Electoral Boundary Determination in Single Transferable Vote Electoral Systems:
the case of Ireland and Scotland















Prepared for

The British Columbia Electoral Boundaries Commission


October 2006





R. K. Carty

4422 West 6th Avenue

Vancouver, BC

V6R 1V3



Electoral Boundary Determination in Single Transferable Vote Electoral Systems:
the case of Ireland and Scotland





This report provides an account of electoral district boundary determination (and revision) in two Single Transferable Vote electoral systems – Ireland and Scotland. It includes a review of the instruments, process and criteria used in each country for this task.

The cases provide two distinct perspectives on the challenges facing STV constituency map-makers. In the Irish case the boundary delineation process is about revising and updating an existing map to recognize the realities of a growing and shifting population. The Scottish case illustrates how a political system has responded to the challenge of creating STV multi-member electoral districts for the first time after the system had been newly adopted.

The report proceeds by considering each case in turn focusing on the context, institutional structure and process of each. A concluding comparison is instructive for there are more similarities than differences between them. Each uses independent commissions that operate with (by Canadian standards) remarkably little input from the public, work to broadly similar criteria, and make authoritative decisions.  

The report draws on a review of the reports of boundary commissions in Ireland and Scotland, the legislative and constitutional framework which orders their existence and work, and from a series of personal, confidential interviews with commissioners and staff in both countries in May of 2006. The observations reported and conclusions drawn are solely the responsibility of the author and should not be attributed to any informant. 

I     Historical Context

Ireland adopted the Single Transferable Vote (STV) after winning independence from Britain after the First World War and has used it for all its general elections to Dáil Éireann, the lower house of its national parliament.[1] A shifting and growing population has led to periodic revisions of the electoral districts and there have been a total of 12 changes since the original STV constituency map was introduced in 1923. Table 1 indicates the size of the house and the organization of constituencies (by district magnitude – the term used to indicate the number of members elected in an individual electoral district) over time. Several features of this record stand out.

     Table 1                        Dáil Éireann Constituency Structure




District magnitude




Number of Districts

Size of Dáil

Population /
Dáil member













































































































































                * The 1959 redistribution was found unconstitutional by the Court; the 1980 redistribution was the
first done by an independent (ad hoc) commission; the 1998 redistribution was the first done by a
statutory commission.

                  Based on the population recorded at the census immediately preceding the seat distribution

 First, the size of the Dáil has generally been relatively stable. For about the first 60 years it did not move much but then jumped by about 20 in 1980 to 166 where it has remained since. This suggests that, except for one occasion when a major adjustment was made, the boundary process has not used the expedient of increasing the size of the house to deal with major population shifts. In fact the most significant changes in the Irish population have been internal. The total population was in continuous decline until the mid-1960s when it turned up for the first time in a century. It then rose through the 1970s but fell again through the 1980s only to turn up again after 1991.[2] At the same time there appears to have been a steady drift to the east of the country. Dublin County, for instance, grew steadily – from 17% of the population in 1926 to 25.5% in 1961 and 30.3% in 2002.[3] Boundary commissions have dealt with this movement by continually restructuring the distribution of seats of different magnitudes within the constitutional and legal constraints governing their decision-making. 

Second, the total number of constituencies has also remained relatively stable since the boundary revision of 1947. That said, it is also apparent that the arrangement of constituencies in terms of their district magnitude has changed in most revisions (except in 1983 and again in 1995) as distinctly different combinations of 3, 4 and 5-seaters have been used. This suggests that a reorganization of the arrangement of constituencies by magnitude has been the principal tool of the boundary revision process.

Third, boundary revisions have become far more frequent in the past two decades: there were five in the first fifty years after 1923 but there have been six in the thirty since 1974. Thus boundary revisions are occurring twice as frequently and this must have consequences for the organization of the electoral machinery and for the stability and shape of local political party organizations through which individual citizens participate in the electoral process.   

Fourth, large constituencies – those returning more than 5 members – were used in the first two STV electoral maps but then quickly discarded. Since 1947 all districts have returned either 3, 4 or 5 members.

