Terje Wessel, Dep. of sociology and human geography, University of Oslo
P.O.Box 1096 Blindern,
0317 Oslo, Norway
In a classic article, the late Simon Kuznets (1955) argued that income inequality would tend to decline once a society had passed a certain level of development. That threshold was supposed to appear after a transitional heightening of inequality during the early phase of modernization. Sooner or later, according to Kuznets, modernization would bring about a steady redistribution in favour of the poor.
Nowadays, in a period of economic restructuring, reports from western societies often describe a resurgence of income inequality. Although Kuznets did not anticipate a new turn, it may be compatible with his celebrated hypothesis. That will be the case if the causal background is a deteriorating economy, for instance owing to a decline in manufacturing employment, soaring trade deficits or a crack in the stock market. If, on the other hand, rising inequality is produced by economic growth, a more harmful conclusion will be justifiable. In a nutshell, such a pattern makes the Kuznets hypothesis patently invalid.
Simon Kuznets is certainly not a familiar name in the current debate on social transformation in western cities. This is probably explained by his naturalistic approach, his preoccupation with nations and his emphasis on industrialization. It is important to notice, though, that the core model has been extended to include education, energy consumption, population features and political rights (Nielsen 1994). An updated version of the Kuznets curve is thus based on a number of factors that are relevant to present-day urbanisation. More important is the fact that later theoretical developments have rendered ‘laws of succession’ so devalued that they seem to belong to a prehistory of trivialities. That reason alone, but also the subject-matter, makes it useful to review contemporary frameworks and theories against the backdrop of Kuznets’ model.
The paper starts by presenting four theoretical frameworks which all relate to the way global economic restructuring affects inequality in western cities. How do the content and empirical implications of these frameworks differ from Kuznets’ model? The next section presents some results from concrete research in European cities, including my own in Oslo, Norway. What empirical and theoretical insights can be drawn from these analyses? The last section returns to the issue of theory construction. Is there any methodological affinity between current theory and Kuznets’ model? What remains of convergence theories when the empirical statements they make are proved wrong by facts? How could frameworks and theories be improved?
Inequality in the city - theoretical frameworks
Kuznets’ notion that a prospering economy contributes to progressive redistribution seems particularly inappropriate from the point of view of polarization theory. In the formulation of its leading proponent, Saskia Sassen (2000, p. 139), the polarization thesis contends that we are entering a phase of ‘centers and marginality’, whereby increasing poverty and increasing prosperity coexist in the same cities, often in adjacent neighbourhoods. For Sassen, the decline of the manufacturing sector and the subsequent rise of the service sector have generated two trends: first, a growing upper tier of high-skilled jobs with high wages and, second, a massive expansion of low-wage, often temporary, and part-time service jobs. The latter trend is recognized as a structural consequence of the first, generated through a demand for consumer services such as catering, cleaning, clothing and surveillance. It is also an explicit assumption that polarization in the labour market leads to increasing overall inequality of earnings. This is of course fundamentally at odds with neo-classical theory, which assumes that a rise in the relative demand for consumer services will strengthen the position of workers within those labour market segments, and thus retard or completely counterweight the pressure towards greater inequality. According to Sassen, such a simple competitive model does not capture the pattern of change. Partly, there are imperfections in the labour market due to a functionally synchronized supply of low-skilled immigrants. Partly, the new economy is based on a downright exploitive institutional setting: “Finally, much work that was once standardized mass production is today increasingly characterized by customization, flexible specialization, networks of subcontractors, and informalization, even at time including sweatshops and industrial home-work. In brief, the changes in the supply evident in major cities are a function of new sectors as well as of the reorganization of work in both the new and the old sectors” (Sassen 2000, p. 118-119). Similar statements are rather frequent in Sassen’s discussion of labour market changes, and they could be rephrased (sympathetically) as follows: firms have changed the way they utilize labour in response to an uncertain economic environment.
