I am very grateful to Bill Speck, Michael Brock, Vernon Bogdanor, David Butler, Geoffrey Marshall, Martin Linton, and Roger Mortimore for their comments on earlier versions.
1. Introduction
Scotland has been represented in the Parliament of the United Kingdom since the Union of Parliaments in 1707. Only between 1885 and 1922 has Scotland's level of representation in the House of Commons lain at the average level for the United Kingdom as a whole. Wales has been represented since its incorporation with England in the sixteenth century. It, too, has swung between under- and over-representation. The purpose of this note is to investigate the under- and over-representation of Scotland and Wales, to inquire why it has arisen, and its implications for current policy on apportionment of seats, procedures for boundary commissions, and devolution. In particular, we evaluate the claim that Scotland was promised overrepresentation in the Act of Union of 1707.
2. The level of representation since 1707
The number of seats in the Commons assigned to Scotland was fixed at 45 in 1707, and remained unchanged until the 1832 Reform Acts, when it was raised to 53. It was raised modestly in the 19th century to 72 in 1885, a peak of 74 in 1918, and either 71 or 72 in all the boundary changes since 1947. These numbers include the Scottish Universities seats where applicable. Change in the total size of the House of Commons has also been rare. It was fixed at 558 in 1707, and raised to 658 with the addition of 100 Irish members in the 1800 Act of Union. An increase to 670 in 1885 was specifically in order to provide an extra 12 members for Scotland without cutting the numbers in any other part of the UK. The size of the House fell to 615 in 1921 with the departure of Ireland (except Northern Ireland) from the UK. In the successive boundary revisions since 1944, it has crept up gradually because of the upwards-only ratchet effect of Rules 4, 5, 7, and 8 of the Rules for Redistribution of Seats (Parliamentary Constituencies Act 1986, Schedule 2). It is expected that the Fourth Periodical Reports of the four Boundary Commissions (one each for England, Scotland, Wales, and Northern Ireland) will recommend an aggregate of 659 seats for the next apportionment, to be in force from 1996.
Table 1 shows Scotland's share of Commons seats, of population, and of the UK-wide electorate over the years. The final four columns show, on various measures, the degree of over- and under-representation of Scotland compared to the average for the UK.
Wales received no special treatment until 1944. Indeed, as most statistical series including the Censuses usually aggregate Wales with England (and until the 19th century count Monmouthshire - now Gwent - with England), it is quite difficult to determine basic population and electoral numbers. Therefore our Welsh series, (Table 2) begins in 1832.
Some explanation of these tables and charts is needed. The fourth column of each table gives Scotland's (Wales's) proportion of the seats in the House of Commons. Scottish and Welsh University seats have been attributed to their respective countries. Each change in either Scottish/Welsh representation or the overall size of the House of Commons from 1832 onwards has been dated from the General Election at which it first took effect. Except at the dates shown, Scottish and Welsh representation never changed. The overall size of the House occasionally varied because of the disenfranchisement of boroughs which had been found to be corrupt, but never by more than 6 in any inter-election period. The ratio of Celtic to UK-wide electorates and population of course varies continuously. The seventh column of Tables 1 and 2 is Scotland's (Wales's) share of the UK population. Population figures are taken from the nearest Census in each case. The tenth column of Tables 1 and 2 gives Scotland's (Wales's) electorate as a proportion of the UK electorate. This has not been done for 1707 or 1800, when it would be meaningless. The Scottish electorate given for 1800 is sufficient to show that the franchise in Scotland was much more restricted than even the pre-1832 franchise in England and Wales. The population figure for 1996 is the most recent projection from the Office of Population Censuses and Surveys (OPCS); the electorate figure is extrapolated from the trend of the 1992 and 1994 figures. The final two columns of the tables show the ratio of Scotland's (Wales's) seat share to its population and electorate shares. A figure of 1 in these columns stands for proportionate representation. (No allowance has been made for any variation in the proportion of plural voters in the four parts of the UK , but this is believed to be a trivial source of error). These final two columns are also shown in graphical form in Figs 1 and 2. On both measures, Scotland was under-represented until the 1867 apportionment inclusive, proportionately represented in 1885 and 1918, and over-represented since then. The story of Wales in the 19th century is more complex. Judging by population share, Wales was over-represented until the 1885 apportionment, under-represented in 1918, and thereafter (like Scotland) over-represented on a gradually upward trend until the present day. Judging by electorate share, Wales was under-represented until 1885. In the 19th century, electorates and populations did not vary closely together as they now do. A relatively high proportion of the Welsh population had the vote in the 19th century and a relatively low proportion of the Scottish population did, hence the divergence between the two series.
