Changes to the way in which estimates of international migration are produced mean that NRS will start to regularly revise our population estimates. Further information on the reasons and our planned approach can be found below.
Summary of background and planned approach to revisions
- Mid-year population estimates, and other statistics produced by National Records of Scotland (NRS) make use of estimates of Long-Term International Migration (LTIM) produced by the Office for National Statistics (ONS).
- Recent changes to the data sources and methodology used by the ONS mean that they will now produce and release multiple estimates of LTIM for a given period, each based on progressively more complete data.
- To ensure a balance between releasing timely statistics and using the most accurate data, NRS plan to revise each set of population estimates the following year alongside the next annual release.
- For example, our mid-2023 population estimates, which were first published in October 2024, will be revised in summer 2025 as part of our mid-2024 population estimate publication. And our mid-2024 population estimates will be published in summer 2025 and will then be revised when the mid-2025 population estimates are published in summer 2026.
- A similar approach will be taken for Small Area Population Estimates (SAPE).
This is different to our current approach which generally involves revisions being limited to a rebasing exercise after each census. This is because in the past international migration estimates tended not to be regularly revised.
- Other revisions to our population estimates may also occasionally be necessary when improvements to the method of estimating international migration are introduced. We plan to limit such revisions to mid-2023 onwards.
Outline of our approach for planned revisions
Mid-year population estimate methodology
Mid-year population estimates are produced each year by NRS. They are based on the census with data on births, deaths and migration then being used to provide updated population estimates for each year.
Data on births and deaths are based on registrations and are considered complete. There is no equivalent data source on migration and so these figures have to be estimated using various datasets that have usually been collected for other purposes.
For migration within Scotland or between Scotland and other parts of the UK the main data source is GP registrations. For international migration other sources are used.
Long-term international migration
An international migrant is defined as someone who changes their country of residence for 12 months or more. Estimates of LTIM for both the UK and Scotland are produced by the ONS.
In the past, estimates of international migration were based primarily on the International Passenger Survey (IPS) which recorded information on people’s intentions.
It had long been acknowledged that the IPS had been stretched beyond its original purpose. This, along with issues caused by the Covid-19 pandemic, led ONS to change the data sources used to estimate international migration. Since 2021, estimates of LTIM have been based primarily on administrative data sources including data from the Home Office and the Department of Work and Pensions.
Using this new approach ONS now release provisional LTIM estimates for the UK around six months after the period to which they refer. At that time, it is not known whether people moving to or from the UK will stay for 12 months or more (in order to be considered an international migrant). ONS make adjustments to estimate how many will stay or leave for 12 months or more.
As time passes, more complete data becomes available. ONS publish a first revision of UK level estimates around 11 months after the period to which they refer and final (or ‘mature’) estimates around 17 months after the period to which they refer. So, for example, LTIM estimates for the year to mid-2024 should be finalised by November 2025. These published estimates then need to be broken down by ONS to provide figures for Scotland and council areas.
Further details on the timing of LTIM publications and revisions can be found in the February 2025 ONS article: Publication schedule for admin-based population and migration statistics - Office for National Statistics.
Consequences for population estimates for Scotland
To produce timely mid-year population estimates for Scotland, NRS needs to use LTIM data before it has been finalised. This allows us to publish population estimates for a given year the following summer (e.g. we plan to publish the first set of estimates for mid-2024 in summer 2025).
Our plan is to then revise these population estimates when we publish the next set one year later (e.g. June / July 2026) in order to incorporate the ‘mature’ LTIM estimates. (Note that we do not intend to make use of ONS’ “first revision” estimates which will be released in the meantime).
This means that population estimates for Scotland for a given period will initially be released around one year after the mid-year point to which they refer and will then be finalised around two years after the mid-year point to which they refer.
Similar revisions will also be required for our SAPE which are based on many of the same data sources.
Other revisions to earlier population estimates
The method used to estimate international migration is still developing. This means that in addition to the regular revisions to LTIM estimates described above, ONS may also make other changes as they introduce improved methods or new data sources. These are not limited to the most recent years of data but would be made to all years back to 2021-22.
Such changes will be assessed by NRS statisticians on a case-by-case basis and a decision will be made as to whether the change warrants further revisions to earlier population estimates. If such revisions are deemed necessary, users will be alerted as soon as possible by ScotStat e-mail. Our current position is that if such changes were necessary we would apply them back to 2022-23 only. This would avoid the need for further revisions to rebased estimates between 2011 and 2022.
We anticipate that this sort of revision will become less frequent and smaller in magnitude in the future as the LTIM methodology becomes more stable. ONS have said that their plan is that from autumn 2026 they will only open the back series for revisions every three years and then, once their methods have matured, only every five years.
