Robert Baynham, Tim Phillips and Jon Simmons.

Q130 The Chair: Welcome everybody to this 10th public oral evidence session of the committee’s inquiry into settlement, citizenship and integration. We are delighted to have three more witnesses. To kick off, I would be grateful if you could each introduce yourselves.

Richard Baynham: Hi. I am Richard Baynham, a senior audit manager at the National Audit Office.

Tim Phillips: I am Tim Phillips, a director at the National Audit Office.

Jon Simmons: Hi. My name is Jon Simmons. Until last month, I was the deputy director in charge of immigration system statistics and refugee analysis at the Home Office. I have just retired.

The Chair: It is lovely to have all three of you with us. Because time is tight, if there is anything you feel we have not picked up, please feel free to contact us afterwards. I have no doubt there will be things we will ask you to write to us about anyway. We will start with Baroness Bertin.

Q131 Baroness Bertin: Good morning. The first question: how does the Home Office monitor compliance with the Immigration Rules? This question is threefold. How does it ensure that visa holders leave when their visas expire? How are the numbers of people who are arriving irregularly estimated? What are the issues with regards to that and the common travel area?

Jon Simmons: That is probably one for me, I think. That is a really big question. I will try to be brief but that may be difficult.

Compliance is a matter for individuals. There is a whole operational aspect to this question. I am here to talk about statistics, which is my speciality. I can talk a little bit about operations. Home Office enforcement has lots of different ways to monitor and check compliance. If someone comes to its attention, it can very easily find out the visa they travelled on, their status, whether that visa is still valid, and so on and so forth. There is a whole operational aspect to that.

It is of course related to the compliant environment and the checks that go on there. Last April, another analytical team at the Home Office published Evaluation of the Compliant Environment: Interim Report, which contains lots of information about the operation of the compliant environment and the sort of checks that go on. I would refer you to that report to understand how the compliant environment operates.

It is probably best that I talk about the exit check statistics and the regular migration statistics, which are most pertinent to this. You are probably well aware that between 2015 and 2020, the Home Office published exit checks statistics. Essentially, those statistics were there to link people’s due departure date from the UK with the visa they held.

The statistics consistently show that the vast majority of visa holders abide by the conditions of their visa. In the final year of those statistics—year ending March 2020—1.9 million visas expired with no extension of further leave. Some 95.2% of those visa holders could be seen to have departed on time. There is a small percentage of people who depart late. Sometimes that is for good reason—they may, for example, become ill and cannot take their flight. In that year, 0.4% were late departures.

We were very clear in the published report that we could not assume that the remaining number from those figures, which is less than 5%, were necessarily people who are still in the UK. You referred to the common travel area, which is one aspect; people who depart via Ireland are not captured in the data because we have, in essence, freedom of movement between the UK and Ireland. But there are also issues around data matching that are quite important in this respect.

The exit checks statistics rely on trying to find a very small number of people within a very large volume of visitors. The example we would often give is that, in 2020, there were around 140 million passenger arrivals in the UK—a very large number. More than half of those were British citizens. Another quarter or so were from the EU. Obviously, at that time we had freedom of movement. So, only about 14%—one in seven—of those people were non-EU nationals. That is still 20 million people. If there is an error in matching the data of, say, just 1%, that would amount to 200,000 of those passengers not having their arrival matched with their departure.

Most of those people are just common visitors and tourists, not long-term migrants, but the impression it would give, if you just took a very simplistic view of that data, would be that there were 200,000 people who we could not say had left the UK and therefore were maybe overstaying. That is completely unrealistic. We did not believe that, we do not believe that, and we do not think that that is a correct view of the statistics. It is important to understand some of the gaps in that data, because we are trying to find a very small difference from a very large pool of people.

We stopped publishing the exit checks data in 2020. That was basically because the Covid pandemic affected people’s travel patterns, particularly around extensions. There was also a new immigration system, and the departure from the EU, which meant that a lot of EU citizens had a different relationship with the immigration system.

Home Office statisticians in my former team have been working on trying to re-establish the exit checks statistics. They have come a long way and are almost there but are not quite ready to publish yet. We want to make sure they are of a good enough quality to publish in a robust and reasonable way.

Baroness Bertin: When will you be ready to publish them?

Jon Simmons: Well, it is not my job any more, obviously, but it is coming to the end of the process. I anticipate they will be published at some point in the coming year. I cannot say when. It could be the spring or the summer. It is hard to say at the moment.

Tim Phillips: I will just come in on a couple of things. If we zoom out a bit on the issue of compliance, it is important to say that compliance happens in many layers across the immigration system and it happens on a risk basis.

The Home Office will try to think about the riskier cases, whether that it is sponsors of visas or people applying for visas. An important difference is that, out of country, at the point of application, the Home Office knows the total population that it is dealing with—every applicant, essentially—and does its risk-based triage on the basis of a knowledge of the whole population. The deeper you get into the immigration system, the more loss of knowledge about people there is. So, although the Home Office is able to estimate the number of overstayers, it does not necessarily know where they are or who they are. Therefore, when it does its work to decide who to go after, it does so on the basis of both risk and opportunity, from information that comes to light.

The second point is that we can see that, even from year to year, where the Home Office puts its efforts in compliance can change quite dramatically. Sometimes, in certain areas of the system, it will be looser; sometimes it will be tighter. An example of that is that, over the two years between 2022 and 2024, there were 10 times more compliance checks on sponsors for skilled worker visas from one year to the next. The number of checks on sponsors went up by 10 times.

The Chair: Jon, you opened your remarks with one sentence: “Compliance is a matter for individuals”. What did you mean by that?

Jon Simmons: I meant that in the sense that we are not North Korea—we do not track people’s movements around the country. We expect them to abide by the terms of the visa that we issued them with. When they do not and then come to our attention, as Tim points out, there are various ways that the Home Office can check their status.

For example, if Immigration Enforcement comes across people in one of its raids and wants to check their status, it can do that on an individual basis. What is much harder, because of the fuzziness of the data around the edges, is to create statistical estimates of the overall number.

Q132 Lord Hogan-Howe: Jon, you explained that of the 1.9 million, at most about 5% overstay, although, as you explained, it may be a smaller number because you cannot count everyone. That is a small percentage, but it is still quite a lot of people—approaching 100,000. Who goes looking for them when they overstay?

Jon Simmons: Obviously, Immigration Enforcement has various ways to find overstayers. It has increased the number of checks on employers for example, because people might be working illegally. They get calls from the public if people have suspicions, which is another way of getting information. They do informed but relatively random raids of different businesses. The majority are probably in the food sector, because a lot of irregular migrants have been found to work in restaurants, takeaways and so on. They do checks on businesses, checks on employment and can check the status of any individual because the individual has to prove who they are, which allows that to be checked against Home Office records.

Lord Hogan-Howe: May I just follow up briefly? I know that you are not here to explain about what customs or immigration do, but it sounds like the best place to start in finding them would be the employer who obtained the visa for them. They would have all the information—their address, telephone numbers.

Jon Simmons: That is true, but most people with work visas who are sponsored are pretty compliant. In the exit check statistics, 95% of all work visas could be found to definitely have left in good time, in the terms of their visa. It is not quite a needle in a haystack that we are looking for, but—

Lord Hogan-Howe: The numbers who overstay are relatively small. It is only cumulative up to 95,000.

Jon Simmons: Yes, the numbers that we are looking for are very small within a very large haystack.

Lord Hogan-Howe: But if you start with the employer, it is not a haystack.

Jon Simmons: No.

Richard Baynham: In March last year we published a report looking at the skilled worker visa system. We looked at the compliance effort at the point of entry and at what happened at the end of people’s visas. On the compliance effort at the point of entry, we raised four issues. First, there was the need to strengthen real-time risk assessment, so that you understand across the sector where the problems lie. We were encouraging the Home Office to increase its activity there.

Secondly, there are finite resources which are often stretched. Over the period of 2021-24 there was a reduction of around a third in the number of compliance officers. That limits the ability to engage and explore the sector. Only 1% of sponsors were referred for compliance checks. That is an issue that needs to be managed.

Thirdly, the system needed to improve data, including data from other government departments, to strengthen compliance checks, and improve co-ordination with other bodies. When you get into what is happening in the labour market, you need to join up with other bodies—sector regulators, DBT and others, to understand what is happening in that sector. We felt that they did not have a sufficient understanding of what happened to people at the end of their visas. We encouraged them to do more evaluation to understand what was going on. They responded to that and were going to evaluate and find out in more depth what happened to people at the end of that.

