Examination of Witnesses (Questions 160
- 179)
TUESDAY 23 OCTOBER 2007
Professor Christian Dustmann and Professor Ian Preston
Q160 Lord Layard:
It does require, as you said just now, that capital is supplied
infinitely elastically, i.e. that the rate of profit is not affected
by immigration, whereas it does not seem altogether unreasonable
to suppose that lying behind the pressure for immigration is the
notion that you can get more profit out of it. Is there anything
that can be said about what the impact of immigration has been
on the rate of profit?
Professor Dustmann: We have not looked at that,
we looked at the impact on wages and I am not aware of any research
which has looked at that. Possibly the two novelties in our particular
study would be to look at wage effects along the distribution,
which so far has not been done because we looked at average wages,
and to give up the assumption that capital is perfectly inelastic,
which was the assumption in most previous theoretical models.
We suggested the possibility that capital may be in elastic supply.
If capital is perfectly inelastic, then any surplus would go to
the owners of capital and you may want to call that profit.
Q161 Lord Layard:
Can I just make sure we all understand it. If there is a constant
return to scale then if the wages have gone up—let us leave
it!
Professor Dustmann: Maybe one thing I should
just add is that what we find for the UK may not be the same for
other countries, Migration affects labour markets in very different
ways in different countries. Why is that the case? Because the
skills structure and capital structure of the receiving economies
are different and the skill mix of immigrants is different. Even
if the skill mix was the same where we would find immigrants in
the wage distribution may differ across countries. So we have
shown that in the UK, while on average they are higher educated
than native workers, we do find immigrants, at least initially,
very much at the low end of the wage distribution, and these things
together show the effects that immigration will have on wages
at different parts of the distribution and will have on capital.
Q162 Lord Turner of Ecchinswell:
I am sorry to go on about these theoretical points but it is very
important that we understand what the theory tells us. I looked
at your document for the Low Pay Commission and indeed the bit
about averages is on pages 28 to 30; I just want to clarify that
I have a correct understanding of what you are saying there and
what you are not saying. Am I right, first of all, that you are
saying that under certain circumstances the average wages of pre-existing
native non-immigrants can rise, but of course that is a different
thing from whether average wages rise because there could be a
compositional effect which pulls it down.
Professor Dustmann: Yes.
Q163 Lord Turner of Ecchinswell:
I just wanted to clarify that. Secondly, I think you were rather
suggesting in what you said a minute ago that there was a bit
of a mystery about a positive effect on the average wages of natives
but what I think you are saying the theory suggests is that as
long as capital is in elastic supply and as long as the skill
mix of immigrants is different, then we would expect a positive
effect.
Professor Dustmann: That is absolutely right.
Q164 Lord Turner of Ecchinswell:
The third point though is a crucial one because what you then
say at the end of it is that the very same conditions which are
required for there to be a positive effect on average native wages
also require that some categories of native workers lose.
Professor Preston: That is correct as well.
Q165 Lord Turner of Ecchinswell:
So the very factors which tell us that on average native workers
might gain would automatically tend to mean that there must be
some category of native workers who lose.
Professor Preston: Yes, that is all true within
an equilibrium.
Lord Turner of Ecchinswell: It is all
within an equilibrium but it is useful to get those points clear.
Thank you.
Chairman: Let us move on. Lord Vallance.
Q166 Lord Vallance of Tummel:
can we come off the theory just for a bit, and what we heard from
some previous witnesses was that there were real practical data
methodological problems in the analysis of immigration issues
and the labour impact. What is your assessment of those problems
and how robust do you feel are your most recent results on the
labour market?
Professor Preston: There are data problems and
there is not much we can do about that if we accept the data.
We can try and make it more robust to non-systematic measurement
errors in our measures of immigration but if there are systematic
errors there then that is going to cause almost insuperable problems.
There are other methodological problems which we recognise completely,
particularly about identifying the direction of causation. The
estimates in the study we have been talking about are all based
on comparing the economic outcomes in different regions that have
different intensities of immigrant inflow. Of course, if a region
is temporarily doing economically well you would expect wages
to be rising there and immigrants to be attracted in, so even
if immigration had no effect on wages you would expect to see
a positive association between wages and immigration. We try as
far as we can to overcome that by not relying on association between
actual immigration and wage changes but association with that
part of immigration that we can predict from past levels of immigration
which you would not expect to be affected by temporary changes
in economic circumstances. That should make the estimates robust
to that sort of problem but there are assumptions behind that
that you cannot test; you have just got to take them on faith
and if they do not work then that is a problem. But we make as
strong an effort as we can to try and make them robust to those
sorts of worries, which are the main things that we have been
concerned about.
