What to look for in the quarterly migration statistics after the referendum

NOTE: This commentary was originally published in November 2016 and has been updated on 23rd of February 2017 to include new data released on the 23rd of February.

For the last 5 years the publication of the ONS quarterly migration statistics has evolved into something of a ritual in media and policy circles. The data are published at 9.30am on a given Thursday, and then pored over by reporters, NGOs, policy-makers, think-tanks (and the team here at the Migration Observatory) to examine what they mean for the migration debate. In part because of the Government’s ‘tens of thousands’ net migration target, the statistic that receives most attention is the estimate of net migration.

But a post-referendum change may be afoot. Because the net migration data arrive with some delay, and because the policies that affect overall UK migration could be poised to change dramatically when the Brexit process is finalised, quarterly net migration statistics are now a side-show next to the main event – Brexit.

What questions will the new data help address?

It will be several years before the effects of any Brexit-related changes to migration policy might be visible in the net migration data. Article 50 is expected to be triggered in March, followed by an anticipated 2-year negotiation with the EU. If restrictions on EU migration were implemented immediately after this period, in early 2019, this would be an ambitious timetable since it does not include a transitional period between the end of negotiations and the implementation of new policies. Even under that timetable, a full year of data on net migration under the new policy regime would not be available until the second half of 2020.

By contrast, the next few data releases will tell us more about how EU citizens are responding to the knowledge that Brexit is on its way. Among the questions the data might help to inform are:

• Does the UK remain an attractive destination—for example, will EU citizens be deterred by the lower value of the pound or, on the contrary, become more interested in moving to the UK in the short run in the knowledge that the opportunities to do so could soon be greatly reduced?
• Will EU citizens already living in the UK emigrate in greater numbers, for example because of a perception that they are no longer welcome; or, on the contrary, will emigration fall because people do not want to lose their residence rights as a result of a long period abroad?
• To what extent will EU citizens seek the security of permanent residence status by applying for it before the Brexit process is finalised; and will they become more likely to seek UK citizenship?

Understanding the datasets

Below is an outline of the key data sets that may inform this debate, when we will see them, and what they can realistically tell us. When interpreting the data, there are several important ‘health warnings’ to bear in mind:

• If the effects of the referendum on migration decisions and applications are not very large, it may be difficult to distinguish any clear trend at all. In this case understanding causation—that is, whether trends are caused by the referendum or by other factors that would have remained in place regardless of the referendum—will be even more challenging than usual.
• It will not be possible to ‘cleanly’ identify separate pre- and post-referendum trends for some time, since the data arrive with a lag. Quarterly data often follow seasonal trends, so a direct comparison of the second and third quarters of 2016, for example, will not give us a simple picture of the impact of the referendum. The next sets of available annual data, meanwhile, will combine pre- and post-referendum periods.
• Several of the datasets on migration are based on surveys and have margins of error, which makes it unwise to read too much into small changes from one period to another.

National Insurance Numbers

The first data giving any indication of new migration to the UK after the referendum were National Insurance Number (NINo) registrations. NiNo registrations are not a direct measure of migration flows, since not everyone who comes to the country applies for a NiNo and some people apply after they have lived here for some time. Nonetheless, they provide a useful proxy for migrants’ entry into the labour market.

NiNo data are released by the Department for Work and Pensions (DWP), usually approximately 2 months after the end of the quarter in which they were collected. This means that data covering the third quarter of 2016, immediately after the referendum, were released later, on 1^st December 2016.

However, caution is required when interpreting quarterly data, as NiNo registrations fluctuate substantially from quarter to quarter (Figure 1). NiNo registrations are often higher during the summer months, although this is not the only source of quarterly variation.

Figure 1

Using annual data that combines four quarters together gives a much smoother trend but does not allow one to ‘cleanly’ identify pre- and post-referendum trends (at least until August 2017, when a full year post-referendum should be available). Comparing quarterly data to the same quarter the year before addresses the problem of seasonal variation, but still does not necessarily enable us to distinguish post-referendum trends from a continuation of medium-term trends in EU migration that have taken place in recent years.

Labour market statistics

The first post-referendum labour market data were published in November 2016 with information on the estimated number of people working in the UK by nationality and country of birth. These data are usually released approximately 6 weeks after the end of the quarter in which they were collected approximately 1-2 weeks before the net migration statistics quarterly release.

The labour market statistics receive attention in part because they are released earlier than other migration datasets, but they are not a measure of migration. They show the number of foreign-born or foreign-national workers employed in the UK. This number depends not just on the scale of new migration, but also on how many of the people already living in the UK are in work. So, for example, if the employment rate of EU citizens increases in a given quarter, the number of people employed could increase even if net migration of EU citizens were zero.

The labour market data are also subject to sampling error, making it unwise to read too much into small changes over time or small differences between groups.

Long-term International Migration Data

The first long-term immigration, emigration and net migration data covering any meaningful part of the post referendum period were published on the 23^rd of February 2017, but it should be noted that while the net migration data is released on a quarterly basis, the period covered by the data is always a full year. So in February the data comprised three quarters of pre-referendum data (Q4 2015 to Q2 2016), together with one post-referendum quarter (Q3 2016). This makes it difficult to identify any specific post-referendum trends.

