1. This paper is provided in response to a request from Cabinet on the use of constituencies and the Single Transferable Vote (STV) electoral system for district health board (DHB) elections [CAB (00) M 19/14 refers]. It requests that Ministers indicate which electoral system should apply to DHB elections in 2001. It should be noted that alternate electoral systems other than STV and constituencies are not considered in this paper.

EXECUTIVE SUMMARY

2. Ministers have asked officials to report back on STV and constituency based approaches for DHB elections in 2001. As Table One shows, none of the options for DHB decisions in 2001 considered in this paper can guarantee that all of the desired objectives for DHB board membership and function are achieved.

3. As Table One shows, none of the options for DHB decisions in 2001 considered in this paper can guarantee that all of the desired objectives for DHB board membership and function are achieved. The voting system which is chosen will involve making trade-offs between community representation, the function of the board and the cost and risk of implementing a new system. Any electoral system design can also only respond to the candidates who stand and to the people who 'turn out' to cast their votes.

4. In summary:
   1. STV is likely to be the best system for promoting minority representation, but there are significant costs and risks involved with implementing this option for 2001
   2. the addition of a constituency system to STV when there are only seven elected members across the whole district takes away almost all of the benefits of STV, and officials recommend against this option
   3. a constituency system can be implemented relatively easily, and will achieve a measure of geographic representation, but makes it harder for minority candidates, including Maori, to get elected, and risks factionalising the board
   4. while an at large election using the current local government voting method is the easiest to implement and presents the least risk, it cannot guarantee geographical or minority representation

5. Ministers are asked to note the associated costs and risks and indicate which type of electoral system is desired for DHB elections in 2001.

BACKGROUND

6. At its meeting on 12 June 2000, Cabinet noted its initial preference for a constituency system for the DHB electoral process which may include either single or multiple member constituencies, and, where there are multiple member constituencies, the Single Transferable Vote system would apply [CAB (00) M 19/14 refers].

7. Cabinet directed the Ministry of Health to liaise with both the Department of Internal Affairs (Local Government) and the Ministry of Agriculture and Forestry (Rural Affairs) officials and provide advice as soon as possible to the Cabinet Social Policy and Health Committee on a constituency system and STV [CAB (00) M 19/14 refers].

8. This paper provides advice on the implementation of, and choice of, STV and constituencies, and identifies the risks involved with using these approaches, for the first DHB elections in October 2001.

DHB and local government elections

9. Seven people will be elected to each DHB board and up to four people will be appointed [CAB (00) M 19/14 refers]. DHB elections will be held concurrently with local government elections [CAB (00) M 11/1A(3) refers] and the first elections will take place in October 2001 [CAB (00) M 2/4 refers].

10. Local government elections are those which are held for the 74 territorial authorities (district councils and city councils), the 12 regional councils and other local bodies such as licensing trusts and community boards. Most territorial authorities and all regional councils operate under a constituency system (constituencies are known as 'wards' in territorial authorities). Under the current voting method, voters are allowed as many votes as there are members to elect in their ward or constituency (or in the district as a whole if voting is at large).

11. Territorial authority elections, regional council elections and DHB elections will be separate elections but will share the same electoral registration data. Elections will be held by a postal vote, where voting papers for all elections will be part of the same package sent out to voters.

12. Each DHB will have to employ a supervising returning officer to run its election. DHBs will meet the cost of their own elections [CAB (00) M 11/1A(3) refers] as do territorial authorities and regional councils.

CONSTITUENCIES FOR DHBS

13. Having constituencies is a way of ensuring the representation of geographical communities of interest. In particular, in large, widespread districts it ensures a degree of representation from rural areas, which the Ministry of Agriculture and Forestry (Rural Affairs) supports, as consistent with Government's rural affairs objectives.

14. Draft DHB constituencies are attached as Appendix One. These differ from the constituencies provided as part of SPH (00) 64 as they now reflect Cabinet's decision that there will be seven elected DHB members in each district.

15. Constituencies have been constructed taking into account:
    1. geographical communities of interest, as defined by territorial authority and ward boundaries (wards are constructed to provide effective representation of communities of interest within the district under s101L of the Local Government Act 1974)
    2. consistency with how the overall DHB boundaries were defined, which was by territorial authority boundaries and wards [CAB (00) M 19/14 refers]
    3. that constituencies should be geographically continuous and not be 'artificially shaped' as it is not sensible to have constituencies which comprise a number of separate geographical areas
    4. the fact that the total number of members across all a district's constituencies should always add up to seven
    5. the population of the constituency, and therefore their proportional 'share' of the seven elected DHB members
    6. the physical size and remoteness of the constituency as it may be difficult for one DHB member to represent a very large geographical area
    7. a reasonable spread of members across the district, so that one constituency, for example, does not have six members and another has only one as this could cause lack of balance in constituency representation on the DHB Board.