Professor Dustmann: We had meetings with the
Low Pay Commission and several government department representatives
before that report was published, and of course these concerns
came up and we tried hard to ease these concerns. We not only
look at data from the labour force survey when we compute wages
but we also use data from NES/ASHE, which is the data set which
is used to look at wage inflation in the UK, real wage increases
and nominal wage increases. We discuss very carefully the issues
Ian has been talking about, but the problem with any empirical
work in economics is, of course, that the data is never as good
as we would want it to be. But I am at least pretty confident
that the analysis we have done here is very robust and I would
not expect future work to show dramatically different results
for the same wave of immigration that we have analysed here. That
is very important because future immigration may have different
effects on the British Labour Market.
Q167 Lord Turner of Ecchinswell:
Can I ask one question which is a direct follow-up to this methodological
point? To the extent that you are trying to infer it from differences
between regions, if capital were fully mobile between regions
of the UK and other categories of labour were fully mobile between
categories of the UK, you would not then expect to see the regional
pattern of results. Does not the regional pattern of results only
arise from the extent to which other factors of production are
not fully mobile?
Professor Preston: Mobility of labour at all
is going to diffuse the effects out across different regions and
we are aware of that issue as well. You can control for that to
a degree because you observe native labour of different types
and you can put that in as controls in your estimation and we
do try and do that.
Professor Dustmann: We do also extend things
and try to publish them, so we have done a lot of additional work
among it trying to get a grip on the degree of mobility of native
workers as a reaction to immigration, and there is very little
evidence for that.
Q168 Lord Turner of Ecchinswell:
Can I ask another point on your empirical results on these questions?
You have commented on pay. Is it your belief that immigration
has had any significant effects on employment or the participation
of British native workers, particularly on those which start off
with low participation rates? For instance, if you have long-established
ethnic minorities that have very low participation rates, do new
waves of immigration affect their employment or are all the impacts
that you have found as it were wage-related—price effects
rather than quantity effects?
Professor Preston: In that particular study
we did not look at employment but we have done a previous study
for the Home Office which was looking explicitly at employment
and participation. It was using earlier years, it was the mid-Eighties
up to 2000, I think, whereas the later study used the late Nineties
and the early part of this decade, and that work has been published.
We do find small, negative effects on employment and participation,
but these are not very well-determined. Reading them at face value,
there are some negative effects on employment and on participation,
but as I say the statistical imprecision on them is so great that
you could not rule out zero effects. We did compare different
groups as well—you tend to find participation going up amongst
the skilled and down amongst the medium-skilled and low-skilled
which fits with the pattern of wage effects. We did look separately
at recent immigrants and minorities and we could not find evidence
that they were worse affected, the actual estimates were that
they were less affected but they were so imprecisely determined
that I would not want to make a strong claim one way or the other.
Q169 Lord Kingsdown:
How have the children of immigrants performed educationally, do
you know, and how have they fared in the British labour market?
Professor Dustmann: I have brought you a paper
we have written on that. Let me start with the problem we have
in the UK which is that there is no data set except for the BHPS—which
is too small to analyse immigrants—which links an individual
to their parent. So the way to address the children of immigrants
has to rely on something else. What we do based on the labour
force survey is we use the different waves of immigrants who have
come to this country which, for the ethnic minority community,
is a pretty recent event. We have linked children who are born
in the UK but who are ethnic minorities to their parents, so we
can identify in the labour force survey second or maybe higher
generation immigrants but only from ethnic minority groups, not
from the white group. To give you the magnitude of what that means,
at the moment about half of the foreign-born in this country are
white in terms of ethnicity and half are non-whites, so the non-whites
would report their ethnicity and we can identify them in the labour
force survey. If you look at their educational achievements, across
all groups educational achievements are at least as high as those
of the white, native-born individuals who, with a very high probability,
are descendants of white individuals, but there is a large variation
on top of it. For instance, 85% of Chinese individuals who are
born in this country go to college, a much smaller percentage
of Pakistani and Indian individuals go to college, but even those
percentages are higher than those of comparable native whites.
So the educational achievement of ethnic minority second or maybe
third generation individuals is remarkably high in this country.
