayed. The largest difference is visible for the Support party, the PVV. They supported 95.5 per cent of other legislation and 83.5 per cent of EU-related legislation, a difference of 12.0 per cent. Of the opposition parties,
only the SGP and CU support more other legislation than EU-related legislation. Whereas the PVDA, GL, D66, SP, and PvdD all support EU related legislation more than other legislation. The parties of the minority cabinet, VVD and CDA, support 100% of EU-related proposals and on other proposals 98 and 99%, respectively.
Figure 3 – Voting Behaviour during Rutte I
An Independent-Samples T-Test was run in SPSS to determine if the differences that were found were statistically significant. First, to determine if the condition of homogeneity of variances was met, a Levene test was run. PvdA, CDA, PVV, SGP, and VVD violated this assumption. When this assumption was violated, a Welch T-test was run, where equal variances were not assumed. The other parties did meet the assumption of homogeneity, and therefore a T-test was run on those.
The test showed that the CDA, a party of the minority cabinet, had higher support for EU-related legislation (M=1) than other legislation (M=0,99), a statistically significant difference, M=0.01, 95% CI [0,00, 0,02], t(487.000)=2.245, p=0,025. Furthermore, the PVV, the Support party, had higher support for other legislation (M=0,955) than EU-legislation (M=0,84), a statistically significant difference, M0,12, 95% CI [-0,20, -0,04], t (108.113) =-3.071, p=0,003. Lastly, the VVD, a party of the minority cabinet, had a higher support for EU-related legislation (M=1) than other legislation (M=0,984), a statistically significant difference,
0 0,2 0,4 0,6 0,8 1 SP PvdD PvdA GL D66 VVD SGP CU CDA PVV Ye s v o te s i n % Parties EU*R-related Other
M=0,02, 95% CI [0,01, 0,028], t(487.000)=2.85, p=0,005. The other differences were not significant as can be seen in figure 4.
Party EU Non-EU Difference
SP 0.88 0.87 0.01 PVDD 0.87 0.85 0.02 PVDA 0.96 0.92 0.04 GL 0.95 0.93 0.02 D66 0.96 0.95 0.01 VVD 1 0.98 0.02** SGP 0.95 0.98 -0.03 CU 0.93 0.94 -0.01 CDA 1 0.99 0.01* PVV 0.84 0.95 -0.12**
FIGURE 4–T-TEST,SIGNIFICANCE LEVELS:***=<0.001,**=<0.01,*=<0.05.(*)=<0.1
In order to further understand the differences that were found, a multivariate test was run. Due to the VVD and the CDA not being part of the hypotheses as minority cabinet parties, they were left out of further tests. Moreover, because of their hundred per cent support on EU-related legislation, including them, could lead to unreliable estimates. Furthermore, all opposition parties were put together to form the opposition variable and the PVV; the support party became the support variable. As VVD and CDA are not included in the dataset anymore, the only parties left in the dataset are either support or opposition, which makes these each other’s inverse.
A binomial logistic regression was performed to determine the effects of the EU, Support, Opposition, Support Agreement, Budget, and the Senate Commissions variables on voting behaviour. Furthermore, intervariation variables were included to test for interaction effects between the EU variable and the Support and Opposition variables and between the Support Agreement variable and the Support and Opposition variables. The logistic regression model was statistically significant X2(15)=350.969, p<.0005. The model explained 17.1% (Nagelkerke R2) of the variance in voting behaviour and correctly classified 92.2% of cases.
Moreover, in figure 5 it is shown that the interaction variables of EU and Support were statistically significant and showed that when the support party, the PVV, voted on an EU-related proposal, they were less likely to support it than other proposals. This confirms that there is most likely an interaction effect between the EU and Opposition/Support variables. The interaction variable with the Support Agreement and Support variables did not turn out to be significant, making it likely that there is no interaction effect between if a proposal is part of the Support Agreement and the Support party. Moreover, the variables Support and Budget turned out significant. The Support variable indicates that if this variable would be increased with one unit the increase in odds that the vote would be in favour would be 2.152 and the Budget variable indicating that this increase in odds would be as high as 13.021.
