Could next Monday present another moment of embarrassment for the government on immigration? As figures are finally due out on net migration – delayed from May – the signs are somewhat ominous, writes Marely Morris (IPPR). He argues that it’s looking increasingly possible that the government’s central estimate of migration – based on its longstanding International Passenger Survey – could be overestimating net flows from outside the EU.
When IPPR published a report two years ago comparing different data sources on international student migration we found early signs of a problem. On the one hand, our analysis showed that according to the International Passenger Survey, which measures migration flows, non-EU students comprise a very large part of overall net migration. This survey indicated that net migration of non-EU students (i.e. the number of non-EU students immigrating minus the number of former non-EU students emigrating) was in the range of 70,000-90,000 per year. On the other hand, the Annual Population Survey, which measures the size and composition of the UK’s population, suggested that the number of non-EU migrants who came to the UK as students hadn’t really changed at all in the past few years. This appeared to us a somewhat odd anomaly, which placed considerable doubt on the accuracy of the government’s migration figures.
Sure enough, when last summer the ONS compared its figures on students in the International Passenger Survey with new data from the Home Office exit checks process, it found that very few international students overstay their visa – and not many legally stay on either, particularly since the closure of the post-study work route. According to the ONS, this meant that the International Passenger Survey under-estimated student emigration. As a result any implied ‘net student migration’ figure would be an overestimate. The problem is, of course, that it was these same questionable estimates that had driven an increasingly stringent government policy on international students for the last 5-10 years.
But the ONS argued that, even if the statistics on students were inaccurate, this would not necessarily have an impact on the overall net migration figures. After all, the figures drawn from the International Passenger Survey tend to broadly align with census data. Any overestimates on students, they argued, could be counterbalanced by underestimates in worker or family migration.
But now next Monday it looks like the ONS is planning to publish further comparisons of non-EU work and family migration from the International Passenger Survey with the Home Office’s new exit checks data. It’s possible that underestimates of net non-EU family migration in particular might help to offset the issues with the student figures. But if these results do not fully counterbalance the analysis conducted last year for the study category, then it will become increasingly hard to sustain the argument that the problems with the data can be contained to just students.
In particular, while the overall net migration figures might be broadly correct, the International Passenger Survey could be overestimating net non-EU migration and underestimating net EU migration. In a recent article, economist Jonathan Portes has taken a similar approach to IPPR’s report and compared International Passenger Survey data on migration flows with Annual Population Survey data on migrant populations. He finds that from 2009 to 2017 the cumulative build-up in net migration of non-EU citizens is significantly higher than the growth of the non-EU born population – a difference of more than 500,000 people over that time period. Conversely, net migration of EU citizens is significantly lower than the growth of the EU-born population. These effects largely cancel each other out when we look at the overall net migration and population figures.

Digging a bit deeper, we can identify the particular nationality groups where there are the biggest discrepancies between the net migration statistics and the population figures. On the one hand, East Asian (largely Chinese) migration is considerably higher in the International Passenger Survey than the population data suggests. On the other hand, Eastern European migration is considerably lower in the International Passenger Survey than the population data suggests. (On top of this, the ONS have suggested that higher than estimated non-EU emigration could be offset by lower than estimated British emigration – so it might be that fewer British emigrants could also help to explain some of the issues in the statistics.)
Of course, all this analysis should be taken with a pinch of salt. The ONS recommends against comparing the International Passenger Survey and the Annual Population Survey because there are important technical differences – for instance, due to differing definitions (whether we define migrants by their country of birth or nationality) or differing coverage (the Annual Population Survey excludes most communal establishments). Still, it is hard to see how these discrepancies could be entirely explained by such issues – and last year’s exit checks data on students should be a warning sign that something seems problematic with the migration figures.
If our suspicions materialise, then Monday’s figures could have important implications for the UK’s migration debate. First, it is likely to highlight the problems with the UK’s current approach to non-EU migration. In particular, it will suggest that the UK might be doing a poorer job at retaining skilled migrants from around the world than it may appear. IPPR has called for a new ‘Global Talent Visa’ to offer highly skilled migrants the opportunity to work as employees, set up businesses or freelance in the UK for a two-year period (with the possibility of extensions). We’ve also argued that the UK should reopen its post-study work route for international students.
Second, it could also raise the stakes in the UK’s debate on the free movement of people. If the composition of net migration leans more towards EU citizens than originally thought, then the impacts of restricting free movement could be of even greater significance to the UK, in particular to its labour market. It’s worth noting, however, that the ONS will not draw any direct conclusions on patterns of EU migration from their analysis on Monday, because exit checks do not cover EU nationals. The accuracy of EU migration figures will therefore probably remain uncertain for some time to come.
But perhaps most importantly, the figures on Monday could put further pressure on the government’s wider strategy on immigration. An approach that is fundamentally directed toward driving net migration down to the tens of thousands – and that encourages highly skilled migrants to leave the UK after short periods – is hard to sustain if there are such serious doubts surrounding the underlying data on which the target is based. Last year IPPR proposed a new approach to post-Brexit immigration, based on addressing our key economic challenges rather simply bringing down numbers. Just days ago Home Secretary Sajid Javid did his best to dodge questioning at the Home Affairs Select Committee on the government’s net migration target; perhaps next week will finally be the time to lay it to rest.
This post represents the views of the authors and not those of the Brexit blog, nor the LSE.
Marley Morris leads on IPPR’s work on Brexit. His research focuses on migration and trade policy, including the alternative options for Brexit, the role of EU and non-EU citizens in the UK labour market, and government policy on immigration and international students.
It is more likely the Government has grossly underestimated the amount of immigration into this country from within and without the EU.
until it starts controlling our borders more proactively and forcefully, immigrants will continue to flow in.
Think the problem is based around a methodology that is no longer adequate. Essentially APS divides tourists from migrants by the expected length of stay in the country. That may have been alight when access to (what are now called) citizen’s rights and pubic services was based on residency of a British country (the Empire) and migrants were described as ‘aliens’ And all aliens needed to be registered.
In labour market terms we started off with the equivalent to the LFS unemployment measure whereas increasingly we need an estimated of claimant unemployment – what eligibility is required to be considered a migrant.
Consequently, I think that much more use needs to be made of admin data to get a better picture of what is going on. (and in fact would shift the headline stats from a survey based measure to admin). Developing the statistics on National Insurance Numbers would seem to me to be the best and most likely source.
This does seem right especially when you look at how low the figures for awarding permanent residency to non-EU immigrants are (60,670 in 2016 and 63,941 in 2017; though note this is lower than expected due to the switch from a 2 year to 5 year qualifying period for family based permanent residency. 80,000 would be a more normal figure, still far less than the alleged non-EU net immigration figures).
One thing that often gets neglected though is how high the figures are for non-EEA nationals coming to the U.K. as family members of EEA nationals. 43,000 such permits were given out in 2017 alone. Many are for South Asian people, but also West Africans and Brazilians.
The international Passenger Survey is a wholly inadequate assessment of Net migration. Other evidence other than the Annual Population Survey is the fact that the number of settlement grants have fallen to about the same levels as before 1997. In 2016 there were 60,670 issued and 2017 63,941 were issued. The amount of work visas issued have remained similar for the past ten years at between 140,000 to 180,000. The amount of visa extensions have fallen from 291,000 by end June 2013 to 192,000 by end June 2017.