Not Settled Yet? Understanding the EU Settlement Scheme using the Available Data

1. Introduction

As the UK prepares to end free movement, EU citizens already living in the UK are required to apply to the UK Home Office’s EU Settlement Scheme (EUSS) in order to retain their residence rights. The basic outlines of the scheme are explained in the Understanding the Policy section above.

By the end of February 2020, the Home Office had received more than 3.3 million applications to the scheme. Because there is no registration system or population register in the UK, the government does not know which UK residents are EU citizens. EU citizens will thus need to come forward of their own accord.

For a host of reasons ranging from lack of information to poor organisation or vulnerability, it is likely that at least some EU citizens will fail to come forward by the deadline, regardless of how well the scheme is designed and communicated (Sumption and Kone, 2018). The number of people who will not apply is not possible to predict, and under current government policy, the default position is that they will become unlawful or irregular migrants.

This has created considerable interest in statistics on how many people are taking up the EU settlement scheme. In addition to understanding its coverage, stakeholders have raised a range of other questions about the performance of the EU Settlement Scheme, which can be broadly grouped into three main categories:

• Take-up: who is applying and not applying to the Settlement Scheme? What share of EU citizens eligible to apply has done so, and which groups are more vs. less likely to have done so (e.g. based on characteristics such as age, gender, ethnicity, location within the UK or vulnerability)?
• Are EU citizens getting the ‘right’ status (settled vs. pre-settled)? Are applicants who have lived in the UK for at least five years receiving pre-settled status rather than settled status? As the scheme progresses, will people ‘convert’ pre-settled status to settled status in the numbers expected?
• How smoothly is the scheme functioning? How long does it take for applications to be processed, do applications from some groups take longer than others? What are the reasons for not being granted status or for experiencing longer delays? How many decisions are challenged (e.g. challenging a grant of pre-settled vs. settled status, or challenging a refusal on eligibility or suitability grounds) and are appeals usually successful?

Although the Home Office has published significant volumes of data about EUSS applications—much more than it does for other immigration routes—it is very hard to use the data to answer these questions. This report explains why, assesses what we do and don’t know, and looks at what it would take to get clearer statistical answers to questions about how EUSS is performing.

2. Take-up: what share of EU citizens has applied to the Settlement Scheme and are some groups less likely to apply?

A key question for understanding the Settlement Scheme is what share of eligible people have applied and how many are yet to apply. However, there is a major barrier to calculating this, which is that the number of eligible applicants is simply not known. This fact is well recognised by the Home Office and Office for National Statistics, whose publications have cautioned readers against making a straightforward comparison between Home Office figures and official estimates of the EU citizen population living in the UK (Home Office, 2020; ONS, 2020a).

Why don’t we know how many people are eligible for the EU Settlement Scheme?

The Office for National Statistics (ONS) estimates that there were 3,725,000 EU citizens living in the UK in the year ending June 2019, in its official estimates of the population by nationality and country of birth. Of these, 322,000 were Irish, leaving just over 3,400,000 non-Irish EU citizens living in the UK. This excludes non-EEA family members whose numbers are not known and difficult to estimate from existing sources.

However, the number of people eligible for EUSS is almost certainly larger than this. In part, this is because the official estimate of EU citizens is, by definition, a snapshot taken at a point in time and tells us nothing about the numbers of people coming and going. The numbers of grants and applications reflect a broader group of people who have ever been eligible for the scheme and went on to apply (even if they are no longer resident here). Specifically, the official estimates of EU citizens exclude several groups of eligible people, including:

• People who are eligible for EUSS but no longer live in the UK. Applying to EUSS is easy for most EU citizens and there is no reason to expect that only people who plan to stay permanently will do it (not least because pre-settled status is valid for five years so long as a person does not have an absence of more than two years, giving people the option to come back later if they want to without applying for a visa). There will also be applicants who no longer live in the UK but qualify for settled status because they previously lived here for five years or more and have been living overseas for less than five years.
• People who receive status through EUSS and later become UK citizens. These people will be recorded in Home Office statistics on EUSS grants but are likely to leave the officially estimated EU citizen population when they become British.
• People who are excluded or undercounted in the official population estimates. This group includes people not living at private addresses, e.g. those in care homes, student halls of residence, hostels or caravan parks, because they are not included in the ONS figures. It also includes some part-year residents, because people who are temporarily away will often not be included in the survey. It includes some dual citizens with both EU and non-EU citizenship, because the ONS data only capture one citizenship. These groups are explained in more detail in Annex 1.

The Home Office figures are also more up to date than the ONS population statistics, which take some time to produce and publish. The ONS figures are published almost 1 year from the mid-point of the time period they describe (e.g. data for the 2019 average published in May 2020), which means that the available figures at any given point in time are usually 1 to 1.5 years old). Because net migration is still positive—or at least was still positive before the coronavirus crisis disrupted international movement—this means that the official estimate is expected to be below the current population size.

Additional uncertainty results from the fact that not everyone agrees to participate in government surveys. For example, there is some evidence that newly arrived EU citizens are undercounted in the population estimates (ONS, 2019b). A comparison between the 2011 Census and the survey-based population estimates for the same period suggests that in the past, surveys have underestimated this population. In 2011, the survey-based estimate of the number of EU-born people living in the UK was 6% lower than the number recorded in the 2011 Census. The estimated number of EU citizens in England and Wales was 8% lower than the number of EU passport holders recorded in the England and Wales 2011 Census (Table 1). The composition of the EU migrant population has changed since 2011 and we cannot assume that any undercount would remain the same today. However, it is reasonable to conclude that the direction of the error is not likely to have changed, i.e. that survey-based estimates are likely to understate the size of the EU population living in the UK.

Table 1

Estimates of the EU migrant population in 2011

Country of birth (UK)
Census estimate of EU-born population, March 2011	2,684,000
APS estimate of EU-born population, 2011	2,517,000 ± 69,000
Difference between Census and APS estimate	-167,000 (6%)
Citizenship (England and Wales)
Census: EU passport holders, March 2011	2,250,000
APS: EU citizens, 2011	2,070,000
Difference between Census and APS estimate	-180,000 (8%)

Source: ONS Census table QS203EW for England and Wales, NISRA table QS206NI for Northern Ireland, NRS (2013) for Scotland; and ONS Population by Country of Birth and Nationality Report August 2012. Original published APS-based estimates for 2011, and not the population estimates that had been revised in light of the Census results. For more information about post-Census reweighting of population estimates, see ONS (2015). For passports figures, dual passport holders with UK passports are not included in EU figures.

Taken together, all of these factors mean that the ONS figures are not a reliable guide to the total number of people eligible for the Settlement Scheme (indeed, this is not what they are designed to measure), and the true figure is likely to be higher.

How many people have received status through EUSS?

The Home Office has published monthly statistics on applications to the Settlement Scheme since April 2019. By the end of February 2020, there had been 3,343,700 applications and 2,972,200 million grants of status (either settled or pre-settled status). This included both EEA and Swiss citizens and their non-EEA family members.

There are two main problems in assessing what these headline figures mean.

First, the grants figures double-count an unknown number of people who were granted pre-settled status and then ‘upgraded’ to settled status. This is because the Home Office IT system was not initially set up to be able to extract data on individuals, only on applications and grants. The figures also count people twice if they are not granted status (e.g. because they make an invalid application) and then reapply. The Home Office (2020) has indicated that ‘initial analysis’ showed less than 2% of applications were repeat applications in early 2020, although this share should increase over time as more people with pre-settled status apply for settled status. If this problem is not addressed, over time the settlement statistics on grants of status will become increasingly meaningless, as it will not be clear how many individuals have received status. See the House of Commons’ library briefing on the EU settlement scheme (Sturge and Hawkins, 2020) for a more detailed discussion of double counting.

Second and most importantly, we do not know how this figure compares to the eligible population, because the Home Office data and official ONS EU citizen population estimates are too different to be compared.

The differences between the two sets of figures are summarized in Table 1. Of the factors listed, almost all suggest that the number of eligible applicants will be higher than the estimated EU citizen population.

The only exceptions are the fact that a share of UK-EU dual citizens ineligible for EUSS will be included in the ONS figures; and that some longer-term EU citizen residents who can show that they already have indefinite leave to remain from a period pre-dating their country’s free movement rights in the UK. Note also that the ONS figures take some times to produce, and thus represent a snapshot in time (the most recent at the time of writing is the year ending June 2019), that doesn’t reflect immigration, emigration, deaths and naturalisations that have taken place since.

