The Electoral Commission

The independent body which oversees elections and regulates political finance in the UK

Section menu

Response to the UK, Scottish and Welsh governments’ policy statement: Proposals for reform of the annual canvass

This response sets out our views on proposals to reform the annual electoral registration canvass in Great Britain. We are responding to this consultation as part of our statutory role to keep electoral law under review and to recommend changes where we think they are needed.

Summary

We welcome proposals to enable Electoral Registration Officers (EROs) to use data to improve the efficiency and effectiveness of the canvass of residential properties that they are required to do each year. We have previously highlighted our concern that the current annual canvass requirements are not sustainable.

The proposed data discernment step, enabling EROs to check the information they hold about a property against national and local datasets, should help EROs to target resources better and identify properties where residents are more likely to have changed and may not be registered to vote. EROs should also be able to use local data to supplement the results from mandatory national data matching.

Using poor quality data for matching could mean, however, that EROs do not identify and make sufficient contact with properties where there has been a change of residents. This risks reducing the accuracy and completeness of electoral registers over time. To mitigate this risk, the UK’s governments should consider how to specify in law the factors that EROs should take into account when deciding whether to use specific local datasets.

Managing implementation

These important and significant changes will need careful implementation and management to ensure that they are delivered successfully. The UK’s governments have a key role to play in ensuring that electoral management software systems can support both local and national data matching, and that EROs and their teams have the technical skills and training they need to use these systems. We also want to ensure that new canvass communications and form designs have been properly tested with the public before EROs use them.

The window of opportunity for implementing these proposals in time for the 2020 annual canvass is small. EROs, software suppliers and the Commission will need certainty about the detailed policy and legislation for canvass reform by not later than December 2019, and indeed some preparatory work will need to begin well before then. The UK’s governments need to ensure that work on developing and agreeing implementation plans is accelerated now to meet this challenge.

The UK’s governments will also need to consider how these important changes will be delivered alongside other significant electoral registration and franchise reforms that are currently being considered by governments and legislatures. These include allowing all British citizens overseas to register to vote at UK Parliament elections, allowing 16 and 17 year-olds to vote at elections in Wales, and extending the franchise in Scotland to include citizens of EU member states.

The UK’s governments should also consider the extent to which areas of electoral law relating to the annual canvass could be simplified as part of annual canvass reform, in line with the principles established by the UK’s Law Commissions’ recent comprehensive review and recommendations on electoral law reform.

Further reforms to deliver a modern electoral registration system

We believe that these changes are an important first step to modernise the electoral registration system. There is, however, more that should be done to build on these reforms to ensure we have an effective year-round registration process. We want the UK’s governments to develop longer-term plans for electoral registration reform that build on the canvass reform proposals.

These should include giving EROs further powers and tools to more easily identify people who may be eligible to register to vote throughout the year, and to give those people every opportunity and encouragement to register. For example, allowing EROs to use a wider range of reliable and trusted information from other public sources to maintain the registers; integration of electoral registration into other public service transactions; and exploring the feasibility and benefits of more automated forms of registration.

Data discernment step

The data discernment step process

We support the introduction of a data discernment step to improve the efficiency and effectiveness of the annual canvass. A modern electoral registration system should use trusted public data to help ensure it remains accurate and complete throughout the year.

EROs will need the skills and capacity to effectively implement both mandatory national data matching and discretionary local data matching.

Properties should only be considered a green match if an ERO is satisfied that all currently registered people in the property are still resident there. If they are not satisfied, they should be required to follow the more intensive route two canvass approach for that property.

Only good quality, accurate locally-held data should be used when matching electoral registration records ahead of the annual canvass. The UK’s governments should consider whether it is possible to specify in law the types of factors that EROs should take into account when deciding whether to use specific local datasets.

Question 1: We are proposing that the national data matching process is mandatory to complete, with local data matching being conducted at the ERO’s discretion. Do you agree that this is the right approach?