16. The constituencies defined in Appendix One are a mixture of single and multiple member constituencies. The majority are multiple member constituencies of two or three members. The maximum number of elected members in one constituency is five. Most of the main cities have been split into separate constituencies, but only in Auckland and Manukau Cities was it possible to take the socio-economic status of different suburbs into account.

17. Within each DHB, the number of people represented by elected DHB members can vary significantly between constituencies (see Appendix One). For example, in the Canterbury DHB the rural Kaikoura/North Canterbury constituency (defined by the Kaikoura, Hurunui and Waimakariri Districts) has one elected member for a population of 45,000. This contrasts with the two urban Christchurch constituencies, where each elected member represents 60,000 or more people. Such differences are a reflection of the need to have exactly seven elected members from across the district.

18. The ward structure of territorial authorities can change every three years, and can even disappear if the authorities choose to elect their members at large. The ward structure on which Appendix One is based is subject to review over the next nine months. A process would need to be put in place to ensure that DHB constituencies were reviewed as well, taking into account any territorial authority or population changes.

IMPLEMENTATION OF A CONSTITUENCY SYSTEM

19. Cabinet has agreed that DHB boundaries be included in a schedule to the New Zealand Public Health and Disability Act with provision for them to be amended by Order in Council [CAB (00) M 19/14 refers]. DHB constituencies could therefore also be included in this schedule.

20. Having DHB constituencies makes little difference to the complexities of the electoral process and has no significant time or cost implications.

RISKS OF A CONSTITUENCY SYSTEM

21. Cabinet has noted previous advice that having DHB districts divided into constituencies for voting purposes may result in boards becoming factionalised and operating in a manner which is inconsistent with the overarching intent that DHB boards assume collective responsibility to the Minister of Health for the health services and disability support needs of the total DHB population [CAB (00) M 19/14 refers].

22. A constituency system can only deliver representation of geographic communities. It cannot ensure the representation of communities which cannot be defined geographically, like Mäori and Pacific Island people. In fact, a constituency system can make it harder for minority groups to gain representation on boards (although some constituencies will have a high concentration of Mäori and Pacific people). This is because only a small number of members, or only one member, will be elected in each constituency which reduces the likelihood of minority candidates being elected.

23. A constituency system together with STV means that the number of members elected in each constituency will be too few for the STV system to be effective (this risk is discussed further below).

SINGLE TRANSFERABLE VOTE

24. Ministers have expressed a preference for the Single Transferable Vote (STV) system to apply for DHB elections where there are multiple member constituencies [CAB (00) M 19/14 refers]. STV can also be used in a one member constituency.

25. STV produces proportionality by the criteria voters choose to prioritise, for example, whether they vote for candidates on the basis of gender, age, ethnicity, rural issues and so on. It is a good system for promoting the representation of minority communities of interest, to the extent that those communities of interest are what determine voters preferences (and to the extent that they can attract the required number of votes). The mathematics of elections shows that STV makes it harder than with the current local government method for majorities to have a disproportionate influence on election results.

26. A description of STV is attached as Appendix Two, but it should be noted that this describes one form of STV and that other systems vary in some respects. STV is used in some public elections overseas (for example, in Ireland, Australia and Malta) but has not been used for public elections in New Zealand since the 1930s.

STV TOGETHER WITH CONSTITUENCIES

27. STV delivers more accurately proportional representation as the number of members to be elected in each electoral district increases. Therefore STV elections tend to be organised so that between three and seven members are elected in each district. DHBs have only seven elected members, however, and dividing the DHB into constituencies means that many constituencies will have only one or two members representing them. Around two-thirds of the constituencies defined in Appendix One have only one or two members.

28. The strength of STV is that it generates a result which is more representative of the different preferences of voters, and makes it easier for minority groups to get their representatives elected than the voting method currently used for local government elections. In an STV election to elect seven members, for example, a candidate would need to get just over 12.5 percent of the votes, including votes transferred to them from other candidates, to get elected. If around this percentage of voters gave their highest preferences to a particular group of candidates (e.g. rural candidates, Maori candidates, or candidates standing on a particular 'ticket') there is every likelihood that one of those candidates would be elected.

