Direct Elections for a Reformed Second Chamber

Posted 01 Sep 2004


The Rt. Hon. the Lord Richard of Ammanford, QC September 2004

I greatly welcome the publication of this paper. It is a valuable and helpful addition to the argument over the nature of the second chamber.

Despite the recent changes in the composition of the House of Lords, its reform still remains unfinished business. It is therefore very useful to have a detailed analysis of the secondary mandate proposals as compared with the direct election model.

The debate on Lords Reform has perhaps been conducted overmuch along general and theoretical lines with little attention being paid to the practicalities of the various options. This paper fills that psephological gap and deserves to be widely read.


Towards a consensus

The debate on the composition and role of the House of Lords may finally be nearing a conclusion after a century of on-and-off debate. What started as a pitched battle leading to an important step in the democratisation of Britain is now fought in a rather languid fashion over a relatively small battlefield. But the second chamber is still an important element of the constitutional structure and the debate is about how far democracy is going to extend.

At the turn of the 20th Century, the House of Lords was a real danger to the emergence of democracy in Britain. It was hereditary, unaccountable, overwhelmingly biased towards a single party regardless of public opinion, and militant in the defence of the interests of one class above all others. To make matters worse its formal powers were equal to those of the elected House, although convention put some constraints in its way. When the Liberal Party won a landslide victory in 1906, the Lords impeded the government at every turn and finally abrogated the convention that it would not interfere with the government’s supply of money.

At the turn of the 21st Century few of these obnoxious features remain. Many reforms followed directly from the confrontation of the ‘peers and the people’ in 1909-11. The 1911 Parliament Act took away the Lords’ powers over money bills and ended their veto over other legislation (except Bills prolonging the life of a parliament), replacing it with the power to obstruct for two parliamentary sessions. Its powers were further reduced in 1949, and by a self-denying convention inaugurated in 1945 that it would not obstruct legislation that had received a ‘mandate’ by featuring in the winning party’s manifesto. Having tackled powers, composition then seemed a somewhat secondary matter and despite a considerable number of proposals for a reformed composition (for a discussion of these see Andrew Tyrie Reforming the Lords: A Conservative Approach, published by the Conservative Policy Forum in 1998) there were few dramatic developments. The most important by far have been two major steps that weakened the hereditary principle – life peerages, introduced in 1958, and the exclusion of all but 92 hereditary peers in 1999. With the departure of the hereditaries went single-party control of the House of Lords. The things that are wrong with the Lords – its somewhat eccentric biases (it seems to reserve its strongest feelings for the preservation of fox-hunting and the suppression of homosexuality), its unrepresentative composition, and the role of party patronage – are but shadows of the faults in their Lordships’ House a century ago.

There are still some important arguments of principle remaining. The question of whether there should be a second chamber at all is a valid one. In February 2003, 172 MPs voted to abolish the Lords entirely. However, the unicameral option has struggled to gain wide acceptance. Many people are not convinced that the Commons and its committees would be able to provide an acceptable level of scrutiny, revision and checks and balances to a majority government. Further, most countries of more than a few million population have one form or another of bicameral legislature (though New Zealand, Sweden, Bangladesh and Ukraine manage without). The British constitution vests a high level of centralised power in a government with a Commons majority, and the electoral system can give that power to a party without anything approaching a majority of public support. So long as the Commons is elected by this system and responsible for sustaining a government it cannot give the depth of scrutiny that is desirable of a legislature.

The case for a second chamber is very strong as long as the electoral system for the Commons remains unreformed.

As well as the unicameral option, there is the opposing, equally pure option of bicameral equality. Some conservative thinkers are the most constitutionally radical on this point; in the 2003 Commons debate Douglas Hogg (Viscount Hailsham, as he now is in an honorific sense) stated that ‘my preference is, in effect, for a wholly elected senate with authority and competencies equal to our own. I acknowledge that the existence of such a chamber would be a challenge to the authority of this House and I welcome that.’

Like unicameralism, full bicameralism is a perfectly respectable intellectual position, but it is if anything a more extreme departure from recent British practice than outright abolition.

However, the most generally acceptable position in the current debate is that a government with a Commons majority should be allowed to get on with governing. There should, it is conceded, be a second chamber but its powers should be limited to allow ultimate authority to remain with the Commons. After the work of the Wakeham committee and the two parliamentary committees (Joint, and Commons Public Administration), a fair degree of consensus about the role for a second chamber exists.


The second chamber is intended to be primarily a revising and debating chamber, with real but limited powers making it an effective part of a constitutional system rather than a source of authority in its own right. It is intended to be a more reflective, less tribal political environment than the Commons, with a measure of independent judgement and seniority. Independence means that, while many members will generally follow their party whip, the ethos and rules of the house should tolerate judgement and dissent and members should not be influenced by patronage (either in gratitude or expectation) or fear reprisals. Independence also, many reformers argue, should mean that the parties are not the only pathways into the second chamber.

It is a matter of general consensus that it should represent the regions and nations of the UK, and that it should fairly represent the diversity (including ethnic diversity) within the UK. No party should have an overall majority and its composition should be roughly representative of the strengths of the parties in the country. It should be a forum where all interests are heard but none dominate, unlike even the present appointed House of Lords.

Appointment, direct and indirect election


Having outlined, in the broadest and least objectionable terms, what the replacement for the Lords might do and what its composition might look like, one must now turn to how this might be accomplished.


There are essentially three methods of filling the second chamber: appointment, indirect election and direct election; a mixture of more than one method is also possible. None of these are worthless ideas. Appointment can be a quick route to representing diversity, although it also – at least as regards the party representatives are concerned – is a quick route to a second chamber dominated by the patronage of the party leaders. Politics suffers from the perception that it is conducted by a closed, self-regarding circle who lack connection to the electorate at large, and an all-appointed second chamber would heighten this perception. As Robin Cook put it:

If we exclude the public from the process of election, we should not be surprised when the public become cynical about those who pose as the people’s peers. Trust is a reciprocal quality. If we want the public to trust politicians, we must trust the public to elect the right people.

However, there may be room for a certain leavening of appointed members to ensure independent and expert representation even in a mostly elected second chamber although any more than about 20 percent of its membership would weaken its claim to be a mainly elected rather than hybrid body. If the defeat of an option favoured by a majority of the elected members because of the balance of opinion among the appointed members, became a routine occurrence this would nullify the whole point of election. Julie Morgan, in the Commons debate, was sceptical about ‘a myth about brilliant achievers taking part in the second Chamber. If their expertise is needed, they can be co-opted on to Committees in the second Chamber.’This might be a tidier method of ensuring that expertise (distinct from independence) is represented during the legislative process.

The classic form of indirect election is used in Germany, where Bundesrat members are delegates from the governments of the Länder. In a federal state there is a case in principle for the representation of the sub-national states in the national capital. However, Britain is not a federal state and even if English regional devolution proceeds quickly and smoothly it will not become one any time soon. The component institutions are not there for the usual form of indirect election. Another possibility would be to compose a house of delegates from local authorities (similar to France), which would guarantee regional diversity but at the cost of skewing composition in favour of a particular interest group.

The lack of a satisfactory basis for the usual sort of indirect election has led to the proposal for a ‘secondary mandate’ system, advocated by Billy Bragg, which has restarted the debate after the fiasco of the Commons’ failure to approve any of the proposals before it in February 2003. However, although the solution advocated by Bragg would be something of an improvement on a system of appointment, it is far from perfect and his arguments should be assessed against the alternative of direct elections, which have been the more usual demand of reformers throughout the long debate on Lords reform.

