The UK's migration statistics are "little better than a best guess," according to a committee of MPs. We explained their concerns in 'How accurate are immigration statistics?' yesterday.
But what can be done to improve the quality of immigration information we have?
If Disney World can do it, why can't the UK?
The Select Committee, and individual politicians of different parties, have their own prescription: count people out of the country and count people in.
As one observed: "I have just come back from Disney World with my kids. You have your fingerprint every time you go in with your pass; if they can do it at Disney World, I am sure we can do it on our borders."
The Deputy Prime Minister took a similar stance, telling a Radio 5 live phone in that he wanted "a simple system that's done in many other countries, where you count people in and you count people out." His spokesman offered the examples of Australia and New Zealand when we asked.
In contrast, the Chair of the Committee observed in an evidence session, "there is no country that formulates their statistics on that basis, because it is an impossible task."
e-Borders to the rescue?
The Committee recommended that "The ONS and Home Office should move as quickly as possible to measuring immigration, emigration and net migration using e-Borders data."
e-Borders is the "huge disappointment" that is intended to electronically collect information on people going in and out of the country. In a tradition of government IT projects, it's late, although it is now at work checking entrants, if not yet counting them. It's due to come fully online in 2018.
The Statistics Authority attaches a lot of importance to it, but stressed in their evidence to the Committee that e-Borders "will not entirely replace the IPS [the current survey], as it does not record migrant intentions and will not in itself provide information on where migrants settle in the UK."
Besides this, the Committee noted that e-Borders doesn't record information about the language ability of somebody arriving, their reason for coming, their job before coming, where in the country they plan to go, how long they plan to stay, and so on.
Problems that counting heads can't solve
Counting people in and counting people out is the migration equivalent of registering births and deaths, introduced in England and Wales as recently as 1837.
But immigration and emigration aren't quite so simple. Not only will some people coming and going want to avoid giving honest answers, but asking people their intentions as they enter isn't proof against subsequent changes of plan; the job that doesn't work out or the love affair that does.
This leaves some big gaps in our data, such as where immigrants go and thus which local authorities need more money to provide their services.
One idea floated when the Statistics Authority consulted on the topic in 2009 was to introduce a 'population register'. Immigration expert Professor Salt told the Committee that it represented the "only… way in my mind that we can bring about a step-change improvement in migration data."
A population register, as UN statisticians describe it, is a compulsory continuous "inventory of the inhabitants of an area and their characteristics, such as date of birth, sex, marital status, place of birth, place of residence, citizenship and language."
Whether or not you think this kind of counting might be "totalitarian", as Vince Cable seemed to suggest, the fate of previous attempts to legislate for ID cards suggests that a population register would be a major political step.
We do get a snapshot of some of the benefits of a register from the census which takes place every 10 years. Even that is politically controversial and may not be repeated but the last one in 2011 changed what we thought about where immigrants were significantly.
The e-Borders machinery for (mostly) counting people in and out of the country is being put in place and will eventually have an important role in future national-level migration statistics, although it won't replace the current survey.
The "simple system" of counting people in and out won't give authorities all the information they need about where immigrants are ending up, and the kind of counting that would be needed to solve this problem definitively is a hotter political potato than you might have realised from the weekend's coverage.