These figures already show that any claim that Scotland was guaranteed over-representation in 1707 must rest on shaky evidence. We next study each occasion on which the seat ratios were changed to consider what arguments, if any, about the proper ratio were used.
3. The ratio of Scottish and Welsh seats in the House of Commons between 1707 and 1944.
3.1 1707
The Act of Union entrenches, or tries to entrench, some features of Scottish government `to be held and observed in all time coming as a Fundamental and Essential Condition of any Treaty or Union to be concluded betwixt the two Kingdoms without any Alteration thereof or Derogation thereto in any sort for ever' (6 Anne c.11, Art. XXV). This entrenchment is primarily designed to protect the Presbyterian government and doctrine of the Church of Scotland. The universities of St Andrews, Glasgow, Edinburgh, and Aberdeen 'shall continue within this Kingdom for ever' as part of this protection . Second in the Scots negotiators' priorities was the Scottish judicial system. Representation in the combined parliament was of lesser concern. Article XXII of the Act of Union provides
That by virtue of this Treaty of the Peers of Scotland at the time of the Union Sixteen shall be the number to sit and vote in the House of Lords and Forty five the number of the Representatives of Scotland in the House of Commons of the Parliament of Great Britain.
Article XXV, section VI, recites and incorporates the Act of the Scottish Parliament providing for the method of election of the representative peers and the MPs. Thirty MPs were to represent counties, with six small counties grouped in pairs, only one of each pair to be represented at any one time. Edinburgh was to be a single-member constituency, and the other 14 Scottish MPs were to represent groups of burghs. Each burgh was to appoint a commissioner to a joint meeting to determine their representation, with the chairmanship, and casting vote, being held by one burgh in each group in rotation. Nothing in these arrangements suggests either that this was a matter ranking with the standing of the Church in its need for constitutional protection, or that the number 45 was chosen for any reason of fairness or proportionality to population. As Table 1 and Fig. 1 show, it was only some half of the number that could have been claimed on an equal population basis. The idea that Parliament existed to represent people was scarcely developed in the 18th century; rather, it existed to represent interests. The best way to measure it, according to the Scottish MP William Seton, was by proportion of tax revenue:
there must be members from both nations and their numbers must be adjusted by some rule, which can be no other than that fundamental in the union of all societies whereby suffrages are computed, to wit the proportion each society contributes for the support of the whole.
Scotland's tax base was less than 45/558ths of the UK's, although her population was much more. The English had originally offered 38 MPs and the Scots had asked for 50. In his Reflections on the Revolution in France (1790), Burke asks:
When did you hear in Great Britain of any province suffering from the inequality of representation?...The very inequality of representation, which is so foolishly complained of [in France], is perhaps the very thing which prevents us from thinking or acting as members for districts. Cornwall elects as many members as Scotland [it elected 44 in the 18th century]. But is Cornwall better taken care of than Scotland?
The idea of equal representation, of which Burke was so scornful in revolutionary France, had no articulate defender in 18th-century Britain.
