Household registration, and multiple registers makes checking accuracy difficult

3.1 There are currently over 400 separately compiled electoral registers covering the country. As the register is collated by address it is very difficult to check accuracy. In England, Scotland and Wales registration is by household, and no personal identifiers, such as date of birth, signature or national insurance number are required. This makes it more difficult to control the accuracy of the register because there is no independent source against which details can be checked.

3.2 The Commission definition of completeness and accuracy are set out below. Completeness means that every person who is entitled to have an entry in an electoral register is registered; that no one not entitled (including the dead and fictional person) are registered and that all relevant fields are completed in relations to every entry on an electoral register. Accuracy means that none of the information relating to any person registered is false or incorrect, intentionally or otherwise.

3.3 In contrast, in Northern Ireland additional information - signature, date of birth and national insurance number - is required to register. Photographic ID is required to vote. As there are no personal identifiers on the Register in Great Britain and because registration is by household it is difficult to confirm from the register the identity of individuals. An individual called "John Smith" at one address, for example, may or may not be the same person as the "John Smith" recorded at a different address. Other sources of data, such as Council Tax records, may not list each individual at an address.

Controls have been strengthened following concerns over electoral fraud

3.4 There have been a number of high profile cases of electoral fraud in for example Burnley, Bradford and elsewhere, leading to public concern about the robustness of the electoral process. [24] These cases arose from alleged misuse of absentee (postal or proxy) voting. Weaknesses in the controls over absent voting stem in part, however, from a lack of controls provided by the registration process. As no identifiers are provided at registration the Electoral Registration Officer had nothing to check a postal vote application against, as no details were sought at registration. It has been found that postal vote applications have been completed on behalf of another individual, without informing them. In the absence of reason to suspect fraud or error the Electoral Registration Officer had no choice but to accept the application for an absent vote at face value, as is the general approach in electoral law. This postal or proxy vote could then be used by the fraudster, without the legitimate voter knowing.

The Electoral Administration Act 2006 has strengthened controls, although it has not eliminated this weakness altogether. The changes made now require new applicants for absent votes to include their date of birth and a signature on their absent vote application form. During an election Electoral Registration Officers must make arrangements for staff counting the votes to check signatures on a sample of at least 20 per cent of postal ballot papers to the signatures and dates of birth on the original applications.3.5

EXISTING CONTROLS OVER REGISTERS ARE NOT SUFFICIENT TO PREVENT FRAUD OR ERROR

3.6 The system as it currently stands operates to a large extent on trust. Unless the Electoral Registration Officer has good reason to think that there is an error (deliberate or accidental), they cannot challenge the return. From our interviews with Electoral Registration Officers we found that their approach, where there are doubts over a registration form, is to ask for more information from registrants. There are powers to do this in the relevant regulations. If no reply is received, however, Electoral Registration Officers' responses suggested they would be more likely to err on the side of leaving someone on the register, rather than removing them.

The principal method of checking is scrutiny of forms by Electoral Services teams for apparent anomalies, such as unusually large numbers of registrants at a single residence or registrations at a derelict property. This approach relies on a combination of judgement, local knowledge and corroborative information supplied by others, such as planning departments. The effectiveness of checking may vary.3.7

The current operation of the system is often effectively to accept the accuracy of the form at face-value. The checks that can be carried out [25] are sending out registration forms, personal canvassing, and inspecting other records available such as council tax records. These checks would not prevent additional voters being added to the register, or absent votes being applied for and subsequently misused. There are new powers for an Electoral Registration Officer to review register entries at any time. 3.8

The perception among Electoral Registration Officers is that there is little fraud

