Jump to ContentJump to Main Navigation
The success paradoxWhy we need a holistic theory of social mobility$

Graeme Atherton

Print publication date: 2016

Print ISBN-13: 9781447316336

Published to University Press Scholarship Online: September 2016

DOI: 10.1332/policypress/9781447316336.001.0001

Show Summary Details
Page of
Subscriber: null; date: 26 February 2017

Social mobility: rising, falling or staying the same

Social mobility: rising, falling or staying the same

(p.29) Two Social mobility: rising, falling or staying the same
The success paradox

Graeme Atherton

Policy Press

Abstract and Keywords

This chapter analyses in detail how social mobility has been defined and then researched by sociologists and economists since the mid part of the 20th century. It shows how the strong association between social mobility and economic progress was only one route that could have been taken in understanding social mobility and it was one shaped to a great extent by the work of sociologists operating in a broad Marxist tradition. The chapter goes onto explore the disagreements between sociologists and economists in recent years regarding the extent of social mobility in the UK. It concludes by arguing that there is a lack of consensus where the extent of social mobility is concerned in the UK and depending on who you believe it may be going up, remaining static or going down.

Keywords:   socialmobility, sociologists, economists, Marxist


For most of those who study social mobility, it is interesting because it is a vital part of a bigger picture. For sociologists, it enables them to be able to better understand how society is structured. One of the most important contributions to such sociological research was Erikson and Goldthorpe’s Constant Flux in 1992. This book was a major comparative study of social mobility across nine different countries. However, one of the most revealing parts of the book with regard to why social mobility should be studied at all is on the very first page: ‘Discussion of social action requires reference to actors who have, in the last analysis, to be recognized as individual men and women: we must as Stinchcombe has put it, see the social structure as being “peopled”’ (Goldthorpe and Erikson 1992: 1).

The question is whether sociological research manages this ‘peopling’ successfully or not, or whether it falls back on a preference for the creation and analysis of structures within which people are slotted. This same question also applies to the work of economists. Throughout the different theories outlined in this chapter, how social mobility is examined is always a product of a broader set of beliefs or principles within which it is positioned. This applies to sociologists with their adherence to providing macro-level structural explanations of how society works, or economists and their commitment to mathematical modelling as the route to understanding better how scarce resources are allocated.

The problem with this approach to social mobility is that, as a concept, it is always defined – and thus trapped – within this broader set of principles. Hence, the understanding of what social mobility is and its importance is always partial.

(p.30) Pitirim Sorokin and the meaning of ‘stratification’

The major pre-war sociological contribution to understanding social mobility came from the work of Pitirim Sorokin. The way in which Sorokin considered social mobility was broader. Less trammelled by the development of an academic discourse in the area (which, as in any area of established academic discourse, can close off debate by setting parameters to discussion), Sorokin was able to consider the meaning of social mobility.

For Sorokin: ‘by social mobility is understood any transition of an individual or social object or value – anything that has been created or modified by human activity from one social position to another’ (Sorokin 1959: 10).

The schema he developed gave equal weight to the importance of downward, upward and horizontal social mobility. Sorokin was equally concerned with social mobility in a range of forms, not just the upward social mobility that concerns our present-day politicians. But he was also concerned with structures of stratification per se, rather than structures defined by social class, occupation or income, unlike our present-day sociologists and economists.

Sorokin did not deny the importance of economic position in defining social mobility, particularly in Western capitalist economies, but he also saw society as politically stratified when its social ranks were hierarchically structured with respect to authority and power as well. Stratification was economic, political or occupational. It was not necessarily a unitary concept but had different facets. As his scope was social mobility and stratification as a concept, Sorokin also looked at different countries and cultures, pointing to the way in which the caste system in India, for instance, as a form of stratification differed from those systems based on occupation in Western economies. This scope allowed him to see social mobility as context-specific.

Sorokin also recognised the critical importance of the movement of groups. While he fully acknowledged that the study of social mobility needed to concern itself with the movements of individuals, it also needed to pay close attention to the impact of these movements – for the social groups who are mobile, those who are not, and for the structures within which these groups exist. This concern for the interaction between social mobility and the society in which it occurs also extended to ambivalence regarding the merits of mobility. Sorokin described a price to be paid for social mobility in dysfunctional outcomes for society, in terms of negative psychological impacts, social isolation and loneliness for individuals who move away from (p.31) their familiar cultural context into one where they feel less of a sense of belonging.

The great value in starting with Sorokin is to illustrate that there is a precedent where different ways of considering social mobility are concerned. By focusing on the importance of context in understanding social mobility, Sorokin highlights that stratification is not static. In the UK, for example, the significant changes in ethnic composition that have occurred at the same time as the rise in the study of social mobility do not appear to have been able to penetrate the academic orthodoxy where social mobility is concerned. A more reflexive ‘Sorokinesque‘ approach may have enabled greater discussion, in the UK context at least.

Finally, this kind of approach also allows what is understood by ‘stratification’ to change over time as societies change. This insight is extremely important. The objective of the holistic approach is to take a forward-looking perspective on social mobility in the 21st century. The likelihood of social, economic, technological or political change shifting how stratification is understood must be recognised.

David Glass and the dominance of class

As Payne (1990) argues, Sorokin’s pluralistic approach did not find favour in the study of social mobility of the UK mainly due to the influence of Glass’s landmark work, which was the first real attempt at defining the field of enquiry in the UK. Social Mobility in Britain, published in 1954 (Glass 1954), drew mainly from national research conducted in 1949 with a random sample or 10,000 adults, which was the largest study of its kind undertaken at the time. In selecting occupation as the indicator of social status and concentrating on intergenerational differences between father and sons as the measure of social mobility, Glass frames the rest of the sociological study that has been undertaken since then and up to the present day, accepting the relatively recent increase in attention given to the position of women. However, Glass’s study may actually take a wider view than Payne gives it credit for. Two of the chapters look at the stratification in voluntary organisations. The rationale for this is the argument that leadership in such organisations offers an alternative form of social prestige. While this line of enquiry was not taken forward by stratification sociologists post-Glass, it does reveal that this study was still searching for the dominant definition of status in the UK, and was at least open to the possibility that there were complementary forms of stratification to class.

