Did They Write Back?

A Mandate Divide in Response to Constituent Casework in Devolved Bodies

Other formats: Plain text, PDF, Kindle .mobi, EPUB.

Version of record

This is a pre-copyedited, author-produced version of an article accepted for publication in Parliamentary Affairs following peer review. The version of record is:

Alex Parsons, Rebecca Rumbul, Did They Write Back? A Mandate Divide in Response to Constituent Casework in Devolved Bodies

Parliamentary Affairs, Volume 72, Issue 3, July 2019, Pages 522–541

Summary

Three devolved bodies in the UK (the Scottish Parliament, the National Assembly for Wales, and the London Assembly) use Multi Member Proportional Representation (MMP) electoral systems. These bodies have two groups of representatives, regional and constituency, elected through different methods. Using a survey of members of the public who contacted their representative, constituency representatives are shown to have a higher response rate across all three bodies, and this effect is isolatable from the role of being a government party. In Scotland this effect is demonstrable when comparing the behaviour of the same people holding different roles, suggesting that individuals adapt their behaviour to the expectations of the role.

Introduction

The election of New Labour in 1997 marked the beginning of formal devolution in the UK. Manifesto pledges made by Tony Blair's new government to provide devolution referendums in London, Northern Ireland, Scotland and Wales were quickly actioned, and following 'yes' votes, four devolved governments were created and implemented between 1997 and 2000. These were the Scottish Parliament, the National Assembly for Wales (NAfW), the Northern Ireland Assembly and the London Assembly. All of these bodies except the Northern Ireland Assembly use a Mixed Member Proportional (MMP) electoral system (called the Additional Member System, or AMS, in the UK). As a result of this system voters in Scotland, Wales and London have two kinds of representatives: a single constituency member and multiple regional members. This study explores the differences between these two kinds of representative in terms of their behaviour in dealing with constituent casework. It uses a survey of constituents who have contacted their devolved representative to examine whether different kinds of representative exhibit differences in their response rate.

The focus of this study was to identify if there are differences in representative behaviour linked to individual members' election to either a regional or constituent seat.

This study first discusses the culture and practice of constituency service and casework in the UK. It then discusses the literature on the structure and different behaviour of representatives in multi-member systems, and a theoretical discussion of factors considered to create divergence in responsiveness rates between list and constituency candidates. In reviewing the literature concerning divergence between list and constituency candidates, the evidence for structural incentives driving competitive behaviours between candidates provided significant potential to explain differences in responsiveness. The culture, resource allocation and operation of devolved systems is considered within the theoretical framework, as is how dual candidacy can lead to 'contamination' of what is considered normal electoral and post-electoral behaviour in list and constituency elections. It discusses how role differentiation has been observed to date in terms of legislative behaviour and relationship with constituents, and the hypotheses are then constructed and presented on the basis of this theoretical framework.

In order to empirically detect differences in the behaviours of regional and constituency representatives, the research examines the practice of casework in the devolved UK bodies. In some bodies there is an obligation on representatives to undertake casework within the area they represent, however representatives are able to conduct this work at their own discretion. Data relating to responsiveness by regional and constituent members was therefore considered to be a neutral method of comparison.

This study finds a statistically significant relationship between mandate type and responsiveness, which can be isolated from alternative explanations such as being a member of a governing party. The results show that a responsiveness difference exists even for the same individuals when in different roles, demonstrating a change in behaviour. It concludes with a discussion on the potential implications for citizens and parliaments concerning the divergence in role profiles for constituency and regional representatives.

Constituency Service and Casework

Constituency service is the understanding of MPs that part of their role involves availability to constituents, both through written communication as well as through events and surgeries run in the constituency. This extends beyond constituents' opinions on legislation, to place the MP in the role of intermediary and advocate for their constituents with other aspects of government or wider society.

The important role elected British politicians assign to constituency service is a relatively recent development, with time spent in constituencies and working on constituent business increasing substantially since the 1950s. Riddel (as cited in Norris, 1997) argues this shift resulted from the post-war expansion of government services, requiring more official mediation between citizens and government. Gay (2005) shows how this behaviour has adapted to changing circumstances once it became accepted as an important part of the role: 'privatisation of major utilities has made no difference to the overall workload of the MP, who became as effective in demanding redress from private companies as from nationalised industries. [The MP] role developed from holding central government to account to holding everyone to account' (2005, p. 59).

Notes (hide):

1: Norris, P. (1997) ‘The puzzle of constituency service', The Journal of Legislative Studies, 3 (2), pp. 29–49.

2: Gay, O. (2005) ‘MPs Go Back to Their Constituencies', The Political Quarterly, 76 (1), pp. 57–66.

The growth of casework was tied in with an increase in direct communication with constituents. Letters MPs received grew from two or three per day in the 1950s, to 25-74 per week in the 1970s to 190 letters per week by 1992 (with half directly from constituents) (Norris, 1997, p. 29-30). One explanation for why MPs have embraced this shift is their motivation to grow a personal vote in order to help secure re-election. King (1991) provides evidence that this is effective in a US-context although the personal vote has been found to be weaker in the UK (Cain, Ferejohn and Fiorina, 1984). There is evidence for the existence of a personal vote for constituency representatives in a variety of countries with MMP voting systems, with the exception of Germany (Moser and Scheiner, 2005).

Notes (hide):

3: Norris, P. (1997) ‘The puzzle of constituency service', The Journal of Legislative Studies, 3 (2), pp. 29–49.

4: King, G. (1991) ‘Constituency Service and Incumbency Advantage', British Journal of Political Science Polity American Political Science Review American Political Science Review American Politics Quarterly, 21 (8), pp. 119–128.

5: Cain, B. E., Ferejohn, J. A. and Fiorina, M. P. (1984) ‘The Constituency Service Basis of the Personal Vote for U.S. Representatives and British Members of Parliament', The American Political Science Review, 78 (1), pp. 110–125.

6: Moser, R. G. and Scheiner, E. (2005) ‘Strategic Ticket Splitting and the Personal Vote in Mixed-Member Electoral Systems', Legislative Studies Quarterly, 30 (2), pp. 259–276.

British MPs have been shown to believe it's a good idea themselves. Norris (1997) found that "members defending marginals devoted almost half their week to constituency work, while 'safe' members spent only about a third of their time on such service". Those elected in 1983 spent less time on constituency service than those elected in 1987 (which was closer to the survey), suggesting that newer MPs spend more time on constituency service to secure a personal vote (Norris, 1997, p. 38). Norris also found a gender divide, where in the late 1990s female MPs were spending more time on constituency work and casework (1997, p. 44).