Our survey found that Electoral Registration Officers feel the current system of regulation has a slightly positive impact on increasing registration, monitoring completeness of the register and detecting inaccuracies. Electoral Registration Officers believe it has been neutral in respect of existing levels of fraud and negative in preventing fraud happening in the future. This perception may, however, increase the risk that fraud could go undetected, as Electoral Registration Officers may not be as watchful for fraud as they would be if they believed it to be more of a threat. None of these findings is an endorsement of the current system, but the findings on detecting inaccuracies and fraud and in preventing fraud suggest that Electoral Registration Officers do not view fraud as a major issue. This may increase the risk of fraud, as they will therefore not be as pro-active in managing the risk. The Commission has a long standing recommendation in place for the introduction of a system of individual registration which would mean that electors would have to provide personal identifiers when registering to vote. And in its 2003 report, The electoral registration process the Commission recommended the introduction of Co-ordinated On-line Record of Electors (CORE) which would provide the potential for the identification of illegitimate duplicate entries appearing and therefore increased accuracy. This development was provided for in the Electoral Administration Act with the intention of allowing registers to be consolidated into a centralised record.3.9

Source: NAO Survey on Electoral Registration

3.10 In evidence to the Committee on Standards in Public Life, some witnesses expressed concern about inaccuracies on the register. (See Annex 8 for a summary of the Report of the 11^th Enquiry, which covered the Electoral Commission.) One witness commented: "The first step is to take the matter seriously and that means recognising that redundant names is just as important a problem as missing names, because if you have people on the register who should not be, the dead and the non-existent, plus a system of postal voting that is, as I said to you in June, a lethal cocktail that leads to electoral fraud."[26] The Commission's guidance to Electoral Registration Officers on their new duties reinforced that the duty to take necessary steps included removing redundant entries.

Other witnesses were less convinced of the risk of widespread fraud. "Electoral fraud has been overstated for some time. I think some of the issues are around culture, around the way in which different communities perceive the operation and management of elections. Some of these issues need to be treated with a degree of sensitivity."[27] In early 2007, the Commission commissioned GfK NOP to undertake a pilot study to examine the accuracy and completeness of electoral registers across Greater London. This pilot will determine the feasibility and most appropriate methodology to employ in advance if rolling the research out to the whole of Great Britain later in 2007 with a separate study in Northern Ireland. The Commission plans to publish the results of this research in late May 2007.3.11

3.12 However, in its report on the 2005 General Election, the Office for Democratic Institutions and Human Rights concluded that "The introduction of postal voting on demand without the need to present a reason for the application, has demonstrated the vulnerability of any trust based electoral process.[28]

3.13 The Committee on Standards in Public Life noted some 20 cases of electoral fraud between 2001 and 2007 in its final report.[29] Going further, on 15 January 2007, in a speech to an elections conference, Sir Alistair Graham reported that he was aware that since 2001, 342 "cases of electoral malpractice" have been reported to the Crown Prosecution Service (CPS). The Commission has since examined the files in question and published its initial analysis on its website. Three per cent of the allegations referred to, a total of seven allegations, related to electoral registration. In the light of conflicting evidence and the perceptions of Electoral Registration Officers, it is crucial that the Commission investigates the accuracy of the Register, whether caused by fraud or error and then starts to address the issues.

3.14 The Electoral Commission has provided guidance on the new duties of Electoral Registration Officers under the Electoral Administration Act 2006. In interviews a number of Electoral Registration Officers stated that this guidance had been issued too close to the annual canvass to be helpful. However the Act was only passed in July 2006, with the provisions coming into force on 11 September 2006, giving little time. This may therefore have an impact on the consistency with which controls are applied.

INDIVIDUAL REGISTRATION IN NORTHERN IRELAND LED TO A DROP OF TEN PER CENT ON THE REGISTER

3.15 In Northern Ireland an identification document, national insurance number, date of birth and signature, are all required to vote. There is also a system of individual, in contrast to household, registration. This process was introduced to address perceptions of electoral fraud in Northern Ireland.

3.16 In August 2003, The Electoral Commission published a report giving a positive overall assessment of the new Northern Ireland regime.[30] Comments included the observation that the anti-fraud measures had had a positive impact in restoring integrity to the electoral process in Northern Ireland. However, the research also showed that the registration rate of 95.5% was likely to have been an overestimation of the actual numbers of people registered.

