Home Political science
Variation in the Strength of Oversight Institutions
Analysing the explanations discussed so far, the literature's favoured approach is comparing levels in the strength of oversight institutions at particular points in time. In line with this approach, Table 4.3 presents three regression models of the strength of oversight parliaments had reached at the end of each of their reform opportunities. The first model includes the factors at the heart of an explanation of institutional adaptation based on constitutional preferences, namely parliamentary EU support and our indicator of existing national legislative institutions. The second model comprises all three variables close to the view that EU oversight institutions arise as a response to
Table 4.3 Models of the strength of oversight institutions at the end of reform opportunities
Note: Ordinary Least Square (OLS) regression with country-clustered standard errors. в: Coefficient estimate. SE: Standard error. p.: p-value. RO: Reform opportunity. AIC: Akaike information criterion. BIC: Bayesian information criterion.
challenges of minority government and coalition conflict along the left-right and EU dimensions of political contestation. The third, 'full', model brings together all these variables and a measure for popular Euroscepticism as an additional control variable. Furthermore, all three models control for the level of oversight that parliaments had reached at the end of their last reform opportunity. Note, in this context, that the models do not directly measure the development of the authority of the EU. A corresponding variable correlates highly with the lagged outcome because both trend upwards over time. The effect of integration is, of course, implicit in the research design which focuses on reform opportunities associated with steps to deepen integration. It is possible to control for EU authority instead of the level of oversight after past reform opportunities, as discussed after the following two paragraphs have introduced the results shown in Table 4.3.
The results support the argument that constitutional preferences shape parliamentary adaptation to the EU, while minority government and coalition conflict appear not to matter. National parliaments' institutional competences in EU policy-making are weaker the more supportive parties are of European integration. Where the EU raises partisan opposition, parliaments have stronger EU-related rights and capacities. Every one-unit difference in parliamentary EU support corresponds to about a 0.15 unit change in the expected direction on the 0-2 scale measuring the strength of oversight institutions. The same holds true of domestic institutions. The more extensive the competences of the domestic committee system, the more extensively parliaments adapt to European integration. For instance, comparing the parliaments with the weakest and strongest committee systems corresponds to a difference of about 0.4 units in the strength of oversight institutions. These results hold in the first model as well as in the full model. None of the other explanatory variables, except for the lagged outcome, as one would expect, has significant effects. Minority government even works in the opposite direction to what is expected.
Comparing the fit of Model 1, which focuses on constitutional preferences, and Model 2, which focuses on incentives related to the partisan configuration of the cabinet, the former prevails. It is the preferable model according to the r-square, Akaike information criterion (AIC) and Bayesian information criterion (BIC) scores. In fact, it performs better not only than the second but also than the full model on the latter two indicators. Figure 4.3 shows how closely the expected levels of oversight of the first and second models approximate the outcomes of parliamentary adaptation to the EU that we actually observe in the data. Many cases are well predicted by both models, which is, however, due to the inclusion of a lagged dependent variable. Since changes over time are often small or absent, the state at the end of the last reform opportunity will often closely anticipate what happens next. Nevertheless,
Analysing Domestic Adaptation to Integration
Figure 4.3 The accuracy of the empirical expectations of Models 1 and 2 compared
Note: The horizontal axis shows how far the outcomes predicted by the models in Table 4.3 and the ones actually observed lie apart. The vertical axis and the bars show the number of observations in a given range of the vertical axis.
model 1 makes more predictions that deviate only slightly from the observed outcome than model 2. Model 2, in turn, significantly exceeds model 1 in terms of more inaccurate predictions. Overall, the model focusing on constitutional preferences fits the data notably better.
Let us return to the question of how the level of EU authority affects parliamentary adaptation to integration. It is clear that significant changes at the European level create the opportunity for domestic policy-makers to consider whether reforms of oversight institutions are in order. But does the extent of European-level change also matter? Table A4.4 in Appendix I replaces the control variable of the past level of oversight with one for the level of EU authority and shows a strong and significant association with parliamentary adaptation to integration. For instance, a change similar to what the Maastricht Treaty brought about in the authority that the EU holds on average across all policy areas, all else equal, corresponds to a 0.22 unit change in parliamentary oversight institutions. Less important treaties, on the other hand, would be expected to lead to more marginal domestic adaptation.
