CHAPTER    SEVEN
Outline of a Theory of Social Mobility*
It would be widely recognised among sociologists that the field of social i bility research is that in which quantitative techniques of data collection ;■ especially, of data analysis have reached their highest levels of sophisticat: However, such recognition could not be taken to imply approval. A serie articles by sociologists both within and outside the field could be cited ( see esp. Miller, 199S) in which it is charged that the concern with techni has become excessive and has had seriously detrimental consequences ii least two respects. On the one hand, it is held, the problems that are j sued by mobility researchers are to an undue degree chosen in the Ugh technique—that is, because they appear readily treatable via some favou procedure; while, on the other hand, the preoccupation with quanrira analysis and the results that it produces has led to a crude empiricism nn< an avoidance, or at least a disregard, of central theoretical issues.
As will be apparent from what I have written elsewhere (see, e.g., \r ch. 6), 1 can have some sympathy with this latter claim. Sociologists emiac in the quantitative analysis of social mobility, or indeed of other macre cial phenomena, have, I believe, often shown an insufficient apprecini of the importance of theory and, in particular, in failing to see that s analysis, no matter how sophisticated it may be, cannot itself substitute theory in providing explanations of the empirical findings that it protlu At the same time, though, I would regard the former claim—that the ra
This is an extended, reorganised, and largely rewritten version of the chapter u; same tide that appeared in the first edition of On Sociology. For help, advice, and i structive criticism, í am indebted to Richard Breen, Robert Erikson, David Firth, Míci Jackson, janne Jonsson, Ruud Luijkx, and Colin Mills.
154
Outline of a Theory of Social Mobility          155
problems treated in mobility research is unduly restricted by technical nsiderations—as being mistaken and indeed as betraying a lack of underlining of what has happened in the field. It is not difficult to show how ccĽSsive technical advances have permitted the more successful treatment problems, which for long antedated them (Goldthorpe, 2005). And, fur-■r, even if the researchers who have made and exploited these advances vľ often neglected the need for their work to be complemented by appro-iiite theory, this does not mean that the results of their analyses have been diout theoretical significance. To the contrary, the new findings that have llten produced have made it far clearer than before just what the focus of a theorv of social mobility should be—that is, just what such a theory should, and should not, seek to explain—and at the same time have revealed major problems with both the content and the form of the theory that has for long been of main influence in the field.
In this chapter I aim, first of all, to develop the above argument and then vo move on to outline a new theory of social mobility. In its general conception this theory is inspired by the findings that advances in quantitative annlvses have made possible, and it has, I believe, the potential to provide a fairly coherent explanation of these findings.
NEW  FINDINGS  AND  THEIR
GENERAL THEORETICAL  IMPLICATIONS
The major illustration of my claim that technical advances in mobility research have proved to be of major theoretical significance is provided by the application, from the mid-1970s onwards, of loglinear (and logmultiplica-tive) modelling and then of logistic regression techniques to the analysis of mobility data.1
Mobility researchers realised at an early stage that where, as is virtually always the case, the marginal distributions of standard mobility tables—that is. the distributions defining the 'origins' and 'destinations' of possible mobility trajectories—were not identical, some amount of mobility would for this reason alone necessarily be displayed: it would not be arithmetically possible for all cases in the table to fall on the main diagonal. An issue which then came to attract much attention was that of how this 'structural' (or 'forced') mobility might be differentiated from that which could be thought or as occurring, independently of any marginal discrepancies, in the form of
Outline of a Theory of Social Mobility
mutually offsetting instances of 'exchange' {or 'circulation') mobility. 1 efforts made to resolve this matter were in one way or another based on accounting identity of the form:
Total mobility — Structural mobility s= Exchange mobility
However, this approach did not lead to any very satisfactory outcor
It entailed an attempt at partitioning total mobility into two notional co
ponents that could be identified only at the supra-individual, or macrosoci
level, whereas the mobility table itself was a record of individual cases, N
until the introduction of loglinear modelling was the difficulty overcor
although not, it should be said, directly by the application of this techmo
but rather by a new conceptualisation that the application prompted (
Hauser et ah, 1975; Goldthorpe, Payne, and Llewellyn, 1978). Instead
distinguishing between structural and exchange mobility as two supposec
different components of total mobility, analysts using loglinear models wi
led to distinguish between absolute and relative mobility rates. The font
were the total rates and inflow and outflow rates that could be derived fro
the standard table by straightforward percentaging, while the latter we
expressed by the odds ratios that defined the pattern of net association i
origins and destinations within the table. (Odds ratios do of course consi
tute the basic elements of loglinear models.) Thus, in place of the identi
given above, it became possible to think of a set of relative rates in the for
of odds ratios, when embodied within given marginal distributions, as tht
implying a set of absolute rates:
Marginal distributions, Relative rates =*■ Absolute rates
By means of loglinear modelling, therefore, mobility tables could 3 analysed in a far more coherent way than hitherto. It became possible ' separate out the impact on absolute rates, or on changes or differences absolute rates, of marginal distributions or 'structural effects', on the oi hand, and of relative rates or 'fluidity effects', on the other.
Subsequently, logmultiplicative models were proposed (Erikson ar Goldthorpe, 1992; Xie, 1992) that allowed questions of the levels of flui ity implicit in mobility tables also to be addressed. More specifically, the models could provide tests of hypotheses to the effect that over a time šerí of mobility tables for a particular society relative rates were becoming mo equal and fluidity was thus increasing (i.e., the odds ratios defining sut
Outline of a Theory of Social Mobility        157
rates were moving generally closer to a value of 1, implying a complete independence of social origins and destinations) or were moving in the opposite direction; or, analogously, tests of hypotheses to the effect that mobility tables for different societies embodied relative rates that implied higher or lower fluidity from one society to another.2
Finally, in a further significant development, it has been shown how loglinear models for the grouped data of mobility tables can be rewritten as logistic regression models for individual-level data (Logan, 1983; Breen, 1994). In this way, it has become possible to move beyond the bivariate analysis of the association between social origins and destinations and to bring additional variables of interest—such as, say, individuals' educational attainment—into the analysis of relative mobility rates and patterns (e.g., Hendrickx and Ganzeboom, 1998; Breen and Goldthorpe, 1999,2001).3
The general theoretical implications of the results produced by these new techniques has become most apparent in the case of mŕergenerational rather than mŕragenerational, or worklife, mobility, and also, it might be added, where intergenerational mobility is studied within a class structural context rather than in the context of a hierarchy of, say, occupational prestige or socioeconomic status.4
The two central findings to emerge could be stated as follows. First, absolute rates of intergenerational class mobility display considerable variation, both over time within national societies and across these societies. But, second, this variation is to an overwhelming extent produced by structural rather than by fluidity effects—in other words, by differences in the ways in which class structures have evolved rather than by differences in underlying relative rates. Relative rates appear to be characterised by a rather surprising degree of invariance: that is, by a large measure of temporal stability and also by a substantial cross-national commonality at least in the general pattern of fluidity that they imply.