If we look then at the wages and if you look at the employment,
it is also more favourable. However, that is due to the fact that
40% of ethnic minorities do live in London and London has higher
wages and also a more buoyant labour market. So if we control
for the fact that those individuals of second generation origin
live in London, then their wages are slightly lower, in particular
for females, and their employment falls slightly. We have done
that very carefully in this report and if you find that interesting
we will leave this here.
Chairman: That would be very helpful.
Lord Sheldon.
Q170 Lord Sheldon:
Migrants have taken certain jobs where they are over-qualified.
Do you expect them to move on to jobs that better match their
skills in due course?
Professor Preston: We can already see this in
the evidence that we have in our report of where they are in the
wage distribution. If you look at very recent migrants they are
quite heavily concentrated down the lower end and if you look
at less recent migrants it is the same sort of pattern; but it
is less pronounced so you can see evidence of them moving up the
wage distribution, which is compatible with the movement towards
more appropriate matching up with their qualifications or skills
and you would expect that to ameliorate the downward pressure
that we found on wages at the bottom end if that continues. I
would expect things to go in that direction.
Q171 Lord Sheldon:
What sort of timescale do you expect for that?
Professor Dustmann: There is a problem again
of measurement and that problem is that many immigrants—and
we come to that a little bit later when we go down your catalogue
of questions—return back to their home countries. So return
is not random if only the good guys or only the bad guys in terms
of educational achievement go back, then we cannot identify a
moving up in the skills distribution due to a particular selective
return migration from changes in the application of their skills
to the British labour market. So that is a little bit problematic.
We could give you precise numbers on that if we did not have this
particular problem.
Q172 Lord Layard:
How does the national minimum wage affect the demand for migrant
labour? Do you think that minimum wages have other implications
we should take into account when we look at the economic impact
of immigration and, in particular, do you think that the current
UK policies ensure employer compliance with minimum wage regulations
applying to migrant workers?
Professor Preston: It is not something I really
have a view on because I do not think I have any evidence on any
of those questions, so I am not sure what I can say.
Professor Dustmann: Given that basically immigration
leads to a downward pressure in that part of the distribution
where actually we do find the minimum wage which is on the sixth
or seventh percentile of the wage distribution, one may deduce
from that that the minimum wage does protect some workers at that
particular part of the distribution. What effect it has on higher
grade immigrants versus natives as Ian says, I do not have evidence.
I can answer it in terms of what I think about it myself, which
is I do think that European labour markets in general, particularly
in continental Europe, have been characterised by a strong rigidity
at the lower end of the wage distribution and employers try to
undercut that by employing immigrants who come illegally. So I
do think there is a lot of that going on in the construction sector
or the agricultural sector in countries like Germany, but that
again is not something on which we have done research.
Q173 Lord Layard:
The implication would be that immigration at the lower end in
the presence of the minimum wage would have increased the unemployment
of the low waged domestics. When you did your cross-regional research
were you able to look at that issue?
Professor Dustmann: We did not look. The Low
Pay Commission asked us to look at wages, not employment, and
we have not looked at employment. We are happy to do another study
looking at it.
Chairman: Let us move on. Lord Vallance.
Q174 Lord Vallance of Tummel:
Does immigration lessen the incentives for employers to train
the indigenous workforce or indeed to invest capital in improving
productivity?
Professor Dustmann: There are two types of training
we distinguish in labour economics. The one sort of training is
general training, that is training you can take with you and every
firm finds it valuable and will pay you the same premium on that.
The other type of training is firm-specific, which is only of
value in a particular firm, so general training in competitive
labour markets of the type we usually study will never be borne
or never be paid by the employer because the employee can take
it with himself or herself. So it will always be the employee
who will pay for general training. For specific training the employer
is willing to pay but that specific training is not worth anything
outside that particular firm, so this is the kind of rudimentary
model of analysing firm-based training and we do not see that
migration would in any way change that. If you train a worker
in general training, you would not be willing to pay for that
training whether that worker is a migrant or a non-migrant; if
you train him in specific training, the specific training is worth
nothing to the worker outside your firm, so you are the beneficiary
of that training and you would finance that training.
Q175 Lord Vallance of Tummel:
As far as capital investment is concerned, if there is availability
of labour through immigration, does that mean there will be less
incentive to invest in productivity improvements through capital
investment?