  • We agree that the national data matching process should be mandatory, with local data matching allowed at the ERO's discretion. EROs will need the skills and capacity to implement both elements of this proposal effectively.

If the national matching is to be an annual, mandatory process, EROs will need to be able to easily send the necessary data to be matched. The results which are returned to them will need to be clearly presented in a format that is compatible with their electoral management software systems. UK governments’ implementation plans should ensure that these systems can support both local and national data matching, and that EROs and their teams have the technical skills and training they need to use the systems.

We support the proposal that EROs should be able to use local data matching to supplement the results from mandatory national data matching. Reliable and current local data should help EROs to identify more accurately where there has been no change of residents at a property. If the national matching is to be an annual, mandatory process, EROs will need to be able to easily send the necessary data to be matched. The results which are returned to them will need to be clearly presented in a format that is compatible with their electoral management software systems. UK governments’ implementation plans should ensure that these systems can support both local and national data matching, and that EROs and their teams have the technical skills and training they need to use the systems.

EROs should, however, only use reliable and robust data sources which contain current data on properties and residents for local data matching. Using poor quality data for matching could mean that EROs do not identify and make sufficient contact with properties where there has been a change of residents. As set out in more detail in response to question 3 below, UK governments should consider how to specify in law the factors that EROs should take into account when deciding whether to use specific local datasets.

Question 2: We are proposing that any property with a red elector must be converted into a red property. A property will only be green if 100% of electors in the property are individually green. Do you agree this is the right approach?

  • We agree that any property with a red elector must be converted into a red property, and that a property should only be considered green if an ERO is satisfied that all currently registered electors in the property are still resident there. If an ERO is not satisfied, they should be required to follow the more intensive route two approach for that property.

The approach proposed in the policy statement would reduce the risk of a negative impact on the effectiveness of the canvass and the resulting accuracy and completeness of electoral registers. We agree that any property should be subject to the full canvass process if the ERO is not satisfied that all currently registered electors are still resident there. We also recommend that properties where the ERO has information to suggest that there may be new potential electors should also be subject to the full canvass process.

The policy statement also suggests that, following the matching process, “The ERO will have the discretion to override an individual’s result if they have a valid reason to do so”. We agree that this is likely to be useful for EROs in certain limited circumstances. Feedback from previous data matching exercises suggested that some people (e.g. someone who has changed their surname after marriage) fail to match with national data sources, even where their change to personal details is not recent. It would be important to try to avoid scenarios where the same elector repeatedly, year-on-year, cannot be matched against a national data source but it is clear to the ERO that they are still resident at an address.

UK governments should consider carefully how to define the power for EROs to use their discretion to override an individual’s match result. We recommend that the scope of this power is defined in legislation, to minimise the risk that EROs could apply inappropriate criteria and override match results without sufficient grounds, which could have the effect that EROs do not carry out appropriate canvass activities to identify actual changes in residents at some properties.

There may also be a role for guidance in supporting EROs to use their local knowledge and judgment in helping to resolve match result queries.

Question 3: Do you think a minimum standard for the accuracy of locally held datasets should be mandated?

  • We agree that only good quality, accurate locally-held data should be used when matching electoral registration records ahead of the annual canvass. The UK governments should consider specifying in law the factors that EROs should take into account when deciding whether to use specific local datasets.

Given the wide range of different, locally held datasets, there is a risk that some EROs could use poor-quality or out-of-date data to match electoral register records ahead of the canvass. Using poor quality data for matching could mean that EROs do not identify and make sufficient contact with properties where there has been a change of residents.

To reduce this risk, UK governments should consider how to specify in law the factors that EROs should take into account when deciding whether to use specific local datasets. This could be similar to the approach which has been taken in Ministerial guidance for EROs on using local data matching in the verification process under IER. This sets out a list of potential data sources, but also specifies criteria which EROs must use to assess each data source before using them.

Exemptions from the standard data discernment step

EROs should be able to follow the less intensive route one canvass process if they are satisfied that a property is currently empty or should remain void because no eligible people are living there.

