Some thoughts on the parliamentary boundary changes

The starting gun has been fired on the latest review of parliamentary constituency boundaries. In one sense, it’s not before time. The current set of boundaries came into force in 2010 (2005 in Scotland) and were based on electorate data from December 2000, and we have had more than two decades of population movement since then. The map is becoming increasingly mis-shapen, as trendy inner-city areas and shire counties have new apartments and housing estates built, while other areas get left behind with declining populations. In the March 2020 electoral register to be used in the review (taken just before the pandemic hit), the largest English constituency (the Isle of Wight, with 111,716 parliamentary electors) has more than twice the electorate of the smallest (Stoke-on-Trent Central, 54,551).

The historical context

Boundary changes used to be the preserve of Parliament itself, and this led to a lot of stasis. After the redistribution of 1885 led to single-member constituencies becoming the norm, there was only one further set of boundary changes in the next sixty years (in 1918). By the time of the Second World War, some constituencies had become grossly oversized. The most extreme example was the Romford division of what was then Essex, which had been filled with houses by the London County Council’s Becontree estate and similar developments; between the 1918 and 1935 elections, Romford’s electorate grew from 37,000 to almost 168,000 electors, far in excess of any other constituency. Examples like this are the reason that we have a boundary review process. An emergency change for the 1945 election spit Romford up into four new constituencies and also divided a number of other outsize seats into two, with a major overhaul following in 1950 at which the last of the two-seat borough constituencies disappeared.

A Speaker’s Conference in 1944 resulted in the periodic boundary review process which we have today, run by four independent boundary commissions. The original rules set out a fixed number of 12 constituencies for Northern Ireland (which was deliberately under-represented due to the existence of the Stormont Parliament), set out that the number of seats in Scotland and Wales should not fall below the 71 and 35 seats those countries respectively had after the 1918 redistribution, set out a presumption that “as far as practicable” constituencies should not cross county, county borough or (in London) metropolitan borough boundaries, and gave the Commissions power to depart from those rules if they needed to in order to draw reasonable constituencies.

With some tweaks over the years, in particular to the frequency between reviews, those rules lasted into the twenty-first century through five redistributions (in 1955, February 1974, 1983, 1997 and 2010). The allocation for Northern Ireland was changed to 16-18 seats in 1983 following the demise of the Stormont Parliament, and the floor of 71 seats for Scotland was removed following the establishment of the Scottish Parliament; the Scottish boundary commission immediately got to work, and a review for the 2005 general election cut the number of constituencies in Scotland from 72 seats to 59. Despite the advent of devolution for Wales, the floor there was not removed and the current 40 constituencies for Wales represents a major over-representation; the review in progress will cut that to 32. Of the ten seats with the smallest constituencies in the UK, seven are in Wales and the other three are in the sparsely-populated Scottish highlands and islands, where “special geographical considerations” could be justified.

The seat allocation formula tended to have a “ratchet” effect, because when counties’ seat allocations were rounded up to the nearest whole number that had the effect of increasing the number of seats in the quota for following reviews. From 625 seats in 1950, the size of the Commons peaked at 659 in 1997 before falling to the present 650 seats in 2010 (mostly due to the Scottish reduction in 2005).

The Commissions also had the power to do interim reviews between major redistributions. This was normally done to tidy up cases where local government boundaries had changed, but one interim review for the 1992 election had the effect of awarding a second seat to the fast-growing town of Milton Keynes.

What changed?

Since the formation of the 2010-15 Coalition government, boundary reviews have got more political. The Conservatives’ 2010 manifesto proposed reducing the size of the Commons and ensuring that all seats were very close to the electoral quota. Following some haggling with their coalition partners, the targets were set at 600 seats and plus or minus 5% (with a small number of exceptions, all relating to offshore islands), and the time between reviews was cut to 5 years from the previous 8-12.

The Boundary Commissions immediately got to work to create a new ward map to be ready for the scheduled May 2015 election. Their preferred unit was the electoral ward, as this was the only readily-available geographical unit which covered the whole country and bore some relation to what might be considered as a local community. Several problems immediately became apparent.

The relationship between wards and communities is rather more tenuous than the Commission might like to think. Local government wards are drawn up by a separate boundary commission using basically the same rules as those for drawing constituencies. It’s primarily a numbers game. If the numbers on the electoral register force the Local Government Boundary Commissions to split a community between wards, or put a number of disparate communities in the same ward, sometimes that’s what they have to recommend doing. The constituency boundary-drawers, drawn from a different set of staff, didn’t have the benefit of knowing which wards represented communities better than others.

Most wards in England are not single-member: many elect two councillors, and three-councillors are the norm in most urban areas. To take one example, let’s look at the Metropolitan Borough of Wirral, which has 66 councillors elected from 22 three-member wards. Since 1983 it has had four parliamentary constituencies (Birkenhead, Wallasey, Wirral South and Wirral West).