The Chair: You encouraged them to do something. Is the evidence to date that they are pursuing that?

Richard Baynham: We follow up all the recommendations from our reports. That is one of the ones that is work in progress at this point.

The Chair: What is the interim evidence?

Richard Baynham: They have responded positively, undertaking to conduct an evaluation, and have put—

The Chair: That sounds a bit “long time after the event”. When were the recommendations made?

Richard Baynham: They were made last March, and the Public Accounts Committee held its hearing in May 2025.

The Chair: That was nearly a year ago, but so far you are saying that it is just work in progress and they are promising to do it.

Richard Baynham: They have published one evaluation of the skilled worker visa system. That is publicly available online. What we have not seen in the public domain yet is its evaluation of what happens to people at the end of the visas.

The Chair: Thank you. I have one other quick question, then I am going to bring in Baroness Hughes, then Lord Tope. The report may come out some time this year. I understand that you are not responsible for that. In the past, the reports were regular. Should we assume that this is just a one-off, or is there a commitment to make regular reports?

Jon Simmons: As far as I know, the intention is to reintroduce the statistical series so that it is a regular annual report on the exit checks.

Q133 Baroness Hughes of Stretford: Picking up on the comment that compliance is a matter for the individual, the Public Accounts Committee report in July 2025 said that the Home Office had failed to capture basic information on whether people leave and showed little curiosity about how the route was operating. This is in relation to the skilled worker visas. How far do you feel it is in the Home Office mindset that compliance is not just for the individual? It is a very important issue of public assurance about the whole immigration system.

Jon Simmons: I think that it is very present in the Home Office mindset. I probably would say that, would I not?

Baroness Hughes of Stretford: You would.

Jon Simmons: However, I do believe that is the case. There are two aspects of enforcement here. One is finding people who have broken the law, gone against their visa regulation in some way and who are sponsoring someone whom they should not sponsor. There is also a deterrent aspect. You do not have to find everybody to stop people from doing things. People must have sufficient confidence that there is a strong possibility that they will be caught out. That will also deter them.

There are two things going on here. It is not just as simple as there being more resources to catch more people doing bad things. It is also a case of the resources that are there and the regular checks that go on with sponsors. Any reputable sponsor will know that if they are found to be employing someone whom they should not be employing, then their sponsorship is at risk and they will not be able to then fill the jobs with foreign employees.

Tim Phillips: I want to frame how you might see things if you are working in enforcement in the Home Office. You might not day-to-day need to be very interested in the compliance of the whole set of people who have overstayed. What might motivate you is to enforce the removal of a greater number of people than last year. If that number goes up, you feel more successful.

For instance, in 2023, 27,000 people were removed. In 2024, 34,000 people were removed. You would target your efforts where you thought you were going to get the biggest return. Also, in terms of risk, we know that the Home Office puts great effort into removing foreign national offenders because it believes that there is an ethical priority to that as well as it knowing where they are compared with other people. However, often those working in enforcement will be thinking, “How can we do more than we did last year?” rather than, “How can we make sure that everybody or nearly everybody complies?”

Baroness Buscombe: Following on from Baroness Hughes, that first sentence struck me as well. Surely it should not be a matter for individuals. Surely the language that we use, our whole approach, should be turned on its head to make it very clear to the sponsors that “We are watching you” and making it very clear that people are not at risk. They will be removed and sponsors will not be able to have anyone else coming in through sponsorship. Do we not have to get tougher?

Jon Simmons: We are tough; sponsors know that. Both are valid.

Baroness Buscombe: I know people who have got away with having illegal migrants coming in. They are able to carry on working; they are able to keep their business open.

Jon Simmons: Yes, and they should not expect that, but the emphasis on the individual is that everyone who comes to this country is issued with a visa. They need to read the visa, understand its terms and that it is serious and that they need to adhere to it.

Baroness Buscombe: It is not working.

Q134 Lord Tope: How does the Home Office measure people leaving the UK? I do not know who I am asking, but you can volunteer.

Jon Simmons: Estimates of emigration are a matter for the Office for National Statistics. We are very much reliant on the collaboration with the ONS. In the past, as the committee is aware, this was carried out by the international passenger survey. Since the pandemic, that information has been derived primarily from estimates from a combination of Home Office border data and HMRC and DWP records. The resultant statistics are published every six months by the ONS. Mary Gregory was before this committee a week or so ago to discuss some of that.

We do not count numbers of people leaving. If we just counted the numbers of people departing, that would be a large number of visitors and tourists—not really very interesting for us. We try to identify whether there is a risk of people who should have left but have not left. We also rely very much on working with the ONS to establish a coherent and robust set of estimates for the overall levels of emigration, so we can see numbers departing in that way, broken down the way that the ONS published its statistics.

Maybe I should stop there to let you ask follow-up questions.

Lord Tope: No, please carry on. I think you were about to say something interesting.

Jon Simmons: The relationship with the ONS is of long standing and we have worked really closely with the ONS. It is really important that statisticians at the ONS are informed by the Home Office’s understanding of the way in which the immigration system works and of regular and irregular migration. We have spent a lot of effort in working closely with the ONS and trying to help it improve the quality of its estimates, which I think have improved a lot over recent years.

Again, you will probably be aware that a lot of revisions have been made to the ONS estimates in recent years. We have now got to the point where we understand the reasons for those revisions. We have changed the process and we think that the estimates we produce now are much more robust and stable. That is as a result of the close working between the two departments.

I will perhaps pause there and let you ask follow-up questions.

Lord Tope: Each time you say you are going to pause it makes it more interesting. You should carry on. How comparable are the statistics in the current system with those in the previous system? Is that a fair comparison or is it meaningless?

Jon Simmons: The ONS has done a lot of work to compare the two and to check differences. There is some understanding—I would not say a complete understanding—of where the IPS was particularly weak. For example, the IPS was particularly weak in measuring certain elements of immigration.

Where the new system is better is that we try to check the new administrative-based estimates with other data and our knowledge of other aspects of migration. I say “we”—we do that collectively, not just the people within the ONS or the Home Office but an expert group that is at the moment chaired by Madeleine Sumption—whom you know from her appearance in front of this committee—which also gets presented with work in progress from ONS, which I have also seen. We get to challenge the ONS on its estimates. If we see trends that do not seem credible or full data, we challenge them. ONS then goes back to look at things and we work with it to try to make sure we have the best, most robust, set of estimates we can get.

I would emphasise that the ONS is producing estimates that are based on a method of collecting administrative data and precise counts but also elements of using assumptions to refine their estimates. This is not a precise measure of the people who depart the country; it is a broad estimate of the trends. That is the purpose and has always been.

Lord Tope: But they are estimates in which you have greater confidence than you had in the previous system?

Jon Simmons: Yes, I would say so.

Lord Tope: Do your colleagues want to add anything, or even disagree?

Richard Baynham: No, I think Jon set it out. On exit checks, it uses the advanced passenger information, which it collects from ports. I think it has already been covered that, between 2016 and 2020, there was a published annual report on the exit checks that has not been produced since. The principle that we would highlight is the need to understand what is happening on your visa routes with regard to what happens to the people when they come to the country and at the end of their visas.

Q135 Lord Bach: My question is about trying to ensure that the public can best understand the migration system. What consideration has been given, or perhaps should be given, to determining the most relevant migration metrics, so the public can best understand? It is clear from the papers we have that the public do not understand and get it very wrong, particularly with regard to the last couple of years—believing that migration is going up much higher than it is and all sorts of things that the statistics prove are wrong. I think you get the gist of my question.

Tim Phillips: As the auditors of the Home Office, we obviously place reliance on Home Office data where we can, rather than generate our own from scratch. We typically find that the Home Office has lots of good analysis going on. Jon has described some of the work that has to go into transforming data that is in a day-to-day management system into something that is good for the purposes of national statistics and is clean and reliable. We see that going on.

The quarterly statistics about immigration and asylum that are published and with which we are all familiar are a long run of statistics that are broadly stable. Obviously, there are changes at the margins over time. An important change, which has been felt and noted by many people, including parliamentarians, since 2020 is the loss of the picture of exits.

I would draw attention to another aspect that is, I think, related to the public’s understanding. We have suggested that, where possible, impact assessments for policy changes should be published as fully as possible, so that the public can understand what the Home Office believes the modelled future will be if a policy comes into place.

