FOTO CREDIT: Shutterstock
This week’s “meaningful vote” in the House of Commons marks one of the most remarkable losses a British prime minister ever had to experience. 68.1% of all representatives voted against the deal negotiated by May’s government with the European Union. MPs rejecting the Brexit deal justify this with very diverse reasons including demands for re-negotiations with the EU to reach a better deal, requests for a more dramatic break in EU-UK relations, concerns about the future of the Irish border or hopes for a second referendum. Even within her Conservative party, the prime minister experienced considerable opposition with 37.6% of the MPs voting against the EU (withdrawal) Act. Such extensive disapproval within the prime minister’s party is exceptional for a Westminster democracy, in which the relationship between government and parliament is usually characterized by a principal-agent relationship (Strøm 2000, Strøm, K. & Bergman, T. 2011): Parliamentary majorities delegate the power to make politics (which they received from the people) to the government. In this brief blog contribution, I aim to identify which conservative MPs were most likely to vote against their own government’s Brexit proposal. Were representatives with certain social characteristics and contextual settings in their districts more likely to vote ‘no’?
I looked into the characteristics of the conservatives MPs who supported and rejected the European Union (Withdrawal) Act main Motion on Tuesday. For that purpose, I merged vote data provided by the House of Commons (2019) with information on MPs’ individual characteristics (terms served, age, gender, ethnicity) and their districts attributes (closeness of race 2017, % leave in Brexit referendum, population density, % population older than 45 years, % population declaring to be ethnically white British). The data sources I used are described in more detail in Endnote .
Table 1 provides t-tests comparing the 195 conservative MPs supporting their government’s bill to the 119 co-partisans voting against it. The groups of ‘yes’ and ‘no’ voters show some systematic differences in their individual characteristics: ‘No’ voters tend to be older (+2.16 years), slightly more frequently male (+6%-points), and more experienced in parliament (elected 0.39 terms earlier) than ‘yes’ voters. The only social background factors that (at the first glance) seems to be of little relevance is belonging to an ethnic minority. The contextual setting in MPs districts further impacts the decision to vote for or against the EU (withdrawal) Act: Those rejecting the motion were elected in constituencies with slightly higher ‘leave’ support in the Brexit referendum (+0.79%-points), in closer races (-2.59%-points distance between winner and first looser), more urban districts with higher population density (+4.03 habitants per km2), a younger population (-1.09%-points habitants over 45), and lower shares of habitants declaring to be ethnically white British in the last census (-2.58%-points). Since many of these factors are eventually correlated, Table 2 provides a logistic regression model predicting the chances to reject the bill for Conservative MPs based on these variables. As Model 1 indicates, the directions of the relationships stay the same, with the exemption of the effect of minority membership and the ethnic composition of the district. In the following, I focus the discussion on three factors I find particularly interesting: Age, gender, and minority belonging, as well as their interactions with the district’s support for Brexit in the 2016 referendum and the closeness of race in the last general election.
Tab. 1: T-tests comparing conservative MPs voting for and against the EU (withdrawal) Act.Annotations: With * p < 0.10, ** p < 0.05, *** p < 0.01.
Age unfolds a u-shaped impact on the likelihood to vote against the proposed policy. Figure 1 (a) displays the effect visually. The youngest and oldest Conservative MPs are most likely to reject May’s Brexit deal, while middle-aged representatives more frequently follow the party line. Two causal mechanisms might be at play here. Firstly, older MPs might be less concerned about future career endeavors, which allows them to toe the party line, while middle-aged legislators might want to show their loyalty to the party to possibly reach top positions in the party and government. Secondly, the youngest MPs (while also having a future career to take into account) might be more likely to protest the establishment similar to the way young voters tend to engage in protest voting in general election.
Women are in general less likely to reject their government’s proposal. Interestingly, the relationship between gender and voting behavior is shaped by the support for the ‘leave’ campaign in the Brexit referendum (as Model 2 in Table 2 reveals), meaning that women react differently to high proportions of leave voters in their electorate than men. Figure 2 (b) shows the predicted probability of male and female Conservatives to reject May’s Brexit proposal depending on the referendum results. The more citizens wanted to leave the EU in their district, the more likely were male representatives to deem the government’s way to realize this endeavor inappropriate. Women, in turn, had a higher likelihood to support May’s strategy to withdraw from the EU, in case large numbers of citizens in their electorate wanted to leave the EU. This finding either suggests crucially different evaluations of the failed Brexit deal by male and female representatives or reveals gender differences in perceptiveness for May’s pleas to support the deal and, thus, in policy style.
