stat

Monday 29 July 2013

Number Of Constituencies Under AMS

I have been looking at the Additional Members System recently, and wondering what would happen if it were introduced to the House of Commons. Now, I'm not going to tackle "who wins what where" issue - that can be for other people to do - but wonder how many constituency MPs there would be.

The Parliamentaty Voting System & Constituencies Act 2011 envisaged 600 constituencies, with 596 of them within 5% of the electoral quota (determined by taking the total number on the electoral roll and dividing it by 596). There were to be 4 preserved constituencies - 2 of them in Scotland (the current Na H-Eileanan An Iar and Orkney & Shetland) and 2 in South East England (drawn from the current Isle of Wight), and these are ignored for the purpose of calculating each nation's/region's entitlement.

Then came the Electoral Registration & Administraton Act 2013 which postponed the boundary review for 5 years so the May 2015 general election will be on the current boundaries.

But what if we followed this method for a House of Commons using AMS. I will assume that there are 3 preserved constituencues (surely under larger AMS constituencies, there would be no justification for splitting the Isle of Wight) and add them in at the end.

As to the proportion of constituency and regional seats then there are 5 combinations that spring to mind. Three of them are those actually in use - the London model (the Greater London Assembly has 14 constituencies and 11 additional), the Welsh model (the National Assembly of Wales has 40 constituencies and 20 additional) and the Scottish model (the Scottish Parliament has 73 constituencies and 56 additional).

This then brings me to the Jenkins Commission report, which suggested a variation of AMS, with between 80% and 85% of MPs elected in constituencies - I will refer to these limits as "Jenkins min" and "Jenkins max".

If we follow this, then Jenkins max is the maximum number of constituency MPs/minimum number of regional MPs to ensure a reasonable measure of proportionality.

With that background, how would 600 MPs be allocated in various systems?:

System Constituencies Regional Southampton Hampshire
Jenkins max 510 90 1.84 14.57
Jenkins min 480 120 1.73 13.71
Welsh model 400 200 1.44 11.41
Scottish model 340 260 1.22 9.69
London model 336 264 1.21 9.57

The reference to Southampton there is the number of constituencies that the area covered hy Southampton City Council would be entitled to, and for Hampshire the number of constituencies that the area covered by Hampshire County Council, Portsmouth City Council and Southampton City Council would be entitled to.

Currently, Southampton has two constituencies - Southampton Itchen and Southampton Test - together with part of Romsey & Southampton North. From the table above, if we adopted one of the Jenkins limits, then we would probably be looking at Southampton Itchen and Southampton Test remaining, but one or both of them expanding a bit outside the city boundaries. Under the Welsh model, then we could have something similar, or one of them being in Southampton (except for about 5 wards) with the other about half-in half-out of Southampton.

Go to the Scottish or London models, then we should expect something along the lines of York at the May 1997, June 2001 and May 2005 general elections - which had York entirely inside the city, while Ryedale, Selby and Vale of York all crossed the city boundaries - so one constituency covering the bulk of Southampton while about 3 wards would lie outside in a constituency (or constituencies) predominantly in New Forest and/or Test Valley or in Eastleigh.

There is a reason I have included Hampshire. During the last boundary review, the Boundary Commission for England divided England into the standard regions, such as South East England and then split them into sub-regions, which could be a single county, such as Hampshire, or a combination of counties, such as East Sussex and Kent.

If we follow the principle that constituencies need to be within 5% of the quota, then we get the maximum and minimum sizes of constituencies:

System Minimum Quota Maximum
Jenkins max 85,590 90,095 94,600
Jenkins min 90,973 95,761 100,549
Welsh model 109,305 115,058 120,811
Scottish model 128,766 135,543 142,320
London model 130,312 137,171 144,030

Now, return to Hampshire. Under the Scottish or London model, assume we consider it as one electoral sub-region. In that case, it would need to have 9 or 10 constituencies. With an electorate of 1,312,952, an issue becomes apparent.

If we give Hampshire 9 constituencies, then the average size is 145,884 - above the maximum allowable size. But if we give Hampshire 10 constituencies, then the average size is 131,295 - which is only just above the minimum allowed (and less than the quota). So, unless we cross a county boundary, we are forced to give Hampshire 10. But keeping the number of constituencies to a pre-determined level would mean somewhere else in South East England having larger-than-average constituencies.