3.2. 1800
By 1800 Scotland was less under-represented, more because of change in the denominator than in the numerator. The population of Scotland had grown since 1707, but that of England and Wales had grown faster. Therefore 45 seats in a 558-member Commons was closer to being a proportionate ratio. The Act of Union with Ireland was formally similar with the 1707 Act. It likewise brought in representative peers and MPs. Ireland was given 100 MPs and the size of the Commons was raised to 658 with no change in the numbers for each of Scotland, Wales, or England. In 1801 Ireland contributed 32.8% of the population of the new United Kingdom but was given only 15.2% of the seats in the Commons. This necessarily further increased Scotland's proportion of seats. No discussion of these matters in the debates on the 1800 Act has been found. Continued protection for the Scottish Church was inserted into the Act of (Irish) Union, which is otherwise silent on any implications for Scotland.
3.3. 1832
The issue of fair representation between the parts of the United Kingdom was raised in the run-up to the three parliamentary Reform Acts of 1832. However, it was raised on behalf of the (over-represented) English, not by the Irish or the Scots. In 1831, the government proposed to cut the size of the House from 658 to 596. In response an ultra-tory, Isaac Gascoyne, gave notice of a proposal that each part of the kingdom should retain its proportionate share of representation as laid down in the Acts of Union. The government first moderated, then eliminated, its proposed cut in House size. Nevertheless, an opposition amendment, based on Gascoyne's proposal and forbidding any reduction in the number of English and Welsh MPs, was carried in April 1831. This led to the fall of the government, and to the return of an even larger majority in favour of reform than before; Gascoyne himself lost his seat. His move was only formally concerned with proportionate representation. Substantially, it was an attempt to resist any change, including change in the constitutional settlement reached in the Acts of Union. However, the signals Gascoyne intended to send may not have been the signals others thought they heard, in an atmosphere of noise and confusion where politicians on both sides thought that Britain might be on the brink of a revolution.
Table 3 shows the representation of each part of the UK before and after the 1832 Acts.
Scotland's increase to 53 seats, together with a further slight decline of Scotland's relative population, raised Scotland's ratio of representation to 0.85. Based on electorate rather than population, it was raised to 0.94, reflecting the highly restricted Scottish electorate (Table 1). Wales fared quite well. Its 27 members of 1831 (14 county and 13 borough) rose to 31 in 1832, with populous counties and the industrial towns of Merthyr and Swansea gaining a member each, while tiny boroughs such as Montgomery and Beaumaris retained their members.
3.4. 1867
F. B. Smith's standard commentary on the making of the Reform Act of 1867 endorses contemporaries' view that the details were a 'Serbonian bog' - a phrase Disraeli borrowed from Milton - into which no sane man would sink. Scottish seats were raised to 60 and Welsh to 33, but we have found no discussion of the principles on which this might be based. As in 1832, the Scottish reforms were enacted separately to those for England and Wales. The Representation of the People (Scotland) Act 1868 unexpectedly annihilated seven English borough seats to create the seven extra Scottish seats; this coup was staged at dinner-time against (the minority) government advice but was not based on any evident principle.
3.5. 1884--5
The passage of the Third Reform Bill saw the first sustained discussion of the proportionate representation of each part of the UK. Ireland now had 103 seats (15.4% of the 670 agreed in 1884), and its share of the UK population had fallen drastically to 14.73%. Since 1801 Ireland had suffered famine, death and emigration while the rest of the UK had grown rapidly. Ireland was now slightly over-represented, but Gladstone was determined to preserve Irish numbers in order to placate the Irish leader Parnell and his party. This led Scottish MPs to complain 'When generosity to one part of the country inflicted, or threatened to inflict, injustice on another part, its Representatives could not hold their peace'. Gladstone offered them 12 extra seats, which they regarded as an 'irreducible minimum'. Gladstone wrote to his reluctant follower Hartington that reduction in Irish numbers would 'entail reduction for Wales: and this would I suppose entail County amalgamation: and this would run into Scotland'. For the first time, the representation package was presented in a unified bill for the whole UK. Neither Gladstone nor any other party leader was yet committed to the idea of equal-population electoral districts within each component part of the UK, but the politics of the bill entailed an effort to get proportionality among them. Tables 1 and 2 shows that Scotland and Wales were proportionately represented for the first time.