Data for the calendar year 2016 will be released May 2017 and this will still, of course, will cover both pre- and post-referendum months. Net migration data for a full post- referendum year are not due to be published until November 2017.

While there are several different sources for examining immigration, these data will be the main source of information on any increase or decrease in emigration of EU citizens after the referendum, when they become available in over the course of 2017.

Short Term Immigration Data

Short term migration is discussed rather less than long-term net migration figures or labour market statistics, although many of the EU citizens who come to work in the UK do so only a short-term basis and so are not included in the more frequently reported long-term migration data.

However the data on short-term migration is slower to filter through. The most recent data – which was published in May 2016 – covers the year to mid-2014, suggesting that post-referendum data on short-term migration may not be available until May 2019 if this timetable remains in place.

Permanent Residence and EU citizenship Data

If free movement comes to an end after Brexit, EU citizens currently living in the UK who wish to remain in the country will need to be able to prove that they are authorised to do so. Although EU citizens currently do not need to register to live or work here, applications for permanent residence status and citizenship give an indication of the desire for a more secure status.

Since November 2015, EU citizens who want to apply for citizenship have been required to apply first for permanent residence. As a result, we are likely to see an increase in permanent residence applications and grants before any corresponding increases for citizenship.

Data on grants of permanent residence and citizenship are released by the Home Office. Data covering for 2016, were released February 2017. They showed 65,195 grants of permanent residence in 2016, , or 53,298 if non-EU family members of EU citizens are excluded. Grants of citizenship to EU citizens during the same period totalled 16,754.

Migration transparency Data

The permanent residence grants data reflects decisions rather than applications themselves. Applications that have been submitted but are still pending are not included, but are separately reported in another dataset known as the ‘Transparency Data’.

These data do not separate out permanent residence applications from applications for registration certificates, which EEA citizens and their family members can obtain to document their status in the UK when they are not yet eligible for permanent residence. They cannot therefore be compared directly to permanent residence grants figures. However, they provide a useful indication of changes in the overall volume of applications over time.

An important caution in reading these data (which are labelled as ‘Euro’ in table InC 01a is that the most recent quarter may not include recent applications that have not yet been entered into Home Office databases. For example, application data for Q2 2016 were revised up from 36,815 to 61,500 between the August and December data releases and the Q3 2016 figure was revised up from of 59,746 to 72,117 between December 2016 and February 2017.

The number of pending applications is also reported in the transparency data (table InC_04). Here, ‘European casework’ again refers to both registration and permanent residence certificates; in Q4 2016 there were 85,242 pending applications plus 6,904 that had been received but not entered into the database.

Cautions regarding terminology on ‘permanent residence’

Note that the data that were labelled ‘permanent residence’ in the transparency data up to and including the December 2016 data release are not, in fact permanent residence data. They refer to indefinite leave to remain (also known as settlement) for non-EEA citizens. This labelling has caused some confusion in recent months. As of the February 2017 statistics, this labelling has been correctly re-labelled as ‘settlement.’

Permanent residence and citizenship in context

When interpreting figures on permanent residence and citizenship, it is worth bearing in mind that even a large proportionate increase in the number of permanent residence or citizenship applications may still represent only a small proportion of the more than 3.5 million EEA citizens estimated to be living in the UK. The government faces a formidable task in processing applications for this much larger population of potential applicants after Brexit. The Migration Observatory pointed out in August this year that at the rate that permanent residence applications were being processed before the referendum, processing applications for all the EEA citizens living in the UK would represent 140 years’ worth of work.

Conclusions

Several different official datasets will be available to inform the post-referendum debate on migration, and the timetable for their likely release (assuming that current schedules are maintained) is outlined in Table 1.

Table 1 – Expected timetable for post-referendum data releases

Dataset	Name and location of table	First post-referendum quarterly data	First post-referendum annual data (full year)
NiNo allocations	Table 1.	1 December 2016 (for Q3 2016)	August 2017 (for YE June 2017)
Labour market statistics	Table 8	16 November 2016 (for Q3 2016)	N/A
Long-term international migration (the ‘net migration’ figures)	Table 1a	N/A	November 2017 (for YE June 2016)
Short-term international migration	Citizenship tables	N/A	May 2019 (for YE June 2014)
Permanent residence grants	Table ee_02	1 December 2016 (for Q3 2016)	August 2017 (for YE June 2017)
EU registration and permanent residence applications received and applications pending*	Table InC 01a and table InC 04	1 December 2016 (for Q3 2016)	August 2017 (for YE June 2017)
Citizenship applications and grants	Table cz_01_q_a and table cz_06_q	1 December 2016 (for Q3 2016)	August 2017 (for YE June 2017)

Note: assumes that releases follow the same schedule as has been customary in recent years.
*See note on terminology above.

Because the next two years are a period of huge uncertainty and—potentially—change, the headline figures on net migration will in some respects be less ‘interesting’ than they have been in recent years. As a result, to extract meaning from the migration statistics in the post referendum world, participants in the debate may need to consider a wider and more nuanced set of data than they have had to for the last five years, where net migration was the only show in town.