How to produce a fair arrangement of constituencies

In boundary changes, boundary commission, redistricting on July 18, 2015 by dadge

Just before the general election, the Political and Constitutional Reform Committee of the House of Commons issued a report on the parliamentary boundary review system. Sensibly, they recommended that the 5% rule be replaced with a 10% rule, i.e. that seats can have electorates up to 10 per cent above or below the average. But otherwise the report is quite conservative in its conclusions.

My input was as follows:

This submission is divided into two parts: recommendations within the current framework; and recommendations for a new framework.


In order to work effectively, Britain’s First Past The Post System requires a fair arrangement of constituencies.


PART A: Recommendations within the current framework


  1. Some of the current problems with seat distribution are caused by the philosophy of “minimum change”. Minimum change leads to the perpetuation of arrangements of seats which contribute to the unfairness of the system. For example, smaller seats tend to stay smaller, marginals tend to stay marginals, and so on. The arrangement of constituencies is not for anyone’s convenience, least of all that of political parties, it’s for the effective and fair representation of voters and their communities, and hence each review should start with a blank canvas, as it were, rather than any attempt being made to keep changes to a minimum.
  2. In order to reduce the disruption this causes, the period between reviews should be as long as possible. Given that each review is allowed to make wholesale changes, and these will take time to get used to, I would suggest that the period be 15 years. Interim reviews can take place where constituencies grow or shrink very quickly.
  3. While the review areas at previous Reviews had often been too small, e.g. Northumberland, the Wirral, and the London boroughs, such that constituencies varied wildly in size, the policy at the last review changed to what I call “amorphous blobbism”: regional review areas with very little respect for the boundaries within them. All sorts of unnatural cross-border constituencies were recommended, many of which were removed on appeal. For reasons of tradition, belonging, and practicality it’s good to use counties and metropolitan boroughs as the basic review units. Just considering the review process, it’s rather unfair on a member of the public who wants to oppose a particular proposed constituency if they have to consider the knock-on effects across a huge region, most of which they won’t know.
  4. The size of review areas needs to be a compromise between working within known units and having areas big enough to give flexibility with regard to constituency electorate and formation. The normal review area should be the top-level local authority, i.e. county, metropolitan borough or unitary authority, but this is not possible with the new 5% tolerance in electorate, since this will only work if every review area has an electorate of 800,000+. Hence the tolerance should be put up to 10%. (At the last review, this would’ve allowed constituencies of between 68,977 and 84,305.) With these limits, review areas need an electorate of at least 414,000, which would once again allow review areas to be based on counties and the larger metropolitan boroughs.
  5. The Commission needs to be expressly told to protect communities from being split by the process whenever possible. During the last Review the English and Welsh Commissions were obsessively committed to completing their duties without splitting any wards. In several review areas there were very few constituency arrangements that were possible without splitting a ward, which meant they had effectively tied their own hands, as well as removing almost any element of choice in the consultation discussions and leaving some areas with constituencies that split established communities and forced unconnected areas together. The classic case was Cheshire, but the stubbornness also detrimentally affected the outcome of the Review in Birmingham, Leeds, Portsmouth, and several other areas.
  6. Wards are useful building blocks but they should be considered less, not more, sacrosanct than town or county boundaries. Despite the ludicrous claim by the Commission that there is no agreed way of splitting wards, polling districts with known electorates exist in every area and can be used in a limited way in order to create better constituencies.
  7. Working on the data from the last Review (with the current 5% tolerance),
    Constituency electorate: minimum 72,810, maximum 80,473. Once ward size reaches 8,000 redistricting is effectively impossible: 9×8,000=72,000, 10×8,000=80,000 and 8×9,000=72,000, 9×9,000=81,000 and 7×10,000=70,000, 8×10,000=80,000.

    There is a way for the Commission to allow ward-splitting without opening the floodgates: allow it only in districts where it is mathematically necessary, i.e. with an average ward size of a certain amount. I calculated this amount as 102.5% of the average constituency size, divided by 10, which at the last review was 7,856.

    Most, if not all, metropolitan boroughs would qualify (Rochdale’s average is 7,867, for example), giving the Commission the necessary flexibility in those places.

    A cap can also be put on how many splits are allowed in a particular review area, and I suggest this could be: number of seats divided by two, rounded down. In Rochdale’s case this would be 1, in Leeds 3, in Birmingham 4, in Cheshire-Wirral 5.

