Written evidence submitted by Professor Toby S. James, Professor of Politics and Public Policy, University of East Anglia, and Dr. Holly Ann Garnett, Associate Professor of Political Science, Royal Military College of Canada, Co-Directors of the Electoral Integrity Project [ELR 006]

Introduction

1. We are responding to the request to submit evidence on voter registration for the Levelling Up, Housing and Communities select committee.  We are the co-directors of the Electoral Integrity Project, a world-leading academic project which produces innovative and policy-relevant research on elections worldwide.  Our joint and individual research expertise includes research on voter registration systems, including an international comparison of systems.  We have also undertaken evaluations of the quality of electoral registration system in the UK.  The recent report Is it time for Automatic Voter Registration in the UK? [1], provides an implementation plan for how automatic voter registration can be implemented in the UK. 

Key measures of electoral register quality

2. To assess the quality of electoral registration in the UK, there are four core considerations:

• The accuracy of the electoral register can be usefully defined as the extent to which there are no false entries on the electoral registers.  Accuracy is therefore commonly measured as the percentage of entries on the registers which relate to verified and eligible voters who are resident at that address. Inaccurate register entries may relate to entries that have become redundant (for example, due to home movement or death), duplicate registrations, ineligible electors, or fraudulent registrations. 
• The completeness of the electoral register is the extent to which every person who is entitled to be registered, is registered.  The electoral register should include all citizens eligible to vote. Low levels of completeness can lead to citizens not being able to vote on election day, but also affects other processes such as the drawing of boundaries and jury service.
• Register equity – the degree of equity in electoral registers refers to the extent to which there is an even distribution in the completeness of the electoral register across educational, socio-demographic, ethnic, gendered or other groups.
• Administrative robustness – electoral registration processes must be deliverable without errors which can lead to citizens not be able to vote or the trust in the system being undermined. This requires sufficient staffing, resource and capacity.  The administrative processes must also be efficient with potential risks identified – otherwise electors may not be able to vote on election day.

What are advantages and disadvantages of the existing system of electoral registration?

3. The strengths of the existing system are that:

• The online voter registration system provides a user-friendly portal for citizens to be able to register – and has proven enormously popular since  introduction.
• Individual ownership over the voter registration record ensures that every citizen can register themselves to vote and is not dependent on a ‘head of household’ or other actor.
• The reformed canvass enables automatic re-registration for citizens who have not moved and that can be data-matched.  This is time efficient for citizens and for administrators.

4. The weaknesses are that:

• There remains low levels of voter registration amongst specific groups and this is a problem which is widening over time.  There is a specific problem with ‘attainers’ – usually 16 and 17 year olds who become eligible to vote in the near future.  The number of attainers have dropped from 541,000 in 2009 to 177,000 in 2021.  This is largely the result of the move to individual voter registration [2].  Prior to this a head of household would have been able to complete the form for younger members of the household.  Research shows that removing this facility contributed towards a decline [3].
• The most common problem at polling stations is not cases of impersonation, but electors who wish to vote, but whose name is missing from the electoral register.  At a general election, more than two-thirds of poll workers report that they are turning away one or more voters because their name does not appear on the electoral register [4].
• There is no complaints system for electors who experience a problem and who may therefore wany to log their problem and seek redress. 
• There is no online portal that allows citizens to check their registration status.  This causes thousands of duplicate registrations.  An ‘am I registered?’ website would prevent this. 
• Voter registration is exceptionally seasonal – with most voter registration applications now occurring in the period following the calling of an election.  This creates huge administrative problems for electoral officials.
• The online system encourages already registered citizens to register to vote, ‘just in case’.  This creates duplicate registrations.
• There are considerable administrative pressures on electoral officials owing to the absence of ‘am I registered?’ website, tight turnarounds at election time and scare resources.

How does the system of Individual Electoral Registration compare to an automatic or assisted system of voter registration?

5. Our research argues that a continuum of voter registration models exists, with two extremes, and a variety of models between the two (Figure 1).

• At one extreme is a model in which the individual is responsible for their own voter registration at each election.
• At the other extreme is automatic registration. Most commonly this is where registration lists are compiled through sources such as municipal or state residence lists. This model provides the greatest ease for the voter, since the state actively insists on their voter registers being up-to-date and accurate.
• An assisted method of voter registration lies in the middle of the spectrum.  The responsibility lies with the individual, but there are a greater range of methods provided to encourage them to register to vote (be that online, canvassing, in-person registration, prompts when interacting with other government services).

Figure 1: A continuum of Electoral Registration Models

6. Globally, we find that the world is roughly divided evenly between cases that may be considered automatic model (49%) and individual (including assisted, with citizen-responsibility) models (51%).
7. Our research shows that there is greater accuracy and completeness in countries with automatic registration (Figure 2). [1]  The main reason for this difference, we think, is that voter registration is much easier for the citizens under automatic voter registration because they need to take no steps to register.  Registration can be a significant burden to some voters: especially those with low-information, low-resources or high mobility. By placing the burden of registration on electoral officials, rather than the citizen, automatic registration reduces the cognitive, time and resource costs of registration for everyone, but especially for these population groups, who are already less likely to be represented at the polls. [2] It is therefore more likely to capture citizens who would find it difficult, time-consuming or costly to register otherwise.

Figure 3: Completeness and Accuracy by Electoral Registration Models

Figure 2: The effects of individual and automatic voter registration systems on the completeness and accuracy of the electoral register.  Source: authors, based on data from the Perceptions of Electoral Integrity Index [6].[3] 

Which countries have high levels of electoral registration, and what lessons can the UK learn from these electoral registration systems?

8. The quality of the electoral registers around the world can be measured using the Voter Registration Index, which is part of the Perceptions of Electoral Integrity Index.  This is an expert survey constructed by the authors [6, 7], building on the work of Professor Pippa Norris and colleagues, to measure election quality globally. [4]

9. The top four scoring countries on the Voter Registration Index are: Finland[5], Iceland[6], Switzerland[7] and Denmark[8], followed closely by Taiwan[9]. All five countries rely on existing population registers (either nationally or locally) for voter registers. 

10. The highest scoring country, Finland, has it’s electoral register compiled by the Digital and Population Data Services Agency using information from the Population Information System.[10] These data are then publicly displayed, and each elector is sent a notice of their right to vote. Any elector who is not on the list may make an appeal before the election to have their name included. This model rests on the usage of an existing population register and is managed by the agencies and departments involved in its administration.

Figure 3: Perceptions of Electoral Integrity Voter Registration Sub-Indices (Top 14 countries listed)

How can existing public data and digital methods be better utilised to create a more joined up electoral registration system?

11. The Joseph Rowntree Reform Trust (‘JRRT’) report, Is it time for Automatic Voter Registration in the UK? [1],  sets out how existing public data and digital methods could be better utilised to create a more joined up electoral registration system.

12. To begin with, there should be one national electoral register rather than the current system of hundreds.  Electoral registers could be merged into one dataset with respective EROs still holding responsibility for the data for their respective areas. This would allow duplicates and inaccurate entries to be identified more easily.  Voters missing from the registers could more easily be identified.  When a voter moves, their records could be updated in one transaction. 

13. Citizens could be prompted to register to vote when they interact with other government services.  This could prompt their data to be sent to the respective EROs on a voluntary ‘opt-in’ basis.  The JRRT report provided estimates on how many new registrations could be added through this method. Up to:
• 6.5 million per year could register when applying for a passport
• 4 million people could register when they update their driving licence address with the DVLA
• 2 million a year could register when applying for Universal Credit
• 2.5 million students could be registered through annual student enrolment
• 800,000 could register when they apply for child benefit for the first time
• 500,000 could register when they provide the Student Loans Company with a new address
• 450,000 could register when they apply for disability benefit

14. Citizens could also be directly enrolled when they interact with other government services.  For example, if they submit an application or change of address to any of the above organisations, this data could be provided to EROS, who could then use this amend or query electoral registration entries.  This would require legislative change to give EROs power to register electors automatically.

What issues exist regarding cyber security, data and privacy, and how can these concerns be addressed?

15. The main cyber security issue and threat to elections is posed by the open/edited register.  This is currently free for anyone to purchase for any purpose.  This means that overseas actors could purchase the data and compile it into a single dataset for malign purposes.  For example, it would be possible to identify marginal constituencies and then micro-target voters with postal misinformation about the voting process to discourage them to vote.  The open/edited register serves no electoral purpose and should therefore be abolished to safeguard against this threat.

What issues do electoral registration officers face in relation to electoral data, including access to and sharing of data?

16. There is a common assumption by citizens that they are registered to vote because the ‘government knows about them’.  Citizens pay their local and national taxes and therefore commonly assume that electoral registration services have access to this information and that they are therefore registered to vote.  However, EROs do not have access to many datasets. 

17. The key dataset which they could be provided access to is the Department for Work and Pensions Customer Information Service dataset.  It is the closest that the UK has to a civil population register – and in many ways already serves the purpose of one.  EROs can currently check applications to register to vote against this dataset – but cannot view the wider dataset to improve the accuracy and completeness of the electoral register. 

What are the challenges presented by event-led registration, and what additional burdens does this place on local authorities?

18. The move to online voter registration has created major seasonal pressures in voter registration.  Whereas the annual canvass brought a steady flow in applications prior to online voter registration, there is now a major rush of applications in the immediate pre-election period.  This is greatest when a snap general election is called because there is then a short window before the voter registration deadline falls.  There are often also a huge volume of unnecessary duplicate applications that come in during this window from citizens who are already registered.  There have been consider cost pressures on local authorities in recent years which has led to reduced staff teams in many places [8-10]. 

19. Policies to ease these pressures would include an ‘am I registered?’ website so that citizens can check their registration status.  These are provided in many other countries.  For example, Ireland provide citizens with a ‘Check the register’ service https://www.checktheregister.ie/.  In Mauritius is it possible to send an SMS message to check whether you are on the electoral register.  Providing automatic voter registration would also increase voter registration rates, which would in turn, reduce administrative pressures at an election.

What are the barriers to eligible electors registering to vote?

20. The barriers to voter registration would include:
1. Educational – there is no systematic voter education in schools, the natural place to first engage and educate citizens about the democratic process.
2. Informational – there remains an assumption amongst many people that they are already registered because they pay taxes.  However, voter registration is not automatic and registers are less accurate and complete as a result.
3. Procedural – there are key points in citizens’ life cycle where they could be registered to vote such as when they are issued with their national insurance number.  Once they were registered.
4. Data access – EROs do not have the key data available to them to automatically register electors.
5. Resources – EROs and their teams have had to minimalise voter outreach work because of limited financial resources.  Civil society groups also operate on small and short-term budgets leading to quickly lost capacity and knowledge.
6. Motivation – levels of political engagement and broader apathy amongst the public is a problem in all countries. 

Why are there so many inaccurate entries and duplications on the register? How can they be rectified in a cost-effective manner?

21. Ensuring complete and accurate electoral registers is a huge challenge.  The challenges are greater where there is no civil population register, as is the case in the UK.  As outline above mechanisms and improve accuracy and completeness rates are:
1. Automatic voter registration
2. Assisted voter registration
3. Funding for electoral officials [5]
4. Funding civil society groups to undertake voter engagement work and registration drives

Recommendations

Overall we would recommend that:

a)      Automatic voter registration is introduced – at least for groups at key moments in their lives.  Registering all citizens at the point of issuing their national insurance number would be the most straight forward and equitable way to do this.  It would also generate cost savings.

b)      Assisted voter registration is introduced – prompting citizens to register to vote when they interact with other key government services.

c)       Setting up a ‘am I registered?’ website so citizens can check their registration status.

d)      Providing a central government fund to support voter registration drives in low-registration areas.

e)      Re-evaluating the case for a central electoral register.

f)        Abolishing the open register.

g)       Using electronic electoral registers in polling stations.

h)      Providing a centralised complaints system for voter registration issues

i)        Supporting legal consolidation for electoral law.

Bibliography

1.              James, T.S. and P. Bernal, Is it time for automatic voter registration in the UK? 2020, Joesph Rowntree Reform Trust: York.

2.              ONS, Electoral statistics for the UK. 2023, ONS: London.

3.              James, T.S., Comparative Electoral Management: Performance, Networks and Instruments. 2020, London and New York: Routledge.

4.              Clark, A. and T.S. James, Poll Workers, in Election Watchdogs, P. Norris and A. Nai, Editors. 2017, Oxford University Press: New York.

5.              James, T.S. and H.A. Garnett. The Determinants of Electoral Registration Quality: A Cross-National Analysis. in Paper for the Annual Political Science Association Conference. 2021. Seattle.

6.              Garnett, H.A., T.S. James, and M. MacGregor, Perceptions of Electoral Integrity, (PEI-8.0), P. Electoral Integrity, Editor. 2022, Harvard Dataverse.

7.              Garnett, H.A., T.S. James, and M. MacGregor, Electoral Integrity Global Report 2019-2021. 2022, Electoral Integrity Project: Kingston and Norwich.

8.              James, T.S. and A. Clark, Delivering electoral integrity under pressure: local government, electoral administration, and the 2016 Brexit referendum. Local Government Studies, 2020. 47(2): p. 186-207.

9.              James, T.S. and T. Jervier, The Cost of Elections Funding Electoral Services in England and Wales. 2017, ClearView Research: London.

10.              James, T.S. and T. Jervier, The cost of elections: the effects of public sector austerity on electoral integrity and voter engagement. Public Money & Management, 2017. 37(7): p. 461-468.

February 2023