People who have recently been added to the register should still be checked against national datasets to ensure that EROs have the maximum amount of information on which to base their canvass management decisions.

The single occupancy tick box should be removed from registration forms. Requesting non-essential information can be confusing or off-putting and should be avoided.

Question 4: Do you agree that empty and void properties should be sent through a data matching process?

  • We agree that empty and void properties should be sent through a matching process using local data. EROs should be able to follow the less intensive route one canvass process if they are satisfied that a property is currently empty or should remain void because no eligible people are living there.

The process outlined in the policy statement is logical. EROs should be able to use accurate local data on empty or void properties to determine which canvass route is appropriate. As with other types of properties, if an ERO is not satisfied that a property is currently empty or should remain void because no eligible people live there, they should be required to follow the more intensive route two canvass approach for that property.

Question 5: Do you agree that recent applications to register should be exempted from the data step and automatically marked as green? If yes, what time period do you think should be defined as “recent” (1 month, 2 months, linked to the last monthly update)?

  • People who have recently been added to the register should not be exempted from the data step and automatically marked as a green match. Instead, EROs should be able to use their discretion to determine that individual entries should be a green match if they are satisfied that the specific person is resident at the property.

Records for people who have recently been added to the register should still be checked against national datasets. This would ensure that EROs have the maximum amount of information on which to base their canvass management decisions. We do not think it would be appropriate for EROs to adopt a blanket approach to determine that all such records should be a green match without any specific consideration of each individual record.

We recognise that in some cases these records may be returned as a red match following data matching because there could be a delay in the national data being updated. In these cases, we think it should be open to EROs to use their discretion to determine that individual entries should be a green match if they are satisfied that the specific person is resident at the property. A recent application could form part of the evidence for this decision, but in some cases there may be other evidence which indicates that the person is not in fact resident.

In our response to question two above we explain that it will be important to consider carefully how EROs should determine whether there is a valid reason to override an individual’s match result. This would include considering the time period during which a recent addition on the register could be considered sufficient evidence to override a red match. In areas of high population movement, for example, a relatively short period might be appropriate, while there may be a case for allowing a slightly longer period in areas with a more stable population.

We also suggest considering further safeguards to minimise the risk that properties where there has been a change in occupation do not receive the appropriate canvass communication. For example, as applications can now be made online, where a recent successful applicant is changed to green (whether automatically or by ERO discretion), they should be sent a route one communication to establish a link to the property rather than being able to respond to electronic or telephone canvassing (see question seven).

Question 6: Do you agree with no longer including a single occupancy tick box on registration application forms?

  • We agree that the single occupancy tick box should be removed from registration forms. All forms should be designed with a clear purpose and intention. Requesting non-essential information can be confusing or off-putting and should be avoided.

Where used appropriately, the single occupancy tick box has been a useful tool under the current system. However, we agree that it would be appropriate for single-occupancy properties to be treated in the same way as other property types under the proposed canvass reforms. It should no longer be necessary to collect this information on application forms.

Route one: the lighter touch

We agree with the proposed process for route one.

EROs should have the flexibility to use electronic contact channels such as email to make initial contact with households under the route one process. This should be followed by a paper contact if there is no response to the electronic contact.

EROs should not be required to follow up after sending the mandatory route one communication. If they are satisfied that there has been no change of residents at the property, there would be no need to require a response.

Question 7: Do you agree that an email contact should be permitted as the first form of contact for households in route one (where an email address is held), followed by a paper contact if there is no response?

  • We agree that EROs should have the flexibility to use electronic contact channels such as email to make initial contact with households under the route one process. This should be followed by a paper contact to the property if there is no response to the electronic contact.

We agree with the principle of allowing email (or other electronic) contact as the first form of contact for households under the route one process. Based on data about the number of email addresses held, likely costs and response rates, EROs may decide that this would be a cost-effective way to begin the canvass process for these properties.

EROs should not be required to choose one person to act as the head of the household for this contact. Instead, they should be required to send an electronic contact to all residents in a property for whom they hold relevant details.