Professor Dustmann: We have done work so far
for Germany and there is work for the US looking at alternative
adjustment mechanisms to immigration. The adjustment mechanism
we are analysing in this particular work (the LPC Report) is wages,
and that is what people have usually in mind when they think about
the way immigration affects the local labour market. However,
there are other adjustment mechanisms. One is by changing the
industry structure and the other is through technology, so the
most interesting example or maybe the most illuminating example
is the wine industry in Australia and in California, which is
highly labour-intensive in California and highly mechanised in
Australia, the reason being that it is very easy to get unskilled
workers in California but not in Australia. This is anecdotal,
but if you look at the evidence there is evidence for the US that,
particularly in the traded sector, technological adjustment is
more important than adjustment in the mix of what companies produce.
We have done the same analysis for Germany; we are still working
on that but the preliminary evidence points in exactly the same
direction with surprisingly similar magnitudes of estimates. So
there is evidence that technology adjusts to the availability
of labour in particular parts of the skill distribution.
Lord Vallance of Tummel: Thank you, that
is interesting.
Q176 Lord Turner of Ecchinswell:
The Government has often stressed the positive impact of immigration
on total GDP. Do you have a point of view on what its impact is
on GDP per capita, which is presumably a balance between the positive
effects on the average wages of native workers minus the compositional
effect simply due to the influx of people on low wages? Have you
ever put that together and worked out whether this is positive
for GDP per capita or negative?
Professor Preston: I have never done that calculation,
but it is a mixture of what we said earlier about the effects
on native wages. The evidence on recent immigrants is that they
tend to be located lower down and you would think that average
native wages possibly go up. The typical recent migrant tends
to be at the lower end, so those two things work against each
other.
Q177 Lord Turner of Ecchinswell:
Presumably, if that overall was a negative effect because of the
compositional effect and if society was committed through, for
instance, the working and family tax credit to redistribution,
it could be the case that the native workers would lose either
all or a significant proportion of their beneficial impact from
higher pre-tax wages. Has anybody ever done that analysis to see
whether they are a net winner after the redistributional effects
to which we are committed?
Professor Preston: I have no more information
on that.
Q178 Lord Kingsdown:
In 2003 you were both among the authors of a report for the Home
Office forecasting that annual net immigration from the eight
countries which joined the EU in 2004 (the so-called A8 countries)
would be between 5,000 and 13,000 a year. What explains the difference
between your estimates then, if I may ask, and the much larger
number of A8 migrants who have come to the UK since May 2004?
Professor Dustmann: First of all, we should
point out that that number is a net number and it was calculated
over a ten year period, so net is the difference between in migration
and out migration. We expected that return migration of these
immigrants would be, over a ten year period, quite large. Secondly,
that number was based and at least half of that report was a critical
assessment of these numbers. We were very, very careful to put
these numbers in the context and in the possibilities we have
actually for estimating these things. So while the study we have
done for the Low Pay Commission may have some data problems, it
is standing on much more robust grounds than any study which was
published at that time on EU Accession, and ours was not the only
one to predict the inflow of immigrants from the enlargement countries,
the reason being that we can only make a prediction if we have
some past information about such migration. In this particular
case, because of the Cold War and the Iron Curtain, there was
no information about how individuals from those countries would
react to the wage differential and to other economic differentials
between the UK in particular, Europe in general and those particular
countries on the other side. What we needed to do, therefore,
was take past migration from other countries to come up with a
prediction, and that of course leads always to estimates which
may get it very wrong. That is the context. We should then point
out that the numbers we predict are net numbers. So if, for instance,
every year about 200,000 individuals would come to the UK, the
stock of immigrants would be increased so that the net average
migration over the period would be 200,000 per year and after
ten years there would be two million immigrants added to the stock.
If, however, they only stay for two years, then after ten years
the average net inflow would be 40,000, so that is a very different
number. We have to take account therefore of return migration
when we look at these numbers and that has not been done in the
entire public debate on EU enlargement unfortunately. The other
issues which we pointed out very carefully in the report are that
our report was based on the assumption that countries like the
large European economies, in particular Germany and Italy, would
open their borders towards the eastern immigration countries as
well. That is something that they announced but then they retreated
on that particular announcement, so we looked at those immigrants
who would basically go to the UK as an alternative to Germany
and we did not consider immigrants who went to the UK because
they could not go to Germany or some of the other larger continental
countries.
Q179 Chairman:
Do you reckon that accounted for quite a lot?
Professor Dustmann: I am absolutely sure that
if Germany had opened its labour market to the accession countries
we would have seen lower inflows to the UK.
| 
