SOCIAL MOBILITY

A DISCUSSION PAPER

Performance and Innovation Unit

April 2001

Contact for this Discussion Paper :

Stephen Aldridge
Chief Economist
Performance and Innovation Unit
Room 4.11a
Admiralty Arch
The Mall
London SW1A 2WH

Tel : 0207 276 1470

Fax : 0207 276 1407

E-mail : stephen.aldridge@cabinet-office.x.gsi.gov.uk
SOCIAL MOBILITY

A Paper by Stephen Aldridge, Performance and Innovation Unit

Summary

What is social mobility?

1. Social mobility describes the movement or opportunities for

movement between different social groups, and the advantages and
disadvantages that go with this in terms of income, security of
employment, opportunities for advancement etc.

Why is social mobility important?

2. Social mobility matters because :

− equality of opportunity is an aspiration across the political

spectrum. Lack of social mobility implies inequality of
opportunity;

− economic efficiency depends on making the best use of the

talents of everyone; and

− social cohesion and inclusion may be more likely to be achieved

where people believe they can improve the quality of life they
and their children enjoy through their abilities, talents and
efforts.

What are the facts about social mobility?

3. Analysis of intragenerational (within own lifetime) mobility shows

that there is a great deal of income and earnings mobility (upward and
downward) in Britain but much of this is short range. For example,
individuals on low incomes or earnings may move one or two deciles up
the income or earnings distribution but not in general much more.

4. There is evidence that :

− earnings mobility in Britain has declined over the past 20-25

years;

− poverty is more persistent in Britain than in, say, Germany; and

 − upward career mobility from manual occupations to higher
status professional and technical occupations has declined, with
entry to the latter higher status occupations taking place
(increasingly) direct from the education system rather than
through mid-career flows from lower status occupations.

5. Sociologists make a distinction between absolute and relative

social mobility in analysing intergenerational mobility. Absolute social
mobility is concerned with the absolute number or proportion of people in
a social group who are upwardly or downward mobile. Relative social
mobility is concerned with the chances people from different
backgrounds have of attaining different social positions.

6. Analysis of intergenerational mobility (children compared with

parents) shows :

− there is and continues to be a considerable amount of absolute

social mobility. Over the long run, increasing numbers of
children have enjoyed upward social mobility compared with
their parents because economic and social change has increased
employment opportunities in the professional classes. There
has, in other words, been “more room at the top”;

− the growth in the rate of absolute social mobility, especially

upward social mobility, appears to have been halted in recent
decades;

− the longer term expansion in the number of people in higher

social classes has not only permitted upward social mobility
from the lower social classes but allowed an increasing
proportion of children whose parents were in the higher social
classes to remain there;

− as a result relative social mobility (the relative chances of

people from different social backgrounds making it to a given
social class) has been fairly stable; and

− though there is evidence that some groups within the ethnic

minority population are catching up with the majority white
population through rapid upward social mobility, ethnic
minorities are still disadvantaged compared with the majority
population.
7. International comparisons are difficult because the defining
features of social classes may vary from country to country. Nonetheless,
there appears to have been considerable absolute social mobility in all
industrialised countries, though absolute rates of intergenerational social
mobility vary widely between countries and over time. Cross-national
variations in absolute rates are explained by specific and historical
factors, such as economic growth, migration and wars, rather than by
sociological factors.

8. Relative rates of social mobility exhibit a remarkably high degree

of similarity between countries and over time. Within the modest
differences that do exist, it is possible to identify a cluster of more fluid
societies (such as Israel, Australia, Japan, Poland, Sweden and Norway)
and a cluster of less fluid societies (such as Hungary, the United States,
Britain, Italy, Germany, France, the Netherlands and Ireland). There is
some evidence that higher rates of relative social mobility are associated
with lower income inequalities.

What are the barriers to more social mobility?

9. The determinants of and barriers to relative social mobility are not

well understood. Potential explanatory factors include:

− educational attainment;

− childhood poverty and associated psychological and behavioural

development;

− the family and the strategies families adopt to support their

children, including access to financial, social and cultural
capital i.e. not just money but values, behaviours and networks
of contacts that affect access to opportunities;

− attitudes, expectations and aspirations, including aversion to

risk; and

− economic and other barriers which some groups use to “hoard

opportunities” e.g. anti-competitive practices which limit access
to professions such as the law and may discriminate against
particular groups.
10. The role of ability (whether inherited genetically or through
socialisation) as a determinant of social mobility is hotly contested, with
different sociological studies coming to widely differing conclusions
about the importance of the role it plays.

What are the implications for policymakers?

11. Policymakers are likely to have a range of objectives, including :

− maintaining and promoting high levels of prosperity;

− promoting a fair and just society, of which there may be several

dimensions :
− equality of opportunity : enabling individuals to
improve themselves through their own individual
ability and effort rather than class background,
gender or ethnicity (a meritocracy);
− social justice, cohesion and inclusion : providing
safety nets and opportunities for the disadvantaged
which reduce inequalities of outcomes and help
build communities.

− promoting individual freedom : either in the positive sense of

enabling individuals to make their own choices (e.g. by
ensuring the rule of law) or in the negative sense of minimising
coercion by the State or some combination of the two.

12. These different objectives may be complementary or in conflict

with each other. For example :

− policies to promote greater fairness might weaken incentives to

work, save, invest and innovate, undermining the goal of
maintaining and promoting high levels of prosperity. On the
other hand, a perception that economic and social outcomes are
fair and just may build trust and facilitate the more efficient
working of markets;

− increases in income inequality may be associated with increases

in inequalities of opportunity and thus lower rates of relative
social mobility. On the other hand, a highly meritocratic society
characterised by high rates of social mobility might have
significant income inequalities and social tensions for the
reasons described below.
13. By a meritocracy is meant a society in which the most able and
committed people can succeed in attaining the most desirable, responsible
and well-rewarded positions. Unless one believed ability and
commitment were (either in whole or in part) determined by one’s social
class origins, the features of such a society would include high rates of
social mobility and the absence of any association between class origins
and destinations i.e. equality of opportunity.

14. Prima facie, a meritocracy described in these terms has many

attractions. But it would also have downsides :

− there would be downward as well as upward mobility i.e. there

would be losers as well as winners. This could create economic
instability and social tensions;

− the losers would have no one to blame for their circumstances

but their own lack of ability and commitment. This could create
a lot of unhappiness and resentment;

− there could be significant income and other inequalities. The

upwardly mobile could “take all”; and

− it could be economically inefficient. The market does not

reward ability and commitment but those who provide the
goods and services consumers want to buy. The more able and
committed may be more likely to succeed in a market economy
but the lazy who have a good idea may be highly rewarded too.
Overriding these market signals could have economic costs.

15. Ultimately, the weight to be given to the different possible

objectives of policy, and any trade-offs that may need to be made
between them, require a political judgement.

16. So far as social mobility and equality of opportunity are concerned,

there is a choice to be made between :

− measures which promote a weak form of meritocracy i.e. which

focus on removing barriers to upward mobility and assisting or
encouraging upward social mobility. This approach would give
primacy to absolute social mobility; and
 − measures which promote a strong form of meritocracy i.e.
which not only seek to facilitate upward social mobility but
which seek positively to reduce barriers to downward social
mobility for dull middle class children. This approach would
give primacy to relative social mobility.

17. Promoting higher rates of social mobility in order to allow

individuals to succeed on the basis of their individual ability and effort is
likely to be challenging. Absolute social mobility has been driven by
structural changes in the economy and society over which governments
may have limited influence. Relative social mobility has been stable over
time and has shown little variation between industrialised countries
notwithstanding widely differing political systems, policies and
institutions.

18. The literature suggests a range of specific policy options for

promoting social mobility. They can be grouped under four headings :

− removing barriers to upward mobility;

− positively assisting upward mobility;

− removing barriers to downward mobility; and

− positively assisting downward mobility.