29. If STV elections were held within constituencies, most elections would be aiming to elect one or two members. This low number of vacancies significantly reduces the power of STV to elect minority groups. This is because, in the two member example, a candidate would need to get just over 33.3 percent of the votes to get elected. In an election of only one member a candidate would require just over 50 percent of the votes. These requirements would make it much harder for minority groups to get one of their favoured candidates elected.

30. Officials consider that the addition of a constituency system to STV when there are only seven elected members across the whole district takes away almost all of the benefits of STV. However, STV used in an at large system has the potential to achieve the objective of constituencies - that is, representation of geographical communities of interest (to the extent that those communities of interest are what determine voters preferences).

31. Advice on equitable representation of Mäori [in SPH (00) 83] has included that adding a constituency system to STV will weaken its ability to achieve equitable representation.

LOCAL GOVERNMENT AND STV

32. The Department of Internal Affairs is conducting a review of all local government electoral legislation, and the results of this review will be reflected in the upcoming Local Government Electoral Bill which has priority three on the legislative timetable (to be passed in 200 if possible).

33. Cabinet has directed Internal Affairs to include in this Bill provision for local authorities or communities to choose STV as a voting system [CAB (00) M 18/8 refers]. How this option would is exercised in a district has yet to be determined (it could, for example, be determined by a poll of electors or by a decision of the local authority).

34. However it is not intended that STV will be an option for the local government elections in 2001 [CAB (00) M 18/8 refers] because of a lack of time to adjust electoral boundaries to suit this option, to consult communities and implement it. Although STV is proposed as an option in 2004, it is uncertain how many, if any, local authorities would use it in that year. Under current proposals, a local authority which took up the STV option would be responsible for all the set-up and implementation costs of running the system although there is clearly potential for savings through co-operation.

35. A Members Bill from Rod Donald MP - the Local Elections (Single Transferable Vote Option) Bill - is currently before the House. This also seeks to have STV as an option for local government elections. However it is unlikely to alter the timing or the likelihood of local authorities adopting STV.

PREPARING AND RUNNING STV AS AN ELECTORAL SYSTEM FOR DHBs IN 2001

36. In order to run a DHB election using STV in 2001 the following establishment and implementation work would need to be done:
    1. policy and legal work to decide on and specify the exact details of the STV system to be used and consequent changes to the electoral system in the New Zealand Public Health and Disability Bill
    2. publicity and voter education
    3. design, development and testing of systems to read and count votes in an auditable manner.

37. The responsibility for this work would rest with the Ministry of Health. It is not feasible to rely on Internal Affairs to have STV legislation passed in time for a 2001 DHB election. The development of systems, publicity, etc., is also not the responsibility of Internal Affairs, since this is not their role, even for local government elections. This development should also be centrally managed, and therefore is not appropriate to be left to individual DHBs.

Specifying the voting system

38. Significant policy work would need to be undertaken in order to use STV for DHB elections in 2001. The details of the voting method would need to replace the proposed schedule to the New Zealand Public Health and Disability Bill dealing with elections (amendable by Order in Council) or dealt with in regulations.

39. Internal Affairs have not yet undertaken any work on the details of the STV option for 2004. Local government STV provisions at this level of detail may well be included in regulations and therefore not be required for the introduction of the Local Government Electoral Bill. In any case, there is no guarantee that this Bill will be passed in time for use in DHB elections in 2001.

40. Rod Donald's Member's Bill on STV is currently before the House but represents ad hoc amendments to current local government legislation and was formulated and promoted before the Government commitment to STV as a future option was announced.

41. Policy work on STV would consist of the following stages:
    1. research and policy development on the exact specification of STV to be used for DHB elections, and how different eventualities are dealt with (for example, what types of votes will be accepted, rules for transferring votes, specification and handling of 'spoilt' ballot papers)
    2. consultation with central government agencies, local government, district health boards, Maori, and other community groups
    3. Cabinet agreement to the policies around STV
    4. regulations, or a schedule amendable by Order in Counsel, drafted by the Parliamentary Counsel Office
    5. making of the Order in Council.

42. This is a substantial piece of policy work which could be completed in time for DHB elections in 2001. Timeframes would be very tight, however, and there would be little, if any flexibility to respond to unforeseen policy issues.

43. A major potential difficulty is that similar issues will need to be addressed as part of the review of local government electoral legislation, but at a later stage and in the broader context of that review. Decisions made now in respect of DHB elections in 2001 should not limit that later policy consideration, but the alternatives of two slightly different systems or a subsequent adjustment to DHB processes are equally unattractive. Work on local government STV issues cannot be brought forward to meet the DHB timetable without major implications for that portfolio's work programme.

