Examination of Witnesses (Questions 1-19)|
23 JANUARY 2008
Q1 Chairman: I welcome Dr John Goldthorpe,
Professor Stephen Machin and Dr Jo Blanden to our proceedings.
It is a pleasure to have you here. You will know why you are here.
We committed ourselves to having a look at social mobility, and
this is by way of a taster to see whether we need to take further
evidence. It seemed sensible, given that the work conducted by
the London School of Economics, particularly for the Sutton Trust,
has opened up an interesting debate about investment in education
and whether it is responsible for social mobility. It raised some
interesting questions that all of us involved in education would
like answered. I declare an interest. I am a governor of the London
School of Economics, which probably means that I shall ask harder
questions rather than easier ones. John Goldthorpe, I have read
much of your work over the years, and I am somewhat of an admirer
of your worklet us get that out in the open. Again, I shall
not ask you any kinder questions. Do any of you want to say anything
briefly to open our discussion, perhaps about your work and where
we are with social mobility and how it affects education? We always
give people a chance to say something if they want to.
Dr Goldthorpe: I would be ready
to join you in discussion.
Q2 Chairman: You want to go straight
into the cut and thrust?
Dr Goldthorpe: Yes.
Professor Machin: We could give
a little overview of the research we have done, if that would
Chairman: Yes, if it is not long.
Professor Machin: No, it will
not be long.
Chairman: Let us get started then.
Professor Machin: We have been
involved in three strands of research in this area. One is the
often-cited work in which we compared what had happened to social
mobility across two birth cohorts. We have extremely rich birth
cohort data in Britain, where people are followed from when they
are born through their lives at various stages. One of the birth
cohorts was just under 20,000 people born in 1958; the other was
about the same size but of people born in 1970. That is where
we have the first finding. We related people's earnings to their
parents' across these cohorts, and this relationship strengthened
over time. That is when we inferred that social mobility had been
falling in Britain, because people's earnings were more strongly
associated with their parents' income. That was for the more recent
cohort rather than the first cohort. That was a comparison between
the 1958 and 1970 cohorts. The second strand of work was to look
at more recent changes. We were not able to undertake the same
exercise of relating people's labour market earnings to their
parents' income, because the people involved were yet not old
enough to be in the labour market. But we can look at the relationship
between a number of intervening factors, like education attainment
and early-age test scores, and how they are related to family
income. We found much more constancy in our recent work. The third
area of research was done by Jo and other colleagues. They have
been comparing what happens if you look at income mobilitylooking
at how strongly income is correlated across generations as compared
to social class mobility. John has done a lot of work on that
in the past. There are two papers out thereone by Robert
Erikson and John Goldthorpe, the other by Jo Blanden, Paul Gregg
and Lindsey Macmillanthat compare what has happened through
changes over time, when you use income or when you use social
class. That is the kind of research, with three strands, that
we have been involved with.
Q3 Chairman: Can I ask you, Stephen
Machin, why did you do this research?
Professor Machin: Why did we start
doing this research? One of my long-term interests is in labour
market inequality and how it has moved over time. A lot of that
looks within generations, to say how wide the income distribution
or the earnings distribution is for the population in a given
year. One key feature that underpins that is people's family background
and whether it maintains any qualities or generates any qualities
across generations as well. My overall interest is in terms of
labour market inequality and how the income distribution is evolving
Q4 Chairman: But you have been commissioned
to do some of this research, have you not?
Professor Machin: For the first
strand, we were funded partly by the Sutton Trust. We started
that work well before then, because we thought it was of significant
academic interest. Of course, it fed into the policy process subsequently,
but we were interested in it from a purely academic point of view
to start with.
Dr Blanden: It followed on from
previous work that Stephen had done looking at the level of intergenerational
income mobility in the UK.
Professor Machin: The key innovation
we made, from the academic perspective, was to start looking at
changes over time. There is a lot of work out there that measures
how strongly correlated people's earnings or income is with their
parents' earnings or income or how their social class is correlated
with their parents' social class at a point in time. The main
innovation in the newer work that we have done was to start looking
at trends or changes over time.
Q5 Chairman: How did you and the
Sutton Trust get together?
