Completeness in Northern Ireland

There was a six percentage point difference in completeness between rural areas (76%) and urban areas (70%).

Our research has always found a connection between home movement and completeness; as registration is residence-based, greater mobility is associated with lower levels of completeness, while the longer an individual has been resident at their property, the more likely they are to appear correctly on the electoral register.

Completeness stands at just 11% among those who have been resident at their property for less than one year, rising to 29% of those who have lived at their address for one to two years. The absence of an annual canvass goes some way to explaining this (the last canvass in Northern Ireland was carried out in 2013), especially when compared to the equivalent figures in Great Britain (36% completeness for those resident less than one year and 71% for those resident between 1 and 2 years).

Completeness rises among longer-term residents with 61% of those resident for two to five years registered and 78% among those resident for more than five years and up to 10 years. For those living at their address for 10 to 16 years completeness was 80% and for those resident for more than 16 years it was 90%.

These figures are also similar to those we found in our 2012 study, at a similar point after the 2006 canvass, when completeness was 7% for those resident less than a year, 34% for those resident between one and two years and 59% for those resident between two and five years.

The return to 2012 levels of completeness emphasises the importance of improving mechanisms for capturing population movement in between each canvass to ensure that the improvements resulting from a canvass can be maintained afterwards.

This pattern, relating to length of residence, is consistent throughout our research. This is why we want to see changes to the registration system to enable the Chief Electoral Officer to access national-level public data, such as that held by the Northern Ireland Housing Executive, which could enable easier identification of when there has been change at a property, allowing the CEO to focus resources on addresses where a change has occurred.

We also want to see greater integration of electoral registration into other public sector transactions, such as applying for or updating a drivers’ licence, to encourage voters to update their electoral registration as part of another administrative task associated with moving house.

Tenure is a variable that has previously been strongly associated with levels of completeness and this research reconfirms this finding.

Homeowners are more likely to be registered than people in other types of tenure.

Completeness has declined in all tenure groups, reflecting the broader trend of decreasing completeness in the register. However, this fall has been greater among some of those groups who were less likely to have a complete register entry in the first place, such as a twelve point drop for private renters (from 50% to 38%), eight points for those owning with a mortgage (from 80% to 72%), six points for social renters (70% to 64%) and five points for those who live in a home that is owned outright (93% to 88%).

The figures for 2018 are, again, similar to those observed in 2012 where completeness was 88% for outright owners, 78% for households buying with a mortgage, 65% for social renters and 26% for private renters.

The lower levels of completeness among private renters is associated with population mobility as private renters tend to move more frequently than those in other tenures. For example, the 2016-17 Family Resources Survey shows that 25% of private renters in Northern Ireland have been resident at their address for less than a year compared to 6% of social renters, 8% of owners with a mortgage and 1% of outright owners.

Levels of completeness varied by age group in December 2018, with older groups more likely to be registered.  This is a finding that has been consistently found in our research across the UK.

In Northern Ireland, the highest level of completeness is seen for those aged 65+ (94%) and the lowest level is recorded for attainers (16 and 17 year olds, 0%), which has dropped from 8% in 2015.

In Northern Ireland the age at which citizens become entitled to vote is 18, but the electoral registers also include records of ‘attainers’ - 16 and 17 year olds who will turn 18 within the twelve month period starting on the 1 December after they make their application.

It should be noted, however, that the sample size for attainers in Northern Ireland is low from study to study so we often see fluctuation in the percentage estimates.

Nevertheless, administrative data in the table below shows the number of attainers on the register in Northern Ireland from 2013 to 2018. In 2018, there were only 574 attainers on the electoral register in Northern Ireland compared to 5,350 in December 2015.

The schools programme previously run by EONI was an effective way to engage with potential attainers but was resource intensive. The availability of online registration is likely to provide the potential for alternative solutions to be explored that would not necessarily require EONI staff to visit schools, as pupils can now be encouraged to register online through other mechanisms.

We understand that Chief Electoral Officer is working with the Education Authority to explore ways to encourage registration in schools without the full programme of visiting all schools.

By taking advantage of the C2k network it should be possible to engage with attainers in their school environment by providing them with the tools and information to register to vote. In doing so this will help to embed electoral registration into the school day.

The Electoral Commission has previously recommended that the registration of attainers presents an opportunity for the development of a more automatic approach to registration, for example when individuals receive their National Insurance number (NiNo).

NiNos are allocated automatically just before eligible people turn 16, and a letter containing details of the NiNo are sent to individuals at the address held by Her Majesty’s Revenue and Customs.