Direct election v the secondary mandate

The secondary mandate is not without its merits. It is a rational formula for obtaining the outcome that is considered desirable by most reformers, namely a chamber whose composition is in rough proportion to the balance of public opinion and in which a single party majority is unlikely. While appointment can also achieve this objective as long as members are appointed to finite terms, the secondary mandate looks more transparent and provides a more effective channel for the representation of the nations and regions of the UK, another widely favoured criterion for a second chamber.

However, it suffers from several disadvantages, some of them common to any system involving the use of closed lists and some of its own. In the former category is the lack of choice for the voter as to whom his or her vote helps elect, and the difficulty of accommodating independent candidates. According to the Times (4 May 2004) the government are considering a kind of assisted places scheme for independents, who would be inserted into party lists. This procedure could produce any number of anomalies and inequities.

The possibility of forcing the parties to take independents onto their lists illustrates quite how closely the secondary mandate can resemble appointment. Bragg himself urges that the parties choose their list candidates by as democratic an internal procedure as possible (although there is no real way of guaranteeing this); the idea of having people on the list who are not nominated by party members directly contradicts this principle. It is merely a device to make the second chamber not appear hybrid in its composition, which suddenly became a government priority in 2003.

The particular electoral system suggested by Billy Bragg has some curious properties, which make it an unsuitable choice (see Appendix A). However, this is peripheral to the central case against the secondary mandate, because the concept could easily be adapted for use with a more logical electoral system.

The most acute problem of the secondary mandate is that it makes the dilemma of the tactical voter even worse. A Liberal Democrat supporter in a Labour/ Conservative marginal in 1997 or 2001 casting a tactical vote for Labour would contribute to the election of a Labour member of the upper house. For supporters of minor parties the situation is also cruel – should a UKIP supporter vote for his or her preferred party, hoping to contribute to that voice being heard in the upper house? Or should he or she vote Conservative in their constituency, and possibly help a pro-European Tory win a seat in the second chamber? The secondary mandate is efficient but unfair.

Direct elections enable voters to exercise subtle choices that are not available with the secondary mandate.To wish for the return of a government of one party, but with its wings clipped, is a not uncommon feeling among electors, and simultaneous direct elections would allow people to achieve this by supporting their favoured party of government in the Commons election and another party for the Lords. Lords elections would also free voters to choose parties that they wish to see in a position of influence rather than power – the Green Party for instance, or the UK Independence Party.They could also use their second chamber vote to encourage opposition parties that they feel are moving in the right direction but are not yet ready for power.

In short, the single constituency X-vote is already doing too much work. It is a vote for an individual candidate and it is also a vote as to which party should be in government; it is often a vote against another candidate rather than a positive expression of support. It is asking too much to also treat it as a list vote for a second chamber.

The most effective and often-made argument against direct elections is that they would produce a second chamber that would threaten the traditional supremacy of the House of Commons. The solution, if this point is accepted, is to deprive the second chamber of any claim to a mandate or democratic legitimacy in its own right. According to Bragg:

If the reformed chamber were to be directly elected, then members would have equal legitimacy with MPs, and so the Commons would lose its primacy.The result would be democratic gridlock as each house claimed that it alone was the true representative of the people.

It is not a trivial point, particularly in circumstances when a First Past the Post election has produced a Commons majority for a party with a relatively low share of the popular vote. A second chamber that reflected the balance of public opinion more fairly might indeed claim that it is more representative of the people. An obvious answer is electoral reform for the House of Commons.

Another obvious answer is to circumscribe the formal powers of the second chamber to prevent gridlock; members of that House could claim what they liked about a democratic mandate, but they would not be able to do anything about it beyond the formal powers as set out in the Parliament Act. Andrew Tyrie argued in the Commons that the argument of gridlock:

– that is, as soon as we have elections up the Corridor, the two Chambers will be locked in mortal combat… is a load of old tosh. We have a Parliament Act, which was framed to set out the rules and limits of that relationship and to prevent gridlock from ever happening again.The Parliament Act 1911 resulted from the only gridlock crisis that we have ever had in this country. No one in their right mind in this place would ever repeal the Parliament Act. It is in place to protect us from gridlock.

Local authorities are also democratic institutions that often have fresher mandates than the House of Commons, but other than in a few isolated instances councils have not attempted to use this to challenge the authority of parliament. Their powers are regulated by law and they do not have the power, much as some may wish, to prevent the decisions of parliament from applying in their localities – even if they could claim a local popular mandate for such a course of action. Very few advocates of Lords reform favour giving a new second chamber any powers in excess of those allowed under the Parliament Act 1949, and many favour further restrictions on powers to bring them more closely in line with those the Lords have exercised in practice.

Another complementary option is to dilute the legitimacy of the second chamber through ‘partial replacement’, i.e. to elect it by halves or thirds, a familiar system in many local authorities. Electing by halves, simultaneously with general elections, would mean that half the second chamber would owe its existence to an election which took place longer ago than the last election for the House of Commons, and the other half to the same election as the Commons. Partial replacement would reduce the power any claim to democratic legitimacy.

The very concept of gridlock deserves some further examination. There are two distinct varieties of gridlock- those that make the continuation of government impossible, and those that merely impede legislation. The two forms involve two different points of principle.

Lethal gridlock

The condition for the existence of a government in Britain is that it sustains the confidence of the House of Commons. In some constitutional systems the executive does not require the confidence of either house, because the presidential executive receives a mandate directly from the electorate as in the United States. In a fully bicameral parliamentary system a government will require the confidence of both houses. The principal case of this is Italy.

Before the 1911 Parliament Act Britain was in an ambiguous position, in that there was nothing to stop the Lords from breaking convention and denying the government supply of funds, in effect terminating its existence. A similar ambiguity persists in Australia, where the Senate attempted to choke off government funds in 1974, prompting an election, and 1975 culminating in the fall of the government. The need for the support of both chambers has been one of the elements encouraging instability in Italian government, at least until the more majoritarian post-1994 dispensation – although the prospect of hostile majorities in Chamber of Deputies and Senate still exists.

It is useless to imagine when designing a constitutional system that drastic circumstances will not happen – from time to time they do. A clear choice should be made between bicameral equality and giving the responsibility for forming a government unambiguously to one chamber, to avoid ‘unconstitutional’ episodes like those in 1909 in Britain and 1974-75 in Australia.

The implication of this position is the carrying-over of the provisions of the Parliament Act relating to money bills, and conceivably – given the relatively small number of bills certified as money bills – a broadening of the class of legislative business that is the sole preserve of the House of Commons. It should be technically and legally impossible for a new second chamber to bring government to a halt.

Disagreements on legislation

Logically distinct from the power of life and death is the power to obstruct the passage of legislation, once one has separated out legislation that is essential for the functioning of government. The distinction between these categories underlies the different treatment of money bills (no veto) and other bills (suspensory veto) in the Parliament Act. It is in this area that a directly elected chamber might make more of a challenge to the Commons, more through an increased feeling of legitimacy than increased formal powers.