Professor Machin: There are many
sources of funding for academic research. In relation to the Sutton
Trust, I cannot quite remember the details, but in the same way
as we get funded by the Joseph Rowntree Foundation and the Economic
and Social Research Council, it was an opportunity to get funded.
Dr Blanden: I think you met Peter
at some point, didn't you?
Professor Machin: Yes, I met Peter
Lampl somewhere, but that is back in history, I am afraid.
Q6 Chairman: The reason why I am
probing is this. The Sutton Trust is a very respected education
think tank and I know that it commissions quite a lot of research.
Indeed, Sir Peter Lampl is a good friend, but he does have a particular
view, which I imagine he had before he came to you for the research,
that there is a lack of social mobility. Did he hear that you
were doing research that would back up his ideas?
Professor Machin: Yes. We must
have been at a meeting somewhere where we spoke about common interests.
I probably told him that we were very interested not just in looking
at a snapshot or point in time, which is how this work currently
tends to be done, but in trying to say something about how social
mobility has changed over time. Of course, you can see why that
would interest Peter.
Q7 Chairman: Yes. If you had got
your money from Coca-Cola or any other commercial organisation,
I would probably probe you on this even harder. What was the research
brief that Sir Peter Lampl or the Sutton Trust gave you?
Professor Machin: We wrote the
research brief. We said we are interested in doing research on
changes over time in intergenerational mobility.
Dr Blanden: And we did have some
Professor Machin: We had some
preliminary results; that is right. We made the case that this
would beand it has subsequently turned out to beoriginal
research. I think that is partly why it has received so much attentionbecause
people just had not done much on changes over time before.
Q8 Chairman: Why do you think that
is? It is a rather important issue, isn't it?
Dr Blanden: It was about data.
Professor Machin: It was about
at least two things, one of which was data availability. In terms
of a consistent comparison over time, the data have not been available
until relatively recently to enable us to do that, in a very representative
way. The other point is that the natural thing to do in the earlier
work was to look at a point in time and just say: how much or
how little social mobility is there?
Q9 Chairman: Right. So, you completed
the research. It has obviously become quite controversial, but
certainly it has been used, by politicians of all kinds and by
groups. When you are reflecting on the impact of your research,
what do you think about it?
Professor Machin: It has been
used by different people in different ways. It has sometimes been
used well; it has sometimes been misquoted badly. Sometimes people
have pushed it much further than you might want to push it. If
we are talking just about the first strand of work, it is a comparison
just across two cohorts, born in 1958 and born in 1970a
12 year periodso it is rather specific in that sense. Some
people have made a lot more of it than perhaps we would have made
of itcertainly than we have made of it.
Dr Blanden: Can I follow that
up? For example, there has been the idea that social mobilitywe
should really call it income mobility, particularly when John
is here. "Income mobility is falling" is often a headline
you will see or a quote from a commentator or politician. Based
on these two snapshots, we would never say that. That is one of
the reasons why we have gone on to look at what has happened more
recentlyso that we can find out whether it is true, because
certainly our earlier research did not actually tell us that.
Professor Machin: I think we would
say that in that particular period there was a fall in social
mobility, intergenerational mobility, income mobility. If you
look at our paper, that is what it says. It does not say anything
about what is going on today. It does not say what was going on
before that either, although there are other pieces of research
out there where you can start to pull together the pieces of a
jigsaw, if you think about the long time period as well.
Dr Goldthorpe: The research that
my colleagues and I have been doing is concerned with social mobility
in terms of social class rather than income. We are interested
in the relationship between children's social class and their
parents' social class. For that purpose, we treat class on the
basis of the National Statistics socio-economic classification,
which has been widely used in official statistics since 2001.
That is one difference from the work that Stephen and Jo have
been doing. The second difference is that we try to work, as far
as possible, using representative samples of the entire active
population. We also use the birth cohort studies that Stephen
has described, but only for want of better data. The position
is that between 1972 and 1992, the General Household Survey (GHS)
collected data that could be used for social mobility analysis
in terms of class. We have analysed that run of data quite extensively.