3.6 1918
Redistribution was the least controversial of the matters settled by the Representation of the People Act 1918. As is well known, female suffrage had been bitterly divisive in the years up to 1914. The extension of the male suffrage was also controversial. And the question of proportional representation, already aired at length in 1884--5, recurred. These four issues were remitted to a Speaker's Conference set up, with all-party approval, in 1916. The device of a Speaker's Conference was thought of, apparently on the spot, and was sprung by Walter Long, the (Conservative) Chairman of the Local Government Board, on his unsuspecting colleagues. The Speaker selected the members, after consulting the party whips, during a shooting party at Nuneham Courtenay, Oxon. They comprised 13 Unionists, 12 Liberals, 4 Irish Nationalists, and 3 Labour members. Of the 27 MPs among them, 4 represented Irish seats, 3 Scottish seats (one of them a university member), and none Welsh seats.
The Speaker's Conference met 26 times between October 1916 and January 1917. No minutes were kept, and only brief conclusions reported. The report was unanimous; it is clear from various sources that logrolling went on among the members to achieve unanimity. Its recommendations on the distribution of seats were 'that each vote recorded shall, as far as possible, command an equal share of representation in the House of Commons', whose overall size should stay substantially unchanged. The target population for an average seat should be 70,000. The resulting instructions to the boundary commissioners for England & Wales, and for Scotland, gave the target population of 70,000 in identical wording. However, the instructions for Ireland forbade changing (that is, reducing) the total number of seats in Ireland. Thus, as in 1885, roughly proportionate representation was secured for England, Scotland, and Wales. The principle of equal representation both within and among the countries of Great Britain was accepted for the first time.
3.7 1922
By the time that most of Ireland left the UK, it was substantially over-represented. Its population had continued to decline both relatively and absolutely, and its MPs elected in 1918 (most of whom did not take their seats at Westminster, but formed themselves into the first Dail Eireann) overrepresented the Irish population by about 50%. The withdrawal of southern Ireland therefore necessarily increased the proportionate representation of England, Wales, and Scotland. Northern Ireland remained in the UK Parliament, but with a deliberately less than proportionate delegation, reflecting the fact that a devolved Northern Ireland Parliament existed under the Government of Ireland Act 1920. Northern Ireland continued to have 12 territorial seats at each Westminster election until 1974 inclusive, since when it has had its proportionate share of MPs. As Northern Ireland is small in relation to the other components of the UK, the effect of this on their representation is correspondingly small.
We have found no discussion of the impact of Irish withdrawal on the other components of the UK. Scottish representation moved into surplus and Welsh representation back to proportionality as an accidental consequence of the withdrawal of Ireland.
4. The 1944 Speaker's Conference and its legacy
Up to 1944, none of the union or redistribution acts had contained any mechanism for periodic review of constituency boundaries. As a consequence, all redistributions, including the 1918 one which started with equal-population constituencies within each component part of Great Britain, aged quickly and constituency sizes became very unequal. In 1942 a departmental committee under the Registrar-General, Sir Sylvanus Vivian, reported in favour of permanent boundary commissions. It suggested some of the features which have become part of the system, such as a commitment to equal electoral districts, to a 'substantially' unchanged House size, and to Boundary Commissions chaired by the Speaker. It expressed the hope that preliminary hearings with the 'Chief or National Officers' of each party before each boundary review would eliminate 'representations arising out of Party interests' from local hearings. In 1944 a Speaker's Conference was again established. In the debate on its establishment, Scottish and Welsh MPs of all parties argued for over-representation. Three arguments for over-representation of Scotland were made. Arthur Woodburn (Lab., Clackmannan) argued that the distinct 'contribution to culture, education, and different trends of thought that can be made by different nationalities' should be recognised, if need be at the expense of the 'population [which had] developed all around the London area'. Major E. G. R. Lloyd (Cons., Renfrew East) argued that the Scottish share of seats should not be decreased because the decline in Scotland's relative population was `not due to the fault of the Scottish people', but rather to industrial decline which was not the Scots' fault. The redoubtable Secretary of State for Scotland, Tom Johnston (Lab., West Stirling), argued:
At the time of the Union of the Parliaments - it is sometimes forgotten that we are here by Treaty rights - we had reserved to us a proportion of the Members of the House. In 1707 we had 45 Members. That was when we had a population of just over a million. It is now well over 4,500,000 and we have 74 Members.