Systems for responding to an electronic contact under route one should require respondents to select the registered address. This would reduce the risk of false positive responses from people who have actually changed address but have not understood the purpose of the electronic contact.

We agree that EROs should be required to send a paper communication to a residential property if they have not received a response to an electronic communication. This would mitigate some of the risk in sending emails to incorrect or unused email accounts.

There are likely to be further risks to be identified and addressed for this approach, such as how the system would reconcile conflicting responses from more than one person in the household. We would like to see this option more fully mapped out to understand how it would work in all cases.

Question 8: Do you agree with the proposed process for the route one?

Question 9: Do you agree with the proposal to exclude mandatory follow up activity (household visits, etc.) with households sent through route one?

  • We agree with the proposed process for route one. EROs should not be required to follow up after sending a mandatory paper communication to a property under route one. If they are satisfied that there has been no change of residents at the property, there would be no need to require a response.

We agree with the proposal to allow a less intensive canvass process for properties that EROs have been able to match through national or local data matching. If EROs are satisfied from the data matching results that there has not been any change in residents at those properties, it would not be an efficient use of their limited resources to follow up a mandatory paper communication.

The success of this process will be dependent on EROs using good quality data for matching, and on well-designed and effective communications with residents at properties under route one. Without these in place, there is a risk that residents at some properties, where there has in fact been a change in people who are eligible to register, do not understand what they need to do to inform EROs about the change. This includes making sure it is clear what people who are not currently registered to vote, but who may be eligible, need to do. Without these, there could be a decline in the accuracy and completeness of registers over time.

Route two: the full canvass process

The proposed route two canvass process is intended to be used only for properties where the ERO believes there may have been some change in occupancy since the publication of the register. We agree that the current canvass requirements should form the basis for this route, to minimise the risk that changes in occupancy – including potential new electors – are not identified.

Personal contact should continue to be a mandatory element of the revised canvass for route two properties, to support EROs in maintaining an accurate and complete register.

The route two process

Question 10: Do you agree with the proposed process for route two?

  • We agree with the proposed canvass process for route two properties. The current canvass requirements should form the basis for this route, to minimise the risk that changes in occupancy – including potential new electors – are not identified.

The proposed route two canvass process is intended to be used only for properties where the ERO believes there may have been some change in occupancy since the publication of the register. We agree that EROs should continue to be required to make several contacts with these properties, to ensure that they obtain the information they need to either identify changes in residents, or confirm that there has been no change.

We agree that the initial contact with properties under route two should be by sending a paper form to the property. Where an ERO believes there may have been some change in occupancy, it would not be logical to send an electronic contact to an individual who is currently registered at the property.

We also support the introduction of some discretion for EROs to vary the way that follow-up contacts are made, allowing EROs to target their resources based on an assessment of what will work best. Our evaluation of the annual canvass pilots in 2017 showed that using different contact methods could result in higher response rates than when the same approach was made repeatedly.

Personal canvassing

Question 11: Do you agree that a personal contact (door knock or telephone call) should be a mandatory element of the revised canvass?

  • EROs should have some discretion to vary the way that follow-up contacts are made for properties under route two, but we agree that a personal contact should continue to be a mandatory element of the revised canvass process, to support EROs in maintaining an accurate and complete register by ensuring that responses are received from as many households as possible.

We recognise the benefits of house to house enquiries in the existing canvass model and believe this to be a vital component of any revised method. Removing a requirement for some form of personal contact would compromise the ability of EROs to maintain an accurate and complete register, and could mean potential new electors are not identified.

EROs should have discretion to choose whether these personal contacts are made in person or by telephone; however while the benefits and risks of door knocking are well understood, telephone calls are not a direct equivalent and thought should be put into dealing with the differences between these approaches. For example, while knocking on a door ensures that the household has been visited, a phone call can be made to an incorrect or unused number; an unanswered call or ‘dead’ dial tone therefore should not be counted as a completed contact attempt and should be followed up.