Thus, at one level the polarization thesis contrasts diametrically with the Kuznets model. The suspected direction of change is opposite, and modernization is depicted in terms of fragmentation, antagonism and split interests rather than integration, harmony and mutual benefits. The latter distinction stems from divergent views on advanced technology. Simon Kuznets placed considerable emphasis on long-term effects of shifts in the labour force from agriculture to the modern sector of the economy. Given that agrarian and industrial technology correspond to a high respectively low level of inequality, a rather simple deduction follows: overall inequality will at any point in time depend on the relative size of each sector, the degree of inequality within each sector and the difference in mean inequality between the two sectors. Polarization theorists, by contrast, perceive a dualism withinthe modern sector, that is, within the new service economy. Although the causal chain is overly complicated, there seems to be in the end a rather simple notion of dual technological requirements. To be competitive, firms need to maintain a pool of highly skilled labour, but they do no longer need to engage in the fate of low-skilled labour.
The ‘skills mismatch thesis’ and its offshoot, the spatial mismatch thesis, are easier to harmonize with Kuznets proposition. Like Sassen’s thesis on social polarization, the variables responsible for increasing inequality are found in technology, economic restructuring and global competition. The basic assumption is that the post-industrial society requires less ‘muscle labour’ and more ‘smart labour’. Because there is a shortage of the latter, at least in a transitional phase, one of the inevitable results is an increase in the wages of workers with high skills. By the same logic, less skilled or inappropriately skilled workers suffer relative wage losses or, worse, are excluded from the labour market. Skills mismatch is therefore seen as a result of demand side changes, particularly related to increased trade flows. Richard Freman (1997, p. 28) has formulated the thrust as follows: “…free trade makes worldwide factor endowments rather than domestic labour conditions the determinant of the relative and real pay of factors of production in trading countries”. The policy implications are straightforward: education and vocational training offer the double effect of higher productivity and a lower payment to high skills. In time, the expected result is a stable and possibly reduced level of inequality. As for spatial mismatch, the thesis suggests public spending in order to increase the supply of local jobs.
A supply and demand explanation of this kind does not evoke notions of essence or fundamental needs: there are no implications in terms of institutional regulation or a permanent shift in the distribution of income, wealth or job opportunities. Instead, mismatch of each kind could be interpreted as ‘short swings’ in a steady-state evolution or, alternatively, as episodes in the contingent development of history. The first interpretation offers a slight possibility of full accordance with Kuznets’ model. It is not hard to imagine a causal chain that starts with economic depression, continues with skills mismatch and increasing inequality, continues further with supply side adjustments, and leads back to square one with economic recovery and declining inequality. The second interpretation provides a common ground, although minimal, in the postulation of transitional effects. Kuznets was quite aware of complications related to the interaction between population growth and the growth of capital. In fact, population growth due to reduction in death rates is probably a major reason why the curve bends.
A third theoretical framework emphasises the importance of welfare state policies. For want of a better term, I will call it ‘the welfare modification thesis’. Many authors, mostly from Europe, have claimed that globalization influences inequality, poverty or segregation according to the institutional context, in particular according to the type of welfare state (cf. Hamnett 1996, Musterd and Ostendorf 1998, Kloosterman 1996, Burgers 1996, Andersen and Clark 2001). To be more precise, these authors point at three types of modification measures, first economic redistribution through taxes and transfers, second provision of welfare state services (health, education, housing, childcare, home help etc.), and third creation of jobs and renewal of industrial relations. The literature also highlights the importance of gender and ethnicity, although it is difficult to spot a unified view on the causal implications, specifically regarding gender. Certain participants in the debate, for instance Chris Hamnett, seem to make light of gender, whereas others underscore this dimension very strongly. Juhani Lehto (2000), who belongs to the latter group, argues that the welfare state is a major provider of female-dominated jobs in European cities. Such jobs, according to Lehto, constitute the functional basis of the dual-career family. Both Hamnett and Lehto, however, put Scandinavia at the top of a ‘welfare state hierarchy’. In the words of Lehto (2000, p. 113): “It might be expected that the impact of the welfare state on cities is strongest in Scandinavia”.