What follows, then, is that if variation in absolute mobility rates and patterns is to be explained, this will have to be primarily by reference to factors exogenous rather than endogenous to processes of class mobility themselves. In other words, the key factors will be those determining the 'shapes' of class structures, in the sense of the proportionate sizes and the rates of growth or decline of different classes, rather than those determining the propensities of individuals to retain or to change their positions within these structures. Moreover, it has also emerged from recent research that
15S         Outline of a Theory of Social Mobility
such exogenous factors are extremely diverse—demographic, economic, p, litical, and so on—and that they have interacted in the histories of particul; national societies in many different ways (see, e.g., Erikson and Goldthorp 1992: ch. 6 esp.; Miles and Vincent, eds., 1993). It thus becomes apparei why attempts that were previously made to develop theory in this regan as, say, by seeking to link variation in absolute rates to levels of econom development or modernisation or to types of political regime (e.g., Lipst and Zetterberg, 1956; Fox and Miller, 1965,1966), were not very successfu And in turn a fairly clear indication is given that insofar as variation in al solute rates cannot be usefully regarded as systematic, explanations of sue variation, whether over time or place, will need to be provided far more i specific historical than in general theoretical terms.5
in contrast, however, in the case of relative rates, the extent of the tempc ral constancy and cross-national commonality that have been displayed ca only be seen as posing an evident theoretical opportunity—and indeed cha lenge. While class structures themselves may often evolve in highly specifi ways, the patterns of relative mobility chances that prevail within them—c what have become known as 'endogenous mobility regimes'—would seer to be determined through processes that to a significant extent are system a tic and also context independent: that is, that operate in much the sam way over a wide range of societies.
In this regard, then, the importance of the new analytical technique is that they have served to reveal extensive social regularities of a largel unexpected and in fact quite opaque kind and, moreover, ones for which ai explanation grounded in theory with some claims to generality is evidentl required.
NEW  FINDINGS  AND  THE  FUNCTIONALIST THEORY  OF  SOCIAL  MOBILITY
At the same time as they indicate that theoretical efforts should focus on thi explanation of relative rather than absolute mobility rates, the new empiri ca! findings also strongly underline the need for a renewal of such efforts that is, by creating a series of difficulties for existing theory. This theor could be said to derive from a larger endeavour, in effect a general theory o industrialism, developed in the 1960s and 1970s chiefly by American soci oiogists of a liberal, anti-Marxist persuasion (cf. ch. 2, this volume). It wai
Outline of a Theory of Social Mobility         159
E
QÍ------------------------------------------d
-----------Weakening association
----------- Strengthening association
Figure 7.1.    Changing relations over time among class origins, educational attainment, and class destinations as expected under the functionalist theory
then elaborated in its application to social mobility by American specialists in the field, notably Blau and Duncan (1967: ch. 12) and Treiman (1970).
The theory is, implicitly if not explicitly, functionalist in character, and thus the explanation of mobility rates and patterns that it provides is one couched in terms of societal responses to functional 'exigencies' or imperatives1.6 In particular, emphasis falls on the way in which the dynamic technologies and economies of modern industrial, or postindustrial, societies impose requirements on their educational and employment systems that in turn carry direct implications for their mobility regimes. As regards relative mobility rates specifically, the three principal claims that are made by the theory can be set out, with reference to Figure 7.1, as follows.'
J. Technological and economic advance creates an inexorably rising demand for highly educated and qualified personnel and thus requires that human abilities or'resources' should be as efficiently utilised as possible wherever within the social structure they may happen be located. To meet this requirement, educational provision is expanded and educational institutions are reformed so as to further equality of educational opportunity and selection on the basis of merit. Thus, the association between individuals' class origins and the level of their educational attainment—the OE association in Figure 7.1—tends to weaken over time.
2.  Considerations of productive efficiency require that selection within labour markers and work organisations should also become 'meritocratic', that is, should reflect achievement rather than ascription; and in this regard educational attainment becomes the leading criterion of merit. Thus, rhe association between individuals7 level of educational attainment and the class positions, or destinations, that they eventually reach—the ED association in Figure 7.1 — tends to strengthen over time.
3.  In consequence of (1) and (2) together, education becomes increasingly dominant in mediating intergenerational class mobility and in turn relative
160        Outline of a Theory of Social Mobility
rates become more equal or, in other words, social fluidit)' increases. Controlling for education, the association between class origins and destinations moves towards zero and, as mediated through education, moves towards thi limit that is set by such genetic or cultural effects exerted by class origins on ability and motivation that education is unable to modify.
Up to the 1970s, this theory could be taken as providing a fairly coh ent way of understanding the evidence that had so far emerged from mot ity research. Or, at all events, the foregoing propositions, deriving from t theory, could not be regarded as being obviously inconsistent with this e dence. However, the far more refined, and also far more extensive, resear findings that have subsequently been produced are such as to call the thee into serious doubt in several different respects.s
The most immediate difficulty is created by the finding that across mc ern societies, as earlier noted, relative rates of intergenerational class mot ity show a high degree of temporal stability. Further, much of the change tt can be detected is not of a directional kind—that is, towards greater or h fluidity overall. The most that might be said in favour of the theory in this: gard is that so far as such directional change does show up, then this is mc often towards greater fluidity rather than less. In other words, the assoc tion between class origins and destinations weakens rather than strengthe; However, such shifts are observed in some societies but not in others a are, moreover, usually related to particular birth cohorts or are otherw episodic rather than sustained. In other words, they could scarcely cou as compelling evidence of a tendency towards greater social fluidity oft 'worldwide' and 'secular' kind that the functionalist theory would predk The functionalist theory also appears highly questionable in croi national perspective. Again as earlier noted, the endogenous mobility i gimes of modern societies reveal a notable degree of commonality in t general pattern of relative rates that they comprise. Sociologically as well statistically significant differences can certainly be demonstrated, includi in general levels of fluidity, but these would appear better regarded as 'van tions on a theme1 rather than as supporting any exceptionalist claims.10 Aj still more damagingly for the functionalist theory is the fact that more tec nologically or economically advanced societies do not tend to show high levels of social fluidity than those less advanced. Cross-sectional analys have found no evidence of an association between fluidity and various i dicators of development or modernisation, and a number of instances c;
Outline of a Theory of Social Mobility         J61
be cited, notably those of Japan (Ishida, 1995; Ishida and Miwa, 2005) and Israel (Yaish, 2000,2004a, 2004b), where relative mobility rates would appear to have remained essentially stable over periods of especially rapid and far-reaching economic and social transformation.
Finally, more recent analyses have examined the role of education in mobility processes, in terms of the OED triangle of Figure 7.1, and have aiso produced results that, overall, are scarcely compatible with the functionalist theory. Such analyses have considered, to begin with, the possibility of a 'global' decline (cf. ch. 2, n. 6} in class differentials in educational attainment, and in fact such a decline, even if often slight, has been detected in a number of societies, or, that is, the OE association weakens, as expected under the theory. However, when attention then turns to the relation between educational attainment and eventual class position—that is, to the ED association—the strengthening association that the theory would here predict is in fact rarely observed. Rather, the most common finding is that this association, too, is tending to weaken. And this is the case, it may be added, not only for countries characterised by a very high stability in relative mobility rates, such as Great Britain (Jackson, Goldthorpe, and Mills, 2005; Goldthorpe and Jackson, 2006) but also for countries in which some increase in fluidity has occurred, at least for a time, such as France (Vallet, 2004a) and Sweden (Jonsson, 1991, 1996, 2004). The indication then is of course that the mechanism producing this greater fluidity must be something other than the development of an education-based meritocracy in the way that the functionalist theory would envisage.11
One further finding of interest in this regard should also be noted: namely, that the OED triangle often comprises an interaction effect. Hout (19S8) first showed this finding for the United States, but it has since been found to occur far more widely (see, e.g., Guzzo, 2002). One possible interpretation of this effect is that the strength of the association between social origins and destinations varies, inversely, with individuals' level of educational attainment. Thus, Hout (19S8) observes that among American university graduates this association entirely disappears—while clearly still present among nongraduates. However, an alternative and, I would believe, preferable interpretation can be advanced: namely, that the association between education and class destination varies with class of origin, being weaker the more advantaged the origin.12 From this standpoint, as will later become apparent, the idea can then further be questioned that in modern societies it
162        Outline of a Theory of Social Mobility
is increasingly through an education-based meritocracy that class mobiljj is mediated.
