Since accelerating the move from travel survey findings to administrative data gathered by government bodies, the Office for National Statistics (ONS) has been continuously improving the way it estimates long term international migration. This transformation and continued development of methods has contributed to the revisions seen in recent times. Here Mary Gregory updates on the latest work to provide a more detailed picture of EU and British international migration.
Historically the ONS used the International Passenger Survey (IPS) to estimate flows of people into and out of the country. Whether they were a student from China, a worker from Poland or a British national leaving the country for a new life in Spain, we’d estimate international net migration using the IPS and what people told us about their intentions for future travel.
However, we’ve long said this sample survey carried out at various ports and airports around the UK was stretched beyond its original purpose as a tourism survey. It is harder to get a robust sample for migration and it couldn’t capture changes to people’s intentions.
Since the pandemic – when the IPS was paused – the ONS has increased its use of administrative data, such as visa data and tax information, to improve estimates of long-term international migration.
At a high level, we currently estimate migration for EU+, non-EU+ and British nationals [EU+ includes EU countries plus Norway, Iceland, Liechtenstein and Switzerland]. However, no single source of admin data captures all three together, so we use different sources of information for each group.
A decade ago, migration to the UK was driven by arrivals from the EU who didn’t need a visa, but that has changed following the new post-Brexit immigration rules. Our latest data shows around 86% of immigration comes from outside of the EU.
Non-EU
For estimating non-EU migration, we use Home Office visa data. We’ve been learning more about the data sources at our disposal and have been updating our estimates as our understanding has improved and we’ve got more complete travel information available, as my blog in November explains.
We’re confident in these methods and believe the granularity of detail they offer add real value when understanding a migrant’s journey. Rather than survey-based intentions, we can now see actual behaviours. For instance, we now know when someone moves from one visa to another [a student moving to a work visa, for example] and we have more information on main applicants and their dependents. This sort of detail is of vital importance for policy makers.
EU migration
We currently use the Department for Work and Pension’s Registration and Population Interactions Database (RAPID) to estimate EU migration by identifying interactions within the tax and benefit system to determine if someone is in or out of the country.
However, there are limitations to this method – including timeliness as there is only one cut of data each year. Because we use interactions with the tax and benefits system to infer an individual’s presence in the UK, it is possible that some of these interactions occur while the individual is abroad (for example, some people stayed in UK employment but worked abroad during the COVID-19 pandemic). It also excludes some populations who don’t interact with the tax and benefits system. While we make adjustments to account for these limitations, there may be instances where we assume people are not in the country and emigrate them erroneously (or don’t immigrate them).
Since January 2021, EU nationals migrating to the UK have required a visa and EU nationals already living in the UK needed to apply for the EU Settlement Scheme (EUSS). This has enabled ONS to start development of an EU long-term migration estimation method using Home Office Borders and Immigration (HOBI) data, which combines visa and travel information, for the first time. This is the same data source we already use for estimating the migration of non-EU nationals.
Unlike RAPID, which is a stock of the population, HOBI estimates flows and is more timely, with quarterly updates. This data source should include international students whether or not they work, and it also includes EU nationals of all ages coming to and from the UK, regardless of their interactions with the tax and benefits systems.
We believe this method will not only give greater and more timely coverage of EU migration but, like non-EU migration, over time it will allow us to gather greater insights into movements. For instance, we’ll have a more accurate picture of reason for migration and nationality for the first time, as well as information on transitioning from one visa to another.
However, it is not without its complexities. For example, we know there are data challenges around the movements of Irish nationals, while there are also coverage issues when looking at travel data, especially through sea and rail routes.
British migrants
The smallest part of the equation is migration of British nationals – that’s British nationals leaving the country or returning after a period abroad, for example after working or studying in another country for several years.
In the absence of an alternative, we’ve continued to use the IPS to measure British migration. However, the IPS arrivals data collection stopped on July 1, 2024, and no longer measures arrivals of British nationals.
To improve our estimates of British national migration, we’ve explored alternative data sources. Currently RAPID is the best available alternative in terms of timeliness and coverage with qualitive improvements in data richness. We are also exploring Home Office Border Data, including the advanced passenger information of British nationals, for potential future use in our estimates.
When to make changes
For the next long term international migration publication in May 2025, we will continue to use methods we used in November 2024 (rolling forward IPS data for British nationals).
Our ambition is to move to using the new methods for estimating migration of EU and British nationals, but we’ll only do so once we’re confident they’re of sufficient quality and we’re confident the methods are robust. As I have made clear earlier in this blog, EU and British national migrants make up a relatively small part of the overall net migration picture. However, any updates to the methods will lead to some revisions of previous estimates.
Revisions
Revisions are an inherent part of our population and migration system. There will always be some revisions as we aim to publish timely estimates which we update as we get new data covering the full 12-month period. This enables us to determine whether someone has stayed in or out of the UK long enough to be considered a migrant. We also have revisions because of method changes like those outlined above and these can lead to sizeable updates, as we saw in November.
Given the importance of decisions made using the migration and population estimates – and the implications of any revisions – we are very conscious of the importance of balancing timeliness, stability and accuracy. As our data and methods mature, we are moving to a more stable rhythm, our proposals are outlined in this paper also published today.
This greater maturity in our approach is one part of our criteria towards moving to Admin Based Population Estimates as our official population estimates.
It is our intention that from mid-2026, the published back series for population estimates will only be open for revisions at fixed intervals that are communicated to users. We are also conscious this approach must be flexible. If we make a significant revision to our migration back series, for instance, we would have to incorporate that into our next population estimates.
Gathering feedback
We are continuously trying to improve our migration and population statistics, including how we gather stakeholder feedback.
To that end, we have reviewed our approach to engagement with key groups interested in migration and population estimates. We are setting up an Advisory Panel on Migration which will provide independent advice to the National Statistician on the quality and utility of migration estimates. I look forward to working with this group in due course.
In the meantime, if you have feedback on any aspect of our work on population and migration statistics, please email pop.info@ons.gov.uk.
Mary Gregory – director of population statistics