In all likelihood a directly elected second chamber would feel more entitled to revise, make significant amendments to and even reject legislation that the Commons has approved. Moves away from the hereditary principle since the introduction of life peers in 1958 and particularly since the exclusion of all but 92 hereditaries in 1999, have been accompanied by an increased willingness to use existing powers, which have remained constant. Both the secondary mandate and direct election are likely to result in an intensification of this trend, and quite probably an effect on a wider range of subjects than currently seem to interest the Lords. With either system, although perhaps particularly with direct election, powers would be likely to be used. It would be part of the package of a more democratic second chamber that the process of legislating would be more thorough than at present, and possibly a bit slower.

The Salisbury convention dating from 1945 and preventing the second chamber from obstructing legislation promised in the governing party’s manifesto was originally an accommodation to the specific circumstances of a Conservative-dominated Lords and a majority Labour government. It has much to commend it, but it is questionable whether it need always apply in a situation where the second chamber is representative to a considerable degree of public opinion – whether by election, secondary mandate or even appointment. It is probably not possible to codify the Salisbury convention in the form in which it has existed in the past, but it would seem desirable to prevent circumstances developing where a government’s programme could be held up for its first year of office. This might be achieved by reducing the suspensory period allowed in the Parliament Act.

There may not be a need for a reconciliation procedure beyond what is specified in the Parliament Act – that if, after a time lag, the Commons passes the legislation again the Lords are not able to obstruct it any further or insist on their amendments. An alternative, permitted in some circumstances in the Australian constitution, is to convene a joint sitting in which the greater numbers in the lower house can generally outvote the smaller upper house.This could be an option worthy of consideration, either as an alternative to the Parliament Act procedure or as a fast- track supplementary procedure that a government could use in, say, its first year of office.

Opponents of direct elections sometimes talk as if paralysis is a permanent, inevitable condition. It is not – the occasions on which it has happened, historically, reveal problems with the definition of powers, or a poorly designed electoral system. The onus should be on opponents of direct elections to find reasons to deny the electorate a direct voice. To his credit, Bragg makes his argument on this point openly, but his concerns about the possibility of gridlock are unfounded. It is quite possible to devise a better alternative that gives control over the composition of a second chamber to the electorate.

Devising a second chamber electoral system

Timing of elections

There are two main options for timing upper house elections:

According to a fixed timescale
Concurrent with general elections

If the fixed option is chosen, a five-year cycle concurrent with the European Parliament elections would be the most logical, to avoid a profusion of election days. But even this would be unlikely to stimulate turnout. William Hague put the argument well in January 2003:

The idea was floated in our Committee that the vote could be combined with the European elections. The more I think about it, the more I think that even that would not work. We can all imagine the scene.There is a knock at the door; a woman stands there with a child in her arms, while another screams in the background. I am sorry—I am slipping into a stereotype. The dinner is on the stove, while the man of the house is working hard on his DIY. The woman is asked, “Will you vote in the elections for the upper House?” She says “No.” “It is combined with the European elections.” She is not going to say “Oh well, in that case, of course!”

The fixed term option is also more likely, depending on the electoral cycle, to produce a chamber at odds with the party of government.1 If the second chamber election falls at a midpoint on the cycle, the opposition party is likely to do well given normal midterm disillusion and differential turnout. In extreme cases this might lead to outright control falling to the principal opposition in a midterm election, even if a fairly proportional system and partial replacement are used.2 The result would be most conducive to gridlock – different party control in each house, with opposition control in the upper house resting on a newer mandate than the government’s in the lower house. Proponents of midterm Lords elections argue that a government which has become sufficiently unpopular to cause such a result deserves to have its powers circumscribed in this way. Because the general election cycle is not fixed, there may also be occasions when the second chamber election takes place during a honeymoon period. A government which wished to have a compliant second chamber and was doing well in the polls may be well advised to call the general election a month before the fixed term election. This might, as in the June 1979 Euro elections and May 1992 local elections, lead to a landslide win in the second election because of the demoralisation of its opponents.

The option of running upper house elections at the same time as general elections would probably lead to a higher turnout than would be likely in a midterm contest, and particularly than a contest just after a general election. This option would automatically mean that the upper house could not claim a fresher mandate, even in part, than the government. It is likely to deliver the outcome that the Bragg scheme and efforts to engineer an appointed system are designed to achieve – a house in which the government is the largest single party but without a majority.3 A fixed- cycle election may not do this.

It may be argued that elections concurrent with a general election will simply replicate the outcome of the general election. This is certainly not the case if FPTP continues to be used for the Commons, and is not even the case with the secondary mandate provided secondary mandate seats are allocated proportionally. However, the evidence suggests that voters will use concurrent direct elections more creatively. In Scotland in 2003 there were significant differences between list and constituency voting. On the same day in June 2001 voters in many constituencies cast different votes for local authority and Westminster elections (in Bristol West Labour held the parliamentary seat while the Lib Dems dominated council election voting). Recent polling evidence shows that voters can distinguish between different levels of election:

% vote for… Euro Local General
Conservative 28 32 36
Labour 27 30 33
Liberal Democrat 18 22 19
UKIP 14 5 4
Green 6 3 2
Others 7 9 6

Source:YouGov for Daily Telegraph 24 May 2004; 1928 online surveys completed, fieldwork May 20-21 2004.

Given the precedents, it seems likely that smaller parties will fare relatively well in the second chamber election, even if it runs concurrently with a general election, while the general election itself will remain to a considerable extent a contest for the major parties.

These two principal options are each worthy of consideration in devising a second chamber electoral system.

Form and function

The functions that a second chamber will have are the subject of a short discussion in the introduction. No less significant, there are two significant functions that a new second chamber will not have, and therefore important consequences for the electoral system.

Sustaining a government

In the light of the assumption that one is not opting for bicameral equality, the traditional, always questionable FPTP argument – that it facilitates the formation of strong and stable governments – is utterly inappropriate in devising a system of direct elections for an upper house.

Constituency representation

Arguments about constituency service are not relevant to an upper house, given that constituency representation is a jealously guarded responsibility of the constituency MP. Preserving this representative role is an important feature of possible reforms to the Commons system and circumscribes the possible systems for that chamber (FPTP, STV, AV, and AMS all preserve the local representative role). A system that creates the potential for a conflict between members of the houses each with a claim to territorial representation is likely to lead to problems. In any case, it is unlikely to attract the support of MPs – which is necessary if it is to be passed into law.

The House of Commons Public Administration Committee reported in 2002 that:

We agree with the Government that elections for the second chamber should be based on the same regional constituencies as those used in the European Parliament elections. These members are to represent the ‘nations and regions’, and it makes sense for their constituencies to match the nations and regions of the UK. It also helps to distinguish them further from MPs in the House of Commons: the larger the area, the less likely they are to do constituency work. We also recommend a formal convention to prevent this.

Many schemes for an elected second chamber also involve term limitation.The Public Administration committee stated:

We recommend that elected second chamber members should serve a single term extending to two Parliaments. No member of the second chamber

should be permitted to stand for election to the Commons for ten years after leaving the second chamber. These restrictions would apply from the next general election. Political parties should not be allowed to nominate for appointment anyone who has served as an elected member of the second chamber.

Its members are, therefore, if this argument is to be accepted, to be rendered deliberately unaccountable in the normal sense. They would not be answerable to their parties, or the electorate, for their voting behaviour in the second chamber. Their individual judgement is to be accorded a higher priority in designing the system than their relationship to the party whip or the electorate. This means that an argument that one electoral system gives more accountability of members to constituents than another is irrelevant in thinking about the arrangements for the new second chamber.