(p.32) Glass’s book also contains chapters looking at the influence of educational background on intergenerational mobility. It was produced in the wake of the Education Act 1944, which ushered in the biggest transformation in the English school system in its history. In his opening commentary, Glass felt that the increasing importance given to education by the Act would lead to education itself driving greater social mobility. Hence, his book also laid some foundations for the present political discourse and its belief in the power of education.

However, Glass was also sensitive to some of the potentially negative implications of social mobility. He argued that the three-track system of grammar, secondary modern and technical schools could lead to some being left behind. He was concerned about the resentment that those who did not make it to grammar school, and hence who did not progress to higher prestige occupations, may feel, compounded by the fact that their failure to do so was not due to some hidden mechanism of exclusion, based in nepotism or economic disadvantage, but due to their public inadequacies as they were unsuccessful in the ‘fair’ selection system of the 11-plus exam. Glass suggested that to reduce this risk, ways of increasing the social prestige of the occupations that such young people did enter should be considered.

The willingness to consider the implications for social mobility on society itself, as well as an acute awareness of the case for action to change what prestige means for the good of society, illustrates that Glass is not only part of the evolution of academic and political discourses, but also of the holistic one being developed in this book.

Michael Young, meritocracy, industrialism and historicism

Payne (1990) also goes on to describe a dearth in research into social mobility in the UK for the 20 years after Glass until another landmark piece of research, the Oxford Mobility Study, conducted in 1972. However, this did not mean that nothing was produced over that period germane to the social mobility debate in the UK.

The Rise of Meritocracy, written by Michael Young in 1958, was a satire warning against the dangers of building a society where status is based entirely on merit, when that society is already stratified by wealth and power. While it is not a piece of sociological research, it has had a significant impact on how social mobility is conceived. Young picks up on the worries that Glass expressed of the danger of frustration and injustice becoming even more pronounced as society moves to becoming ‘fairer’, with rewards not distributed by ascription but by education. His narrator describes how, over the mid-part of the (p.33) 21st century, society became ossified into a number of self-selecting classes based on measures of intelligence. As what Young describes as the ‘bureaucratic system’ becomes more efficient, intelligence can be measured with greater and greater accuracy in ever younger children. By 2030, the potential of children would be accurately identified in children as young as three. The continual advances in the ability of society to identify and reward merit only serves to produce a self-recruiting elite, who use their moral authority to further cement their power and rewards. The narrator describes the breakdown of society as those whose inferior abilities exclude them from the higher classes’ revolt.

This dystopian vision of meritocracy, while widely read and quoted, has had something of the opposite impact on the discourse around stratification, social mobility and education that Young intended. Rather than acting to warn against the dangers of continual educational expansion in a stratified and unequal society, the need to create a more meritocratic society in Britain has been at the centre of the political social mobility discourse since the 1990s. Writing in The Guardian in 2001, Young says: ‘It would help if Mr Blair would drop the word [meritocracy] from his public vocabulary, or at least admit to the downside’ (Young 2001).

The inability or unwillingness of politicians to confront the extent of change necessary to make meritocracy viable is Young’s major concern. Several of the developments that Young foretold in fiction have come to resemble reality. As well as the earlier identification of potential in children, which is quite prescient of the extensive work that has been undertaken into measuring ability among infants from different groups in the last 20 years (see Chapter Four), Young’s book also describes how higher education would come to be essential for admission to higher class groups, which it now is. The rise of a powerful, self-recruiting, self-aggrandising elite is exactly the problem that tasks the majority of politicians as well as many academics looking at social mobility today. Other elements of the book remain fictional, however, and have yet to resemble fact. Young describes how competition would wither away, and the discontent that comes from a system that is unequal but now justly so has yet to come to pass, and probably will not do so by 2030. There may well still be the kind of negative outcomes from a justly unequal system, in terms of alienation, frustration and resentment among those who see themselves as failures in this kind of society but who are not minded to revolt – at least not in an organised, political way. In highlighting so clearly the particular nature of meritocracy, and its difference from equality of outcome, Young’s book is extremely (p.34) important. As with Sorokin’s work and, to an extent, Glass’s work, it reinforces the idea that the study of social mobility should place at its centre the impact (or lack of impact) that it has on the society in which it occurs.

In a similar vein to Young, a number of US writers from the 1960s onwards sought to make an explicit link between upward social mobility and industrialisation (see, for example, Blau and Duncan 1967, Kerr et al 1973). In the US, a theory was also being developed which had a significant impact on the understanding of social mobility, and which attempted to place it in the context of the ‘development’ of society. There is an inherent ‘logic’ in the shift from pre-industrial to industrial societies that brings with it a common pattern of mobility across such societies. This common pattern is an increase in both absolute and relative mobility, and these changes are necessary parts of the development of industrial society. It is an approach driven by a historicist account of human development. There is an inevitable move to pluralist democracy, as it is the form of political arrangement most functional for the equally inevitable progression of all societies from the industrial to the pre-industrial. Social mobility takes the particular form it does within this context due to the role of technology in increasing the skill requirements of employment, creating more of the kind of managerial, administrative positions that Young also describes. This shift is coupled with a move to achievement and indeed meritocracy as qualifications become far more important in the recruitment of staff, and as we move away from an ascription economy where upward social mobility is based on social backgrounds towards one based more on the qualities of the individual.

Finally, while mobility is only part of a more comprehensive story, it is an important one. Social mobility contributes significantly to the stability of liberal democracies, by providing a justification for their existence. By being based on achievement, inequality is legitimised, and with the bias towards upward mobility, presumably the majority of the population will feel satisfied with what the state offers them.