Furthermore, only one Senate Commission variable turned out to have a significant influence on voting behaviour, which is the SZW Commission (Sociale zaken en Werkgelegenheid/Social Affairs and Employment). An increase in one unit of this variable would mean a decrease of 0.287 of the odds of voting in favour of a proposal. The other Commissions did not have statistically significant results.
95% C.I. For Exp(B)
B S.E. Wald df Sig. Exp(B) Lower Upper
EU 0.248 0.180 1.896 1 0.169 1.281 0.900 1.823 Support 0.823 0.268 9.429 1 0.002 2.278 1.347 3.852 EU*Support -1.770 0.402 19.390 1 0.000 0.170 0.077 0.374 Supp Agr. 0.269 0.164 2.692 1 0.101 1.308 0.949 1.803 Support*Supp Agr. -0.484 0.410 1.393 1 0.238 0.616 0.276 1.377 Budget 2.555 0.288 78.709 1 0.000 12.876 7.322 22.644 VROM/V&W/IMRO -0.031 0.511 0.004 1 0.951 0.969 0.356 2.638 SZW -1.265 0.525 5.801 1 0.016 0.282 0.101 0.790 FIN 0.139 0.554 0.063 1 0.801 1.150 0.388 3.405 BZK/AZ/KOREL/WWI -0.547 0.514 1.129 1 0.288 0.579 0.211 1.587 OCW -0.768 0.523 2.153 1 0.142 0.464 0.166 1.294 VWS/JG -0.721 0.557 1.676 1 0.195 0.486 0.163 1.449 JUST/I&A/JBZ/V&J 0.010 0.517 0.000 1 0.984 1.010 0.367 2.781 EZ/LNV/EL&I -0.164 0.533 0.094 1 0.759 0.849 0.299 2.413 ESO/BDO -1.036 0.621 2.789 1 0.095 0.355 0.105 1.197 Constant 2.298 0.512 20.181 1 0.000 9.959
After determining the significance of the variables, the predicted odd change was studied. What this research set out to do is look at the support party and the opposition parties and their difference in voting on EU-legislation and other legislation. The difference that the variables of Support and Opposition have on the average voting of all votes is described by EXP(B) meaning the increase in odds of a yes vote by an increase of a variable by one unit. As shown in figure 5, when it came to the support party, there was a 2.152 increase in the odds that the vote would be in favour per increase of one unit. The differences between the variables Support and EU described in percentage of voting in favour of proposals is shown in figure 6. Based on a Zelig test the predicted values and their significance were analysed. This showed that only the influence of the combined Support, EU, and interaction variable EU*Supp to the Support variable is statistically significant, with a difference of 11 per cent as can be seen in figure 6.
EU = 0 EU = 1 Difference SUPP = 0 0.90 (0.81-0.96) 0.92 (0.84-0.97) 0.02(n.s.) SUPP = 1 0.95 (0.89-0.98) 0.84 (0.69-0.94) 0.11(*) Difference 0.05(n.s.) 0.08(n.s.)
FIGURE 6-PREDICTED VALUE DIFFERENCES,SIGNIFICANCE LEVELS:***=<0.001,**=<0.01,*=<0.05.(*)=< 0.1
Due to the insignificance of the variables dealing with the Support Agreement, as shown in figure 5, the decision was taken not to include a similar analysis between Support and Support Agreement as these significance numbers indicate that there is no strong relationship between these variables.
DISCUSSION AND CONCLUSION
This study sought to further understand the functioning of parties during a supported minority cabinet by looking at the clash between Rational Choice Institutionalism and Normative Institutionalism. The case of Rutte I in the Netherlands is arguably a case in which the perfect situation for Normative Institutionalism is present, there was a long time of majority cabinets, and now for the first time, there is a minority cabinet. In the logic of that theory, the parties would act as if there still was a majority cabinet. This would mean that the support party would act as a de facto member of the government. However, if parties would act rationally to enact their preferences their would be a need for ad hoc support on differing legislation. Therefore, this study sought to answer the research question: To what extent do
supported minority cabinets rely on ad hoc support to implement legislation?
