Rent, Class Conflict, and Class Structure: A Commentary on Sørensen1
John H. Goldthorpe
Nuffield College, Oxford University
I much appreciate Aage Sørensen’s article, “Toward a Sounder Basis for
Class Analysis.” I am entirely in sympathy with his argument that, follow-
ing on the effective collapse of the Marxist tradition, class analysis needs
to be informed by better theory. I also agree that attempts to develop
theory on supposedly “Weberian” or “neo-Weberian” lines have not been
2
impressive. More specifically, I find helpful Sørensen’s attempt to “bring
some order” into class concepts by differentiating them according to the “level of theoretical ambition” with which they are associated (pp. 1525– 26). This attempt does in fact provide me with the starting point for my commentary on his article.
Sørensen distinguishes three main types of class concept, hierarchically ordered, that I would myself, with only minor glossing of his own formula- tions, characterize as follows:
- Classconceptsthatprovideapurelynominalcategorizationofpopu- lations, according to a particular “dimension,” or possibly a combina- tion of dimensions, of stratification and that can be used to display class inequalities in life conditions and life chances or class differ- ences in attitudes, beliefs, values, and patterns of action
- Class concepts that go beyond point 1 above in aiming to delineate sets of class positions among which individuals or families within a population are distributed and may thus come to form collectivities with a greater or lesser degree of demographic, subcultural, and so- cial identity
- Class concepts that go beyond 1 and 2 in aiming to specify actual
1 I am indebted to Robert Erikson for helpful conversation and for comments on an earlier draft. Direct correspondence to John Goldthorpe, Nuffield College, Oxford University, Oxford OX1 1NF, England.
2 My own approach to class analysis has often been characterised as Weberian, but I would not myself regard this labeling as being very informative. I would in fact share Sørensen’s view (p. 1527, n.3) that the importance given to Weber in the literature on class analysis is “a bit curious.”
2000 by The University of Chicago. All rights reserved. 0002-9602/2000/10506-0003$02.50
1572 AJS Volume 105 Number 6 (May 2000): 1572–82
Symposium: Goldthorpe
collectivities within populations whose members have the motiva- tion to engage in conflict with members of other comparable collec- tivities in consequence of their interests being structurally opposed; that is, opposed by virtue of the class positions that they hold
Thus, those sociologists deploying type 2 concepts tend to be concerned not just with revealing class inequalities or class differences but further with explaining the conditions under which varying degrees of class for- mation occur and, in turn, with giving accounts of what Sørensen nicely calls (p. 1526) “the geography of social structure.” And those deploying type 3 concepts seek yet further to explain not just how inequalities and differences arise between classes but also antagonistic interests that create the potential for conflict.
In his article, Sørensen clearly aims to provide “a sounder basis for class analysis” by developing type 3 concepts and pursuing the theoretical ambitions that they are devised to serve. But, if I understand him cor- rectly, he would—consistent with the hierarchical nature of his typol- ogy—also wish the theory that he outlines to be one that can contribute to our understanding of the geography of social structure and to the analy- sis of the nature and extent of class inequalities and differences. At all events, I would myself want to argue that any well-constructed theory of social class should in fact have the potential to operate effectively at all three of the levels of theoretical ambition that Sørensen specifies.
My own work in this connection began with type 2 class concepts and related theoretical concerns that arose out of a dissatisfaction with type 1 concepts (in particular, with that of “socioeconomic status,” as used in much American sociology). Together with a number of colleagues, I elabo- rated a “class schema,” for use in empirical research, which derived from the idea that in modern societies class positions are best understood as being defined by differences in employment relations (see esp. Erikson
3
and Goldthorpe 1992; Goldthorpe 1997a). However, as Sørensen notes
(p. 1527, n.2), I have more recently—encouraged by results from studies of the criterion as well as the construct validity of the class schema (Evans 1992, 1996; Evans and Mills 1998)4—moved on to try to explain just why a
3 In consequence of its rather complex genesis, the schema is known by several differ- ent names. In a British context, it is usually referred to as the Goldthorpe schema; in an international context, as the EGP (Erikson-Goldthorpe-Portocarero) schema, the Erikson-Goldthorpe schema, or the CASMIN schema, after its use in the Comparative Analysis of Social Mobility in Industrial Nations project.
4 I refer here to studies that show that not only does the class schema perform well in displaying theoretically expected class inequalities and differences but further that, as operationalized through information on occupation and employment status, it does
1573
class structure comprising positions differentiated in terms of employment relations should be an apparently abiding feature of modern societies and give rise to persisting class inequalities and differences and further, sub- ject to certain conditions and limits, to various expressions of class conflict
5
the adequacy of Sørensen’s proposals to serve as the basis of class theory of a “comprehensive” kind in the sense I have indicated, and I seek to develop my arguments by reference to both similarities and differences between Sørensen’s approach to class analysis and my own. Finally, I note how our two approaches lead to contrasting expectations (all else being equal) regarding the future of the class structures of modern capital- ist societies, so that the possibility exists of some eventual empirical assess- ment of their respective merits.
My first reservation concerns the conceptual viability, from the stand- point of a comprehensive class theory, of taking what Sørensen calls the “rent distribution” within a society (as distinct from mere class inequalities in “total wealth”) as being the source of opposing class interests and in turn of the formation of antagonistic classes—exploitation classes—that engage in collective action directed toward the creation and preservation
6
or the destruction of rents. I would not wish to deny that action of the
kind in question is widespread in modern societies, and I would indeed agree with Sørensen that sociologists should pay more attention to it than they do. However, what I would still query is whether it is helpful either
in fact capture in a generally satisfactory way the differences in employment relations that it is conceptually supposed to capture.
5 This chapter in On Sociology is a revised version of Goldthorpe (1997b) as cited by Sørensen (n. 2). I would add to his note that as well as drawing on transaction cost economics I draw to at least an equal extent on recent work in organizational and personnel economics. My efforts in developing the theoretical basis of the class schema were stimulated by the fact that, following reports of a committee of the British Eco- nomic and Social Research Council, set up on behalf of the Office of National Statistics (Rose and O’Reilly 1997, 1998), the schema was accepted as the basis for a new official social classification for the United Kingdom, which has now replaced the Registrar General’s Social Classes, in operation since 1911.
6 I am puzzled by Sørensen’s use of the concept of exploitation—a word that I would myself gladly see disappear from the sociological lexicon—since it is not clear that it plays any vital role in his arguments. Its function in Marxist thought was to allow a fusion of normative and positive claims in a way that I would find unacceptable, as also, I believe, would Sørensen. For example, it is evident that he would regard the elimination of rents, and thus of exploitation, from labor markets as tending often to have highly negative human consequences. If invoking exploitation is no more than a way of flagging the presence of structurally opposed class interests that lead to zero- sum conflicts, then its use is innocuous but scarcely necessary.
(Goldthorpe 2000, chap. 10).
In what follows, I set out two main reservations that I have regarding
1574
Symposium: Goldthorpe
to regard all class conflict as being conflict over rents or to regard all conflict over rents as being conflict between classes.
As regards the first proposition, my difficulties arise from the (very broad) definition of rent that Sørensen adopts: rent, it would appear, is any return to an asset that exceeds the return that would be gained in an economy operating under conditions of perfect competition—as these would be defined in a standard economics textbook. The perfectly compet- itive economy is Sørensen’s baseline for determining whether rents exist. There are, then, two important implications of Sørensen’s position, and ones which—I think—he is ready to accept. First, since rents mark, as it were, deviations from the functioning of a perfectly competitive economy, they must lead to a shortfall from the standard of efficiency that such an economy sets. And second, since all class conflict is conflict over rent, it must be the case that in a perfectly competitive economy no class conflict could arise.