Unfortunately, for reasons that have never been clear to me GHS
ceased to collect the relevant data in 1993. So, we have the rather
ironic situation that at a time when social mobility has become
much more important as a political issue than previously, we do
not have good dataat least, not as good as that from between
1972 and 1992. In order to get information, we have resorted to
the birth cohort studies. As Stephen has explained, there were
two birth cohorts only 12 years apart. So far, we can compare
their members only up to their early 30s, which is not entirely
satisfactory. I am currently working on extending the nationally
based analyses through to 2005, when GHS was required to include
an EU module on standards of living. In that module, there is
information that is relevant to social mobility. Unfortunately,
it is not strictly comparable with the data that we had for the
period between 1972 and 1992, but we are working hard to make
it as comparable as possible. We have had some preliminary results
from that work. Taking the whole set of analyses together, the
main findings are as follows. First, in the period between 1972
and the present, we have found no change in the total mobility
rate. That is to say that we find no change in the proportion
of children in different class positions from their parents. Neither
have we found any weakening during that period, in the net association,
or net stickiness, between parents' class position and children's
class position. That might seem to be a more optimistic conclusion
than the one that Stephen outlined on declining income mobility,
but in another sense, it is more pessimistic, because we find
that, throughout, class mobility seems to have been at a lower
level than income mobility. We can use the birth cohort studies
for that. We think that looking at mobility in terms of social
class captures more of the intergenerational continuity in economic
circumstances. Finally, we have found one change. Although the
total mobility rate is unchanged, the composition of the rate
is changing. In the middle decades of the last centuryfrom
about 1940 to the 1980swe saw steadily rising rates of
upward mobility and steadily falling rates of downward mobility.
From about the 1990s, those trends tended to level out. Now, especially
with men, we find that if anything, rates of upward mobility have
flattened out and may even be declining a little, while rates
of downward mobility are no longer decreasing and may even be
increasing a little. For women, the situation is not quite so
bad. For us, that is the important change. If you wish, I can
try to explain why I think that change came about.
Chairman: I had better not ask you that,
because my colleague is going to drill down on that in a moment.
Dr Goldthorpe: I am sorry?
Q10 Chairman: I had better leave
that to some of my colleagues. We will drill down on that in a
moment. I have one more question for the moment. We have seen
great change in the social class composition of our country have
we not? Most of us believe that the number of people describing
themselves as being in the middle classes and living a middle-class
lifestyle has grown.
Dr Goldthorpe: Yes.
Q11 Chairman: How does that impact
on the very small numbers of people who would now be classified
as unskilled workers? I think that the Leitch report said 3.2
million, and that the number is going to down to 600,000 by 2020.
Something fundamental is happening in our class structure while
you are researching it. Is that all taken into account?
Dr Goldthorpe: Oh yes. You have
put your finger exactly on the explanation for the change that
I referred to. During the middle decades of the 20th century,
there was very steady growth in the proportion of the active population
in professional and managerial employment, and a contraction in
the proportion in wage-earning, mainly manual employment. You
could say that there was simply growing room at the top. That
drove the steady increase in rates of upward social mobility over
that period. From the 1990s, that growth in the proportionate
size of the professional and managerial salariat, as one might
call it, has slowed down. It is still going on, but it has clearly
slowed down, especially insofar as higher-level managerial and
professional positions are concerned. That is really behind the
tailing off of the increase in upward mobility. That is especially
marked in the case of men, because they are facing greater competition
from women for professional and managerial positions. The change
in the shape of the class structure is absolutely crucial to the
changing patterns of upward and downward mobility. One point I
would add is that over that whole period of major changes in the
shape of the class structure and corresponding differences in
patterns of upward and downward mobility, what is constant is
the inherent what I call stickiness of the relationship between
parental class position and children's class position. Once you
net out, as we can, statistically, all the effects of the structural
change, that inherent, underlying stickiness shows a remarkable
Q12 Chairman: What do you say, Stephen,
to people such as Stephen Gorard who really criticises your methodology
and conclusions? From my reading of the literature, I take it
that he is your fiercest critic?
Dr Goldthorpe: I do not think
that Stephen has criticised my
Q13 Chairman: No, sorry, Dr Goldthorpe,
I have moved on. Professor Gorard has fundamentally criticised
the work of Stephen Machin and Jo Blanden. I was asking them.
He is your most consistent critic is he not, Jo?
Dr Blanden: Yes. I can say something
Q14 Chairman: Has he got it right
or is he wrong?
