Assessing democratic representation in multi-level democracies

Abstract

This article introduces the special issue ‘Assessing democratic representation in multi-level democracies’ from a conceptual perspective. It adapts Powell’s chain of responsiveness – as a model of the democratic process on the national level – to the context of multi-level systems and discusses conditions that might facilitate or hamper responsiveness in regional democracies. The theoretical reasoning identifies added complexity, multiple actors sharing the same label and cross-level interdependent decisions as the key challenges to multi-level democracy. Empirical illustrations focus on the first stage of the representation process. Here voters should form rational policy preferences and take informed voting decisions, and parties are expected to offer coherent policy platforms that are tailored to the specifics of the regional level. While the analysis of party manifesto data suggests that regional parties cover regional issues, and strategic incentives to focus on other levels appear limited, survey-based information illustrates the cognitive burden of multi-level democracy on voters. In combination with a synopsis of the other contributions assembled in this issue, the findings suggest that information, responsibility and accountability problems may be particular obstacles to responsiveness in multi-level systems.

Keywords:

• Regional democracy
• responsiveness
• multi-level systems

Acknowledgements

This paper introduces a special issue, which has resulted from a workshop on ‘Vices and Virtues of Decentralization: Challenges in European multi-level democracies’, held at the Mannheim Centre for European Social Research (MZES) on 25‒26 November 2016. We would like to thank the MZES for financial support of this event, and we are grateful to all authors and other workshop participants for the stimulating discussion and their interesting contributions. We also thank two anonymous reviewers for their valuable suggestions.

Notes

1. A similar and somewhat related development can be observed with regard to the local level: based on an impressive data collection effort, Ladner et al. (2016: 328) find ‘an increase of local autonomy on almost all variables measured’ from 1990 to 2014.

2. In line with other work, we believe that this constitutes an important research agenda that goes well beyond what is covered in this special issue. The many implications of federalism and decentralisation, which we do not attempt to discuss here, are investigated, for instance, by Bednar (2011), Benz (2015), Benz and Sonnicksen (2017) and Treisman (2007).

3. Evaluating the whole chain of responsiveness is important because a correlation between citizens’ preferences and public policies might merely be the result of ‘[g]ood luck or advantageous circumstances’ or ‘the good will of policy makers’ (Powell 2004: 92).

4. Powell (2004: 97) mentions the importance of ‘nationally coherent policy agendas’: if parties remain ‘mere collections of local elites’, citizens might be confused and unable to identify a party that is likely to represent their preferences. He does not develop this point more systematically, though.

5. Similarly, we consider regionalisation and decentralisation as synonymous.

6. Arguably, for the second linkage it is likely that by and large the same subversions and facilitating conditions as at the national level matter for the regional one (see Powell 2004: 97–9). This is not to say that the linkage is less vulnerable in multi-level systems, but we expect the fundamental logic regarding the translation of votes into seats, the functioning of party groups, and government formation to be similar.

7. With regard to research at the national level, Powell (2004: 101) emphasises that comprehensive studies ‘trac[ing] the full causal chain … using measures that are comparable across countries’ might be hard to realise – particularly in a cross-national setting. He therefore also discusses more pragmatic approaches, such as the identification of ‘red flags’, which could point to obstacles in the translation of citizens’ preferences into public policies.

8. On the other hand, parties might have incentives to present ambiguous policy platforms to the voters (Bräuninger and Giger 2016).

9. Please note that by speaking of regional parties we refer, unless explicitly differentiating between the two, to regional party branches of state-wide parties on the one hand and stateless nationalist and regionalist parties (SNRPs) on the other (see Hepburn 2011).

10. What may be seen as a clear vice for the quality of democracy at the regional level may, in fact, have positive effects on democratic responsiveness at the national level. If the performance of national governments is constantly evaluated in regional elections, it might provide an accountability mechanism that also works between national elections and facilitates dynamic representation (Stimson et al. 2013). Yet negative consequences of a permanent state of campaigning are also conceivable.

11. The values reflect the share of ‘don’t know’ responses, so they do not represent the ability to identify the correct tier.

12. Differences within countries (e.g. between England and Scotland) are neglected.

13. Also included are three Scottish and three Welsh manifestos, which we do not consider here.

14. For the graph, statements that refer to the regional level but ask for a reduction of authority on that level are recoded to ‘other’. The overall pattern of results is the same without that recoding or when looking only at statements not asking for a change in authority levels. The graph was created with the help of the R package ggtern (Hamilton 2016).

15. The plots exclude documents that do not include any statements referring to the respective policy area (one out of 292 documents for economic policy, 21 for foreign policy).