Fifth, a revision conducted in 1959 was found to be unconstitutional by the High Court. This is a reminder that the process in Ireland, though then instituted by the government of the day and accepted by the elected members of the Dáil, is subject to judicial review and must meet constitutional standards as identified by the courts.

Sixth, the revision of 1998 was the first conducted by a statutory commission as provided for in the Electoral Act (1997). This completed the process of removing the boundary revision process from the direction of the government.

Gerrymandering under STV is not unknown and there is reason to believe that in the early decades governments sought to manipulate the constituency map to their advantage. Under First-Past-the-Post systems gerrymandering is a matter of drawing constituency boundaries to partisan advantage (maximizing one’s own vote-seat ratio while forcing opponents to waste votes). The proportional character of STV ensures that parties can not be so dramatically punished by the electoral map but governments will be tempted to order district magnitudes to their advantage. For instance, a party winning 50% of the vote in a given area could expect 2 of 3 seats in a 3 seat district but only 2 of 4 in a 4 seat constituency – it is thus likely to prefer a 3-seat district in that case. Similar calculations can be made for other regional vote shares and from that a political party is able to devise a preferred constituency map. But as in single-member district gerrymandering exercises this is an inherently risky business for any adverse turn in popular opinion is likely to cost the would-be gerrymanderers more heavily than their opponents. This appears to have been the case in 1974 when a coalition government redrew the electoral map in ways it anticipated would advantage itself. (This involved a decrease in the number of 4-seaters and a corresponding increase in 5-seaters as well as a major redrawing of the Dublin city constituencies.) The exercise, commonly called a ‘Tullymander’ after the minister (James Tully) responsible, proved to be a political disaster and the government was swept from office at the next (1977) election. It also proved so controversial that subsequent constituency revisions have been conducted by independent commissions, a process that was institutionalized by statute in the Electoral Act of 1997.


II    Provision for Electoral Boundary Revisions

The basis for redrawing the electoral map is multi-dimensional and is rooted in the Constitution of the country (Bunreacht na hEireann), subject to the standards set out by the High Court in its interpretation of the relevant sections, and the more precise provisions of the Electoral Act,1997.

The constitution sets the limits to the size of the Dáil (requiring one member for between a minimum of 20,000 and a maximum of 30,000 in the population), indicates that there must be a redistribution at least once every twelve years, and directs that the ratio of population to Members (known as Teachta Dála or TDs) “shall, so far as it is practicable, be the same throughout the country”. While the courts have dealt with the implicit equality provision in a fashion that provides some latitude for boundary commissions, it has made it clear that this equality provision is a dominant principle, qualified only by the somewhat slippery notion of practicability. Unlike courts in other countries, the Irish High Court has not set a numerical definition that would define the limits of “practicability”. However, recent commissions appear to have informally (but explicitly) defined it in practice.[4] Constituencies that do not vary from the national average by more than about 7 to 8% are generally deemed to be acceptable, that is not requiring revision, but when changes do need to be made then the target is tighter, with a variance of 5% being the norm.

These general constitutional provisions have been considerably tightened by the provisions to the 1997 Electoral Act that provides for a Constituency Commission. The statute requires that an independent commission be established after every census, now held on five-year intervals, which explains the frequency of recent redistributions. The Act is also more specific in its direction to the commission than is the Constitution. It provides that:

a)      The total number of members is to be between 164 and 168 (a number considerably different from the more expansive constitutional provision which would allow for between 130 and 196)

b)      Constituencies shall return 3, 4 or 5 members

c)      Breaching county boundaries is to be avoided as far as possible

d)      Constituencies are to be composed of contiguous areas

e)      Regard shall be given to physical features and to the ‘extent and density’ of population

f)        Efforts shall be made to maintain continuity in the constituency map

These directions constrain the commission and set the terms on which any revised constituency map will be drawn.