This information could be shared with the Chief Electoral Officer enabling 16 year olds who have been issued with a NiNo to be added to electoral registers as attainers (who are not able to vote in elections until they are 18), provided that the Chief Electoral Officer is satisfied that the individual is eligible and meets the residence requirements for registration. The falling number of attainers presented above supports exploring such a reform.

This chart shows that there have been reductions in completeness across several age groups. Completeness for 25-34 year olds is 50% (from 64% in 2015), among 20-24 year olds it is 61%, (from 71% in 2015) and it is 31% among 18-19 year olds (from 51% in 2015).

The 35-44 age group is closest to the overall average of 73%, on 70% (79% in 2015). Above this point all age groups have broadly higher levels of completeness, although there have been declines since 2015 (81% for those aged 45-54, down from 87% in 2015; 85% for those aged 55-64, down from 91% in 2015).

While past studies have noted that women are more likely to be registered than men, in this study no statistically significant difference has been found by gender. Completeness is 72% for men and 73% for women.

People with a long standing physical condition or disability are more likely to be registered (83%) than those without a disability (72%) or those with a longstanding mental condition or disability (71%).

Levels of completeness are associated with social groupings. The main division is between those in social grades AB and other groups.

While completeness stands at 80% for AB households, it is in the mid-seventies for social grades C1 and C2 (72% and 76% respectively). Since 2015, completeness for those in DE households has declined considerably (from 77% to 63%).

Previously, people in this group were as likely to have complete register entries as those in social grade C2, but there is now a thirteen percentage-point difference.

Completeness is highest among those with degrees or further qualifications such as BTECs, as well as those with other types of qualifications or none at all (perhaps related to age). This reflects the pattern observed in 2015, when this question was first asked.

The groups least likely to have complete register entries are those with A Levels or Scottish Highers (65%) or GCSEs (68%) as their highest form of academic qualification.

Single adult households remain the least likely to have complete register entries. Completeness for this household type was 72% compared to 74% in 2015, though this difference was not statistically significant.

However, the difference between this type of household and those with more adults has shrunk since 2015.

Completeness for two and three to five-adult households stands at 73% and 72% respectively, declining from 80% for both in 2015.

Households with six or more adults are also more likely to have complete register entries – however the base size for this group is small and not suitable for drawing representative conclusions.

The lead participant in each household interview was also asked some questions around their attitudes to registering to vote and voting itself. While this means it is not a representative sample of the population, the results provide another angle to consider the relationship between registration and attitudes towards elections.

Those who feel it is everyone’s duty to register to vote are most likely to have a complete electoral register entry (77%).

There have been shifts in other attitudes since 2015, most notably a fall in completeness among those who say it is not worth registering to vote (from 70% to 54%).

There has also been an increase in completeness among those who believe that people should only register if they care who wins an election, from 58% in 2015 to 68% in 2018.

Attitudes towards voting show a similar pattern, with those who feel it’s everyone’s duty to vote having the highest level of completeness (77%) compared to those who feel it’s not worth voting (62%) and those who say people should only vote if they care who wins an election (64%).

Although the distance between the December 2018 electoral registers and the most recent UK-wide election (the 2017 UK Parliamentary general election) is greater than it was for the December 2015 registers, the same relationship can be observed.

Those who reported voting at the most recent UK general election showed higher levels of completeness: in 2015 the level was 91%, and in 2018 it was 82%.

Completeness was lower among those who said they did not vote (although eligible to do so), at 49%.  The equivalent figure in Great Britain was 72% which is further evidence that an annual canvass helps to encourage people who are usually disengaged to register to vote.

The 2016 referendum on the UK’s membership of the European Union was notable for its higher turnout compared with the UK Parliamentary general election which preceded it.

As this is the first study of accuracy and completeness since the referendum occurred, a question was asked to understand the relationship between participating in the vote and the completeness of the registers.

This reveals a very similar pattern to participation in the 2017 UK general election, with completeness highest among those who said they voted in the referendum (81%).

For those who did not vote completeness stood at 51%.

As was the case in relation to attitudes towards registration and voting above, only the lead participant in each household was asked questions about their political views.

This means that, again, the sample is not fully representative of the population, and these results are subject to notable margins of error and so the differences should be treated as indicative rather than absolute.

Those who identify as Nationalist (76%) and those that identify as Unionist (80%) both have higher levels of completeness than the average. The apparent difference in completeness between these groups is not statistically significant.

Completeness was lowest (65%) among those that said they had no political view (when given the option of identifying as unionist, nationalist, loyalist, republican, other or no political view).