Publicity and voter education

44. Nationally-organised publicity would be required for DHB elections regardless of the electoral system adopted. For an STV election, however, publicity and voter education will be crucial as STV is a significantly different and more complex voting method than the current local government (or general election) voting method. Votes are cast in a different way, and understanding how individual votes translate into election results is extremely complicated (see Appendix Two).

45. At the very least, publicity for DHB elections should include a newspaper advertising campaign covering the whole country, a web-site and a free-phone line (0800 number) for further advice. The best publicity would be generated by television advertisements, but effective advertising in this medium is very expensive.

Systems to read and count votes

46. Modern STV votes need to be counted by computer but there is no 'off the shelf' computer package available to use. STV software in New Zealand is only for demonstrating the voting method on personal computers. Before an STV election in 2001, software would have to be written to capture the exact details of the voting system and extensively tested. Such development would need to be out-sourced.

47. Reliance on computer technology to produce electoral results also raises policy issues concerning the extent to which such results can be audited and reviewed by the Courts in the event of disputes. Privacy considerations will also be important. These are significant issues at both policy and software development level which need to be resolved, in part at least before regulations are finalised, and completed well before the elections are conducted.

48. Officials consider that, in an STV election for DHBs, for reasons of efficiency and economy, the technology and expertise required for counting votes should be centralised. This means that all voting papers for concurrently held elections would be sent first to territorial authority returning officers and DHB voting papers would then be sent to a central location for counting.

49. STV voting papers will contain a large amount of information. Manual entry of voting papers into a computer would take a good deal of time and resources. Papers could be scanned into a computer, but this creates difficulties with developing suitable voting papers and issues of compatibility with the printing of other local government voting papers. The development of a voting paper which is simple, allows people to unambiguously rank a number of candidates, and can be scanned and processed presents a significant challenge.

50. It should be possible to have technological solutions in time for 2001 DHB elections, if the policy development process was completed in time. The costs and risks involved with these solutions are discussed below.

COST OF IMPLEMENTING STV FOR DHBs IN 2001

51. The cost of conducting DHB elections using STV voting will be higher than using the current local government electoral process. Costs will be incurred both in the capital development required and in the operation of the election itself. These costs are likely to be incurred on an ongoing basis.

52. Officials have undertaken preliminary costing work. It is not possible to develop detailed costs ahead of policy decisions about the detailed structure of the STV system to be used.

53. The cost of 'piggy-backing' DHB elections onto local government elections, under the current local government voting method, has been estimated at about $4 million per election. This cost would be met by DHBs.

54. It is estimated that the extra cost of using STV for DHB elections could be in the region of approximately $3 million for one-off and ongoing operating costs and up to $1.5 million for capital. This marginal cost results from capital and development costs and the operational cost of the election.

55. One-off operating costs relate to establishing a DHB electoral policy unit in the Ministry of Health which would undertake the necessary policy development and consultation work and manage the infrastructure development process. The development of electoral regulation for inclusion in health legislation is not the Ministry of Health's core business and expertise in electoral issues would need to be brought into the Ministry for this task.

56. Ongoing operational costs would be incurred each election and relate to:
    1. Publicity and voter education
    2. Running the election - scanning voter forms, data entry, and staffing a central processing unit.

57. Capital costs relate to:
    1. Design, development and testing of computer systems to run an STV election. This work would need to be contracted out to a software designer
    2. Investment in computer hardware to run the election.

58. It is not possible, on the basis of the information available to date, for officials to provide other than a broad indication of potential costs. Of necessity, a number of key assumptions have been made which, following detailed policy development, could change the costs of implementing STV for DHBs in 2001. Should Ministers wish to implement STV as an option for DHBs in 2001, officials will undertake more detailed costing work and provide further advice to Cabinet as part of a 31 December report back on the fiscal implications of health sector change.

RISKS OF INTRODUCING STV IN 2001

59. There are risks of introducing STV for DHB elections in 2001 in each of the areas of development identified:
    1. policy work to decide on and specify the exact details of the STV system to be used
    2. publicity and voter education
    3. design, development and testing of systems to read and count votes.

60. These are risks for the DHB elections themselves and to the credibility of the newly-established DHBs. Difficulties - actual or perceived - with the implementation of STV for DHB elections in 2001 would also reflect badly on the voting system itself and would be likely to impact significantly on the willingness of local authorities and/or their communities to adopt STV for the 2004 or subsequent local body elections.