Such ‘liberal theories of industrialism’ (Bell:1973) have been heavily criticised by other sociologists, in particular Goldthorpe and Erikson (1992), on the grounds of their historicism and inaccuracy. The relationship between economic growth and progression and social mobility is more complex than the liberal industrial theory assumes. For example, the role of technology in social mobility, assumed to be benign, may in reality be anything but (as argued in Chapter One).

(p.35) John Goldthorpe and the importance of measurement

The 1972 Oxford Mobility Study was reported in the 1980 book Social Mobility and Class Structure in Modern Britain, by John Goldthorpe and colleagues, which marked the beginning of the period of dominance in the study of social mobility in sociology which continues to the present day. Goldthorpe et al’s analytical techniques, conception of what the study of social mobility entails and their arguments surrounding the extent of it have set their own parameters around how the vast majority of sociologists look at social mobility (Goldthorpe et al 1980; Goldthorpe 1996, 2007, 2012; Goldthorpe and Jackson 2007; Goldthorpe and Mills 2008), as later acknowledged by Payne and Roberts (2002).

The Oxford Study was not dissimilar to Glass’s study in the early 1950s, as it was a major national representative survey of over 10,000 men aged 20 to 64 in England and Wales. Respondents were asked to report their current occupation and employment status and those of their father when the respondents were about 14 years old. An assumption was made regarding the degree of career mobility across the lifespan. Goldthorpe et al take father’s occupation at the age of 35 as what they describe as ‘occupational maturity’. Beyond that, it is assumed there would be relatively little career mobility.

Occupation and employment status data were then coded into a seven social class schema designed by Goldthorpe himself, based on market situation (source and level of income, security of employment, promotional aspects) and work situation (degree of control and authority in job). There was, however, a simplification of these social classes into three groups: service class, intermediate class and working class. While the service class included experts and specialists filling important positions in the dominant institutions of society, it is far broader than what could be understood as a conventional elite; Goldthorpe, however, felt that there was a significant level of social distance between the two groups with an intermediate group in between.

Goldthorpe’s major conclusion is that at the time of this study there was significant upward social mobility in the absolute sense as the occupational structure was changing. Service and intermediate group jobs were expanding creating more ‘lower middle’ and ‘middle’ class’ jobs which needed to be filled. But the relative levels of mobility were not increasing. The chances of these new jobs being filled by those from the existing working and middle class locations were not changing. It (p.36) was still the case that the chances of middle class jobs being taken by working class people was far lower.

This distinction between absolute and relative mobility is a crucial one for Goldthorpe, and the failure (in his eyes) of politicians to grasp it is his major bugbear. He defines these two concepts as: ‘Absolute rates refer to the actual proportion of individuals of given class origins who are mobile to different class destinations, while relative rates compare the chances of individuals of differing class origins arriving at different class destinations and thus indicate the extent of social fluidity’ (Goldthorpe 2012: 3).

It is the relative, intergenerational mobility of those from different social classes that has been Goldthorpe’s major concern, and for him, it is what social mobility means. The 1972 Oxford Mobility Study showed that while in the early 1970s there may have been ‘room at the top’, some social groups were more likely than others to take advantage of the new opportunities that were emerging. Whatever the chance of a working-class boy reaching the service class, a boy from the intermediate class had twice the chance and a boy from the service class four times the chance. Relative mobility rates had remained remarkably constant – there had not been reductions in class inequality or increases in social mobility. Those in higher social positions were still far better off than those in the working classes. Overall, though, Goldthorpe did find a significant degree of this absolute mobility. Over 70% of the children in the study were in different social classes to their fathers. On this reading there was significant social mobility in the UK, but it did not, in Goldthorpe’s view, imply greater equality of opportunity.

This ‘room at the top’, however, was not a permanent phenomenon. As the century progressed, later cohort research shows a levelling out of absolute mobility. This led Goldthorpe to conclude that the ‘change observed in absolute rates was in very large part structurally determined’ (2012: 6). In the same 2012 article, Goldthorpe summarises his position on social mobility across the 20th century, and it is clear that it has also remained relatively consistent over a 30-year period. On the basis of his later work, for example, Goldthorpe and Mills (2008), which brought together data from 13 representative surveys of the adult British population carried out between 1972 and 2005, and that of others such as Paterson and Iannelli (2007) who looked at both men and women born between 1937 and 1946 comparing them to those born between 1967 and 1976, Goldthorpe states:

Absolute rates of inter-generational class mobility … appear quite high … rates of upward mobility increased in the (p.37) course of the twentieth century primarily as a consequence of class structural change ie of the expansion of professional and managerial positions creating more ‘room at the top’. However, immobility at the ‘top’ also increased, and there were indications that towards the end of the century rates of upward mobility among men were stabilising, as the growth of professional and managerial positions began to slacken and as men faced greater competition for these positions from women. Relative rates of intergenerational class mobility … showed a basic constancy over most of the twentieth century, or at all events no sustained directional change, with the possible exception of some recent slight increase in fluidity among women.

(Goldthorpe and Mills 2008: 88)

Those sociologists who do not agree with Goldthorpe come at his work from a number of angles.

Geoff Payne, for example, has been writing about social mobility in the UK for about as long as Goldthorpe, but he parts company with him mainly by stressing the importance of class – as opposed to occupational change – in understanding inequality and social mobility, refuting the lesser importance that Goldthorpe places on absolute mobility. Payne argues that the fact that large numbers of people have moved out of lower socioeconomic groups is of greater consequence than Goldthorpe allows for. Its major implication has been that the actual size of these different groups has changed, so that there are actually far fewer people at the lower end of this stratification system. These changes in the composition of the overall class structure are important for both the individuals involved and society itself. As Payne states (1990: 297): ‘mobile people and indeed the immobile too, do not encounter mobility as a relative chance, but rather as mobility or non-mobility.’ This attention to the lived experience of mobility and the role of occupational change in both driving and defining lifecourse mobility makes Payne’s contribution very valuable in the context of the ideas being developed here. Understanding the nature of changes in the structure of the economy and labour market are legitimate and vital parts of the study of social mobility. Payne (1990: 295) describes the problem with Goldthorpe’s work here as being not that it is ‘interested in class, but to narrowly interested in class’. This narrowness means that while Goldthorpe’s work is, in many ways, the starting point for understanding social mobility in the UK in the last 50 years, it should not also be the end point.