The Act provides for a commission to implement these provisions that is composed of five members – a judge nominated by the Chief Justice (who serves as chair) and four others, appointed by virtue of other offices held, including the Clerks of the Dáil and the Seanad, the Ombudsman and the Secretary-General of the Department of the Environment, Heritage and Local Government. The latter is the government department with the technical expertise utilized by the Commission in its work so its Head provides an organizational link to necessary staff support. The actual work of analyzing the population data provided by the Irish Central Statistics Office is done by a small group in the franchise section of the Department who provide constituency boundary mapping support for the Commissioners. The conduct of Irish elections is very decentralized (it is left in the hands of individual constituency returning officers) so there is no connection between the authorities responsible for boundary revision and those who subsequently manage the electoral process in those constituencies. 

Finally, it is important to note that the report of the Commission is, in law, an advisory document. When it finishes its work its report is presented to the Chairman of the Dáil and when it does so the Commission is automatically dissolved. The report then goes to the Dáil and is subjected to the normal legislative process to formally constitute the new constituencies. In practice, all agree that the report is a definitive document and it is fair to say that there is now a constitutional norm that politicians do not alter any of the recommendations as they proceed to legislation.


III   Redrawing the Irish STV Map

 Irish electoral boundary commissioners work to a number of well defined criteria that clearly structures their activity. They agree that the equality provision is paramount but that it has to be interpreted in terms of two other important criteria – a desire to respect county boundaries and to minimize change (items c and f above). This means that STV map-making in Ireland is an incremental process. With minimizing change held as a primary value, redistribution begins with the assumption that the map will be left as is unless the equality provision (noted above) has been breached. Where population growth and shifts have led to greater than acceptable variances, commissioners work to rebalance the affected constituencies with as little disruption to the rest of the constituency map as possible. The second value, that of adhering to the country’s historic (26) county boundaries wherever possible, is a second, conservative force. It leads commissioners to prefer to alter district magnitudes rather than existing boundaries. Thus, if an existing constituency has suffered a loss of population the commissioners will first look to explore the possibility of shrinking the number of members returned from the constituency rather than shrinking the physical size of it if the latter would entail abandoning a county boundary.[5]

It appears then that of the two techniques open to STV boundary makers – changing the actual physical boundaries OR altering the district magnitude of a constituency – the mandate to favour continuity and respect county boundaries means that altering a constituency’s district magnitude is the preferred option in Ireland. When boundaries do need to be moved a critical issue is the scale of the change and one commissioner suggested to me that, as a rule of thumb, numbers equal to about the total of one quota[6] are an appropriate unit to work with for in principle they would allow an area moved to a new constituency (and thus often linked to a different county) to elect their “own” member in the new district.

As Table 1 indicates, there has been no consistent pattern in the preference given to smaller or larger district magnitudes over time. The most recent electoral map has a higher proportion of 3-seaters than in the 1980s and 1990s but a distinctly smaller proportion than in the 1960s and 1970s. It may be that this most recent increase in the proportion of 3-seat districts is a response to the fragmentation of the party system, and the instability of Irish governments during the 1980s, for theoretically one would expect more proportional results when the constituency map has a higher proportion of electoral districts with larger magnitudes. However electoral outcomes also depend upon the number of candidates and parties contesting an election and the geographic distribution of their vote shares. In the period since 1989 the share of the vote going to independents in Ireland has nearly tripled and they won 14 seats (8.4% of the total) in the most recent (2002) general election. Five of them were elected in 3-seat constituencies so it appears that smaller district magnitudes have not worked to their great disadvantage as compared to the candidates of the large parties.

Commissioners and staff both indicated that they believed that the formal requirement that limited district magnitudes to 3, 4 or 5 was a good one. It provided a reasonable range of flexibility in creating districts while not presenting them with a set of alternatives that was so large as to be unmanageable. It also provided a framework that did not lead to large alterations in the map (the continuity principle –f  above) that would be disruptive to the political parties that had to organize the nomination of candidates and conduct elections. Their view was that these magnitudes provided for very desirable levels of proportionality and, as the experience of Independents in recent years indicated, keeps the process open to all interests.  