Table 2

Key differences that prevent comparison between Home Office and ONS figures on population by nationality

Home Office figures on applications and grants in the EU Settlement Scheme	ONS figures on the number of EU citizens living in the UK
Double counts grants of status to people who received pre-settled status then converted to settled status.	Includes individuals, and counts each person once.
Includes people who no longer live in the UK (e.g. people who received PSS and then left the country, or those with a previous 5-year residence period).	Only includes current UK residents.
Includes all eligible applicants resident in the UK	Excludes most people living in communal establishments e.g. student halls or care homes. Includes only residents who consider the sampled household to be their main residence. Undercounts part-year residents. Undercounts dual EU/non-EU citizens. Likely to underrepresent migrants (e.g. new arrivals) because based on voluntary survey.
Includes people granted status before naturalising.	Naturalized citizens expected not to be classified as EU citizens, though some may remain in the estimates.
Includes those who choose to apply, where doing so is optional.	Includes long-term residents with ILR who are not obliged to apply to EUSS.
Updated monthly.	Published approximately 1 year after mid-point of data.

Source: Author’s analysis.

In other words, if we simply compare the Home Office and ONS numbers to create a percentage of EU citizens who have applied or been granted status, that percentage is likely to be too high. In fact, it is likely that the total number of EUSS cases will eventually significantly exceed the estimated EU citizen population and this would still not reveal much about the number of EU citizens living in the UK who had not applied.

Estimates of application rates by citizenship

The hazards of estimating what share of EU citizens have applied to EUSS by comparing Home Office and ONS statistics are particularly clear when we examine figures for specific nationalities. Figure 1 shows the number of EUSS applications by January 2020 by nationality, as a percentage of the officially estimated number of citizens living in the UK. Well over 100% of the estimated number of Bulgarians living in the UK had applied to EUSS. Five other nationalities also exceeded 100% (Figure 1), making it clear that such calculations are not reliable.

There are various possible reasons for this, including the factors outlined in Table 1. For example, it is possible that short-term migrants from Bulgaria are more likely to have applied to EUSS before leaving the country; and/or that population figures underestimate the total number of Bulgarians (e.g. because they are more likely to have arrived in the UK recently or more likely to live at non-private addresses). It may also be that some nationalities are more likely to have applied twice (Sturge and Hawkins, 2020).

Figure 1

Even a comparison of take-up rates across nationalities is likely to be hazardous. Poland has the lowest estimated take-up rate if one simply compares the Home Office and ONS figures. However, Polish citizens have on average been resident in the UK for longer than other groups of EU citizens (as discussed further in the next section). It is therefore possible that the population estimate for this group is simply more accurate than the estimate for Bulgarians, for example, because a higher share of Bulgarians recently arrived.

Do we know what other groups have higher vs. lower application rates?

Other than citizenship, the Home Office currently provides information on applicants’ locations (England, Wales, Scotland and Northern Ireland, plus separate figures for each local authority), and by age. There are currently no published breakdowns by gender, ethnicity or other characteristics.

In theory, the shares of applicants from different groups can be compared against the estimated shares of these individuals in the EU population, although statistical uncertainty again makes it unclear how reliable these comparisons are.

For example, by the first quarter of 2019, an estimated 20% of non-Irish EU citizens living in the UK were children under 18, compared to 14% of applicants in the Home Office data by the end of September 2019 (Table 2). Note that for UK-born children in particular, it may be difficult for parents or guardians to know if the child needs to apply to EUSS or are already a UK citizen, because of the complexity of citizenship rules for EU citizens.

Adults age 65 and older made up an estimated 3.6% of the EU population, compared to 2% of applications. This in principle suggests that both children and those age 65+ were underrepresented among applicants by the end of September 2019. However, these comparisons should be treated with caution, due to the statistical uncertainties described above (for example, it is possible that families with children are better represented in the population estimates than young adults without children).

Table 3

Non-Irish EU citizens by age, provisional estimates from Q1 2019 LFS

Population estimates	Estimate	Lower bound	Upper bound	Margin of error
Children under 18	683,000	615,000	751,000	+/- 68,000
Children under 16	614,000	554,000	673,000	+/- 60,000
Adults age 65+	125,000	90,000	159,000	+/- 34,000
Population estimates
Children under 18	20%	18.4%	21.2%	+/- 1.4 pp
Adults age 65+	3.6%	2.6%	4.6%	+/- 2pp
Home Office: EUSS applications by September 2019	Number		% of total
Children under 18	254,800		14%
Adults age 65+	37,900		2%

Source: Migration Observatory analysis of LFS Q1 2019 and Home Office EU Settlement Scheme Statistics. Note: EUSS applications data includes Irish citizens, although overall numbers of Irish applicants are small. Lower and upper bound estimates represent the 95% confidence intervals.

