 Rwy'n gweithio y cwestiynau cyfnodd. Why is it so hard to get detailed migration statistics for the UK? Here are a lot of debate, there are a lot of people talking about migration in the news. Often they're using numbers, and actually they're numbers that we are not very certain about. I'm going to come to those two things in a moment but so you know where I'm going and so you can relax a bit, I basically only have three points. One point is I'm going to show you a graph from the Office for National Statistics about recent migration and we'll try and decode it a bit. Then we're going to play a little thought experiment, see what would happen if we decided to close the doors and declare ourselves in here an independent republic and then we had the problem of accounting migration, we'll see how we might go about it and then we'll put the two together and see how it works when you try and do it for the UK. So when you know I'm on that bit, you know I'm nearly at the end. Two statements from the media, different politicians, same week earlier this year. Theresa May sticks to a 100,000-year net migration target. Sustainable level means tens of thousands, Prime Minister says. It's kind of vague, isn't it really? 100,000 and tens of thousands. Day before, Home Secretary refuses to commit to bringing immigration below 100,000. Amber Rudd said there was a lot of discussion around where the student numbers should be included in the target. So the message to begin with here is a lot of important people are talking about it, but no one will commit to any very exact numbers. That's partly because the meaning of those numbers are very loaded. They refer to real people, real people who may be in the UK right now wondering how to read the debate, but also it is extremely difficult to be confident about those numbers once we narrow them down and fill in some detail. That is my only graph, my one statistical artefact for the evening. And that is what the Office for National Statistics published. It's their latest quarterly migration statistics. So it's from August this year it was published and it's an attempt to show what we think long-term international migration was doing for the last 10 years up to March this year. I'll decode it just a little bit. The first thing to realise in here is that the magic number you just came up with, that 246,000 that you correctly guessed, is over on the right-hand side here. And that was the best estimate from the last quarter of how many people had stayed in the UK over the last year once we have taken the top line, which is the people coming in, subtracted the bottom line, the yellow one, which is the people going out and assumed that the remainder is the net addition to the population from international migration. And are there any international students in the room? Anyone who described themselves as an international student? Welcome. You're the only guy owning up to it, I suspect there may be more. The reason that matters is because it was in Amber Rudd's statement there, and how many of those people are international students, whether they should be included in these numbers at all, and what long-term means for students are all really contested issues, but some of the figures might say that they contribute about 80,000 to that number, so it kind of matters a lot as to whether you're going to hit any particular target, certainly a few tens of thousands. The other thing to realise about this 246,000, I'm aware now you've told me what you've done, that there's quite a lot of you with a degree. The only concept I want you to realise is that when we do a survey sampling, often we have a small survey and we're trying to guess about a big population. None of these lines are counts of people. They're all estimates based on a survey. It's quite a small survey and we'll say something about it in a minute. When we do that survey, we might not get the number quite right, and if we did it again, we might get a slightly different answer. If we did it again, we might get a different answer again. The confidence interval around that 246,000 is such that we're 95% certain we're within 42,000. This is quite a bold estimate, but it's got a big level of variation around it. If you then take a look at how much the line changes, it goes up and down there, often the change is much less than the confidence we've got in the number. It might be in the right ballpark, but it's not going to be much use for getting very anxiety-ridden about whether the number has gone up by 10,000 or 20,000 or gone down, but it's actually within the margin of error of the whole exercise in the first place. Let's decode it a bit further. The first thing is this is long-term international migration, and you've already established that you know that that means we're talking about people intending to stay in the UK for a year or more, so the way the United Nations do it is international migration, so it's helpful if all the countries measure it the same way. So we're talking about a year or more. Most international students would be counted as in long term in that definition, it's the way it's done. The other thing about it is, of course, that the most recent data are provisional, because you can't really be sure about what's happened if people have stayed for more than a year until you've waited at least a year since they arrived and then had another go at working out what's happened. So whenever we look at migration, we're always looking at what happened sometime back, and if you've got a lively political debate going on, people making decisions about, well, Brexit, not sure about what these politicians are saying, I don't like the feel of how it's going, I might make a decision now. We're always quite a way lagged when we understand what's been going on. We're looking back in the past a bit. Another thing here, I mentioned that we work out what we've got as the immigration and what we've got as the immigration, we take one from the other and that's how we get our net migration figure. That's fine, makes sense, except as I've already hinted, we don't actually have an exact count of either the immigration or the immigration. We've got a lot of conflicting different data sources that give us a clue about what those numbers ought to be and then we try and put them together to get an overall answer on that we will unpack in the next couple of slides. Last thing to realise about this one, it's a very peculiar graph when you look at it harder. So around about 2011, those bars at the bottom are drawn in a different colour, they're done at a different frequency and they're labelled as revised net migration and that's because in 2014, when we knew the results of the census from 