What is meant by ‘impact’, what is the causal logic involved, what are the relevant criteria for empirical demarcation, are not at all clear. There is obviously an expectation that major cities in Scandinavia move in a less regressive direction than comparable cities in ‘liberal’ welfare states. But does this suggest a ‘fattening middle’ in Scandinavian cities? Should a moderate rise of inequality count as a refutation? Due to its open nature, the thesis falls well short of providing answers to such questions. A further twist concerns aggregate variations. Pure logic implies that we might be facing three options: a segmented downward trend, a segmented upward trend or a more blurred pattern with cities spreading in diverse directions. Only in the first case could the welfare modification thesis be fitted into or combined with Kuznets’ model. This reading is not as strange as it might seem. Miles Simpson (1990), for example, has argued that political democracy allows the lower classes to improve their share of economic rewards. Similar ideas have been strongly emphasised in comparative welfare state research, particularly by Walter Korpi (1983) and successors within ‘the power resources school’.
The taste for divergence is even more pronounced among supporters of ‘new regional geography’. One of its most prominent spokesmen, Michael Storper (1997), has directed a tough criticism against ‘machine conceptions’ of the post-industrial city. His request for a more place-specific approach is underpinned by empirical analyses that indicate huge variations in the localization pattern of advanced services industries. Comparative advantages and limitations have also been exposed in the European context, and with Burgers and Musterd (2001, p. 130), one might entitle this proposal ‘the place specificity thesis’. Its basic suggestion is that the social and economic character of specific places impinges on processes of economic restructuring and further on the distributive outcomes. With this approach, the room for stylised facts is rather narrow. A possible subtype, though, is one where local history and local autonomy are constrained by distinctive patterns and structures of economic organization. Analysts working within such a framework (‘Post-Fordist city research’)) generally emphasise that global economic processes are themselves uneven, and that it is not only a question of how specific places are situated in relation to global business hierarchies and capital flows. Thus, implicit and explicit empirical statements tend to be opposite of Kuznets’ model.
These frameworks have inspired a substantial number of case studies. The nest section briefly presents some key results emerging form that research. Two reservations, however, are pertinent. First, the discussion of frameworks and theories is not exhaustive. The research on inequality and polarization includes, at least a gender sub-theme (e.g. Bruegel 1996, McCall 2000), an ethnic sub-theme (e.g. Waldinger 1996) and a household/family sub-theme (e.g. Pahl 1988, Pinch 1993). Second, the different strands of theorizing can of course be combined. Efforts to explain economic inequality in a particular context tend to draw on several theories.
Empirical points of departure
Much of the theorizing presented above apply broad concepts that are hard to ‘unpack’ empirically. One particular source of misconception concerns the difference between inequality and polarization. It is by now generally acknowledged that increasing inequality does not necessarily imply polarization, and vice versa. The middle of a distribution can be ‘hollowed out’ (i.e. polarization) while simultaneously improving the Lorenz criterion, which is an indisputable criterion of increasing equality. Such anomalies crop up because polarization has to do with clustering of attributes (two sub-groups), whereas inequality is independent of sub-group frequencies. Some theorists, however, tend to use the two concepts interchangeably. Partly, they neglect the difference related to sub-group frequencies, partly they refer to differing units of measure, such as jobs, social groups, income, share of income and wage-level. A certain amount of fuzziness is probably difficult to avoid for anyone who attempts to capture the Zeitgeist, for instance ‘the informational age’ (Manuel Castells). But fuzziness is definitely not conducive to sound empirical research. Possibly for this reason, there is now an increasing consensus that ‘polarization’ refers to a changing distribution of jobs or people at both tails. ‘Inequality’, on the other hand, concerns dispersion of income (or levels of income). This is a concept with numerous alternative measures, but none of them demonstrates (at least not effectively) the shape of the distribution. More descriptive methods have therefore been developed to capture substantive details of each theory. These methods basically expose the changing shape of the distribution by breaking it into sections for a baseline year and tracking the percentage of people in each section over time. Conveniently, two of the theories have fairly distinctive implications – growth at both tails (the polarization thesis) and growth at the top (the skills mismatch thesis). The welfare modification thesis has sometimes been linked to ‘professionalization’, and one version of it might thus imply a ‘fattening middle’.