Notes (hide):

7: Norris, P. (1997) ‘The puzzle of constituency service', The Journal of Legislative Studies, 3 (2), pp. 29–49.

8: Norris, P. (1997) ‘The puzzle of constituency service', The Journal of Legislative Studies, 3 (2), pp. 29–49.

There is also a sense that constituency service is an aspect of an MPs time that they will focus on in the absence of other responsibilities. In a cross-national study, Heitshusen et al. found that as portfolio responsibilities (such as ministerial duties) increased, MPs were far less likely to have a high constituency focus (Heitshusen, 2005a, p. 40). They also found a strong New Zealand effect, with New Zealand constituency MPs very likely (92%) to assign constituency work as their highest priority. Anagnoson noted MPs in New Zealand put a high priority on constituency work even prior to the shift to MMP and argues the strong party system with tight control of national policy and parliamentary votes led MPs to focus on an aspect of their job they had more control over (Anagnoson, 1983, p. 172).

Notes (hide):

9: Heitshusen, V. (2005a) ‘Electoral Context and MP Constituency Focus in Australia, Canada, Ireland, New Zealand, and the United Kingdom', American Journal of Political Science, 49 (1), pp. 32–45.

10: Anagnoson, J. T. (1983) ‘Home Style in New Zealand', Legislative Studies Quarterly, 8 (2), pp. 157–175.

Differences and conflict between different types of electoral mandate

There are two families of mixed member systems, Mixed Member Majoritarian (MMM) and Mixed Member Proportional (MMP). In the first, the results of one election does not affect the results of the other, in the later the results of the Proportional Representation (PR) election to calculate what the allocation of seats by party should be for the entire body and then, based on how many seats are already filled through Single Member Districts (SMD), allocates seats from party lists to correct the imbalance. While there are published examples of role differentiation in both MMM and MMP systems, these systems have different effects on the relationship between the two roles and comparisons from both are useful in examining different behaviours of representatives in the UK (Krauss, Nemoto and Pekkanen, 2012).

Notes (hide):

11: Krauss, E., Nemoto, K. and Pekkanen, R. (2012) ‘Reverse Contamination : Burning and Building Bridges in Mixed-Member Systems', Comparative Political Studies, 45 (6), pp. 747–773.

The two kinds of representatives in multi-member electoral systems are commonly referred to as two kinds of 'electoral mandate' who have differing electoral incentives which might lead to different kinds of behaviour. Prior to the establishment of the Scottish Parliament Henry McLeish (then-Scottish Office Minister of State) noted that he hoped to avoid two tiers of MSPs, but that there might be role differentiation and 'list MSPs might get more involved in policy and less in constituency matters' (Cowley and Lochore, 2000, p 176). This has generally been borne out, with Battle (2011) finding a statistically significant difference in effectiveness on committees, with list MSPs spending more of their time on committee activity and attending more meetings.

Notes (hide):

12: Cowley, P. and Lochore, S. (2000) ‘AMS in a cold climate : The scottish parliament in practice', Representation, 37 (3-4), pp. 175–185.

13: Battle, M. (2011) ‘Second-Class Representatives or Work Horses ? Committee Assignments and Electoral Incentives in the Scottish Parliament', Parliamentary Affairs, 64 (3), pp. 494–512.

The differing electoral incentives for different mandates might in some systems lead to differing levels of loyalty of the representative to the party. Sieberer (2010) shows that in Germany district MPs are more likely to defect on party votes than list MPs, with constituency MPs being 1.5x more likely to break from the party line. Thames (2005) argues that the mandate divide in terms of different legislative behaviour depends on the level of institutionalisation of the party system - examining the Hungarian National Assembly the Russian Duma and the Ukrainian Rada he found evidence for a mandate divide in the Russian Duma, which had the weakest institutional parties.

Notes (hide):

14: Sieberer, U. (2010) ‘Behavioral consequences of mixed electoral systems : Deviating voting behavior of district and list MPs in the German Bundestag', Electoral Studies. Elsevier Ltd, 29 (3), pp. 484–496.

15: Thames, F. C. (2005) ‘A House Divided: Party Strength and the Mandate Divide in Hungary, Russia, and Ukraine', Comparative Political Studies, 38 (3), pp. 282–303.

Kerevel (2010) examining the behaviour of members of Mexico's Chamber of Deputies found constituency members spent more time on 'credit-claiming activities' to benefit constituents while PR members 'spent more time drafting legislation that serves national party interests', while being little different in party discipline. Jun and Hix (2010) point out that effects on party loyalty aren't strictly a result of the mandate type but that this is an indirect result of candidate selection practises. Korea counter-intuitively has list members who are more independent and less loyal than constituency members. This results from very safe constituency seats that the party can control access to, and the habit of using list seats to draw in national figures who would only expect to serve one term (and hence are less bound by party control) .

Notes (hide):

16: Kerevel, Y. (2010) ‘The legislative consequences of Mexico's mixed-member electoral system, 2000-2009', Electoral Studies. 29 (4), pp. 691–703.

17: Jun, H.-W. and Hix, S. (2010) ‘Electoral Systems, Political Career Paths and Legislative Behavior: Evidence from South Korea's Mixed-Member System', Japanese Journal of Political Science, 11 (2), pp. 153–171.

Different mandates mean representatives can develop different conceptions of what it means to do their job. Bradbury and Mitchell (2007, p.125) found that constituency MSPs were 'more likely to rate various forms of local representations as "very important" than list members, while list members were more likely to rate working on parliamentary committees as "very important"'. In Wales, almost all 'constituency and list AMs considered helping to solve constituents' problems to be "very important", exceeding the perceived importance of all of the other roles open to members' (Bradbury and Mitchell, 2007, p 125-126). These figures are based on surveys in 2004 (and so a short time into the existence of these bodies) but are reflective of an immediate divide in terms of beliefs and behaviour of list representatives in the developed bodies.

Notes (hide):

18: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

19: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

Lundberg (2006) found that constituency MSPs were spending more time helping voters deal with government agencies, while list MSPs were spending more time in contact with interest groups (2006, p. 67). Bradbury and Mitchell (2007) found constituency MSPs were spending on average half a day more on casework than list MSPs (but there was no significant difference between AM mandates in Wales), and Parker and Richter (2016, p. 28) found that 'constituency based members are more likely to hold surgeries and less likely to attend to legislative affairs, while the [reverse] is true for regional members'. Similarly, constituents were treating their different representatives differently, with list members in Scotland and Wales receiving lower median communications from individual constituents, but a higher mean average, reflecting that some list members were contacted far more often (Bradbury and Mitchell, 2007).

Notes (hide):

20: Lundberg, T. C. (2006) ‘Second-class representatives? Mixed-member proportional representation in Britain', Parliamentary Affairs, 59 (1), pp. 60–77.

21: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

22: Parker, D. C. W. and Richter, C. M. (2016) ‘The Effect of Mixed Member Electoral Systems on the Development of Representational Styles: The Case of the Scottish Parliament', in Political Studies Association Annual Conference.

23: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

This mirrors findings in other countries. In 1999 Klingemann and Wessels (1999) found that 25% more German constituency MPs than regional MPs aimed to represent all citizens of the constituency. Almost all constituency MPs had contact with people from their district at least once a week, while only 84% of regional MPs did. Indeed based on an early 2000s survey, 'on indicators of constituency and partisan orientation that measure behaviour, there were more significant differences between constituency and list representatives in Germany than in Scotland and Wales' (Lundberg, 2003, p.21). Patzelt (2007) found that there is a 'fairly uniform core to the role of a German MP' but that the '"level of parliament" and the "type of seat" are factors that shape quite significantly the overall role of different MPs' with constituency members being 'closer to citizens and less dependent on party structures than list members' (2007, pp. 66-67).

Notes (hide):

24: Klingemann, H.-D. and Wessels, B. (1999) ‘Political consequences of Germany's mixed-member system: Personalization at the grass-roots?', WZB Discussion Paper, (No. FS III 99-205).

25: Lundberg, T. C. (2003) ‘The Impact of Mixed-Member Proportional Representation in Scotland and Wales: Lessons from Germany?', Keele European Parties Research Unit, (Working Paper 20).

26: Patzelt, W. J. (2007) ‘The Constituency Roles of MPs at the Federal and Länder Levels in Germany', Regional and Federal Studies, 17 (1), pp. 47–70.

These kind of role distinctions hold when looking at MEPs elected to the European Parliament (where different countries use different electoral system), Bowler and Farrell (1993) found that MEPs elected from individual districts had higher levels of contact with voters than those elected from regional-lists (and those were again higher than those elected from national list).

Notes (hide):

27: Bowler, S. and Farrell, D. M. (1993) ‘Legislative Shirking and Voter Monitoring: Impacts of European Parliament Electoral Systems upon Legislator-Voter Relationships', Journal of Common Market Studies, 31 (1), pp. 45–69.

Complications in examining role mandate

Contamination

The idea of 'contamination' complicates the picture of a divide between the mandates. As Crisp (2007) defines it, contamination is the idea that 'the existence of one tier prevents legislators from the other tier from behaving as if they were elected in a "pure" system made up of their tier alone'. Having a position on a list might affect how a candidate behaves (and how voters perceive them) in both their campaigning and representative activities, in that they may tend towards higher levels of party loyalty and interest in internal party politics. Alternatively, standing in a constituency might lead the same candidate to adopt different behaviours that cultivate closer links to their community.

Notes (hide):

28: Crisp, B. F. (2007) ‘Incentives in Mixed-Member Electoral Systems: General Election Laws, Candidate Selection Procedures, and Cameral Rules', Comparative Political Studies, 40 (12), pp. 1460–1485.

A candidate standing simultaneously in a constituency and on a party list is called 'dual candidacy'. Papp (2016) groups representatives in mixed systems into three tiers: 1) those elected in a Single Member District (SMD), 2) those elected through a list but who stood in an SMD district, and 3) those elected through a list who did not stand in an SMD. Examining the Hungarian National Assembly, she found that it was the second tier that asked the largest number of questions about local issues. Papp also found that previous experience in an SMD district only led to a greater constituency orientation when elected as a constituency representative. Olivella and Tavits (2013) (also looking at the Hungarian National Assembly) found that when a legislator moved from a List to SMD seat, they rebelled against the party more and became more constituency oriented, but found no equivalent for representatives taking the reverse journey. They suggest this as a different kind of contamination, where the personal electoral history of a candidate changes how adaptable they are to institutional incentives. While there are discrepancies between this and Papp's finding, both found different behaviour for representatives elected by list who had either tried, or previously had been elected, as an SMD representative.

Notes (hide):

29: Papp, Z. (2016) ‘Shadowing the elected. Mixed-member incentives to locally oriented parliamentary questioning', The Journal of Legislative Studies, 22 (2), pp. 216–236.

30: Olivella, S. and Tavits, M. (2013) ‘Legislative Effects of Electoral Mandates', British Journal of Political Science, (44), pp. 301–321.

Contamination can also change the behaviour of voters. Hainmueller and Kern (2008, p. 36) found that the incumbency bias in Germany's SMD elections also had an effect on additional proportional representation votes in the district, which 'possibly lead to the close victory of the SPD/Green coalition in the 2002 Bundestag election'. Krauss, Nemoto and Pekkanen (2012) found that contamination had different effects for MMM and MMP systems, with candidates and parties trying to encourage different behaviours from voters. Examining Japan and New Zealand, they found that the MMP system in New Zealand encouraged parties to rank key members highly in lists, and that SMD voters found a high-ranking local figure desirable. In MMM systems the opposite would occur. For Japan they found a 'reverse-contamination' which created an incentive for SMD candidates to 'burn bridges' with the party to minimise the chance of election in the PR tier. Their argument was that a lower placement on the list might encourage more local voters in the SMD contest, as if supporters believed that a member would be elected anyway, they would be less likely to turnout to vote for that member.

Notes (hide):

31: Hainmueller, J. and Kern, H. L. (2008) ‘Incumbency as a source of spillover effects in mixed electoral systems: Evidence from a regression-discontinuity design', Electoral Studies, 27 (2), pp. 213–227.

32: Krauss, E., Nemoto, K. and Pekkanen, R. (2012) ‘Reverse Contamination : Burning and Building Bridges in Mixed-Member Systems', Comparative Political Studies, 45 (6), pp. 747–773.

