1 of 25

Playing by (or with) the rules? Bias and UK General Elections

Professor Charles Pattie

University of Sheffield

2 of 25

The 2024 UK General Election: �an unusual contest?

3 of 25

The 2024 UK General Election: �an unusual contest?

4 of 25

The 2024 UK General Election: �an unusual contest?

5 of 25

The 2024 UK General Election: �an unusual contest?

  •  

6 of 25

The 2024 UK General Election: �an unusual contest?

7 of 25

The 2024 UK General Election: �an unusual contest?

8 of 25

So what happened?

  • How did the electoral system produce such an odd result?
  • Not simply disproportionality (that’s baked in to FPTP)
  • Electoral bias – does the electoral system treat parties in the same way if the get the same vote… or doesn’t it?
  • Ralph Brookes and two-party bias:
    • If two parties get the same vote shares, they should get the same number of seats in an unbiased electoral system
    • But if one gets more seat than the other, the system is biased towards the more successful party… and the extra number of sets it gets provides a measure of that bias

9 of 25

Applying Brookes’ method in 2024

  • Rerun the 2024 election, assuming:
    1. There is a nationally uniform swing from Labour to Conservative of 5.1%, tying their national votes at 29.5%
    2. All other parties’ votes remain unchanged.
  • Some seats will change hands: at equal national votes, how many more/less MPs would Labour have won in 2024 compared to the Conservatives?

10 of 25

Applying Brookes’ method in 2024

99-seat advantage!

11 of 25

Applying Brookes’ method in 2024

  • How did that happen? Decomposing the bias into…
    • Efficiency effects
    • Constituency size effects
    • National effects
    • Turnout effects
    • Third party votes effects
    • Third party wins effects

​

​

12 of 25

Applying Brookes’ method in 2024

13 of 25

Applying Brookes’ method in 2024

14 of 25

Applying Brookes’ method in 2024

  • Labour was ‘helped’ in 2024 because:
    1. Its vote was more efficiently spread in 2024 than was the Conservatives;

15 of 25

Applying Brookes’ method in 2024

  • Labour was ‘helped’ in 2024 because:
    1. Turnout tended to be lower in Labour seats than in Conservative seats;

16 of 25

Applying Brookes’ method in 2024

  • Labour was ‘helped’ in 2024 because:
    1. ‘Third party wins’ (largely Lib Dem) came more at the expense of the Conservatives than Labour. But where 3rd parties lost, their greater support in Conservative than in Labour seats helped the Conservatives a bit)

17 of 25

How unusual was 2024: Con-Lab bias since 1950

Pro-Labour

18 of 25

How unusual was 2024: Con-Lab bias since 1950

Pro-Labour

19 of 25

How unusual was 2024: Con-Lab bias since 1950

Pro-Labour

20 of 25

How unusual was 2024: Con-Lab bias since 1950

Pro-Labour

21 of 25

How unusual was 2024?

  • In some ways comparable to 1997-2005, a previous ‘high-point’ of pro-Labour bias.
  • But in other ways very different:
    1. Labour is much more vulnerable now.

​

22 of 25

Median majority = 30.4%

Median majority = 17.4%

23 of 25

How unusual was 2024?

  • In some ways comparable to 1997-2005, a previous ‘high-point’ of pro-Labour bias.
  • But in other ways very different:
    1. Labour is much more vulnerable now.
    2. The Lib Dems were treated fairly by the electoral system: 12% of the vote, 11% of the seats
    3. A disaster for the Conservatives: their lowest general election vote share (23.7%) and number of MPs (121) ever – and things have become much worse since (opposition support usually grows!)
    4. A much more volatile electorate: 28% changed their vote between 19902 and 1997; 44% did so between 2019 and 2024!
    5. Reform UK as wild card, leading in polls a year after the election…

​

24 of 25

So where next?

  • Into the unknown?

​

​

​

​

​

(Source: Electoral Calculus poll of polls https://www.electoralcalculus.co.uk/polls.html)

25 of 25

So where next?

  • Into the unknown?