2011, everybody had a hard look at the numbers and worked out that it simply didn't add up what we'd been doing from 2001 to 2011 because the series of the counts of migrants coming in and migrants going out simply did not add up to the stock of people who the census found who told us that they'd come from another country in the last year or been born in another country and as a result, the method for doing the counting was revised, the numbers were restarted and they're completely inconsistent at that point. The net migration figures are an official revised answer but the top line and the other line haven't been changed. So when we look into the past and we change methodology, we don't have a consistent picture and that's really important as well and it will be really important for any of us interested in this debate if you go forward a few years because what I'm going to paint for you is a picture in which the methodologies are all going to change and the thing we can be sure about is that we will then disagree about the numbers we've got now because the new methodologies will almost certainly give us a different answer. So, 0.1 dealt with, 0.2 thought experiment. Let's think about migration to a small venue. Let's just imagine that about 7.30 we shut all the doors, had a referendum and declared ourselves an independent republic. When I walked down the road, I walked past Barcelona and I thought, what a shame we're not there. Never mind. We'll do it for here. We could imagine that there are people who've come here from somewhere in Southampton City Centre this evening. Anyone come from somewhere around Southampton City Centre? Maybe stretch as far as Highfield? Quite a few. A few people who came from somewhere else in Hampshire did withdraw anybody from Basingstoke, Winchester and over. A few. How far? Tottenham. Tottenham? OK. Well, it's almost the rest of Hampshire. Point is, we've got different domains out there where people may have come from. If we think about what's happening, some of those people may have come in, joined us from Central Southampton, maybe after work, and then when they go home they'll go back somewhere further away. I'll be doing the reverse journey. Some people will come in and go back where they started from. Effectively, it's already becoming a slightly more complicated picture. Of course, in our independent republic we've declared someone would have to be charged with managing official statistics and this would be quite a problem for them because we're all going to be demanding that we want to set up and run our new country properly. We're going to want information about where are people from and what are they doing and what services do they want? Are they old or young and are they intending to stay for a long time? So this kind of stuff will matter to us. Now, it hasn't happened. I guessed that one or two curious people might be walking up and down above Bar Street and they thought, those people look like they're having a really good time. I might just go in and see what they're doing. So some people might come in just for a quick look but then they'll realise it's me talking and they'll go again. Other people may be here and I'd notice, and it was maybe 6% of the people who said they knew a speaker so it may be that some of you are just here because you're desperate to hear what Julie's going to say later on and some people will just come in for Julie's talk, there'll be a big surge and they'll listen and then they'll go again and they'll kind of be with us for a while but then they'll be gone. And some, stars like most of you, I'm guessing, are here for the whole event. So our migrants are coming on different timescales from different places going to different destinations and they're all playing different roles so I'm currently speaking but I will, for your keeping you happy, shortly sit down and become a member of the audience. John there, I suspect, is being tech support for the entirety of the evening but people's roles and reasons for being somewhere change. So we can obviously play this game for a while but you see that it would be quite a tricky job for us to work out exactly what's going on. How could we measure this? We could freeze everything and count everyone so we say okay, nobody move, Sylvia can't move, can't do that. I need to count everyone. We get everyone exactly where they are and we go round and we say where did you come from, where are you going, what kind of person are you, why are you here and we have really good evidence base but the problem would be that shortly afterwards some people would leave, some people would come in, some people would change roles, it will all be out of date again. No use. Kind of helpful but didn't last very long. Another thing we could have done which is a bit like what actually happened is we could get somebody to stand at the door and count who comes in and out. A sort of entrance and exit survey it might work for this venue but it's going to be tricky if our nation becomes very popular a lot of people join us. We have actually got more than one exit because you all checked, fire escapes didn't you? We're going to have to man all the doors and then we're going to have to be there and we'll count all the people in, all the people out and work out what's going on. We certainly couldn't just rely on the tickets because there may be people who are here without a ticket on one of them and there may well be people with tickets that didn't come so people's intentions and tickets and visas and things like that are not reliable. Or a third strategy we could put a big list of everybody work out where they were and follow where they go except we haven't got a list so that one's not going to work. Count everyone, we get the stuff in the middle count them all in, we get one half count them all out, we get another half simple. So if you put those two together that was point two done only one to go how can it be so complicated to do the Republic of Wales 104 job for a small country? Why is it that we are not really very sure about any of those numbers if the counting process isn't inherently that complicated? So here we go again same boxes in the same places although this time the bars of the UK the rest of Southampton has become the EU and the rest of Hampshire has become the rest of the world loosely speaking. We have visitors coming into the UK more than 100 million people come in of the UK every year the enormous majority of them are not migrants we've already seen those numbers they're in the hundreds of thousands so a huge churn people coming in visiting relatives seeing Tower Bridge going on trips Stonehenge Scottish Islands loads of people come in they do short courses