We have a worked example of what can happen if you do not do an impact assessment. When the Home Office changed the rules for care workers during the pandemic to allow more care workers to come in through the skilled worker visa route, it did not do an impact assessment. That went on to have a really dramatic effect on the assumptions that had been made and the reality that happened. The original assumption for skilled worker visas in 2020 was that 360,000 people were using it. It ended up being used by 930,000. Obviously, if an impact assessment had been done, it might have revealed some of that and caused some thinking about the desirability of that.

My final point on that is that it is important to say that it is not just about the numbers, but rather it also completely changes the compliance effort you need to make—for 900,000 people rather than 360,000. It changes what happens with regard to some people who at the end of their visa claim asylum—if you have a population of 900,000 people, the number of those claims will be greater, et cetera.

So, away from the headline figures about how the line graph is changing over time in this or that category, getting into the meat of, “What will this policy change result in?” as best as we can foretell or think can contribute to a better and more complex public debate.

Jon Simmons: I definitely agree with the plea that policy should have good impact assessments. I do not think the impact assessment process is particularly easy to work with; it can be quite burdensome, which is part of the problem if you want to produce things quickly. But the idea that one should assess the impact of policies is clearly right.

Certainly, although there was not a full published impact assessment on new immigration system in the way that we might do for other policies, it was certainly assessed and there was a lot of surprise. There was, for example, a lot of discussion with the department of health about introducing the health visa and trying to understand the department’s view of the sector and where there were gaps and job vacancies, and how many people it expected to need to bring in. It vastly underestimated the number and we believe that some of that was because it was a very liberal system that allowed some people to come in on the health visa who perhaps did not then do a full health job. The Home Office has been cracking down on that with the help of the Department of Health and Social Care.

It is probably worth going back to how we explain things to the public, which is really the core of the question. As you point out, we know from a recent survey that one in six of the public thinks that levels of immigration are coming down at present, even though they have been coming down for the last two years, which has certainly been very clear in the ONS’s statistics. It takes a long time for the public to get this message and we, as statisticians, need the help and support of lots of other people, including parliamentarians and, obviously, the media in communicating this consistently to the public in a way that is credible and believable.

It is not just an issue with the immigration statistics. I used to run the crime statistics. Crimes have fallen in the UK since the mid-1990s. Thirty years on, I do not know whether there is yet a majority in the public who believe that. Maybe there now is; I am not sure. But it is pretty clear from the data and statistics, and we all need to work together to, hopefully, convince the public in a credible way about what is actually going on.

Lord Bach: Forgive me for interrupting part of your answer. Could the Home Office itself do more in making it clear to the public which way it is going?

Jon Simmons: There were good reasons why the ONS and the Home Office published immigration statistics on a quarterly basis. I think that started in the mid-1990s—maybe it was the early 2000s, I am not sure—but it was essentially because there was lots of random statistical information we presented to the public on the subject of immigration, and ONS and the Home Office got together and said, “We should provide a more coherent picture to the public. Why do not we just publish the information on the same day once a quarter, and at least we can then tie things together and help explain things more clearly?” That was an entirely reasonable objective at that time.

Prior to the pandemic, when ONS was publishing estimates of net migration on a quarterly basis, that was somewhat misleading, because the quarterly estimates were essentially an estimate of an annual change in net migration. With a year’s worth of data every quarter, essentially every quarter it was a repeat of three-quarters of the information you got the quarter before. I do not think that was being honest with the public; it just encouraged lots of newspaper headlines every single quarter. It was not really a productive discussion.

Frankly, I would rather that we move to a system where we publish the big numbers once a year, as most countries do on immigration. That relates to the fact that immigration is just one of the key elements of measuring population change, along with births and death; it is a key contributor. Essentially that is what we are trying to do when we use population data to estimate demand in local councils, for example, on housing or on schools. We do that once a year with good annual estimates of the population. We do not expect that to change and be revised every single month or every single quarter.

One of the good things that the Home Office did over the last couple of years was introduce a model where, if we were to move to a slightly more longer-term perspective or an annual perspective on immigration trends, we can still publish on a more regular basis things that are very high profile and of key relevance to the public. Obviously we publish daily statistics on small boat arrivals. As a statistician, I am not keen on daily statistics; I think weekly or monthly would be sufficient, frankly. But we do, and they are available to the public.

One of the best things we introduced recently was the monthly reporting on visa trends. For the last couple of years, there has been a monthly report on key visa numbers for work visas, study visas and family visas. That provides a really clear, very short and very focused presentation to the public on the trends in the main categories of legal migration. That sort of model would be possible to roll out for other things that are of key interest to the public, so that we do not have to put so much effort of our statistical teams into this regular quarterly cycle. They could actually produce those high-level numbers relatively simply and focus on a big annual report on the key big picture, with all the data tables that they currently produce, but also do more analysis in the interim over the course of the year on other aspects of migration, which we simply do not have the resource and time to do.

So there is a different way of doing it that would be very productive. It is controversial, of course, because any change in statistics, particularly such important statistics as the immigration statistics, is controversial and reliably liable to be criticised. But it could be very positive.

I will just carry on a little bit on what good measures of migration look like. There are essentially three categories of measures. Measures of migration flow are the net migration estimates, annual numbers of people coming in and annual numbers leaving. It is like the visa statistics; that is numbers entering the country, essentially. There are also measures of stock, which are underused and would be really valuable. It would be interesting to know much more about the numbers of migrants that are resident within our local communities, for example, and we do not put as much effort into understanding that—the composition of our communities and which migrants are going to different places. We have done some of this in the past, and it is done in pieces, but it is not done routinely and it is not done in as much depth. We have done a bit more—for example, some good analysis is now being produced on the migrant contribution to different economic sectors and different businesses. That is great. Again, that is a measure of stock. We are looking at the composition of particular sectors. How much do migrants contribute? How needed are migrants in those particular economic sectors?

We should do more on stock analysis than we do on flows. Obviously, there has been a lot of focus and a lot of public discussion has been on flows. The net migration target, introduced in 2010, was all about the measure of flows. But we have seen, particularly in the last few years actually—since the pandemic—that flows can be very misleading. We saw that very large spike in net migration and immigration. We know the reasons for that. Some of that was due to policy change and the relatively liberal immigration system that was introduced, but some of it was also a result of the pandemic and people delaying their travel. For example, if you wanted to get married during the pandemic, you could not. You could not come here, so you postponed the wedding. If you wanted to take up a job, again, you could not; you had to postpone that and you started later. It was the same with study, and so on. We have seen some of that with the sharp spike—lots of people coming, fulfilling their ambitions—and then of course things go down again, as we have seen in the last couple of years. Flow measures can be very misleading, and there would be a lot of value in putting a lot more effort into understanding our stock of people that live in the country and what they actually do.

The other measure which is very important is operational data. We need to be aware that operational data—things the Home Office does within the immigration system—are important for understanding the performance of the department. They are incredibly valuable. There are examples like the monthly visa statistics, which have been really informative in helping us understand trends in the number of people under different categories that are coming to the UK on a very regular and very immediate basis. But they can also be misleading. For example, a classic problem in enforcement statistics is: if we see the number go up, is that a good thing or is that a bad thing? Does it mean that there is more activity to address a particular problem and therefore the number rises, or does it mean that there is more of that problem, so more to find? It is a classic problem in policing statistics, as with immigration enforcement. There are issues with operational data and it needs to be seen in context.

Tim Phillips: I just wanted to associate myself with the remarks about stock and give one additional way of looking at it. If you look over time at the objectives that the Home Office has set for the immigration and asylum system—I do not have them verbatim as they are today—often they include protecting access to the privilege of citizenship and access to the UK. They include often trying to reduce the number, and those are things that are often captured through the flow data that Jon was describing. One area where we have typically seen less data to illustrate performance is on supporting the British economy, which is one of the objectives of the immigration system. I do not have an easy answer as to how you illustrate that through data, but it is a question for the Home Office: how does it illustrate its performance in meeting that objective that it has set itself?

Q136 Lord Hogan-Howe: My question is just, Jon, on that point about performance management. One of the benefits of performance management is that it should drive the data to get better. A concern I would have is that it seems to me that you are able to establish the broad flows or stock—whichever it happens to be—but there are some really important tactics in the middle of that. Do the people who have committed murder and are just about to leave get out or not? If I was Home Secretary, I would want to know where the gaps were. Did the police not tell the ports? Did the ports not find them? It is an indicator, not a strategic issue in that sense. We talked about whether or not, at the end of somebody’s visa, somebody goes looking for them and just hopes that they are going to find them one day by having a random search of various employers who happen to employ people who have gone beyond their visas. That is the area I am not sure about—why the performance management is not driving the data to get better.