Similar uncertainties arise when comparing other characteristics, such as the distribution of applicants by region and particularly by local authority.

3. Is the EU Settlement Scheme giving people the right status?

One of the key questions that stakeholders have raised is whether EU citizens who have lived in the UK for at least five years are receiving pre-settled status instead of the full settled status to which they are in principle entitled. This could be because they do not have evidence of living in the UK covering the whole five-year period.

Whether EUSS applicants are getting the ‘right’ status is also surprisingly difficult to answer using available statistics, however. By January 2020, 41% of EU citizens’ applications had been granted pre-settled status. This share has increased over time, from 32% of people in the initial testing phases to 43% by the end of January 2020 (Figure 2).

Figure 2

How does this compare to the actual share of recent (less than 5 year) arrivals according to ONS population data? This depends to some extent on how the share is measured, but in the first quarter of 2019, an estimated 34% of EU citizens who were born abroad had arrived in the past five years (Table 4). This estimate comes with a margin of error, because it is from a sample survey. We therefore present the likely share of EU citizens who arrived in the last five years as a range, from approximately 32% to 37%.

For most groups of EU citizens (Poland, other EU-8, and EU-2), the share of people granted pre-settled status is roughly in line with what one would expect based on population statistics. The exception is EU-14 (minus Ireland) where the share of grants of pre-settled status is higher than expected.

Table 4

Share of foreign-born EU citizens age 5+ who have lived in the UK for up to five years, Q1 2019

Country/region	Central estimate	95% confidence interval	Share granted PSS (to December 2019)
All non-Irish EU citizens	34%	31.6% to 36.3%	40%
EU-14 excluding Ireland	36%	31.9% to 41.0%	45%
EU-8, of which	17%	13.8% to 19.8%	20%
Poland	15%	11.0% to 18.2%	18%
Other EU-8	20%	14.9% to 24.7%	24%
EU-2	63%	57.4% to 69.1%	63%

Source: Migration Observatory analysis of Labour Force Survey, Q1 2019. Note: Duration of residence calculated based on year of most recent arrival. Confidence intervals are calculated from one quarter of the LFS, and will be larger than those found in official population estimates, which use the APS. Data on the month of arrival is not available for most respondents who have been in the UK for around 4 to 5 years, so making the calculation from data collected in the first quarter of 2019 allows us to observe arrivals during a period quite close to 5 years.

However, this comparison too must be taken with caution, for all the reasons discussed throughout the report. We cannot be sure that the population data from which the figures in Table 2 are calculated are representative of the population of people eligible for EUSS status –including those who no longer live in the UK. For example, people who have been in the country longer may apply sooner, while those who do not yet qualify for full settled status wait until they are eligible, so that they do not have to apply twice. But it could also be the case that longer-term residents feel more secure and thus have not been in a hurry to apply, or do not realise they need to apply for a range of reasons (Sumption and Kone, 2018). In addition, if substantial numbers of people who have received pre-settled status are no longer in the country, this would mean that the share granted this status should in theory be higher than the share observed in the population. Other factors in Table 2, such as long-term residents with ILR who do not need to apply, may also contribute.

From late 2019, the online EUSS form asked applicants if they had been in the UK for five years or more, which means that it should be possible to derive some estimates from internal Home Office data on the extent to which the status ultimately granted matches the individual’s stated period of residence. However, figures from this part of the application have not been published.

Are too many children being given pre-settled status?

The share of people granted pre-settled status varies by age, with older applicants much more likely to receive settled status (this is not surprising as most will have been resident for longer). For children, it is difficult to calculate from population data what share of applicants should be granted pre-settled vs. settled status, because their eligibility depends on their parents’ duration of residence in most cases. By the end of December 2019, 37% of children under 18 had received pre-settled status, compared to 42% of adults age 18 to 64. However, it is difficult to know what share of children should receive settled status, given that EU parents with children have, on average, lived in the UK for longer.

