Desktop version

Home arrow Political science

  • Increase font
  • Decrease font

<<   CONTENTS   >>


With the limited number of observations for which data on the explanatory variables and the outcome of interest is available, the analysis presented in the following must necessarily be regarded with caution. At the same time, there is no reason to think that the limited number of cases favours an argument based on constitutional preferences over alternative accounts such as those stressing party political exigencies arising from minority and coalition government. The analysis can at least compare the relative ability of different explanatory approaches to account for cross-national diversity in parliamentary adaptation to the ESM, and in their preferences as to the Article-13 conferences. Along these lines, Table 7.2 presents two regression models. These models are set up in a similar way to those in Chapter 4 in the sense that the first brings together the variables of interest for an explanation based on constitutional preferences, while the other includes measures central to the view that incentives and constraints arising from minority and coalition government shape EU-related parliamentary institutions. One might object that these variables should include at least some additional control variables such as popular Euroscepticism or the factors, discussed above, capturing the

Table 7.2 Models of parliamentary rights in ESM decision-making









EU support




Budget rights Minority government







Cabinet conflict (left-right)




Cabinet conflict (eU)

















Note: Ordinary least square regression models. в: Coefficient estimate. SE: Standard error. p.: p-value.

political-economic salience of the Eurozone reforms. Table A7.1 in Appendix

shows corresponding models without, however, finding an effect for any of the control variables, or relevant changes in the effects of the variables of interest.

The results reinforce the visual impression derived from the plots in Figure 7.3. In Model 1, parliamentary EU support, taken to stand for the inclination of parliamentary parties towards federal constitutional preferences, has a negative impact on parliamentary ESM rights, albeit, as theoretically suspected, inconsistently, so that this effect is not significant. In line with the argument, existing budgetary rights have a positive and significant effect on parliamentary adaptation to the ESM. A one-unit increase on Wehner's (2006) index, which ranges from 16.7 to 66.7 in the data, corresponds to a 0.075 unit increase in the strength of parliamentary ESM rights, which are measured on a 0-4 scale. In other words, the largest observed difference in budget rights corresponds to a 3.75 unit difference in ESM-rights. None of the variables in model two has a significant effect on parliamentary adaptation efforts, although the coefficients do point in the right direction, except for cabinet conflict on the left-right dimension that appears, if anything, even to depress the ESM-related rights parliaments decide on.

Comparing the explanatory power of the two models, the first point to note is that the r-square value of the first model, which focuses on constitutional preferences, is more than twice as large as that of the second model, despite the fact that the latter includes one additional variable that should enhance its r-square, notably in such a small sample. Figure 7.4 goes one step further. It is based on the rationale that, in such a small sample, we should care most about a model's ability to pinpoint successfully what kind of ESM-related competences exist where. This, of course, is to some extent made impossible as the models are linear regressions despite the fact that, strictly speaking, the outcome is categorical: ESM rights take on only integer values from 0 to 4 whereas the models

Observed and predicted parliamentary rights, based on the results in Table 7.2

Figure 7.4 Observed and predicted parliamentary rights, based on the results in Table 7.2

Note: Crosses: Observed values of parliamentary rights. Diamonds and triangles: Fitted values of the respective models, rounded to the next integer value.

predict rational numbers from 0 to 4. The reason to employ linear models is pragmatic and owed to the limited number of available observations. What the figure does, however, is to round predicted values from the two models to the next integer value. In other words, it tells us whether the models get the outcomes of the cases approximately right. The results speak out in favour of model 1. It is closer to the observed outcome than model 2 in six out of twelve cases.[1] Model 2 does better in only two cases, while both models are equally close to the observed outcome in four cases. What is more, model 1 accurately predicts the strength of ESM-rights in six cases (although two of these six cases are shared with model 2). Overall, these results suggest that an explanation based on constitutional preferences, and existing budget competences in particular, is more helpful in understanding why parliamentary parties create institutional rights in ESM decision-making compared to an account focusing on exigencies arising from minority and coalition government.

Finally, one might object that, given the limited number of cases, the results could be driven by individual, influential observations. Table A7.2 re-estimates model 1 in Table 7.2, while excluding individual countries. Far from being

Table 7.3 Comparison of parliaments with different Article-13 conference preferences

Article-13 conference



Difference as % of largest



possible difference*

EU support






Budget rights






EU oversight institutions






Cabinet conflict (left-right)






Cabinet conflict (eU)












Budget transfers






Public Debt






Note: Cell entries in the first two columns are the mean values of the explanatory variables. The third column shows the level of significance of the difference between the groups preferring a weak or strong Article-13 conference, respectively. * The absolute difference in the values of the explanatory variables for the two groups (weak and strong supporters) relative to the difference between the minimum and maximum value of the respective variable that exists in the data. At the extreme all countries would have either the minimum or maximum observed value. The mean difference between supporters and opponents would then be equal to (i.e. 100 per cent of) the range in the data.

driven by selected countries, however, this robustness test reveals stability in the findings, regardless of which country is left out of the analysis.