The study was built around two hypotheses, the Support party and the Opposition party hypotheses. First, the Support hypothesis, when comparing the Support party votes without the influence of the EU variable with the Support party votes together with the EU variable and the intervariate variable of EU*Support there is an 11% difference in predicted values, which is just statistically significant. This difference seems to indicate that the support party has rationally chosen to vote based on their preferences concerning EU-proposals, as the PVV are Eurosceptic. This would confirm the Support party hypothesis. That the support party seems to have rationally chosen not to support specific EU-related legislation means two things for the Opposition parties hypothesis, First, it seems logical that if the support party can act outside of the norms of majority cabinets, the opposition can do the same and choose to support rationally or not support measures based on their preferences. Second,
due to the lower level of support of the support party, the minority cabinet would need opposition support to create legislative majorities on EU-related proposals.
However, when looking at the results of the opposition variable, which is the inverse of the support variable, the predicted value difference between opposition on EU-related legislation and other legislation, as shown in figure 6 does show a small difference supporting the Opposition hypothesis with 2% difference, but this number is statistically insignificant. This seems to indicate that the opposition parties do not support more EU-legislation than other legislation. This seems to go against the results of the Support hypothesis that indicated that parties would be able to choose to vote for their preferences rationally. All opposition parties showed higher support for the EU, as shown in figure 2, and most opposition parties seemed further away from the minority cabinet parties on the left-right dimension, except the SGP. Based on this information, it would be rational for these parties the opposition parties to support more EU-related legislation than other legislation. However, the results seem to indicate that there is no significant difference between the two.
Considering the support party did support less EU-related legislation, and therefore, it became necessary for the minority cabinet to seek ad hoc support for EU-related proposals, why does this difference not show up in the results? One reason could be the overall high form of consensus on all votes of the opposition. Based on figure 6, the opposition has predicted support for EU-related proposals at 92% and other proposals at 90%. These high levels of support by opposition parties confirm studies that the Netherlands has a high level of consensus voting (Andeweg, 2000; Andeweg, De Winter, & Müller, 2008). This high level of support could weaken the difference if there is any between opposition voting on different issues. This makes it unclear as to the perspectives of RCI and NI and what this would mean, as opposition parties would by norm vote against a large part of the proposals of the cabinet,
which the results deny. However, the results also do not confirm that the parties acted rationally to support their preferences. Due to this unclarity in the results, it would be wise to further study opposition behaviour during supported minority cabinets in states with a lower level of consensus voting. In this way, a potential difference between voting on different issues could be better shown.
The study chose to not only look at the influence of the EU variable on the Support and Opposition variable but to also take into account a Support agreement variable. This was done to not only look at the RCI perspective and see if there is evidence for that but also take into account the NI perspective and look at the explanative value of that variable. Figure 5 shows that the variables of the Support Agreement and the intervariate variable of Support*Support Agreement do not have statistical significance. This seems to indicate that there is no strong relationship between the voting behaviour of the support party and the support agreement. This seems to contradict the NI perspective, based on that perspective; it would be logical to find a relationship between those variables.
This study took into account the influence of the Senate Commissions and the Budget. The Senate commissions seemed to have a small influence on the voting behaviour, as only one commission showed significant results, SZ, the others did not. If the variable of SZW were increased by one, the odds of that vote being in favour would decrease by 0.282. This could indicate that the proposals coming through SZW are more contentious than others and therefore, a higher number of votes are against. Furthermore, the budget variable showed that when it would increase by one, it would significantly become larger by 12.876. This shows a high level of consensus on budget proposals.
The case study of the Netherlands has indicated that parties can act rationally to enact their preferences as the PVV/Support party seems to have done, but also leaves room for
discussion and further studies as to whether opposition parties did not show similar results. Moreover, it can be said, that it seems likely that if one party can act rationally during a supported minority cabinet, others can as well, therefore, it is important to study this issue with other case studies further. Furthermore, this study seems to indicate that there is no strong relationship between the Support Agreement that was signed by the minority cabinet parties and the support party and the voting of the support party.