MPs belonging to ethnic minorities (a total of 19 in the Conservative party) were 63% more likely to reject the motion than British-origin Conservatives. However, this effect is driven by those elected in rather save seats (as Model 3 reveals). Figure 2 (c) visualizes the interaction effect between minority belonging and closeness of race. Minority representatives elected in close races tended to support the EU (withdrawal) Act, while those who won clear victories at the last election tended to toe the party line and reject it. For British-origin representatives, the relationship is just the other way around. The saver their seat, the more likely are they to support their government’s proposal. This insight might be explained by different worries about punishment by the party. Minority MPs might be more concerned about their re-nomination than non-minority representatives, not daring to work against their parties interests if their seat is not save.
Tab. 2: Logistic regression predicting the chance to reject the EU (withdrawal) Act with regional fixed effects.
Annotations: Figures show exponentiated coefficients; with * p < 0.10, ** p < 0.05, *** p < 0.01; All models include regional fixed effects.
Fig.1: Predicted probability to reject the motion by (a) age, (b) gender and % leave in Brexit referendum, (c) MPs ethnicity and closeness of race in 2017 election.Annotations: (a) Figure based on Model 1 in Table 2; (b) Figure based on Model 2 in Table 2, (c) Figure based on Model 3 in Table 2.
The “meaningful vote” is not only an interesting case for British and European politics, but actually bears the potential to shed light on party discipline as a broader phenomenon. It provides a unique example to study representatives’ voting behavior in Westminster democracies in which representatives tend to vote according to party blocs. Perhaps most strikingly, social background characteristics unfold different effects on the likelihood to toe the party line depending on the contextual setting in the district. Nevertheless, one has to be aware that the characteristics discussed in this contribution explain only about 7% of the variation in the voting behavior, meaning that other factors than social background characteristics and district attributes matter just as much to understand the outcomes of the Brexit vote in the House of Commons.
By Corinna Kroeber in January 2019
 For a detailed discussion of the role of parliaments in Westminster democracies see an interesting contribution by Matthew Shugart in Fruits and Votes (available at https://fruitsandvotes.wordpress.com/).
 Data on terms served and gender is available at the website of the Parliament (2018). Data on MPs ethnicity was provided by British Future (2017). Closeness of race was provided by the Electoral commission. The remaining data on the share of leave votes in the Brexit Referendum, population density, percent population older than 45 years, the population declaring to be of white British ethnicity, and the region was retrieved from Norris (2017). All population data provided by Norris refers to the 2011 census.
 Even though not all differences reach statistical significance, one has to keep in mind that this is data for the full population. The figures themselves are thus informative, but the level of statistical significance reveals variation in the sample and thus deviation from the general trend.
 However, there is considerable variation in the voting behavior of male and female MPs (the confidence intervals overlap).
British Future (2017). 52 minority MPs to sit in ‘most diverse UK parliament ever’. Available at http://www.britishfuture.org/articles/52-minority-mps-to-sit-in-most-diverse-uk-parliament-ever/.
Commons (2019). Votes and proceedings for 15 Jan 2019. Available at https://publications.parliament.uk/pa/cm201719/cmvote/190115v01.pdf.
Electoral Commission (2018). Electoral data files and reports. Available at https://www.electoralcommission.org.uk.
Norris, Pippa (2017). British Parliamentary Constituency Database 2010-2017. General Election results file version 1_2. Available at https://www.pippanorris.com/data/.
Parliament (2018). MPs. Available at https://www.parliament.uk/mps-lords-and-offices/mps/.
Strøm, K. (2000). Delegation and accountability in parliamentary democracies. European Journal of Political Research 37(3), 261-289.
Strøm, K. & Bergman, T. (2011). Parliamentary democracies under siege? In K. Strøm & T. Bergman (eds). The Madisonian Turn: Political Parties and Parliamentary Democracy in Nordic Europe. Ann Arbor: University of Michigan Press, 3-34.