Johnston went on to point out that, if a standard electorate was imposed throughout the UK, 17 of the Scottish seats would be too small and would have to be amalgamated or expanded. Johnston had a considerable reputation as a historian as well as a politician. Thus he may have given birth to the myth that the Union guaranteed Scottish over-representation. He pointed out that Scotland's population had quadrupled and its representation had less than doubled. He did not add that the population of the rest of the UK had gone up by more and its seat numbers had gone up by less. Thus he gave (intentionally or not) the impression that Scotland was less favourably represented in 1944 than in 1707. If he did not create the myth, he certainly gave it new life. Immediately after him, D. R. Grenfell (Lab., Gower), made an 'us-too' claim for Wales.
The 1944 Speaker's Conference was appointed and run on very similar lines to its predecessor of 1916--17. It once again had 32 members besides the Speaker, who stated that its membership was 'roughly in proportion to Party strength in the House of Commons and also intended to secure, as far as possible, representation of various shades of opinion, different types of constituency, and all parts of the country'. It comprised 17 Conservative representatives, 9 Labour, 2 Liberal, one National Liberal, one ILP, and two independents. Unlike its predecessor, it overrepresented Scotland (5 members) and Wales (3 members). Each party sent a list of candidates, from whom the Speaker made a selection. The Conservatives, Labour, and Liberal Parties each named one Welsh member. The Conservatives named three Scots, and the Labour Party two. The Speaker selected four of these, and the ILP member he selected was James Maxton, a Scot.
Like its predecessor, it published only its conclusions. The minutes and evidence were withheld from public consultation for 50 years, and have only just been released. In 1948, when the Labour Government abolished university seats and the business vote, some members vigorously complained that this was to undo the logrolls of 1944, in which each group got something and sacrificed something in order to get a unanimous report. However, the special treatment of Scotland and Wales was not part of any logrolling. The issue was addressed at the 9th meeting of the conference, on 19 April 1944. A briefing note, written by its secretaries 'at the direction of the Speaker', first pointed out that Northern Ireland had less than its proportionate share of MPs, but that its 12 territorial seats had been specified in a schedule to the Government of Ireland Act 1920. Turning to Scotland and Wales, the document notes that on 1939 figures, the average English constituency had 54,775 electors, compared to 44,642 in Scotland and 47,220 in Wales. However, 'It is clearly out of the question to instruct the Boundary Commissioners to apply only a strict mathematical test' in Scotland and Wales. Apart from the problem of sparse population in the Highlands, the document does not explain why the Speaker thought it was out of the question.
The minute of the committee discussion is fuller.
[I]t was pointed out that a strict application of the quota for the whole of Great Britain would result in a considerable decrease in the existing number of Scottish and Welsh seats, but that in practice, in view of the proposal that the Boundary Commissioners should be permitted to pay special consideration to geographical considerations [sic], it was ... unlikely that there would be any substantial reduction. It was strongly urged that ... it would be very desirable, on political grounds, to state from the outset quite clearly that the number of Scottish and Welsh seats should not be diminished. The absence of any such assurance might give rise to a good deal of political feeling and would lend support to the separatist movements in both countries.
Accordingly, the conference resolved not to cut the number of seats in either Scotland or Wales and to establish a separate Boundary Commission in Wales. The latter move went beyond the recommendations of the 1942 Vivian Committee.
At the next meeting, somebody pointed out the logical inconsistency of the conference's position: that its decisions to hold the total size of the House constant while not cutting the numbers of seats in Scotland, Wales, and Northern Ireland 'might be represented as meaning that only England could suffer any reduction in representation. It was pointed out, however, that in fact England's position would be adequately covered by the formula for calculating the quota by dividing the electorate by the number of seats, and the Conference agreed that it was unnecessary to amend the Rules to meet this possible line of argument'. Politicians are sometimes accused of wishing everybody to be richer than average. Here they were content to agree that on average every constituency in England should be larger than average.