Both the confirmation of the Support party hypothesis and the founding that there is no strong relationship between the support agreement and voting behaviour of the support party contradict the findings of Otjes and Louwerse. Furthermore, the unclarity around the Opposition party hypotheses does not confirm or contradict the value of both this study and Otjes and Louwerse’s study. The overall results do seem to indicate that at least the support party can act rationally and unbound by the support agreement, which would logically create the need for ad hoc support by the opposition, however, this is not clearly shown in the results. These results do seem to indicate that supported minority cabinets do at least on some issues need to rely on support to implement legislation, which seems to match with the results Denmark. This would mean that the ability of supported minority cabinets to rely on ad hoc support does not stem from norms, as Otjes and Louwerse suggest, but stems from the rationality of parties to try to enact their preferences.
REFERENCES
Andeweg, R. B. (2000). From Dutch Disease to Dutch Model?Consensus Government in Practice. Parliamentary Affairs, 697-709.
Andeweg, R. B. (2011). Purple puzzles: the 1994 and 1998 government formations in the Netherlands and coalition theory. In R. Andeweg, L. De Winter, & P. Dumont, Puzzles
of Goverment Formation: Coalition Theory and Deviant Cases (pp. 147-164). Abingdon:
Routledge.
Andeweg, R. B., De Winter, L., & Müller, W. C. (2008). Parliamentary Opposition in Post-Consociational Democracies: Austria, Belgium and the Netherlands. The Journal of
Legislative Studies, 14(1-2), 77-112.
Bale, T., & Bergman, T. (2006). Captives No Longer, but Servants Still? Contract Parliamentarism and the New Minorit yGovernance in Sweden and New Zealand.
Government and Opposition, 41(3), 422-449.
Best, H., & Vogel, L. (2014). The Sociology of Legislators and Legislatures. In S. Martin, T. Saalfeld, & K. Strøm, The Oxford Handbook of Legislative Studies. Oxford.
Bovens, M., & Yesilkagit, K. (2010). The EU as Lawmaker: the Impact of EU Directives on National Regulation in the Netherlands. Public Administration, 88(1), 57-74.
Budge , I., & Laver, M. (1986). Office Seeking and Policy Pursuit in Coalition Theory. Legislative
Studies Quarterly, 11(4), 485-506.
Budge, I., & Herman, V. (1978). Coalitions and Government Formation: An Emperically Relevenant Theory. British Journal of Political Science, 8(4), 459-477.
Christensen, J. (2010). Uncertain Steps towards a European Public Policy. Public
Administration, 88(1), 3-17.
Christiansen, F. J., & Pedersen, R. B. (2012). The Impact of the European Union on Coalition Formation in a Minority System: The Case of Denmark. Nordic Political Science
Daalder, H. (1971). Cabinets and Party Systems in Ten Smaller European Democracies. Acta
Politica, 6, 282-303.
de Vries, C. E., & Marks, G. (2012). The Struggle over Dimensionality: A Note on Theory and Empirics. European Union Politics, 13(2), 185-193.
Dinas, E., Matakos, K., Xefteris, D., & Hangartner, D. (2019). Waking Up the Golden Dawn: Does Exposure to theRefugee Crisis Increase Support for Extreme-RightParties?
Political Analysis(27), 244-254.
Döring, H., & Manow, P. (2018). Parliaments and governments database (ParlGov).
Information on parties, elections and cabinets in modern democracies. Development version.
Douma, W. T., Pieters, K., Feenstra, K., & Koch-Hartmanová, R. (n.d.). Pilot Monitor EU Invloed. T.M.C. Asser Instituut.
Downs, A. (1957). An Economic Theory of Political Action in a Democracy. The Journal of
Political Economy, 65(2), 135-150.
Eerste Kamer der Staten-Generaal. (n.d.). Eerste Kamer. Retrieved from https://www.eerstekamer.nl/home
Field, B. N. (2009). Minority Government and Legislative Politics in a Multilevel State: Spain under Zapatero. South European Society and Politics, 14(4), 417-434.
Gedoogakkoord VVD-PVV-CDA. (2010).
Grande, E., Kriesi, H., Lachat, R., Dolezal, M., & Timotheos, F. (2008). Globalizing West European Politics: The Change of Cleavage Structures, Parties and Party Systems in Comparative Perspective. In H. Kriesi, E. Grande, R. Lachat, M. Dolezal, S. Bornschier, & T. Frey, West European Politics in the Age of Globalization (pp. 320-344). Cambridge: Cambridge University Press.