19. The possible measures include :

(a) removing barriers to upward mobility

− raising educational standards;

− ensuring access to further and higher education for all;

− changing attitudes, expectations and aspirations, in particular

encouraging bright children from deprived backgrounds to take
full advantage of the educational and other opportunities open
to them;

− improving social mixing and the formation of networks;

− reducing/removing barriers to geographical and occupational

mobility;
 − greater disclosure of employers’ recruitment practices – gender,
race composition of workforce etc;

− removing non-competitive barriers to entry to professions such

as the law;

(b) positively assisting upward mobility

− tackling poverty and social exclusion, particularly childhood

poverty;

− increased investment in education;

− improving access to financial capital;

− building better routes for upward mobility within professions

and workplaces;

− giving people second chances e.g. through access to education

in later adult life;

(c) removing barriers to downward mobility

− reducing the weight given to geographical catchment area as a

determinant of access to the best State schools;

− higher rates of taxation of investment income and wealth;

− higher rates of inheritance tax; and

(d) positively assisting downward mobility

− abolition of inheritance.

Performance and Innovation Unit

April 2001
SOCIAL MOBILITY

Introduction

1. Equality of opportunity is a goal common to a broad spectrum of

political opinion. The Government’s first Annual Report on tackling
poverty and social exclusion [1] states that “everyone should have the
opportunity to achieve their potential”. A society characterised by
equality of opportunity will reward people according to their abilities,
talents and efforts rather than on the basis of their social background. It
might also be expected to exhibit high levels of mobility between social
groups.

2. The purpose of this paper is :

− to define in more detail what is meant by social mobility;

− to set out why social mobility is important;

− to summarise the key facts about social mobility – how much is

there, what have been the trends over time and how does social
mobility in the UK compare with that in other countries;

− to discuss the main drivers of or barriers to greater social

mobility; and

− to set out some tentative policy implications.

What is social mobility?

3. Social mobility describes the movement or opportunities for

movement between different social groups and the advantages and
disadvantages that go with this in terms of income, security of
employment, opportunities for advancement etc.

4. The academic literature draws a distinction between several

different kinds of social mobility :

− the sociological literature takes a definition of social mobility

defined in terms of movements between social classes or
occupational groups whereas the economics literature generally
focuses on income and income mobility. Income has advantages
as a direct measure of command over resources (at least at a
 point in time) but social class may be a better measure of life
chances;

− social mobility can be examined both intergenerationally (how

far the opportunities open to children are determined by the
social class or income of their parents) and intragenerationally
(to what extent individuals’ social position or income changes
over their own lifetime). Sociologists generally focus on
intergenerational mobility and economists on intragenerational
mobility (reflecting the nature of the data sets they use);

− sociologists make a further distinction between :

− absolute social mobility, which refers to the
absolute number or proportion of a social class
who move into another social group; and
− relative social mobility, which refers to the degree
of inequality in the likelihood of people from
different social class origins moving into a
particular social class.

5. Understanding these different definitions is important to the

description of the key facts about social mobility in the section below.

Why is social mobility important?

6. High rates of social mobility may be desirable for a number of

reasons :

− politically, because equality of opportunity is an aspiration of

politicians across the political spectrum;

− economically, because it is inefficient to waste the talents of

even one single person. Economic growth depends on the extent
to which everyone’s talents can be fully utilised; and

− socially, because social cohesion and inclusion may be more

likely to be achieved in a society where people believe they can
improve themselves through their abilities, talents and effort
than in a society where opportunities and quality of life depend
on social background.

7. Over the past 20 years, income inequalities have widened

significantly. Between 1979 and 1998/9, the real incomes of those in the
bottom decile of the income distribution rose by 6% in real terms whereas
the real incomes of those in the top 10% rose by 82%. Mean incomes rose
by 55% (also in real terms).

8. These figures are not compiled from a sample of people whose

incomes have been tracked in each and every year for 20 years. Rather, a
new sample is taken each year and the incomes of each decile of the
population compared. The figures are therefore potentially consistent both
with large numbers of people remaining at the bottom end of the income
distribution for long periods or with large numbers of people permanently
moving out of spells of low income. Large and growing income
inequalities may be more acceptable if there are significant opportunities
for social mobility, which allow individuals and their families to change
their position in the income and wider social distribution.

9. Social mobility is not without potential disadvantages, however :

− social mobility can be downward as well as upward i.e. there

are potential losers as well as gainers. A society with a lot of
upward and downward social mobility could be characterised by
considerable variation in individual and household income.
Significant change in individual and household incomes,
especially over short periods, could create economic instability
and social tensions unless there were economic and social
policy interventions which helped individuals and households to
cope with such income fluctuation;

− a society characterised by high levels of social mobility in

which the greatest rewards go to those with the greatest merit
(however defined) might not be an attractive one. Those
individuals with certain abilities or talents might scoop all the
rewards (the “winner takes all” society). By contrast, the
downwardly mobile (whose expectations and aspirations about
their social position and economic status will have been shaped
by the social position and economic status of their parents and
peers) are likely to be unhappy and resentful. This could
threaten social cohesion; and

− there is a danger that focusing on social mobility will lead to the

circumstances of the socially immobile being ignored.
Improving living standards may depend as much on increasing
opportunities for the immobile as breaking down barriers to
social mobility.
What are the facts about social mobility?

Intragenerational mobility

10. The academic literature gives only a partial picture of

intragenerational mobility in Britain. There are two main reasons for this.
First, sociologists have generally focused on intergenerational rather than
intragenerational mobility. Second, there have, until recently, been only
limited longitudinal data from which to draw conclusions. However,
several datasets have begun to provide a picture (mainly) of income and
earnings mobility in Britain in recent decades.

11. For example, the Department of Social Security’s analysis of

Households Below Average Income [2], uses the British Household Panel
Survey (BHPS) to analyse income mobility in Britain during the 1990s.
The BHPS was designed as a nationally representative sample of the
population of Great Britain in 1991. This sample has been re-interviewed
each year since then and there are now eight waves of data. The key
findings of this analysis are :

− there is a great deal of movement out of the bottom of the

income distribution. Half of those in the bottom two deciles in
1991 had estimated incomes above the bottom two deciles in
1998;

− but this movement is generally short range. Of those in the

bottom two deciles in 1991 less than one in four was in the top
60% of the income distribution in 1998.

12. Abigail McKnight [3] has used the New Earnings Survey Panel
Dataset and other sources to analyse trends in earnings mobility in
Britain over the 21 year period from 1977 to 1997. This finds :

− an overall decline in earnings mobility over the period,

particularly amongst prime age men (the earnings progression
of women appears to have improved over the period);

− a considerable degree of persistence of low pay with between 40

and 50% of employees, who remain in employment over a six
year period, remaining in the lowest quarter of the earnings
distribution;
− evidence of a low-pay/no-pay cycle with low paid employees
more likely to go on to experience spells of unemployment than
 higher paid employees and the unemployed most likely to enter
low paid jobs; and

− a steadily increasing proportion of employees in the highest

quartile of the earnings distribution who remain in that quartile
one year later, comparing 1977/78 with 1986/87 with 1995/96.

13. Other studies of income and earnings mobility (e.g. Dickens [4])
have made similar findings. Compared to, say, Germany there is some
evidence that poverty is more persistent in Britain.

14. Jonathan Gershuny [5] has examined a different dimension of

intragenerational mobility : career or occupational mobility. Using the
1986 Social Change and Economic Life Survey, he finds decreasing
upward mobility from manual occupations to higher status professional
and technical occupations over the four decades from 1941. During the
post-war period, recruitment to higher status occupations has increasingly
been direct from the education system, for those with the necessary
formal educational qualifications, and decreasingly dependent on mid-
career flows from lower status occupations.

Intergenerational mobility

The sociological approach and its limitations

15. There is an extensive literature (mainly sociological) on

intergenerational mobility. This draws on a combination of specially
conducted surveys and the analysis of pre-existing survey information
(such as the British Election Surveys and the General Household Survey)
to compare the social class of children with that of their parents.

16. This paper draws largely on an article by Anthony Heath and Clive
Payne [6], which provides a particularly clear analysis of the key facts
about intergenerational mobility in the UK. Any important differences
with other studies are noted, however, as the discussion proceeds.