One effect of cross-system contamination in the UK is the ability for an electoral conflict to continue after the election, where a candidate defeated in a single constituency may still be elected on the basis of their position on the party list. Regional MSPs trying to become a constituency MSP have an incentive to perform constituency service to achieve better name recognition in the next election. They might even concentrate their efforts in one constituency of their region to better help their future election chances in that area.

This is called 'shadowing' - when a regional representative tries to raise their profile in an area to have a better chance at taking the seat. Bradbury and Russell (2005, p. 6) found that 'Constituency MSPs, predominantly Labour, have strong perception of targeted activity for electoral purposes by list members' while list members 'accept that such perceptions have some truth to them but refute their general validity or significance'.

Notes (hide):

33: Bradbury, J. and Russell, M. (2005) ‘The Local Work of Scottish MPs and MSPs: Effects of Non-coterminous Boundaries and AMS', University of Wales Seansea, The Constitution Unit (UCL).

There is evidence that regional MSPs do make decisions based on the vulnerability of the local constituency MSP. Carman and Shephard (2007) found that the marginality of the seat, the population density and the unemployment rate were significant predictors of where list members would put their offices, with a 50 percent probability that the most marginal constituency would also contain an office of a regional list member (as opposed to 6 percent for the least marginal).

Notes (hide):

34: Carman, C. and Shephard, M. (2007) ‘Electoral Poachers? An Assessment of Shadowing Behaviour in the Scottish Parliament', The Journal of Legislative Studies, 13 (4), pp. 483–496.

The negative attitude of constituency MSPs towards dual candidacy in the UK (especially in interviews closer in time to the establishment of the devolved bodies) might simply reflect attitudes grown in one context being transposed to another, rather than an enduring feature of the system. In New Zealand, Mcleay and Vowles (2007) found that MPs who had been MPs prior to MMP were 'less tolerant of the new roles emerging under the mixed member rules'. The system is well established in Germany and while half of the constituency members surveyed said they were being shadowed and 89.9% of list members said they were shadowing, constituency MPs did not express that there was anything untoward about this behaviour. List members were just doing their job (Lundberg, 2003).

Notes (hide):

35: Mcleay, E. and Vowles, J. (2007) ‘Redefining Constituency Representation : the Roles of New Zealand MPs Under MMP', Regional and Federal Studies, 17 (1), pp. 71–95.

36: Lundberg, T. C. (2003) ‘The Impact of Mixed-Member Proportional Representation in Scotland and Wales: Lessons from Germany?', Keele European Parties Research Unit, (Working Paper 20).

Different Resources

While there is no formal requirement on Westminster MPs to perform constituency service, the Code of Conduct for the Scottish Parliament developed a specific requirement to 'be accessible to the people of the areas for which they have been elected to serve and to represent their interests conscientiously' (Gay, 2005, p. 61). There was no legal contrast in terms of what constituency and regional MSPs/AMs were allowed to do regarding casework, but differences did develop in the resources they would be given to do it (Bradbury and Mitchell, 2007, p 119).

Notes (hide):

37: Gay, O. (2005) ‘MPs Go Back to Their Constituencies', The Political Quarterly, 76 (1), pp. 57–66.

38: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

Only a few MMP systems have formal differences in resources. In the National Assembly for Wales the list members were given the same salary, the same time allocation for constituency work and the same allowance to establish an office as constituency members. In Scotland there was a recognition of difference. Where a party had multiple regional MSPs there was reduced funding (with the assumption that they would share office space). Regional Members were also required to inform the constituency members of constituency work and prove they were working in more than two consistencies (Bradbury and Mitchell, 2007, p 119), although in 2008 it seemed that this was often not happening (Lundberg, 2012, p. 13). While at the time this was unique in MMP systems, New Zealand implemented a version several years later allocating $NZ66,000 for constituency MPs to set up two offices while regional MPs were only allocated $NZ42,000 for one office (Mcleay and Vowles, 2007, p. 74).

Notes (hide):

39: Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

40: Lundberg, T. C. (2012) ‘Tensions Between Constituency and Regional Members of the Scottish Parliament Under Mixed-Member Proportional Representation: A Failure of the New Politics', Parliamentary Affairs, pp. 1–20.

41: Mcleay, E. and Vowles, J. (2007) ‘Redefining Constituency Representation : the Roles of New Zealand MPs Under MMP', Regional and Federal Studies, 17 (1), pp. 71–95.

This system of differential funding was implemented by the Labour-Lib Dem coalition in the first Scottish Parliament, at least partially in an attempt to undermine the ability of regional MSPs to compete on casework (of concern because the coalition was primarily made up of constituency MSPs) (Lundberg, 2012). The logic behind this was that regional MSPs would have less casework from constituents, and so required less office funding. The underlying political hope was that it would be a self-fulfilling prophecy and undercut the ability of regional MSPs to compete on casework. Early guidance appeared to the public to help citizens determine which MSP they should contact. A public information leaflet - 'You and Your MSPs' - advised they are free to approach any of their MSPs with problems but that the code of conduct 'envisages that in most circumstances, individual constituents will approach the relevant constituency MSP' (Lundberg, 2003, p. 19).

Notes (hide):

42: Lundberg, T. C. (2012) ‘Tensions Between Constituency and Regional Members of the Scottish Parliament Under Mixed-Member Proportional Representation: A Failure of the New Politics', Parliamentary Affairs, pp. 1–20.

43: Lundberg, T. C. (2003) ‘The Impact of Mixed-Member Proportional Representation in Scotland and Wales: Lessons from Germany?', Keele European Parties Research Unit, (Working Paper 20).

This reflects that, while different mandates are a result of the electoral system, the exact expected responsibilities and resources emerge from political concerns. Different bodies will end up with different solutions, and an attitude being present does not mean it will be successful (for example. the failed 2008 push to reduce the office staff allowance of regional MSPs (Lundberg, 2012), or lasting (the Welsh ban on dual candacies). Restrictions on resources might reinforce electoral incentives with more direct practical restrictions.

Notes (hide):

44: Lundberg, T. C. (2012) ‘Tensions Between Constituency and Regional Members of the Scottish Parliament Under Mixed-Member Proportional Representation: A Failure of the New Politics', Parliamentary Affairs, pp. 1–20.