A second question one could ask about the impact of developments in European integration is whether it matters in what kind of policy areas the EU obtains competences. In this case, however, we know empirically that parliaments across the member states have adopted a set of oversight institutions to deal with EU affairs across the range of policy areas, instead of tailoring their arrangements to the characteristics of particular areas. Thus, it is not possible to test whether parties are especially concerned about EU intrusion into selected domains. However, the following qualitative chapter provides additional information on this point. Moreover, in the context of the Eurozone crisis, the EU has create decision-making procedures in the context of the European Stability Mechanism that stand outside of the regular treaty framework and that were highly salient domestically. Later chapters will examine parliamentary reactions to this new challenge and reform opportunity in more detail, allowing us to see, amongst other things, whether there are issue-specific deviations from the patterns of adaptation that parliaments have pursued.
Appendix I further shows that it is possible to replace parliamentary support for European integration with a measure of parliamentary support for the EP without producing noteworthy changes in the results (see Table A4.5). This indicates that relying on EU support as an operational indicator, instead of support for the EP is suitable, even though the latter is likely to be a more accurate approximation of whether parties tend towards intergovernmental or federal constitutional preferences.
Consider a final issue: accession. Several scholars suggest that EU accession is a particularly important moment for the development of EU-related oversight institutions in member state parliaments (Saalfeld 2005; see also Hamerly 2007). When countries join the EU, they have to make first choices about new legislative institutions. As the overview of the development of institutional adaptation in Chapter 2 showed, the first reforms that countries implement are often the most substantive ones. However, knowing that accession reforms are especially pronounced does not necessarily imply that they follow a different logic from subsequent reforms. What one wants to know is whether variation in accession reforms requires a different explanation from subsequent variation. If so, it would not be appropriate to analyse these initial reforms together with later reforms.
Due to the low number of accession cases, it is not possible to examine variation in accession reforms in a multivariate analysis. However, Table 4.4 shows bivariate correlations between the explanatory variables and the strength of oversight institutions in national parliaments after their EU accession. The correlations point in similar directions as the results discussed so far. EU support
Table 4.4 Correlates of reforms of oversight institutions at EU accession
Note: The cell entries result from bivariate correlations.
and committee strength correlate with the strength of parliamentary adaptation strongly and in the expected direction. There is no noteworthy association between oversight institutions and, respectively, minority government or cabinet conflict on the left-right dimension. As one would expect, the changes parliamentary parties implement after their countries become EU members are, on average, more extensive the more powerful the EU is at the time (Saalfeld 2005). There are two discrepancies between the results of the multivariate analysis and the correlations in Table 4.4. First, Euroscepticism appears to have a bearing on accession reforms, albeit, if anything, in the opposite direction of what the literature has suggested. This effect results from the more Europhile attitudes prevailing in the populations of Eastern Enlargement countries, whose parliaments have tended to create strong oversight institutions. If one were to compare the EFTA and Southern enlargements only, with countries such as Finland or Sweden on the one hand, and Portugal and Spain on the other, Euroscepticism would work in the expected direction. However, overall, the fact that Euroscepticism works inconsistently across enlargement rounds reinforces the non-finding of the previous multivariate analysis. The final discrepancy is that, during accession, cabinet conflict over the EU appears to encourage the creation of stronger parliamentary oversight institutions. We will return to this effect further below. At this stage, the main point is that accession reforms do not seem to be fundamentally different from subsequent reforms except that they are stronger while later reforms are more incremental—logically, given that they build on the oversight institutions created previously.
The analyses presented so far build on the most common analytical approach in the literature, namely the comparison of levels of oversight in different countries. The results are largely in line with the argument put forward here and with a number of previous studies that emphasise the importance of parliamentary support for the EP (Winzen etal. 2015), domestic institutions (Dimitrakopoulos 2001; Raunio 2005; Karlas 2012), and the deepening of EU authority (Raunio and Hix 2000; Winzen 2013). The findings are incompatible with claims as to the importance of incentives arising from minority government (e.g. Bergman 1997: 381; Martin 2000: 154; Saalfeld 2005: 357), and coalition conflict over policy (Saalfeld 2005), while it cannot be fully ruled out that coalition conflict over the desirability of European integration encourages parliamentary reform albeit, as the previous section showed, only in the few cases of cabinets that are actually characterised by noteworthy EU conflict. The next section takes a different perspective, focussing on the probability that parliaments realise reform opportunities that arise as a result of developments in European integration.