Table 5

Duration of residence for non-Irish EU citizens age 18+, for those in families with and without dependent children, Q1 2019

Non-Irish EU citizens…	Share resident for less than five years	95% confidence interval
… without dependent children	40%	36.5% to 43.2%
…. with dependent children	23%	19.2% to 26.1%

Source: Migration Observatory analysis of Labour Force Survey, Q1 2019. Note: Duration of residence calculated based on year of most recent arrival. Confidence intervals are calculated from one quarter of the LFS, and will be larger than those found in official population estimates, which use the APS. Dependent children include all those under age 19; data are provided for all those recorded as head of family or partner, living in a family with or without dependent children.

While in principle there may be other relevant differences between those granted pre-settled vs. settled status—such as those who proved their residency in different ways (HMRC or DWP automated checks vs. uploading their own evidence, for example), the data is not available to measure this.

4. Looking forward: will we know if people are maintaining their status?

One concern that some stakeholders have raised with EUSS is whether people who have been granted pre-settled status will successfully ‘convert’ to full settled status.

People who hold pre-settled status must reapply for settled status, and will not move to this status automatically.

The process of upgrading from pre-settled status to settled status could bring complications that it will be difficult to measure. First, individuals do not always understand their own immigration status. For example, during early testing of the EUSS scheme, around 12,750 of the 200,000 applicants “mistakenly believed that they had documented PR status or existing ILR” (Home Office, 2019c), when in fact they held a different type of status such as a registration certificate. This raises the question whether EU citizens with pre-settled status will understand that their status is temporary and that they are set to lose it if they do not reapply to the EUSS to be granted settled status later. The Home Office has said that it will remind applicants before their pre-settled status expires (Home Office, 2018b), though it is possible that not all will be easily contacted.

Second, unlike the initial EUSS application, there will not be a single deadline for people to upgrade to settled status. Instead, there will be many different deadlines depending on when the person made their initial application. This makes public communication around the need to apply harder.

Third, the evidence requirements for settled status are more stringent than for pre-settled status. This is because an applicant who is not covered by the automated checks and lacks the necessary paperwork to prove their residence can currently receive pre-settled status with just one piece of evidence, such as a single invoice issued in the past six months; however, once the main EUSS deadlines have passed, applicants will need a full five years of evidence retrospectively to qualify for settled status.

At the time of writing, the data required to monitor transitions from pre-settled to settled status are not published by the Home Office. This is because there are no statistics on grants of settled status to people who already held pre-settled status. These data will be important in the future for understanding the experience of people who receive pre-settled status.

Even if and when these data materialise, it will still be difficult to know whether significant numbers of people have ‘fallen off’ the path from pre-settled to settled status, however. We should expect that some people granted pre-settled status will leave the UK and no longer need settled status. Published emigration statistics do not identify these people, so it will not necessarily be clear how many there are. For example, imagine that by time the deadline arrived, 1.5 million people had been granted pre-settled status, and new statistics showed that five years after the deadline, 1 million of those had upgraded to settled status. It would not be immediately obvious how many of the remaining 500,000 had left the country (or died or perhaps received status under another immigration route such as a work visa) vs. remained in the UK but failed to reapply and thus lost their residence status. To understand this trend, new approaches to data collection would be required, as discussed in section 6.

5. How smoothly is the settlement scheme functioning?

The EU Settlement Scheme is arguably the highest-profile immigration programme the Home Office has implemented for decades, and there has been widespread interest in what the ‘customer experience’ has been for applicants. The former Home Secretary, Amber Rudd famously declared that it should be ‘as easy as setting up an online account at LK Bennett’.

Official statistics on the various different immigration routes to the UK tend to focus on headline figures, such as the total numbers and nationalities in each route, and provide relatively limited insight into day-to-day operations and applicants’ experience (although some operational information is provided through ‘transparency data’. This means that in any immigration programme, understanding applicants’ experience using published data is usually quite difficult (see Migration Observatory (2018) for more discussion of this issue).