However, the perfectly competitive economy is an abstraction. Outside of economics textbooks, economies do not exist and function in vacuo but as “embedded” in complex institutional contexts. And in modern capitalist societies, I would then argue, it is in fact over the particular form that such supporting institutions should take that class conflict very largely occurs: most prominently, perhaps, in regard to social welfare rights and provisions and the manner of their financing, as these bear upon the de- gree of class inequalities of both condition and opportunity.
Sørensen would, of course, wish to treat all such conflicts over welfare rights and such as being conflicts over rents. But an important issue then is that of how far welfare and other institutions, and the rents (in Sørensen’s sense) associated with them, can be understood simply as rep- resenting so many deviations from the principles of a perfectly competitive economy and therefore as sources of economic inefficiency—or have, rather, to be seen as significantly contributing to the viability of any real world economy that seeks to approximate the competitive ideal. Most so- cial democratic interpretations of the “institutional” (as opposed to the “residual” or “marginal”) welfare state would see this as an achievement of the “democratic translation” of the class struggle that in fact promotes more than it undermines economic efficiency. An advanced welfare state serves not so much as a “leaky bucket” of rent redistribution but more as a vital “irrigation system” that prevents the waste of human capital and the underutilization of productive resources in general (Korpi 1983, 1985; see also Rieger and Leibfried 1998). Unless, then, Sørensen wishes totally to reject such interpretations—and he makes no evident effort in this di- rection—his restriction of class conflict to conflict over rents would seem somewhat artificial and, in so far as the creation rather than the destruc-
1575
tion of rents can be taken as serving efficiency, risks losing theoretical
7
tween classes—I find yet more problematic in its implications. As Sørensen makes very clear, conflicts over rent, in his sense, are highly heterogeneous. They are such that they can readily arise among capitalists and among workers, even workers of similar grade, as well as between capitalists and workers. Thus, capitalists (or their agents) may engage in conflict over the securing or terminating of monopolies, licenses, or patents that afford rents; likewise, workers may engage in conflict over the rents that can be obtained from organization or from involvement in firm-inter- nal labor markets or from minimum-wage legislation.
Now Sørensen would presumably wish to argue that we should be ready to break with our established ways of thinking and to treat such conflicts as indeed being class conflicts, no less than those that divide capi- talists and workers; he is of course fully entitled to propose such a concep- tual innovation. However, it is then difficult to see how a theory of class could be built on this basis that would also be adequate to the other con- cerns of class analysis previously noted: that is, in regard to class inequali- ties and differences and the formation or decomposition of classes as fea- tures of the geography of social structure. Rather obviously, the lines of class conflict over rents that would be drawn among capitalists (and their agents) and among workers would not, in the same way as those drawn between capitalists and workers, map onto the main lines of class inequali- ties (including those in “total wealth”) or of class differences that can be empirically established, more or less clearly, on the basis of any existing type 1 or type 2 concepts. Would it be possible, following Sørensen’s ap- proach to class analysis, to arrive at type 1 or type 2 concepts that would in fact be serviceable in empirical research?
Moreover, the heterogeneity of rents and of the conflicts to which they give rise would serve to inhibit the degree of formation of Sørensen’s “ex- ploitation classes,” as in turn would the degree of shifting and cross-cut- ting of the lines of conflict that would surely be found. Many individuals would be exploited in one context, but act as exploiters in another. Thus, within modern societies at least, it would seem unlikely that conflict over rents could ever become conflict on a “totalizing,” societal scale. Sørensen
7 These considerations lead me to suspect that Sørensen might have been better served by a somewhat more restricted understanding of rent. The history of the concept in economic analysis is indeed one of progressive generalization. But a comment of Schumpeter (1954, p. 679) on an earlier stage of this process is of interest: “There are generalizations that spell additional success for a theory: they enrich and extend, but do not endanger, its original interval of application. But there are others that spell or foreshadow the break-up of the theory.”
coherence.
The second proposition—that all conflicts over rent are conflicts be-
1576
Symposium: Goldthorpe
does in fact appear to recognize this, at least as regards conflict that has the potential for radical social transformation. For he remarks (p. 1541) that “class conflict should be more prevalent under feudalism than capital- ism” and then goes on directly to observe, in support of this view, that “no revolution has occurred in an advanced industrial society.”
However, for non-Marxists at least, it would not appear either neces- sary or desirable to envisage all class conflict as being directed toward revolution. I would myself regard the occurrence of certain forms of class conflict as being entirely compatible with the acceptance by all parties involved of the institutions of liberal democracy; or, in other words, I would recognize the possibility—and, in some historical instances, the actuality—of the “democratic translation” of the class struggle, as previ- ously referred to. At the same time, though, I would still wish to preserve the conventional distinction that Sørensen’s notion of “exploitation classes” opposing each other over rents would seem necessarily to blur: that is, between class conflict that through the solidaristic transcendence of special interests can possibly find expression at a societal level and “in-
8
terest-group” conflict that remains in its nature sectional and localized. My second reservation about Sørensen’s proposed “sounder basis for class analysis” concerns the reluctance he shows—along with most neo- classical economists—to accepting the distinctiveness of the labor market and, more specifically, of the employment contract. For Sørensen, “em- ployment rents” are not essentially different from other rents. They arise insofar as a labor market is in some way “closed” rather than being “open,” that is, perfectly competitive. Such closure may result from the restrictive practices of unions or professional associations or from various kinds of governmental intervention; but it may also result directly from employers entering into certain kinds of employment contracts, chiefly ones that re- flect the bargaining advantages that workers enjoy via problems of work monitoring or human asset specificity. Employment rents then lead, just like all other rents, to conflict between classes (in Sørensen’s sense) that could only be resolved by the destruction of such rents, whatever their origin might be, and by labor markets becoming open in all respects: that is, becoming ones in which there is no restriction on supply, no elements of subsidy and so forth, and in which all employment contracts take on a “rent-free” form. This would then, so far as I can see (though Sørensen is not himself explicit on the point), imply straightforward “spot” contracts
8 On the basis of this distinction, there has of course been much discussion, especially in regard to the working class, of the way in which interest-group activity, including activity that could clearly be regarded as “rent seeking,” inhibits class formation and reduces the possibility of effective strategic action being taken at a societal level, through class organizations.
1577
in which so much labor, treated simply as a commodity, is exchanged for so much money, with renewal of the contract being entirely contingent. All other forms of employment contract, in their propensity to give rise to rents, would be inefficient: that is, they would fall short of the standard set by perfect competition.
In the approach to class analysis that I would myself follow, recognition of the distinctiveness of the employment contract has, in contrast, a crucial role. Because labor cannot be physically separated from the persons who provide it, what is bought and sold through this contract, I would argue, is not a commodity, in the sense of some definite, objective thing, but rather a social relationship. Employment contracts are contracts through which employees agree, in return for remuneration in some form, to place themselves under the authority of an employer or of the employer’s agents. A consequences of this, I would then hold, is that spot contracts need not be, and often in fact are not, the most efficient form of contract.
The initial lines of demarcation within the class schema that my col- leagues and I have developed are those drawn—following the general idea of class positions being defined by employment relations—between employers, self-employed workers, and employees. But the numerically preponderant category of employees is further differentiated according to the kinds of contracts on the basis of which their work is regulated. At one pole are those employees, most characteristically unskilled manual and entirely routine grades of nonmanual worker, whose work is regulated by the basic “labor contract” that in fact comes as close as is possible to a spot contract in the sense previously referred to. At the other pole, how- ever, are those employees, most characteristically professionals, adminis- trators, and managers, whose work is regulated by a contract of a signifi- cantly different kind establishing a “service relationship.” While the basic labor contract entails a specific wage-based, “money-for-effort” bargain, the service relationship entails a more diffuse and longer-term exchange in which, in return for service to their employing organization, employees receive “compensation” that comprises important prospective elements re- lating to salary increments, promotion, employment security, and so on. Modified versions of the basic labor contract and service relationship are also recognized, as are “mixed” forms of contract, and these can likewise be associated with different types of work (Goldthorpe 2000; see chap. 10 for a fuller exposition).
