Examination of Witnesses (Questions 64
- 79)
TUESDAY 16 OCTOBER 2007
Ms Karen Dunnell, Professor Paul Wiles CB and Mr
David Frazer
Q64 Chairman:
Good afternoon. Thank you for coming to help us with our inquiry.
We have a number of questions, as you know, but I wondered if
each of you might like to introduce yourself so that it is in
the record.
Ms Dunnell: Thank you. I am Karen Dunnell, the
National Statistician and Head of the Government's Statistical
Service.
Mr Frazer: David Frazer, Head of Statistics
for the Department for Work and Pensions.
Professor Wiles: I am Paul Wiles, Chief Scientific
Adviser at the Home Office but I have ultimate responsibly for
statistics at the Home Office.
Q65 Chairman:
Thank you very much. Whoever answers is up to you, whether you
all want to do it or as is necessary. Please do not feel you have
to say the same thing as your colleague has said, but if you think
of something to add please do. Last year the ONS described immigration
data as "inadequate". Could you give us some examples
of inadequate data? Would you also comment on the scale and significance
of the problem and how this situation came about.
Ms Dunnell: First of all, the situation has
come about because the scale of travelling across the world and
the movement of populations across the world has been increasing
over the years. This particularly became a problem when we got
rather larger than we expected migration from the new countries
joining the European Union. It was at that point, as the National
Statistician, that I decided I really needed to join colleagues
in the GSS and have a task force to look at the issue and come
up with an agreed way of going forward. The main difficulties
we have are the following. One of the most difficult things to
measure in this country, and, indeed, in many countries, is not
the number of people arriving in the country but the number of
people leaving the country. That is quite a challenge for the
survey that we have at all airports and seaports. We have already,
as a result of the task force, put in place a new sampling scheme
that captures more people leaving the country. The other problem
is that we can get a pretty good national figure of who comes
in and who goes out, but of course we really need to know where
they go when they arrive in the country. Every year we have to
use these data to produce population estimates for every single
local authority, and that is another area where we did not feel
it was good enough. Particularly since we have had the European
Union and a very open labour market and many, many cheap flights
between Europe and the United Kingdom, there has been a lot more
of what we call "short-term migration", particularly
for work reasons, into the United Kingdom and of course our population
estimates and our definition of migration has traditionally been
that of the United Nations of "coming intending to stay for
12 months". We needed to start doing more work on that, which
of course is quite complex, working out who wants which definition.
Those are the main areas.
Q66 Chairman:
That is a clear answer as to what the problems are. Are you able
to give us any indication of the scale of differences that have
arisen and how much you think it has changed from what you thought
it might have been?
Ms Dunnell: At this point it is quite difficult
to do that. When we have the 2011 census we will get a benchmark
and that will be a test. One of the reasons for doing the work
now, of course, is that we do not want to get a census which is
out of kilter with the estimates that we produce every year. The
main problems are around how many people have moved from the port
of entry to which part of the country, and we are assembling quite
a lot of information about that now, and also this issue of how
long people stay when they arrive. That is where the big issues
are.
Professor Wiles: I am sure your Lordships know
this but just to make it clear: embarkation controls were abolished
between 1994 and 1998. The effect of that is that from 1993 onwards
we really do not have any data about people leaving the country.
As Karen has already said, with 90 million, roughly, passenger
journeys per year in and out of the country, the consequence of
that is that we really no longer have adequate data on net migration.
That is obviously quite important because that means we do not
have as much of a detailed understanding about the nature of the
population of the United Kingdom as we really ideally want to
have. Remember that migration is the biggest single driver of
population change. Those removals of embarkation controls really
had some significant impact. The problem we have now, given we
now have those 90 million journeys, is that if we were to go back
to the kind of embarkation checks we had in the early 1990s we
would simply bring Heathrow and many of the other ports to a halt;
it would be hugely expensive because of the people who would be
involved; but it would also be too slow. We would gum up the whole
system. The result of that is that we can only solve this problem,
particularly of measuring people coming in and coming out, by
a more technological solution. Perhaps later on we can talk about
that, because of course that is what we are trying to move to
in terms of our E-Borders Programme, so that we can reintroduce
embarkation counting but without having to employ huge numbers
of people and effectively bringing everything to a halt.
Q67 Lord Griffiths of Fforestfach:
The Statistics Commission has said publicly that there has been
something of a reluctance on the part of departments to share
information and to develop information together. I wonder whether
you could tell us where we stand on that issue today, what you
feel the obstacles are to greater coordination, and whether the
information held by local authorities might be helpful.