In the light of this, single-member majoritarian systems (FPTP and AV) are therefore unsuitable systems for a second chamber. Electoral districts for a second chamber election are therefore principally for the purpose of ensuring a spread of regional representation and giving voters a realistic degree of knowledge and hence informed choice about the options.

Partial replacement

Partial replacement and long terms have also become nearly a matter of consensus, to the point where schemes that do not incorporate it (such as Billy Bragg’s) are designed to attempt to replicate its effects. Directly elected Senates in Argentina, Chile and Brazil depend on partial replacement to establish a greater degree of legitimacy for the lower house, and a greater degree of wisdom and experience in the upper house. The same principles were intended to govern the US Senate and House. French Senators, indirectly elected, have 9-year terms and come up for election in thirds.

The Commons Public Administration Committee considered the case for electing by thirds or by halves, and came down favouring halves. The principal difficulty of electing by halves concurrently with general elections is that parliaments have indeterminate terms and when elections follow each other in quick succession this can result in rather short terms for members of the second chamber. Historically, someone elected in June 1970 would finish their term in October 1974 and not be eligible for another go, while someone elected in June 1987 would be able to serve until May 1997. There is a case for making some sort of stipulation that someone whose two-parliament term ends up being particularly short should be eligible for re-election. Perhaps 5 or 6 years would be a suitable threshold, below which would not count for term limitation. Five years would have allowed those elected in 1970 to face re-election if they wished; a six year rule would also allow those first elected in 1950, 1964 and February 1974 another crack of the whip.

Partial replacement does provide continuity of membership and helps to prevent parties gaining outright majorities on the basis of particularly strong election performances in one year. It is also an important mechanism in reducing the potential for a conflict of mandates. Running it concurrently with general elections also makes it possible for governments, provided they have public support in their stand, to flush out the oldest section of a troublesome upper house by means of a dissolution.


Partial replacement can of course be applied to secondary mandate as well as direct elections.The following tables illustrate the results of a regionally elected secondary mandate, using a proportionate allocation of seats to regions and the D’Hondt formula (as used in European Parliament elections) of allocating seats to parties.4 A membership of the second chamber of 300 is assumed; the issue of size is dealt with later, and the proportions will not vary much with different numbers of members. The results of direct elections may well have been different.


300 seats






Lib/ LD







1983 65 42 38 1 4 0
1987 64 (129) 47 (89) 34 (72) 2 (3) 3 (7) 0
1992 65 (129) 54 (103) 24 (58) 3 (5) 4 (7) 0
1997 48 (113) 73 (127) 22 (46) 4 (7) 3 (7) 0
2001 49 (97) 66 (139) 27 (49) 4 (8) 4 (7) 0

Note that Labour would have come quite close to winning a majority of seats in 1997 and that in 2001 Conservatives plus Lib Dems is not sufficient for a majority. Looking back, it is probable that the Conservatives would have enjoyed an overall majority on a secondary mandate during the 1959 parliament because their vote was over 49 percent in two successive elections.

Size of the second chamber

The first Joint Committee Report suggested a House of approximately 600, on the grounds that independent appointed members with specialised knowledge would not wish to attend every debate and large numbers would be necessary to ensure a spread of expertise. This proposal was heavily criticised in the Commons debate in early 2003, with speakers from all sides arguing for a smaller chamber. Robin Cook noted that on international comparisons very few second chambers are larger than 200. The precise figure is dependent on technical factors to do with the functional efficiency of the chamber, but it does have consequences that relate to the electoral system to be used.

One problem of having a chamber much smaller than 200 with partial replacement is that too few people will be elected at one time in the smaller regions like the North East and Northern Ireland. It would be a serious disservice to the complexities of politics in Northern Ireland if a single (or even two) members were all that could be elected at a time. In a wider sense, proportionality depends on electing several members at a time and to have too small a chamber would mean that the overall result might be unfair between the main parties, as well as setting an unacceptably high threshold for smaller parties and independents.

In devising a system a number in the region of 300 members should be borne in mind as a happy medium between too large a second chamber and too crude an electoral system. It is also quite possible that an appointed element will persist as a mechanism to get independent and expert opinion into the second chamber (few reformers would have much difficulty with an 80% elected, 20% appointed chamber). An elected element of 300 might mean a total size of 350-400 members.

Which system?

Given the criteria of proportionality, and large base electoral units, the regions used for European Parliament elections (and other governmental business) are the obvious basis for second chamber elections. The House of Commons Public Administration Committee and the government both agreed with using these regional boundaries.

There are essentially two systems that are compatible with the criteria for electing members of the second chamber – regional lists and the Single Transferable Vote (STV). This also was recognised in the Public Administration Committee’s report:

We recommend that any voting system for the second chamber should satisfy the following general It needs to:

  • be complementary to the voting system for the House of Commons;
  • minimise the risk of one party gaining an overall majority;
  • maximise voter choice, by enabling voters to vote for individual candidates, within and across parties;
  • encourage a more diverse chamber; and
  • encourage the election of independent-minded people.

These principles will best be realised by using multi-member constituencies, and a proportional voting system. This could be either STV or regional lists, so long as the lists are fully open lists, which maximise voter choice. We would not support limited open lists, which present an appearance of choice for the voter, but almost never affect the outcome.

Broken down more precisely, the choice of the Single Transferable Vote (STV) or lists gives four real options for a second chamber electoral system:

  • Closed regional lists
  • Open regional lists
  • STV with regional constituencies
  • STV with sub-regional constituencies.

Closed and semi-open lists

The House of Commons Public Administration Committee report argued, for sound reasons, against closed lists and list systems that are theoretically open but in practice make

varying the party’s chosen order extremely difficult. Closed lists would be likely to expose the second chamber even further to allegations that it was a patronage chamber for party hacks. Closed lists allow the parties to guarantee some people membership of the chamber regardless of what voters think of them as individuals. The Independent Commission on PR found that voters disliked the lack of choice in closed list systems and (even if the element of choice is illusory) an open system stands a greater chance of public acceptance. A genuinely open list system would be acceptable, but at the cost of greater complexity and the possibility that like closed lists it would make life difficult for independents.5

Whole-region STV

STV in whole regions is also possible, but in some regions this would involve very long ballot papers. The South East, for instance, would be entitled to 41 seats in a 300-member house, and 20 or 21 would be elected in one go. This is a much larger number than is generally considered acceptable for use in STV. Even the median region would be electing 12 or 13 at a time, with ballot papers almost certainly featuring more than 50 candidates (and conceivably over 100 in the larger regions). STV in large regions like this would be likely to produce significant ballot order effects and a high spoilage rate. Candidate promotion and giving voters sufficient information to enable an informed choice would be difficult, particularly in a second-order election.

One option for circumventing the problem of large numbers would be to add the possibility of casting a pure party vote, as in the 6 and 12 member elections for the Senate in Australia. Since 1984 Australian voters in Senate elections have been given the choice of two ways to vote. They can continue to vote in the classical STV fashion by numbering individual candidates, what is called ‘below the line’ voting; or they can vote ‘above the line’ by choosing a single box and supporting the party ticket. Above the line voting is still nominally STV, in that by choosing the party’s box the voter’s preferences are deemed to have been cast in the favoured order of that party. But for the individual voter it is equivalent to a party list vote. Over 95 per cent of voters choose to vote above the line. While this proportion is lower in Tasmania where STV is more familiar because it is used for state elections, even there party line voters outnumber choice voters by more than 2 to 1. If the extent of party line voting is anything like this, the system as a whole behaves much like a list, at least as far as the major parties are concerned. Any non-approved choices by individual voters are generally swamped by the overwhelming torrent of party votes cascading down the party’s preferred ordering. While the Australian Senate is a more partisan and powerful institution than is proposed for a UK second chamber, and British voters probably less party-minded than Australians, STV in whole regions could easily devolve into being just a variant on a closed list.