Let us turn to parliamentary preferences for a narrow or broad mandate for the Article-13 conference. In this case, the outcome of interest is binary and, therefore, cannot usefully be analysed with a linear regression model, while the number of observations is too small for a logistic regression. Table 7.3, therefore, takes the more modest approach of examining whether there are differences in the mean levels of the explanatory variables between the groups of parliaments that, respectively, favour a narrow or a broad mandate. Notwithstanding the limits that this method necessarily entails, the results nonetheless are relatively clear-cut. A helpful way to look at the table is to think about the largest possible difference that could exist between the two groups of supporters and opponents of a broad conference mandate. The largest possible group difference on any given explanatory variable would exist if all parliaments in one group had the minimum value of a variable that we observe in the data, and all parliaments in the other group the maximum value. The average difference between the two groups of supporters and opponents would then correspond to the observed range of the explanatory variable. Taking this as the benchmark, there is a categorical gap between three explanatory variables and the rest. The gap between supporters and opponents in terms of the strength existing EU-related oversight institutions, and the political-economic salience of the Eurozone reforms amounts to between 18 and 21 per cent of the maximum possible. For all other explanatory variables, the gap is very small, not exceeding five per cent of what it could be. Unsurprisingly, the first three variables are also the only ones that come close to conventional statistical significance thresholds. Parliaments that seek a broad mandate for the new inter-parliamentary conference have weaker EU-related oversight institutions at home, receive less from the EU budget (or even make net contributions), and are indebted at around 71 per cent of GDP, as compared to 50 per cent for the opponents of a broad mandate.

These findings are comparable to the ones of Chapter 6. There, it also turned out the EU-related oversight institutions existing in the member states constrained the interest of parliamentarians in a strong direct role in EU policy-making. In the case at hand, domestic opportunities to engage with the EU lead parliamentary parties to settle on modest suggestions for the mandate of the Article-13 conference. Whether partisan compositions of national parliaments tend towards federal or intergovernmental constitutional preferences does not make a difference. As argued, this might be the case because neither federalists nor intergovernmentalists are 'friends' of far- reaching inter-parliamentary cooperation, albeit each camp for its own reasons. Having said this, while Chapter 6 also found that domestic parliamentary rights unrelated to the EU shaped parliamentary demands for a direct European role, this appears not to be the case in the context of the Article-13 conference. In this sense, the findings reinforce particularly the point that existing EU competences at the domestic level depress parliamentary interest in a direct European-level role.

The results, furthermore, suggest that issue-specific factors can shape parliamentary reform preferences for a direct European role in addition to the institutional conditions considered so far. In countries in which economic and budgetary policy-making at the European-level is particularly salient as a result of a large public debt in excess of the EU's prescribed limits, and due to limited benefits from or even contributions to the EU budget, parliamentary parties tend to support a strong Article-13 conference, even if institutions at home would suggest otherwise. The German Bundestag, for instance, has far- reaching EU-related oversight institutions. Its constituent parties and member should be expected to have little interest in inter-parliamentary conferences. Yet, with Germany being a net contributor to the EU budget with a considerable public debt level, an inter-parliamentary conference to monitor European-level policy-making appears to be valuable to German representatives.

Summing up, the analysis has produced insights in line with the theoretical argument. Institutional conditions in particular have shaped national parliamentary rights in ESM decision-making and their preferences regarding the mandate of the Article-13 conference. In the former case, however, domestic budgetary competences are central whereas existing EU-related oversight institutions matter in the latter. Additionally, issue-specific political-economic factors apparently work to enhance parliamentary interest in monitoring European-level economic policy-making. Whether parliaments are composed of federalist or intergovernmentalist parties does not consistently affect their adaptation to the challenges posed by the ESM treaty. As argued, this might be the result of the strongly intergovernmental traits of this treaty, which finds favour among intergovernmentally oriented parties, while also encouraging federalists to look for sources of parliamentary authority other than empowering the EP. Finally, an argument focusing on constitutional preferences and the closely related domestic institutions hold more explanatory power as to parliamentary responses to the new challenges arising from the recent Eurozone reforms than an account focusing on party political exigencies under minority and coalition government.

  • [1] The models in Table 7.2 each have 13 observations. However, one of these thirteenobservations is not shared between the two models and not included in Figure 7.4.
<<   CONTENTS   >>

Related topics