17. Sociological studies analyse social mobility in terms of social class.

Heath and Payne use a 7 class schema (a modified version of the class
schema devised by John Goldthorpe) :
− I – professionals, managers and administrators in large
enterprises
− II – semi-professionals, managers and administrators in small
enterprises
− III – routine white-collar workers
 − IV – petty bourgeoisie (farmers, small employers and self-
employed workers)
− V/VI – higher working class (manual foremen, technicians and
skilled manual workers)
− VII – lower working class (semi-skilled and unskilled manual
workers)

18. This class schema is not completely hierarchical. Movements

between classes III, IV and V/VI are treated as horizontal rather than
vertical. For some of the analysis, classes I and II are grouped to form
what is called the salariat, and classes V/VI and VII are grouped to form
the manual working class.

19. Sociologists prefer to measure social mobility in terms of social

class rather than income because income is only one dimension of social
position. The Goldthorpe class schema is derived from measuring an
occupation according to 3 criteria : market situation (wage, pension, sick
pay, benefits); status situation (status of job); and work situation (level of
autonomy/control).

20. On the other hand, this approach also has a number of

disadvantages :

− the characteristics of social classes may not remain constant

over time. Andrew Adonis and Stephen Pollard [7] have
written, for example, about the emergence of a new, highly
prosperous, largely private sector employed “superclass” within
the middle class which stands in increasing contrast to the
traditional middle class of, for example, teachers and doctors.
Consistent with the Adonis and Pollard thesis, Abigail
McKnight [8] has found the growth in earnings inequality in
recent decades to be greater in the top half of the income
distribution than in the bottom half. More generally, there is
evidence that the greater part of the increase in earnings
inequalities is accounted for by the increasing variance in
earnings amongst those with similar education and experience
[9]. An individual may enjoy upward social mobility without
necessarily enjoying upward income mobility (to the same
degree);

− there are problems of classifying the social class of many

individuals. Women who have never had a job (e.g. because
they have been “looking after the home”) and economically
 inactive men who cannot be assigned a class on the basis of
their last job are excluded from many analyses;

− individuals’ social position may vary over their life cycle. It is

therefore important to compare the social class of parents and
children at similar points in their respective life cycles. Heath
and Payne do this by restricting their sample to respondents
aged over 35 (when it is argued they will have reached
“occupational maturity”) for comparison with their father’s
social position when they were growing up; and

− the extremes of the social classes (the elite and the underclass)
are probably under-represented in the various surveys.
Movement in and out of these extreme groups may be quite
different from mobility between the more broadly defined social
groups used in most sociological studies.

21. Heath and Payne’s study uses data from the 1964-1997 British
Election Surveys. These surveys were not designed for social mobility
analysis but the information they generate about the social class of
parents and children has been used extensively in social mobility
research. By dividing the data into cohorts of people determined by their
date of birth it is possible to establish how much social mobility there has
been in Britain during the 20th century and how it has changed over time.

Absolute social mobility in Britain

22. Trends in the absolute social mobility of men and women in Britain
are summarised in tables 1 and 2. They suggest that there is and has been
a considerable amount of movement between the social classes and that
increasing numbers of children have enjoyed upward social mobility
compared with their parents.
Table 1: Trends in the Absolute Social Mobility of Men During the
20th Century
(Column percentages)

Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

Percentage stable 43 39 38 35 34 28 35
Upwardly mobile 27 29 30 39 38 42 42
Downwardly mobile 20 21 20 17 18 19 13
Horizontal movements 10 11 12 9 10 11 10
Sample size 397 890 1303 1778 1572 1074 533

Notes: results presented by ten-year cohorts of the date of birth of respondents. Since only respondents
aged 35 and over are included, this effectively takes the analysis in this and the other tables from Heath
and Payne’s paper up to the mid-90s.

Source: Heath and Payne

Table 2: Trends in the Absolute Social Mobility of Women During

the 20 th Century
(Column percentages)

Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

Percentage stable 31 29 26 28 22 20 17
Upwardly mobile 24 22 27 23 29 32 36
Downwardly mobile 30 32 30 28 27 26 27
Horizontal movements 15 17 18 21 22 22 20
Sample size 219 602 1000 1403 1164 886 298

Source: Heath and Payne

23. This pattern of increasing upward social mobility has been made
possible by economic and social changes, which have increased
employment opportunities in the professional classes. Over the last
century the middle class has grown substantially in size relative to the
working class (table 3). This has created “more room at the top” and
permitted considerable upward social mobility in Britain. Sociologists,
such as Peter Saunders [10], have argued on the basis of these figures that
: “it is difficult to avoid the conclusion that Britain seems a remarkably
open society”. And, as income inequalities have increased in recent
decades, so the rewards for those who make it to the top have increased
too.

24. Looking, however, at the later decades of the 20th century, John
Goldthorpe and Colin Mills [11] find that the previous tendency for the
absolute rate of social mobility, especially upward social mobility, to rise
appears to have been halted with some kind of break around 1979-81. For
men, there appears to have been a small decrease in the rate of upward
absolute social mobility and a small increase in the rate of downward
absolute social mobility, with the reverse the case for women. Goldthorpe
attributes this to two factors : slower growth in the number of
professional and managerial jobs and a contraction in skilled manual
work (which meant there were fewer opportunities for short range social
mobility from the unskilled to the skilled working class).

Table 3: The Changing Class Structure of Britain – “More Room at

the Top”
(Column percentages)

Class Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

I 7.0 10.1 11.6 16.1 17.1 21.6 23.2

II 11.0 9.2 11.2 13.1 15.0 16.3 19.1
III 7.9 8.3 9.1 6.9 5.8 4.3 4.5
IV 12.2 10.3 8.7 8.9 13.3 15.0 15.3
V/VI 31.9 30.3 30.8 31.9 25.6 22.6 22.0
VII 30.0 31.7 28.7 23.1 23.2 20.2 15.8
Sample size 17 941 1393 1840 1617 1095 565

Source: Heath and Payne

25. The creation of “more room at the top” has allowed more children
from higher class backgrounds to remain in the same social class as their
parents alongside increasing upward social mobility. Thus :

− table 4 shows that 46% of sons with fathers in social class I are
themselves in social class I and a further 23% in social class II;

− table 5 shows that this has happened even though only 20% of
all people in social class I had fathers who were in social class I;
and

− more dramatically, table 6 shows increasing intergenerational

stability amongst the higher social classes. There has been
declining downward social mobility amongst the higher social
classes for much of the past century. Saunders [12] concludes
that, whilst the barriers against bright working-class children
succeeding are quite low, the safeguards against failure enjoyed
by dull middle class children are still quite strong.
Table 4 : Social Class Destinations of Men
(Row percentages)

Son’s class
Father’s class I II III IV V/VI VII Sample size

I 46 23 5 11 8 6 447
II 30 29 10 10 12 9 535
III 27 21 12 9 19 13 381
IV 15 13 7 29 20 17 1103
V/VI 12 12 7 8 38 23 2626
VII 9 9 5 7 32 38 2204
AII 16 14 7 11 28 24 7296

Source: Heath and Payne

Table 5 : Social Class Origins of Men

(Column percentages)

Father’s class Respondent’s class

I II III IV V/VI VII All

I 20.7 13.4 8.6 7.7 3.0 2.3 7.7

II 15.7 16.8 11.0 6.5 4.4 3.6 8.7
III 8.5 7.8 7.4 4.3 3.4 3.2 5.4
IV 13.4 14.0 13.5 36.2 9.9 11.5 14.1
V/VI 26.7 29.9 35.1 25.8 46.3 36.2 35.3
VII 14.9 18.1 24.6 19.5 33.1 43.1 28.8
N 1885 2589 3388 1367 3673 4589 17490

Source: Heath and Payne

Table 6 : Trends in Intergenerational Stability (percentage remaining

in the same class as their father): Men

Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

Salariat (I + II) 51 49 61 69 67 62 73
Working class 76 75 73 67 60 53 50
(V/VI + VII)

Source: Heath and Payne

Relative social mobility in Britain

26. Absolute social mobility is concerned with the absolute number or

proportion of people in a social group who are upwardly or downward
mobile. Relative social mobility is concerned with the chances people
from different backgrounds have of attaining different social positions.
These chances are measured by “odds ratios”.

27. Table 4 shows that 46% of men from class I origins stayed in class
I while 6% were downwardly mobile to class VII. This can be expressed
as odds of 7.7:1 (46/6). By contrast, 9% of men from class VII reached
class I while 38% stayed in class VII, giving odds of 1:4.2. The ratio of
these two odds is 32.3:1.

28. An odds ratio of 1 would indicate that the two classes had equal
chances of reaching one destination and avoiding the other. The odds
ratios would all be 1:1 in a society where social origins were unrelated to
class destinations. The larger the odds ratio the more unequal the
opportunities and the relatively less mobile is society.

29. Tables 7 and 8 summarise the trends in relative social mobility in

Britain during the 20th century. They look at relative social mobility
between:

− the salariat (classes I and II combined) and the working class

(classes V, VI and VII);

− the salariat and the petty bourgeoisie (class IV); and

− the petty bourgeoisie and the working class.

30. The tables suggest that the odds ratios may have declined over time
i.e. relative social mobility may have increased. But these odds ratios are
statistical measures constructed from sample surveys and therefore
subject to sampling error. Heath and Payne carry out some formal
statistical tests to establish whether the changes in the odds ratios are
greater than can be explained by random sampling error. They conclude
that there is some suggestion that the odds ratios have declined, albeit
modestly, over time for men but not for women.

31. Goldthorpe and Mills [13] have, however, called this finding into
question. Using arguably superior General Household Survey (GHS) data
(because of its larger sample size and higher response rates) for the period
1973-1992, they find no evidence of a statistically significant change in
relative social mobility in the later decades of the 20th century,
particularly in the case of men. Richard Breen and John Goldthorpe [14]
come to similar conclusions using data from the 1958 National Child
Development Study and the 1970 British Cohort Study.
Table 7 Trends in Relative Social Mobility During the 20 th Century:
Men

Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

Salariat: working class 16.0 10.0 19.0 14.0 10.3 5.6 7.7
Salariat: petty bourgeoisie 5.5 7.0 6.7 9.4 6.8 4.1 7.3
Petty bourgeoisie:
Working class 9.5 11.0 9.6 8.2 5.8 3.2 4.9

Source: Heath and Payne

Table 8: Trends in Relative Social Mobility During the 20 th Century:

Women

Pre-1900 1900-09 1910-19 1920-29 1930-39 1940-49 1950-59

Salariat: working - 17.2 15.2 13.4 7.3 10.6 5.8

Salariat:
routine non-manual - 2.1 1.0 1.5 1.7 1.3 0.6
Routine non-manual:
working - 6.5 4.5 1.9 3.3 2.7 3.2

Source: Heath and Payne

32. The economics literature is less rich in terms of the analysis of

intergenerational mobility than the sociological literature. However, there
have been some studies.

33. For example, Lorraine Dearden, Stephen Machin and Howard Reed
[15] have looked at the extent to which there is a correlation between a
parent’s position in the earnings distribution and that of his/her children
using data from the National Child Development Study. Zero correlation
would suggest complete mobility. A perfect correlation (1) would suggest
complete immobility. Dearden et al find correlations of between 0.4 and
0.6 between fathers’ incomes and sons’ incomes, and between 0.45 and
0.7 for daughters. Similar results are also obtained for a small sample of
adopted sons.

34. Consistent with the sociological literature, Dearden et al find that

the highest proportion of sons in the same quartile of the earnings
distribution as their fathers is to be found at the top of the earnings
distribution. In other words, upward mobility from the bottom is more
likely than downward mobility from the top.
Social mobility amongst women and ethnic minorities in Britain

35. The pattern of social mobility amongst women differs from men.
Women are more concentrated in routine white collar jobs. More room
for women has been found in the “middle” rather than “at the top” (which
may not really constitute a move up for working class women).

36. On the other hand, better education, increased labour force

participation and improved opportunities at work have had a positive
effect on women’s social status. The previous section suggested that, in
recent decades at least, women have, to some extent, enjoyed more
favourable social mobility trends than men, if mainly as part of a catching
up process.

37. There have been relatively few studies of social mobility patterns
amongst different ethnic groups. The main findings of those which have
been undertaken may be summarised as follows :

− the first generation of black migrants to Britain generally

experienced downward social mobility. Many of those who had
held white collar positions in their country of origin or whose
fathers had held white collar positions in their country of origin
were forced to take manual work on their arrival in Britain;

− in aggregate, the black population is more socially mobile than

the white population. Ethnic minorities are more likely to
experience both upward and downward social mobility than the
white population (see table 9 from a paper by Vaughan
Robinson [16]). However, there are significant variations
between different ethnic groups (most of the difference is
explained by Indians and Pakistanis) and there is increasing
polarisation within ethnic groups (with Indians and Pakistanis
more likely to experience both rapid downward and rapid
upward social mobility than the white population);

− though there is evidence that some groups within the ethnic

minority population are catching up with the majority white
population through rapid upward social mobility, ethnic
minorities are still disadvantaged compared with the majority
population. Table 10, taken from a paper by Yuan Cheng and
Anthony Heath [17], shows the proportion of graduates by
ethnic group in the (upper) service class. It shows that, whereas
90% of white British graduates are in service class occupations,
the corresponding percentages are 56% for Pakistanis, 62% for
 Indians and 68% for West Indians. Only Chinese and Irish
graduates are as likely to be in the service class. Using a formal
statistical model confirms these findings. Cheng and Heath
conclude that Indians, Pakistanis and West Indians suffer a
significant “ethnic penalty” even after controlling for education
and other factors. Anthony Heath’s work also suggests,
however, that the effect of ethnicity on social class destinations
is smaller than that of social class origins; and

− these points are broadly confirmed by interim findings from an

analysis of the experiences of immigrants (as opposed to the
ethnic minority population) using data for England and Wales
from the OPCS/ONS Longitudinal Study for 1971-1991. This
study, by the Migration Research Unit at UCL [18], finds that
people born in the Indian Sub-continent and Sub-Saharan Africa
have rates of net upward social mobility over twice that for
people born in the UK over the period (17% compared with
8%). However, those born in the Indian Sub-continent are less
likely to move into professional, managerial and technical
occupations (despite higher overall levels of mobility) than
those born in the UK whilst the reverse is true for those born in
Sub-Saharan Africa.

Table 9: Social Mobility by Ethnic Group, 1971-81

(percentages)

Indians Pakistanis West Indians Non-NCWP*

Stable 42 37 54 57
Mobile:
Rapid upwards 15 13 11 12
Slow upwards 15 17 12 11
Rapid downwards 15 17 11 10
Slow downwards 14 17 12 10
Total 100 100 100 100

*effectively white

Source: OPCS Longitudinal Study quoted in Robinson

Table 10: Access to the Service Class* by Ethnic Group, (a) for the
Total Population and, (b), for Graduates Alone, 1983-89

West African British-born

Chinese Indian Pakistani Indian Asian Irish whites

Males
All 25.6 21.4 8.7 9.6 26.0 20.2 33.3
Graduates 84.6 62.3 56.4 68.8 68.1 88.6 90.7

Females
All 30.3 13.4 17.1 34.5 18.8 28.3 26.1
Graduates 62.5 40.9 53.3 82.6 67.7 87.2 82.6

* large employers/managers of large businesses; professionals; managers of small businesses; and

intermediate non-manual workers.

Source: Labour Force Survey quoted in Cheng and Heath

Mobility in Britain compared with other countries

38. International comparisons of social mobility need to be made with

a degree of caution because the defining features of different social
classes are not necessarily the same in all countries. Nonetheless, studies
by Robert Erikson and John Goldthorpe [19] and by Gordon Marshall,
Adam Swift and Stephen Roberts [20] allow a number of conclusions to
be drawn :

− absolute rates of intergenerational social mobility exhibit wide

variations between countries and within particular countries
over time. There is no consistent ranking of countries, therefore,
according to absolute rates of social mobility;

− cross-national variation in absolute rates is explained by a range

of factors including (e.g.) economic growth, shifts in trade
patterns, changes in fertility, migration and wars. The
explanations are specific and historical rather than sociological.
One common feature can, however, be found : industrialisation
appears to bring about a significant step increase in absolute
rates of social mobility in its early phases when the outflow of
individuals from agricultural to non-agricultural employment is
at its most rapid. Thereafter, absolute rates of intergenerational
mobility show some tendency to fall;

− absolute rates of intragenerational mobility also tend to fall over

time in mature industrialised societies i.e. employment and class
position in early life become more predictive of employment
 and class position in later life (consistent with the findings of
Jonathan Gershuny’s work reported in paragraph 14);

− relative rates of social mobility exhibit a very high degree of

similarity between countries and over time within particular
countries. Such differences as exist are modest. Claims of
national “exceptionalism” – that the US is a distinctively “open”
or “fluid” society or that England is unusually “sclerotic” – are
not supported by the evidence (both countries in fact having
very similar patterns of relative social mobility [21]);

− within the modest differences that do exist two main clusters of

countries can be found : more fluid societies, such as Israel,
Australia, Japan, Poland, Sweden and Norway, and less fluid
societies, such as Hungary, the United States, Britain, Italy,
Germany, France, the Netherlands and Ireland. Since relative
rates of social mobility are comparatively stable over time, the
allocation of societies to these clusters does not change very
much; and

− Goldthorpe’s work suggests there is no statistically significant

association between type of political regime or, in democracies,
the party composition of governments and relative rates of
social mobility. Of the former Communist countries, for
example, Poland is in the more fluid category whilst Hungary is
in the less fluid category. Goldthorpe does, however, find an
association between relative rates of social mobility and the
degree of income inequality : the lower the degree of inequality,
the higher relative rates of social mobility.

What are the explanations for these facts and what are the barriers
to more social mobility?

39. There is a much broader consensus, and much better evidence,

about the causes of absolute social mobility than relative social mobility.
The rise in absolute social mobility is attributable to the structural
changes in the economy and society brought about by economic growth
and industrialisation. These have opened up “more room at the top” and
permitted considerable upward social mobility.

40. The main barriers to and/or determinants of relative social mobility

are more numerous. It is worth noting, however, that whatever
explanatory variables are included in formal statistical analyses of social
mobility (class origins, education, ability etc) the total variance explained
tends to be rather low, rarely over 20-25%. This suggests that
idiosyncratic factors and sheer luck (e.g. being in the right place at the
right time) play a very large part. Nonetheless, the key barriers and
determinants identified in the literature include :

Education

41. Education plays a crucial role in explaining social outcomes and is

especially important in accounting for long distance social mobility.
Breen and Goldthorpe [22], for example, find that educational attainment
has a substantial impact on occupational outcomes and mobility chances.

42. But the sociological literature also suggests that the role of
education should not be overstated and that its importance may be
weakening, for a variety of reasons :

− during the 20th century, the expansion in the number of

professional and managerial jobs preceded the expansion of
educational opportunities. Employers took on people to fill
these jobs whether or not they had the necessary education or
other qualifications. They then trained them as appropriate. In
other words, upward social mobility was driven by demand
from employers not by the supply of better qualified labour;

− John Goldthorpe’s work on cross-national differences suggests

there is no statistically significant association between
variations in relative rates of social mobility between
industrialised countries and differences in type of education
system, level of educational participation or measures of
educational opportunity;

− Anthony Heath’s work has shown that people from all social
origins have been able to take advantage of increased
educational opportunities. There is little evidence that
educational expansion has enabled the working class to catch
up; the middle class buy the best education either through
private schools or by moving into areas with the best State
schools;

− Goldthorpe and Mills [23] find that, even when individuals’

level of educational attainment is controlled for, a significant
association between an individual’s class origins and class
destinations remains;
 − there is evidence that women and ethnic minorities experience
less favourable occupational outcomes even when they have the
same or similar educational qualifications to their white, male
counterparts; and

− Breen and Goldthorpe’s [24] comparison of intergenerational

class mobility among individuals in the 1958 National Child
Development Study and the 1970 British Cohort Study suggests
that education has played a (modestly) diminishing role in
determining class destinations in the later decades of the 20th
century than in the earlier decades. This may seem at variance
with the finding of John Bynner et al [25] that the likelihood of
being in employment appears to be strongly and increasingly
associated with obtaining qualifications. But, as Saunders [26]
has said about higher education and which may hold for
education more generally, “… higher education is a classic
“positional good” – the more people who have it, the less
valuable it becomes as a means of entering top positions, but the
greater is the penalty for those who fail to get access to it at all
…”.

43. The reasons why educational attainment might have become less
important in explaining class destinations in recent decades have been set
out by Goldthorpe [27] :

− as any particular level of educational qualification becomes

more widely held across the workforce, the less information it
provides about the potential of those who possess it to
employers who may then give increased attention to other
indicators in making employment decisions;

− social and people skills, personal style, adaptability, team

working and other softer skills have become more important to
employers. This has been driven in part : by the growth in
service sectors, particularly personal services such as leisure
and entertainment; by the sale of an increasing range of high
value goods and services in a personalised way such as travel,
insurance and real estate; and by the expansion in what
Goldthorpe describes as various forms of people processing
such as customer service centres, call centres, help-lines etc.
Many of the softer skills required by employers may be
correlated with social class background rather than formal
educational attainment, which may help to explain the increase
 in the variance of earnings at any particular level of educational
attainment (see paragraph 20); and

− the deregulation of labour markets and the decline in trade

union membership may have led to reduced emphasis being
given to formal qualifications as recruitment criteria.
Historically, education has most strongly determined
employment prospects in unionised industries and in the public
sector.

44. There is considerable discussion in the literature about the role of

private education. Private schools appear to have been highly successful
in improving the standards of educational attainment of their pupils,
notwithstanding parallel improvements in State education. Higher levels
of educational attainment mean that private schools nowadays offer
access to good universities through academic qualifications rather than
through contacts or “old boy” networks.

45. A recent analysis by Anthony Heath and Sin Yi Cheung [28]

concluded that : “the independent schools have now largely adapted to the
new “credentialised” regime operated by universities and now ensure that
their pupils, who can no longer rely on connections to secure entry to
privileged universities, acquire the A-level grades necessary”. The fact
that relative rates of social mobility are similar across industrialised
countries in which private education plays widely differing roles may also
caution against giving too much weight to private education as a causal
factor.

46. Goldthorpe and others attribute the strong association between

class origins and class destinations, after controlling for educational
attainment, to differences in the material and other advantages enjoyed by
children from different backgrounds. Goldthorpe argues that the evidence
of an inverse relation between income inequality and relative rates of
social fluidity suggests that policy might concentrate on reducing
inequality of condition since, at least beyond some point, this appears to
militate against equality of opportunity. In this context, factors such as
childhood poverty; access to financial, social and cultural capital;
attitudes to risk; and opportunity hoarding may assume considerable
importance. These factors are discussed in the following sections but,
first, the paper addresses the question of the extent to which ability and
effort may differ between social classes.
Ability and effort

47. Sociologists such as Peter Saunders argue it is erroneous to assume

that, in a socially mobile society, there would be no association between
class origins and class destinations. Such an assumption would only make
sense if one believes that there exist no differences of aptitude and
motivation between members of different social classes.

48. The extent to which intelligence is inherited (whether genetically

or through socialisation) and causally related to class background is hotly
contested. Estimates range from zero to 80%. Some recent work suggests
that genetic and environmental factors interact in complex ways to
determine IQ and that this interaction underlies the significant long run
rise in IQ scores in industrialised countries. Nonetheless: “… assuming
… intelligence is normally distributed in the population, and that there is
a correlation of 0.5 between the intelligence of parents and that of their
children”, Saunders suggests [29], “ … existing patterns of social
mobility in Britain correspond almost exactly with the patterns which
would be found if class recruitment were based solely on differences of
intelligence between individuals”. It could also explain the
intergenerational correlation of earnings found by Dearden et al
(paragraph 33).

49. Saunders [30] has tested this proposition using data from the
National Child Development Study. He finds that intellectual ability (as
measured by test scores at age 11) is the single most important
determinant of class destinations and that its effect is independent of class
origins. Parents’ social class is a factor too but considerably smaller in
impact than either the intellectual ability or measured motivation of the
child.
50. Saunders’ findings have, however, been strongly disputed by other
sociologists. For example, Breen and Goldthorpe [31] have reanalysed
the NCDS dataset used by Saunders and found that, while ability and
effort play a part, the effect of class origins on class destinations is in fact
much stronger.

Childhood poverty : psychological and behavioural development

51. Why might there be a strong association between class origins and
class destinations even after controlling for educational attainment and
intellectual ability? One reason suggested by the literature is childhood
poverty. There is evidence (psychological as well as sociological) from a
number of studies that outcomes in later life are affected by the
experience of poverty in childhood but that carefully targeted
interventions (e.g. compensatory education programmes) can affect an
individual’s life experiences.

52. Abigail McKnight’s analysis [32] of the 1958 National Child

Development Study and the 1970 British Cohort Study has found a labour
market penalty is associated with childhood poverty :

− young people in low income households at age 16 are more

likely to be unemployed in their early 20s than young people
from higher income households;

− young people from poor backgrounds are disproportionately

observed at the lower end of the earnings distribution where
they are in work;

− even after controlling for differences in educational attainment

there remains a negative influence associated with growing up
in poverty. McKnight estimates that, at age 26, young adults in
the 1970 cohort experience an earnings penalty of 9% if they
were brought up in a household with an income below half the
average and 6% if they were brought up in a household with an
income below average (all after controlling for educational
attainment); and

− it appears that the labour market penalty associated with

growing up in poverty has increased over time. Young adults
from low income backgrounds in the 1970 cohort face greater
disadvantages in terms of the probability of being in work and
the size of the earnings penalty they face than young adults in
the 1958 cohort.

53. Childhood poverty may affect outcomes in later life in a number of

ways. At one level, physical effects may be at work e.g. the effect of diet
on brain development. At another level, childhood poverty may have
various psychological and behavioural effects.

54. A recent paper by Leon Feinstein [33] has highlighted the

importance of psychological and behavioural factors in early childhood
using data from the 1970 British Cohort Study. He finds that :

− psychological characteristics (such as feelings of self-esteem

and the sense of control over one’s destiny) and behavioural
 qualities (such as anti-social behaviour, peer relations,
attentiveness and extraversion) at age 10 are strongly associated
with social class background. Feinstein attributes this partly to
the impact of different material circumstances but also to
differences in child rearing abilities and psychological support
which impact on the self-confidence of children from different
social class backgrounds; and

− these psychological characteristics and behavioural qualities at

age 10 affect labour market outcomes in later life. Thus,
children with higher scores for self-esteem, for example,
experience shorter spells of unemployment and enjoy higher
wages in adulthood. Feinstein notes that candidates with greater
self-esteem are likely to be more confident and better able to
sell themselves at job interviews; candidates able to get on well
with their peers are likely to work better in teams and so on.
Employers may therefore quite rationally have a preference for
such candidates notwithstanding the potential association of
these characteristics with social class background.

The family : financial, social and cultural capital

55. Another possible explanation for the association between class

origins and class destinations is the family. The previous section has
already noted that parental upbringing may affect a child’s psychological
and behavioural characteristics and through those the child’s labour
market outcomes in later life. But this is only one possible channel
through which the family may affect social mobility. Goldthorpe argues
that the family has very powerful effects (despite all the talk of its
breakdown). Families throughout the world do all they can to help their
children and to prevent downward social mobility in particular. Indeed,
the dilemma for policymakers is that it is highly desirable in a civilised
society that parents should do this but the consequence may be to
reinforce inequalities of opportunity.

56. Families give their children access to various kinds of capital,

which may affect their class destinations :

Financial capital

57. Middle class parents can give their children access to financial
capital. There is evidence that access to modest amounts of financial
capital at an early age can have a major impact on outcomes in later life.
For example, Gavin Kelly and Rachel Lissauer [34] quote a study (using
data from the National Child Development Study) which found that
someone who was 23 in 1981, who had received a £5,000 inheritance (at
1981 prices), was approximately twice as likely to be self-employed as
someone who had received no inheritance.

Social capital

58. Social capital consists of the values and networks of contacts which
parents pass onto their children. Such social capital may work to the
advantage of children from middle class backgrounds by giving them not
only the motivation to succeed but access to extensive and diverse social
networks of individuals who can offer information or other support.

59. Perri 6 [35] has pointed out that the social networks of the middle
class are much more diverse than the social networks of the working
class. Working class sociability tends to revolve around close contacts
with kin and a small set of friends all of whom are relatively closely
connected with each other. By contrast, the middle classes have extensive
weak ties with, for example, former colleagues, acquaintances and friends
of friends. Such weak ties have been found to be more important than
strong ties for success in finding employment and other labour market
outcomes.

60. Middle class parents can give their children access to these
networks of weak ties, which may be a product of “white collar”
employment. Middle class children have further opportunities to develop
their own weak ties at university and elsewhere.

Cultural capital

61. The concept of cultural capital has been pioneered by France’s pre-
eminent sociologist Pierre Bourdieu. His work, drawing on extensive
surveys of attitudes, behaviour and occupational status, has shown that
cultural capital and social conventions (such as familiarity with particular
types of music, cuisine or sport) can serve as a powerful barrier to
mobility and are crucial to understanding how people internalise class
distinctions. His work has demonstrated :

− the importance of understanding strategies for trading economic

for cultural capital and vice versa (e.g. the nouveau riche
property developer who feels a need to give to the arts); and
 − the tensions between different parts of the elite (those whose
status primarily rests on cultural capital and those whose status
is primarily economic).

62. Analyses of this kind highlight the limitations of purely economic

models of class and mobility, and the importance of institutions such as
the family, private schools and universities in passing on cultural capital
and thus in determining life paths. Cultural factors may also go some way
to explaining the different social mobility experiences of different ethnic
groups (paragraph 37).

Aspirations and aversion to risk

63. Middle class parents are highly ambitious for their children and
adopt a range of strategies to support them. Working class parents may
lack the means to be as supportive, even if they are equally ambitious for
their children.

64. But being ambitious (e.g. about educational choices) has costs and
risks as well as potential rewards. There is evidence, for example, that the
labour market returns (the earnings premia) from a university degree
depend on the social class of the student, the type of school attended
before university, the subject studied and the class of degree obtained
[36]. This implies that there are risks involved in investing in higher
education, which may deter participation by poorer families.

Economic and other barriers : opportunity hoarding

65. The sociologist Charles Tilly [37] has focused on how particular
groups “hoard opportunities” as a potential barrier to social mobility. This
can take a variety of forms. For example :

− constructing barriers to entry to jobs (such as professional

qualifications); or

− adopting strategies to give one’s children access to a good

education e.g. by moving to the areas with the best State
schools. Such strategies and associated opportunity hoarding
may become more important in societies where income
inequalities are rising.

66. Increased competition tends to promote more meritocratic (i.e.

socially mobile) outcomes. The effect of deregulation in the City of
London is often cited as an example of this.
What are the implications for policy?

Policy objectives

67. In order to draw out the implications for policy of the preceding
sections of this paper, it is necessary first to consider what might be the
objectives of policy. Social mobility per se implies nothing for policy
since it is merely a descriptive term for the movement or opportunities for
movement between different social groups (or the lack of them).

68. Policymakers are likely to have a range of objectives, including :

− maintaining and promoting high levels of prosperity through

policies to, for example, raise economic efficiency, encourage
entrepreneurship and stimulate innovation;

− promoting a fair and just society, of which there may be several

dimensions :
− equality of opportunity : enabling individuals to
improve themselves through their own individual
ability and effort rather than class background,
gender or ethnicity (a meritocracy);
− social justice, cohesion and inclusion : providing
safety nets and opportunities for the disadvantaged
which reduce inequalities of outcomes and help
build communities.

− promoting individual freedom : either in the positive sense of

enabling individuals to make their own choices (e.g. by
ensuring the rule of law) or in the negative sense of minimising
coercion by the State or some combination of the two.