Dr Blanden: Well, we obviously
think he is wrong.
Professor Machin: With very good
Dr Blanden: With some quite good
reason. He took a summary paper that we wrote in 2005 for the
Trust, which was a simplification of our work. Perhaps we did
not present things in the first version as clearly as we should,
because we were talking about two different things: changes over
time, and differences across countries. We did not lay it out
clearly enough, in essence, and we responded to his comments and
tried to be clearer, but the thing is that the Sutton Trust report
in 2005 was a summary of a wealth of papers that we have written,
and we also summarised other people's work to give that picture
to the Trust. However, Professor Gorard seems to refuse to look
at any of our other papers, or any other work that is behind this,
and he sticks to what was said in the 2005 paper. He also criticises
us for using a small sample of the cohort studies, because we
have to look at only a situation involving income and earnings,
and he is absolutely right that that is a concern. It is one that
I looked into during my PhD. When you restrict the sample that
you look at, you always have an element of doubt about whether
you are misrepresenting the overall picture. One good thing that
has come about from our work, looking at the issues in tandem
with John Goldthorpe and Robert Erikson, is that they use a much
larger sample, and when they restrict their larger sample of social
class to that which we used for income, the patterns do not change,
so it seems unlikely that our results are driven firmly by the
restriction in sample. So, although that is a reasonable point,
we do not think that it applies in this case.
Q15 Chairman: But the data are patchy,
are they not? You are not comparing like with like on data sets
and the different countries that you take. Some of the countries
are a bit strange, are they not? You say in one passage of your
work, "major industrial countries", but many of them
are Nordic countries and Canada. In this Committee, I am always
happy and safer when comparing Great Britain as a 60 million population
country with Germany, France, Italy and Spain. But that is not
who you compare with, is it? You have the US, but many of them
are quite small countries.
Dr Blanden: I have written a chapter
for a book that some colleagues at the LSE are putting together
on education and income inequality. In it, I try to look more
widely at where the UK fits on intergenerational income mobility,
and to link it with the role of education. I do not know whether
Stephen would necessarily agree with me, but if we look only at
the countries from which we can get really good data, we are possibly
picking countries where the UK and the US are often at the low
end of mobility. For the UK and the US, that is probably broadly
true, but there are probably many other countries down there,
too. I would argue that France, Italy and Spain are, but that
Germany is a little bit unclear.
Q16 Chairman: Because most of your
data are from West Germany, are they not?
Dr Blanden: Yes, which does not
help. If you try to look around for as much data as you can, even
if they are not completely comparable, I would say that several
large European countries are equally as immobile as the UK. But
then there is other evidence from other sources, such as the PISA
study, which show that educational opportunities in the UK are
very strongly related to family background, so there is evidence
on either side. We have some holes in our evidence; that is completely
the case. In that particular Sutton Trust study, we picked only
the countries where we had data that were quite similar, so we
compared with the UK and the US, Nordic countries, which are probably
fairly extreme cases.
Professor Machin: I would add
that Stephen Gorard is out on a limb in what he says, if you compare
what we say with other leading experts in the area. Gary Solon
is the leading US economist on intergenerational income mobility.
He has written a survey piece bringing together the international
evidence and he reaches very much the same conclusions as we do,
by drawing on studies from different countries. However, the US
and the UK are towards the bottom of the international league
table for intergenerational earnings or income mobility. If you
take Stephen Gorard's criticism literally, because he thinks that
he has this methodological point, which is not correct, about
the way in which we presented the evidence, he is actually saying
that he thinks that the extent of intergenerational mobility is
rather like that in Norway, and I do not think that anybody would
think that that was the case.
Q17 Chairman: Is the heart of the
problem that he is attacking your truncated paper, or perhaps
even the gloss that the Sutton Trust put on your research, because
in a sense it did not do the research, but was making a case for
Professor Machin: To be fair on
him, I think his observation is that the findings have been misused
in certain quarters and he, therefore, is commenting on it because
it is a piece of research that has received so much attention
that he thinks that he should be saying something about what his
position on it would be.
Chairman: Okay. That has been some very
Mr Carswell: I have three questions,
if I may?
Chairman: On something that has come
up already? We are going through the questions in sections.