Uncertainty also exists regarding the relationship between income, wealth, education, life chances etc. It is increasingly evident that the middle population might be growing in one sense and declining in another. A growing section of middle-class professions and titles should not be interpreted as a general sign of upgrading.
A third question concerns the choice of unit, and a fourth the relation between distributive change per se and residential changes. Briefly, the polarization and mismatch theses relate to individuals (in employment) and propose increasing socio-spatial inequalities. The welfare modification thesis and the place-specificity thesis are concerned both with individuals and households and make no claim regarding urban morphology.
Published pieces in this field of research are often based on one city. As far as I know, there are no reports covering a large number of cities in different countries. This is a pity because the theoretical dispute centres on convergence tendencies, either across a vast array of national/institutional contexts or within specific ‘welfare state regimes’. The existing record may nevertheless provide useful clues on the subject. I have chosen to organize the presentation in three sections, relating to polarization, inequality and residential changes.
A significant part of the social polarization literature does not relate explicitly to economical issues, but rather to skills, health, mortality, ‘quality of life’ or to indicators of social exclusion such as unemployment, homelessness, crime, fear of crime and stigmatization. One may also encounter reports on economic dualism that neglect the clustering aspect. But at least three studies have utilized a technique that effectively reflects population frequencies. Robert Kloosterman (1996) analyses changes in the wage-class distribution in Amsterdam and Rotterdam during 1980-92. Both cities suffered job losses in the middle-paid, manufacturing segment, as anticipated. These losses were accompanied by increases in low-paid segments of the consumer services, but there were also a number of unexpected observations, most notably a substantial job-growth in the highest wage-classes of manufacturing, trade, hotel and transport. The resulting outcome was a limited polarization of the labour force in Rotterdam and a marked downward shift in Amsterdam. Yet another pattern emerges in a study from London. Using data both at the household and the individual level, and covering the period 1979-95, Hamnett and Cross (1998) conclude that polarization is highly asymmetric, with growth at the top end far greater than growth at the bottom. The upward trend was particularly marked at the individual level, with “…no evidence of earnings polarization for male full-time workers, female full-time workers, or female part-time workers …” (Hamnett and Cross 1998, p. 677).
My own research in Oslo adds complexity to the picture. The data analysed so far have not been disaggregated into specific sections of the economy, nor to specific occupations. I should therefore make a strong reservation regarding economic and technological determinants. The preliminary results (Wessel 2001) do however point in several directions, depending on the observation unit, the period of time, the definition of income and demographic characteristics. The number of households falling into the 4th to the 8th decile decreased during 1986-96, largely because of a downward shift. A second set of data for the period 1993-98 showed an opposite trend for full-time male workers (figure 1) and a perfect symmetrical polarization for non-western immigrants (figure 2).
Figure 1 Changes in the distribution of full-time workers across deciles 1993-98. Market income. Cut-off points in 1998 deflated by the general rate of inflation. Oslo
Figure 2 Changes in the distribution of non-western immigrants across deciles 1993-98. Market income. Cut-off points in 1998 deflated by the general rate of inflation. Oslo Statistical analyses by Angelo (1995, Storper (1997) and Fainstain (2001), and probably many others, support the view that polarization is not an inevitable, maybe not even a likely outcome of technological change and sectoral shifts. Michael Storper (1997, p. 231-232), comparing Los Angeles, New York and the United States, is uncompromisingly clear: “This appearance of polarization is false: the whole distribution followed a more generally upward direction than in the economy as a whole, hardly the catastrophic picture painted by the global-dual city theory”.