An interesting recent development in Australian Senate elections has been an increase in the number of minor party tickets. Some of these are fronts intended to grab votes from the unwary. A voter for a party with an innocuous, non-partisan sounding name can find that vote, once the bogus minor party is eliminated and preferences transferred, fed into the mainstream of the major party that has put up the front party. Another development has been the agreement between genuine minor parties, even those with wildly different ideologies, to trade preferences with each other rather than give them to the major parties. In 2001 in New South Wales, Mick Gallagher of the fringe ‘No GST’ party started with 25,734 votes and levered this up thanks to transfers from other micro-parties to 191,583 by the time he was eliminated. He was only 15,000 short of knocking out the One Nation party and successively riding its transfers and then Green transfers to a seat in the Senate. All this suggests that there is something to be said for leaving transfers of preferences to the voters to decide.

Having eliminated closed lists and macro-STV as unsuitable, this leaves two reasonably good systems – genuinely open lists and sub-regional STV.

Open regional lists

If elections are to take place in constituencies defined purely as the English regions, Scotland, Wales and Northern Ireland, a list system would appear to be the principal contender. However, short of appointment or the secondary mandate, a closed list would be the system most likely to deliver control over the second chamber to the party machines and eliminate any spark of interest or enthusiasm for second chamber elections. If lists there must be, then voters must have a genuine choice as to which candidates are elected under each party’s colours. The price of the greater voter choice represented by open lists as opposed to closed is an increase in complexity.

Some reformers have considered the Belgian form of list suitable for UK use, for instance by Charter 88 before the European elections of 1999. Voters are permitted to vote with an ‘X’ either for a party list as a whole or as a personal vote for a candidate, including those low on party lists.

Candidates whose personal vote exceeds a quota (i.e. the party’s total vote divided by the number of seats won by the party plus 1, rounded up) are elected. Once these have been taken out of the equation, the party’s list vote is distributed among candidates in the party’s preferred list order.The threshold for winning a seat on the personal vote is therefore variable, dependent on the size of the region and the strength of the party’s vote. In a good year for the Conservatives, the largest party in the largest region, it would be relatively low – perhaps 1 in 12 Conservative voters would be enough to elect someone from lower down the list in the South East. But in practice the hurdle will usually be higher, and one should have few illusions about how many changes in list order will take place. It does not expand voter choice very much.

However, the parties would be made more conscious of the desirability of putting up a balanced list of candidates with some degree of personal popularity. A more radical form of open list election would be to abolish the party ticket vote and the threshold for personal election and leave the ordering of the list entirely up to the voters who choose that party. The Finnish system takes personal votes as being representative of the party vote – the number of seats going to each party is determined by the proportion of the total vote going to its candidates. Which people represent the party is determined by the voters – they are elected in order of personal votes. This represents a big increase in voter choice on closed or semi-open lists, but it still retains a key feature of any list system. By voting for a party candidate, you are increasing that party’s entitlement to seats, and the candidate you are helping, at the margin, to elect may not be one that you find attractive.

It may be easier to elect a candidate from lower down a major party’s list using open lists than with modified STV, or even pure STV in large districts. It is, however, more difficult to show support across party lines and when minor parties and Independents are a factor can result in tactical and wasted votes.


  1. Of course, the composition of any elected second chamber would be likely to differ from that of the Commons, particularly if one election is proportional and the other It is possible that even if support for each party were exactly the same, the government would be outnumbered by its opponents in the second chamber. But this is different from a situation which would leave government supporters an embattled minority in the second chamber as a pair of results like those of the 1999 and 2004 Euro elections would do.
  2. The most likely historical case, projecting the European Parliament cycle backwards, would be a Conservative win in the upper chamber elections in June 1969 of sufficient size as to transfer control immediately and maintain it until June 1979 or even June
  3. It does not always produce this outcome, although the secondary mandate would in every parliament from October 1974 Oddly, the Conservative government of 1970-74 would face a Labour-led upper house, while the Labour government between the two 1974 elections would see the Conservatives lead the upper house. In 1964-66 the Conservatives would also be the likely leading party of the upper house. There are two reasons for this. In 1970 (and 1964) the new government’s election victory was by a smaller margin than the outgoing governments’ leads in 1966 (and 1959). The Labour advantage based on 1966 would be more than the Conservative lead in 1970, and the combination would produce a narrow Labour lead. The counter-intuitive change in Feb 1974 would be because the pro-Labour delegation of 1966 would drop out to be replaced by a closely balanced new crop from 1974.
  4. For those interested in electoral formulas, the allocation using Sainte- Lague divisors is shown below:


300 seats Con Lab Lib/ LD Nats NI Others
1983 64 41 38 3 4 0
1987 62 (126) 47 (88) 35 (73) 3 (6) 3 (7) 0
1992 63 (125) 52 (99) 27 (62) 4 (7) 4 (7) 0
1997 47 (110) 68 (120) 27 (54) 4 (8) 3 (7) 1 (1)
2001 49 (96) 64 (132) 28 (55) 4 (8) 4 (7) 1 (2)
  1. In list systems people considering voting for an Independent or minor party have to come to a view as to whether the candidate stands a chance of being elected, e. of getting about one quota. If not, the vote may well be wasted. If there are too few other supporters, those votes have no way of influencing the result. If there are too many, the vote is also wasted because no matter how many votes an Independent gets, only one person can win a seat as a result.

The preferred option: STV

The Single Transferable Vote

The obvious solution to the problems of macro-STV is to subdivide the regions into electoral divisions allowing a sensible number of members of the second chamber to be elected at each election.

STV is an electoral system which is suitable for use in many sorts of election, but it may be particularly well suited to second chamber elections. When compared with the criteria of the House of Commons Public Administration, STV fits extremely well:

  • As a proportional system, STV will certainly minimise the risk of one party gaining an overall majority, particularly if it is elected by partial replacement.
  • STV is the system best suited to maximising voter choice. Unlike list systems it does not require complex calibrations to ensure that popular individuals within each party are elected, rather than those most favoured by the party organisation. It is a most unusual system in that it allows voters to express approval across party lines through the transfer of votes to second and lower preferences.
  • STV offers parties an incentive to appeal to different sections of the population within each region, in order to attract the maximum number of first preference votes.The sub-regional constituencies would encourage an even geographical spread of members.
  • STV is centred on the candidate, not the party. If people who are independent of any of the parties stand for election, they are able to do so and attract votes on their own merits. It is quite possible to imagine a situation in which a highly popular person stands in one region on a non-party basis for the second chamber. Voters who normally support a party are free to choose that independent, confident in the knowledge that if he or she is elected overwhelmingly (or defeated) their vote can then be used to help their regular party.

How large should the constituencies be?

The precise district magnitude is not something to be unduly dogmatic about at this stage. There are two broad possibilities, assuming about 300 are to be elected in two tranches of 150: a pattern of smaller districts electing 3-6 at a time (mostly 3 to 5), and larger districts electing 5-8 at a time (mostly 6 and 7). The merits of a larger scheme is that it is more proportional (and hence more permeable for independents and smaller parties), and that it is more flexible to respect natural boundaries and less likely to need future alterations. On the other hand, it would involve long ballot papers and might seem more remote.