69. These different objectives may be complementary or in conflict

with each other. For example :

− policies to promote greater fairness might weaken incentives to

work, save, invest and innovate, undermining the goal of
maintaining and promoting high levels of prosperity. On the
other hand, a perception that economic and social outcomes are
fair and just may build trust and facilitate the more efficient
working of markets;
 − Goldthorpe’s work on cross-national differences (paragraph 38)
suggests that, at least beyond a certain point, increases in
income inequality may be associated with increases in
inequalities of opportunity and thus lower rates of relative social
mobility. On the other hand, a highly meritocratic society
characterised by high rates of social mobility might have
significant income inequalities and social tensions for the
reasons described below.

70. By a meritocracy is meant a society in which the most able and

committed people can succeed in attaining the most desirable, responsible
and well-rewarded positions. Unless one believed ability and
commitment were (either in whole or in part) determined by one’s social
class origins, the features of such a society would include high rates of
social mobility and the absence of any association between class origins
and destinations i.e. equality of opportunity.

71. Prima facie, a meritocratic society, as described in the previous

paragraph, seems a highly attractive one. But, it could also have
unattractive features :

− it would be characterised by a high rate of downward as well as

upward social mobility. As noted in paragraph 9, this could
create economic instability and social tensions. To the extent
that the subjective cost of losing a given amount of status or
income exceeds the subjective benefit of gaining the same
quantity of status or income, increased social mobility could
lower the mean level of social welfare. This loss might be offset
if increased social mobility raised the rate of economic growth
but it is not clear that it would necessarily do so;

− the losers or downward mobile in a meritocracy would have no

one to blame for their circumstances but their own lack of
ability and commitment. This could create a lot of unhappiness
and resentment;

− a meritocracy could be characterised by significant inequalities.

The winners or upwardly mobile could scoop all the rewards
leaving little for the rest. Egalitarians, who would prefer there to
be no losers, are therefore generally critical of the concept
meritocracy; and
 − though a meritocracy would be a highly competitive society, it
would not necessarily be economically efficient. The market
mechanism does not reward ability and commitment but those
who provide the goods and services consumers want to buy.
The more able and committed may be more likely to succeed in
a market economy but the lazy who have a good idea or the just
plain lucky may be highly rewarded too. As Saunders [38]
points out, libertarians such as Friedrich Hayek see this as a
price worth paying for the individual freedom (absence of
coercion) and economic dynamism offered by a market
economy.

72. Ultimately, the weight to be given to the different possible

objectives of policy and any trade-offs that may need to be made between
them require a political judgement. Studies of public opinion suggest the
public believes hard work, need and luck should be rewarded as well as
merit (aptitude and ability).

73. So far as social mobility and equality of opportunity are concerned,

there is a choice to be made between :

− measures which promote a weak form of meritocracy i.e. which

focus on removing barriers to upward mobility and assisting or
encouraging upward social mobility. This approach would give
primacy to absolute social mobility; and

− measures which promote a strong form of meritocracy i.e.

which not only seek to facilitate upward social mobility but
which seek positively to reduce barriers to downward social
mobility for dull middle class children (see paragraph 25). This
approach would give primacy to relative social mobility.

The challenge

74. Earlier sections of this paper suggest that using policy measures to
promote higher rates of social mobility may be a considerable challenge.
Goldthorpe’s view is that changing social mobility through policy
intervention is likely to be very difficult. Absolute social mobility has
been driven by structural changes in the economy and society over which
governments may have little direct influence. Relative social mobility has
been remarkably stable over time and across countries with widely
differing political systems, policies and institutions :
 − Goldthorpe has emphasised the role of the family in preventing
children from experiencing downward social mobility and in
adopting strategies to counter attempts by government to
promote more open societies;

− Tilly has emphasised how the upwardly mobile “hoard”

opportunities and how they seek to pass them onto their
children; and so on.

75. Significant interventions may be needed to overcome these

barriers.

76. A recent paper by John Roemer [39] contains an extremely

interesting analysis of what sort of education policy would be needed to
equalise the earnings capacity of blacks and whites in the United States.
Though concerned with race, the scale of intervention indicated has
intuitively plausible parallels with what might be required to tackle social
class differences.

77. Roemer estimates that to compensate for all the other

disadvantages associated with growing up black in the United States, it
would be necessary to spend at least 10 times as much per capita on the
education of black children as white children to equalise future wage
earnings. In practice, the amounts could be very much larger since the
marginal cost of reducing the impact of (say) social class on life chances
could rise exponentially, as its influence is driven towards zero, whilst the
benefits could be subject to the law of diminishing returns.

Future prospects

78. The implications for policy also depend on what are the future
prospects for social mobility in Britain. So far as relative rates of social
mobility are concerned, their historical stability across industrialised
countries for much of the 20th century might lead one to expect
comparatively little change in future. How absolute rates will evolve is
unclear. Some of the factors that may explain the levelling off in the rate
of absolute social mobility during the 1980s and 1990s may still be in
play (e.g. continuing de-industrialisation).

79. Structural economic changes may also have an impact. The “e”
revolution may open up more room at the top for people in professional
occupations. But it could equally lead to deskilling (with less skilled
people able to perform previously professional jobs such as accountancy
or giving legal advice) and growing numbers of unskilled or semi-skilled
jobs e.g. “white van men”.

80. Perri 6 [40] has set out how personal profiling, permitted by the
increasing flow of information about individuals, may allow employers
and others to make much more sophisticated assessments of the merits of
different individuals. This may require policies to give those who lose out
from such assessments second chances of various kinds.

Specific measures

81. The literature suggests a range of specific policy options for

promoting social mobility. They can be grouped under four headings :

− removing barriers to upward mobility;

− positively assisting upward mobility;

− removing barriers to downward mobility; and

− positively assisting downward mobility.

82. As noted in paragraph 72, the appropriate measures will depend on

the objectives of policymakers and their judgement about the trade-offs
that should be made between them. Measures to assist downward
mobility might, for example, carry high costs in terms of impinging on
individual freedom and/or in terms of maintaining or promoting high
levels of economic growth and prosperity.

83. The possible measures include, for example :

(a) removing barriers to upward mobility

− raising educational standards e.g. through league tables,

measures to tackle poorly performing schools etc;

− ensuring access to further and higher education for all;

− changing attitudes, increasing expectations and raising

aspirations e.g. by taking children to see what other walks of
life are like; cultivating greater self-confidence in individuals;
providing mentors and career advice for bright children from
 deprived backgrounds which encourage them to be
educationally ambitious etc;

− improving social mixing, both to influence attitudes and to help

people form networks with people from different backgrounds.
National service and citizens service, and projects like City
Year in the US are relevant here;

− reducing/removing barriers to geographical and occupational

mobility;

− greater disclosure of employers’ recruitment practices – gender,

race composition of workforce etc;

− removing non-competitive barriers to entry to professions such

as the law and medicine;

(b) positively assisting upward mobility

− tackling poverty and social exclusion, especially childhood

poverty. A variety of measures are relevant here : maintaining
high levels of employment; raising levels of in-work income;
and improving public services (both through increased
investment and through greater rights of exit where poor
services are delivered);

− increased investment in education, covering softer skills as well

as traditional academic subjects;

− giving young people better access to financial capital. For

example, David Nissan and Julian Le Grand [41] have proposed
the introduction of capital grants of £10,000 for all young
people at the age of 18;

− creating pathways for individuals to progress within the

workplace or within professions – for example, favourable
access to medical degrees for nurses; options for secretarial and
administrative staff to become managers; general improvements
in workforce development and training etc;

− giving people second chances by : encouraging employers to

bring in and bring on talented individuals who may not have
formal academic qualifications; more opportunities for
 education and learning later in adult life; and reforming the
insolvency laws to reduce the stigma of bankruptcy and
stimulate entrepreneurship;

(c) removing barriers to downward mobility

− reducing the weight given to geographical catchment area as a

determinant of access to the best State schools (to counteract the
scope for middle class parents to buy a good education for their
children by moving to the right area);

− higher rates of taxation of investment income and wealth;

− higher rates of inheritance tax; and

(d) positively assisting downward mobility

− abolition of inheritance.