Compared to the rest of the immigration system, the operational data available about EUSS is more extensive, with published reports on the initial pilot phases of the scheme (Home Office 2018a, 2019b, 2019c) and quarterly statistics providing more detail on characteristics such as age, location and application outcomes. For example, as of early 2020, published statistics or other government documents told us the following:

• Success rates: The large majority of applications to EUSS have been granted either settled or pre-settled status. Around 26,000 applications decided by the end of February 2020 had not been granted status. This included just over 19,000 applications that were either withdrawn by the applicant or were considered void because the applicant was ineligible to apply, for example because they were (also or solely) a British citizen. A further 6,800 were considered invalid because they did not include the required proof of identity and citizenship or other mandatory information. And 300 were refused, in most cases because they were not able to prove that they lived in the UK or (if they were a non-EEA citizen) that they had a family relationship with an EEA citizen.
• Waiting times: The amount of information on waiting times has gradually increased. In January 2020, the Home Office website said that it ‘usually takes around 5 working days for applications to be processed if no further information is required, but it can take up to a month’ (Home Office, 2020). It notes that some cases are ‘likely to take more than a month’, including non-EEA family members, children whose applications are not linked to an adult, people submitting paper applications, and people with criminal records. More detailed waiting times data appears to be available (ICIBI, 2020, pp21-22), but is not routinely published.

Beyond these headline figures, there is relatively little data showing how applicants experience the scheme (including how this has changed over time as larger numbers of applicants come forward and the Home Office adjust the scheme in response to feedback). For example, we do not know:

• Outcomes for different groups of applicants: currently statistics are available by citizenship, three broad age groups, and place of residence within the UK. There are no published breakdowns by gender, or other characteristics such as whether someone previously held a permanent residence card, whether they applied from overseas, whether they had a criminal record, whether they lacked a valid ID, or derivative rights cases. Paper applications and their outcomes are excluded from EUSS statistics.
• Detail on waiting times: there is no data on the actual processing time for specific groups other than the broad averages mentioned above, nor on the distribution of waiting times (for example, how long it takes for the slowest 5% or 10% of applications to be processed).
• Why some applications are pending: over time, a gap has opened up between the number of applications received and the number concluded. At the end of February 2020, for example, around 345,000 more applications had been received than had been concluded. It is difficult to tell from current statistics how much of this gap is simply because recent applications are in the ‘queue’ waiting for a caseworker to look at them, vs. are waiting for the applicant to provide more information, have been referred to immigration enforcement, or other reasons. According to the Independent Chief Inspector of Borders and Immigration’s February 2020 report, ‘an estimated 40%’ of pending applications at the end of June 2019 were awaiting further evidence from the applicant. (Note that some of these pending applications will be from people with valid claims who have not responded, while others may fraudulent claims in which no further documentation is likely to be forthcoming.) More recent figures are not published.
• Administrative reviews: there is limited information on the outcomes of administrative review, although snippets of information have been released in response to FOI requests. For example, the Public Law Project calculated based on multiple FOI requests that around 90% of administrative reviews had been successful, although relatively few people had applied for administrative review (Welsh and Tomlinson, 2019).
• Reasons for not granting status: current statistics lay out the total number of applications that did not lead to a grant of status, and separately list how many of these were refusals on suitability or eligibility grounds. There is no breakdown for the other reasons and application might not be granted (e.g. the applicant did not provide identity documents, was ineligible because they had become a UK citizen, was ineligible based on absence from the UK, or other reasons).
• Evidence burden: During the public beta testing phase from mid-January to late March 2019, 73% of adult cases that had been decided did not need to provide evidence of their residence in the UK, either because they relied on automated checks of government databases, or because they already had permanent residence/ILR documents. However, we do not know how this has evolved over time (and there is reason to think it may have changed given that other things such as the share of people granted pre-settled status have also changed and the initial pilot phases were only open to certain types of applicants).

6. Can the data be improved?

In the next 2-3 years, data on the underlying population of EU citizens in the UK may improve due to new data developments, including the 2021 Census (results available in 2022-23), a programme of work at ONS to improve migration data, and changes to the way that citizenship is measured. These are explained in Annex 2. At the same time, however, the COVID-19 crisis has begun to disrupt some data collection, particularly where it relies on face-to-face interviewing (ONS, 2020c), such as the Annual Population Survey on which official population estimates rely. This could mean that in the short run there is less statistical information rather than more. In the event of very long periods of disruption, it is possible that aspects of the 2021 Census could be affected, although this remains too early to assess.

Improved data is unlikely on its own to address the fundamental problems that face analysts who want to understand the take-up of the Settlement Scheme. This is because without different sources of data, it will continue to be necessary to compare apples and oranges – that is, to compare data sources that are produced in different ways and still measure different populations. Many of the problems outlined in Table 2 above will remain regardless of any improvements in the quality of the different data sources.