Overlap between mandate and government

When identifying a behaviour of 'constituency' representatives, it is likely that this also refers to the behaviour of 'government' representatives (as the government is likely to be mostly based in constituency seats). Cowley and Lochore (2000) argue the link between mandate type and party might sometimes mean that conclusions drawn about one might possibly be caused by the other, '[regional] MSPs have a slightly lower participation rate in the Parliament's votes than do constituency MSPs. But this is almost certainly caused by the fact that the two parties with the highest participation rates are Labour and the Liberal Democrats, both of which predominantly consist of MSPs from constituencies' (Cowley and Lochore, 2000, p. 177). Herron found that in Ukraine's Verkhovna Rada 'legislator's seat type [mandate] does not directly influence factional cohesion when party membership is taken into account' (Herron, 2002, p. 378).

Notes (hide):

45: Cowley, P. and Lochore, S. (2000) ‘AMS in a cold climate : The scottish parliament in practice', Representation, 37 (3-4), pp. 175–185.

46: Cowley, P. and Lochore, S. (2000) ‘AMS in a cold climate : The scottish parliament in practice', Representation, 37 (3-4), pp. 175–185.

47: Herron, E. S. (2002) ‘Electoral Influences on Legislative Behavior in Mixed-Member Systems : Evidence from Ukraine's Verkhovna Rada', Legislative Studies Quarterly, 27 (3), pp. 361–382.

They might be more effective in response to constituent problems if they have greater power to intervene on decisions via the ear of friendly ministers. Conversely, they may also receive more communication dissatisfied with party/government policy than would otherwise be the case, which may be less likely to get a response.

Hypotheses

This study investigates two hypotheses about how members of devolved bodies behave in the UK.

Mandates and responsiveness

While it is expected that some regional representatives may be very responsive to casework, across all regional representatives it is expected that the institutional incentives to focus on party priorities will lead to regional representatives being less responsive on average to messages from constituents than constituency representatives. If the difference is only apparent in the Scottish Parliament, it would suggest that resource difference plays a substantial role in driving the effect. Hypothesis one is therefore:

H1:Constituency representatives will have higher average responsiveness rates than list representatives across all three bodies.

Mandate as proxy for personal preference

A possible explanation for a difference in responsiveness is that the personal character of the representatives might be different, with party selection practice sorting candidates into those who have an affinity for constituency engagement. Policy-focused candidates might be put off by the work of winning over a constituency party, but happy to do other kinds of party work that is rewarded with a list position, and then to focus on committee and policy work over constituents when elected.

This would not invalidate any responsiveness difference, but would reframe it as party strategy rather than individuals reacting to different incentives.

This can be examined by looking at candidates who have occupied both kinds of mandates. There have been none of these in the London Assembly, and the ban on dual candidates in the NAfW from 2006-2014 means there have only been 4 AMs who have sat as both. However, in the Scottish Parliament there are 45 MSPs who have been both constituency and regional MSPs. This provides the opportunity for a natural experiment to detect whether the same people behave differently towards constituents when they hold different electoral mandates.

H2:For Scottish MSPs who have held both kinds of mandates, the same representatives will have higher average responsiveness rates when elected in a constituency mandate.

Methods

To test these hypotheses, this study extracted data from the results of a survey sent to users of an online email platform used to contact elected officials in the UK (WriteToThem.com). WriteToThem is a website created by the civic technology charity mySociety. It is designed to simplify the process of constituents contacting representatives.

To keep it focused on the constituent-representative connection, users cannot directly contact a representative by name, but must enter a postcode to show representatives for this area. While this doesn't represent an ironclad barrier to non-constituent communication - generally email communication through WriteToThem should represent a more constituent-focused set of communication than the representative's more general email inbox.

Two weeks after users send an email to their chosen representative, they are sent a survey to ask if their representative responded. If there is no response, the questionnaire is resent a week later. While this does not provide information about the quality of the response, it provides sufficient meaningful data to construct a response rate metric.

For emails sent from 2005 to the end of November 2017, this survey collected 20,388 responses for users contacting Members of the Scottish Parliament, with 5,399 responses for the NAfW, and 5,247 for the London Assembly. A survey of users contacting their national MP will also be used for validation. The date range for this survey is from 2005 to 20th June 2017 (to avoid an unusually short set of memberships after the 2017 election) and received 777,679 replies.

An additional complication is that WriteToThem lets users write to multiple regional representatives at once. For the bodies in question, regional group messages represent the majority of messages sent, while being a minority user action, because each action sends several emails. Only 17.67% of users writing to the Scottish Parliament sent a group message, but this accounts for 59.80% of emails sent, because each action triggers seven emails. As there are no survey responses collected for group messages, this should not directly impact comparisons between constituency and regional representatives who were written to directly. While it is possible that there is an indirect effect via increased overall work, this represents an average of 62 emails a year per regional representative (just over one a week), whereas a survey in 2004 showed that the average regional MSPs were receiving was 147.6 per week (with a median of 60) (Bradbury and Russell, 2005, p. 34). However, the total number of emails received will be included as a variable in the regression models to calibrate for any effect of regional members also receiving larger numbers of group messages.

Notes (hide):

48: Bradbury, J. and Russell, M. (2005) ‘The Local Work of Scottish MPs and MSPs: Effects of Non-coterminous Boundaries and AMS', University of Wales Seansea, The Constitution Unit (UCL).

WriteToThem data contains information about the seat type, and this was double checked against membership information of the respective bodies from EveryPolitician.org, as well as additional datasets for terms not covered by that dataset. These additional datasets were the source of gender and age information.

There have been minor changes to the survey questions over the whole time period. The survey response currently allows 'yes', 'no', 'unsatisfactory' and 'not expected'. These last two were introduced in October 2016 later and make up a small amount of responses (for the Scottish Parliament survey, these last two represented 146 and 58 respectively). 'Unsatisfactory' was read as 'yes' for the purposes of this research, responses for 'not expected' were dropped. The question text has always informed respondents to treat acknowledgement emails as 'no response'. In 2016 the wording was changed to provide separate links for 'it was only an acknowledgment' and 'no reply' - where both are counted as 'no reply' in the database.

Aggregating these responses into individuals creates a list of representatives with a corresponding 'responsiveness' percentage. Individuals are unique on each 'membership', so being re-elected, or changing party creates John Smith-1 and John Smith-2. For the Scottish Parliament, NAfW and Westminster Parliament, this information was integrated with a dataset stating which parties were in government to create an is_government variable for each representative.

