Bargaining in legislatures, portfolio allocation, and the electoral costs of governing

Abstract

What motivates political parties in the legislative arena? Existing legislative bargaining models stress parties’ office and policy motivations. A particularly important question concerns how parties in coalition government agree the distribution of cabinet seats. This article adds to the portfolio allocation literature by suggesting that future electoral considerations affect bargaining over the allocation of cabinet seats in multi-party cabinets. Some parties are penalised by voters for participating in government, increasing the attractiveness of staying in opposition. This ‘cost of governing’ shifts their seat reservation price – the minimum cabinet seats demanded in return for joining the coalition. Results of a randomised survey experiment of Irish legislators support our expectation, demonstrating that political elites are sensitive to future electoral losses when contemplating the distribution of cabinet seats. This research advances our understanding of how parties’ behaviour between elections is influenced by anticipation of voters’ reactions.

Keywords:

• Government formation
• legislatures
• Gamson’s Law
• electoral costs

Acknowledgements

Earlier versions of this paper were presented at the Conference on Institutional Determinants of Legislative Coalition Management, Tel Aviv, November 2015, the Essex Comparative Political Economy Seminar and at APSA 2017. I wish to thank participants and in particular Cristina Bucur, Giannetti Daniela and Indridi Indridason for very helpful comments.

Notes

1. Research on the economic vote (the notion that positive national economic performance results in the incumbent government being re-elected, while negative economic performance leads voters to vote against the incumbent government) has tended to look at the overall electoral fortunes of the incumbent government or the performance of the party of the prime minister. The clarity of responsibility literature (the economic vote is conditional on voters’ ability to assign credit or blame) views coalition government as an obstacle to accountability because voters are unable to reward or punish a single party (Hobolt et al. 2013; Powell and Whitten 1993). In this paper, we explore the bargaining consequences where electoral costs arising from joining the coalition are pooled unequally amongst government parties.

2. Following the formation of the first post-World War II coalition government in Britain, support for the Liberal Democrats dropped in the very first post-government formation opinion poll. In contrast, Conservative Party support remained static in the opinion polls (Whitaker 2011: 1164).

3. Gamson’s Law may also predict the distribution of cabinet seats between factions within a party (Ceron 2014; Mershon 2001).

4. We follow the convention in coalition studies by treating policy as a payoff of office. As Müller and Strøm (1999: 8) note, ‘policy pursuit is typically presented as a supplement to, rather than a substitute for, office seeking’. Our focus here and throughout this paper is on the quantitative distribution of cabinet seats. We do not consider qualitative issues, such as the fact that some parties may be particularly keen to secure certain portfolios, such as a green party seeking the environment portfolio.

5. Note that Δv_i need not refer to losses but could refer to the expected variation of votes, which can be negative or positive.

6. We assume, as discussed below, that membership of the government may result in electoral costs but that electoral costs do not increase with the number of cabinet seats held by the party. We also assume that parties care equally about electoral losses while acknowledging that in the real world, the same electoral loss from government participation can be more or less costly to a party depending on how much the party is policy-oriented or vote-oriented.

7. Exploring the causes of coalition termination, Lupia and Strøm (1995) argue that positive electoral expectations actually enhance bargaining power. Rejecting the conventional wisdom that electorally popular incumbent parties desire to call an election, they argue that ‘a party with favourable electoral prospects will also consider the option of extracting advantages through non-electoral means (e.g. bargaining with parties that have less favourable prospects)’ (Lupia and Strøm 1995: 655). Thus, the prospects of electoral gain cause a party to reopen bargaining during the lifetime of a coalition. Although their argument appears opposite to this paper’s argument, the logic is not contradictory given their focus on incumbent governing parties’ behaviour related to breaking the coalition once, in effect, electoral costs have been awarded.

8. Moreover, the politicians who form a critical component of any political party are highly sensitive to re-election concerns and what it takes to get re-elected, as a well-developed literature in legislative studies attests to (for an overview see Andre et al. 2016).

9. http://johnbruton.com/coalition-experience-in-ireland-what-works, accessed 29 March 2017.

10. This makes sense given that Baron and Ferejohn’s base model focuses on the allocation of particularism in a legislature. For each player, only benefits are derived from receiving a slice of that pie. This is not the case, we suggest, for political parties considering whether or not to enter a coalition government.

11. Parties are assumed to be unitary actors. However, as Martin (2016) has argued, the costs and rewards of governing are not necessarily shared equally between legislators within governing parties. Instead, perquisites of ministerial office may be used by ministers in preferential electoral systems to insulate themselves electorally from any voter displeasure with the party.

12. Employing observational data to test our argument creates significant identification issues in terms of causal inference. For instance, those circumstances in which ‘office-shy’ parties eventually get into government might be situations in which, for whatever (non-)observable reason, there is no other alternative for the formateur but to seek the support of these parties. Then, the ‘office-shy’ party, knowing that, may demand extra cabinet seats. In such circumstances, the bargaining context may be the omitted variable actually explaining why these parties are able to extract more.

13. While the lower chamber comprised 166 members, at the time of the survey experiment one seat was vacant.

14. Under the Labour Party Constitution, the ‘approval of the Party Conference, which can be given only on the recommendation of the Party Leader, is required for the Party’s entry into Government’ (source: https://www.labour.ie/party/constitution/).

15. I am very grateful to a reviewer for alerting me to this point.

16. http://www.oireachtas.ie/ViewDoc.asp?DocId=-1&CatID=138.