Boundary determination

The problem of boundary determination is one that needs to be addressed once one goes beyond using whole regions. Desirable features for second chamber electoral districts would include:

  1. Equal level of representation for all
  2. Infrequent alteration.
  3. Coherent boundaries that reflect some sort of community or geographical
  4. Coterminosity with Commons
  5. Coterminosity with local government

The first is a given starting point in any democratic system, although it can be modified by federalism and other factors. In the UK situation seats should be allocated to regions as proportionally as possible. However, there is a case for adding a ‘floor’, stating that in that in no case would the number of people representing any region be less than, say, 8 or 10 (this would produce a slight degree of overrepresentation for Northern Ireland). Once one starts subdividing regions, the possibility of anomalies in allocation emerge, in that the entitlement of a region could be more (or less) than the entitlement of the sum of its parts. As with the current Parliamentary Boundary Commissions’ treatment of counties, the arithmetic should first arrive at a whole-number entitlement for the entire region, which is then parcelled out into the subdivisions.

The second is a desirable feature in giving the second chamber some element of continuity. It is supposed to speak for the nations and regions of the UK, not a rather arbitrary and often- changed constituency. Frequent alterations to boundaries would cause confusion to electors and elected representatives alike, particularly if partial renewal is to be a feature of the system. It is easy enough to adjust entitlements up and down in steps of 1 while preserving the same boundaries in multi-member systems; boundary reviews and all-out elections should be unnecessary provided that the initial designation of districts commanded widespread acceptance.

Although electoral districting is secondary to providing an acceptable composition of the chamber as a whole, the legitimacy of the system would be improved by having recognisable rather than arbitrary districts as the basis for elections. This would assist candidate promotion and the sense of a real election taking place, as opposed to the rather remote and technocratic European Parliament regional lists. Administrative convenience, the avoidance of confusion and legitimacy all militate for respecting existing local authority and Westminster boundaries where possible.

In the light of 2, 3, 4 and 5 a starting point of counties within England, whole boroughs in London and the List regions in Scotland and Wales, would seem to offer the most attractive solution. Where possible this should be ‘old’ counties, i.e. grouping unitary authorities with their parent county. If delineation by county causes problems, the first option is to split unitary authorities (or divide former metropolitan counties as long as boroughs remain whole), and only if the argument is strong, arithmetically and in terms of community ties, to divide counties by having district council areas in different senatorial seats.

Westminster constituencies are an unsuitable building block for an STV districting scheme for a second chamber. The principal problem arises from the periodic reviews of the Boundary Commission, which take place on what amounts to a 12-year cycle. There is no guarantee that reviews will not create non-coterminous boundaries between senatorial districts and parliamentary constituencies. The alternatives are either to alter senatorial districts at the same time (which makes long terms and partial replacement nearly impossible), or to tolerate the existence of confusing and irrational anomalies for sustained periods.

An alternative basis for senatorial districts that has more continuity than Westminster constituency boundaries is therefore required. Local authorities are more suitable, as their boundaries are revised much less frequently than constituencies. They also have the merit of better reflecting natural communities, while constituencies are sometimes rather arbitrarily drawn in order to achieve units of roughly equal size. In due course it might be possible to draw subdivision lines based on groupings of single-tier local authorities below the level of regional government, which can be presumed to have continuity for a long time thereafter.

Periodic revisions to the second chamber electoral areas may take place on a more infrequent basis than for the Commons. The principal means of rectifying inequalities in representation would usually be adjustments of plus or minus 1 to the constituency’s entitlement, changes which can easily be accommodated in the partial replacement cycle.

An illustrative scheme is provided in Appendix B to show how STV constituencies within the regions might work. No great significance should be attached to any boundary arrangement shown. If STV in sub-districts were to be adopted as the second chamber electoral system, the borders of each constituency should be drawn with a maximum of consultation and sensitivity to ensure that the boundaries would respect natural units and be an acceptable basis for a long-lasting system.

Transitional arrangements

Introducing any new system does involve a transitional problem of electing all the members of the house in one go. This applies just as much to lists and the Bragg scheme as STV. As in Australian whole-Senate elections, establishing the whole chamber at once means a temporary lowering of the effective threshold for representation. It is likely that parties and candidates would be elected in the first election who would not be successful when regular competition, for half as many seats, began at the second election. The first election would also involve double-length ballot papers with the consequent possibility of confusion. It is asking a lot of voters to struggle through a ballot for 12 members to be elected by STV.

Billy Bragg’s means of getting around the transitional problem would be to keep 150 life peers in the House for the first term, in proportion to the party votes in the general election. This has much merit in alleviating the problem of transition and smoothing the process. There are ample House of Lords precedents for ‘representative peers’ from disqualified categories (Scottish, Irish, and hereditary peers at different stages) continuing to sit for a transitional period. Perhaps, as well as being representative of the parties’ strength, the 150 holdovers from the old Lords might include the most active and a fair measure of independents (perhaps even independents in proportion to abstention at the general election!).

This noted, transitional problems – unless they are extremely severe – are poor arguments against what is intended to be a lasting settlement.

The benefits of STV

  1. STV is broadly proportional. The precise level of proportionality will depend on the number of elected members in each seat. A larger number will mean a more proportional final result. Because voters rank candidates in order of preference, votes which cannot be used to benefit an elector’s first choice candidate can transfer to their next choice and so on. This tends to mean that the number of votes required for election is less rigidly fixed than in list elections and tends to mean that extremist parties do not win seats unless they can fill an entire quota.
  2. STV offers increased voter choice. Because voters can choose between candidates of the same party as well as simply between parties, they have far more choice than any other electoral system. Whereas in an open list election voters might get to plump for a single candidate (and thus have a certain amount of choice) preferential voting allows them to say ‘if I cannot have my first choice, then I would like my second choice instead’.Voters are not limited to ranking the candidates from a single party but can give preferences that cross party lines. Whilst parties still retain a say over which candidates are put forward to represent them, it is the voters who decide which of these are elected. This is all the more valuable as a feature of an electoral system for a second chamber, where an excess of party discipline defeats the purpose of the institution.
  3. STV is fair to independents. Most systems of proportional representation demand some form of party allegiance in order to guarantee election and they therefore discriminate against candidates running as independents. STV is different because it places all candidates on an equal footing and they run as individuals. It is the system for second chamber elections that gives genuine independents a fair chance, and may render unnecessary the small appointed element that many deem the only way of getting independent expertise into the second chamber.
  4. STV ends safe seats. Because voters have the choice between candidates of the same party, there is no longer any form of safe seat. Party loyalists have to compete with mavericks for the support of the electorate and parties cannot manipulate their lists to effectively guarantee the election of favourite sons or daughters.
  5. STV avoids the need for tactical voting. Because electors know that if their vote cannot be used to help their first choice candidate win election it will be transferred to their next choice, they will be more likely to vote for their genuine first choice rather than feel the need to vote
  6. STV allows electors to vote against unpopular candidates and parties. Because voters can rank candidates in order of preference, they can effectively discriminate against candidates as well as for This is particularly valuable in cases where there are candidates representing extremist and racist parties. Preferential voting allows electors to effectively say ‘I would have any candidate other than the extremist’.