A t-test (unequal variance) was employed to query if the difference between the averages of the two groups was sufficiently statistically significant to draw conclusions on the differing responsiveness of the two groups in the different devolved bodies.

Representative-memberships which had received no replies to surveys were ignored.

A multiple linear regression was then used to examine the effect while controlling for additional variables that might also explain the result. These factors were:

  • If the representative's party forms part of the government (as this might be affecting the volume and type of correspondence received).
  • The gender of the representative (following the suggested idea that female MPs had more constituency focus),
  • Their age at start of membership (checking for any generational effects in representatives),
  • The time since they were first elected at start of membership (accounting for the suggested change in responsiveness as representatives accumulated other responsibilities in their career).
  • How many emails they received through the platform (accounting for any effect that receiving more emails from WriteToThem has on responsiveness to emails received through WriteToThem).

As the membership is the unit of analysis, these are not independent as the same individual will be represented multiple times. Standard errors were thus clustered by unique individual. To exclude outliers with an 100% responsiveness rate resulting from very few surveys, individuals with less than four survey responses were excluded.

While the key question (effect of mandate type) cannot be examined in data on the Westminster Parliament, the effect of these additional factors will be validated by a regression against the survey results for people writing to their national MPs.

To check for possible selection effects leading to individuals with different priorities ending up in different roles, the t-test will be re-run restricted to just the group of MSPs who have been elected to both constituency and regional members to see if the same individuals behave differently responsive under different circumstances.

The regression will then be expanded with information about this group of crossover MSPs to examine if future roles are a guide to responsiveness. This will explore the possibility of contamination changing regional MSPs view of their role in advance of seeking election.

Results

Mean Difference

Using t-tests to check for a mean difference in responsiveness between constituency and list representatives in the Scottish Parliament, NAfW and London Assembly revealed a significant difference in all three cases.

Table 1 shows that constituency representatives have higher average responsiveness rates than regional representatives in all three bodies. That bodies with and without difference resources assigned to regional members show a difference is suggestive it is a result of the mandate divide rather than different resources.

% Mean Response RateConstituency%Regional Direct%Mean DifferenceT scorep-value

Scottish Parliament

59.09

46.04

-13.04

7.8590

<.0001****

Welsh Assembly

66.99

53.23

-13.76

4.9385

<.0001****

London Assembly

52.64

34.65

-17.99

5.2181

<.0001****

****<0.0001 ***<0.001 **<0.01 *<0.05

Table 1.Mean Regional Responsiveness

Regression Analysis

Using a multivariate regression analysis the component contribution of these factors for both Scotland and Wales was examined (with London excluded both for the lack of an executive drawn from the legislature and for having less data). Also included is the Westminster Parliament as an additional regression model to validate the results of the other two regressions.

Model 1 applies personal properties of representatives: 'age' (in years at start of the membership) and gender, the time since first elected (in years at start of membership) and the number of emails received by the representative in that membership. Model 2 adds the variables under investigation - if the member was part of a governing party and if they held a constituency seat.

Measures of significance on the change in R2 in Table 2 result from an wald test of each group of models..05

Predictor VariablesNAfW M1NAfW M2SP M1SP M2WP M1WP M2

(Constant)

87.22***

70.29***

48.11***

34.49***

48.11***

48.55***

[50.64,89.94]

[49.91,90.67]

[23.53,4 5.45]

[24.01, 44.97]

[42.04,55.06]

[42.02,55.08]

Age

-0.34

-0.32

-0.01

0.08

0.14

0.14

[-0.68, 0.05]

[-0.70, 0.07]

[-0.14, 0.29]

[-0.13, 0.28]

[-0.00, 0.28]

[-0.00, 0.28]

Gender

-0.71

-1.66

-7.84***

-7.48***

-7.12***

-7.21***

(0 = Male, 1 = Female)

[-10.46, 7.13]

[-10.32, 7.00]

[-11.57, -3.40]

[-11.34, -3.62]

[-9.76, -4.67]

[-9.78, -4.64]

No. of Emails

-0.05***

0.00

0.01

0.04***

0.01***

0.01***

[-0.02, 0.03]

[-0.02, 0.03]

[0.03, 0.05]

[0.02, 0.05]

[0.00, 0.01]

[0.00, 0.01]

Time in Parliament

-0.43

-0.61

1.27***

0.74***

-0.14

-0.15

(years)

[-1.30, 0.08]

[-1.27, 0.05]

[0.31, 1.16]

[0.31, 1.16]

[-0.31, 0.02]

[-0.31, 0.02]

Government Seat

-

-0.89

-

-6.19**

-

-0.85

[-11.31, 9.52]

[-9.95, -2.43]

[-2.66, 0.96]

Constituency Seat

-

18.62**

-

19.48***

-

-

[6.94,30.30]

[15.34,23.62]

Adjusted R2

0.08

0.16

0.10

0.23

0.04

0.04

Change in adj. R2

-

0.08**

-

0.13***

-

-0.00

N

144

144

511

511

2,146

2,146

****<0.0001 ***<0.001 **<0.01 *<0.05

Table 2.Regression results for Scottish Parliament, National Assembly for Wales and Westminster Parliament

While Gender, the Time in Parliament, the number of emails received, holding a Government Seat and Seat Mandate were significant at the five percent level or higher in the Scottish Parliament model, only the Seat Mandate and time spent in parliament was significant in the NAfW model. This could represent a genuine difference, or it could result from having less data for the NAfW.

The Westminster model finds the same signed effects for gender (female representatives are less responsive in both cases) and for if more emails affect responsiveness (small positive effect). Where they differ is if there is an effect on length of time a representative has spent in Parliament - with this returning a small positive effect year-on-year in the Scottish Parliament and a negative effect in Westminster.

Where the Scottish Parliament and NAfW models are in agreement is that holding a constituency seat has a large effect on responsiveness, with a 19.48% [CI: 15.34, 23.62] increase in the Scottish Parliament and a 18.62% [CI: 6.94, 30.30] increase in the NAfW. H1 is validated when controlling for other possible contributing factors.

Self-selection

To examine if the difference between regional and constituency MSPs can be attributed to a self-selection effect (where representatives inherently are likely to respond in different ways, and chose the legislative career that suits them), the mean difference analysis is limited to just the 45 MSPs who have 'crossed-over' and represented both kinds of mandates. 35 were initially in regional seats, with the remaining 10 starting in constituency seats. All only changed role once.