Directly measuring groups who have not applied

If the government wants to understand how many people have not yet applied to the Settlement Scheme, it will thus be necessary to find a way of collecting data on non-applicants directly, and not just looking at the gap between applicant numbers and population estimates. In principle this could be done in different ways, although all options have significant limitations.

Linking administrative records

First, the Office for National Statistics or the Home Office could link data on grants of status under the Settlement Scheme with administrative records on people who were EU citizens when they registered for a National Insurance Number, and who appear in current records held by HMRC and DWP. This approach is consistent with the ONS existing ‘direction of travel’ as it attempts to make greater use of administrative data in its population estimates (see Annex 2) and could help to understand whether there are substantial numbers of EU citizens active in the UK (i.e., paying taxes and/or receiving benefits), who do not appear to have applied to the Settlement Scheme. The main limitation of this approach is that some of the more vulnerable groups who lack status may also be those not covered in HMRC or DWP records.

While understanding the EU Settlement Scheme may be the most urgent application of a data linking exercise of this kind, it would bring a range of other benefits for the evidence base on migration, as the Migration Observatory pointed out in the past (Sumption, 2019). This is because one of the major limitations in migration statistics currently is the absence of usable statistics on how people with different immigration statuses fare in UK society. The potential for such statistics to improve the evidence base are wide ranging, from a better understanding of the economic impacts of people receiving work visas to more knowledge about how the economic outcomes of family migrants change over time.

Collecting official survey data on residence status

Another approach would be to add a question to the Labour Force Survey and Annual Population Survey, on which the ONS population estimates are based, to ask respondents what kind of residence status they have, including whether they have been granted a permanent status such as settled status/ILR. It would also enable some analysis of the characteristics of people who do and do not have an EUSS status, assuming that the groups in question are large enough (since margins of error around small estimates in this data source can be quite big). This sort of analysis is not possible using existing data sources. Annex 3 provides more detail on how this might work in practice.

Collecting self-reported information in this way would of course have significant limitations. This is because individuals do not always understand their own status very well. For example, someone who does not realise that pre-settled status is temporary may report having a permanent status. However, it may still enable a better understanding of overall trends and characteristics of applicants and non-applicants than is possible based on current data.

Again, the statistical uses of such an approach would extend beyond EUSS, as data on migrants’ type of leave to remain in the UK could facilitate analysis on other topics, from the health and well-being outcomes of different migrant groups to their labour market and fiscal impacts.

7. Conclusions

There are many gaps in the evidence base about the EU Settlement Scheme, ranging from serious uncertainty about how many people have not applied to limited information about the scheme’s day-to-day operations and the ‘customer experience’ for those who apply.

Several of these can, in principle, be addressed by releasing or collecting new data (e.g. Home Office operational data or linked administrative data to assess the non-applicant population). Even without data improvements, there will remain some uncertainty, however. For example, administrative data will likely miss some migrants who are eligible for EUSS but have not engaged with public authorities that collect the data (e.g. people with no tax liability who are not receiving benefits or using the NHS). Survey data could significantly improve understanding but would remain self-reported and thus affected by individuals’ own uncertainty or knowledge limits.

For policy purposes, a key question is therefore how to proceed in the face of uncertainty. For example, it is not likely to be possible to specify a point at which application numbers are high enough that the ‘job is done,’ and that almost all EU citizens living in the UK have secured status. This suggests the need to plan for a range of outcomes, including the possibility of substantial numbers non-applicants to EUSS who have not been measured through available data sources.

Finally, the 2020 COVID-19 crisis may create further challenges both in the operation of the scheme itself and in measuring its performance. Operational challenges include the loss of capacity among organisations that assist applicants (e.g. the government help centre that assists applicants was closed in March 2020, and voluntary organisations funded to support vulnerable groups will struggle to provide services without face-to-face contact); and the need to address absences where applicants are unable to return to the UK due to travel restrictions. COVID-19 is also disrupting data collection (i.e. official surveys that use face-to-face interviewing), though it is too early to assess the consequences of this for monitoring the EU Settlement Scheme.

Acknowledgements

Thanks to Georgina Sturge, Jon Simmons, Chris Benn, Gabrielle Monk, Bex Newell, Debbie Curtis, Becca Briggs, Kuba Jablonowski and Oliver Carlisle for helpful comments on a previous draft. This report was produced with the support of the Joseph Rowntree Charitable Trust. Analysis of the limitations of survey data sources draws on research funded by Oxford University’s John Fell Fund.