Ms Dunnell: The first thing to say is that colleagues
across the Government Statistical Service are working very closely
together at the moment. My task force, which reported in December
last year, had several government departments actively involved
in it and we are taking forward the planning work together. That
part of it is going very well. Of course between us—and
we can all talk about it—we have already started the process
of improving things. As I mentioned we have already increased
the sample on the port survey. We have also carried out a review
of the 21st century requirement for a port survey—because
of course this is quite an old survey—so that we can bring
the whole idea up to date and we will be reporting on that before
Christmas, and we have already put in place our population estimates,
using our large-scale household surveys to apportion more correctly
the total number of migrants across local authorities. David,
do you want to start to say what you have been doing?
Mr Frazer: Yes. DWP has information on people
who come to this country who have the right to work. They have
to register for national insurance purposes so that they can build
up contributions for future benefit entitlements. The first thing
we are doing is making as much of that information as we can available
to the ONS at lower level detail. We can drill down those numbers
into some very low geographical levels, local authority districts,
and we can split them by country of origin, age and sex, so that
we have a better flow of information coming here. That is about
people coming here who intend to work; those numbers alone do
not tell you about those who go on to do so. This is where we
can begin to do more linking of administrative records, clearly
only where we have a clear legal purpose to do so. We are talking
about being able to link records from people who come in and get
a national insurance number with other information on the contribution
system and on employment from the tax system, so that we can start
to get some idea about how many come here and work for a shorter
period and who works for longer periods and maybe who works for
a spell and then does not work. That work is underway now. There
are lots of issues about quality of linking and understanding
information but we have made a start, and by the end of the year
will be in a better position to say what we can do going forward.
Professor Wiles: Perhaps I might add something
to that. First of all, in terms of sharing our data: yes, we do
share our data with ONS. In terms of working together on the future:
yes, not only Home Office statisticians but also representatives
of the National Statistician sit on the E-Borders Programme, so
we are working together on that. I think however there is another
issue behind it. I have no doubt that we could get a better sense
of, as it were, the migration career and what that looks like
longitudinally, if we were able to analyse any data that government
holds, but I do not need to tell you that there are serious concerns
about whether we should be allowed to do that. This is not the
only area in which I could make a similar statement of how, if
we were able to share more government data, we could get a better
understanding on something or other, but the same problems are
there. As you know the Council for Science & Technology called
for a debate on just this issue and advised the Prime Minister
there ought to be a debate because there is a very difficult balance
here between how far should we be enabled to bring data together
in order to understand problems as opposed to how much right should
the individuals have to have privacy from us doing so. That is
a difficult balance. In the end it is for Parliament to decide
but also the public need to be part of the debate that leads into
that decision. There was a Cabinet Committee, MISC 22, I think
it was, that was looking at just this issue. That Committee of
course disappeared with the change of Prime Minister. But I think
the issue is still there of the need to generate that debate and
for Parliament to decide how far we should be able to share the
data we possess.
Q68 Lord Griffiths of Fforestfach:
Do you think there is data available in local authorities which
could be better used?
Ms Dunnell: Yes. In fact we have done some investigation
into a number of local authorities. We are pulling all that together
and publishing that towards the end of the year. One of the difficulties
with local authority information is that some local authorities
may develop methods that are very, very good, but, in order to
use them at a national level, we need every local authority to
comply with the same methods. But we are doing some work on looking
at that and looking in particular at whether they have information,
for example, about children and schools and so on that would be
useful. We are doing quite a lot of work on that.
Q69 Lord Lawson of Blaby:
I gather there was an interdepartmental task force last year which
made various recommendations about improving migration statistics.
Ms Dunnell: Yes.
Q70 Lord Lawson of Blaby:
Have those recommendations been implemented or is there a problem
of adequate funding for it? Do not forget those questions, but
perhaps I may add a few supplementary questions from your interesting
answers. You mentioned that a lot of information you get is from
a sampling scheme. Can you tell us how large the samples are,
particularly in relation to the total inflows, and whether you
think those samples are of an adequate size? Secondly, you mentioned
that there are other ways of doing more technological solutions
and so on. Are there any other countries in the western world,
any other countries which are magnets for migration/immigration,
which you think might do a better job than we do in getting the
necessary information for policymaking? If so, tell us. Finally,
on this interesting question which your Home Office colleague
has referred to about the balance between getting the necessary
information and individual privacy, do you think at the present
time the balance is struck in the right place?