  The actual process of drawing and redrawing the constituency map is best described as a ‘top-down’ process. The Constituency Commission starts by identifying problems – that is districts that violate the equality principle by having a Member-population ratio that exceeded 5% from the national average. The Commission then asks its staff to explore options to rectify these discrepancies within the limits imposed by the other constraints on their discretion (i.e. the factors, listed a-f above, specified in the Act). Where this is particularly difficult to do they are prepared to tolerate continuing variances up to about 8%, anything larger is not acceptable and will force a change even if it means trumping other imperatives such as maintaining a county boundary.[7] The staff work supporting analyses and assessments of alternate electoral maps is done with GIS software using very long-standing electoral divisions (there are between 3 and 4 thousand in the country) as the basic building blocks for the constituencies.

 The process is also reasonably ‘closed’ in that the Commission does not hold any public hearings on a current or proposed map. It does invite written submissions, by means of one notice in the daily newspapers, from the public and it specifically solicits submissions from interested parties – members of the Dáil, Seanad and European Parliament, the registered political parties and the Dáil constituency returning officers. The most recent redistribution exercise (2004) generated 99 submissions. It appears that the largest proportion of these submissions are: a) focused narrowly on local issues, b) rarely integrated with submissions from others –including those from others in the same national party, and c) ignore any ripple effect on adjacent constituencies proposed changes might make. While the Commission takes note of all submissions it appears that they find them of very limited use in its work.

Public hearings have never been used in the process. The general view appears to be that this protects the Commission from undue influence and certainly any appearance of responding to or even giving in to political pressures. In this sense the lack of any public hearing dimension to the process is held to strengthen the independence of the Commission. At the time of the last redistribution (2004) the Commission established a website to, in its words “facilitate public access to material relating to its remit” but the site was not used to solicit or receive submissions. However it may have helped those who wished to prepare a submission for it recorded 626 visits in a five month period.


IV  Summary

The Irish process of drawing STV maps for elections to Dáil Éireann is essentially a conservative, top-down, process governed by both constitutional and statutory provisions which set clear expectations that favour minimizing change. The independent Commission works largely in private, makes incremental changes to correct population discrepancies and, where possible, appears to prefer altering district magnitudes rather than redrawing boundaries. When boundaries must be redrawn, the aim of not disrupting established patterns of political representation suggests that moving populations in terms of full electoral quotas serves as a rough rule of thumb. There is no appetite for expanding the range of district magnitudes in use and experience suggests that creating constituencies returning from 3 to 5 members has provided for broadly proportional outcomes that satisfy parties and independents alike. The balance among constituencies of different size varies over time as successive commissions seek to juggle the competing imperatives of their statutory mandate.           








I     Context

Citizens in few democracies face the variety of institutional arrangements for electing their representatives and governments that the Scots do. As a result of the constitutional evolution of that part of the United Kingdom, voters in Scotland now elect representatives to their Local Government Councils, the Scottish Parliament, the UK House of Commons, and the European Parliament. The mix of political parties that effectively contest these elections differs as does the time at which voters typically go to the polls. But the most striking feature of contemporary electoral politics in Scotland is that elections to each of these different representative bodies are carried out under a quite distinct and different electoral system.

 These provisions are summarized in Table 2.

Table 2        Electoral Systems in Scotland, 2006

Political Level

Electoral System

No. Electoral Units

No Elected Representatives

Local Authorities

Single Transferable Vote

32 Local Authorities
300-400 multi-member

~ 1222 Councilors

Scottish Parliament

Mixed-Member Proportional

73 single-member districts 8  7-MSP regions

129 MSPs

Scottish seats at Westminster

Single-Member Plurality

59 single-member districts

59 MPs

European Parliament

Regional Party Lists (d’Hondt)

7 regions

7 MEPs


 Elections to the House of Commons at Westminster are carried out under the old, familiar single-member plurality system; those to the other levels are comparatively new. The issue of whether the proliferation of systems has led to voter confusion was the subject of the recent Commission on Boundary Differences and Voting Systems (commonly referred to as the Arbuthnott Commission after its chair) and it has recommended that, for the time being, the present arrangements should be maintained.[8]

So new is this institutional structure that the country has yet to hold its first local authority STV elections. That will happen in the spring of 2007 on the same day voters go to the polls to elect a new Scottish Parliament (under the MMP system – referred to locally as the Additional Member System or AMS). This means that Scotland is in the process of creating its first set of STV electoral maps. This stands in contrast to the Irish case where STV map-making has, for some decades now, been a question of revising an existing map.

It should be noted that responsibility for creating electoral maps in Scotland falls to different authorities. There is a Boundary Commission for Scotland (provided for in the Parliamentary Constituency Act, 1986) that is responsible for establishing constituency boundaries for elections to the parliaments in London and Edinburgh. A quite separate body, the Local Government Boundary Commission for Scotland (LGBCS) is charged with creating maps for elections to local government authorities and it is that commission which has been engaged in establishing STV constituencies over the past year and whose experience is reported here.

II    Provision for Electoral Boundary Determinations

The Local Boundary Commission for Scotland is an independent body created by Section 12 of the Local Government (Scotland) Act 1973. It is fully funded by the Scottish Executive and serves in an advisory capacity. Despite the fact that it merely recommends boundaries, and does not have the power to legally set them, the working assumption of those involved is that the recommendations will be accepted by the relevant minister. Thus, for all practical purposes the Commission can be said to determine the boundaries.

The Local Governance (Scotland) Act 2004 provides for the use of the Single Transferable Vote form of proportional representation for local authority elections. In creating electoral districts for this system the legislation stipulates that the following conditions should govern map-making within the existing local authority areas:

     a) electoral parity (that is the number of councilors per elector should be “as nearly
                 the same as may be”)
            b) concern should be given to maintaining local ties that define community
            c) electoral districts to return 3 or 4 members

This last condition was to have a major impact on the work of the commission as it sought to create the initial STV districts.

The Commission is made up of six citizens (one of whom is designated as chair) appointed by the government to fixed terms. In practice, when vacancies occur they are advertised and individuals are free to apply indicating their interest and expertise. The current commission includes individuals with long experience in local government, academic experts (in geography, economics and land use planning) and a former college principal. They are supported by a small staff of technical experts headed by a professional public servant. 

II   Drawing a Scottish STV Map

The Scottish Local Boundary Commission started by considering whether the new multi-member districts required by STV could be created simply by merging existing single-member wards within the various local authorities. They soon came to the conclusion that there were too many anomalies in the old maps and so a completely fresh start would have to be made in each local government area. Given that there were some 1222 single-member wards in the 32 authorities, this meant that they would be creating between 300 and 400 new STV districts. In effect this involved creating 32 new and distinct electoral maps with a varied number of constituencies (whose district magnitudes were to be 3 or 4) in each and the commission proceeded to do so local authority by local authority.

The process in each local authority followed the same legally prescribed pattern. It can be summarized in point form as follows:

1)                The LCGBS informs the relevant local authority Council of its intention to start the boundary delineation process; identifies a proposed timetable (governed in part by government instruction); and seeks a meeting with the Council

2)                At an initial meeting the Commission explains to the Council how it intends to approach the task and asks for the Council’s views on the particular issues of its local area

3)                The Commission considers this advice and creates a set of initial proposals for electoral districts

4)                The Council is given the opportunity to respond to the initial proposals

5)                The Commission considers the Council’s response and publishes a set of provisional proposals

6)                The public is invited to make written representations with regard to the provisional proposals during a 12 week period

7)                Final recommendations are made by the Commission

This is a process that involves intimate and continuing conversation between the Commission (staff) and the officials of the local governments. Although interested public official and political parties are informed, there is very little opportunity or expectation of wider public involvement. The Commission reports receiving few public representations (it doesn’t hold public hearings on its proposals at any stage) and indicates that most that are offered refer to one or two obvious issues of particularly local concern.

At the heart of their decisions is the principle of parity. In practice the Commission informally defines this as keeping districts (on the same map) to within a 5% range in the ratio of electors to representatives. Within that framework there is a concern to maintain the local ties of community and association that have been established in the past and continue to structure public communication and activity. However the Commission is quite restricted in establishing electoral districts by the requirement that all must return either 3 or 4 members. Staff report that, in a number of instances, this resulted in far from optimal arrangements being made for the organization of constituencies. Their suggestion was that the possibility of creating 5 member districts would introduce a much needed flexibility: in several cases a 3 and a 5 member set of districts would have made much more local and geographic sense than two 4s. When asked if district magnitudes of even larger size would have been an aid in establishing these new STV maps their considered view was no – they believe that a range of 3 to 5 would be sufficient for their purposes.

The local boundary commission did all of its mapping with highly sophisticated GIS systems that took advantage of large-scale Ordinance Survey digital maps that had been matched to electoral register data. This allowed for precise electronic mapping and the creation of formal descriptions that are more accurate than in previous constituency mapping exercises.


IV  Summary

The Scots have moved quickly and effectively to create 32 new STV electoral maps. The process is a highly structured one that involves little public input (except as it may be understood to be expressed through the elected Councils). The decision to limit the district magnitudes of the constituencies on these maps to 3 or 4 appears to have made (in the minds of the map-makers at least) the process more difficult than necessary. However it is also fair to note that until the country holds its first set of elections under this system and with these maps it may be too soon to assess the extent to which that limitation constrains the wider

political process. 
Comparative Observations


Despite the considerable difference in the situation facing contemporary STV constituency map-makers in Ireland –where the task is revising an established map– and Scotland –where their task is to create maps de novo– there are a number of striking similarities in the approach taken.


On process mechanisms, both systems can be fairly described as elite-managed and top-down with little public involvement. Though both processes are formally advisory only, in practice both produce authoritative outputs. The public and political class accepts their independence which legitimates the process and forestalling challenges to it. The widely accepted independence of the respective commissions, made up of individuals appointed for their positions (Ireland) or personal qualities and experience (Scotland), allows them to use government public servants as staff support. In the case of Scotland the magnitude of the task of creating so many individual STV maps has led officials being seconded to the Commission for an extended period. In both countries new GIS technologies have made the actual consideration of alternate boundary configurations much easier and more efficient than in the past.


On criteria, both countries adopt a fairly strict definition of parity (± 5% of the average district elector to representative value) as the central principle. In Ireland this has been given constitutional weight by High Court judgments, in Scotland it has simply been adopted as the working standard. Both countries also adhere to a principle of respecting local community ties as much as possible. In Ireland this has been reinforced by the requirement that the historic county boundaries be respected. The smaller scale of the local authorities in Scotland contributes to the same impulse. The practical reason for this is to sustain the stability of local organizations, especially political parties which are the instruments of citizen electoral involvement and which must operate the competitive system, as much as possible. Finally it is noteworthy that both countries use a narrow range of district magnitudes. In Ireland the early use of larger districts has been abandoned, in Scotland an even narrower range has been adopted for a new system. Neither Commissioners, staff nor the wider publics appear to believe that districts larger than five are either necessary or appropriate.  


[1] The Single Transferable Vote is also used for elections to the Seanad (the upper house) but elections to it are fought on a very restricted franchise in vocationally defined constituencies. STV is also used to return Irish members to the European Parliament and the same process is used to draw European Parliament districts as for Dáil Éireann. In this report I focus on an account of the more extensive Dáil constituency revision process.  

[2] The population per member in 1961 appears to have been just within the constitutional limit of 20,000. Had the calculation been made on the 1961 census figure it would have fallen below – to 19,571. Much of the change in population is driven by dramatically shifting emigration and immigration rates.

[3]  This undoubtedly underestimates the recent concentration of the population in the Dublin conurbation for much of the capital’s effective population growth has spilled into neighbouring counties.

[4] The Constituency Commission’s Report on Dáil Constituencies, 2004 discusses these variances in Sect. 3.1and 3.4.

[5] Another option is to put two counties together in a single constituency. Such an amalgamation does not violate the principle of maintaining the integrity of the county boundaries themselves.

[6] One quota being the number of votes (on any count) sufficient to elect a Member.

[7] The 2004 report illustrates this. See the discussion (Sect 4.5) justifying the maintenance of the Cavan-Monaghan constituency at –7.50%.  Note that it concludes with a comment indicating that should population trends continue that there would likely have to be a seat reallocation at a subsequent redistribution.

[8] The commission’s report Putting Citizens First: Boundaries, Voting and Representation in Scotland was published in January 2006 at The Stationary Office, Edinburgh.