(p.38) The commitment to a class-based understanding of social mobility is based in Goldthorpe positioning his approach within a broad Marxist context. This commitment, though, leads him, in Payne’s opinion, not to tackle some of the implications of his findings. The extent of social mobility that the 1972 study found, for example, calls into question the cohesiveness of the different class groupings. If they are subject to what Goldthorpe himself described as ‘constant flux’, then how could they develop the kind of class consciousness that is a prerequisite for collective action, which is the driver of social change in any Marxist-inspired framework?

Payne’s critique is again very useful, in turning the spotlight on how class is understood by those who experience it. Goldthorpe remains firmly committed to the primacy of class as the best way of understanding stratification and the major determinant of life chances into the early 21st century. This is a contentious commitment in a number of ways, but none more so than in the area of consciousness. The evidence suggests that ‘the subjective significance of class has declined considerably over the past 30 years’ (Park et al 2013:180), that the class that people think they are in is far less useful in understanding their attitudes. Nor is socioeconomic group as measured by occupation a very strong predictor of their views on welfare or politics, for example. Rather than the invisible hand of the market in the 21st century (which, given the rampant competitive ethos now extending to all aspects of life, is now in quite open view), it may make more sense to talk of the invisible hand of the class system, which is structuring our everyday lives but in a way that we are far less conscious of. To the success paradox we must add what Savage (2000) describes as the ‘paradox of class’.

Payne does not confine himself to offering a critique of Goldthorpe. In 2002, Payne looked at a different set of data to that of Goldthorpe’s in 1972, and argued in Payne and Roberts (2002) that social mobility, in both absolute and relative terms, was increasing at a greater level than work in what the paper describes as ‘the Nuffield tradition’ (inspired by Goldthorpe and colleagues from Nuffield College) allows. Expanding on the critique offered in 1990, Payne and Roberts argued that the longitudinal survey data they used in their paper, drawn from the British Election Study (BES) carried out at each general election, was superior to the ‘one-off’ cohort analysis used in the 1972 study. The problem with such cohort comparison work (and it is a criticism that Goldthorpe himself applies with some force to the economistic approach to social mobility pioneered by Machin, Blanden and others, as we shall see later in the chapter, is that it is difficult to hold age constant, so that those of different ages are compared when they are (p.39) actually at very different points in their careers. Payne and Roberts also point to the extreme limitations of a survey that focuses only on the experiences of men. While later work by Goldthorpe and others does look at women separately (Goldthorpe and Mills 2008), Payne and Roberts bring home the importance of maintaining this distinction in the very nature of social mobility analysis.

Absolute versus relative social mobility

The final critique of Goldthorpe offered by Geoff Payne and Judy Roberts returns to the issue of absolute and relative mobility. They argue that the separation of these two forms of mobility is taking the study of the field down the wrong direction, and that it is the interaction of these different phenomena that should be the focus: ‘Equally, although the odds-ratios used to express “relative mobility” may be statistically calculated independent of the marginal totals, the numbers of people following particular origin-to-destination trajectories are not sociologically independent of changes in the distribution of origins and destinations’ (Payne and Roberts 2002: 13).

Looking at the BES, Payne and Roberts find that Britain was more open than the work from 1972 suggests. At the same time, however, their conclusions do not look hugely different in some ways to the overall summary of the UK mobility history offered by Goldthorpe in 2012:

‘The patterns of mobility that we have reported suggest that Britain became a more open society between 1972 and 1992 for men but then the picture changed. On the basis of the BES evidence, opportunities for upward mobility for men decreased in the mid-1990s. The changes in openness are demonstrated both by absolute and relative mobility measures. The gate to the service class was pushed wider open for a while, but then swung-to again.’

(Goldthorpe 2012: 8)

What Payne and Roberts do, though, is place the changing nature of the occupational structure as the key explanatory variable, and it is the detailed understanding of specific occupational changes that is important. Payne builds on this theory in his 2012 paper Labouring Under a Misapprehension. In it he presents a different way of measuring social mobility, which combines the measurement of absolute and relative mobility that he calls a ‘conditional disparity ratio’. He argues (p.40) that the probability of being in a particular class considerably affects one’s chance of leaving it. Again, the explanation that this begs is not completely dissimilar to what Goldthorpe has suggested before. Goldthorpe has explained the resilience of relative mobility in terms of differential decision making at particular points in the educational and employment trajectory by different class groups. Such an explanation could fit with a number of sociological explanations of differences in mobility, but it also remains in the realm of the broadly descriptive. Payne argues that the working class is decreasing in size, but there is a danger of those in it being left behind and experiencing frustration, alienation and presumably (although Payne’s 2012 paper does not mention this explicitly) financial hardship.

Searching for consensus

What is becoming clear is that while Goldthorpe and his colleagues may have exerted a great influence, there is certainly no complete or comfortable consensus among sociologists regarding social mobility. It makes sense to talk of a sociological viewpoint grounded in class analysis, biased towards the intergenerational and descriptive, but this does not imply that there is agreement among those examining social mobility. It is possible to identify research that describes relative mobility and absolute mobility as increasing, decreasing or constant, and it appears that all scenarios are a possibility, depending on the data analysed, the techniques used and the position on what mobility means taken by the researchers. However, none of those researching the field point to big changes, excepting the upward mobility into the service classes in the mid/late 20th century. Any change is gradual and takes time.