The 1944 recommendations have provided a template for all subsequent legislation, and created some problems with which the Boundary Commissions have had to deal ever since. The number of seats in Great Britain should remain 'substantially as at present', i.e., 591 excluding the university seats (recommendation 6). There should be no reduction in seat numbers for Scotland, or for Wales and Monmouthshire (7). There should be a Great Britain-wide quota, or target electorate, for each seat, calculated as (GB electorate/GB non-university seats); the maximum deviation of any seat from this target should be 25% (9-10). The City of London should retain two seats (13); Northern Ireland should continue to have 12 territorial seats (14). Boundary Commissions might 'depart from the strict application of these rules' if necessitated by 'special geographical considerations, including the area, shape, and accessibility of a constituency' (15). There was to be one permanent Boundary Commission for each of England, Scotland, Wales, and Northern Ireland (17), which should hear representations from the 'Chief or National Officers of the Party organisations' before commencing each major review (20).
The Redistribution Act 1944 (7 & 8 Geo. VI c. 41) implemented these rules. The guaranteed seat minima for Scotland and Wales were defended but not attacked in the Second Reading and Committee debates. An ad hoc redistribution was made for the 1945 General Election. During 1946 and 1947 the Labour Government announced that the 25% rule was too restrictive and was leading the Commissioners to break up historic communities. This conservative argument was accepted by the Conservatives; an Act of 1947 removed the explicit 25% rule, and placed equal constituency size below respect for local boundaries in the Commissions' rules.
The subsequent history of the interaction of the rules, which are now consolidated into the Parliamentary Constituencies Act 1986, has been well traced elsewhere and is not repeated in full here. . The most important change for our purposes has been the replacement of an electoral quota (i.e., target mean electorate for each seat) for Great Britain, as enacted in 1944, by a separate electoral quota for each component part. This change was made in the House of Commons (Redistribution of Seats) Act 1958. It relieved the Boundary Commissions of a mathematically impossible task, but further entrenched the over-representation of Scotland and Wales.
Rule 1 of the 1944 Rules sets a target number of seats for Britain (originally 591, raised to 613 in 1948). It sets minima but not maxima for Scotland and Wales. It is completely silent as to England. If interpreted literally, it would require the English Commission to delay starting work until the numbers of Scottish and Welsh seats were known in order to give the English Commission its target number of seats. This would be impractical. But the English Commission must make some assumption about its target number of seats and their mean electorates. In its first two reports (1947 and 1954) it took its target number of seats as that set by Rule 1 minus (71 + 35). The average seat size for England, as for Scotland and Wales, was supposed to be determined by the GB-wide electoral quota laid down in the 1944 Act. But to average the English constituencies around that number would have created more constituencies than the number the Commission started with. The Commission squared this circle by setting an English electoral quota as the target number of seats divided by the English electorate. In Harper v. Secretary of State for the Home Department (1955 Ch. 238), the Court of Appeal rejected a Labour complaint against this procedure. The plaintiff had obtained an injunction from a lower court on a Friday against the redistribution proceeding to an Order in Council, on the grounds that the English Commission had not followed the rules. On the Monday, the Court of Appeal overturned this and ruled that the Commission had proceeded properly. However, the English Commission did fail to carry out the rules. If the matter had been considered in less of a hurry and/or by people who understood arithmetic, it could have been pointed out that the rules were so nakedly self-contradictory that the Commission was bound to fail to carry them out.