The Wirral

The December 2000 electoral register gave the Wirral 247,183 electors (an average of 11,236 per ward) which entitled it to just 3.53 seats. For the review implemented in 2010 the Boundary Commission originally felt that a continued allocation of four seats wasn’t justified by the numbers, but they were unwilling to draw a constituency crossing the Merseyside-Cheshire boundary. The result was a provisionally-proposed seat called Wallasey and Kirkdale, consisting of territory either side of the Mersey estuary connected only by the Wallasey tunnel. This proposal went down like a cup of cold sick and was thrown out by the public inquiry, and the Commission were eventually forced to apply the “special geographical considerations” rule to retain Wirral’s four seats. As there were 22 wards to divide between 4 constituencies, two seats had to get six wards each with the other two getting five each which created a bit of a disparity and two seats (Wirral South and Wirral West) with very low electorates. Given the choice between that and “Wallasey and Kirkdale”, fair enough.

Fast-forward to 2011. The Wirral’s entitlement had by now fallen below 3.5 seats, but because of the new tolerance rules it was not low enough to get three seats of its own. Given the furore “Wallasey and Kirkdale” had created, there would have to be a seat crossing the Wirral-Cheshire boundary in the Boundary Commission’s next review which was due to report in 2013. And that seat – along with every other seat in the country – would have to be within 5% of the quota.

Then we get to the problems around what to do with the Wirral itself. This was a borough whose wards had an average of around 11,000 electors, with (because they were drawn to be approximately equal in electorate) not much variation in that figure. The Boundary Commission were faced with the task of using building blocks of 11,000 electors to hit a 5% tolerance level which was less than 8,000 electors wide. Things were similar in Cheshire, Greater Manchester and Merseyside. The Commission were unwilling to split wards between constituencies. You don’t have to be a mathematician to realise that this is going to run into trouble.

This was what the Boundary Commission came up with for the Wirral and related parts of Cheshire in their original proposals:

Let’s take a look at what’s going on here. We have not one but two constituencies crossing the county boundary: one ward from Cheshire in constituency 27, two wards from Wirral in constituency 42. We have Bidston ward, an integral part of Birkenhead, in the Wallasey-based constituency 59 to make the numbers fit. We have Ellesmere Port divided between constituencies 18 and 42.

In fact, just look at that constituency 42. This is not the answer to life, the universe and everything which the Commission clearly thought it was. We have two suburban/industrial wards from the Wirral, most but not all of industrial Ellesmere Port, the rural towns of Frodsham, Helsby and Weaverham which have no connection to Ellesmere Port without going through constituency 18, and to the north the villages of Hale and Ditton which are separated from everything else in the seat by several miles of Mersey estuary, Manchester Ship Canal and saltmarsh. Seemingly at a loss for anything else to call this Frankenstein’s monster of a constituency, the Commission gave it the name “Mersey Banks”.

The 2011 Mersey Banks was by any metric the worst proposal in a car-crash map for North West England which gave little impression of respecting any of the Commission’s statutory criteria except the 5% tolerance. It brought the whole process into disrepute. Your columnist attended the first day of the public inquiry into these proposals, which deservedly threw the entire map out and substituted a counter-proposal of its own which was a heck of a lot letter.

That counter-proposal, however, was never used for real because the Liberal Democrats killed off the 2013 boundary review in protest at losing House of Lords reform. The next review, which reported in 2018, never saw the light of day either: the Government sat on the report without bringing it to Parliament for approval, and the 2019 snap election means that the Commissions are due to report again before the next scheduled general election in May 2024. (Which is slightly galling for your columnist, because the public inquiry accepted my counter-proposal for those boundaries.)

The new review

Following their election victory in December 2019, the Conservatives have passed some tweaks to the current boundary review which is due to report in 2023. The reduction to 600 seats is gone; the Commission is now able to take account of new ward boundaries which had been finalised by December 2020 but are yet to be used in an election; the 5% tolerance remains. Because of the tolerance requirement, we are going to see a number of seats crossing county boundaries and many more seats crossing London and metropolitan borough boundaries than previously. The “Devonwall” seat which has caused an awful lot of grief at the last two reviews is no longer necessary, but we are likely to see:
– a seat in the Morecambe Bay area containing parts of Lancashire and Cumbria
– at least one seat crossing the Wirral/Cheshire boundary
– a seat crossing the Staffordshire/West Midlands boundary, probably in the Dudley area
– a Warwickshire seat containing all of Staffordshire
– a seat crossing the boundary between Berkshire and one of its neighbouring counties (probably Hampshire)
– a seat crossing the boundary between Bedfordshire and Hertfordshire
– a seat crossing the boundary between Suffolk and either Essex or Norfolk.

The 5% tolerance is going to cause issues all over the place. In the metropolitan borough of Wigan, the current Makerfield (74,400) and Wigan (75,607) seats are within the required tolerance, but the Leigh constituency (77,416) is 354 voters over the upper limit. This can be fixed, but the least disruptive fix involves transferring the two wards covering Golborne and Lowton (19,091 electors) from the Leigh seat into the Makerfield seat while Hindley and Hindley Green (18,422 electors) go the other way; that’s a minimum of 37,513 electors moved between constituencies to fix one seat being 354 voters too large.

And no doubt this sort of thing will be repeated in future reviews. We will see more disruption caused by perfectly reasonable seats falling narrowly out of tolerance. The cross-county seats are unlikely to be popular and unlikely to last very long, as populations continue to move and entitlements change.

In the meantime, the Boundary Commissions are hard at work on their provisional proposals. The English Commission is hoping to be able to open consultation in the early summer, with public hearings scheduled for early 2022. This column awaits the initial report with interest and your columnist will certainly be making representations about his own area. Hopefully you will do so too.

Andrew Teale