Jon Simmons: I think it is. An awful lot of data flows around the immigration system. I have said in public that I believe the UK publishes more information about the operation of its immigration system than literally any other country in the world.

Lord Hogan-Howe: Just stick with people leaving. We do not know, because it has not been published, but it does not appear to be getting better.

Jon Simmons: No country has good measures of emigration as far as I am aware. It is a classic thing that is difficult to measure and, to be honest, most countries actually are not that fussed. They just think, “People are gone. Great—not our problem any more, not our issue”.

Lord Hogan-Howe: Well, it is if they have murdered your child.

Jon Simmons: Yes, absolutely. There are separate issues around specific aspects of the operation of the immigration system which require different sorts of solutions. The Home Office has committed to publish more information on foreign national offenders, and that is something the team are working on and they expect to publish a report in the next few months. It is a very specific and key issue there. In the past, people talked about, “We have removed so many foreign national offenders”. Well, for me as a member of the public, that is not that interesting because I want to know, “Have you removed the most serious ones—the ones who committed the most serious crimes?” We need to get into that, and to understand that there is quite a complex task of matching the immigration data we hold with that from the Ministry of Justice, for example, on people who have been prosecuted, the sentence that they are given and so on and so forth. It is a complex statistical task, but it is one that we can achieve if we focus on specific areas of public interest.

The Chair: I am just going to remind all colleagues that we are now half way through our time, and we are nowhere near half way through all of the issues we want to raise. I am going to ask my colleagues and our witnesses if we can be fairly brief. As I said earlier, that will give you the opportunity to write to us with all the things you did not get to say. However, we will stick to this issue for a few more minutes. Lord Dubs?

Lord Dubs: In view of what you said, no.

Q137 Lord Anderson of Ipswich: I have a very contained question, following on the theme of people leaving the country. Every time that I leave the country, I have to provide advanced passenger information to the carrier. Richard, you mentioned that. Are we getting all the value that we can out of advanced passenger information, which ought to be able to unlock an awful lot of details about the person travelling out? If we are not getting the value out of it, are there any other questions that carriers should ask that would improve our attempts to monitor who is leaving the country?

Jon Simmons: I am sure that there is more information that we could collect at the border, but there is always a balance regarding the amount of information. We got rid of landing cards because they were a burden that was unnecessary and not adding a great deal of value to the information that we collected. There will always be a balance when you design any statistical system, between the demand for information and the need for it.

We get a lot of information from API data. When someone comes to the attention of Immigration Enforcement, it can check very readily whether there is a record of them departing. That is readily available. The issues are with the fuzziness around the edges, such as the common travel area and data-matching issues. A relatively small proportion of individuals can have a big impact when you are looking at overstayers—who are only a small proportion of the total number of people who come to the UK.

The API data is well used. On ways in which it can be used, we published some research recently on refugees who have been granted refugee status in the UK and been allowed to stay in the UK permanently. We looked at lots of outcome indicators for those refugees and through lots of complex data matching with lots of different systems. We looked at their housing situation, their health situation, their employment and so on. One of the things that we looked at and we have published information on is whether refugees who are granted status in the UK leave the UK. We can use API data to do that.

I can tell you with confidence that of the refugees between 2015 and 2021 whose status was granted through an asylum application, about 6% could be shown to have departed the country and on what appeared to be a long-term basis. That does not mean that they have gone back to the country that they originally came from. They may have families somewhere else and have moved somewhere else, for example.

For resettled refugees, the figure is much lower. Only 2% left the UK. We did all that through matching the information that we hold on refugees with the API data and drawing various assumptions depending on how long people have been away from the country and so on.

Q138 Baroness Cash: I would like to hear a bit more about data collection. Jon said that we do not have certain data and nor do other countries about who is leaving, nor are some other countries bothered. That is not the same as evidence that we have heard. I am concerned about data sharing across the Home Office, picking up on what Lord Hogan-Howe was asking about. In the last week, we have had some very serious investigative reports in newspapers about large-scale fraudulent schemes selling work visas. What do we do about any of this? Where is it collected? Where is it investigated? I want to understand better what the data sharing is across the Home Office between our border controls and our police and what we could do better. It is a very big question again, but it has emerged as a result of evidence given this morning.

Jon Simmons: It is a very big question that would probably be better addressed by operational teams within the Home Office—Immigration Enforcement and so on. Data sharing has come on leaps and bounds since I started managing the immigration statistics 16 years ago. We do a lot of data sharing with the Home Office. I gave you a very simple example of the work that we did as researchers with refugees just now. When things are brought to the attention of the Home Office Immigration Enforcement operational teams, while this is not a statistical task, there are good analysts in the operational teams who look at the data and draw some assessment.

It is complicated. I do not know the degree to which we look at the information that we hold on particular sponsors and the visas that are associated with those sponsors and then know about the departure of visa holders, but that kind of connection can be made with the data that we hold now. We are reliant on the information that comes to the attention of the Home Office through Immigration Enforcement operations, but sometimes the media do a good job and discover things or are told things about the public—which may be the first time that the Home Office finds out about it. There then has to be a process. We pick that up as an important issue, see what the data can tell us, and investigate. There is then a process of applying that.

It is not automatic that we will find out about everything. Often we read about things in the media because one or two stories make us think, “Oh, hold on. That may be a signal of something much bigger”. We then need to investigate it. It works in lots of different ways.

Baroness Cash: I am conscious of an expert on this committee sitting opposite, but it has troubled me for some time that we may not have adequate data sharing between border control and our domestic law enforcement. What is your view about that?

Jon Simmons: I have been involved in that aspect of the work for some time and presented to colleagues at the Met many years ago the kind of data sharing that we can do. It is complicated because the police are involved at the coalface of crime. When somebody commits a crime, immigration status may not be the most important thing. It may be something that the police can resolve relatively straightforwardly and does not result in a prosecution, further action or conviction.

As a statistician, I am much more interested in giving the public a clearer view of the more serious criminality, of foreign nationals who do get prosecuted being convicted, resulting in prison sentences—some of which are very serious. We have talked about murderers, rapists and so on, who we want to be very clear are being removed from this country once they have completed their sentence, and in some cases before that if they can be punished overseas.

It is very easy to get distracted by the broader, bigger numbers. I would rather that we as a country focus much more on the serious elements of criminality and robustly say that we are addressing the more serious elements.

Tim Phillips: I do not have a great deal to add on this, but the concepts of things being material by value or nature is relevant here. We talked earlier about trying to do more removals than were done last year. That is a driver for the teams in the Home Office. However, we know from our work that the Home Office will often spend years trying to remove a particular foreign national offender who is perhaps asserting rights and raising appeals and trying to prevent removal because they believe that it is in the public interest for them not to be here. An interesting question for the Home Office would be how, day to day in this year, it strikes that balance—to go after the target while getting a bigger number in total than last year. It is prioritising a very intensive, heavy caseload with a small number of cases that it believes are so important that, even though it would mean only one person gone, it is worth devoting resource to it. How does it make that balance-of-investment decision? It would be really interesting to hear about that.

Q139 Baroness Buscombe: I got a tip-off from my local police and crime commissioner and so asked a Written Question: “To ask His Majesty’s Government how many convicted foreign nationals awaiting deportation have been released from HMP Huntercombe in the last two years owing to the Home Office’s failure to provide the necessary deportation documents on time”—in other words, people being released rather than deported. In essence, my answer says that, yes, people are being released in certain circumstances and that it is not always possible to match the paperwork with the release of these foreign national criminals. We are not telling people in the local vicinity that we have these foreign national criminals who are going back into the community. There is something very wrong going on here, is there not?

Jon Simmons: That is probably a question for other people in the Home Office.

Baroness Buscombe: It is important to state this. As Lord Hogan-Howe said, the problem with foreign national criminals is serious but it seems that we are letting them out into the community because we have not got their papers straight.

The Chair: That is an issue we will pursue. Thank you for raising it. It is now on the record.

Time is tight. We wanted to ask Jon about international comparisons. I would be enormously grateful if you could take a bit of time and drop us a note about that. If either of the other two has anything on that issue, we would be very grateful. Can we move on to a rather different issue with Lord Anderson?

Q140 Lord Anderson of Ipswich: I have a series of linked questions for the National Audit Office about cost. We have been told that the costs associated with all stages of the visa settlement and citizenship process are well above those applied by comparable nations and that they are profit-making for the Home Office. What is this money used for? How much does it bring in? Which are the profit-making fees? Do they include, for example, the national health surcharge? Are there others which are not currently profit-making but could be profit-making? Finally, what would be the consequences of charging for cost recovery rather than profit or, alternatively, introducing fee waivers for certain categories of applicant?

Tim Phillips: It is often good when you get a question and discover that your colleagues have published a report on the exact thing recently. We published a report last year on fees across Government, so I can talk with some authority to the position last year.

You are right: the Home Office charges applicants for visas more than the cost of processing those visas. It has a specific waiver from the Treasury to allow it to do that. The principle in Government is that a fee should be set at a level that is expected to meet, but not exceed, the cost of delivering that service. In our work, we looked at driving licences, courts and tribunal fees, passport fees and immigration visa fees. All are set at a level that is 100% of the cost of processing, as understood at the time of setting it, except for immigration fees, which are set at 212%. It is way beyond. They are intended to generate revenue.

The agreement with the Treasury is that all that money has to be reinvested in the immigration and asylum system. The total raised in 2023-24 from immigration fees was £2.6 billion. That is considerably less than the total cost of the asylum and immigration system. The asylum system alone costs £5 billion a year. The Home Office is meeting its arrangement with the Treasury that allows it to charge more. The total that comes in is less than the cost of the immigration and asylum system—which is the basis on which it was allowed to charge it.

It is interesting to note that of all the fees that we looked at, the Home Office has been most likely over the last 20 years to change the fees for immigration visas. The price went up 13 times between 2005 and 2025. That compares with nine times for passports, six times for court fees in the Ministry of Justice and three times for driving licences. This is a highly price-sensitive thing that has gone up quickly in recent years with the intention of generating revenue.

The various surcharges operate slightly differently, one from another. The health surcharge generated £2.4 billion last year. It is set at just above £1,000 per year. That money goes to the Department of Health and Social Care in England or the relevant department where the person is applying for a visa to come to, in Northern Ireland, Scotland or Wales. Whether that is profit-making is difficult to say because we all access the health service in different ways at different times. Research indicates that the average immigrant is younger, healthier and therefore less likely to make a heavy use of the health service. I do not know whether that means that, once you add in taxes they might pay et cetera, they are overpaying. You can see that it is not set up with that intention, but the money does go to the health services.

The skills charge is different. At the time of the report, it was £1,000 per year. It has just gone up to £1,320 a year. That is an example of a surcharge increasing. As far as we could see, that money does not go to the skills department; it does not go to DWP; it stays in the Treasury. It would be interesting to know whether there is a hypothecation there for what that money is spent on. In the White Paper there is a proposal of a levy on English universities for the income that they generate from foreign fees. As you would expect, there is not detail on how that would operate—to whom would that money be paid? What would it be for? Is it just for general funds? Is it for the Home Office to defray the costs or for the Department for Education and higher education establishments in some way? I will stop there.

The Chair: That is enormously helpful. We can look in more detail at the figures.

Q141 Lord Henley: I will stick with the NAO’s skilled worker report, which was critical of the Home Office for failing to consider the consequences of policy proposals. Is policy being generated that cannot be implemented? For example, does the Home Office have the ability to identify whether migrants have met the requirements of shorter route to settlement—for example, meeting income thresholds—or check people’s immigration status every two and a half years? Is enough time being allowed to consider responses to the consultation?

Richard Baynham: I will start, because I worked on the report. We found that there was not enough collaboration between the Home Office and relevant departments to understand the potential consequences and impacts of opening up the route to allow more workers in. That brings the responsibility on to the other government departments. They need to bring the evidence in the labour market that they were considering—where are the gaps, what are the shortfalls, what are the domestic strategies to grow the workforce in those areas, how does immigration fit in and complement that? Over what period do you want to use immigration to fill skill shortages in the country? We pointed to the need for—

Lord Henley: Which other government departments?

Richard Baynham: It depends on the sector. Skilled worker visas cover right across the board from healthcare to fishing to the MoD. The sectors need to work collaboratively to understand the balance between domestic workforce strategies and the need for immigration and what that means for the visa system that you operate. In the checks that you then make, you must go to the eligibility criteria that you have set. It is important then for the caseworkers to understand what evidence they are going to collect and where they will get the evidence from to check eligibility against the set criteria. They introduced new checks in the skilled worker visa to allow caseworkers to make those judgments more easily. If you do not get that right, there is a burden of evidence that the caseworkers have to get to prove or test the eligibility of the applicant.

It is important to understand with other government departments what the need is and what immigration can do. It is important for caseworkers to get the eligibility criteria right and understand the evidence that is needed to test eligibility against those criteria.

The Chair: Summing up what was said in an earlier discussion and what has just been said, it seems that although some work is done, for any new policy, particularly in the areas that we are talking about, there needs to be a detailed impact assessment and, in advance, a decision made on how an assessment will be carried out to see whether that predicted impact has been achieved. At the moment that happens on a rather ad hoc basis—for some things and not others. Is that a fair summary of what I am learning?

Richard Baynham: Yes, I think that is fair. You need to undertake modelling to understand future demand, because that has an impact on compliance, compliance resource and how you assess the risks in the sector. You have to understand the operational impact as well, in the flow of cases through the system. We gave them credit for the workflow modelling, and they would reallocate resources. But that was finite. When we saw the skilled worker route opened up to the care system, the numbers increased massively. That has an impact on operations. You need to understand the flow of cases from a compliance risk perspective and an operational perspective.

The Chair: You gave the example that the number of those who came in was threefold the predicted number. In your conversations with the Home Office, is there any evidence that we are moving more towards an impact assessment approach in the way that I have described? It seems a large number. Because it seems to me that a large number of proposals have been made in recent times—Green Paper statements by the Secretary of State and so on—for which there appear to be no impact assessments. There is very little data suggesting what the impact of these proposals might be, just claims that they will make things better.

Richard Baynham: Our work, really, has encouraged them to do the analysis up front.

The Chair: I am sure you encouraged them. That is not my question. My question was: given all of your encouragement, both in the past and now, have you got any evidence that this is happening?

Tim Phillips: I think we do not know. I will deal with the things that are subject to consultation through to 12 February—the measures about earned settlement. From our previous work on some of those measures and the way in which indefinite leave to remain is proposed to change into a much broader menu of lengths of time linked to certain things that would need to be demonstrated through documentation and data, it is clear that if you were to proceed on that basis, modelling would need to be done to understand what that would do to the workload of the finite resource that is currently in the Home Office. I cannot tell you whether we think that is taking place or not.

The Chair: Common sense tells us that the modelling of that needs to be done, but one would assume that the modelling has been done before a proposal has been put forward. My question to you is: from your organisation, do you have any indication as to whether that work has been done?

Tim Phillips: I think we have evidence from the past that sometimes that work has been done—looking at the consequences for workload and how the system functions—and presented to senior decision-makers and Ministers, and sometimes they would decide, “That’s very suboptimal but I still want this policy”. We have seen that. At other times, as we know and have discussed, the evidence base of what will happen—the modelling—has not been there. So we do not know in this instance. I cannot answer that.

The Chair: Okay. We will turn to Jon.

Jon Simmons: There is a lot of modelling that goes on within the Home Office, and the Home Office also has access to the Migration Advisory Committee, which does a lot of analysis of economic sectors, for example, and demand for labour.

I think we are talking about two things here. One is the impact of policy changes on society or the population or a sector, and the other is being able to model the impact on Home Office operations and the ability of the Home Office to deal with the number of visas it might receive, for example.

Both of those things occur. I do a lot of modelling on the demand for visas and where that will rise. It is complex and difficult, and you cannot always predict things, but they do the work, and that drives recruitment, for example. If they know that a particular area will require more caseworkers, they can plan for it. That might mean moving people from other types of cases or it might mean recruiting more caseworkers. Both of those things can be seen to have occurred in the past.

The other aspect of the modelling is, of course, the impact of things, which is all about impact assessments now. All I would say is that the White Paper set out a strategy—an approach—but it did not set out specific policies. That has to come and, as part of developing those policies, the analysis of the various impacts should be the bread and butter of what goes on in the Home Office and should occur, I think.

The Chair: You used the word “should” on more than one occasion in that paragraph.

Jon Simmons: Only because I am not there and I am not involved, so I cannot guarantee it.

The Chair: I know you are not, but you can tell us whether it has happened on a regular basis in the past.

Jon Simmons: Oh, absolutely, it has happened in the past.

The Chair: So you would be confident that, before final policies are put forward, they will have had a detailed impact assessment and procedures would be in place ready to monitor whether that impact has happened or not?

Jon Simmons: That is a very good and very focused question. I will try to answer it honestly. I think that there are lots of examples where that has happened. I think we can also find lots of examples where it has not happened quite the way it should. A more regular approach to policy-making that engaged with that and that we could guarantee was consistent, as well as a published impact assessment on the impact of policies as a routine part of policy, is a reasonable expectation that we should have.

The Chair: “Reasonable expectation”—I will bear that phrase in mind.

Q142 Lord Moraes: The inquiry is also about integration, so this question about where these stats lie is mainly to Jon. What do you have direct control of in the Home Office? Where integration stats are spread among other departments, how they are co-ordinated? I think earlier you mentioned in passing that there were some stats on migrant contributions, but also people are concerned about benefits take-up and so on. Where are these stats located and how do departments work together?

Jon Simmons: Integration is a very diverse thing, and a lot of people struggle with trying to measure it exactly and how to measure it. In 2019, the Home Office published a report called Indicators of Integration, which was a very thorough report on the way in which you might measure integration. Essentially, it did two things—it provided a very structured way of measuring integration, with clear examples of ways to find the data that might help you to measure it, and it also provided a structured approach to how you might address integration. So if you wanted to write a strategy for integration, it provided you with a framework to do that.

It was a popular report and very well known in lots of different areas, including in local government. I was also very pleased to hear from Welsh colleagues, for example, that they used the Indicators of Integration report as the basis for their Nation of Sanctuary, which they published a few years ago. I also heard recently that Norway used it as the basis for its strategy, and I do not think that it is the only country in Europe that has done so, because it was popular both within the UK and internationally.

As I say, it provides a structured way of measuring integration and lots of different examples of where data and different types of metrics are available. It will vary, depending on the sphere, obviously. Data on work, housing, education, health and so on is collected in lots of different places. One of the things we have started to do within the Government Statistical Service is to try to use data matching to provide a more coherent picture of integration outcomes, particularly for refugees at the moment.

One of the things we had in mind when we designed the Indicators of Integration report is that it should not just be focused on refugees, because there is loads of academic work that has looked at refugees and integration over the years and we thought, actually, integration in our communities could be as relevant to other types of migrants and potentially other people as well—ex-offenders, for example, because exactly the same issues apply to people who have left prison after a number of years, and they too need support locally to help them integrate back into our communities. So that report is available and is able to support a range of different initiatives.

One of the things we have published recently is some emerging findings from data matching that looks specifically at refugees and their integration. We have matched a cohort of refugees that the Home Office has managed and granted status to. We look at both people who applied through the asylum route and people who came as resettled refugees. The difference is very interesting and had never been done before. There are very different qualities of data. We know a lot more about resettled refugees—and the data quality is much more reliable for them—than we do about those who came through the asylum route, where a lot less information is collected and there is a lot less follow-up. Obviously, when we resettle refugees, the Home Office follows up on people and supports them for the first year of their resettlement. We do things differently for people we just grant refugee status to.

We did publish statistics on employment rates, for example. We looked at employment rates over the first eight years of people’s status and we could see very clearly that, initially, people in both groups had very low employment rates but they were significantly higher for people who came as asylum seekers rather than people who came as resettled refugees. Resettled refugees struggle to enter the labour market much more. Over the course of the first seven or eight years, which is the period we looked at, those employment rates rise, but certainly to nothing like the extent we would expect for other members of our communities. Employment rates for refugees are still very low. However, we do not just look at employment rates. We look at housing, education, achievement, health, all those things. I will not go into all the detail here because there are published reports now, but they provide a wide range of information about refugees.

This is such a complex issue. We could spend the next hour talking about it, but there are other issues that you might want to ask about. How do we measure how well migrants integrate in our communities? Do they get on well? Are they abiding by the law? These issues are very complicated. This needs to be looked at in very distinct ways. It is a matter for bespoke research rather than regular statistics, but it is something that we look at from time to time in different ways.

Q143 The Chair: I will come back to you, Lord Moraes, in a moment. Given that we have strayed into an issue that we were going to look at in a bit more detail, Lord Hogan-Howe, is there anything that you want to add on that?

Lord Hogan-Howe: Is there any data or evidence available that shows when integration is working well?

Jon Simmons: That is a very interesting question. I do not think that there has been enough research on that question. It would be very interesting to carry out a survey, for example, of different local communities and ask them whether they think that there is an issue with the people who have arrived in their community and, if so, on what basis.

The Chair: Is that an issue that you have been looking at?

Jon Simmons: No.

Tim Phillips: We have not looked specifically at that. We have found in the course of our work many issues around the integration of data and how that can affect the handling of cases. Behind each case sit individuals. Pertinently, a case in the asylum system has no unique identifier between the Home Office, the Ministry of Justice and local authorities. A workaround has been created between the Home Office and the Ministry of Justice to ease that friction, but there is not such a workaround with local authorities. If you get granted asylum or your case gets finished, as a single person you have 28 days to vacate Home Office accommodation. If you are a family, you have 56 days. The friction point there of letting the local authority know—

The Chair: I am sorry to interrupt you, but I want to come on to this issue in a bit more detail. I think Baroness Buscombe wants to ask about it. My question was whether any work has been done by you on how the Home Office might gather data that would best assess the effectiveness of integration policies.

Tim Phillips: No.

The Chair: Thank you.

Q144 Lord Moraes: I am flummoxed by that last one. I want to ask about hard integration data. If you talk about integration generally, it becomes complex and vague, but if you want hard data, what does hard data mean? If you are a business owner in hospitality or the care sector and you are a migrant, both want to know, with easily accessible stats, what your contribution is. Do we need that contribution? Do you claim benefits? Are you less likely to claim benefits or access to health service? These are hard integration stats that concern people. Where are they located? Are they easily accessible? Do they go into this? As you said earlier, we do not know where your stats go. Migration is falling; some crime stats are falling but nobody would know this. It is a Confucian state where we must wait 40 years to get to that point. I get that your stats are not going to the right place, but where are the hard integration stats located? Are you satisfied that they are of decent quality?

Jon Simmons: No, we could make a lot more effort. A lot of this is about the issues around stocks of migrants in our communities, what we understand about them, what they are doing, how they survive, how they are being received and all these issues. It is very varied. There are lots of different types of stats that we might look at.

Part of the problem with the UK statistical system is that it is very well established—it is very old—so we are reliant on lots of well-established systems that have been developed by lots of individual departments at different times. As statisticians, we are dealing with lots of stovepipes of information that we are then trying to match together. We would like a more person-centric system where, when the Home Office grants someone refugee status, we have a little identifier and can track those people onward when they are looked after by local authorities, enter the benefit or housing systems or whatever. We can then marry that data up. The UK is not in that place as a society for historic reasons. We are having to play catch up.

There is data on lots of these issues available, dispersed in lots of different places. The DWP, for example, recently published information on benefit claims by people with different migrant statuses. There is probably less information from the health service because traditionally, if people come to accident and emergency, nobody is going to stop people and ask for their passport. The job is to fix people and get them going again. That sort of information is quite hard to come by and we do not have a unique identifier or way of reliably matching individuals who turn up at A&E, for example, with the immigration record. That is really difficult to do. It is the same on the streets. If police come across someone and do not arrest them, it does not become a court record. They do not enter the prison system. The information that we hold to match people with is that much less, so it is more difficult.

If I was prioritising the use of statistical resources, I would prioritise on the more serious cases such as the more expensive hospital episodes. Perhaps we ought to stop and ask people about their status. Regarding more serious crimes, we should find a way of identifying those who are convicted of them and reassuring the public that we understand who those people are and what happens to them next. That does not mean everything but is my perspective as a statistician.

Q145 Baroness Buscombe: I find this deeply depressing. We must be the only country in the world that does not ask questions of people when they go into hospitals, et cetera. Wherever else that we go, we are asked questions and we have to pay. That is just a side point to what you have been saying.

My question has in part been answered already: what are the challenges with linking Home Office visa data and other data sources? For example, could the Home Office feasibly link visa records with public service data?

Jon Simmons: We have done that in various places. DWP linked visa information with the information from universal credit claims and published that recently. We have done the same for refugees in a variety of different environments. It is not something that we can do routinely. ONS had big plans for an integrated data system that would enable them to bring in data from across the public sector and make a start at linking that in a more routine way. They have abandoned those plans now because it would have been extremely expensive and probably not have delivered anything quickly in the short term. That is no longer part of our planning for the statistical system.

We are having to deal with things in a more ad hoc way, focusing on the things that are the most important. If we genuinely believe that issues of the use of the health service are more important, then we should be asking the Department for Health and Social Care to find ways of capturing that information on a more routine basis. There are challenges with that, but it is something that we could focus on as a statistical system.

Baroness Buscombe: I was struck by your suggestion that we should have a more person-centric system. That might be part of the solution.

We have a Home Secretary who is wanting to be radical, wanting to find solutions. Should we suggest a moratorium on legal migration until we get some of this sorted out? We have rising unemployment. I was struck by what Richard was saying about the lack of link between looking at our needs and giving people visas where there may not be jobs now. We need a more flexible system. Can you respond to that please?

Tim Phillips: I will just respond simply: the National Audit Office does not comment on the merits of policy proposals. We only look at the policies that the Government are pursuing at a given moment in time.

Baroness Buscombe: Okay, so let me try another way. Can I put it to you, Jon: do you think it would be sensible to have a moratorium until we sort some of this out, because there is a genuine lack of data?

Jon Simmons: I think a lot of employers would find that problematic. We do not have high unemployment at the moment. We have relatively high employment rates in the UK, certainly compared to levels we have seen in the past. Yes, at the margins they may be up or down a bit, but they are still relatively high. It does mean that, in some sectors, some employers do struggle to fill jobs—we know this—so they use foreign workers as a way of filling those jobs because there is not a ready pool of people just waiting for employment in the UK in their area who have the right skills for their sector.

We can debate that. We can obviously do lots more in developing skills. The Government are committed to doing that, I know. We can do a lot more to get people, for example, who may not have been in the labour market for a while—maybe for reasons of, I do not know, disability or other things—trained or supported in a way that enables them to re-enter the labour market. There are lots of things Government can do. But from my perspective, and certainly in all the analysis I have seen, that does not necessarily mean that we would not still need to allow other people from outside the UK to come and fill jobs on occasion.

Obviously the Home Office has the Migration Advisory Committee, which does routinely analyse these issues. It is certainly open to the Home Secretary to put a question to the MAC to ask about these sorts of things. They would have the skills and professional knowhow to look at these issues.

Q146 The Chair: Just before we move on, Tim and Richard, you have said you do not comment on government policy—the look on your face was fascinating when you said it—but you were talking earlier about the integration of data, which we have heard about from Jon. We appreciate that there are huge difficulties, and mega computer systems to try and do it have been abandoned and so on. What is your assessment at the NAO of the Home Office’s attempts at linking other data to help it fulfil its policy objectives?

Tim Phillips: The most frequently used join is the join into the Ministry of Justice, because so many people, after an initial decision, appeal and those appeals go through to the Ministry of Justice. They have created workarounds and there is an aspiration for something that is more like a single ID. In the report we published just before Christmas, what we see is quite a lot of potential sources of useful intelligence and information not being exploited. The Ministry of Justice does not know what the caseload is at the Upper-tier Tribunal. It does not know how many of the same people are putting in repeated appeals; it could not tell us the number of those people. We saw local evidence of Home Office teams learning the lessons from cases where they had failed—where the original decision was overturned—but we did not see a national strategy for learning the lessons, and that is obviously a vital source that you can potentially feed back into how casework is being worked. This is a universe of activity, and you do see promising things in many places when you go and look at it. We have seen that there is now more testing to see if the Home Office’s case before proceeding to object to an appeal is well founded. We have seen things like that.

I will just complete the thought I started earlier. On the local government side, there is not that join—so, decision made; the clock starts running, 28 days till the person becomes theoretically homeless in the local area and the local authority has a duty to house them. Because there is no seamless flow of data, when does the local authority learn of that duty to house? How does that come about? That is a friction point. Another friction point which we were interested to discover is that the Home Office’s contractors to create asylum accommodation are often in competition locally with local authorities to procure housing. That is undesirable, because you can imagine it bids up costs and all the rest of it. There is more to do in integrating data.

You will be familiar with the “all hands to the pump” attitude that happens in the Home Office. In the last two years it has been looking at getting initial decisions on asylum cases to go up, because they were taking a very long time—and they have shot up. But the problem with that in any system—it is the same problem if you suddenly turn off the tap—is that the administrative people reorientate themselves on the thing you have told them to go and do, but work flows to the next stage and suddenly the backlog you have moved moves forward to the next stage, and at some point that crosses into the boundary of the Ministry of Justice. Then you find that you need a tribunal system that has many more judges than it currently has, where many people are beginning to represent themselves because there is not legal aid available, and that slows down the process.

This is not to criticise the effort to increase the number of initial decisions. This is just to say that the asylum system is a motorway, and people are being paid for by the taxpayer for the entire time they are on the motorway. If you widen one junction, that does not necessarily mean that people come off the motorway. You need to look at it as a whole system—all the junctions; all the friction points; and how you are working as a whole Government, because it does cross departmental boundaries, to make that system flow so that you are not making people wait for longer than they need to for a refugee status that they are entitled to, but also so that the taxpayer is not paying to accommodate them for longer than necessary.

The Chair: This is absolutely moving into an issue that I know Baroness Hughes wants to explore. Baroness Hughes, do you want to just pick it up from where Tim had got?

Q147 Baroness Hughes of Stretford: Thank you, Chair; I will. Tim and Jon, I actually wanted to focus on the report that you have just outlined on asylum and the NAO noting the wasted public funds and inefficiencies and all the rest of it in the system that you have just talked about. You have mentioned a couple of the issues already that I was going to touch on, like the shortage of judges and how that creates backlogs further down the line in the system. Did you make any estimate at all for these problems in the lack of a whole-system approach—for instance, moving the initial decision-making backlog, just shunting that into appeals so that you get cases having to be looked at multiple times successively and ricocheting around the system? Did you make any estimate of how much that contributes to the cost of the asylum system? Presumably it must add to the cost because, instead of somebody flowing through and just being looked at at the right points, they are having to wait and they are having to be looked at again and again and again.

Tim Phillips: We have not put a figure in pounds on that, but what is implicit in your question is absolutely correct. All of this—what I have just been talking about, what you have just been talking about—is covered by the concept of failure demand. You are failing to make the decision right the first time, you are failing to make it quickly enough, and therefore you are increasing demand somewhere else. One thing that is relevant in this is that the Home Office, particularly one to two years ago—it has somewhat improved in the time we were looking at it—had a real problem with a high turnover of caseworkers. It was 35%, I think, in 2024. It then also takes seven months for a caseworker to be fully operational and able to handle whatever is thrown at them. You can see how, even if you are not telling people to rush, just the fact that you have that turnover and you have so many green recruits could drive a culture in which you get suboptimal outcomes the first time so it gets looked at again.

Baroness Hughes of Stretford: That links into just another point to touch on—you raise it in the report—about, if you like, the balance between the speed of processing and the quality of decisions, which also links into your point about the greenness of new recruits and all the rest of it. Do you think that there is enough focus, or has been enough focus, in the Home Office on the quality of decisions, given that that is obviously important for the applicant, but it is also important in the system, resulting in a lot more appeals and so on?

Tim Phillips: I will say first—I do not know if Jon wants to come in—that I do not think you can say that there is enough focus on quality when the observable outcome of the full process is that so many cases are appealed and, at appeal, so many decisions are changed.

Jon Simmons: That is a key issue, the fact that virtually everyone who is refused asylum appeals. It is significant that the courts seem to disagree in probably half the cases with the decisions of the Home Office. That does not necessarily mean that the courts are right, but with the information that is available to them they come to a different conclusion. It is a complex process judging these claims. It is quite an undervalued process. I am not surprised that there is a high turnover of staff. It is a very difficult job to do. I respect the people who are trying to do that job.

Coming back on some of the earlier comments, there are two things that we are trying to do. Part of the problem with the appeal system, for example—not just asylum appeals but a general problem across Government—comes back to the point about needing person-centric data. If we were to design systems on the basis of these sorts of questions and having person-centric data across Government, then as statisticians we would be much better placed to develop proper statistics. Too often the idea is to develop systems very quickly because all they want to do is measure casework performance—how many appeals we processed. That is it. We know that there is a need for richer information. We could pause and do IT system development much better if we did that.

We would then have much better data available for the modelling of the system and the various impacts around the system, as the data would be better integrated. A lot of modelling goes on. There is a big team in the Home Office, which does a lot of modelling of different impacts of policy changes on the immigration system, but it is constantly crying out for better-quality data. The appeals system is just one angle to that.