Evidence on income inequality, on the other hand, is tidy and continuous, containing few imprints of political intervention, industrial relations or cultural values. There are of course variations in the extent of change, but the general conclusion is that present-day urban society displays widening income inequality. A clear ‘trend’ emerges from numerous studies, including cities and conurbations such as New York, Tokyo, Toronto, London, Paris, the Randstad, Brussels, Hamburg, Stockholm and Oslo (Fainstain 2001, Murdie 1998, Hamnett and Cross 1998, Kesteloot 2000, Friedrichs 1998, Borgegård et al. 1998, Wessel 2001). Susan Fainstain (2001, p. 286) is worth quoting on this point: “Analysis of the social structures of the global city-regions of New York, Tokyo, London, Paris and the Randstad supports an argument of increasing inequality, but it is much more ambiguous on the issues of the declining middle and of growth at the bottom”.
The third type of outcome, segregation, is marked by diversity. The link between household income and spatial location is diluted by diversification in family types and housing preferences, regulations and subsidies in the housing market, moving behaviour and adoption of post-modern planning policies (Van Kempen 1994, Wessel 2000). Thus, one should not be surprised that residential developments are harder to theorize than social stratification (cf. Murie and Musterd 1996, Murie 1998, Preteceille 2000, Wessel 2000). ‘Place’ as a causal factor is of course further obscured by the existence of neighbourhood effects, that is, escalator effects that arise from socialization, stigmatization, institutional deficiencies and functional disadvantages. Such complexities have been increasingly exposed in European urban research (cf. Andersson 2001, Urban Studies no. 12, 2001), and although the exact policy implications are hard to establish, the overall message is one of optimism. The following statement is typical rather than exceptional in its trusting approach to political agency: “Nevertheless, (...), state spatial policies can have a dramatic effect on the divisions of the city” (Marcuse and Van Kempen 2000, p. 273).
Three of the theses discussed above seem to share a nomothetic ambition: they are based on the postulate that it is possible to make propositions regarding inequality, differentiation and residential change that are not limited in validity to specific places. The polarization thesis makes a conditional statement of the type ‘if A, then B’, or in a watered-down and probabilistic version, ‘if A, then usually B’. Thus, the trend towards increasing inequality and polarization is not conceived as irreversible or linked to economic growth as such. It is only through the emergence of computer-based technology, removal of geographical constraints and a new mode of political regulation that social polarization becomes a far-reaching phenomenon. This is a rather different way of theorizing than modernization theory, to which Kuznets’ model belongs. Both Kuznets himself and the research tradition he inspired emphasise evolutionary stages that represent progress and that are intrinsically linked with the past. Change is seen as a cumulative process in which earlier developments more or less determine later developments. The polarization thesis, by contrast, is preoccupied with cause-and-effect relationships within a restricted time-span. There is no suggestion that a pre-industrial past, for example the distinction between an agrarian and a horticultural heritage, has a bearing on present-day trends. Not even the preceding modernist era seems to make a vital difference. Saskia Sassen (2000, p. 44) argues that the new economic trends are associated with existing economic functions and not with impacts of urban system balance/imbalance or naturally grown industrial traditions: “The new inequality differs from the long-standing forms of inequality present in cities and national systems because of the extent to which it results from the implantation of a global dynamic, be it the internationalization of production and finances or international tourism” (Sassen 2000, p. 41).
The mismatch thesis contains the proposition that ‘if A, then temporarily B’. Technological and economic heritage are important as implicit premises, but they are not part of the ‘causal logic’. It is also significant that the thesis describes one causal factor, skills, and not a system of factors. The nomothetic bias is thus far from ‘structuralist’.
Yet another position is suggested by the welfare modification thesis. In this case, there is no specific succession of events, but rather a certain area of reality that produces a complex outcome of divergence and convergence, that is, convergence among cities in similar welfare states and divergence among the various groups of welfare states. Whereas ‘A’ is rigorously defined, ‘B’ is fraught with ambiguity. Much the same applies to certain interpretations of the place specificity thesis, above all attempts to establish ‘urbanness’ as the driving ‘force’ of urban development (cf. Storper 1997, Soja 2000).
