Compliance Check of Household Projections for England

Dear Iain

HOUSEHOLD PROJECTIONS FOR ENGLAND

As you are aware, we recently completed our check of the compliance with the Code of Practice for Statistics of ONS’s Household Projections for England. I am pleased to confirm that these projections should continue to be designated as National Statistics subject to addressing the findings from this review.

We initiated this check following our public commitment in March to review these projections. This was in light of the transfer of responsibilities for producing the projections from the Ministry of Housing, Communities and Local Government (MHCLG) to ONS in early 2017; and in light of the recent methods changes introduced by ONS following public consultation. We considered the Trustworthiness, Quality and Value of the projections in relation to the Code.

We’ve appreciated the positive and constructive way that the team has engaged with us during the development of these projections and our compliance check.

We found many positive examples around ONS’s production and presentation of the latest (2016-based) Household Projections for England. These include:

ONS publicly consulting users on its proposals to update the methods used to produce these projections, and seeking users’ feedback about what additional information would support the projections’ interpretation and use.
Setting up a research programme to explore options in response to the consultation, and establishing a Household Projections Collaborative Group, comprised of independent experts, to advise ONS on its research. ONS has signalled its intention to continue to seek advice from this group for its future developments, which we welcome.
ONS’s openness and transparency about its proposals, how it responded to users’ feedback, and publishing detailed guidance on the methods used, the quality of, and quality assurance of the projections.
The main statistics report presents the key messages clearly and impartially, with clear signposting to supporting guidance.
ONS taking steps to assess and publish information about the accuracy of the projections, comparing them with previous projections and other sources of household estimates, and explaining the likely reasons for differences, to help inform future developments.

Overall, we welcome the team’s proactive and open approach to develop these projections.

We identified some areas where we consider that ONS could enhance the clarity and insight presented by these projections, and improve the accessibility of key assumptions used, in order to meet the high standards required of National Statistics:

The team should review and refine the range of methods guidance published to enhance its accessibility to users. Specifically, we consider that ONS could enhance users’ understanding of quality by presenting the key assumptions used in the 2016-based projections, how they compare to previous assumptions, along with clear justification about their choice. This should be clearly accessible from the main statistics report.
In light of lower population projections and methodology changes including ONS’s decision to base the Household Representative Rates used to produce the projections on two Census points – 2001 and 2011, there is a noticeable difference between the 2016-based projections compared to the previous 2014-based projections. This difference is evident both overall and at Local Authority-level. ONS should investigate Local Authority-level projection outliers, and provide supplementary analysis and guidance about the key drivers for these results. ONS should also continue to monitor users’ feedback to identify what other analyses would help to aid users’ interpretation.
ONS had originally planned to produce and publish additional ‘variant projections’ alongside its Stage 2 projections [1] published on 3 December. The team told us when it started this planned work, it became apparent that user needs were varied and more time was needed to ensure that additional variants would best meet these needs. Due to the importance of these projections, we encourage ONS to prioritise publishing these additional ‘variant projections’ in the near future, in a way that is most helpful to users.
We welcome that you have started to, in collaboration with others across the UK, revisit and update published guidance about the coherence and comparability of projections produced for each of the four countries of the UK. We look forward to these outputs being published in the near future.

We have presented these findings in more detail in Annex 1 of this letter.

Thank you for engaging effectively with us during this review. I would welcome an update from you by end June 2019 about how you have addressed these areas. Please let me know if there is any aspect of this letter that you wish to discuss.

I am copying this letter to Rich Pereira, ONS Deputy Director Centre for Ageing and Demography, Sarah Crofts Head of the ONS Centre for Ageing and Demography and Joanna Harkrader, responsible Statistician for Household Projections.

Yours Sincerely,

Mark Pont
Assessment Programme Lead

[1] ONS published the 2016-based projections in 2 stages: Stage 1 presents projected numbers of households in England by age and sex of the Household Reference Person; Stage 2 presents projected numbers of households by household type.