1. Overview of our user engagement on national and subnational population projections

This article summarises responses to our user engagement exercise on the future timescales for national population projections (NPPs) and subnational population projections (SNPPs).

The user engagement exercise, which ran from 15 October to 12 November 2020, gave users the opportunity to provide feedback on how we might approach the publication and timing of the next set of NPPs and SNPPs.

NPPs are produced by the Office for National Statistics (ONS) on behalf of the National Statistician and Registrars General for each UK constituent country and for the UK as a whole.

SNPPs are produced separately by each constituent country of the UK:

  • the ONS for England
  • National Records of Scotland (NRS) for Scotland
  • Northern Ireland Statistics and Research Agency (NISRA) for Northern Ireland
  • Welsh Government (WG) for Wales

Household projections were not included in the scope of this user engagement exercise and we will provide an update on these separately.

We are grateful to everyone who responded to the user engagement. User feedback is helping us make a more informed decision on when we should produce the next set of projections.

Nôl i'r tabl cynnwys

2. Background to user engagement

In the 2018-based national population projections (NPPs) and subnational population projections (SNPPs), we indicated a proposal not to produce 2020-based projections and that the next set of projections would be based on 2021 Census data. This would enable us to produce more accurate projections using updated base populations from the 2021 Census and a revised back-series of earlier years of data. Since then, we recognised growing interest in how projections might be impacted by the coronavirus (COVID-19) pandemic and in the announcement that Scotland’s census will now take place in March 2022.

We asked users:

  • for their preferences on timescales for future projections, taking into account potential availability of census data
  • how they would use the projections
  • whether they needed any variants

We also asked whether a shorter release would meet user needs and whether they would be happy to use a new set of projections that used trends (assumptions) from earlier releases. Lastly, we welcomed other general comments about the projections.

As outlined in our December 2020 update, a total of 65 responses were received from across the UK. These came from a variety of organisations including local authorities, county councils, central government, health organisations, local councillors, national parks, consultancies and community groups. We also received responses from several academics and individuals with an interest in projections.

Nôl i'r tabl cynnwys

3. Approach to assessing feedback

Responses were categorised into groups based on the intended use of the projections. Four groups were identified with this approach, consisting of those informing national level policy, those informing local level planning, academic research and other uses.

Table 1 shows the total number of responses for each of these user groups.

Nôl i'r tabl cynnwys

4. Summary of responses about use of projections and future timescales

This section presents a summary by the main questions asked in our invitation for feedback.

Users were asked for information about their use of the projections and to state their preferences around the timescales for future projections.

Informing national level policy

Nearly all users in this group (12 out of 14) stated a need for 2020-based projections, seven of which also requested 2021-based or 2022-based projections. The main reason given for needing 2020-based projections was that users required the data on a cyclical basis to inform national level planning decisions and that it would be too long to wait for 2021-based projections. Two respondents from this group stated that it would be preferable to wait for 2021-based projections (and not have 2020-based projections).

Informing local level planning

The majority of these respondents (19 out of 34) also stated a need for a set of 2020-based national population projections (NPPs) and subnational population projections (SNPPs), with eight users in this group also requesting 2021-based or 2022-based projections. As with users identified as informing national level policy, respondents in this group also stated that the main reason for requiring 2020-based projections was a pressing need for regular and accurate data to inform planning decisions. Fewer than half of these users (14) were in favour of waiting for a set of 2021-based projections.

Academic research

Among this user group there was no common preference on the release timetable. Just two respondents out of seven requested 2020-based projections, with just one of these users also requesting a set of 2022-based projections. Two respondents expressed a preference to wait for 2021-based projections.

Other users

There was a similarly mixed response among this group as for the academic research group. Three respondents out of seven expressed a need for 2020-based projections, two of these also requesting 2022-based projections. Two other users also stated a preference for a 2021-based set of projections. A further two users in this group requested 2020-based, 2021-based and 2022-based projections.

In response to these questions, respondents in Scotland noted a particular issue about the potential length of time until the results from the 2022 Census in Scotland would be reflected in the population projections.

Nôl i'r tabl cynnwys

5. Summary of responses about variant projections and benefits of a coronavirus (COVID-19) variant

Users were asked to state what they required from the projections, including specific variants and whether some form of separate coronavirus variant projections would be beneficial.

Across all user groups of our existing variant projections, migration variants were the most requested with 28 respondents expressing a need for these (requests associated with EU exit have been included in this total, since many users combined these in their feedback). There were also a high number of requests in all user groups for the principal projection (21) and several requests for life expectancy variants (11), fertility (9), all variants (6), age structure (5) and population variants (4).

Informing national level planning

The majority of respondents in this user group (11 out of 14) stated that a coronavirus variant would be useful. Only one user stated that it would not, and one further respondent did not specify a preference. One of these users commented that the impact of the pandemic and EU exit could be incorporated into assumptions for a number of different projection variants.

Informing local level policy

Most responses in this user group (24 out of 34) expressed the view that a coronavirus variant would be useful. Ten responses did not specify a preference, with some of those expressing the view that a separate variant was not necessary as its impact would be reflected in population trends or existing variants.

Academic research

Among academic users there was less support for a coronavirus variant, with just under half of users in this group (three out of seven) stating that a separate variant would be beneficial. The remaining four respondents either did not specify a preference or did not provide a response.

Other users

Half of responses in this group (5 out of 10) stated a coronavirus variant would be beneficial. One respondent was against this approach on the basis that such a variant would be difficult to calculate. Four respondents either did not specify a preference or did not provide a response.

Other variants

Some respondents from England requested variants that are not produced as part of the national population projections (NPPs) or subnational population projections (SNPPs) and were outside the scope of this user engagement. These included variants reflecting the impact of self-employment rates, marital status, changes to work patterns, economic growth, coronavirus infection rates and vaccine success, projections by ethnicity and a comparison between the 2011 and 2021 Censuses.

Nôl i'r tabl cynnwys

6. Summary of responses about publication of a shorter projections release

Users were asked for their views on whether a shorter release that contained only datasets and information on how the assumptions were set would meet their needs.

Informing national level policy

Nearly all respondents in this group (13 out of 14) stated that a shorter release would meet their needs. One respondent was more equivocal in their feedback, cautioning that a shorter release might be used by some users without a full appreciation of the inherent uncertainty in the data.

Informing local level policy

The majority of users in this group (27 out of 34) were agreeable to a shorter release, while three users stated that a full bulletin was important for their understanding of the data.

Academic research

Only one academic user out of seven respondents stated that a shorter bulletin would be acceptable. Three respondents cautioned against a shorter release on the grounds that an accompanying summary aided user understanding and minimised errors in the production of the projections.

Other users

Half of users in this group (5 out of 10) stated that a shorter release would be acceptable for their needs while just two respondents stated that an accompanying bulletin was needed.

Nôl i'r tabl cynnwys

7. Summary of responses about reusing 2018-based assumptions

Users were asked for their views on the potential reuse of 2018-based assumptions because of uncertainty over demographic trends arising from the coronavirus (COVID-19) pandemic.

Informing national level policy

Just over half of users in this category (8 out of 14) stated a preference for not reusing 2018-based assumptions and advocated the use of fully updated assumptions in new projections. Just four users were agreeable to reusing previous assumptions.

Informing local level policy

Just under a third of respondents in this user group (11 out of 34) were against the idea of reusing 2018-based assumptions. Some of these respondents stated that this would negatively impact local planning as well as damage the credibility of the projections. A small number of users (six) were supportive of reusing the assumptions.

Academic research

Two academic users out of seven were agreeable to the reuse of 2018-based assumptions. The same number were against this proposal, citing concerns about previous assumptions deviating from historical outcomes, which could risk devaluing decisions based on the projections.

Other

The majority of respondents in this user group (7 out of 10) were against reusing 2018-based assumptions with many expressing concerns that circumstances had changed significantly in the past two years with the impact of the coronavirus pandemic and EU exit.

Nôl i'r tabl cynnwys

8. Additional comments on population projections

Users were also invited to provide further comments on the national population projections (NPPs) and the subnational population projections (SNPPs).

A wide variety of answers were received in response to this question with some respondents expanding upon the answers they had already given, while others made specific requests or suggestions.

Informing national level policy

The comments provided by this user group generally consisted of requests for additional variants or data, such as, for a back series of projections data, breakdown of different ethnic groups and updated marital status projections. Others highlighted changes between 2020-based and 2022-based projections and real time estimates, along with feedback related to the timing of the projections publication to avoid clashing with the timing of users’ own analysis.

Informing local level policy

Feedback from this group was wide-ranging, such as requests for projections data as soon as possible, critique of the methods used in projections and highlighting the long-term implications of the pandemic on demographic trends. Some feedback was complimentary about the projections while others stressed the need for more information on future plans. This highlighted current strengths in our general approach to projections but also areas for review in order to further meet user needs.

Academic research

Comments from academic users referred to the importance of the 2021 Census for rebasing the projections to ensure the credibility and value of the data. Feedback also included exploring inter-relationships between trends in migration and fertility, including the possibility of projecting inequalities in mortality, which have been highlighted by the pandemic. There were also comments on our methodology from this group suggesting that this is important to review.

Other

Feedback from this group included requests for variants by ethnicity at the national level, for a finer geography than local authority and the 10-year migration variant to be unconstrained in the subnational population projections.