As different MSPs spend different time in each role, this analysis is limited to the membership before and after their role switch (making sure there are equal numbers of the same individuals on both sides - although the results show a similar pattern without this step). Table 3 shows the mean difference between constituency and regional memberships only appears for MSPs who were initially regional MSPs and later transitioned into constituency roles. MSPs who were initially in constituencies do not exhibit the mean difference.

% Mean Response RateConstituency%Regional Direct%Mean DifferenceT scorep-value

All

61.84 [57.66, 66.02]

49.48 [42.98, 55.97]

-12.36 [-19.79, -4.94]

3.2647

.0020**

Regional First

61.98 [57.39, 66.56]

44.91 [38.61, 51.21]

-17.06 [-24.39, -9.74]

4.5651

<.0001****

Constituency First

61.02 [48.64, 73.39]

56.11 [42.53, 69.70]

-4.90 [-20.82, 11.02]

0.6037

.5536

****<0.0001 ***<0.001 **<0.01 *<0.05

Table 3.Mean Regional Responsiveness Crossover

Table 3 shows that the same individuals show a change in responsiveness when moving from regional to constituency seats, and that the same isn't true for the inverse movement suggests this isn't an inherent feature of changing between the two roles. This may however reflect that very few MSPs have moved from constituency to regional seats (7 out of 10 showed some decline in responsiveness, with 3 showing an increase) and this is reflected in a wide confidence interval.

Re-running the Scottish Parliament regression (Table 4) on this reduced dataset showed this difference was robust against any effect of time in government, or changes in government before and after the switch.

The follow-up question would be if this subset of MSPs differ from their peers who stay in a single role. Returning to the complete Scottish Parliament regression, variables were added to check if an MSP being a crossover MSP, being a constituency MSP who would later become a regional MSP, and being a regional MSP who would later become a constituency MSP were predictive of email responsiveness. None of these variables showed a significant difference - indicating that controlling for the factors already expressed in the model, the regional and constituency MSPs who change roles do not exhibit different email responsiveness than their peers in the same role.

Predictor VariablesEstimateStd. Errort valuep value

(Constant)

54.12

13.82

3.92

<0.001***

[26.63,81.61]

Age

-6.34

4.51

-1.40

0.164

[-15.31,2.64]

Gender

-0.17

0.24

-0.71

0.480

(0 = Male, 1 = Female)

[-0.65,0.31]

No. of Emails

0.01

0.01

0.56

0.579

[-0.02,0.04]

Time in Parliament

0.55

0.42

1.31

0.194

(years)

[-0.29,1.38]

Government Seat

-3.00

4.72

-0.64

0.527

[-12.38,6.39]

Constituency Seat

15.84

4.89

3.24

0.002**

[6.10,25.58]

Adjusted R2

0.13

-

-

-

N

88

-

-

-

***<0.001 **<0.01 *<0.05

Table 4.Crossover MSP Regression Model

For the Scottish Parliament, H2 can be partially confirmed - the mean difference in responsiveness remains when looking at exactly the same individuals in different roles, but only when they changed roles in one direction.

These results do not conclusively address the self-selection issue but do suggest that personal preferences can interact with the role. As MSPs who would later be constituency MSPs (and be indistinguishable from that cohort) behave no differently when they are regional MSPs - this suggests that this group of MSPs adjust behaviour to fit the expectations of the role they are currently in.

However, that former constituency MSPs do not universally reduce their responsiveness when they become regional MSPs is suggestive that there is a personal component at work. While those moving from regional to constituency seats will be actively working towards this end (and this filters out any regional MSPs with no interest whatsoever in constituency work), those moving in the opposite direction will probably do so accidentally - losing their constituency election but being re-elected through the lists. This might include a selection of MSPs whose view of their role is less flexible, and so does not adjust through change of mandate.

Discussion

One possible factor shaping these results is that people email their representatives for a diverse set of reasons, and some categories of letter may be more likely to go to one kind of representative than another. For instance, constituents might seek out a politically similar MSP for a policy problem, but go to the local MSP with better name-recognition for a personal problem.

Constituency and regional representatives might have comparable reply rates to different kinds of letter, but regional representatives might be less likely to receive letters about local, practicable-actionable problems. That some former constituencies MSPs can maintain higher responsiveness rates suggests that this is not entirely a matter of content - but this group may also receive a different profile of communication as a result of increased name recognition.

As the content of communication is out of scope of this study (and is not stored by WriteToThem for privacy reasons), it is not possible to validate this one way or another without additional data. What can be stated is that either constituents treat members differently based on their mandate and send them different kinds of communication, or that representatives with different mandates treat the same kinds of letter differently.

Conclusion

This study has shown that there is a responsiveness gap between constituency and regional representatives in UK devolved bodies using MMP electoral systems. Constituency-based members had an average responsiveness to emails from constituents higher than regional members in the Scottish Parliament, National Assembly for Wales and the London Assembly. Regression analysis showed that this effect could be separated from alternative explanations such as being an effect of the parties in government, time since elected, or the number of messages received.

Narrowing the analysis to just MSPs who had held both mandates showed a divide when the same individuals changed roles - but only for MSPs migrating from regional to constituency seats. When information about crossover is introduced to the regression, future or past role changes do not appear as a factor shaping responsiveness.

This provides further evidence of role differentiation. As Parker and Richter (2016, p. 29) noted 'the creators of the Scottish Parliament did not abandon the question of how to represent to the vagaries of MSP choice; rather, they ensured that both types of representation would emerge naturally from the differing incentives regional and constituency MSPs face'. Different responses to constituent communication fits into other evidence suggesting regional MSPs spend less time on constituency service and more time on committee work.

The key remaining question is if these results show different responses to the same input, or if constituency and regional representatives are receiving different kinds of correspondence. This could be investigated by a study of casework in the devolved bodies that quantifies types and amounts of correspondence received from constituents.

This study raises questions about the ongoing development of different models of political representation in the UK. While the devolved legislatures consciously imported norms of constituent service from the Westminster Parliament, the differing electoral system creates career incentives for representatives that run counter to this model and encourage a different relationship with voters. From the constituent's point of view, they may exist in an area where different representatives compete, or where regional representatives are comfortable to leave constituent work with the constituency representatives and focus on interest groups. This combined with the overlapping representation to the Westminster Parliament paints a complicated picture of how constituent preferences filter up to the legislative process.