The 1958 Act made the contradiction less obvious but did not remove it. It gave each of the four Boundary Commissions an electoral quota, comprising the number of seats in that part of the UK at the start of the review divided into the aggregate electorate for that part of the UK. As the rules now stand, the instruction to restrain the size of the House of Commons (now rule 1) contradicts the effects of the seat minima for Scotland and Wales (rule 1), the rounding-up effects of rule 5, the saving for remote areas in rule 6, and the instruction to take the number of seats at the end of the previous review as the starting-point for the next (rule 8). These rules, taken together, must tend to increase the size of the Commons at each review, for both mathematical and behavioural reasons. Because there are separate Boundary Commissions for each component part of the UK, the seat minima for Scotland and Wales cannot themselves be reviewed under the present law. The compromises of 1944 to 1958 have created a defective system with mutually contradictory rules.
5. Historical summary and policy implications
We have seen that the principle of proportionate representation for each part of the UK was not raised until 1831, and then probably for strategic rather than sincere reasons. No attempt was made to meet it until 1885. The redistributions of 1885 and 1918 represented England, Scotland, and Wales on the same proportionate basis. The system created in 1944, and still with us today, intended to achieve the same end by its recommendation of a Great Britain-wide quota. However, this was inconsistent with its recommendations of an overall ceiling size for the Commons and guaranteed seat minima for Scotland and Wales. When these principles came into practical conflict, it was the Britain-wide electoral quota which was sacrificed. As the 1944 system was avowedly the outcome of compromise among conflicting political interests, it is not surprising that it contains mutually contradictory rules.
Four arguments in favour of over-representing Scotland and/or Wales are currently used, and were used in 1944. In this summary they are identified by reference to their proponents in 1944.
1. Scotland was guaranteed over-representation by the Act of Union (Tom Johnston). It was guaranteed, and given, representation; it was neither guaranteed nor given over-representation.
2. Scotland and Wales are separate communities. Their over-representation recognises their separateness (Arthur Woodburn, D. R. Grenfell). This argument is open to the challenge that England is also a separate community, or a series of communities, which also deserve over-representation. The argument then becomes self-contradictory.
3. Scotland's population loss is the fault of UK government policy. To deprive Scotland of seats would be to punish it twice (Major E. G. R. Lloyd). This has been a powerful emotional argument over the years, and was also used in relation to Ireland in the nineteenth century. Like argument 2, it tends to fall into self-contradiction if generalised.
4. Scotland and Wales contain disproportionately many scattered rural areas, which must be given smaller seats (the 1944 Speaker's Conference). This is true, and recognised by what is now rule 6 of the Rules for the Redistribution of Seats. However, this rule should entitle Scotland and Wales to a number of extra seats which is a function only of population movement into and out of their sparsely populated areas. Probably, the tendency is towards their further relative depopulation. But this could not justify increases in the number of 'rule 6' seats to be awarded to Scotland and Wales, and hence cannot be the reason for the trend of increasing over-representation shown by Figs 1 and 2. Lord Davidson, the deputy chairman of the Scottish Commission, told the Home Affairs Select Committee enquiry in 1986-7 that the Scottish Commission regarded rule 6 as subordinate to rule 1: in other words, that extra seats for the thinly populated parts of Scotland should be at the expense of the rest of Scotland, not taken as an addition to Scotland's statutory minimum. The Scottish Boundary Commission has again taken this line in its Fourth Periodical Report (Cm 2726/1995), which holds the number of Scottish seats to 72. The Welsh Commission appears to take a different line, and raises the number of Welsh seats to 40 in its Fourth Review recommendations.
Thus the over-representation of Scotland and Wales arises not from considerations of principle, but from the bargained compromises of 1944, which have been frozen into the legislation governing the allocation of seats. Because the legislation provides for no means for the seat shares of each component part of the UK to be reviewed, the problems identified in this paper will remain.
It is not for the historian or the political scientist to say what ought to be done. But some policy implications arise from these facts, whether or not Scotland and/or Wales are granted any measure of devolution in the future. They appear to include:
1. It is not a good idea to give a public body, such as the Boundary Commissions, mutually contradictory instructions. From a contradiction, anything follows; therefore nobody can know in advance what a Commission may recommend; and there is no basis after the event for saying that one apportionment is better or worse than another. It might be thought a good idea to rewrite the rules to remove their internal contradictions.