As Lambert et al argue (2007), the studies that look at longitudinal data as opposed to examining cohorts tend to point to increasing social mobility. They go on to perform a meta-analysis of 31 different contemporary social surveys going back to 1800, arguing that this shows that social mobility has increased, albeit slowly, over this period. The rate of increase appears relatively constant when comparing data from those born in the late 18th and early 19th centuries to data concerning those born in the 20th century. While the focus of much of the article is on the rates of social mobility per se, they recognise that it is important to try to differentiate between downward and upward mobility: ‘increasing magnitude of upward moves has fallen off over the period, whilst the downward mobility magnitudes have increased [this] suggests that younger cohorts tend to experience relatively more (p.41) long range downward mobility than long range upward mobility, in contrast to the experience of older cohorts’ (Lambert et al 2007: 14).

The legitimacy of downward mobility as a cause for concern may seem a non-issue from the political standpoint (or, you could argue, even from the non-sociological standpoint): it is clearly a bad thing if some people lose social status. However, from the pure meritocratic position, this is not a problem if it means the most able are rising to the top, which will actually strengthen society. There may also be other factors at play here, meaning that downward mobility is not portrayed as the problem in sociology that it is in politics. First, there is an almost inherent tendency in sociology to see advancement of the interests of those in lower socioeconomic groups as more important than any detrimental outcomes for those in higher positions. This tendency, rooted in the history of the discipline, will affect in some way what researchers say and how they look at an issue like this. And second, the desire of sociologists to position themselves as scientists reporting on phenomena in the social world as opposed to quasi policy makers can make them shy away from any detailed attempts to explore the merits of either downward or upward mobility. This does not stop socioligists criticising politicians for their lack of understanding of social mobility. Sociologists see the obsession with upward mobility as being to the detriment of a better understanding of the issue.

There are those who offer a rapprochement with the views of politicians, in contrast to the cynicism with which most of the other sociologists treat political interventions in the field. Li and Devine take what has to be a more productive approach, as they state in the conclusions to their 2011 paper ‘Is social mobility really declining?’.

[T]his preoccupation with upward social mobility facilitated by major structural change may be somewhat galling for academic sociologists who place a high premium on the distinction between absolute and relative mobility and are exasperated that most politicians and media commentators do not acknowledge that a genuine meritocracy involves downward mobility as well as upward mobility on the basis of merit … it is more constructive to recognise that politicians, policy makers, media pundits and, most importantly, voting parents have different preoccupations to academic researchers on social mobility. Upward mobility, rather than downward mobility, is the name of the game. The ongoing challenge for social scientists of whatever discipline is to understand how and why the debate on social mobility is conducted in the way that it is and to contribute to discussions of policy that improve the life chances of working-class men and women in the (p.42) lower echelons of the class structure in contemporary Britain. (Li and Devine 2011: 17)

Li and Devine do, however, revert to sociological type in the final sentence, and place a concern with how those from the lower classes can be helped to move up as the priority, as opposed to helping the middle classes who may move down. It could be argued that if social scientists are serious about recognising the different preoccupations of politicians and others, they need to confront the fact that the need to avoid downward mobility is at least of equal political importance to improving the life chances of the working classes.

Li and Devine’s paper makes another valuable contribution beyond their thoughtful conclusions. They look at cohort surveys from 1991 and 2005, comparing these to try to ascertain whether social mobility was improving or getting worse at these two points in time. They found that upward mobility for men had declined when comparing the two surveys, and downward mobility had increased, but the upward mobility still outweighed the downward mobility for men (40.6% in 1991 compared with 31.3% in 2005). However, while women’s upward mobility was increasing over the period, downward mobility remains slightly higher than upward mobility among women (36.8% in 1991 compared with 35.9% in 2005). Women, then, are more likely to be downwardly mobile than men.

Li and Devine also point to increasing relative mobility in line with what Lambert et al argue, but in contrast to Goldthorpe’s position. In particular, they find an increase in long-range upward mobility for men. So, while in one way this research may lend weight to the concerns of politicians, in another way it does not, as it shows that the rags to riches journey that politicians so readily associate with social mobility is not in decline but actually increasing.

Finally, Devine and Li also share Payne’s concern with the changing sizes of different groups, and show that men and women occupy different parts of the occupational structure. Men still dominate the positions in the top two socioeconomic groups, for example, and the proportion of all women in working-class positions is 40%. The percentage of men in such positions is 27%. This analysis shows clearly that social mobility is located in a different context for men and women.

The entry of the economists

The sociological analysis of social mobility in the UK has a history stretching back to the 1950s, while the economistic tradition is more recent. It is the contention of several sociologists, in particular (p.43) Goldthorpe (2012) and Saunders (2012b) that this is not a tradition as such. They argue that the claims regarding social mobility that economists have made are based on a limited amount of data, that these data are analysed incorrectly and that the use of income is inferior to occupation-based class categorisations as a basis for understanding stratification.

The research that has sparked the ire of the sociologists so much has been produced in the main by a group of economists from the Centre for Economic Performance at the London School of Economics and Political Science. Building on a growing international interest in intergenerational mobility research based on income and not on social class (some of which is discussed later), research began to emerge in the early 2000s which looked at mobility in the UK in this way (Blanden et al 2001, Blanden and Machin 2004). The main finding, based on comparisons of children born in 1958 and in 1970 using the National Child Development Study and the British Cohort Study respectively, was that ‘social mobility’ had declined. For those born later, in 1970, there is a stronger relationship with family incomes when they reach their early 30s than for those born in 1958. More specifically, Blanden and Machin (2004) found that 42% of sons born into the richest quartile of families in 1970 achieved incomes at this level themselves when they grew up, while just 11% ended up with incomes in the bottom quartile. However, for those born in 1958 the corresponding figures were 35% and 17%. Rather than the odds ratios, which define the relative mobility analysis of the sociologists, the economists use elasticity as their measure. On this basis the ‘coefficient of elasticity’ rose from 0.21 for the 1958 cohort to 0.29 for those born in 1970. There was, therefore, a slight increase in the association between family income and child income across the two cohorts.

