Boundary Review: missing voters, missing seats?

The four Boundary Commissions have started to publish their initial proposals for new constituency boundaries. The new boundaries will be based on the electorate of 1 December 2015. Yet by June 2016 (at the time of the EU referendum), the number of registered voters had increased by 1.75 million. Would using the June electorate change how seats are allocated within the UK?

The 2018 Review of constituency boundaries – what and when?

The Boundary Commissions for England, Wales, Scotland and Northern Ireland are required by law to propose plans for 600 seats (down from the current 650) with roughly even-sized electorates, based on the electoral registers published on 1 December 2015. These seats must fall within 5% of the ‘electoral quota’, which is calculated by dividing the UK electorate, minus the electorate of four protected seats (Orkney and Shetland, Nah h-Eileanan an Iar, and two seats on the Isle of Wight), by the total number of seats, minus the island seats.

The initial proposals for Northern Ireland were published on 6 September, and the proposals for England and Wales on 13 September. The proposals for Scotland are expected around the 20th of October. After a process of consultation and revision, and subject to Parliamentary approval, the new boundaries will come into force at the 2020 General Election.

An increase in registered voters since December 2015

Some have argued against the use of the December 2015 registers, which were the first under the new Individual Electoral Registration system where everyone is responsible for their own registration – there were about 600,000 fewer voters on the registers in December 2015 than in December 2014. As people registered themselves to participate in the EU referendum, the registers had grown by about 1.75 million voters in June 2016. For the purpose of the boundary review, these voters simply do not exist. But are they distributed unevenly, so that some regions have more ‘uncounted voters’ than others?

The table below shows how electorates in the regions used for the EU referendum changed between 1 December 2015 and 23 June 2016. The largest percentage increase was in London, the smallest in Northern Ireland.
Change in the registered electorate by electoral region

A small effect on the number of seats per electoral region

So if the Boundary Commissions could use the electoral registers from June 2016 instead of from December 2015, would the increased number of registered voters have an impact on how seats are distributed within the UK?

The table below shows an estimate of how seats would be allocated within the UK if the Boundary Commissions used the June 2016 electorates instead of the December 2015 electorates.

2016 Boundary Review: estimated allocation of seats

The table shows that if the Boundary Commissions used the June 2016 electorates, three seats would be transferred between regions. London would gain two and the South West one, while Scotland would lose two and Northern Ireland one.

It would now be difficult to change the legislation that requires the Boundary Commissions to use the December 2015 electorates. The 2018 Review is well underway and aborting it (again) would come at a significant cost. However, the Boundary Commission for England has stated in its Guide to the 2018 Review that it will take changes in electorates into consideration (para 40):

“the BCE does not take the view that it is obliged to shut its eyes entirely to growth (or decline) that has occurred since the review date, which it may be aware of from the annual updates of electorate figures it receives, or that it is satisfied is likely to occur. Such a factor may be taken into account in choosing between two or more competing options for the same area that satisfy the statutory rules.”

Technical note: the estimate produced above does not use the Sainte-Laguë formula to distribute seats between countries. 

Picture credit: Straight scale, compasses and map (2) by dewframe * alt-n-anela; Creative Commons Attribution 2.0 Generic (CC by 2.0)