The long debate

Perhaps, at last, the end is in sight for the long debate about the House of Lords. The hereditary principle went from being all-powerful, to being successfully challenged and supplanted, and then to being an embarrassing remnant of pre-democratic times. Appointment may suffer the same fate. It is not appropriate in a democratic society for a chamber of the legislature to be hand-picked by the political establishment, whether that be the Prime Minister, an august quango of some sort or the cover of the secondary mandate. There is ultimately, when it comes to claiming the authority to legislate and govern, no substitute for the endorsement of the electorate in a direct election. Reformers should settle for nothing less than a predominantly elected second chamber, and press for a system of election which maximises the power of the electorate to determine its composition. We in the Electoral Reform Society believe that system to be the Single Transferable Vote.

Appendix A

The Bragg electoral system

The Bragg scheme proposes as a starting point taking the 12 regions used for European Parliament elections and regarding them as being of equal standing with each other and having a theoretical entitlement of 25 seats. The reason for this disproportional allocation is in order to ‘counter the concentration of political power in London and the South East.’

  Proportional entitlement Bragg effect Electors per senator Value of a vote (SE=1)
Northern Ireland 7 +18 42,880 5.7
North East 13 +12 76,863 3.2
Wales 15 +10 89,080 2.7
East Midlands 22 +3 128,936 1.9
Yorkshire/ Humber 26 -1 149,880 1.6
South West 26 -1 153,232 1.6
Scotland 26 -1 155,706 1.6
West Midlands 27 -2 160,441 1.5
Eastern 28 -3 165,307 1.5
London 34 -9 202,403 1.2
North West 35 -10 207,815 1.2
South East 41 -16 243,445 1.0

The proposal does reduce the influence of London and the South East combined from 75 to 50 (in practice a bit less because of the allocation system – see below); the principal beneficiary is Northern Ireland. There is also, in effect, a large transfer from the North West to the North East.

Whether it is just to do this is highly debatable. An incidental effect of giving London less than its entitlement might well be to under-represent ethnic minorities; a problem aggravated by the fact that the three principal beneficiaries all have much lower than average ethnic minority populations.

In federal systems representation in the upper house is commonly based on the regional units rather than on population. In the foundation of the United States and Australia it was an explicit part of the bargain between small states and large, to safeguard the interests of smaller states from being steamrollered by the more populous states. In Germany, overrepresentation of small states is also part of the constitutional order. 18 US Senators for the 9 most populous states represent more people than the other 82 combined! This is starting to prove a significant problem in the US, in that a majority of small-state senators from nearly all-white, rural states can effectively block measures that are clearly in the majority interest.

Malapportionment may or may not be an acceptable price to create or preserve a federation but Britain is not and shows no sign of becoming a federal state; and a House of Lords reform is much more likely to happen before federation than vice versa, to put it mildly. Apportionment should be on the basis of population.

However, voting in the second chamber is likely to be based around party rather than region on most issues. The net effect of giving an equal number to each region is to give a bonus to the Northern Ireland parties, principally at the expense of the Conservatives and Liberal Democrats. The secondary mandate outcome for the 2001 would be, assuming all else is held constant (300 members elected using the same allocation formula, for current purposes D’Hondt):

  Proportionate by region Bragg Bragg effect
Labour 131 129 -2
Conservative 99 88 -11
Liberal Democrat 54 49 -5
Nationalist 8 9 +1
UKIP 1 0 -1
Northern Ireland 7 25 +18

There were over 10 times more Conservative voters than Northern Ireland voters in 2001, but they would receive only 3.5 times as much second chamber representation.

However, the Bragg scheme proposes an unusual allocation formula. A party’s regional list is entitled to one seat for every whole Hare quota it polls (i.e. 4 percent of the regional vote). Fractional entitlements are ignored. This means that the regional distribution will leave a certain number of seats unfilled.

The number of seats left vacant after the regional distribution will be determined by the number of parties – even minor parties – contesting the election, as well as an element of chance. The minimum, unless there are more than 25 parties, will always be zero, in that it is theoretically possible that each party might get a percentage vote exactly divisible by 4. In practice, it is likely to be 1. The maximum left vacant will be, a little arithmetic should show, the number of parties standing in the region, minus one. In an election such as 1951, dominated by two major parties, there would probably be only 12 or 13 vacancies unfilled from the regional lists. In the 1997 and 2001 elections there would have been 33 and 31 seats because of the greater fragmentation of the party system.

The seats left unfilled after the regional distribution are allocated to the parties according to their national shares of the vote. The national list, according to the secondary mandate campaign, enables parties to get ministerial candidates elected without imposing them on the regional lists. However, the indeterminacy of the size of the national list constituency makes this a relatively unattractive proposition – a party winning 40 per cent could gain as few as 6 seats (if there were 15 remainder seats) or as many as 14 (with 35 remainders).

The national list might also serve as a backdoor to representation that evades the 4 percent regional threshold. Using largest remainders (the system most closely approximating the regional distribution rule used by Bragg) in 1997 0.8 percent would suffice to gain a national seat for the UUP. 6

A further problem is caused by the mechanism which attempts to produce partial replacement of members even though all 300 members are elected at one go. Bragg proposes:

At the end of each parliament, the 50% of members who had come to the end of their second term would leave, while the remaining 50% would be placed at the top of their regional party list in the order in which they were originally selected. New candidates would then be added to the list below them.

This does, more or less, work for the larger parties. However, any smaller party that only manages to elect a single member, or gains, say, 9 percent in a region in one general election but only 3 percent in the next, will risk being wiped out after a single term.

The electoral system proposed by Bragg is unduly complicated and inequitable. Simpler regional list systems are available and it would make the system more legitimate to relate regional representation to population. However, even if these problems with the Bragg scheme are fixed, the secondary mandate is still inferior to direct election.


  1. Using the Sainte-Lague formula gives seats to parties with less than 2 per cent support, such as the SNP in both years, and the Referendum Party in 1997. D’Hondt would not give any national seats other than to the three main parties in 1997 or 2001.

Appendix B


Two schemes for sub-regional senatorial districts to produce a house of 300 members.

Smaller districts (3-6 elected at a time)

Eastern (28)

One district electing 11 seats (6:5) comprising Norfolk, Suffolk and Cambridgeshire.

One district electing 8 seats (4:4) comprising Bedfordshire and Hertfordshire.

One district electing 9 seats (5:4) comprising Essex; this district’s entitlement is only marginally 9.

East Midlands (22)

One district electing 8 seats (4:4) comprising Northamptonshire and Leicestershire.

One district electing 7 seats (4:3) comprising Derbyshire and the Ashfield, Mansfield and Bassetlaw districts of Nottinghamshire.

One district electing 7 seats (4:3) comprising Lincolnshire and the remainder of Nottinghamshire.

London (34)

One district (A) electing 7 seats (4:3) comprising the London boroughs of Havering, Redbridge, Barking & Dagenham, Newham, Waltham Forest, Hackney and Tower Hamlets.

One district (B) electing 9 seats (5:4) comprising the London boroughs of Enfield, Haringey, Islington, Barnet, Camden, Harrow, Brent, Westminster, Kensington & Chelsea and the City of London.

One district (C) electing 9 seats (5:4) comprising the London boroughs of Hillingdon, Ealing, Hounslow, Hammersmith & Fulham, Richmond-upon-Thames, Kingston-upon-Thames, Sutton, Merton and Wandsworth. One district (D) electing 9 seats (5:4) comprising the

London boroughs of Bexley, Bromley, Greenwich, Lewisham, Southwark, Croydon and Lambeth.