61. The risks in each of these areas are inter-related. An undetected error in the policy development phase, for example, will be carried through into the development of technical systems. A failure of voters to understand the voting system and the prospect of an increased number of invalid votes will jeopardise the objective of greater minority representation but could also impact significantly on local government elections.

62. The risk of problems with the implementation of STV is heightened when the timeline for policy and systems development is tight and has little flexibility. This is coupled with the fact that modern STV electoral systems are untested in New Zealand and the technology to implement such a system is as yet undeveloped.

63. Some areas of potential failure are:
    1. delays in providing an election result, or votes being lost, because of the logistics of a centrally based counting system
    2. delays in providing an election result, as a result of technical problems, for example, because of a complication in the scanning of votes or a problem with the computer software or hardware which generates the election result
    3. an invalid election result as a result of a computer error; while the risk of this happening should be low, if it did occur it may not be detected straight away due to the complexity of the counting algorithm and the large number of voting papers being processed
    4. a large number of invalid votes due to voter confusion with different voting systems occurring concurrently which require them to make different types of responses (this is also a risk for all other elections occurring concurrently)
    5. poor voter response because of the confusion with voting systems (although if voters felt that their votes 'counted for more' then the response may increase)
    6. Court proceedings challenging the validity of the voting and vote counting mechanisms.

64. The combination of electoral options discussed in this paper for DHB elections in 2001, and a summary of the discussion of these is shown in Table One.

65. As Table One shows, none of the options for DHB decisions in 2001 considered in this paper can guarantee that all of the desired objectives for DHB board membership and function are achieved. The voting system which is chosen will involve making trade-offs between community representation, the function of the board and the cost and risk of implementing a new system. Any electoral system design can also only respond to the candidates who stand and to the people who 'turn out' to cast their votes.

66. In summary:
    1. STV is likely to be the best system for promoting minority representation, but there are significant costs and risks involved with implementing this option for 2001
    2. the addition of a constituency system to STV when there are only seven elected members across the whole district takes away almost all of the benefits of STV, and officials recommend against this option
    3. a constituency system can be implemented relatively easily, and will achieve a measure of geographic representation, but makes it harder for minority candidates, including Maori, to get elected, and risks factionalising the board
    4. while an at large election using the current local government voting method is the easiest to implement and presents the least risk, it cannot guarantee geographical or minority representation.

CCMAU COMMENT

67. CCMAU considers that this paper should address a problem with the constituency system, which is that it does not sufficiently highlight or address a problem with the constituency system, which is the significant inequality of voters per member for any one DHB. This was less of a problem with area health board elections because elected members could number up to 17 in special cases.

68. Creating special purpose constituency boundaries would seem to be the only practical way to achieve closer suffrage equality and overall process integrity. The Local Government Commission would be the most suitable body to determine or formally advise on such matters, on the lines of section 101L of the Local Government Act 1974, and would be best placed to minimise any necessary divergences between DHB and local government constituency boundaries.

CONSULTATION

69. This paper has been prepared by the Ministry of Health in consultation with the Department of Internal Affairs (Local Government), the Ministry of Agriculture and Forestry (Rural Affairs), The Treasury, CCMAU, and the HFA. Te Puni Kokiri and the Ministry of Pacific Island Affairs have also been given the opportunity to provide comment.

FINANCIAL IMPLICATIONS

70. The cost of running DHB elections in 2001 under the local government voting system, and where candidates are elected at large, has been included in the paper titled 'Fiscal implications of Health Sector Change' to be considered by Cabinet on 27 June.

71. Introducing a constituency system will not add to these previously identified costs.

72. Introducing STV for DHB elections in 2001 will be more expensive than using the current local government voting method. Indicative costs have been provided in this paper and indicate that the marginal cost of using STV for DHB elections in 2001 is about $3 million for one-off and ongoing operating costs and up to $1.5 million for capital. However these are subject to change following more detailed policy work and will also depend on the speed with which the implementation of STV is desired. Officials will provide further advice to the Ministers of Health and Finance on cost, should STV be desired for DHB elections in 2001.

LEGISLATIVE IMPLICATIONS

73. The New Zealand Public Health and Disability Bill will require a section on electoral processes. Since provisions for an STV election do not currently exist, the Bill would have to include a provision for the electoral processes to be defined in regulations or schedules at a later date.

HUMAN RIGHTS ACT 1993

74. There are no Human Rights Act 1993 implications in this paper.

REGULATORY IMPACT STATEMENT

75. A regulatory impact statement prepared in accordance with CO (98) 5 is attached.

PUBLICITY

76. Any publicity on matters related to this paper is being managed as part of the Communications Strategy that forms part of the wider work on health sector change.