they do placements with their employer doing all kinds of very valid reasons for coming into the UK for a week they may be around for more than three months in which case they'll fall into a category that we call short term migrants or they may last for more than a year in which case they'll fall into that long term migration category and we've got them all going on and we've got of course all of the reasons and more why people would come they might come to work they might come to study might come for many other reasons work and study are the biggest ones I'm speaking of the migrants into the UK in the period in the last bar of that graph 558,000 people came in in ONS's latest estimate about 250,000 round numbers of them came to work and about 135,000 of them came to study that was their stated intention when you scale up the survey data so those are the two biggest groups but of course something very important happens here that matters a lot and that is that people switch around between the reasons they're in the country so someone may come in as a student and then stay on to work they may come in to get married and then stay for years they may think they're just coming to stay for years and it doesn't work out and they're gone again and one of the challenges we have is the reasons people come in and out of the country once they're here we can't capture any of that stuff there is no border survey we said we're going to do this and now you're going to do that we simply don't know and therein is one of our big sources of uncertainty and one of the others is that we are not very certain because we've got a very small sample of a survey taken at a few places driving these numbers how do we count them same as here we can freeze everything and count everyone we try to do that with a census last one in 2011 migration and out migration data were quite wrong and we tried to correct the methodology but of course that was a long time ago a lot has happened since 2011 so we've changed the rules we've had a big political debate we've had a referendum all sorts of other stuff has gone on in the world those data are not going to be very useful to us we're not doing another one to 2021 who knows where we'll be by 2021 and what we learn there may only be valid and helpful for a year or two and then it will be gone again so that's nice, it works it's brilliant, really rich data source but not very useful in the long run we could stand at the entry and exit points and count who comes in and out so this is the like checking tickets at the door except we don't do it what we do is a bit of a hybrid we have an international passenger survey which is currently the main source of those statistics I will tell you a little bit more about it but it only covers a very small proportion of the people coming in and out of the country and only a very small proportion of them are going to be migrants or we could monitor everyone on a list but just like the lounge 104 we haven't got a list but we are working on it and where we're heading with migration statistics is a bit of a hybrid of those last two census does that just like it would for the people in the room international passenger survey bit like what we did counting them all in and counting them all out except we're not counting them all we're counting a very small fraction it's face-to-face interviews at airport anybody here ever been interviewed by the international passenger survey a couple at the back I wanted to be interviewed by the international passenger survey and I could see them doing it at Heathrow and I walked round the cosmetics but they wouldn't select me because it's quite a carefully regulated exercise but look what happens it's a really complicated survey 700,000 completed interviews most years and that's still 0.4% of the international travellers 80% response rate is fantastic only 4,000 long-term international migrants revealed in that way and all we're discovering is what people tell us their intentions are when they come in of the country or go out again someone who quite rightly tells us I'm coming here to do a masters props out of their program after two weeks and he's gone that piece of information is kind of lost to the system likewise the person who says I'm coming to do a masters they do their masters, they do their PhD they get a fantastic career and they stay and work in the UK this really only captured the intention that they were going to be a student we know nothing about the people who switch intentions so in there is a mystery about that huge uncertainty with the numbers that are based on the passenger survey it's probably not too bad for the absolute count but it doesn't tell us much about the characteristics of the people we're almost at the end, where are we going well those lists what if we had everyone on a list we've got a stack of lists we've got lists of asylum seekers people have applied for visas from outside the EU in order to come in the UK only in 2015 did we start routinely checking everybody's passport as they exited the country and we have the beginnings of a system to work out what the numbers of the flows are you might think that's crazy we stopped doing it on paper in 1998 we only restarted in 2015 there's a few other systems there none of them are joined up just like you'd expect from British Government nothing is connected lots of IT problems we've been talking about making an integrated system for a decade and we haven't done it yet we may be on the cusp of getting there but what will happen is that when we get there all the numbers we think we now know will be rather wrong because the new system will measure it all differently and the students and the people who switch intentions will always remain very very difficult last slide we haven't got any single comprehensive or compulsory way of measuring the migration numbers we do have a decent survey that we stand and do at the exits and entry and it gives us a reasonable ballpark number but it's got a very big confidence interval around it what exactly you count depends on your definitions so if the government makes their minds up about the students then it'll change the number a lot whether we count them in or we count them out cut them one way or another the administrative data sources are like people doing a giant jigsaw which is being done but too soon to see the picture and my message for you is just like us in the republic of lounge 104 all of those migrants are real people they change their minds they have complex lives they meet folk, they get married they decide to return they come back again make some friends this evening don't fall out with anyone over arguing exactly what the numbers are there you go, finished