84. There is little hard evidence on what might be the comparative

(cost-)effectiveness of these or other measures in promoting increases in
social mobility. However, the PIU currently has three projects underway
or in the process of getting underway which are relevant to the social
mobility issue:

− a project on future strategy for workforce development, which is

due to report by the Autumn of 2001;

− a project on the role of government loans, including possible

options for extending the use of loan mechanisms to support
adult learning, due to report by the Autumn of 2001; and

− a project on the current position and prospects of ethnic

minority communities, particularly in the labour market, which
is due to get underway later in the summer.

References

[1] Opportunity for All – Tackling Poverty and Social Exclusion, Cm

4445, September 1999;

[2] Households Below Average Income, Department of Social Security,

2000;
[3] Trends in Earnings Inequality and Earnings Mobility, 1977-1997 : the
Impact of Mobility on Long-Term Inequality by A McKnight, DTI
Employment Relations Research Series, No 8, February 2000;

[4] Caught in a Trap? Wage Mobility in Great Britain : 1975-1994 by R

Dickens, Economica, 2000;

[5] Post-Industrial Career Structures in Britain by J Gershuny in

Changing Classes : Stratification and Mobility in Post-Industrial
Societies edited by G Esping-Andersen, Sage, 1993;

[6] Social Mobility by A Heath and C Payne in Twentieth Century British

Social Trends edited by A H Halsey with J Webb, Macmillan, 2000;

[7] A Class Act – The Myth of Britain’s Classless Society by A Adonis

and S Pollard, Penguin Books, 1998;

[8] see [3];

[9] What Has Happened to Wages? by A Gosling, S Machin and C

Meghir, The Institute for Fiscal Studies, Commentary No 43, June 1994;

[10] Might Britain Be a Meritocracy? by P Saunders, Sociology,

February 1995;

[11] Trends in Intergenerational Class Mobility in Britain in the Late

Twentieth Century by J Goldthorpe and C Mills, December 2000
(working paper);

[12] Social Mobility in Britain : An Empirical Evaluation of Two

Competing Explanations by P Saunders, Sociology, May 1997;

[13] see [11];

[14] Class Mobility and Merit : the Experience of Two British Birth
Cohorts by R Breen and J Goldthorpe, 2001 (working paper);

[15] Intergenerational Mobility in Britain by L Dearden, S Machin and H

Reed, The Economic Journal, January 1997;

[16] Roots to Mobility : the Social Mobility of Britain’s Black Population,

1971-87 by V Robinson, Ethnic and Racial Studies, April 1990;
[17] Ethnic Origins and Class Destinations by Yuan Cheng and Anthony
Heath, Oxford Review of Education, 1993;

[18] International Migration and the United Kingdom : Recent Patterns

and Trends, interim report to the Home Office by the Migration Research
Unit, UCL (forthcoming);

[19] The Constant Flux – A Study of Class Mobility in Industrial Societies

by R Erikson and J Goldthorpe, Clarendon Press, 1992;

[20] Against the Odds? Social Class and Social Justice in Industrial
Societies by G Marshall, A Swift and S Roberts, Clarendon Press, 1997;

[21] Are American Rates of Social Mobility Exceptionally High? New

Evidence on an Old Issue by R Erikson and J Goldthorpe, European
Sociological Review, May 1985;

[22] see [14];

[23] see [11];

[24] see [14];

[25] Obstacles and Opportunities on the Route to Adulthood by J Bynner,

H Joshi and M Tsatsas, The Smith Institute, 2000;

[26] Unequal But fair? A Study of Class Barriers in Britain by P

Saunders, IEA Health and Welfare Unit, July 1996;

[27] Mobility, Education and Meritocracy by J Goldthorpe, Maxwell

Cummings Lecture, McGill University, Montreal, March 2001;

[28] Education and Occupation in Britain by A Heath and S Y Cheung in

From School to Work : A Comparative Study of Educational
Qualifications and Occupational Destinations edited by Y Shavit and W
Muller, Clarendon Press, 1998;

[29] see [26];

[30] see[12];

[31] Class Inequality and Meritocracy : A Critique of Saunders and An

Alternative Analysis by R Breen and J Goldthorpe, British Journal of
Sociology, 1999;
[32] From Childhood Poverty to Labour Market Disadvantage : A
Widening Gulf Between the Prospects of Rich and Poor Kids by Abigail
McKnight (undated working paper);

[33] The Relative Economic Importance of Academic, Psychological and

Behavioural Attributes Developed in Childhood by Leon Feinstein in
What Do We Know About Brain Development and Childhood
Interventions?, Centre for Economic Performance, London School of
Economics, July 2000;

[34] Ownership for All by G Kelly and R Lissauer, Institute for Public
Policy Research, 2000;

[35] Escaping Poverty – From Safety Nets to Networks of Opportunity by

Perri 6, Demos, 1997;

[36] Sheer Class? The Extent and Sources of Variation in the UK

Graduate Earnings Premium by R Naylor, J Smith and A McKnight,
January 2001 (working paper);

[37] Durable inequality by C Tilly, University of California Press, 1998;

[38] see [26]

[39] Equality of Opportunity by J Roemer in Meritocracy and Economic

Inequality edited by K Arrow, S Bowles and S Durlauf, Princeton
University Press, 2000;

[40] Divided by Information? The “Digital Divide” That Really Matters

and the Implications of the New Meritocracy by Perri 6 with Ben Jupp,
2001 (working paper);

[41] A Capital Idea – Start–up Grants for Young People by David Nissan
and Julian Le Grand, Fabian Society, February 2000;

Acknowledgements

This paper was prepared originally as background for a seminar on social

mobility held by the Performance and Innovation Unit on the 20th March
2001. It has been revised in the light of the discussion and comments at
that seminar. A list of seminar participants is given at Annex A.
The paper also draws substantially on a series of helpful discussions and
correspondence with : Andrew Adonis, John Bynner, Joanne Drean,
Stephen Glover, John Goldthorpe, Anthony Heath, John Hills,
Christopher Jencks, Alison Kilburn, Julian Le Grand, Abigail McKnight,
Nick MacPherson, Peter Saunders, Perri 6 and Suzy Walton, none of
whom are responsible for the conclusions drawn or any errors made.

Performance and Innovation Unit

April 2001
ANNEX A

SOCIAL MOBILITY
Participants in the Seminar held by the Performance and Innovation
Unit on the 20th March 2001

Chair :
Geoff Mulgan Director PIU

Presenters :
Stephen Aldridge Chief Economist, PIU
Lynne Berry Chief Executive, Equal Opportunities Commission
John Goldthorpe Professor of Sociology, Nuffield College, Oxford
Anthony Heath Professor of Sociology, Nuffield College, Oxford
John Hills Professor of Social Policy, LSE
Julian Le Grand Richard Titmuss Professor of Social Policy, LSE
Graham Mather President, European Policy Forum
Peter Saunders Professor of Sociology, University of Sussex
Perri 6 Director, Policy Programme, Institute for Applied
Health and Social Policy, Kings College, London

Attendees
James Bowler PS/Ed Balls, H M Treasury
John Bynner Professor of Social Science, Institute of Education
Suma Chakrabarti Head, Economic and Domestic Affairs Secretariat,
Cabinet Office
Selina Chen Head of Research, Social Market Foundation
Phil Collins Director, Social Market Foundation
Stephen Donnelly Equality Unit, Office of the First Minister & Deputy
First Minister, Northern Ireland
Lucy De Groot Director, Public Services Directorate, H M Treasury
Althea Efunshile Head, Children and Young Person’s Unit, DfEE
Naomi Eisenstadt Director, Sure Start Unit
Jonathan GershunyProfessor of Economic Sociology, University of Essex
Tony Giddens Director, LSE
Elouise Hayward Economic Adviser, H M Treasury
Gavin Kelly Senior Research Fellow, IPPR
Steven Machin Professor of Economics, UCL
Gordon Marshall Chief Executive, Economic and Social Research
Council
Abigail McKnight Toyota Research Fellow, LSE
Max Nathan Futures Researcher, Industrial Society
Tariq Modood Professor of Sociology, University of Bristol
Nicky Roche Policy Innovation Unit, DfEE
David Russell Policy Innovation Unit, DfEE
Maeve Sherlock Council of Economic Advisers, H M Treasury