Underlying the functionalist theory, two key assumptions can be nota The first is that, prior to the exigencies of modern technologies and ecoi omies exerting their effects, human abilities were systematically umíniv: ploited: that is, were often not allowed to develop into merit, and especial in the case of individuals of less advantaged social origins. The second a: sumption is that, under the pressure of these exigencies, both public policy i the field of education and the personnel policies of employing organisatior significantly change—indeed are forced to change—so as to enable merit t be more fully expressed and thus to play a steadily growing part in determir ing individuals' chances of mobility, whatever their social origins might be
To the extent, then, that the theory does now appear increasingly inac equate in the light of empirical findings, these assumptions must be treate as suspect. What is most obviously suggested is that the exigencies to whic the theory appeals, even supposing that they do exist, lack the transforrni tive power that is attributed to them. It is true that educational expansio and reform have everywhere been the concomitants of technological an economic advance. But what has not come about is the envisaged restruc turing of the chances of upward, and the risks of downward, mobility, a a result of educational systems creating greater variation rather than cor tinuity in the class positions of families over generations. Moreover, it is i this connection further important to note that, because of its very form, th functionalist theory does not at any point provide specific accounts of ju: how the exigencies that it postulates are expected to work out actually t the level of individual action and interaction. As in all functionalist theoriei the explanatory strategy is simply, as Coleman {1990: ch. 1) would put i 'macro-to-macro', and individual action, insofar as it is considered at all, i reduced to little more than epiphenomenal status.13
In attempting to develop a more successful theory of social mobility, . therefore abandon all functionalist assumptions and also take it as a prime requirement that the theory should have as explicit micro-foundations as can be provided. That is to say, it should aim to spell out how the regularities that are empirically demonstrable in relative mobility rates and in the nature and extent of their mediation via education result from central tendencies in courses of action that are followed by the different categories of actor involved. I opt furthermore, as in previous chapters, to proceed on the basis
Outline of a Theory of Social Mobility        163
0f rational action theory on the grounds that significant explanatory and interpretive advantage is gained insofar as action can be treated as rational, at all events in a subjective and bounded sense, rather than as being understood onIy in terms of actors1 internalisation of the values of particular cultures or subcultures or their commitment to social norms (cf. Boudon, 2003a, 2003b). This consideration would seem especially important in the present case in view of the evident need to construct a theory of a rather high level of generality.
I now to go on to outline a theory of the kind indicated that starts from the idea of 'mobility strategies': that is, courses of action that are pursued by individuals of differing class backgrounds, typically, though not necessarily, in conjunction with their families of origin, in moving towards their own eventual class destinations.1'1 However, before I attempt to characterise such strategies in detail and analyse their implications for relative mobility rates, I need first to take up the questions of the resources that are available for their pursuit and of the actual goals towards which they are directed.
MOBILITY  STRATEGIES:   RESOURCES  AND   GOALS
Resources When social mobility is examined within the context of a class structure, this structure has to be seen as conditioning rates and patterns of intergenerational mobility in two different ways. First, and as earlier implied, the shape, and changes in the shape, of the structure in the sense simply of the proportionate sizes of different classes will determine the extent and nature of what might be called 'objective' mobility opportunities. For example, a structure in which higher-level class positions are expanding relative to lower-level ones could be said to offer increasing opportunities for entry into the former positions, regardless of the class origins of the individuals who actually take up these opportunities. In fact, as I have emphasised, empirical results clearly show that both temporal and cross-national shifts in absolute rates of intergenerational class mobility do overwhelmingly reflect the evolution of class structures, with changes or variation in relative rates making only a very minor contribution.
Second, though, and more relevant for the present concern with relative rates, class structures can also be seen as conditioning mobility via the typi-
Outline of a Theory of Social Mobility
cal attributes of the positions they comprise. In this case, what is affectei not the overall situation as regards mobility opportunities but, precisely, relative chances of mobility that individuals of differing origins have wit the class structure, whatever shape or pattern of change it may display, t ferent classes, considered as classes of origin, provide varying degrees a forms of advantage to those individuals who are born and grow up wit) them. Or, one might say, from class to class, the resources that parents h; available to help support their children's mobility strategies, or stratég that they themselves conceive on behalf of their children, will vary in b< amount and kind, and such strategies will in this way be to a greater or tes extent facilitated or constrained. In short, the class structure not only crea! more or less favourable ground for the mobility stakes, it also plays a ma--part in determining the runners' handicaps.15
The nature and significance of this second way in which class struct!:-influences mobility can, í would argue, be best appreciated if class positio: are understood as being differentiated in terms of employment relations ( further chs. 5 and 6, this volume}. An initial distinction has thus to be ma among employers, self-employed workers, and employees. But more con quential, at least in the context of modern societies, are the further distil -. tions that can then be introduced within the numerically preponderant c •■ egory of employees in regard to the form of regulation of their employme
To recapitulate briefly the argument of Chapter 5, the major division ■■ be recognised is that between, on the one hand, a working class, compr. ing employees in broadly manual and lower-grade nonmanual occupatio and, on the other hand, a salariat, comprising professional and manager.: employees. The former are typically engaged by their employer or empli ■ ing organisation through a contract that implies a short-term and spec   . exchange of discrete amounts of labour in return for wages calculated o1-piece or time basis. The latter are typically engaged through a contract tl " implies a longer-term and more diffuse exchange in which the employee n : ders service to the employing organisation ín return for compensation tl as well as a salary and various perquisites also includes important prosp ■. tive elements—regular salary increments, some expectation of continuity   ' employment, and, above all, career opportunities. Further employee clas    ■ may in turn be distinguished that are intermediate between the worki    . class and the salariat in that the regulation of employment in their consti ': ent positions tends to be of a mixed form, in which elements of both i! -
Outline of a Theory of Social Mobility
Í65
ibour contract and the service relationship occur. These classes typically comprise routine nonmanual employees on the fringes, as it were, of professional and managerial bureaucracies and employees in lower-grade technical or manual supervisory positions.
From this standpoint, it is then, the salariat that appears capable of conferring the greatest degree of advantage as a class of origin, and the working class the least, specifically in respect of the resources that their members can command by virtue of the class positions that they hold. What is important here is not simply that salaried professionals and managers will have higher average earnings than rank-and-file wage workers. The former have also to be regarded as being advantaged over the latter, through the form of regulation of their employment, in at least three other highly consequential ways. First, they have greater economic security through their better protection against the risk of job loss and subsequent unemployment; second, they enjoy greater economic stabilit)' in that their incomes from employment will show less short-term fluctuation in relation to amount of work performed and will be less subject to interruption as a result of sickness or accident; and third, they have more favourable economic prospects in that their incomes will tend to follow a rising curve until a much later stage in the course of their working lives, peaking in their fifties rather than in their late twenties or thirties (see further Goldthorpe and McKnight, 2006). On all of these grounds, therefore, mobility strategies pursued from origins in the salariat are to be regarded as those least likely to be constrained by the availability of resources, while strategies pursued from working-class origins will be those most likely to be so constrained.
Matters become rather more complex when intermediate classes are considered, and especially when these classes are taken to include not just further employee classes, as referred to above, but also small employers and self-employed workers.16 The advantages that such intermediate classes offer as classes of origin are not in fact readily ordered but can be better understood as differing qualitatively. For example, small employers and self-employed workers will rypically be exposed to greater economic uncertainty and insecurity than members of intermediate employee classes but at the same time will have greater possibilities for the accumulation-—and thus intergenerational transmission—of capital in some form. Or again, while employees in manual supervisory and technical positions may have higher average earnings than employees in routine nonmanual work, their earnings
Outline of a Theory of Social Mobilit}'
will tend to be more variable and perhaps to show Jess long-term progj sion. Although, then, the mobility strategies of individuals of intermedi; class origins may be taken as somewhat less constrained than those o£ dividuals of working-class origins, the degree of constraint is likely to v greatly with the appropriateness of the kinds of resources available to particular strategies that are conceived.