  8. Orphan wards. The Commission’s current policy of not splitting wards has led to the proliferation of orphan wards: a single ward from a particular authority is tacked on to a constituency otherwise formed entirely from wards in a different authority. This is unfair to electors since the MP will spend most of their time interacting with the “main” authority and will be seen as that authority’s MP. Therefore the policy should aim to reduce the number of such wards.
  9. Packing/cracking. If a town has an electorate of say, 120,000, the decision must be made whether to have an inner and an outer seat or a west and an east seat. In the jargon, to pack or to crack. This makes an important difference to the way the area is represented, and often makes a difference to which parties win seats. In these cases the Commission should listen particularly to the views of local people in general, rather than to the (biased) views of local parties. And there should be careful analysis to ensure that the national balance between packing and cracking is such that there is no national bias towards a particular party.
  10. Islands and sparsely populated areas. The vote of every elector should be of equal worth. Therefore it’s unacceptable to have places like the Isle of Wight where an elector’s vote is worth less, and places like the Western Isles where an elector’s vote is worth more.
  11. In the case of the Isle of Wight, part of the island should form a constituency with part of the New Forest.
  12. In the case of the Northern and Western Isles, I suggest that each sends a “proportional member” to parliament. This would mean that the Northern Isles would send an MP with 0.5 of a vote in parliament, and the Western Isles would send an MP with 0.33 of a vote. This is the best compromise, since it reduces the over-representation without attempting to combined these islands with the distant mainland.
  13. In the case of other sparsely populated areas (generally distant from Westminster), the rules have already correctly been changed to stop their over-representation. Funds should be available for extra staff for MPs in such areas in order that electors get the same level of service that is available to voters in more compact constituencies.
  14. The electoral register. Boundary reviews rely on correct electoral registers. Most (or all?) electoral registers have not been correct for many years. From my own experience of canvassing for Birmingham City Council I know how hard it is for local authorities to get everyone onto the register. If parliament is serious about accuracy in its parliamentary boundary reviewing, it needs to allocate more resources into getting everyone onto the register, especially as registration varies so widely between areas according to such factors as income and nationality. Among the poor practice that comes to mind from my researches is the extremely inaccurate register that Preston had a few years ago (I seem to remember that almost 20% of voters were missing), and the unavailability of any evidence for Leeds’s ward electorate data at the last Review. These issues are just the tip of the iceberg.
  15. Mapping. The Commission’s mapping was quite good at the last review, but pdf’s are unwieldy, and it was shameful that it was left up to third parties to come up with interactive mapping. If the Commission is serious in wanting members of the public to submit counter-proposals it needs to provide an online mapping tool to help them do this.
  16. The Commission should promote co-operation over confrontation. Why not invite proposals right at the outset and incorporate those ideas into its own initial proposals, rather than coming up with its own, often bizarre, ideas without consultation and then getting all precious about them when they’re criticised?
  17. The Review hearing system is still terribly biased in favour of political parties. They use their resources to produce comprehensive counter-proposals that appear to take up at least half of the process, even though they only provide about 10% of the counter-proposals by number. Although some local parties are quite good at representing local views, my main conclusion from sitting through hearings at the last and the previous Review is that, by and large, parties’ proposals are mainly directed at keeping their MPs (and gaining new ones).
  18. Make-up of Commission. At the last review the Commission was made up of two lawyers and a council manager, only one of whom appeared to be an expert in electoral law, and none of whom was a geographer or a statistician. This meant it was difficult to have faith that they knew what they were doing. The Commission should be increased in size to ensure a greater breadth of expertise.

PART B: Recommendations for a new framework


B1. A much simpler and easier-to-manage framework for the redistribution of parliamentary constituencies is a rolling review. This would operate as follows.


B2. Initial allocation and distribution. A constituency electorate would be decided. Let’s say 80,000. (This would give several fewer MPs than now, but it’s a good idea to use a round number in order to keep future data analysis as simple as possible.) Each review area receives the number of constituencies it’s entitled to. For example, if the electorate in Area P is 420,000, it would have 5 seats (420/80=5.25); and if the electorate in Area Q is 460,000, it would have 6 seats (460/80=5.75). In every area, the distribution of seats would be decided by an independent body. This could be a national body, as now, or a local body for each area.


B3. Subsequent reviews. There would be no national or regional boundary reviews. When the electorate of a review area increases or falls such that its entitlement increases or falls, then a review takes place in that review area. But not immediately. In order to avoid frequent reviews in areas which have an entitlement of approximately half a seat (For example, if the electorate in Area R is 441,000, it would have an entitlement to just over 5 and a half seats (441/80=5.51)) the entitlement would need to have increased/decreased for 3 consecutive years before the review takes place.


B4. Within a review area, if the electorate of any constituency goes beyond acceptable limits (the limits could either be +/- 10% of the average in the review area or +/- 10% of the national average, or whatever tolerance is agreed on) for 3 consecutive years, the boundary of that seat should be reviewed. This will of course require the review of at least one neighbouring seat, and possibly several others. Furthermore, if any seat in the review area remains over 10% smaller or larger than any other seat in the same review area over a period of 10 years, there should be a review of the seat (and its neighbours). This will avoid the long-term over- or under-representation of electors.


17 October 2014


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: