Voter ID at the May 2023 local elections in England: interim analysis

At local elections in England on 4 May, voters needed to show photo identification (ID) to vote in polling stations. These were the first elections in Great Britain where this requirement was in place. Elections took place in 230 areas in England and around 27 million people were eligible to vote.

Our interim analysis provides information and evidence currently available about how the new voter ID requirement was implemented and how voters found taking part. We have looked at evidence from large-scale public opinion research carried out before and after the elections, and polling station data collected from the majority of local authorities that held polls this year. 

There are several areas where further analysis is necessary to establish a clearer and fuller picture and they are highlighted through the report. We will publish our report on the May 2023 elections in September and that will include this additional analysis.

Our September report will also draw on a wider range of information sources, including feedback from candidates, Returning Officers, polling station staff, election observers, charities and civil society organisations, and the police.

If you would like to find out more about the Electoral Commission's views on our interim research, you can read our press release.

Summary

Awareness of the need to bring ID to vote at a polling station was high 

  • Our research found that immediately before polling day, 87% of people in England (excluding London, where there were no elections) were aware that they needed to show photo ID to vote at a polling station. In the research we carried out immediately following the poll, this was 92% of people in areas with elections. 
  • Awareness varied across the population and was lowest among younger age groups (82% for 18 to 24-year-olds), Black and minority ethnic communities (82%) and those who said they never vote in local elections (84%).
  • Awareness was significantly lower among people who said they did not have an accepted form of ID (74%) compared with those who did have ID (94%).
  • While overall awareness levels were high, some groups of people were significantly less likely to know about the requirement. This means that some people may not have known that they needed to show ID until they arrived at the polling station. Those people who did not have any of the accepted forms of ID would not have been able to obtain ID (for example the Voter Authority Certificate) on polling day itself.

Awareness and take-up of the Voter Authority Certificate was low

  • Approximately 89,500 people applied for a Voter Authority Certificate before the deadline on 25 April. Around 25,000 certificates were used as a form of ID on 4 May. 
  • In May 2023, awareness of the Voter Authority Certificate was 57% both among the overall population and those who said they did not already have photo ID.
  • The overall number of Voter Authority Certificates applied for and used was low compared with estimates of the number of voters who might not have any other accepted ID (250,000 to 300,000). 

At least 0.25% of people who tried to vote at a polling station in May 2023 were not able to because of the ID requirement

  • Data collected in polling stations shows that at least 0.25% of people who tried to vote at a polling station were not issued with a ballot paper because of the ID requirement. At least 0.7% of people who tried to vote at a polling station were initially turned away but around two-thirds of those people (63%) returned later in the day and were able to vote. 
  • 0.25% of polling station voters at these elections is approximately 14,000 voters who were not issued with a ballot paper because they could not show an accepted form of ID.
  • However, this is an underestimate, partly as a result of data quality issues but also because some people will have been reminded of the ID requirement before they could be recorded in the data. 

Around 4% of all non-voters said they didn’t vote because of the voter ID requirement

  • Some people who would have wanted to vote at a polling station may have decided not to try, because they realised they did not have accepted ID before attempting to vote. To understand the broader impact of the voter ID requirement, we carried out a representative public opinion survey across the areas with elections in May. Our survey asked people if they voted in the elections and, if not, why they had chosen not to do so.
  • We found that 4% of people who said they did not vote in these elections gave an unprompted reason related to the ID rules – 3% said they did not have the necessary ID and 1% said they disagreed with the need to show ID. The proportion of non-voters giving an ID-related reason rose from 4% to 7% when survey respondents were selecting from a list of reasons. 
  • There is evidence that some people found it harder than others to show accepted voter ID, including disabled people and the unemployed. However, further data collection and analysis are needed to establish a clearer picture and we will include additional information in our full report in September.

The UK Government and the wider electoral community should work to improve the collection of data at polling stations

  • We recommend that the UK Government should ensure that Returning Officers are able to collect and report monitoring data on the impact of voter ID at future elections, and work is also needed to improve the quality and reliability of data collected at polling stations in future.

Levels of voter confidence and satisfaction were similar to previous elections

  • There were high levels of satisfaction with the process of voting, in line with previous comparable elections; 89% of polling station voters said they were very or fairly satisfied. 
  • Voters were significantly more likely than non-voters to say that voting is safe. 90% of all voters said voting in a polling station is safe compared to 79% of non-voters. These results are also similar to those recorded after previous comparable elections.

Awareness of the voter ID requirement

Parliament approved the legislation for the new voter ID requirement on 22 December 2022. We delivered a public awareness campaign about the requirement across England which ran from 9 January to polling day. 

Awareness of the need to bring ID to vote at a polling station was high

Our research found that immediately before polling day, 87% of people in England (excluding London, where there were no elections) were aware that they needed to show photo ID to vote at a polling station.1 In the research we carried out immediately following the poll, this was 92% of people in areas with elections.2 This was a significant increase on the 23% that were aware in December 2022, before the start of our public awareness campaign. 

 

In our post-election research, awareness was higher among older age groups with 98% of over-65s aware in May 2023 compared to 82% of 18 to 24-year-olds. There was also a difference between white respondents (93%) and those from Black and minority ethnic communities (82%). We will be looking at these results for specific ethnicities in more detail in our September report in order to explore any patterns.

These differences in awareness had become more marked closer to the poll. This may be related to the variable level of turnout among different demographic groups, with some being more engaged with the polls and therefore more likely to be interested in or to recall communications about the election as it approaches. 

Our research found that overall awareness of the requirement was significantly lower among people who said they did not have an accepted form of ID (74%) compared with those who did have ID (94%). We also found that those people who said they always or sometimes vote at local elections had higher levels of awareness (94%) compared with those who do not tend to vote (84%).

While overall awareness levels were high, some groups of people were significantly less likely to know about the requirement. This means that some people would not have known that they needed to show ID until they arrived at the polling station. Those people who did not have any of the accepted forms of ID would not have been able to obtain ID (for example the Voter Authority Certificate) on polling day itself.

 

Awareness and take-up of the Voter Authority Certificate was low

Voters who did not have one of the accepted forms of ID could apply for free voter ID (called a Voter Authority Certificate) from 16 January until 5pm on 25 April.

Research that we carried out in early 2023 found that: 

  • 4% of the population in Great Britain either did not have any photo ID (3%) or did not have ID where they thought the photo would be recognisable (1%) 
  • 1% said they did not know whether they had any of the forms of photo ID we asked about.

Based on these figures and expected levels of polling station turnout at local elections, we estimated that between 250,000 and 350,000 applications for a Voter Authority Certificate might be expected.

The overall number of certificates applied for before the deadline on 25 April (around 89,500) was low compared with these estimates of the number of people who might not have any other accepted ID.

In May 2023, awareness of the Voter Authority Certificate was relatively low among the overall population and those who said they did not already have photo ID (both at 57%). This means that just under half of people who did not have any other accepted ID did not know that they could have applied for a form of ID to enable them to vote on polling day. 

There is limited data available on the demographics of people applying for a Voter Authority Certificate. However, applicants were more likely to come from older age groups.

 

 

Our public opinion research found that 84% of those that applied for a Voter Authority Certificate said the application process was easy.3

Approximately 25,000 Voter Authority Certificates were used as a form of ID on 4 May. It is not clear why a significant number of people applied for a certificate but did not use it on polling day. We will continue to analyse the available evidence to identify any patterns that might help to explain this.

Impact of the voter ID requirement on polling day

Polling station staff at the ballot issuing desk collected data on the number of people that could not be issued with a ballot paper because they did not provide accepted ID. They also recorded if any of these people returned later, and whether they were then able to show accepted ID and vote. The figures in this report are based on data received from 226 of the 230 local authorities with elections in May.

At least 0.25% of people who tried to vote at a polling station in May 2023 were not able to because of the ID requirement

The data provided to the Commission indicates that at least 0.7% of people who tried to vote at a polling station in May were initially not issued with a ballot paper because they did not have an accepted form of ID. Just under two-thirds (63%) of these people returned later in the day with an accepted form of ID and were able to vote. By the close of poll, 0.25% of people who tried to vote in person had not been issued with a ballot paper.

The proportion of people who were not issued with a ballot paper varied across local authorities from 0.02% of those who tried to vote at a polling station through to 1.04%. We intend to publish detailed local authority data alongside our September report. We will also be looking at the patterns of higher and lower levels of people turned away across local authorities in order to identify any trends.

0.25% of polling station voters at these elections is approximately 14,000 voters who were not issued with a ballot paper because they could not show an accepted form of ID.

Returning Officers were required to separate out data for polling stations with and without staff acting as greeters. Where that data could be provided it shows that polling stations with greeters recorded that a smaller proportion of people were initially turned away (0.55%) compared to those without greeters (0.8%). Overall, the data shows that 38% of polling stations had a greeter.

Among those recorded as being turned away from a polling station, 68% had not brought any ID and 28% brought a type of ID that was not accepted. The remainder were refused a ballot paper where the photo on their ID was not recognisable, the polling station staff suspected the ID was forged or where the person had failed to answer the statutory questions which can be used to ask a voter to confirm their name and address.

Any voter showing ID could ask to have their identity checked in a private area of the polling station. Approximately 2,250 voters were recorded as having asked to have this check done in private.

Recommendation: Polling station staff should continue to collect data on the impact of voter ID at future elections

Analysing data from the first scheduled elections where this requirement has been in place can help to improve the experience of voters and polling station staff at future polls. There is no legal duty or explicit power for Returning Officers to report data for future elections until the next UK Parliamentary general election, which must be held by January 2025.

It would be a significant missed opportunity to learn and identify further improvements if Returning Officers did not collect and report data at the scheduled May 2024 elections, which will cover all areas of England and Wales.

The UK Government should ensure that Returning Officers are able to collect and report monitoring data on the impact of voter ID at future elections, including specifically at the scheduled May 2024 elections and at any UK Parliament by-elections held during 2023 and 2024.

The data from polling stations underestimates the impact of the new rules on voters for two reasons. 

Firstly, not all people who wished to vote in person will have got to the ballot issuing desk and been recorded by staff before realising they did not have accepted ID. For example they may have been put off by polling station staff who greeted voters and gave them information about the requirement before they got to the ballot issuing desk.

Secondly, some of the data returned to the Commission is incomplete or inaccurate. The most common issues we found are:

  • data returns being completed incorrectly, for example polling stations with more people returning to vote than were initially turned away
  • missing individual pieces of data, for example on the reason why a voter was turned away
  • missing returns from polling stations, for example where some stations in a local authority did not submit a data return
  • blank returns from polling stations where it is unclear if a blank is equivalent to zero.

This was the first time data on photo ID needed to be captured. Issues with data quality and consistency were inevitable across some 20,000 polling stations and a larger number of staff. It is not possible to quantify the level of inaccuracy in the data but, overall, these types of errors would result in an underestimate of the impact of the ID requirement.

Recommendation: The electoral community should work to improve the collection of data at polling stations for future elections

Accurate data recording and reporting is essential to ensure there is a clear and reliable picture of the impact of the policy. This is needed to support informed debate and help identify areas for improvement.

However, initial feedback from Returning Officers, polling station staff and observers suggests that some polling station staff found it difficult to record this additional data.

We will work with the UK Government, Returning Officers and electoral administrators to review the data collection forms and guidance for polling station staff so they are clear and user-friendly.

We will also work with local authorities and their training providers to make sure the ballot paper issuing procedure and data recording processes are fully explained to polling station staff. This should include guidance for polling station staff on the appropriate role of greeters and tellers.

Around 4% of all non-voters said they didn’t vote because of the voter ID requirement

Some people who would have wanted to vote at a polling station may have decided not to try, because they realised they did not have accepted ID before attempting to vote. To understand the broader impact of the voter ID requirement, we carried out a representative public opinion survey across the areas with elections in May. Our survey asked people if they voted in the elections and, if not, why they had chosen not to do so. 

We found that 4% of the people who said they did not vote gave an unprompted reason that was related to the ID rules – 3% said they did not have the necessary ID and 1% said they disagreed with the need to show ID. Overall, the most common reasons given for not voting were a lack of time (15%) or a lack of interest (13%).

We also asked all non-voters a further prompted question (where they were given a number of answers to choose from) to understand whether the ID requirements had played a role in their decision not to vote. In response, 7% of non-voters indicated that the ID requirement was the reason they did not vote.

 

Of those who told us they voted in a polling station, 98% said they were able to vote the first time they went. 2% said they brought either no ID or the wrong type of ID the first time but later returned.

There are challenges in using these estimates to give an accurate, absolute number of people who were unable or chose not to vote in May.

Survey data is an estimate which can be subject to two separate errors. One is sampling error. This is limited by the robust, representative sample we have used particularly for the population as a whole but it can be more of an issue for sub-groups of the population (such as non-voters).

The second issue is non-sampling error. This includes respondents not answering a question and/or not answering it accurately. We can see one of the impacts of this in the difference between the unprompted and prompted question responses, where respondents may be less likely to give an unprompted answer (even if it would be true) and more likely to give a prompted one (where choosing from a list can invite a response which may or may not be true).

The public opinion survey findings are therefore most useful for giving us a more rounded view of the likely impact of the ID requirement than the data from polling stations provides alone. The figures support our assessment that the polling station data underestimated the impact.

Data on overall levels of turnout in May is not yet available and will be included in our full election report in September. This full report will look at how turnout in 2023 compares to previous local elections so we can see if it is possible to attribute any impact to the ID requirement.

There is evidence that some people found it harder to show accepted voter ID

We know from our previous research that some people were less likely to have photo ID and would therefore have faced more of a barrier to voting than those with accepted ID. However, it was not possible to capture reliable demographic data on people who were not able to vote because of the ID requirement; electoral registers do not record any demographic information, and electoral law did not allow polling station staff to collect demographic information about individuals who were turned away.

The small overall proportion of people who said they did not vote because of the ID requirement also means that it is difficult to say definitively, from our public opinion survey, if the ID rules were more likely to deter specific groups from voting than others.

However, the public opinion research does suggest that disabled people and those who are unemployed were more likely than other groups to give a reason related to ID for not voting. As set out above, our research on levels of awareness of the ID requirement also showed variation across different groups.

Where the relevant data is available, from a few local authorities, there is some correlation between the numbers turned away and specific socio-demographic factors, including ethnicity and unemployment. This evidence is also consistent with the findings from the 2018 and 2019 voter ID pilot schemes.

We also know that other organisations, such as the accredited observer organisation Democracy Volunteers, highlighted concerns that there was a disproportionate impact for some groups of voters at the May 2023 elections.

Further data collection and analysis are needed to be able to explore these and other concerns; we will include additional information in our full report in September. This will include consideration of information and data provided by charities and civil society organisations.

The majority of people were able to use their preferred method of voting

One potential impact of the new ID rules could have been that more people chose to vote by post instead of at a polling station. We do not know if that was the case, as data on levels of postal voting at these elections is not yet available. We will include an assessment of any change in our full election report in September.

However, in our public opinion survey we did ask voters if they had voted using their preferred method. The vast majority (95%) said they had.

Views on the voter ID requirement

After each election we ask people who were eligible to vote for their views on voting and elections. This helps us understand if views have changed since the last comparable set of elections.

Levels of voter confidence and satisfaction were similar to previous elections

Our research with the public found that more than two thirds of people (68%) were confident that the May elections were well run. This is slightly higher than the level recorded after the last set of comparable elections in 2019 (64%) but slightly lower than the most recent elections in 2022 (73%). We asked respondents why they did or did not feel confident and it is clear that the ID requirement had positive and negative impacts:

  • among those who said they were not confident, the most common reason selected (by 46%) was that “some people were unable to vote due to the ID requirement”.
  • among those who said they were confident, 30% selected the reasons that the requirement to show ID had “reassured me about the safety of the voting system”.

There were high levels of satisfaction with the process of voting, in line with previous comparable elections; 89% of polling station voters said they were very or fairly satisfied.

Satisfaction with the process of voting varied across different demographic groups, with the highest seen among the oldest age groups. We cannot make any link between this variation and the new ID rules, however, because the patterns are in line with the results we have seen after previous polls.

We also asked voters to rate the ease of participation in the poll on a scale from one to five. Nearly nine in ten polling station voters (88%) rated it as easy (four or five). Again, there is clear variation in the level of ease reported by different groups but no indication that this is different to previous polls.

Voters thought the elections were safe and secure

Our survey of the public asked everyone who said they had voted to rate the security of voting in the election on a scale from one to five. 84% of polling station voters rated it as secure (four or five). We also asked polling station voters how safe they felt polling station voting was from fraud – 94% said it was safe.

Voters were significantly more likely than non-voters to say that voting is safe. 90% of all voters said voting in a polling station is safe compared to 79% of non-voters.

These results are similar to those recorded after previous comparable elections and should not be directly linked to the introduction of the voter ID requirement.

In our full report in September we will explore further the demographic differences we see in the data on public attitudes.

No cases of personation in polling stations were reported to police

Early data from 26 of the 37 police forces in England where elections took place in May 2023 shows that no allegations of polling station personation fraud were reported following polling day. Police forces identified one public order incident relating to the voter ID requirement that led to an arrest. This was where a person who could not show accepted ID was arrested after allegedly assaulting a member of polling station staff.

We will provide updated data on allegations in our September report.

Next steps

As highlighted throughout this report, there are areas where we have more work to do in order to understand the impact of the new voter ID requirement, and which will be covered in our September election report. These include how the implementation of voter ID:

  • may have differently affected people across society who wanted to vote and what actions can be taken ahead of the next set of elections to address these issues. This includes looking at varying levels of awareness of the ID requirement, patterns in the numbers of people being turned away from polling stations and demographic / socio-economic differences in the data we have collected through our public opinion research
  • impacted on the administration of the polls. This will incorporate views from Returning Officers and their teams responsible for planning the elections, as well as the polling station staff responsible for managing the processes on 4 May
  • changed how parties and candidates approached their campaigns for these elections

Background

On 4 May, local elections took place in 230 councils in England, covering almost 5,000 individual wards. The areas holding elections included:

  • 152 district councils (out of 164 in total) covering largely rural areas and cities or larger towns in areas that also have county councils
  • 46 unitary authorities (out of 62 in total) covering cities or larger towns in areas without county councils
  • 32 metropolitan district councils (out of 36) covering predominantly urban areas in Greater Manchester, Merseyside, South Yorkshire, Tyne and Wear, the West Midlands and West Yorkshire
  • Local mayoral elections in Bedford, Leicester, Mansfield, and Middlesborough.

Scheduled elections did not take place in several areas of England in May 2023, including both London and Birmingham.

The Elections Act 2022 introduced a new requirement for voters to show an accepted form of photo ID to vote in person at a polling station for certain types of elections in Great Britain.

The requirement applies at local elections in England, Police and Crime Commissioner elections in England and Wales, UK Parliamentary by-elections and recall petitions. From October 2023, it will also apply to UK Parliamentary general elections.

Page history

First published: 19 June 2023

Last updated: