Introduction

Abstract

This article introduces the subject of constituency representation in states with systems of multi-level government and/or mixed member proportional (MMP) electoral systems. It introduces the interest in this subject in the UK following devolution in Scotland and Wales. It specifies research questions on the approaches of representatives to constituency work and the relationships between them. It puts forward five hypotheses as a focus for comparative analysis. The paper then raises the need in developing comparative analysis to take account of different national institutional and political contexts. It explores different institutional forms of multi-level and MMP systems, and then different ideas about the roles that different types of representative should play in constituency work as well as contexts of party organization and party system.

Keywords:

• Constituency representation
• devolution and mixed member proportional electoral systems

Notes

¹Consideration does not extend to the rest of the UK. The devolved Northern Ireland Assembly poses an interesting case where multi-level government is combined with the use of a Single Transferable Vote (STV) electoral system. However, the Assembly has been suspended for longer than it has sat since its creation in 1998, its last suspension being in October 2002. In England, representing 85% of the UK, there is only administrative regionalism.

²Three of Norton's six hypotheses are investigated here in revised form. These relate to commitment to service provision, comparison of constituency and list member contact with constituents, and comparison of constituency member provision in unitary and multi-level states. To these are added others relating to representatives' perceptions of relationships over constituency representation. Norton's other three hypotheses relating to the growth of constituency load over recent decades, the impact of constituency service on the personal vote, and other costs and benefits for representatives require consideration of different kinds of data and lie outside the focus of this study.

³After detailed analysis of constituency work by a selection of UK MPs, Pippa Norris concluded that “those most committed to service work were women backbenchers with a high case-load in inner-city seats, who strongly identified with their social work function. In total this model explained 16 per cent of variance in constituency service. In contrast other factors were found to be insignificant, including measures of selectoral and electoral incentives” (Norris, 1997: 45–46).

⁴The articles take different approaches to structuring their arguments. Franks and Patzelt pose research questions and provide answers. McLeay and Vowles, Russell and Bradbury, and Bradbury and Mitchell explicitly discuss hypotheses and test their validity against the evidence. In part this is a reflection of the fact that the UK and New Zealand are recent examples of multi-level government and MMP, and it is more obvious that research should build on the theoretical arguments provided by analysis of more longstanding cases.