Reconciling Parliamentary Size with Personalised Proportional Representation? Frontiers of Electoral Reform for the German Bundestag

Abstract

In the 2017 Bundestag election, the recently reformed mixed-member proportional system (MMP) produced an enlarged parliament with 111 additional seats. As could happen in the future, the Bundestag may even increase in size, so the parliamentary parties committed themselves to a renewed reform in order to curb seat enlargements. This was nonetheless to be done without major changes of the existing MMP. The article analyses the frontiers of such “minimally invasive” reform. Each modification of “personalised PR” that reliably maintains the Bundestag size would compromise one or other of its fundamental features: personalisation or proportionality. We illustrate this by exploring the effects of various single-member district (SMD) reforms that are widely considered the “least harmful” minimally invasive approach. Comprehensive simulations based on survey data demonstrate that the SMD share would have to decrease substantially for maintaining parliamentary size as reliably as two-member districts would do. Thus, any SMD reform would either substantially affect the MMP’s conventional personalisation or curb seat enlargement less effectively.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

Notes

1 Major electoral reforms included the introduction of the two-vote system in 1953 and the national seat allocation in 1956. Thereafter, only the PR formula was changed from d’Hondt to Hare-Niemeyer in 1985 and to Sainte-Laguë/Schepers in 2008. Moreover, in the first post-unification election of 1990 the five-percent threshold was applied separately to West and East Germany (Saalfeld 2005, 201–212).

2 This was supposed to be reached by cutting the connection between the Länder lists of each party (“regionalised MMP”; see Weinmann and Grotz 2020). However, as a final nationwide calculation was included – to avoid disproportionate disadvantages for the FDP –, the initial plan was thwarted and negative voting weight continued.

3 Only some Christian Democrats publicly opted for a mixed-member majoritarian system that would completely separate the SMD seats and the PR seats (Krings 2019). However, as this proposal would massively favour the Christian Democrats only, it would not have any chances to find the other parties’ support.

4 This applies to all MMP including those in the German Länder and elsewhere (see Weinmann and Grotz 2020).

5 This is functionally equivalent to nation-wide lists which would also deplete all list seats if the party wins a number of surplus seats greater or equal to its PR seats.

6 A change from plurality to absolute majority in SMDs or from the two-vote to a single-vote system might lead to fewer surplus seats, but could also increase their number. Therefore, these reform options would not effectively prevent enlargements.

7 In mid-2020, both CDU and CSU for the first time proposed a minor reduction of SMDs (combined with a partial toleration of seven surplus seats; FAZ 2020a)

8 To obtain separate values for CDU and CSU – which is necessary for the simulations, but not provided by nationwide polls –, we use polls in Bavaria (for Bundestag and Land elections, source is again Wahlrecht.de) and interpolation to approximate the CSU share. For readers interested in the more recent past, the Online appendix shows the results for the period 2013–2020 (N = 151).

9 The formula for simulating first votes is: first vote share (in a district) 2017 + own party’s electoral trend + splitting rate to be simulated * supporter party’s simulated second vote share – splitting rate 2017 * supporter party’s second vote share 2017.

10 Without surplus seats, the actual electoral system leads to a random scatter with an average oversize of around 20 seats. Thus, when calculating with the initial distribution, it is impossible to systematically analyse the potential of reform options to reach the regular size.

11 With only one district, regional strongholds play no role at all. For parties other than the largest, an increasing number of districts means more opportunities to win at least one of those, while districts also become smaller and more local thereby increasing the probability that the supporters of those parties are concentrated in such districts.

12 When we employ the risky simulation method, we find 42 cases of preserving the regular size for this reform option (see Figure A1 and Table A1).

13 Simulating this option includes multiple assumptions at once potentially leading to a multiplication of estimation errors. However, as we improved the simulation method for the reduction of SMDs and the method for TMDs is rather conservative, the combined simulation should not be too risky.

Additional information

Notes on contributors

Philipp Weinmann

Philipp Weinmann is postdoctoral lecturer at the Helmut-Schmidt-University / University of the Federal Armed Forces Hamburg. His research interests include electoral systems (with a focus on mixed-member proportional systems), direct democracy and quantitative research methods.

Florian Grotz

Florian Grotz is Professor of Comparative Government at the Helmut Schmidt University/ University of the Federal Armed Forces Hamburg. His research interests include political institutions and party governments in European democracies.