Nôl i'r tabl cynnwys

9. Overview of responses and conclusions

We would like to thank everyone who responded to the user engagement. The feedback from all respondents has been important in helping the Office for National Statistics (ONS) and devolved partners at National Records Scotland (NRS), Northern Ireland Statistics and Research Agency (NISRA) and the Welsh Government (WG) consider options on when to publish future rounds of projections.

The user engagement revealed a wide range of needs and expectations from respondents relating to both the content and timing of future national population projections (NPPs) and subnational population projections (SNPPs). There were particular concerns expressed about the impact of the coronavirus (COVID-19) on demographic trends and a need for a greater understanding of the effect of EU exit on migration.

Just over half of all respondents (36 out of 65) expressed a clear need for a set of 2020-based projections to facilitate planning decisions rather than waiting until 2022 for census data to be incorporated in forming a 2021-based set of projections.

Many respondents also raised concerns about the potential quality of producing projections at this stage due to the unknown effects of the coronavirus pandemic on population trends as well as its impact on the collection of data. Some users expressed a preference to wait until a better understanding of the effect of the pandemic has been developed in addition to being able to include data from the 2021 Census.

Two-thirds of respondents (43 out of 65) stated a strong preference for a separate coronavirus variant to be included in the next set of projections. This underlines the need among users for guidance and a greater understanding of the impact of the pandemic on demographic trends in order to be able to inform planning and policy decisions.

Across all groups, just under half of respondents (28 out of 65) were against reusing 2018-based assumptions in the next set of projections while only a fifth (14 out of 65) were agreeable to such a proposal. Opposition to the reuse of 2018-based assumptions was primarily due to concerns about the impact of coronavirus and EU exit on demographic trends with users stating that to not take into account such factors would erode confidence in both the quality of the data and decision-making based upon it.

The majority of respondents (across all user groups) stated that a shorter release containing just datasets and information about assumptions would meet their needs. This suggests that a scaled-back set of projections documentation might be adequate for the majority of users, although several respondents expressed a need for an accompanying bulletin and summary to aid their understanding of the data. Providing the right level of support and guidance that also refers to the impact of the coronavirus was considered to be important for the next set of projections.

Overall, the feedback from all users highlighted a clear need to strike the right balance between satisfying user needs and timeliness on the one hand, and the quality and completeness of the data on the other. Balancing these factors was central to the decision-making process, in combination with the timing of the 2021 Census.

Nôl i'r tabl cynnwys

10. Final decision

A final recommendation on the production of 2020-based National Population Projections (NPPs) was made by the National Population Projections Committee, comprising the Office for National Statistics (ONS), National Records Scotland (NRS), Northern Ireland Statistics and Research Agency (NISRA) and the Welsh Government (WG).

This recommendation was presented to the UK Census Committee (UKCC) for a final decision. The UKCC, chaired by the National Statistician, co-ordinates UK-wide statistical issues and reviews alternative approaches to meeting future user needs for population and small area socio-demographic statistics. Membership of UKCC includes the Chief Executive and Registrar General of NISRA and NRS and Chief Statistician of NRS and the WG.

The UKCC decided that, in order to meet user needs identified through this user engagement, a principal national population projection only will be published for each UK constituent country and for the UK as a whole, with no variant projections. This is in response to the coronavirus (COVID-19) pandemic and to support the forthcoming State Pension Age Review. We plan for these to be published in December 2021 and be titled ‘2020-based interim national population projections’ using an updated set of demographic assumptions.

The publication of these projections will include full details, guidance around their use and accompanying data. The use of the term “interim” in the publication title is to reflect the interval between the 2020-based principal projection and subsequent projections which will incorporate data relating to population statistics resulting from the 2021 Census. It also recognises this as a period of uncertainty in the mid-2020 base year and in setting long-term demographic assumptions following the coronavirus (COVID-19) pandemic.

This approach will contribute towards meeting user needs for statistics by:

  • using the most recent base year (mid-year 2020) population estimates
  • recognising the uncertainty arising from: using a base year at the furthest distance from the 2011 Census and impacted by the coronavirus; setting demographic assumptions while the full demographic impacts from the coronavirus continue to unfold; and risks around some input data caused by the pandemic

After the 2020-based NPPs, we propose the next round of NPPs be published in winter 2022 to 2023, using 2021 Census data and to include projection variants. At this stage, this is not a definitive policy, and we cannot be certain of exact timings. Provisional first results from the Scottish Census are due to be made available in spring 2023 and will be incorporated in a subsequent round of the NPPs.

Decisions on future timescales for the subnational population projections (SNPPs) and household projections will be taken separately by the ONS, NRS, NISRA and the WG. These will be informed by user feedback to this engagement exercise as well as local user engagements, which may have taken place in each country.

Nôl i'r tabl cynnwys

Manylion cyswllt ar gyfer y Erthygl

James Robards
pop.info@ons.gov.uk
Ffôn: +44 (0)1329 444661