Ms Dunnell: The first thing, as I said earlier,
is that we have begun work across departments on some of the recommendations
of the task force. We have a list of other things we need to do
and these will cost money that ONS has not received in its settlement.
We had our financial settlement in March, so we have already been
able to work out how we are going to spend our money over the
next few years and we have identified several Millions of pounds
to spend on these improvements. I have just last week, following
the settlement for all other departments, written to my permanent
secretary colleagues to ask them if they can top up, between them,
between £8 and £9 million to help us do this work properly.
Q71 Lord Lawson of Blaby:
What did they say?
Ms Dunnell: I have asked for the replies by
the end of October. Then we will come back to it and see what
the results are. I am quite hopeful that they will. In some ways
it is quite a small amount of money. That is the answer to that.
But there are quite a lot of things that we have already said
we can do in our budget and, of course, the other departments
have programmes of work which are already committed, so it is
a top up that we are looking for. On the sampling, the sampling
on our port survey I have already said we did not think it was
adequate and that is one of the reasons for doing a review, but
in the meantime we have upped the samples for people leaving the
country and that will improve accuracy quite a lot. The other
survey that we use extensively is the Labour Force Survey. Again
the task force identified that that has some deficiencies; for
example it does not interview in communal establishments. Workers
from the A8 may be living, for example, in hostels or caravan
parks and things like that and one of the things we need to do
is to improve the survey by going and asking Labour Force Survey
questions at those establishments. That is one of the things we
need some extra funds for. The other thing is changing the definition
of who is eligible for the survey. That we are planning to do
from next year. At the moment if you have not been in the country
for six months you do not need to comply and we are changing that.
Those will all bring about improvements. That is the answer on
sampling. On your other question about how good we are, I am sure
Paul can say something, but it is interesting that OECD have done
a report comparing OECD countries, particularly the European countries,
and come up with the conclusion that our migration estimates were
on the European average. Before I come back to privacy, perhaps
you could say something, Paul, about the experience of other countries
on this.
Professor Wiles: I think your question was:
Is there a country that does it better? I think the answer is
yes, and it is Australia. That is probably the pre-eminent example
of a country that does it much better, but there are rather special
reasons as to why Australia manages to do it so much better: it
is long way from anywhere else; it is quite difficult to sneak
in as an illegal—as you know the Australians have been extremely
robust with anybody who has tried to sneak in as an illegal—and
the relative flows in numerical terms are small compared to those
of the UK. For those reasons they have been able to put in place
probably what I think is the best system in the world among the
developed countries for doing that kind of counting, and they
have, for that reason, more reliable data. I think that is the
answer to the question is anywhere better. In terms of funding,
if I may, the E-Borders Programme at the moment is out for tender
to supply the systems that will run underneath it. That is a funded
programme. That will go ahead, because, of course, it is not just
there to produce statistics, it is critical to the Border and
Immigration Agency's operations. It is necessary for their operations
to work, so it will be funded and go-ahead for those reasons.
We have also been looking to see if we can make short-term improvements
to landing cards. Those are the cards that non-EEA citizens fill
in when they enter this country. At the moment they collect limited
information and we only analyse them on a sample basis, and we
are looking at whether we can actually collect slightly more information
on those as we go ahead. As the National Statistician has just
said, she has indeed written to permanent secretaries, and I have
a note in my inbox from my permanent secretary saying: What do
you advise? I will no doubt advise him that generosity is a virtue,
but we will see! On the balance question—and Karen will
no doubt want to come back to this—I think there are clearly
areas where the difficulty we have in analysing across datasets
is not only in the national interest but even the individual interest.
I think the most obvious examples of that are in the medical area—and
I know this is a different topic but it really is all part of
the same problem—where, often, if data is being collected
for one reason it cannot be used for another purpose, even if
that would be manifestly to the benefit to those who supplied
the data. I think we need to do some sensible rethinking in these
areas and I think we also need to make sure, wherever the balance
is struck at the end of the day, that we also have a clear framework
to ensure that balance is kept and that balance is policed properly,
so that people can be reassured that we are indeed sticking to
whatever balance Parliament decides.
Ms Dunnell: I agree with that. One of the difficulties
is making sure that the people who provide the data understand
that we will only be using data for statistical purposes. That
is the thing we have to get across and at the moment that is quite
hard to do because of some of the legal requirements of individual
datasets. Things are looking more hopeful now. As you know, we
now have a new Statistics and Registration Act, which will come
into being in April, and that will enable us to seek permission
from Parliament to obtain better access for statistical purposes
to some of these datasets. That will start the process. Also,
within the Government Statistical Service we have done a lot of
work over recent years in setting up anonymised datasets, in setting
up safe settings for researchers and colleagues in other government
departments to come and look at individual data. We have created
an environment within the Statistical Service where we can demonstrate
that this is a safe thing to do from the public's point of view.