Viewing relative mobility rates in a class structural context definec terms of employment relations is by now of proven value in empirical search. Such an approach would be widely accepted as in various respf more revealing than that in which mobility is treated simply within a q dimensional hierarchy of, say, prestige or socioeconomic status. Wl though, for present purposes is of added significance is the possibility t the approach may be given a more developed theoretical basis. That is, arguing, as í sought to do in Chapter 5, that, among the body of employ the association that exists between the kind of work they carry out and typical form of regulation of their employment is the outcome of decisi made by employers that have a clear rationale in considerations of orgŕ sational effectiveness. For, if this argument holds good—if the differentiat of class positions among employees can indeed be understood as deriv from employers' attempts to deal with such highly generalised organ tional problems as work monitoring and human asset specificity—then I in itself gives grounds for expecting that endogenous mobility regimes i tend towards uniformity. The shapes of class structures, determining overall distribution of mobility opportunities and ín turn conditioning solute mobility rates, may vary widely over time and place. But what mi be called the principle of differentiation of these structures that constrř individuals' mobility strategies and in turn conditions relative mobility ra could be thought to have a far more enduring character.
Goats
Mobility strategies pursued from different class origins will be backed b>: varying levels and kinds of resources. It has, though, further to be recognised that such strategies may differ in the goals towards which they are directed. Insofar as the question of individuals' goals has been previously taken up b) analysts of mobility processes, two contrasting views can be identified.
The first view—and it is this that would seem to be taken over in the functionalist theory—is that the goals that individuals pursue can be treated
Outline of a Theory of Social Mobility        167
- always and everywhere the same. On account of universal psychological impulses, defined as 'ego needs' or whatever, individuals aim to move, so i'ar as they can, from positions that are less desirable to positions that are more desirable in terms of the various rewards that they offer (see, e.g., Lip-.ijjt and Zetterberg, 1956; Kelley, Robinson, and Klein, 1981). The second view is that individuals' orientations towards social mobility, even within more advanced societies, show wide, subcuSturally determined, variation. In particular, it has been suggested, working-class subcultures may be inimical to the 'success ethic' that more generally prevails in such societies, either hecause these subcultures embody alternative values to that of individual ■ichievement—for example, values of family or community solidarity—or simply because they engender fatalism and a 'poverty of aspirations' (see, e.g., Hyman, 1954; Richardson, 1977; Willis, 1977).
However, it is possible to suggest an alternative approach to this question that is at least as compatible with the empirical evidence as is either or the foregoing views,1'' and that would appear to offer greater explanatory potential. This approach requires that mobility orientations should be uhought of as being basically similar across different social classes but at the same time as involving priorities, so that, given the differing degrees of constraint that are imposed by class origins, some systematic variation may indeed be observed in the actual goals that are pursued. More specifically, the suggestion is that what should be treated as common to individuals of all class backgrounds alike is a concern, in the first place, to maintain a class position that is no less desirable than that of their parents or, in other words, to avoid downward mobility. A concern to secure a more desirable class position, or, that is, to achieve upward mobility, is then to be regarded as a secondary objective, even if, perhaps, a still important one.!ii
If mobility is envisaged as occurring within a single, well-defined hierarchy of positions and if, further, educational attainment, understood in a linear, more-or-less, fashion is taken to be the crucial determinant of mobility chances, then the issue of priorities in mobility orientations may well appear of little importance. For in this case the same mobility strategy, that of maximising educational attainment, could be regarded as equally appropriate to improving the individual's chances both of avoiding downward and oi achieving upward mobility: more education is always better. However, if mobility is envisaged as occurring within the more complex context of a class structure, if other factors than education are accepted as significantly
'168        Outline of a Theory of Social Mobility
influencing mobility chances, and if educational systems are in any evei seen as diversified—that is, as providing options not just for more or le education but also for education of differing kinds—then a quite diľľerei perspective is gained. It is now at least conceivable that the strategie-, th, would best serve to ensure that individuals of a given class origin ma i n m their class positions intergenerationally need not be the same as those rh; would best serve to promote their chances of upward mobility. Or, in orlu words, in this perspective the possibility can be recognised that signifies choices of strategy may have to be made, implicitly if not explicitly, and oik that will entail differing degrees of risk. In pursuing one goal, the chances < realising another may be jeopardised.
Insofar as such situations do in fact arise, then, to repeat, the has assumption that I would propose, and that will underlie the analysis ih; follows, is that the avoidance of mobility downwards will tend to be give priority over the achievement of mobility upwards. Thus, while it will noi I supposed that individuals of working-class or other less advantaged bad grounds reject the goal of social advancement or are effectively preclude from pursuing it as a result of their subcultural conditioning, neither will be supposed, as in the functionalist theory, that an equal commitment 10 rh goal characterises all individuals.
MOBILITY  STRATEGIES,  EDUCATIONAL ATTAINMENT, AND  ASCRIPTION
If it is taken that individuals engage in mobility strategies, subject to coi straints and directed towards goals as outlined above, the general problei to be addressed is the following. How does it come about that this actio actually generates, in aggregate outcome, the degree of constancy and con monality in relative mobility rates and also the empirical regularities i the mediation of mobility through education that must, for any theory c. mobility, constitute the major explanandat Or, more specifically, one could ask: if it is indeed the case that much unexploited ability does exist among the members of less advantaged classes, why have educational expansion and reform and increased pressure for meritocratic selection not produced a more clear and consistent movement towards greater social fluidity, concomitantly with technological and economic advance? Or, again, why, gi\ en the educational systems of modern societies, have able children from less acl-
Oatline of a Theory of Social Mobility        169
ViJniaged origins not come to compete more successfully for more desirable clw positions, at the same time as less able children from more advantaged (jiíľ.1 ris more often end up in less desirable positions?
In pursuing the agenda thus set, it will be a useful and, I believe, a not [■no misleading simplification to regard mobility strategies as being of two main kinds: that is, strategies 'from below' and strategies 'from above1. The former are strategies pursued from less advantaged class origins, which, fol-lowi i ig the discussion of the preceding section, will be taken as origins in the working class or in the various classes, which, in terms of the typical employment relations of their members, were considered as intermediate. The latter are those strategies pursued from more advantaged class origins or, that is, from origins within the salariat of professional and managerial employees.
I will now consider how these two kinds of strategies operate, first, via educational attainment—or what is usually regarded as achievement—and, second, via processes of ascription. In so doing, my aim will be to provide theoretical accounts or narratives that can show the courses of action involved to be ones capable of generating the empirical regularities in relative mobility rates that require explanation, and at the same time to be rational, and thus intelligible, responses by individuals to the situations in which they find themselves. As these narratives are developed, I will consider evidence that is, or that, if produced, could be, relevant to assessing their validity.
Mobility Strategies and Educational Attainment
I earlier suggested that within the conceptual approach to mobility that I here adopt, the choice of mobility strategy may be problematic. In particular, the strategy that would best guarantee the avoidance of downward mobility need not be that which would give the best possibility of achieving upward mobility. What I would now further argue, essentially on lines already developed in Chapters 2—4, is that it is in the case of educational choice within strategies pursued from below that this difficulty is most marked.