This is a pre-copyedited, author-produced version of an article accepted for publication in Parliamentary Affairs following peer review. The version of record is:

Alex Parsons, Rebecca Rumbul, Did They Write Back? A Mandate Divide in Response to Constituent Casework in Devolved Bodies, Parliamentary Affairs, Volume 72, Issue 3, July 2019, Pages 522–541 https://doi.org/10.1093/pa/gsy020

References:

Anagnoson, J. T. (1983) ‘Home Style in New Zealand', Legislative Studies Quarterly, 8 (2), pp. 157–175.

Battle, M. (2011) ‘Second-Class Representatives or Work Horses ? Committee Assignments and Electoral Incentives in the Scottish Parliament', Parliamentary Affairs, 64 (3), pp. 494–512.

Bowler, S. and Farrell, D. M. (1993) ‘Legislative Shirking and Voter Monitoring: Impacts of European Parliament Electoral Systems upon Legislator-Voter Relationships', Journal of Common Market Studies, 31 (1), pp. 45–69.

Bradbury, J. and Mitchell, J. (2007) ‘The Constituency Work of Members of the Scottish Parliament and National Assembly for Wales : Approaches , Relationships and Rules', Regional and Federal Studies, 17 (1), pp. 117–145.

Bradbury, J. and Russell, M. (2005) ‘The Local Work of Scottish MPs and MSPs: Effects of Non-coterminous Boundaries and AMS', University of Wales Seansea, The Constitution Unit (UCL).

Cain, B. E., Ferejohn, J. A. and Fiorina, M. P. (1984) ‘The Constituency Service Basis of the Personal Vote for U.S. Representatives and British Members of Parliament', The American Political Science Review, 78 (1), pp. 110–125.

Carman, C. and Shephard, M. (2007) ‘Electoral Poachers? An Assessment of Shadowing Behaviour in the Scottish Parliament', The Journal of Legislative Studies, 13 (4), pp. 483–496.

Cowley, P. and Lochore, S. (2000) ‘AMS in a cold climate : The scottish parliament in practice', Representation, 37 (3-4), pp. 175–185.

Crisp, B. F. (2007) ‘Incentives in Mixed-Member Electoral Systems: General Election Laws, Candidate Selection Procedures, and Cameral Rules', Comparative Political Studies, 40 (12), pp. 1460–1485.

Gay, O. (2005) ‘MPs Go Back to Their Constituencies', The Political Quarterly, 76 (1), pp. 57–66.

Hainmueller, J. and Kern, H. L. (2008) ‘Incumbency as a source of spillover effects in mixed electoral systems: Evidence from a regression-discontinuity design', Electoral Studies, 27 (2), pp. 213–227.

Heitshusen, V. (2005a) ‘Electoral Context and MP Constituency Focus in Australia, Canada, Ireland, New Zealand, and the United Kingdom', American Journal of Political Science, 49 (1), pp. 32–45.

Herron, E. S. (2002) ‘Electoral Influences on Legislative Behavior in Mixed-Member Systems : Evidence from Ukraine's Verkhovna Rada', Legislative Studies Quarterly, 27 (3), pp. 361–382.

Jun, H.-W. and Hix, S. (2010) ‘Electoral Systems, Political Career Paths and Legislative Behavior: Evidence from South Korea's Mixed-Member System', Japanese Journal of Political Science, 11 (2), pp. 153–171.

Kerevel, Y. (2010) ‘The legislative consequences of Mexico's mixed-member electoral system, 2000-2009', Electoral Studies. 29 (4), pp. 691–703.

King, G. (1991) ‘Constituency Service and Incumbency Advantage', British Journal of Political Science Polity American Political Science Review American Political Science Review American Politics Quarterly, 21 (8), pp. 119–128.

Klingemann, H.-D. and Wessels, B. (1999) ‘Political consequences of Germany's mixed-member system: Personalization at the grass-roots?', WZB Discussion Paper, (No. FS III 99-205).

Krauss, E., Nemoto, K. and Pekkanen, R. (2012) ‘Reverse Contamination : Burning and Building Bridges in Mixed-Member Systems', Comparative Political Studies, 45 (6), pp. 747–773.

Lundberg, T. C. (2003) ‘The Impact of Mixed-Member Proportional Representation in Scotland and Wales: Lessons from Germany?', Keele European Parties Research Unit, (Working Paper 20).

Lundberg, T. C. (2006) ‘Second-class representatives? Mixed-member proportional representation in Britain', Parliamentary Affairs, 59 (1), pp. 60–77.

Lundberg, T. C. (2012) ‘Tensions Between Constituency and Regional Members of the Scottish Parliament Under Mixed-Member Proportional Representation: A Failure of the New Politics', Parliamentary Affairs, pp. 1–20.

Mcleay, E. and Vowles, J. (2007) ‘Redefining Constituency Representation : the Roles of New Zealand MPs Under MMP', Regional and Federal Studies, 17 (1), pp. 71–95.

Moser, R. G. and Scheiner, E. (2005) ‘Strategic Ticket Splitting and the Personal Vote in Mixed-Member Electoral Systems', Legislative Studies Quarterly, 30 (2), pp. 259–276.

Norris, P. (1997) ‘The puzzle of constituency service', The Journal of Legislative Studies, 3 (2), pp. 29–49.

Olivella, S. and Tavits, M. (2013) ‘Legislative Effects of Electoral Mandates', British Journal of Political Science, (44), pp. 301–321.

Papp, Z. (2016) ‘Shadowing the elected. Mixed-member incentives to locally oriented parliamentary questioning', The Journal of Legislative Studies, 22 (2), pp. 216–236.

Parker, D. C. W. and Richter, C. M. (2016) ‘The Effect of Mixed Member Electoral Systems on the Development of Representational Styles: The Case of the Scottish Parliament', in Political Studies Association Annual Conference.

Patzelt, W. J. (2007) ‘The Constituency Roles of MPs at the Federal and Länder Levels in Germany', Regional and Federal Studies, 17 (1), pp. 47–70.

Sieberer, U. (2010) ‘Behavioral consequences of mixed electoral systems : Deviating voting behavior of district and list MPs in the German Bundestag', Electoral Studies. Elsevier Ltd, 29 (3), pp. 484–496.

Thames, F. C. (2005) ‘A House Divided: Party Strength and the Mandate Divide in Hungary, Russia, and Ukraine', Comparative Political Studies, 38 (3), pp. 282–303.