An alternative, which equalises the electorates more exactly and arguably makes each district more coherent, but involves splitting current Westminster seats, exists. It would transfer Hammersmith & Fulham from district C to district B, and Wandsworth from district C to district D.; this would give revised district B 10 seats, revised district C 7 seats and revised district D 10 seats.

North East (13)

One district electing 6 seats (3:3) comprising Cleveland and County Durham

One district electing 7 seats (4:3) comprising Tyne & Wear and Northumberland.

North West (35)

One district electing 10 seats (5:5) comprising Lancashire and Cumbria.

One district electing 6 seats (3:3) comprising the Greater Manchester boroughs of Rochdale, Oldham, Bury, Bolton and Wigan.

One district electing 9 seats (5:4) comprising the Greater Manchester boroughs of Salford, Manchester, Tameside, Stockport and Trafford, plus the Cheshire districts of Macclesfield, Congleton and Crewe & Nantwich.

One district electing 10 seats (5:5) comprising the remainder of Cheshire plus Merseyside.

South East (41)

One district electing 8 seats (4:4) comprising Kent.

One district electing 8 seats (4:4) comprising East Sussex and West Sussex.

One district electing 9 seats (5:4) comprising Hampshire and the Isle of Wight.

One district electing 7 seats (4:3) comprising Surrey and the Berkshire authorities of Bracknell, Windsor & Maidenhead and Slough.

One district electing 9 seats (5:4) comprising Buckinghamshire, Oxfordshire and the remainder of Berkshire.

South West (26)

One district electing 8 seats (4:4) comprising Devon and Cornwall.

One district electing 7 seats (4:3) comprising Dorset, Somerset and the Wiltshire authority of Salisbury.

One district electing 11 seats (6:5) comprising Avon, Gloucestershire and the remainder of Wiltshire.

West Midlands (27)

One district electing 11 seats (6:5) comprising Staffordshire and the West Midlands boroughs of Wolverhampton, Walsall, Dudley and Sandwell.

One district electing 10 seats (5:5) comprising Warwickshire and the West Midlands boroughs of Coventry, Solihull and Birmingham.

One district electing 6 seats (3:3) comprising Shropshire, Herefordshire and Worcestershire.

Yorkshire and the Humber (26)

One district electing 11 seats (6:5) comprising West Yorkshire.

One district electing 7 seats (4:3) comprising North Yorkshire and the Humberside boroughs of East Riding of Yorkshire and Hull.

One district electing 8 seats (4:4) comprising South Yorkshire and the Humberside boroughs of N Lincolnshire and NE Lincolnshire.

Scotland (26)

One district electing 7 seats (4:3) comprising the Scottish Parliament electoral regions of Glasgow and West of Scotland. One district electing 9 seats (5:4) comprising the Scottish Parliament electoral regions of Highlands & Islands, North East Scotland and Mid Scotland & Fife.

One district electing 10 seats (5:5) comprising the Scottish Parliament electoral regions of Lothians, South of Scotland and Central Scotland.

Wales (15)

One district electing 9 seats (5:4) comprising the Welsh Assembly electoral regions of South Wales West, South Wales Central and South Wales East.

One district electing 6 seats (3:3) comprising the Welsh Assembly electoral regions of North Wales and Mid & West Wales.

Northern Ireland (7)

Elects as a whole region (4:3).

This scheme therefore provides for STV elections of 3-7 members; 3-6 if the North West is divided into 4 rather than 3 districts. This would permit ‘normal’ STV to run effectively and without any of the problems that flow from large districts, although with some cost as regards proportionality.

In this scheme, most counties are maintained as wholes, and unitary authorities are attached to their ‘parent’ county. The exceptions are two non-administrative counties, Humberside (split along the line of the Humber) and Berkshire (eastern section hived off); and three full counties, Cheshire, Wiltshire and Nottinghamshire. Of these last three, only the division of Nottinghamshire is essential for the working of the scheme. Even though most district lines can follow county boundaries, the ones that do not may end up not being coterminous with Westminster constituencies.

Larger districts (5-8 elected at a time)

An alternative scheme would involve using larger STV constituencies, which would mean a more proportional

outcome at the expense of making the districts less cohesive. Geographical considerations mean that three seats (Northern Ireland and the two in Scotland) would be outside the 5-8 band.

Eastern (28)

One district electing 14 seats (7:7) comprising Norfolk, Suffolk, Cambridgeshire and Bedfordshire.

One district electing 14 seats (7:7) comprising Essex and Hertfordshire.

East Midlands (22)

One district electing 11 seats (6:5) comprising Northamptonshire, Leicestershire and Lincolnshire.

One district electing 11 seats (6:5) comprising Derbyshire and Nottinghamshire.

The theoretical entitlements for these sub-districts are

11.34 and 10.43 respectively, so taken on their own they would be awarded 11 and 10 seats, but in order to

preserve the region’s entitlement to 22 the number of seats in the Derbyshire-Nottinghamshire district has been rounded up to 11.

London (34)

One district electing 10 seats (5:5) comprising the boroughs of Havering, Barking & Dagenham, Redbridge, Waltham Forest, Newham,Tower Hamlets, Hackney, Islington, Haringey and Enfield.

One district electing 11 seats (6:5) comprising the boroughs of Barnet, Camden, Westminster, the City of London, Kensington & Chelsea, Hammersmith & Fulham, Brent, Harrow, Ealing, Hillingdon and Hounslow.

One district electing 13 seats (7:6) comprising the boroughs of Richmond, Kingston, Merton, Wandsworth, Sutton, Croydon, Lambeth, Southwark, Lewisham, Greenwich, Bexley and Bromley.

North East (13)

Elects as a whole region (7:6)

North West (35)

One district electing 10 seats (5:5) comprising Lancashire and Cumbria.

One district electing 13 seats (7:6) comprising Greater Manchester.

One district electing 12 seats (6:6) comprising Merseyside and Cheshire; this district’s entitlement is close to 13.

South East (41)

One district electing 12 seats (6:6) comprising Kent and East Sussex.

One district electing 13 seats (7:6) comprising Hampshire, Isle of Wight and West Sussex.

One district electing 16 seats (8:8) comprising Buckinghamshire, Oxfordshire, Berkshire and Surrey.

South West (26)

One district electing 12 seats (6:6) comprising Cornwall, Devon and Dorset.

One district electing 14 seats (7:7) comprising Gloucestershire, Avon, Somerset and Wiltshire.

West Midlands (27)

One district electing 13 seats (7:6) comprising the West Midlands county.

One district electing 15 seats (8:7) comprising Staffordshire, Shropshire, Warwickshire, Herefordshire and Worcestershire.

Yorkshire and the Humber (25)

One district electing 10 seats (5:5) comprising North Yorkshire, Humberside and the borough of Doncaster. One district electing 15 seats (8:7) comprising West Yorkshire and the boroughs of Sheffield, Rotherham and Barnsley.

Scotland (26)

One district electing 9 seats (5:4) comprising the Scottish Parliament electoral regions of Highlands & Islands, North East Scotland and Mid Scotland & Fife.

One district electing 17 seats (9:8) comprising the Scottish Parliament electoral regions of Lothians, South of Scotland, Central Scotland, Glasgow and West of Scotland.

Scotland is problematic when it comes to devising large districts based on Scottish Parliament electoral regions.

Wales (15)

Elects as a whole region (8:7)

Northern Ireland (7)

Elects as a whole region (4:3)

Read more publications...