There can be little question that a strategy for upward mobility from relatively disadvantaged class origins into the ranks of the salariat will be most effectively pursued—that is, with the greatest chances of success—via the route of relatively high educational attainment (see, e.g., Ishida, Müller, and Ridge, 1995; Müller and Shavit, 199S; Goldthorpe and Jackson, 2006), and it would, moreover, seem reasonable to assume that there is wide public awareness of this fact. However, it does not follow that for children of disad-
170        Outline of a Theory of Social Mobility
vantaged origins, attempting to maximise their educational attainmen always be the most effective route towards what I would take to be thei priority, that of ensuring that they at least maintain the class position of parents. Thus, for children of working-class origins, the safest option i regard—to ensure, say, that they remain within the skilled as opposed 1 nonskilled working class or in relatively continuous employment of kind as against the possibility of becoming long-term or recurrently u ployed—may well be that of leaving mainstream education relatively and of taking up vocational training, whether full-time or in conjun with some appropriate kind of employment.19 In addition, such a str could also give chances of short-range upward mobility into intermei class, especially manual supervisory or lower technical, positions. In trast, continuing in general education could involve some significant d of risk. In particular, if this option could not, for whatever reasons, bt sued through to the point at which relatively high level qualifications actually achieved, it could weil prove costly in terms of earnings fore and of other—less ambitious but more realisable—opportunities m Moreover, it has in this connection also to be recognised that pursuing strategies from below are likely to be inclined towards educat choices of a more conservative kind—that is, ones more relevant ti achievement of class stability than of decisive upward mobility—simp account of the economic constraints that their class backgrounds im Although reform programmes have in most modern societies removed substantially reduced the direct costs of education, at least up to secor ary level, the opportunity costs of remaining outside the labour market; still present. And, further, the successful passage of students through terth education is still much facilitated by, even if it does not actually demai parental economic support in some form or other—the alternative to wh-' is usually debt. One must then expect that the costs in question will of greater consequence for individuals, the less advantaged are the cl; backgrounds from which they come. Especially in the case of children frc working-class families, where income is relatively insecure, often subject significant fluctuation and tends to peak early in parents' working liv educational options that offer relatively short-term payoffs can be expeci to have attractions over ones from which the economic rewards, thou potentially larger, are longer delayed and carry a greater risk of not bei successfully carried through.
Outline of a Theory of Social Mobility        171
The problematic nature of educational choice in mobility strategies rom below is in fact highlighted if the comparison is made with strategies nipiLied from above: that is, by individuals of salariat background. These uli\ iduals, and their families of origin, face a situation, the logic of which ^ fairly straightforward. Given that intergenerational maintenance of class ■osirion is the prime goal to which their strategies should be directed, it is ir clearer than in the case of strategies from below that attainment within k educational mainstream will be the safest and potentially most effective :>me to follow.20 Thus, families within the salariat can be in general ex-ccR'd to give their offspring every encouragement and support to continue i full-time education beyond the minimum leaving age, to take up more c.ulemic options in secondary education, and to continue through to the .■rriary stage. In other words, their strategy will be that of translating their hildren's ability as far as ever is possible into actual educational attainment s represented by formal qualifications.
Moreover, within these families a strategy of this kind can usually be acked by more or less adequate resources. What is important in this regard ; not only the relatively high level and stability of their incomes but also ie fact that at the same time as children are passing through the crucial ícondary and tertiary stages of their educational careers, parental earning ower will typically be moving up towards its highest point. Thus, invest-lent in children's education is encouraged since the costs involved, direct and indirect, may well be absorbed without any seriously detrimental effect on established family living standards. However, even if some degree of sacrifice is involved, parents can still be expected to commit resources in various ways so as to improve their children's educational chances: for example, by buying homes in high status residential areas that provide good quality state schools, by supplementing state education with private tuition, or by opting out of the state system altogether and placing their children in private educational institutions.
It is, then, in these ways that children of more advantaged class backgrounds are given a clear competitive edge in seemingly meritocratic selection processes—or, as Halsey puts it (1977: 184) that'ascriptive forces find ways of expressing themselves as "achievement"'. Moreover, the differential availability of resources can be seen as underlying the tendency (as already noted in ch. 2) for children of salariat background to be pushed to the very limits of their academic ability, or even beyond, while working-class children
172         Outline of a Theory of Social Mobility
may decline educational options in which they would have good chi of success. For if strategies from above are threatened by academic lire, further resources can then often be deployed: for example, secor third attempts at relevant qualifications can be underwritten or altern courses or institutions explored. In other words, a range of possibilities for what in French discussion has been aptly referred to as récupératio contrast, for children pursuing strategies from below, and thus backe fewer resources, a greater degree of caution is to be expected. For fail u a relatively ambitious educational option could well leave them and families less well placed, in terms of opportunity costs and perhaps ■ than if they had never attempted it.
To the extent, then, that mobility strategies as outlined in the foreg can be regarded as central tendencies in action within modern societiei resistance to change that is shown by class differentials in educational at ment—or, to revert to Figure 7.1, the lack of any decisive weakening h OE association—is made comprehensible. And while further research the actual processes of educational choice within differing class conte: certainly needed, it can at all events be claimed (see further ch. 4 above) a mode! of choice that seeks to capture the key arguments I have depl appears more consistent with existing empirical findings than others have so far been proposed or implied.
In addition, evidence reviewed by A4iiller and Shavít (1998) is of ticular relevance as regards the tendency for those pursuing strategies . below to leave mainstream education relatively early, even when, perl they have good chances of successful continuation. In the light of rese findings on the transition from education to employment in 13 adva societies, these authors are led directly to question the idea that mc general—education is always better. They note that the completion of v cational courses does usually give individuals the best chance of enterij skilled rather than nonskilled manual work, and, further, that it is qui often the case that the probability of being found in unemployment is low for those with vocational qualifications than for those with more acaderr education of a comparable or even a somewhat higher level. In other wore insofar as class stability is prioritised, then for many children of less adva taged origins the choice of vocational over academic educational optio may well reflect a large degree of realism and rationality—far more in fa than is allowed for in culturalist explanations that invoke a poverty of asf
Outline of a Theory of Social Mobility
173
icicins or unduly short time horizons and an inability to 'defer gratification' .,»., Schneider and Lysgaard, 1953; Rosen, 1956).21
A.S was earlier remarked, the functionalist theory turns crucially on the lea that the 'release' of previously unexploited ability, through increasingly „jntocratic social selection in education, and thence in employment, will ľ[-ve as the main driving force behind steadily increasing social fluidity. 11 we ver, the main implication of the present analysis is that such a release isinot be expected to occur in the quite unrestrained way that the function-ist theory would envisage. Even supposing that educational systems do lerate on strictly meritocratic principles, the very nature of the class struc-ire—the basis on which class positions are differentiated—is still capable of h i biting the full exploitation of ability. That is, by creating circumstances which individuals pursuing mobility strategies from below can have good iUions, in the light of perceived costs and benefits, for not in fact attempting i use educational channels to the fullest extent that would be open to them id that their ability would warrant.
Mobility Strategies and Ascription
modern societies, educational attainment may be regarded as the single
ost important factor in determining class mobility chances—although, as
dicated above, it would be naive to equate such attainment with purely in-
vidual achievement, uninfluenced by the ascribed advantages or disadvan-
ges of differing class origins. Furthermore, it needs to be recognised that
cription may play a still more direct part in mobility strategies, whether
irsued from below or from above. As well as contributing to children's
lucational success, various resources associated with family background
ay serve to help children in gaining access to particular class positions
independently of their educational attainment. Where ascription is in this
way involved, its importance is of course likely to be greater in strategies
that are aimed at class stability rather than at major social advancement
(cf. Goux and Maurin, 1997). And, in turn, for those engaging in strategies
from below, the possibility of thus using ascriptive resources may itself be
reckoned as a further factor favouring the adoption of goals of a relatively
conservative kind.