Route three: Exemptions for certain types of property

We agree that EROs should be allowed to adopt a different approach to canvassing specific categories of property where there are multiple occupants at the same address and the people who are resident there may change relatively frequently. This could allow more effective approaches to ensure that the register is accurate and complete in respect of these addresses, while avoiding inefficient canvass activity by EROs.

We recommend that the UK’s governments include military bases in this category.

We agree that where EROs are unable to obtain the necessary data, properties with multiple residents at the same address should be directed to route two. Given the nature of these properties and their residents, this would be a necessary safeguard to reduce the risk that changes are not fully picked up.

Question 12: Are there property types in addition to those detailed above that you believe should be directed to route three?

  • We recommend that the UK government considers including military bases in this category.

The UK Government should work with the Ministry of Defence to establish whether military bases (which are often characterised by the frequent movement of personnel) could be directed to route three.

Question 13: Do you believe this is the correct process to deal with these properties? If no, can you suggest an alternative approach?

  • We agree that EROs should be allowed to adopt a different approach to canvassing specific categories of property where there are multiple occupants at the same address and the people who are resident there may change relatively frequently. This could allow more effective approaches to ensure that the register is accurate and complete in respect of these addresses, while avoiding inefficient canvass activity by EROs.

EROs are likely to continue to find it difficult to obtain the necessary information from some categories of residential properties using current canvassing approaches. Given the nature of some categories of property – where there are multiple occupants at the same address and the people who are resident there may change relatively frequently – we agree that EROs should be able to use a different approach to obtain information more effectively.

We would welcome confirmation that the UK’s governments are content that existing powers for EROs would be sufficient to require relevant organisations or individuals to provide information as part of the route three process. If not, these powers would need to be established in order for the process to work effectively. The UK’s governments should ensure that any relevant national-level organisations (such as professional or trade associations, or other government departments) understand and support the requirements on local organisations and individuals.

Question 14: Do you believe that sending these properties into route two, the full canvass, if the ERO is unable to obtain data, is the correct safeguard for these properties?

We agree that where EROs are unable to obtain the necessary data using route three processes, relevant properties should be canvassed under the route two process. Given the nature of these properties and their residents, this would be a necessary safeguard to reduce the risk that changes of residents who are eligible to register are not fully picked up.

Treatment of ‘pending’ and ‘potential’ electors

We do not agree that ‘pending’ or ‘potential’ electors should be included in the data matching process or in any canvass communication. This would risk confusing people who may need to provide or update information as part of the canvass or registration application process, and increase the risk that EROs do not receive the information they need.

Question 15: Do you agree with the proposal that pending/potential electors should be included in the data matching and canvass communication?

We do not agree that ‘pending’ or ‘potential’ electors should be included in the data matching process or in any canvass communication. This would risk confusing people who may still need to provide or update information as part of the canvass or registration application processes, and increase the risk that EROs do not receive the information they need.

Potential or pending electors would still be contacted directly by EROs through the invitation to register process. Once this process has begun, the ERO should continue until the required stages have been completed. Including their names in a canvass communication could give the misleading impression that they are already registered to vote, and may deter them from completing the application process.

Because ‘pending’ or ‘potential’ electors are not included in the register, we are also concerned that it would be difficult to design a clear and effective canvass communication which adequately explains the difference between them and other residents who are included in the register. Including details of ‘pending’ or ‘potential’ electors is likely to mean that the design of canvass communications is not as simple or effective as possible.

While excluding ’pending‘ or ’potential‘ electors from the data matching process or any canvass communication would mean that some EROs may be required to contact properties under the more intensive route two process, the benefit would be reduced risk of confusion and EROs therefore not receiving the information they need.

Form design

Any annual canvass communications, including forms, should focus on sharing and collecting essential information required to help EROs understand who is resident at properties in their areas.

We have identified some elements of the current canvass forms which are not necessary to fulfil this function and should be removed or amended. This would make them easier for electors to complete, increasing responses and, in turn, registrations.