However, overall registration in Northern Ireland has fallen, from 1,069,160 (84.8 per cent of voting age population) in 2003, to 1,047,601 (83.1 per cent) in 2004. This number compares with a registration figure of over 90% for England and Wales in the Commission's September 2005 report. Low registration was also found to be particularly prevalent among certain groups, such as voters aged 18-24 (71 per cent registration rate). 3.17

The Commission's Report attributes the fall in Northern Ireland registrations to four factors: cessation of "carrying forward"[31]; identification of multiple registrations; removal of former residents of Northern Ireland no longer eligible; and potential fraudulent entries which were removed.3.18

3.19 The cost of individual registration in the first year was estimated at approximately £1.90 per elector. This was calculated by the Electoral Office of Northern Ireland, based on an overall cost of administration of £2.044 million. There is currently no reliable comparable data on the cost of registration in England, Scotland and Wales, although the data received from Electoral Registration Officers in our survey suggested an average cost of £1.39 per elector. (See para 2.24 above.)

Lack of research on accuracy makes it difficult to assess the extent of error and address it

The Commission recognises the value of further research into the accuracy of the electoral register. The pilot study referred to above in Greater London will enable the Commission to determine the most appropriate methodology to be employed for research planned later in 2006 to determine the accuracy of the electoral registers in Great Britain. The Commission has also worked with law enforcement agencies and the Association of Electoral Administrators to identify and tackle incorrect or potentially fraudulent registration and to actively promote best practice in electoral registration. In the meantime the development of performance measurement and a standard setting regime should help to identify and address immediate concerns over accuracy stemming from acknowledged weaknesses and known instances of alleged fraud. 3.20

Standard setting and performance measurement provide an opportunity to address both the completeness and the accuracy of the register

3.21 During our visits to local authorities, all Electoral Registration Officers supported the introduction of standard setting and performance measurement. They were especially interested in data and knowledge sharing between different local authorities to allow meaningful comparisons to be made and evaluations of different methods of encouraging registration and enhancing accuracy to be carried out.

3.22 Such regimes are not easy to design and operate. The Electoral Administration Act 2006 gives the Commission power to set performance measures and request information, but they have no direct control over Electoral Registration Officers. They are keen to avoid overburdening Electoral Registration Officers with too bureaucratic an approach to standards and performance measurement.

3.23 Our survey and field visits to Electoral Registration Officers found that they want performance measures to have the effect of improving performance and sharing best practice, rather than being used to punish "underperforming" authorities. One possible method of approaching performance measurement would be to use a risk-based system of peer review. These views have also been provided to the Commission and their plans are reflective of them.

The Commission has consulted widely as it develops its performance management framework

3.24 The Commission has used consultants, and consulted local authorities, as it establishes its performance management framework. The framework is still work in progress, and no firm decisions on individual indicators and standards have yet been made. The Commission has decided to concentrate first on indicators and standards for electoral registration. The Commission has decided that it needs to reflect the lighter touch regulation and performance measurement regime being established in local government, to achieve support from local authorities. If the performance measures are to achieve improvement in performance, the local authorities need to accept them, and feel that they add value. The relevant local government audit bodies and government departments should be consulted, to enable development of proposed measures which are in line with other performance measuring regimes.

3.25 Guidance published jointly by the NAO, Audit Commission, and others sets out good practice in performance measurement and pitfalls to avoid. [32] The guidance includes checklists which the Commission should use to challenge the performance measures it devises. We have offered the Commission guidance from our Performance Measurement team, once the potential development of measures is advanced further.

25   Section 9 of the Electoral Administration Act 2006 Back

26   Evidence from Dr Pinto-Duschinsky to the Committee on Standards in Public Life 14 September 2006 (11th Enquiry into the Electoral Commission) Back

27   Evidence from Sir Howard Bernstein, Manchester City Council to the Committee on Standards in Public Life 14 September 2006. Back

28   Office for Democratic Institutions and Human Rights United Kingdom Of Great Britain And Northern Ireland General Election 5 May 2005 OSCE/ODIHR Assessment Mission Report  Back

29   Committee on Standards in Public Life Review of the Electoral Commission January 2007. See Appendix 8 for a summary of the report. Back

30   The Electoral Fraud (Northern Ireland) Act 2002: An Assessment of its First Year in Operation, Electoral Commission. Back

31   i.e. if no registration form is returned, the name is removed from the register, rather than being carried forward for a number of years Back

32   Choosing the Right Fabric: A Framework for Performance Information. HM Treasury, 2003 Back