This point is perhaps most readily brought out in the case of the children of small employers and proprietors, and self-employed workers. For these individuals, the maintenance of their class positions may well appear to be
Í 74        Outline of a Theory of Social Mobility
best guaranteed not through educational attainment, at all events be; some threshold level, but rather through their direct inheritance of hi businesses as going concerns or through the intergenerational transmi' of capital sufficient to enable them to start up enterprises of their own. T is in fact clear cross-national evidence to show, first, that the propensit intergenerational stability within the 'petty bourgeoisie' is relatively his: least for men (Robinson, 1984; Erikson and Goldthorpe, 1992), and, ond, that educational qualifications play very little, if any, part in sustai this stability (Ishida, Müller, and Ridge, 1995). —
In addition, not only economic resources but also cultural and s resources can be transferred from generation to generation, and this in various other cases of those starting out from relatively disadvanr origins encourage strategies that have class stability as their prime objet For example, traditions of family employment in particular skilled trad specialised branches of commerce may endow individuals brought up w these traditions with knowledge and capacities and with contacts and s-networks that are of greater value in maintaining class stability than q fixations obtained through the educational system.
For those pursuing strategies from above, however, the significant ascription has to be seen in a somewhat different light. As earlier noted surest means of maintaining class stability within the salariat is throu high level of educational attainment, and family resources of every kirn likely to be exploited in order to help children towards such attain n A more direct reliance on ascription is then to be expected chiefly \\ children's educational performance does not in fact reach the requisite ? dard: that is, as an alternative means of trying to secure their class stal or at very least to save them from any decisive downward movement.
Thus, children of salariat background may be able to benefit from ' nections' within the occupational and wider social milieux of their pai in order to find openings for employment that would not be avaílab those from other backgrounds and that their own educational qualificai would scarcely justify. There seems no reason to suppose that the req ments of modern societies for meritocratic selection in employment entirely eliminated such possibilities or indeed the straightforward practice of nepotism. However, what is probably of more general importance is the fact, quite overlooked in the functionalist theory, that various ascriptive attributes of children of more advantaged class origins—that is, attributes
Outline of a Theory of Social Mobility        175
chat derive directly from their upbringing in particular family and community' contexts—may themselves represent 'merit', at all events in the eyes of employers.
In the functionalist theory, employers have only a shadowy roSe. It is in effect assumed that what constitutes merit within the educational system— r ha t is, the achievement of formal qualifications—will likewise constitute merit for employers. But it is now becoming apparent that this assumption is far too simplistic. For example, analyses of job advertisements (Jackson, 2001; Jackson, Goldthorpe, and Mills, 2005) indicate that while employers ■;or their agents) do indeed attach high importance to formal qualifications in the case of professional and ancillary positions, for many other types of employment qualifications appear less important than a range of other desiderata. In particular, for positions in the rapidly growing personal services and sales sectors, and including positions at managerial level, employers would seem less concerned with the certification of primarily cognitive abilities than with broadly defined social skills and with various personal or lifestyle characteristics: for example, appearance, self-presentation, savoir faire, manners and accent, and in general 'looking good and sounding right' {Warhurst and Nickson, 2001). And indeed it can scarcely be found surprising that it is attributes such as these that should be regarded as chiefly relevant to successful careers in, say, the leisure, entertainment, or hospitality industries, or in high-value sales, customer services, or public relations.
The implication then is that certain positions remain available within the salariat, and are even increasing in number, that, while not requiring a very high level of formal qualification, do call for attributes with which children from more advantaged class backgrounds will tend to be endowed simply as a consequence of their upbringing. And research findings are now in fact emerging to show, on the one hand, that individuals holding managerial positions in services and sales do on average have lower levels of formal qualification than managers in other sectors as well as professionals (Jackson, Goldthorpe, and Mills, 2005) and, on the other hand, that children of salariat origins with only modest educational attainment do gravitate towards such positions (Jackson, 2004).
The account earlier given of mobility strategies pursued through education led to an explanation for the resistance to change that class differentials in educational attainment display or, in other words, for the failure of the OE association of Figure 7.1 to weaken as clearly and consistently as the
176        Outline of a Theory of Social Mobility
functionalist theory would require. The account now provided of tlie that ascriptive factors may play in mobility strategies points to an exp tion of two other empirical regularities earlier noted that likewise c difficulties for the theory: the tendency, at least over recent decades, fc ED association to weaken rather than to strengthen and the presence i overall OED association of an interaction effect.
As regards the ED association, the suggestion is that employers not automatically serve as the agents of an education-based men toe as is supposed in the functionalist theory. Rather, employers must be as having their own ideas of what represents merit—or at least produ value—in employees, which, not unreasonably, will vary from one rv employment to another.23 Thus, if employers believe that attributes thi-more ascribed than achieved are nonetheless those most relevant to ce kinds of work, it is these attributes that they will select for in hiring, re ing, and promoting employees. There is, then, no overriding reason the ED association should steadily strengthen. This is likely to happen insofar as those positions for which employers regard educational qua I tions as crucial are in expansion—as they probably were in many soc over, say, the middle decades of the twentieth century. Today, howevc most rapid growth in employment is often found in the services sector i economy where employers' requirements for high-level qualifications even in regard to managerial positions, would appear to be least stringc And thus, the actual weakening of the ED association that has recently observed in a number of societies could be the outcome of a simple coi sitional effect.25
As regards the interaction effect typically found in the OED rein ship, it would seem best to interpret this, I earlier argued, as indicating the strength of the association between educational attainment and destination varies with class of origin. Viewed in this way, the interactioi then be seen as resulting from the differing part played in mobility strat from below and from above by ascriptive factors, relative to educaf attainment, on the lines I have set out. Thus, in the case of the childre at least the sons, of the petty bourgeoisie, where the direct íntergenerat transmission of economic resources is an obvious possibility, it has alread been observed that education is of little importance in maintaining class si;: bifity and, in turn, that the ED association is especially weak (Ishida, Al Lille and Ridge, 1995). Similarly, one would expect this association to be weakc
Outline of a Theory of Social Mobility        1 //
nong individuals with origins in the salariat than among those originating i the working class or in other intermediate classes apart from the petty aiirgeoisie. For while, as suggested, the former can draw on various family isources to compensate for poor educational performance in their efforts to preserve their class position, the latter have few means of securing upward mobility other than through educational attainment. And again, empirical results would appear to confirm this expectation. Education is found to be of generally greater importance in mediating upward mobility into the salariat from relatively disadvantaged origins than it is in maintaining intergenerational stability within this class (Guzzo, 2002; Goldthorpe and Jackson, 2006).2f;
GENERAL  IMPLICATIONS   FOR  RELATIVE  MOBILITY  RATES
Exploiting the idea of typical mobility strategies from below and from above, I have suggested explanations for an OE association more persistent than the functionalist theory would allow; for the failure of the ED association to strengthen as the functionalist theory would expect; and further for the tendency for the latter association to vary in its strength with class origins. Insofar as these explanations hold good, then an explanation is in effect also provided for the widespread failure of the overall OD association to weaken in any substantial way—for the absence across modern societies of any general and sustained movement towards more equal relative mobility rates or greater social fluidit)'. Or, to be more precise, an explanation is provided for the absence of any such movement as the result o^the development of an education-based meritocracy on the lines indicated in Figure 7.1.
Class structures generate unequal resources among families differently located within them in ways that reflect the adaptation of employment relations to highly generalised problems of work organisation. Thus, the class structural constraints that bear on the mobility strategies in which individuals engage can be seen as in themselves making for temporal constancy and cross-national commonality in endogenous mobility regimes—and despite new opportunities that may be created by educational expansion and reform.
On the one hand, individuals of less advantaged class origins may have good reason not to seek to exploit these opportunities to the full—not to engage in supposedly meritocratic competition for more desirable class positions to the extent that the functionalist theory would envisage. Given the
Outline of a Theory of Social Mobility
class-linked constraints to which they and their families are subject, t may, rather, favour strategies aimed primarily at achieving class stab or only modest social advancement, and even when their ability woujc consistent with more ambitious educational and occupational aspiratíi On the other hand, and likewise with good reason, individuals of more vantaged origins largely do engage in such competition via the educatk system, while, however, being able to draw on family economic resource order to raise their chances of educational success to some extent indej dently of their ability or, in the event of failure, being able to exploit o' aspects of their family backgrounds so as still to avert any radical décla ment. Moreover, in the case of strategies from below a reliance on asc tion rather than achievement typically implies making a choice in fav of stability rather than mobility, but in the case of strategies from ab no such choice is required. Stability is the one goal to be pursued by me of achievement or ascription, so that the latter can in fact complement former or, if need be, serve as substitute for it.
The constraints imposed by the class structure, one might then say, inc rationally adaptive responses from the individuals subject to them—t mobility orientations and the related courses of action that they folio1 which serve to reinforce and to perpetuate the effects of these constraint: the mobility regime. To argue thus is not to underwrite the idea that am individuals oi less advantaged class origins culturally grounded resists exists to the idea of upward mobility. It does not rule out the possibility many such individuals, and especially ones of relatively high ability, will gage in competition for more desirable class positions; nor yet the possib that, at least where these positions are in expansion, they may do so \ increasing relative success.27 But what is implied is that any potential pact on the mobility regime that might in this way arise will tend alway be offset by individuals of more advantaged origins likewise exploiting the favourable conditions that prevail so as to improve their chances of preserving their class positions rather than experiencing downward mobility. Arid the degree of success that they thus achieve will of course itself operate as a further constraint on the extent to which those pursuing mobility strategies from below aimed at long-range upward mobility can in fact achieve their goals.2S In other words, the point that is underlined is that greater social fluidity entails individuals of less advantaged class origins making gains vis-a-vis individuals of more advantaged origins in the'positive' competition for
Outline of a Theory of Social Mobility        179
ore desirable class positions without losing out to a more or less similar
rtent in the 'negative1 competition to avoid less desirable positions. How-
•er, insofar as mobility strategies from below and from above do tend to
; pursued under the constraints and according to the rationales that I have
suggested, there would seem little basis for expecting greater social fluidity
to be produced simply in response, as it were, to economic and technological
advance.-11
CHANGE AND VARIATION IN RELATIVE RATES
It is the degree of temporal constancy and cross-national commonality in relative rates of intergenerational class mobility that poses the main theoretical challenge arising from mobility research and that in the foregoing has chiefly concerned me. In the field of social stratification, and indeed more generally, there is a danger that a preoccupation with change and variation, sometimes quite slight or indeed uncertain, diverts attention away from the more fundamental question of the source of the regularity by reference to which change and variation are identified (Lieberson, 1987: 99-107). However, as I have recognised, instances of shifts in relative mobility rates over time as well as cross-national differences can be demonstrated, and it is therefore pertinent to ask here to what extent the theory that T have outlined can apply in this regard also. I conclude by considering this question.
Two preliminary points need to be made. First, change or variation in relative rates need not imply greater or less social fluidity overall but only perhaps differences in the pattern as opposed to the level of fluidity. Differences of this kind, I would see as falling outside the scope of the theory, and, most probably, outside that of any other theory that has aspirations to generality, simply because of the extent to which they appear to be shaped by highly specific institutional or cultural influences (cf. Erikson and Gold-thorpe, 1992: chs. 3 and 5 esp.).3ü Second, it is important to recognise that change or variation in relative rates that does pertain to the general level of fluidity can come about in diverse ways, some of which the theory will again not cover. For example, largely contingent factors, such as may be associated with wars, frontier changes, mass population movements, and the like, can impact on fluidity (cf. Breen and Luijkx, 2004b); or again what has been called 'perverse fluidity' (Goldthorpe and Mills, 2004; and cf. Duncan, 1968) can be created intergenerationalSy within class structures, in the case,
Outline of a Theory of Social Mobility
say, of women or of ethnic minorities, as a result of unequal opportun! experienced in the course of working life.
However, since the theory that I have outlined sees the degree of inv. ance of relative rates of class mobility as being grounded in generic featn of the class structures of modern societies, one expectation regarding cha; or variation clearly follows: namely, that fluidity will be greater, the m that class-linked inequalities in resources, or the immediate outcomes such inequalities, are in some way modified. In other words, equality opportunity, in the sense of more equal relative rates of mobility, should more manifest within the class structure, the less the inequality of condil that derives from this structure.
In empirical research that is relevant to testing the theory in this resr: one approach taken is that of cross-national comparisons. Analyses h been made of the extent to which levels of fluidity across different socic are associated with the degree of inequality in their (personal or househc income distributions—this being the only generally available indicator class-linked differences in resources. An early study of 24 nations (Tyi Semyonov, and Hodge, 1979) produced results to suggest that such an sociation was in fact present: nations with more equal income distribute tended to show (for men) more equal relative rates of intergenerational cl mobility. Some later and more sophisticated analyses have confirmed t finding (e.g., Erikson and Goldthorpe, 1992: ch. 12) but others have (e.g., Breen and Luijkx, 2004b). However, there is general agreement r the approach in question is fraught with methodological problems and [ the significance of results derived from it, in whichever way they may po is difficult to evaluate.31
The main alternative and more promising approach is that which cuses on changes over time in particular nations where high-quality d and evident theoretical interest happen to coincide. Two such cases are rh of Hungary and Sweden, in both of which a good basis exists for analys relative rates of class mobility over periods in which significant change. ... class-linked inequalities occurred—and indeed, were largely brought about through political action.
In the Hungarian case a series of large-scale and well-designed mobility surveys have been carried out from 1973 through to 2000 and have been subject to much analysis (see esp. Andorka, 1990; Simkus et ah, 1990; Krik-son and Goldthorpe, 1992; Wong and Hauser, 1992; Szelényi, 1998; Róbert
Outline of a Theory of Social Mobility         1 8 1
ind Bukodi, 2004). What emerges of interest in the present connection is the ol lowing.
To begin with, in the period after the communist takeover in Hungary in [ 947 there are clear indications of some increase in social fluidity. In particular among men and women who were born in the interwar years and who entered employment during the 1940s and 1950s relative mobility chances became more equal. This shift can, moreover, be plausibly seen as being, at least in part, the outcome of various forms of state intervention that were directly aimed at 'destratification'. The old land-owning and capitalist classes were eliminated, and, in the context of a command economy, inequalities in family incomes and wealth were significantly reduced. Further, a policy of creating a 'people's intelligentsia' led to formal discrimination in educational selection in favour of the children of peasants and workers and against children from 'bourgeois' or other supposedly privileged backgrounds. At the same time, relatively strict relationships were established between educational credentials and type of employment. An increase in social fluidity thus appears to have been brought about through an improvement in the relative chances of upward mobility of the children of peasants and workers that went together with a decline in the capacity of the professional and managerial salariat to maintain its intergenerational stability.32
However, in the later years of communist rule, and especially from the time of the so-called normalisation, following on the uprising of 1956, the trend towards greater fluidity appears steadily to weaken. In this period, a new understanding between the regime and the intelligentsia led to the progressive abandonment of educational discrimination, and then under the 'goulash communism' of the 1980s income differentials were allowed to widen substantially. Finally, in the years after the ending of communist rule in 1989, which saw the full transition to a market economy, the trend towards greater equality in relative mobility rates disappears, and the latest evidence for the 1990s points to an actual reversal of this trend, at least for men, or, that is, to declining fluidity within the Hungarian class structure (Róbert and Bukodi, 2004).