But it is the perception, at the end of the day, which is very
important, so we need to take these steps very carefully.
Q72 Lord Lamont of Lerwick:
When you were giving your reasons why estimates of immigration
might be a bit wide of the mark, you did not mention illegal immigration
law, which I would have thought was one of the obvious problems.
I notice that Professor Wiles, when he was talking about why had
Australia a better system for measuring immigration, referred
to perhaps easier problems with illegal immigration. That is my
first question. You have outlined in your submission a package
of reforms and my second question would be how much more accurate
do you think data will become as a result of that? How accurate
do you think it can become or are we always going to have to live
with a degree of uncertainty surrounding these statistics?
Ms Dunnell: I will answer that question and
then I will pass to Paul to answer the question about illegals.
Yes, the package of improvements I think will go a long way to
helping get the situation much better. As I explained, one of
the key problems is understanding, first of all, where people
go to in the country. The things we are proposing to do—the
better linkage of data between departments, not necessarily linkage
but the sharing of data between departments and the improvement
of information that we get from surveys by including these communal
establishments, changing the residence qualification and so on—will
all help in dividing up the totals. Because, at the end of the
day, we do not think the totals are too far off the mark and some
of the emigration information, when we improve that, will improve
the accuracy of the inflow and outflow. The other area where it
will be very difficult to keep up with things is understanding
changing patterns, particularly different patterns of people coming
here for education and work. That is where we need to do much,
much more research and use longitudinal information to understand
those patterns. At the moment, that is quite difficult. We are
publishing our first experimental estimates of short-term migrants
later this month, but we cannot do that at a local level at the
moment. That is one thing that we are going to have to do more
work on to improve. Paul, would you like to deal with illegals?
Professor Wiles: Perhaps I could continue with
that part of the answer first and then I will go on to illegals.
Perhaps I ought to say a little bit more about the E-Borders Programme
because it is, in a sense, one of the keys to some of these questions.
Essentially, the E-Borders Programme will require carriers to
provide details of all passenger journeys to and from the United
Kingdom. They will have to supply that information to the Border
and Immigration Agency. That data will then be stored as a travel
history for each individual passenger, so we will have travel
histories for each person who has entered or left the country.
That will further be supported by checks and data capture when
individuals arrive or leave the country on a risk basis; in other
words, we will not check everybody but we will use intelligence
sources and other things to carry out checks, and of course that
will involve biometric checks. In so far as people have biometric
identity documents, whether they are ID cards or passports, we
will also check those biometrics. That programme will give full
coverage of everybody entering and leaving the UK by 2014. However,
since that is an ongoing roll-out from next year onwards, we will
have a 95% coverage by 2010, so you can see that this ramps up
quite quickly and it is that last small percentage at the end
that really takes the time, inevitably, to get hold of. Once we
have that in place—in 2014 or, as I say, we will have pretty
good estimates before then, in 2010—for the first time that
will mean we have an accurate count of both immigration and migration
and hence of the net migrant number, which, as I was saying in
answer to your first question, is the thing which has been missing.
That E-Borders Programme is under way now. It is out to tender.
Perhaps I could turn to your question on illegal migrants. The
first thing to say is that countries that do have illegal migration
problems—and you are correct to say that Australia is not
really one of those for rather special reasons—I do not
think any of those countries have managed to produce what they
would regard as absolutely reliable estimates of the size of the
illegal population. A number of years ago we looked at all the
different methodologies that had been used by other people, to
evaluate them and see which we could use in the United Kingdom.
Apart from methods that, in a funny way, are not methods, like
the Delphi method, where you simply go around and ask people you
think are experts—which is hardly, I think, a method—the
method that some countries have used is the capture/recapture
methodology, which essentially is a methodology that was developed
for counting things like bird populations. You count the number
of people who are captured in one particular dataset who are illegals
and then you compare that with those captured in another dataset.
It is quite a good method. The problem we have in the UK is many
things like arrest data from the police and so on do not record
if somebody is an illegal immigrant. For that reason, we decided
we could not use the capture/recapture method. That left us with
what is called the residual method, which essentially takes the
population census data as a base point and then tries to work
out where there are residual numbers in that. We did that based
on the last census and we published that in a paper called Sizing
the Unauthorised Illegal Migrant Population in the United Kingdom
in 2001. The central estimate that came out of that was 430,000,
with a range of 310,000 to 570,000—so it is quite a large
range.
Q73 Lord Lamont of Lerwick:
You would always be years behind.
Professor Wiles: We can only do it when we have
a census, that is the problem, because it is the residual method
from the census. Until we have another census, we cannot really
do it. Interestingly, that is about 0.7% of the population. If
you look at the estimates produced by other countries, that is
about in the middle of the range that other countries have, so
it is probably not too far out. At the moment, that is the only
method we have available. I would not pretend it is terribly accurate,
but it is the best we can do on the data available and it does
give some kind of estimation of the size of the illegal population.
Q74 Lord Griffiths of Fforestfach:
In reading the submission, this figure of 939,000 between mid-2001
and mid-2006 comes up. What is the confidence around a figure
like that? Is it 90% or 80% or 99%? It is a very specific figure,
939,000, after everything we have been talking about in relation
to inadequacy.
Ms Dunnell: Basically, we have to come up with
a figure every year because we have to produce population estimates
for every health and local authority, so we have to use every
piece of information we have to make that estimate. That is the
best that we can do. As I have said, we are doing an awful lot
at the moment to improve our methods. I do not really want to
put a percentage on how confident I am about it, but it is absolutely
the best we can do with the data that we have available at the
moment. We have been very open about where we feel we can improve
the data, so that is what we are going to do. Some people will
dispute that it is too high and others will say it is too low,
and there we go, but it is actually quite difficult. Given all
the different things we are doing and the ways we are comparing,
across the piece, the information we have in government, that
is our best estimate.
Q75 Lord Paul:
I can imagine the problem of not being able to estimate properly
the illegal immigrants, but the variations come in which newspaper
you read. Is there no way of working out what is the percentage
of illegal immigrants from the total number of migrants, or a
rough estimate, for five years?
Ms Dunnell: First of all, they are illegal and
they generally know they are illegal, so they avoid contact, generally
speaking, barring accidents, with anybody who might be interested
in their legal status. That, in a sense, causes a problem for
all our administrative systems. Also, because they are a very
small minority of the population—and the Health Service
and the National Insurance Service and so on are not going around
asking absolutely everybody, "Are you an illegal immigrant?"—it
is quite difficult to get a handle on it.
Q76 Lord Paul:
I understand the difficulty, but is it 5% of people who are illegal
immigrants or 1%?
Professor Wiles: I think it is a very interesting
question. The thing you have to bear in mind is that probably
the proportion of illegals to legal migrants will be different
for different migrant groups, because the relative ease of being
an illegal obviously varies depending where the migrant group
is coming from. I think we will probably find a different percentage.
I think it would also vary over time depending on what I think
of as the "push factors", what drives immigration. If
you have something like the Bosnian conflict, then you have very
powerful pressures driving desperate people out of countries,
who are going to be much more desperate to get into a country
legally or illegally. That is why I think there is a significant
variation on these illegal flows. You are right, however, that
you would expect that variation also to show up in the legal flows
and, therefore, yes, we could use that to mirror to some extent.
We have not done that. As I explained, we simply went for a methodology
that we thought would give us some idea of the total of illegal
migration, but, you are right, that would be another way of trying
to get a finer grained look at the illegal population.
Chairman: I hate to say this because
the answers are fascinating and very interesting and I am most
grateful to you, but we are getting a little behind the clock.
If all of us, including on this side of the table, could move
on as fast as we can we will finish somewhere near the proper
time.
Q77 Lord Layard:
I have a question on illegals. This is a stock of people who have
come in illegally, as I understand.
Professor Wiles: Yes.
Q78 Lord Layard:
Do we know how recently those people came in?
Professor Wiles: No.
Q79 Lord Layard:
Do we know by what method they came in? Did they overstay a tourist
visa or did they creep in illicitly? Do we know anything about
them?
Professor Wiles: We do know something but not
what the numbers are, if I may put it that way. We do have some
idea, both from intelligence reports and from fairly obvious things,
about the different routes by which people become illegals. Some
of course become illegals because they hide in the back of a lorry,
or whatever, which is the thing that most people think of, but,
you are quite right, probably the biggest single component of
the illegal population are overstayers and many of those are overstayers
from countries that would probably be surprising, like Americans
or Canadians who come here as students or to work or whatever,
change their status, but do not formally apply for visas to do
so. There is that mixture, undoubtedly.
|