Green-Pedersen, C. (2001). Minority Governments and Party Politics: The Political and Institutional Background to the "Danish Miracle". The Journal of Public Policy, 21(1), 53-70.
Herman, V., & Pope, J. (1973). Minority Governments in Western Democracies. British Journal
of Political Science, 3, 191-212.
Jitka, D. (2015). Economic Crisis and Growth in Vote Share for Extreme Left and Extreme Right Parties. Revie of Economic Perspectives, 15(3), 269-290.
Louwerse, T., Otjes, S., & van Vonno, C. (2018). The Dutch Parliamentary Behaviour Dataset.
Acta Politica, 53, 149-166.
Otjes, S., & Louwerse, T. (2014). A Special Majority Cabinet? Supported Minority Governance and Parliamentary Behavior in the Netherlands. World Political Science Review. Page, E. C. (1997, November 20). The Impact of European Legislation on British Public Policy
Making: A Research Note. Blackwell Publishers Ltd.
Parlement. (n.d.). Commissie voor Verkeer en Waterstaat. Retrieved from https://www.parlement.com/id/vhy1lj1nqzp6/commissie_voor_verkeer_en_waterst aat
Polk, J., Rovny, J., Bakker, R., Edwards, E., Hooghe, L., Jolly, S., . . . Zilovic , M. (2017). Explaining the salience of anti-eltism and reducting political corruption for political parties in Europe with the 2014 Chapel Hill Expert Survey data. Research & Politics, 1-9.
Riker, W. (1962). The Theory of Political Coalitions. New Haven: Yale University Press.
Seawright, J., & Gerring, J. (2008). Case Selection Techniques in Case Study Research. Political
Research Quarterly, 61(2), 294-308.
Staatscommissie Parlementair Stelsel. (2018). Lage Drempels, Hoge Dijken: Democratie en
Rechtsstaat in Balans. Amsterdam: Boom.
Steunenberg, B., & Toshkov, D. (2009). Comparing Transposition in the 27 Member States of the EU: the impact of discretion and legal fit. Journal of European Public Policy, 16(7), pp. 951-970.
Strøm, K. (1984). Minority Governments in Parliamentary Democracies: The Rationality of Nonwinning Cabinet Solutions. Comparative Political Studies, 17(2), 199-227.
Strøm, K. (1990). A Behavioral Theory of Competitive Political Parties. American Journal of
Political Science, 34(2), 565-598.
Strøm, K. (1990). Minority Government and Majority Rule. Cambridge: Cambridge University. Strøm, K., & Müller, W. C. (2010). Chapter 1 - Political Parties and Hard choices. In K. Strøm, & W. C. Müler, Policy, Office, or Votes? How Political Parties in Western Europe Make
Hard Decisions (pp. 1-35). Cambridge: Cambridge University Press.
Töller, A. E. (2008). Mythen und Methoden. Zur Messung der Europäiserung der Gesetgebung des Deutschen Bundestages jenseits des 80&-Mythos. Zeitschrift für Parlamentsfragen, 39(1), 3-17.
Töller, A. E. (2010). Measuring and Comparing the Europeanisation of National Legislation: A Research Note. Journal of Common Market Studies, 48(2), 417-444.
Toshkov, D. (2010). Taking Stock: A review of quantitative studies of transposition and implementation of EU law. NIG Annual Work Conference. Maastricht.
van Kessel, S., & Krouwel, A. (2011). Van Vergankelijke Radicale Dissidenten tot Kwelgeesten van de Gevestigde Orde: Nieuwe Politieke Partijen in Nederland en de Toekomst van de Representative Democratie. In R. Andeweg, & J. Thomassen, Democratie
Doorgelicht (pp. 301-317). Leiden: Leiden University Press.
APPENDIX
Another binominal logistic regression was done without the intervariation variable. The decision was made not to include this analysis within the results as the insignificance of most of the variables made the results unreliable for any further discussion.