Q148 Lord Bach: I will home in on the legal advice and lack of representation issue that your report mentioned. That seems to be one of the really important issues here. Put aside justice. To get the system working better, asylum seekers getting legal advice—which could often be “You have got no chance”—and representation if there is a case, which will shorten appeals, seems to be pretty vital under our system and should be pressed. I am sure it has been by the NAO. Ministers should then act on it. I am not asking you whether Ministers should act on it. There have been recent amendments to Bills going through this House which have asked for that. Would you agree that this is quite an important part of this whole argument? Lack of judges is very important too.

Tim Phillips: My understanding is that the two have similar causes. It is not mostly the lack of entitlement to legal aid but that marketplace—

Lord Bach: The marketplace is nonsense.

Tim Phillips: You mentioned casework being undervalued. This is difficult territory. It is hard work and people find it emotionally taxing so think twice before deciding to become an immigration appeals judge or work in the immigration legal aid market. Those are functions that the system requires to exist. The system needs to ask itself how to attract what we need to get the whole system flowing as we want it to.

Q149 Lord Dubs: I have a supplementary on appeals. If a significantly large proportion of the cases are reversed on appeal, does that not suggest that either the original decision was not right or that there is something wrong with the basis on which appeals decisions are made? If so, should we not be doing something about it, or are we just accepting that this is how it is going to be?

Tim Phillips: The assumption is correct. Just to put some data out there, between 2019 and 2022, over that three-year period, 43% of initial decisions went to appeal. The original decision was overturned in 42% of cases—almost half. The confounding factors which are worth mentioning are that people’s circumstances can change and that a certain number of different decisions may be affected by that. Also, some individuals have access to certain protected statuses that are not the ones that they applied for but mean that they cannot be returned or asked to leave because they have some kind of humanitarian protection. Basically though, you are correct that there needs to be a look at what the different people, the caseworkers on the Home Office side and the people in the tribunal system on the justice side are looking for and why it is different.

Lord Dubs: It seems to me that it gives a person every incentive to appeal. If the success rate of appeals is high, why would anybody not appeal, no matter how much advice they are given in the way that you suggest? Is this being looked at or are we just accepting it?

Jon Simmons: Some of this is a data issue. Again, it comes back to how we do not have good data on the appeal system. We have looked at these sorts of things in the past and there is a reasonable view that the Home Office might prioritise the legal representation in appeal cases for certain things but not necessarily want to follow through on others. If the courts take a view that, for example, they are sympathetic to any family with small children and may overturn a Home Office decision on them, the Home Office might want to look at that at an earlier stage and maybe not follow through to appeal.

On the other hand, there are categories where people probably do use the appeal system even though there is very little chance of any reasonable person, in the Home Office or the appeal system, supporting those cases and may well reject them. The peak in asylum arrivals in small boats that happened a couple of years ago might be a good case in point. Very few of those will have been granted refugee status. Of those that were refused, very few would then be successful at appeal.

Why do we allow that process to go on? It feels like a waste of resource. You might say that every case needs to be dealt with individually, that people need the freedom to make their case in any way that they can, but that is the sort of debate that we need to be having—on how we prioritise those cases.

Lord Dubs: Is this being looked at?

Tim Phillips: One thing that is worth finding out more about is that the Home Office was aware that there was more that it could learn from the tribunal end of the process. Nearly two years ago, in April 2024, it created its litigation strategy board. I cannot give you any more information about what that board has discovered or done, but if they are learning lessons, you would expect them to be learning them through that forum.

The Chair: We will pursue this, thank you.

Q150 Lord Dubs: Looking at the overall position, 10.7 of the UN sustainable development goals states that “Nations should facilitate orderly, safe, regular and responsible migration and mobility of people, including through the implementation of planned and well-managed migration policies”. Is the Home Office living up to those ideals?

Jon Simmons: I will start. I am a big fan of the UN SDGs. They are a great innovation providing a structured approach for the world on what good governance looks like, because they cover a massive range of things. They are not about migration but about every aspect of society. They apply to every country and every country is supposed to report on all these measures. The UK does for those that are relevant.

In the immigration sphere, it is just 10.7 that relates. A few others may be vaguely relevant. I cannot remember how many SDGs there are, but probably hundreds. However, 10.7 is about migration and it comes in four parts. There are four specific measures set out in the UN documentation. The first one is about recruitment costs, saying that if you are going to employ people from overseas, you should not charge the employee; it should not be a cost on the employee. The UK is pretty good on that score because that is something that we do not do. Sponsorship would probably be relinquished by an employer that sought to place some kind of fee or charge on the people whom it is trying to recruit.

Then 10.7.2 is the “proportion of countries with migration policies that facilitate orderly, safe, regular and responsible migration and mobility of people”. We have policies, we have loads of them and we have lots of strategies. I do not think that the UK is in any way short of a strategy. This kind of measure is really directed at countries that have not thought about their immigration policies for years or maybe do not have any strategy, and encouraging them to think about migration in a sensible way and come up with policies.

Then 10.7.3 is “Number of people who died or disappeared in the process of migration towards an international destination”. That is a very important thing to monitor. We want to ensure that there is a strategy for addressing the risk to people engaged in migration. Obviously, if you are comparing one country to the next, it very much depends on the circumstances. We have unfortunately seen a not insignificant number of people who have died in crossing the English Channel over the recent years. This is clearly a tragedy and something the Government would be keen to address, but obviously those numbers pale in comparison to the numbers who died crossing the Mediterranean—the central Mediterranean from Libya to Italy, for example. It is a very different measure. It is something that we do monitor. The IOM is very good at publishing information on fatalities. But the situation is very much dependent on the migration routes that are involved and on whole areas of migration globally. The people crossing Africa, for example; we just do not know how many people die crossing the Sahara. We do not know how many people are killed by smugglers. We do not know how many people die because of the treacherous conditions they go through in order to leave one country and try and get to another.

The fourth element is proportion of the population who are refugees by country of origin. Obviously, we publish lots of statistics on refugees and where they come from in the UK, so the data is available. I am not sure that there is a right proportion of the population, so I do not think there is a target, if you like, for that metric. But in terms of information that is available, we know a lot about our refugee population. Sorry, I do not know if that answers the question.

Q151 The Chair: Thank you very much. We are well over time, and I am grateful to you for being here. Just before we come to an end, I am going to ask, as it were, the two bodies represented here: if you were in our position and you had to come up with one fundamental recommendation that this committee should make to the Government in light of all the conversations we have had, what would it be? Tim?

Tim Phillips: From the National Audit Office’s perspective and recent work, where policies are being proposed to be implemented, there should be visibility—at least internally, but preferably publicly—of what it is believed the impact of that will be throughout the system from end to end for people working in the Home Office on cases, through to the Ministry of Justice as we have talked about in tribunals, and also through to local authorities, who will have people living in their communities and working there and with responsibilities and duties towards them. That should be understood in a deliberate, bordered and distinct way, and therefore it could not be said to have come as a surprise. Obviously, this is modelled things and is not a prediction of the future; it is just an attempt to understand. From our point of view, that whole-system view and giving that a status and a priority, hopefully publicly, would help for a better-planned and better-managed system in the future.

Richard Baynham: I just want to add, in line with that, that we should invest more in understanding how your system is working and how your route is operating. Is your route operating as you intended it to operate? Are there unintended consequences from the way it is operating? Are there increased risks in compliance and response? What we have not seen yet is enough evaluation and investment in understanding how visa routes are operating. Linked to Tim’s comments, I would encourage greater investment in that understanding.

Jon Simmons: I agree with all of that and on the idea of having clear and coherent impact assessments. I was taken by the comments about a person-centric system. As a statistician, I would love the committee to encourage Government to focus much more on a person-centric approach to collecting data and not see things in stovepipes. The other point that I would add, just to give you a different angle, is the one that we were talking about earlier, about how the statistics are presented to the public to encourage a proper public conversation and public discourse on whether we are doing things in the right way that helps with that. There is a balance between having, as I said, annual data on more strategic issues that provides a more coherent and broader picture, and then the detail that you might want more frequently on things that are of particular public concern or public interest. That might be another area you want to focus on as well.

The Chair: Thank you very much, all three of you, both for that and for all your answers during this particular session. We have already suggested things that we would like you to write to us about. Of course, if there are any others, we look forward to receiving them. But on behalf of the entire committee, I thank you all three very much indeed; it has been a fascinating time. Order, order! The public evidence session is now concluded.