Mr Carswell: Can I ask three questions
at some stage?
Chairman: If you do not repeat the questions
that people have already bid to ask, yes.
Mr Carswell: Sorry. You have been asking
questions, do you want me to ask questions now?
Chairman: Could you hold back for a moment?
Let me give the people who asked to do the opening session a chance
Q18 Mr Slaughter: That was a very
interesting introduction. When we were talking in private before
you came in, I sensed a slight frustration among some of my colleagues
that some of us, like me, are statistically challenged, but probably
what we are more interested in are the more political questions
about whether social mobility is a good thing, and the role of
education within that. In order to get to that stage, it is always
useful to have a coherent and consistent statistical base. Although
there appears to be some level of agreement between the various
studies, that may be an unfair question. How would you reach a
consensus between your views, or what points would you pick out
from the studies that we have had, which would give us that clear
base from which to go forward to the next stage? In other words,
can you summarise what consistent findings you believe there are
between the various studies that you have done? Or indeed, more
controversially, do you think we should disregard points that
have been made?
Dr Goldthorpe: The first thing
to say is that Stephen and Jo are looking, as we have established,
at mobility in terms of income. In effect, they are comparing
the earnings of children with the incomes of the families from
which they came. My colleagues and I are looking at mobility in
terms of social class. Those are different phenomena; there is
no reason in principle why they should give exactly the same results,
although of course it is interesting if results differ to try
to ask why that is so. We have had a long series of exchanges,
trying to work out why this is so and these occasions get rather
technicalwe could go into that. My own reading is that
intergenerational income mobility is more subject to short-term
social change than intergenerational class mobility. In the case
of income mobility you can get relatively short-term changes of
some magnitude, and that is related to the fact that changes in
the structure of incomes, in income equality, can occur rather
more quickly than changes in the shape of the class structure.
However, as far as the political and policy implications of our
research go, the differences are not that great.
Professor Machin: I would tend
to agree with much of what John has said. Both the social class
findings and the income findings suggest that social mobility
is a problem in Britain. It should therefore, have a high priority
in policy debates and the fact that debates about the research
are giving it a higher profile is both good and important. In
response to the question about why I was interested in intergenerational
mobility, I stated at the start that my interest is in labour
market inequalities, income disparities across the population,
and trying to ascertain whether they are too big or too small.
We know that in Britain over the last 25 to 30 years, the gap
between the highest and the lowest paid has widened massively,
with a particularly big increase during the 1980s. The findings
on social mobility suggest that this opening up of income disparities
is not just restricted to people's own generation but can be tracked
back to where they were born and the incomes of their parents
at that point in time. That suggests that this work is even more
important, as growing inequality at a point in time is reinforced
across generations. When we have looked at education, we have
looked at it as a transmission mechanism underpinning the extent
of intergenerational social mobility. As we saw this episode of
falling income mobility, we discovered that one of the key factors
underpinning it was an increased sensitivity of education to family
income. Basically, the expansion of education, particularly higher
education (HE), that took place in the late 1980s and early 1990s,
disproportionately benefited people from richer families. That
is one of the reasons why social mobility has fallen. In terms
of a policy discussion and the political debate, these issues
are becoming more important than they were in the past.
Q19 Mr Slaughter: Yes. From your
comments so far it appears that there is a greater degree of consensus
that I had first thought. That is good. However, the findings
are rather concerning in the way that you have just indicated.
In relation to the educational side of your last point, obviously
one of the things that we will be concerned with is the chicken
and egg aspect of education. In your view, is access to education
based on social class or income a barrier, perhaps to increased
social mobility, or is the way that the education system is used
widening that gap? If I understood your point correctly, are you
saying that what could be perceived as a widening of educational
opportunity, or something that in a simplistic way may increase
opportunities within income and wealth bands, may actually have
had the opposite effect?
Professor Machin: Yes, it can
go either way. One easy way to think about it is that if people
who benefit from the expansion of education systems come from
above average income families, it reinforces the inequalities
that are already there. If they come from below average income
families, it will narrow the inequality that was there. It appears
that the expansion of HE disproportionately benefited people from
above average income families, and that has been one of the mechanisms
as to why intergenerational income mobility has fallen during