B S.E. Wald df Sig. Exp(B)
95% C.I.for EXP(B) Lower Upper EU 0.040 0.163 0.059 1 0.808 1.040 0.756 1.432 Support 0.241 0.183 1.742 1 0.187 1.273 0.889 1.822 EU*Support 0.216 0.156 1.919 1 0.166 1.241 0.914 1.684 Supp Agr. 2.567 0.288 79.537 1 0.000 13.032 7.412 22.911 Support*Supp Agr. -0.028 0.502 0.003 1 0.956 0.973 0.363 2.603 Budget -1.241 0.521 5.670 1 0.017 0.289 0.104 0.803 VROM/V&W/IMRO 0.134 0.544 0.061 1 0.805 1.143 0.394 3.319 SZW -0.528 0.510 1.072 1 0.301 0.590 0.217 1.602 FIN -0.745 0.521 2.050 1 0.152 0.475 0.171 1.317 BZK/AZ/KOREL/WWI -0.707 0.551 1.645 1 0.200 0.493 0.167 1.453 OCW 0.019 0.509 0.001 1 0.970 1.020 0.376 2.766 VWS/JG -0.161 0.523 0.094 1 0.759 0.852 0.305 2.376 JUST/I&A/JBZ/V&J -1.035 0.603 2.947 1 0.086 0.355 0.109 1.158 EZ/LNV/EL&I 2.330 0.510 20.887 1 0.000 10.282
Furthermore, a first keyword search was done to establish what is EU-related legislation and what is not, based on using the same keywords as mentioned in the research design, however, searching for them only in the full titles of the proposal. This search found 79 out of 585 proposals instead of the 97 out of 585 as found by the more comprehensive test. The decision was later made only to use the more comprehensive test like that one is likely more reliable. However, the results where the other test was used can be found here:
FIGURE 8-VOTING BEHAVIOUR DURING RUTTE I
Party EU Non-EU Difference
SP 0.86 0.87 -0.01 PVDD 0.84 0.85 -0.01 PVDA 0.95 0.92 0.03 GL 0.95 0.93 0.02 D66 0.96 0.95 0.01 VVD 1 0.98 0.02** SGP 0.94 0.98 -0.04 CU 0.91 0.94 -0.03 CDA 1 0.99 0.01 PVV 0.79 0.96 -0.17*** 0% 20% 40% 60% 80% 100% SP PvdD PvdA GL D66 VVD SGP CU CDA PVV Ye s v o te s i n % Parties EU-Related Other
FIGURE 9-T-TEST
95% C.I. For Exp(B)
B S.E. Wald df Sig. Exp(B) Lower Upper
EU 0.248 0.180 1.896 1 0.169 1.281 0.900 1.823 Support 0.823 0.268 9.429 1 0.002** 2.278 1.347 3.852 EU*Support -1.770 0.402 19.390 1 0.000*** 0.170 0.077 0.374 Supp Agr. 0.269 0.164 2.692 1 0.101 1.308 0.949 1.803 Support*Supp Agr. -0.484 0.410 1.393 1 0.238 0.616 0.276 1.377 Budget 2.555 0.288 78.709 1 0.000*** 12.876 7.322 22.644 VROM/V&W/IMRO -0.031 0.511 0.004 1 0.951 0.969 0.356 2.638 SZW -1.265 0.525 5.801 1 0.016* 0.282 0.101 0.790 FIN 0.139 0.554 0.063 1 0.801 1.150 0.388 3.405 BZK/AZ/KOREL/WWI -0.547 0.514 1.129 1 0.288 0.579 0.211 1.587 OCW -0.768 0.523 2.153 1 0.142 0.464 0.166 1.294 VWS/JG -0.721 0.557 1.676 1 0.195 0.486 0.163 1.449 JUST/I&A/JBZ/V&J 0.010 0.517 0.000 1 0.984 1.010 0.367 2.781 EZ/LNV/EL&I -0.164 0.533 0.094 1 0.759 0.849 0.299 2.413 ESO/BDO -1.036 0.621 2.789 1 0.095(*) 0.355 0.105 1.197 Constant 2.298 0.512 20.181 1 0.000*** 9.959