I would then see this variation in the form of employment contracts as providing the main basis for explaining why the class schema in general performs well in displaying theoretically expected inequalities and differ- ences—that is, the class schema can claim good construct validity. Mem- bers of the classes it distinguishes, as well as having differing sources and levels of income, also have differing degrees of stability of both income
1578
Symposium: Goldthorpe
and employment and differing expectations as to their economic futures that together condition both their life chances and many aspects of their attitudes and patterns of action. Moreover, the class schema serves well to display the rates and patterns of inter- and intragenerational class mo- bility that are crucial to processes of class formation (or decomposition), at least at a “demographic” level and, further, the theoretical basis of its construction provides at least a starting point for the explanation of these
9
in which my view of the significance of variation in employment contracts and that taken by Sørensen diverge. As noted, Sørensen would apparently regard all employment contracts that fail to approximate spot contracts— chiefly because of employers’ problems over work monitoring and human asset specificity—as being contracts that give rise to rents and in turn to conflicts of a zero-sum kind (see, e.g., p. 1544). In contrast, I would wish to understand such contracts as being ones that are designed and imple- mented specifically to overcome problems of the kind in question—prob- lems that inevitably occur with many types of work—and, furthermore, that can overcome them in ways that contribute to organizational effec- tiveness and increase the total value of the contract. Thus, while in the case, say, of contracts that set up a service relationship there may well be Marshallian “higgling and bargaining” over their particular terms, the general form of such contracts offers advantages to employer and employ- ees alike. In turn, then, I would again wish to question the association that Sørensen supposes between rent and inefficiency. From the standpoint of organizational effectiveness, it is in fact the basic labor contract that I would see as being the form of regulation of employment with the most limited range of application: that is, to types of work where little more than labor in its elementary physical sense is involved and employers can thus come as close as is actually possible to “commodification.”
This divergence of view, it should be observed, does then lead on to issues that are of more interest and importance than mere disputes over conceptual preferences (with which, I am uncomfortably aware, much of the foregoing may appear to have been concerned). Sørensen’s contribu- tion has the great merit, in contrast to much other writing on class analy- sis, of moving from nominal to real propositions: that is, from propositions about how we might see the world to propositions about how the world actually is. More specifically, having argued the case for treating rents as
9 The linkage of class, as conceptualized in the schema, to educational decision making is further treated in Goldthorpe (1996) and Goldthorpe and Breen (1997); to voting, in Evans (1993), to health in Bartley et al. (1996), and to mobility in Goldthorpe (2000, chap. 11).
1579
rates and patterns.
However what, for present purposes, is of chief relevance is the way
the basis of class division and conflict, he spells out what he takes to be the substantive implications of his position in the context of contemporary capitalist society.
In this regard, a central feature of the scenario that Sørensen presents as already unfolding is a drive by employers, exposed to intensifying glob- alized competition and subject to the tyranny of stock market valuations, to eliminate employment rents. This they seek to do not only through weakening organized labor or limiting governmental intervention in labor markets, but further by an avoidance of the kinds of employment contract in which such rents are implicit For example, employers aim progressively to reduce problems of work monitoring and human asset specificity— and the bargaining advantages that these problems give to workers—by introducing new technologies, by redesigning jobs, or by subcontracting or otherwise outsourcing work. It is in the interests of capitalist employers, Sørensen contends (p. 1554), “to produce a labor market conforming to the assumptions of neoclassical economics”—that is, a market in which all employment contracts approximate spot contracts—and this is exactly what they are in process of doing. Thus, the eventual outcome will be that employers will stand opposed to a merely “structureless” mass of workers (p. 1555), within which individuals will be differentiated by the returns they receive to their productive capacities but not by the class positions that they hold in virtue of the particular ways in which their employment relations are contractually mediated. At this stage, therefore, class conflict within the labor market will be at an end.
My own approach to class analysis also gives rise to expectations in the areas on which Sørensen focuses, but to ones of a significantly different
10
I would certainly believe that employers, under the kinds of pres- sure that Sørensen identifies, will attend closely to contractual arrange- ments and that the regulation of employment on the basis of the service relationship, where this represents no more than “organisational slack,” is very likely to be attenuated. But I would at the same time regard the scope for such action as being subject to clear limits. In modern business organizations, there will always remain—and whatever degree of down- sizing or delayering may be achieved—substantial areas of work, notably of a professional, administrative, and managerial character, where the ser- vice relationship in fact best serves productive efficiency. Problems of work monitoring, especially those arising out of complex principal-agent relations, and problems of human asset specificity, especially those of giv-
kind.
10 It may, however, be noted that I share with Sørensen a rational action perspective: i.e., like him, I suppose that capitalist employers, and other parties involved, are, to the best of their ability, acting in a rational—and thus intelligible—way in the pursuit of their goals.
1580
Symposium: Goldthorpe
ing incentives to both organizations and workers to invest in training and “planned experience,” create the need for contractual arrangements that have a more sophisticated rationale than that of the discrete, short-term money-for-effort bargain. And even if new technology can in certain re- spects provide the means of reducing these problems, in others it would seem likely only to accentuate them.
Thus, I would see little possibility of the entropic decline of modern capitalist societies into what, in a seminar presentation, Sørensen once memorably described as “neoclassical soup.” To the contrary, I would re- gard a relatively complex class structure as being an abiding, integral fea- ture of such societies: class will not cease to be a salient feature of their structural geography. In addition to the primary divisions between em- ployers, self-employed workers, and employees, further, and in many re- spects yet more important, divisions will persist within the body of em- ployees. That is to say, divisions reflecting differing forms of the regulation of their employment, the logic of which derives ultimately from the dis- tinctive nature that the labor market, even under the most advanced man- ifestations of capitalism, will still possess.
Of these divisions, the most obvious is that between, on the one hand, the working class, or, more precisely, the class of wage workers, manual and nonmanual, and, on the other, what I would label as the “service class” or salariat of professionals, administrators, and managers (although, as previously intimated, forms of employment contract that mix elements of the basic labor contract and the service relationship can be associated with various “intermediate” class positions). These classes will engage in conflict, whether via individual action—such as voting—or via organiza- tion, over what Sørensen would treat as issues of rent but what I would view generically as issues of class inequality of both condition and oppor- tunity. For while the class structure itself will not be dissolved, there will remain significant scope for inequalities between members of different classes either to widen or to narrow, in consequence of political as well as economic actions and circumstances. As but one illustration of this point, it may be noted that the increase over recent decades in earnings and income inequality in the United States, to which Sørensen gives great emphasis, has not been experienced in other capitalist countries, such as France (see further Atkinson 2000). In sum, I venture to suppose that the class schema that my colleagues and I have developed, in capturing differences in employment relations and including those that derive from the differentiation of employment contracts, provides a conceptual basis appropriate to all three levels of theoretical ambition that Sørensen recog- nizes within the project of class analysis and that it will continue to do so for as far ahead as it is reasonable to look.
There remains, of course, the ultimate question of whether it is 1581
Sørensen’s prognostications or my own that are gaining most support from the emerging empirical evidence. I cannot take this question up in the space here available to me (but see Goldthorpe 2000, chap. 10), and it might in any event be argued that the evidence of a directly relevant kind that is so far available is still too limited to warrant any strong con- clusions. However, thanks to Sørensen’s forthright article, and to his readiness to have it openly discussed, the issue has been fairly joined and future developments can be awaited with interest.