For Sweden also it is possible to track changes in class mobility over several decades, in the course of which a concerted political effort was made to create a more egalitarian form of society, although in this case through electoral rather than authoritarian politics. From the mid-1930s through to the 1970s, the Swedish Social Democratic Party was more or less continuously
1S2        Outline of a Theory of Social Mobility
in power, and promoted the development of a distinctive political ec and of a comprehensive welfare state that together created a greater of equality in incomes, economic security, and living standards across than was found in most other modern societies. Evidence from n. national surveys does then indicate that over the period in questinr fluidity showed a steady increase (Erikson, 19S3; Jonsson, 1991). 1970s, Sweden could in fact claim not only a rather distinctive pnsi regard to the reduction of class-linked inequalities of condition bur li in regard to the degree of'openness' of its class structure—which more de tailed analysis showed to result chiefly from an unusually low propensity fa /»/mobility within all classes alike (Erikson and Goldthorpe, 1992).
However, further survey evidence then reveals that from the I97(]i onwards, as the Swedish model of economic and social policy losr its coherence, and its egalitarian impact was reduced, the trend towards greater fluidity clearly weakened {Jonsson and Mills, 1993; Jonsson, 1996). And although Sweden has still to be counted as one of the most fluid of modern societies, findings for the 1990s (Jonsson, 2004) would suggest thar, at least among men, relative mobility rates have now tended to restabilise.
Both the Hungarian and Swedish cases do thus provide evidence that is at all events consistent with the theory í have advanced. Changes in fluidity are associated with changes in inequalities of condition across classes in rhe way that would be expected. At the same time, though, these cases must also be seen as underlining the degree of resistance to change that endogenous mobility regimes are likely to offer. The increases in fluidity that occurred were not large but still required political intervention of a determined and systematic kind, and, moreover, they appear not to have been sustained once egalitarian policies were relaxed or lost their force. Further, the effectiveness of educational reform alone is again called into question, unless, perhaps, as in early communist Hungary, it entails a degree of class-related discrimination and of state intervention in training and recruitment policies thai would scarcely be acceptable in a liberal democratic context. For while in Sweden educational reform did indeed achieve its aim of reducing class differentials in educational attainment (cf. ch. 2, this volume), the effect of this in then raising fluidity remains uncertain, and especially in view of the weakening association between educational attainment and class destinations rhar was earlier noted (cf. Jonsson, 1991, 1996).
A final question that may be raised is that of whether endogenous mo-
Otitlhie of a Theory of Social Mobility
i S3
■ regimes would show a similarly strong resistance to change in the o/?-,oKjte direction to that so far chiefly considered: that is, to change towards l0re »»equal relative rates or reduced fluidity. The theory I have suggested Wies that this would not be the case, or at least not if this trend were .rompted by a widening of class-linked inequalities of condition. For indi-iduals pursuing mobility strategies from below would thus face yet greater onstraints and greater risks in engaging in competition for more desirable lass positions, while those pursuing strategies from above would be stiil more advantaged in the resources that they could use so as, by one route or another, to ensure their class stability.
One striking, if perhaps somewhat special, case that illustrates the argument here advanced is that of post-communist Russia. In the years of transition to a market economy, class inequalities in both incomes and degree of economic security and stability widened sharply (Gerber, 2002), and, over rhe same period, a marked decline in fluidity within the class structure can be demonstrated (Gerber and Hout, 2004). A key factor in this change would appear to be that class origins have come to exert an increasing influence on the worklife mobility chances of men and women already in the active population as well as on those of children moving through the educational system and into employment for the first time. Consequently, a strong 'period' effect shows up in contrast to the more usual tendency where fluidity increases for this to occur predominantly through gradual 'cohort replacement' effects.33 Turning to the western world, one may here too note instances of a reversal in trend in economic inequality. Over the long boom of the postwar decades most societies experienced decreasing inequality in incomes, unemployment was at historically low levels, and at the same time governments extended citizens' rights to social welfare—a process aptly referred to by Marshall (1947} as one of'class abatement'. Tt is, perhaps, against this background that one can best understand why insofar as directional changes can be detected in relative mobility rates, they are for the most part ones towards increasing fluidity, and also why such changes have shown up rather more frequently towards the end of the twentieth century than earlier—that is, as those birth cohorts that entered into employment during the long boom have progressively replaced older cohorts. However, since the mid-1970s trends in income distributions have become much more variable than before, in many societies the movement towards greater equality has been checked and in some, notably the United States and the UK, inequality has strongly
184         Outline of a Theory of Social Mobility
increased (Gottschalk and Smeeding, 1997; Atkinson, 1999). Over the sa period, double-digit unemployment rates have at times returned and rh. has been a general tendency for social welfare policy to become less el; redistributive in its aims.
Under the theory that I have outlined the expectation must then be tl at least in those societies in which economic inequality has widened m. sharply, some decline in fluidity will eventually show up, and especialiv birth cohorts entering the labour market since the 19S0s replace those w lived through the long boom. Research over the course of the next few t cades will determine whether or not this expectation is met, but there are fact already some suggestive findings to hand. Thus, for the United Stat Beller and Hout (2005) report preliminary results indicating that anióny men bom between 1970 and 1979 a turndown in fluidity has indeed occurred, and one that can be at least in part associated with widening economic inequality. For the UK clear evidence has been produced of reduced intergenerational income mobility among men and women entering the labour market in the 1980s (Blanden et al., 2004), and this would seem likelv to be paralleled in work on intergenerational class mobility, using the same data-sets, at least as regards the relative chances of long-range mobilitv.!-;
In this chapter I have argued that recent technical advances in social mobility research have not had the intellectually narrowing effect that various critics have claimed. I have sought to show how they have in fact led to empirical findings that have greatly clarified what the focus of theoretical effort in the field should now be. That is, to explain the degree of temporal constancy and cross-national commonality that are revealed in relative rates of intergenerational mobility, especially as viewed in a class structural context, and also the regularities observed in the mediation of such mobilitv In-educational attainment. The extent of change and variation in absolute rates can be shown overwhelmingly to reflect the differing ways in which the class structures of particular societies have evolved, and will therefore require explanation of a primarily historical kind in which the invocation of contingencies and singularities is likely to play a major role. In contrast, the degree of invariance over time and place displayed by relative rates, or endogenous mobility regimes, clearly calls for theoretical explanation—and at a level of generality that can transcend the specificities of particular temporal or societal contexts and that can also provide a basis for some understanding
Outline of a Theory of Social Mobility        1S5
"the more systematic aspects of such change and variation in relative rates ; can be demonstrated.
í have sought to make a start on the development of such an explanation id have opted for one that is grounded in a theory of action, in contrast i rhe functionalist grounding of the most important attempt previously ;ide to account for what were believed to be—often mistakenly, as it now 1[)Cars—the main trends and patterns of relative mobility rates observable modern societies. I do not intend at this point to recapitulate my outline leory but simply to make two final remarks for the benefit of potential crit-s and of those who might try to take the theory further (these categories ;ing, I would hope, largely overlapping).
First, 1 have attempted to preserve coherence by relying throughout on rational action theory. It might at various points have been possible to give my argument at least a more immediate plausibility by shifting ground and treating action as, say, being ultimately shaped by individuals' adherence m cultural values or social norms. But this would in fact have been of little value if no more than ad hoc modifications were involved. I would therefore invite anyone inclined to improve the theory through such eclecticism to see their theoretical task as then extending to justification for so doing.
Second, while I would regard the existing empirical evidence as being generally supportive of both the direction that the theory takes and of expectations that in turn follow from it, I recognise that others might, in certain respects, wish to disagree. And it would, moreover, be only reasonable to suppose that either the reevaluation of old findings or the production of new ones will, sooner or later, create problems for the theory, of one kind or another, that cannot be denied. But I would then urge that theoretical efforts should nonetheless continue, whether on similar lines or different ones, in order to provide explanations of the highly significant results that social mobility research has produced and will, I believe, continue to produce, rather than the field being allowed to lapse into a condition of more or less unleavened, even if increasingly sophisticated, empiricism—for which ii could indeed be rightly criticised.