UK Edition
The missing million: Are we undercounting Britain's immigrants?
As cameras flashed and a video crew dangled its microphone just above his head, Keith Vaz welcomed Britain’s newest immigrant. Shaking Victor Spirescu by the hand, the chairman of the Home Affairs Select Committee peppered him with questions about his plans.
“I don’t come to rob your country,” the 30-year-old assured Mr Vaz, and the gaggle of reporters. “I come to work and then go home.”
It was 1 January 2014 – the first day Romanians and Bulgarians were granted unrestricted access to the UK – and Mr Vaz had come to Luton Airport to greet some of the first arrivals in person. He even took some for coffee.
Not every arrival since then has been so well charted – or charted at all.
Now concerns have been raised that Britain’s official system of measuring immigration – based on a survey initially designed to count tourists rather than migrants – could be significantly understating the number of people moving from the European Union to live and work in Britain.
Critics of the system have highlighted a difference of more than 1.2 million between the official figures for the number of EU nationals moving to Britain in the past five years, estimated by the Office for National Statistics (ONS), and the number of National Insurance numbers issued to those from the EU during the same period.
In the five years from July 2010, the ONS estimated that just under a million EU nationals moved to Britain. But during the same period, the Government gave arrivals from the EU more than 2.2 million National Insurance numbers. In four of the last five years, the gap between the two figures has widened.
Working out the reason for dry statistical discrepancies might normally be the preserve of academics. But this time it could have significant political consequences: the ONS is due to publish a study explaining the difference – together with fresh estimates of short-term migration– on Thursday, a little over a month before the EU referendum in which immigration has become a crucial battleground.
The size of the gap has also thrown a spotlight on the difficulties of accurately measuring Britain’s total population, with academics suggesting a variety of improvements that could be made.
Why now?
The Government has long promised to cut net migration to the “tens of thousands”, but the issue has become more politically charged this year, ahead of the EU referendum.
In March, the UK Statistics Authority, which regulates national statistics, wrote to the ONS pointing out the importance of accurate statistics to “inform public debate on immigration levels ahead of the EU referendum”.
In the letter, Ed Humpherson, director general for regulation at the authority, warned: “There is a significant risk that a lack of progress in reconciling and explaining the differences over the coming weeks could undermine public confidence in official migration estimates.”
Nigel Farage, the Ukip leader, has capitalised on the uncertainty. “They are pulling the wool over our eyes,” he has said. “[National Insurance numbers] are a simple and clear reflection of the real numbers of people in this country, as without them you can neither legally work, nor claim benefits.”
The discrepancy could also have consequences for the provision of public services. Councils and Government bodies will often take into account population statistics before approving planning permissions or deciding the number of school places needed, for example.
Prof David Coleman from Oxford University, who co-founded MigrationWatch UK, said the accuracy of statistics would not matter so much if overall immigration figures were low. But he added: “When migration is the dominant driver of population change, it matters a great deal.”
Why do the figures differ?
There are two important reasons why statisticians expect the ONS figures to differ somewhat from the number of National Insurance numbers issued.
Firstly, the ONS figures only account for “long-term migration”: people who move to the UK for a year or more. By contrast, anyone who wants to come to the UK to work – even for just a few weeks – needs a National Insurance number.
“Someone can come and pick strawberries for three months and then leave,” said Dr Jakub Bijak, a demographer at the University of Southampton. “They would be counted in the [National Insurance] figures but shouldn’t be counted in the population estimate.”
Secondly, while the ONS statistics are based on a survey the first time an immigrant enters the country, an immigrant can apply for a National Insurance number whenever he or she likes. If they arrive as children or students, they might live in Britain for years before applying for a number. So numbers issued in any given year do not necessarily reflect the number of new arrivals to Britain.
But neither of these explanations accounts for why the gap between the two figures is increasing. “The question is not why there is a gap – we know why,” said Dr Madeleine Sumption, director of the Oxford University migration observatory. “The question is: why has that gap got larger?”
One explanation is that the ONS has begun to undercount the number of immigrants. According to Jonathan Portes, principal research fellow at the National Institute of Economic and Social Research, this might be because of changes to the pattern of immigration. As EU immigration has increased, he argues, it has become harder to tell whether new arrivals will stay for just a few months or more than a year. This means passengers from the EU who went on to live permanently in Britain might have been incorrectly designated as visitors and not counted towards the total.
“The survey [the ONS uses] asks people if you plan to stay for more than a year: if you say yes, you’re a migrant; if no, you’re not,” he said. “If you’re coming from India, you probably know the answer: you need a specific type of visa, you don’t come here on a whim and then suddenly decide to stay.
“If you’re an EU migrant, that is very different – you don’t need a visa. There is nothing to stop you saying ‘I’ll come here, look for a job, maybe I’ll stay here, maybe I won’t.’ There’s no way [the survey] can measure that.”
The ONS figures are based on the International Passenger Survey, which interviews about 800,000 passengers at ports and airports each year. But Prof John Salt of the University College London migration research unit points out that only a relatively small proportion of those interviewed are classified as migrants, making any conclusions about immigration much less reliable than the information about tourism the survey was designed to produce.
“You could increase the size of the total [survey] sample enormously and still only get a relatively small increase in the subcategory classed as migrants,” said Prof Salt.
Giving evidence in Parliament last month, Sir Andrew Dilnot, chair of the UK Statistics Authority, said it was “conceivable” that the report into the discrepancy would show the two numbers “are consistent with one another but measuring something different”.
But he added: “It is conceivable that they will not show that… and we will make a judgment about what we think about their quality.”
How else could we measure population?
The census has traditionally been the gold standard for measuring population. But even this is not infallible and is only carried out every decade.
There are three other ways population could be measured:
Other official sources
The ONS could draw more heavily on so called “administrative data” already held by Government departments. For example, there are 68 million patients registered with a GP in the UK, compared to an official population of only 64.5 million.
But each of these potential sources carry their own health warnings. For example, statisticians point out that patients can be mistakenly registered with multiple GPs or can forget to de-register when they move abroad.
Border information
The ONS is already examining whether passport checks when people enter and leave the country could be used to more accurately assess immigration.
“They are moving towards that but they are not there yet,” said Mr Portes. “The primary purpose of having your passport checked is checking you against watchlists not to provide immigration statistics.”
A population register
A Labour proposal to give every Briton an ID card was defeated. But some demographers argue assigning a unique person number to everyone at birth or the first time they visit Britain is the only way to accurately assess immigration.
“We are trying to count population using a mixed bag of unsatisfactory measures,” said Prof Coleman. “That will continue to be the case until we grasp the nettle of a population register.”
Even such a register would have its flaws, though. It could actually lead to overcounting of net migration since people might forget to deregister when they leave the country.
“Even good registers are not perfect,” concluded Dr Bijak. “No single source of information can be 100 percent accurate.”
What other data could be used to measure population? Email population@telegraph.co.uk