Canvass reform provides an opportunity to reconsider the underlying purpose of communications with citizens during the canvass – any new communications and forms should be designed to deliver a clear function and intended response.

Question 16: What do you think the issues with the current HEF are?

Question 17: Is there information that can be taken out of the HEF?

  • Any annual canvass communications, including forms, should focus on sharing and collecting essential information required to help EROs understand who is resident at properties in their areas. We have identified some elements of the current form which are not necessary to fulfil this function and should be removed or amended. This would make them easier for electors to complete, increasing responses and, in turn, registrations.

Since 2013, we have been responsible for designing forms for electoral registration in Great Britain, subject to relevant Ministerial approval. Much of the content of the forms is prescribed in law, which means we would not be able to remove elements which might be considered non-essential by EROs.

Based on our experience in designing and testing forms with the public, and taking into account other feedback we have received, we agree that there is scope for the current canvass forms and communications to be simplified and improved:

  • Extra elements that do not fulfil the primary purpose of identifying who is resident at an address and may be eligible to register to vote should be removed.
  • The UK’s governments should explore whether the word ‘optional’ could be removed from the request for phone numbers and email addresses. This would increase the number of people who provide this information.
  • Plain language and shorter, simplified sections should be used to improve the accessibility of the form.

More broadly, canvass reform provides an opportunity to reconsider the underlying purpose of communications with citizens during the canvass – any new communications and forms should be designed to achieve a clear function and intended response. We will continue to work with UK governments, EROs and electoral management software suppliers to ensure that canvass and registration application forms are well-designed and tested to understand how ordinary citizens use and respond to them.

Other proposed reforms

The canvass reform proposals should be in place at least six months before the 2020 canvass to allow for effective planning.

UK governments should consider carefully the potential overlap between canvass reform and other planned registration reforms, including the challenges of EROs delivering these simultaneously or consecutively.

UK governments should explore broader options for electoral registration reform, which build on the canvass reform proposals, to ensure that the system keeps pace with change.

Question 18: Is there any further feedback you would like to provide in relation to the proposed new model for the annual canvass that has not already been covered in another question?

We note that both the Scottish and Welsh Governments are planning to introduce legislation to extend the franchise for Scottish Parliament, Welsh Assembly and local government elections, and have also proposed other changes to the registration system. In addition, the UK Government is planning to implement its “votes for life” manifesto commitment, which will abolish the 15-year rule for overseas electors. It is possible that the timetable for implementing some or all of these significant reforms will coincide with the implementation of canvass reform. The UK’s governments will therefore need to consider carefully the potential overlap of these reforms, in particular the capacity to deliver simultaneous or consecutive implementation among EROs and their teams, as well as their electoral management software system providers.

It will also be essential to ensure that the canvass reform proposals are in place by no later than December 2019, at least six months before the commencement of the 2020 canvass, so that EROs in England, Scotland and Wales are able to plan effectively for the changes and have the skills and capacity in place to be able to deliver them effectively. Effective implementation planning will also be essential to enable the likes of timely user testing and translating of forms and materials, as well as the provision of revised guidance for EROs.

While the immediate focus should be to ensure the successful implementation of canvass reform, the UK’s governments should also be thinking longer term about how a more data-driven, targeted approach could serve as the foundation for further reform of the electoral registration system.

The focus should be on giving EROs further powers and tools to more easily identify people who may be eligible to register to vote throughout the year, and to give those people every opportunity and encouragement to register. These include allowing EROs to use a wider range of reliable and trusted information from other public sources to maintain accurate and complete electoral registers; integration of electoral registration into other public service transactions; more automated forms of registration; and consideration as to how duplicate applications might be more effectively detected and managed within the system.

These measures will help to ensure that our system keeps pace with change, while meeting voters’ needs and expectations in an increasingly digital society. We therefore recommend that the UK’s governments develop longer-term plans for electoral registration reform that build on the canvass reform proposals.

Publications