2. In practice, the rule which has been sacrificed has often turned out to be rule 5 - the equal electorate or 'one vote, one value' rule. Is that what Parliament intended? Ought it to be sacrificed to other principles, such as restraining the overall size of the House of Commons, not violating county boundaries, or preserving the relative representation of Scotland and/or Wales? If so, why? Where should the proper trade-off lie? Do the arguments for equal representation contained in the US Supreme Court's judgements in Baker v. Carr (369 US 186, 1962), Wesberry v. Sanders (376 US 1, 1964), and White v. Weiser (412 US 783, 1973) have any resonance, or indeed relevance, for the UK? In the Court of Appeal in 1983, Lord Donaldson summarily dismissed American cases as irrelevant to the UK.
3. Should politicians be involved in seat apportionment, or should it be done wholly by disinterested expert bodies? The courts have up to now been singularly reluctant to become involved in apportionment, arguing that that is a task for Parliament. Arguably, however, Parliament neither does nor can regulate itself well. The nineteenth-century debates, in which MPs directly bargained over their own seats, are not encouraging. The device of the Speaker's Conference, and the convention of consulting political parties on apportionment, are designed to strike a balance between an over-partisan apportionment, and excluding elected politicians altogether from something that is vital to their futures. However, the 1944 Speaker's Conference wrote mutually contradictory rules backed by bad reasoning.
Parliament may make a partisan decision, as it did in 1969 when the majority party voted to delay a review which would have lost it seats.. (For a candid account of the gerrymandering by the incumbent Cabinet, see Richard Crossman's The Diaries of a Cabinet Minister, III: Secretary of State for Social Services 1968--70 (London: Hamish Hamilton and Jonathan Cape, 1977), entries for 9.6.69, 12.6.69, 9.7.69, 10.7.69, 24. 7. 69, and 23.10.69). The case for self-regulation is that a truly disinterested expert electoral commission is unobtainable, and that it is better for political influence to be controlled through the usual channels than uncontrolled and subterranean. However, Parliamentary self-regulation is now being queried in other fields, notably by the appointment of the Nolan Commission. The usual channels are rather silted. By contrast, the Australian Electoral Commission seems to manage to be genuinely non-partisan, and to complete its boundary reviews within a year of starting them.
4. How many seats should Scotland and Wales have if they are given devolved assemblies? This touches the thorny 'West Lothian question' of the powers of Scottish and Welsh MPs to determine policy for England, compared to the powers of English MPs to determine policy for Scotland and Wales. There is no easy answer to the West Lothian Question; in particular, the failure of the Irish Home Rule Bills of 1886, 1893, and 1912 to resolve satisfactorily the numbers, powers and duties of Irish MPs at Westminster contributed to their stormy fates. One `solution' to the West Lothian Question is to cut the numbers but not the powers of MPs from the area that is given devolved government. Although, logically speaking, it is not a solution, it might reduce English resentment at Scottish `privilege'. As we have seen, Northern Ireland had 12 territorial seats, some 2/3 of its proportionate share, from 1920 to 1979. If the 2/3 formula were applied to Scotland and/or Wales, Scotland would have some 40 seats and Wales some 22 in the House of Commons after devolution. However, perhaps Article XXII of the Act of Union entrenches at least 45 seats for Scotland, and this number may be regarded as a floor even after devolution. In that case 1707 is relevant after all, but not in the way Tom Johnston said it was. Such an interpretation of the Act of Union would also prohibit the Higher Education Funding Council for Scotland from closing any of the Scottish universities that existed in 1707.
NOTES Maxwell, 1988).
Especially Evershed MR in Harper, and Donaldson MR in R. v. Boundary Commission for
England ex parte Foot (1983 QB 600). See Commonwealth of Australia, The People's Say:
Elections in Australia (Canberra: Australian Government Publishing Service, 2nd ed. 1994).
Iain McLean is professor of politics, Oxford University and an affiliate of NEXUS. Since 1970, he has conducted extensive research on issues of devolution and distribution in the UK.
NEXUS