In terms of explaining this change, Blanden et al (2007, 2011) point to widening income inequality in Britain in the 1970s, caused by children from higher income groups benefiting disproportionately from the increase in numbers going on to higher education. Further research undertaken by this group of economists shows that the earnings premium related to graduate study has remained significant despite the expansion in numbers, hence the argument that higher education expansion is actually driving a slowing down in social mobility (Blanden et al 2005).

Goldthorpe is rigorous in his critique of the position of this group of economists in Understanding – and Misunderstanding – Social Mobility in Britain (2012). In this critique (or more accurately, caveat, as it does not in itself render the work of Blanden, Machin and (p.44) colleagues invalid), he is correct to point to the limitations in the size of the body of work compared to the sociological studies. However, the strength of this research has to be based on the rigour to which these academics have subjected the data to analysis, rather than its existence within a wider set of complementary work that looks at different data sets over longer periods of time.

The more substantive methodological critique concerns missing data and the biases that might result, and the fragility of the measure of family income variable. Stephen Gorard (2008) takes the missing data criticism forward. The first thing that he reminds us, which is especially pertinent to the argument being developed here is that the economistic research does not measure ‘social’ mobility but ‘income’ mobility’: ‘A child who became a university lecturer, born to a father who worked as a fireman and earned as much in real terms as a lecturer, might be an example of social mobility based on education. But this would not show up as income mobility?’ (Gorard 2008: 320).

Gorard goes on to examine the actual number of cases analysed from the two cohort studies, and outlines that when a number of categories of respondent are removed, including, for example, the long-term unemployed, we are left with only around 2,000 cases in each cohort, which represents only just above 10% of the whole cohort for each one (Blanden et al 2005). Gorard goes on to argue that the majority of the difference in incomes of those across the two cohorts is actually attributable to factors other than parental income, and finally, that a detailed reading of the research by Blanden et al indicates that there is actually a high level of income mobility. As Gorard states:

… 17% of those born to the poorest 25% of families end up in the richest quadrant, and vice versa. If there were no financial inheritance, no inheritance of talent, no nepotism, and perfect social mobility then the maximum this figure could be is 25%. The difference from the ideal of perfect mobility in these tables containing 2,000 cases is represented by only about 25 cases in each of the ‘wrong’ extreme cells. Taken at face value, a key policy message could be that Britain has a quite staggering level of social mobility.

(Gorard 2008: 322)

Finally, not satisfied with deconstructing the actual research, Gorard turns his attention to the explanations. His issue with the argument that this alleged slowing down in social mobility was due to increased university attendance is quite simply that it does not square with the (p.45) evidence. An acceleration in university attendance occurred in the early 2000s, which would have been too late for the majority of those born in 1970.

These criticisms add up to a quite formidable set of caveats regarding the strength of the work by Blanden and colleagues. Gorard’s main point, also echoed by Goldthorpe and Saunders, is that this research is just not robust enough to justify a claim that social mobility in the UK has stalled. This is a crucial point to make when a whole political bandwagon has been built around it, which is influencing a whole range of policy areas.

Goldthorpe’s criticisms of this work go beyond caveats. He argues that the measures of income used are ‘one shot’ – they are taken from one point in the lives of correspondents in the survey, and may not reflect their income over their individual lifetimes. The specific problem is that the transitory component of family income in 1958 is higher than in 1970 (Erikson and Goldthorpe 2010). The picture in 1958, then, of family income is less reliable than that in 1970, which means the ability to compare cohorts is impaired significantly. Blanden et al respond to this in a 2011 paper where they build a model of permanent income to get over the ‘one shot’ problem. They then claim that current earnings are a good predictor of permanent income, and even if there are greater transitory errors in the 1958 research (which there may be some evidence of), it may reduce the fall in mobility but it does not imply that the data are not statistically significant. They also seek to address Goldthorpe’s assertions that class is a better predictor of income, and a superior mechanism for delineating social stratification.

This debate between the different academic groups will continue. Much is down to interpretation and belief in the relative superiority of different analytical techniques, but this aspect of it has a wider resonance. The issue of whether class or income per se is a better measure of stratification is a broader one, and while it will not be solved by this exchange, the separation of the two ideas as proposed by Blanden et al (as opposed to one being the product of the other) does have some appeal. They offer a conciliatory conclusion to their 2011 paper: ‘intergenerational income and social class mobility measure different things. Social class reflects wider job autonomy and wider social capital while income and earnings reflect economic opportunities. In this study we find limited common ground between the two approaches’ (Blanden et al 2011: 26).

The extent to which we can be satisfied with this conclusion is important. Accepting the significant caveats placed by Gorard, it is necessary to decide whether the area of income-based social mobility (p.46) analysis is in itself a separate thing from work based on social class. While it may be much less mature than its sociological cousin, should work be done to support its growth? The decision is actually already being made, as this work continues to be produced and, as outlined later, seems to have gained a significant foothold in the political psyche. Despite the imperfections in this work, looking at income as a factor in the stratification mosaic in its own right is valuable and essential.

The UK, the ‘sick man’ of social mobility

The political foothold that the work of Blanden et al has secured is also a consequence of how their research findings on UK income mobility sit in the international context. Comparing the UK to other countries, they have produced several pieces of work suggesting that the UK is not doing well, and is at best mediocre in this regard. Blanden (2009) examines a range of different studies across the world and produce a composite table that positions the UK as the second most immobile of 12 countries behind, notably, France, Canada and Scandinavian countries, with Brazil and the US above the UK. There are considerable caveats to allow for here, however. This is not the outcome of one distinct international cohort survey, but the product of assessing 12 different pieces of work with their own methodologies, delivered at different times.