Is there a way around this? Under the old Parliamentary Constituencies Act 1986, the basic rules were:

  • Great Britain would have around 613 constituencies
  • Scotland was to have 71 or more constituencies (this was removed by the Scotland Act 1998)
  • Wales was to have 35 or more constituencies
  • Northern Ireland was to have between 16 and 18 constituencies

With regards to the Scottish Parliament, in January 2006, the Arbuthnott Commiitee published its report, which looked inter alia at the way AMS was used. And it came up with some rules which it suggested be used:

  • Every local authority in Scotland should contain at least one constituency
  • Every constituency should be contained wholly within one local authority
  • ....the number of constituencies in each local authority should be such that their average electorate is as close as practicable to one seventieth of the total electorate for Scotland (later on clarifying the "one seventieth" by adding The average of the electorate, excluding the reduced numbers caused by their special circumstances of Orkney, Shetland and the Western Isles
  • Scotland should be divided into parliamentary regions which so far as possible reflect natural communities
  • Each Scottish parliamentary region should consist of a local authority or an aggregation of two or more local authorities, provided that no region contains less than five constituencies

There are other rules, which need not concern us. The calculations they gave showed the island councils/constituencies as Orkney (electorate of 15,954), Shetland (17,000) and Western Isles (21,937) to be treated as special cases. The smallest mainland local authority was Clackmannan (35,647), which they suggested has 1 constituency, although they did say it might be possible to combine it with next door Stirling (64,578), for which they also suggested 1 constituency.

For the mainland the average constituency size was recommened to vary from 71,139 (for East Lothian with 1 constituency) down to Clackmannan - although if Clackmannan was combined with Stirling, then the smallest would be East Dunbartonshire with its 2 constituencies having an avergae sixe of 40,917. So, assuming that Clackmannan and Stirling are combined, and ignoring the island constituencies, this would give the smallest constituency being around 57.5% the size of the largest.

Scotland has unitary authorities, so the term "local authority" is used slightly differently to England.

It seems from this that using AMS gives the Boundary Commission the ability to allow large differences in constituency size, as the AMS will help even things out a bit.

The problem with the Parliamentary Voting System & Constituencies Act 2011 was that it gave little room to manoeuvre. The constituencies had to be between certain size limits and there was a set number of them in each region or nation.

In the earlier system, the Boundary Commission would divide the electorate by the quota in each area undetr consideration (as we did in the first table for Southampton on its own and for Hampshire) and then see which average size was nearer to the quota. The following table shows this:

System Hampshire (min seats) Average Above quota Hampshire (max seats) Average Below quota
Jenkins max 14 93,782 3,687 15 87,530 2,565
Jenkins min 13 100,96 5,235 14 93,782 1,979
Welsh model 11 119,359 4,301 12 109,413 5,645
Scottish model 9 145,884 10,341 10 131,295 4,248
London model 9 145,884 8,713 10 131,295 5,876

Basically, this looks at the average size of a Hampshire constituency and how far this would be from the quota. For each model, the one closest is in bold, and under the old rules, this is probably the option the Boundary Commission would have gone for.

How would we give the Boundary Commissions more flexibility without compromising the requirement to keep the House of Commons at 600 MPs? Well, the Jenkins Commission felt that the correct ratio should be somewhere between the two limits, and this gives the flexibility:

Region/Nation Entitled to Constituenices (Jenkins min) Regional (Jenkins min) Constituencies (Jenkins max) Regional (Jenkins max)
South East England 82 66* 16 70* 12
London 69 55 14 58 11
North West England 68 55 13 58 10
Eastern England 56 45 11 48 8
West Midlands 54 43 11 46 8
South West England 53 42 11 45 8
Scotland 52 42** 10 45** 7
Yorkshire & Humberside 50 40 10 43 7
East Midlands 44 35 9 37 7
Wales 30 24 6 25 5
North East England 26 21 5 22 4
Northern Ireland 16 12 4 13 3

[* Includes 1 preserved constituency]

[** Includes 2 preserved constituencies]

This could be a way this works - the Boundary Commissions have an electoral quota based on Jenkins min, and need to have at least that number of constituencies in a region/nation, but have the flexibility to go up to the number of constituencies under Jenkins max if necessary, with the number of regional MPs reduced accordingly.

No comments:

Post a Comment