In his 2012 pamphlet Social Mobility Delusions, Saunders is virulent in his criticisms of Blanden (2009). Despite acknowledging the limitations in the comparisons that can be made internationally where income mobility is concerned, Saunders feels that the Blanden paper still stretches too far in its claims, going on to say that, on all measures examined, the UK ‘tends to the immobile end of the spectrum’ (Saunders 2012: 20). The UK only appears in one of the measures – looking at the relationship between educational achievements and income and children, the UK appears to be performing around average rather than poorly.

Gorard (2008) also has something to say here. He points to the decision to compare Britain with others, using data from the 1970 Oxford Mobility Study work as opposed to the 1958 survey undertaken by Glass when several of the other countries used surveys from 1958 or closer to it. Given that the 1970 survey shows greater immobility in the UK, it is fair to assume that it will inevitably mean the UK looking as if it is doing worse than it would be doing if the 1958 results were used.

It is worth lingering on these points. They are more than just another disagreement between academics because, as both Saunders and Gorard (p.47) point out very well, this research is inherently politically appealing. In such globally competitive times, it compares the UK to its rivals and shows bad news, which always offers more political capital. Because of this, and possibly because (as Saunders argues) this research has benefited from influential sponsorship in the guise of the educational charity the Sutton Trust, the research by Blanden and colleagues over the mid to late 2000s has enjoyed much interest from policy makers.

The constant flux

There certainly appears to be a view among several of the sociologists listed here that the economists could have been more circumspect and considered in the conclusions they have drawn from their work. The sociologists (unsurprisingly) take a different view where comparative social mobility is concerned.

Goldthorpe and Erikson, in their 1992 book The Constant Flux, looked at data from nine different countries in the 1970s. They argue that absolute mobility varied between these countries but there was no trend as such to this, while relative mobility showed no such variation and was far more stable. The factors accounting for the changes in absolute mobility are related to the nature of the political regime and to differences in how countries adapt to external influences in the international political economy. In effect, Goldthorpe and Erikson (1992: 388) argue that cross-national variation in social mobility largely reflects ‘effects specific to particular societies at particular times’. The ‘constant flux’ idea supports the conclusions that Goldthorpe has reached over his career on social mobility in the UK. The comparative position of the UK where social mobility is concerned is less of an issue in this context, as there is much less difference between nations – or at least any differences that are subject to systematic explanation. Goldthorpe and Erikson were particularly concerned in The Constant Flux to confront the industrialisation ideas of Blau and Duncan (1967), and find conclusively against the idea that as societies become more economically developed, they become more fluid.

However, as with Goldthorpe and Erikson’s domestic conclusions, their views on international social mobility have not been accepted uncritically. In 2004 Breen and Luijx look at later data from the same countries as Goldthorpe and Erikson going up to the mid-1990s, but over a longer time span, utilising several different surveys and including information on women. Breen and Luijx argue that there is a tendency for increases in absolute social mobility over time as occupational structures change. While there may be big differences (p.48) between countries in these rates, they are systematic, and Breen and Luijx come out broadly in favour of the industrialisation thesis. They also give some weight to the role of politics. In countries where there is greater ‘direct political intervention of the kind associated with state socialist and social democratic societies’ (Breen and Luijx 2004: 401), mobility is higher.

Further work looking at data collected since The Constant Flux also supports the industrialisation thesis. Yaish and Andersen (2012) utilise survey data from 20 countries collected in the 1990s as part of the International Social Survey Programme (ISSP). Their focus is on the relative importance of individual and macro-level variables in explaining any differences in rates of intergenerational occupational mobility between countries. They also support the idea that greater economic development exerts a positive impact, but that levels of economic migration – and also having a communist past – have a role to play. In their conclusions they compare the relative contributions of individual-level influences, in particular educational attainment and political/historical differences. They argue that their model can account for almost 80% of the differences in absolute mobility between countries. Unfortunately, the UK does not feature in the survey that informed their work, but the interesting aspect of their study is this attempt to look at the individual and societal levels in explaining social mobility.

The debate over structure and agency – or the extent to which individuals have the ability to shape their own destinations relative to the frameworks within which they live – should be at the crux of understanding social mobility. The preponderance of sociological research, though, looks at measuring social mobility rather than explaining it.

The fact that the sociological research above is dated, that some of it does not include the UK, and that it does not enjoy the influential patronage of income-based measures of inequality all combines to mean that, while it may offer a strong rejoinder to the view that the UK is the ‘sick man’ of Europe (or the world), it has not taken hold like that of the economists.

Additional work from influential sources has supported this ‘sick man’ position. Miles Corak, a Canadian economist, has written a number of influential papers on international mobility (Corak 2006). He has produced what he describes as ‘the Great Gatsby Curve’, which shows the relationship between intergenerational elasticity in income and national income inequality, as measured by the Gini co-efficient for 12 of the richest countries in the world, which is reproduced in Figure 2.1. (p.49)

Social mobility: rising, falling or staying the same

Figure 2.1: The Great Gatsby Curve

Source: Corak (2012)

This diagram features in President Obama’s 2012 economic report (White House 2012).On both sides of the Atlantic it is the income-based analysis that has entered the political mainstream, shaping the view on social mobility, and in both cases portraying the UK and the US as problem cases. It shows there is a strong relationship between higher levels of inequality and lower levels of intergenerational economic mobility among fathers and sons. Predictably in social mobility analysis there are those who disagree with this link between inequality and immobility. Scott Winship (2012) from the Brookings Institute in the US, points to the fact that (as with the work of Blanden and colleagues discussed earlier) the graph is assembled from different surveys with contrasting methodologies, done at different times. Nevertheless, despite these criticisms, the power of the message that income-based social mobility in the US has stalled shows no sign of diminishing.

The momentum of this argument is less surprising when it is supported by analysis from heavyweight international research bodies such as the OECD. In 2010 the OECD examined the influence of parental background, as measured by educational achievement of parents, on wages and educational attainment of various cohorts of adults for 14 European OECD countries. They also examined the influence of parental background on the cognitive achievements of 15-year-old students in 30 OECD countries, and report a strong (p.50) link between parental and children’s earnings for the UK, that is low mobility. However, at the same time they recognise that:

… intergenerational social mobility is measured by several different indicators, since no single indicator provides a complete picture of social mobility and the United Kingdom seems to be less mobile in terms of wage persistence (measured by the summary indicator of wage persistence) than in terms of education persistence.

(Causa and Johannson 2010: 24)

This seems only to reaffirm the point that, as with studies of intergenerational social mobility that look at the UK alone, there is a lack of consensus where the extent of social mobility is concerned.

Is inequality a problem?

What unites these sociologists and economists is the view that social mobility reflects some form of structural inequality in society, and these inequalities are things to be combated. But there are alternative perspectives that reflect differences in natural ability and effort and, as such, are far less problematic.

Peter Saunders has been writing on social mobility since the 1990s (1997, 2010a, 2010b, 2012). In recent years, he has subjected the work of Blanden et al to heavy criticism, siding far more with the view that social mobility is relatively static over time, in keeping with Goldthorpe et al. Where he departs from the majority of the sociologists and economists whose work has been described so far, is in his explanations for the stability of social mobility. He argues that this is a function of differences in IQ, and that sociologists and politicians have ignored research showing the importance of IQ and hereditary factors because it does not fit with their beliefs. Differences in IQ are solidified by the increasing likelihood that couples will form among those of similar educational and occupational backgrounds. This leads to middle-class homes where children are both more intelligent and also work harder, as they are inculcated with a belief in the importance of education. As social mobility patterns are the result of genetics (which cannot be affected) and effort (which should not be affected), there is little that policy makers can or should do here. If there is a problem, according to Saunders, it is with an underclass whose ‘feckless’ behaviour is placing a burden on the rest of society and the dilution of the quality of elite (p.51) universities, which are being forced to admit students who are just not good enough to be there.

There is a voluminous critique of the use of IQ as a way of measuring or determining social outcome (Heckman 1995, Korenman and Winship 1995, Kinchelloe et al 1997, Breen and Goldthorpe 1999). It is hard to measure, and the relationship between IQ and future outcomes is fragile. Saunders, however, remains robust in his views. There remains too much doubt regarding the use of IQ and natural ability for it to be a form of external justification for differential outcomes. Saunders’ interpretation of the relationship between IQ and outcome also needs to take into account research showing the malleability of genetic predisposition (Lee Hotz 2011, Collins 2012). In a sense he does, when he talks about differences in effort and mating patterns between social groups, but accepting the existence of this malleability then places greater weight on nurture as opposed to nature. What Saunders’ work does do, though, is demonstrate the need to think through questions regarding the extent to which parental influence should be countered (or not) to produce greater social mobility and where lines should be drawn. Hence, what level of immobility is acceptable and what is not?

Swift (2004) offers one of the few serious philosophical interventions into the social mobility discourse. He considers what the actual goals for social mobility should be, and examines some of the assumptions that underpin the social mobility academic and political discourse but that are not adequately theorised. He asks whether perfect mobility, which he defines as ‘where a person ends up is random with respect to where she starts out’ (Swift 2004: 4), is both possible and even desirable. In his 2004 article he examines five different issues that those researching social mobility need to consider when thinking about the normative implications of their work. The most interesting of these issues is his analysis of where natural ability figures in social mobility and where the legitimate boundaries regarding family intervention in the life chances of children are concerned. What is the difference between an individual’s chances, depending on social background and on natural ability? Both are due to factors out of the control of the individual but one is seen as fair and one not. For example, few begrudge David Beckham his millions, based on his natural talent; compare this with the resentment felt towards many old Etonians. It is where the blurring of talent, background and effort occur that the difficulties emerge for social mobility.

Swift adopts the position that perfect mobility would not be ‘perfect’ because of the sacrifices it would demand. Where the transmission of advantage is concerned, Swift argues that the moral cost of expecting (p.52) parents not to speak to their children at dinner, for example, or not to read to them at bedtime would be too high (as well as impractical to police). Swift’s work identifies the limits to mobility and forces academics and policy makers to engage with more difficult terrain; hence it is in keeping with the goals of this book. It does not provide any answers as to where the boundaries between acceptable and unacceptable transmission of disadvantage are concerned – in terms of either talent or parental support. But his work does make the examination of these issues the legitimate business of social mobility analysis.


The academic study of social mobility is a growing field, drawing in researchers from different disciplines who are taking up a more diverse range of positions.

On one level there is little consensus about the discourse. Depending on who you believe and the definition of social mobility you adopt, the discourse has been increasing, decreasing or static.

However, on another level the discourse remains both consensual and also conservative. It is constructed within a conception of stratification that is based around narrowly economistic measures of welfare. In one way, this is entirely appropriate. Occupation and/or income exert a fundamental hold on the relative welfare of individuals and groups. However, there is little appetite for assessing the extent of this relative hold. The emphasis is predominantly on the measurement of rates of change in social mobility within schemas that are taken as read by their protagonists or criticised by their opponents on the grounds of their methodological weaknesses. For sociologists in particular, the importance of social class as the anchor around which the whole discipline is secured implies that more fundamental questions regarding what social mobility may actually mean – or how important it is – are largely untouched.

It takes the intervention of those operating mainly outside the field, such as Swift, for the issue of how much mobility is desirable to be debated. The inability or unwillingness to invite such questions, and the narrowness of focus around class or income alone, mean that while the existing approach to social mobility study does a comprehensive job in explaining part of the picture, it does not portray the whole of it.