It follows from this discussion that a methodological affinity exists between Kuznets’ model and current theory on distributive change. The evolutionary approach, suggesting that inequality and social differentiation move in a particular direction, is dumped. Some scholars may even have dumped the nomothetic ambition all together. Those working within the frameworks of ‘polarization’, ‘mismatch’ and ‘welfare state modification’, however, still expose an aspiration to transcend the local level. The determinism in these frameworks, and one might add ‘the synekism’ of Soja and ‘the economic reflexivity’ of Storper, bears little resemblance to Laplace’s demon. I nevertheless prefer a somewhat different approach, one that is characterized by abstract conceptualisation and absence of black boxes.
My core argument is that we need explicit models of distributive change in western cities. Much of the current literature alternates in a confusing way between empirical and theoretical statements. References to investments, world trade, employment shifts, wage trends etc. are mingled with conceivable explanations, often with little or no linguistic distinction between the fuzzy world of real-life matters and the idealized world of conceptual constructs. Using the same language for both purposes may not be a problem in itself; this is of course what characterizes dense explanations in humanistic and social research. It is however a problem when virtuality and reality are densely packed within nomothetic frameworks. One imminent danger with such texts is the well-known ‘fallacy of misplaced concreteness’ (mistaking the abstract for the concrete), or oppositely, reading too much into a purely descriptive account.
Nomothetic propositions may sometimes be justified, but the history of social sciences indicates that it is a risky research strategy. A formal perspective in the tradition emanating from Marx, Weber and Simmel avoids trouble by focussing on highly selective assumptions – a stripped-down picture of reality. It is a general framework, but it doesn’t imply any empirical conclusions, and it can only be matched to reality indirectly. Constructing such models makes it possible to explain various processes and various outcomes, rather than frequent or universal processes/outcomes. In the course of time, through extensive empirical research, one may even acknowledge that regular patterns exist. A strict nomothetic concept, by contrast, entails the danger that valid and productive ideas are rejected due to invalid inferences.
It might be informing at this point to bring forward a notion in the theory of knowledge. ‘The Simmel effect’, according to Boudon (1997, p. 27), appears when “…we combine with theories or acceptable ideas, implicit propositions (a priori) based on good reasons but which can lead us to dubious conclusions”. A common version of the effect is to exaggerate the validity of a sensible model, in other words, to assert more than the explicit arguments allow. Such statements are “…generally detected when the theory happens to suffer setbacks” (Boudon 1997, p. 98). A revelation of the excessive validity is often followed by complete rejection of the conceptual argument.
As far as I am concerned, this is exactly what has happened to the polarization and mismatch theses. Ten years ago both constructs spread from their American origins, stimulating empirical research in other parts of the world. Gradually, we all discovered that welfare state politics, customs, cultural values, industrial traditions, and so on, mattered. One possible response to that empirical chaos is illustrated in a much quoted and valuable article by Jack Burgers. Observing that unemployment influences social stratification in Dutch cities, and relying on welfare state theory, his conclusion is that “…social polarisation appears in different guises in different countries” (Burgers 1996, p. 102). The main figure in the polarization debate, Saskia Sassen, has recently subscribed to this position: “Secondly, the development of global city functions is filtered partly through thick local institutional environments and legal/administrative frameworks, so it is not simply a standardised implant that looks the same everywhere” (Sassen 2001, p. 2538). An alternative response is to look for less incriminating concepts, for example replacing ‘social polarization’ with ‘social fragmentation’ (van Kempen et al. 2000). Yet another alternative is a postmodernist transgression of ‘binary thinking’, emphasizing disorder, “thirdspaces” and anarchistic epistemologies (Sibley 2001).
A fourth solution, and one I support, is to restore confidence in the beliefs that aspired mismatch and polarization theorists. The polarization concept, in particular, is a wonderful analytical creation. It directs our attention to “…the generation of tensions, to the possibilities of articulated rebellion and revolt, and to the existence of social unrest in general” (Esteban and Ray 1994, p. 820). It is an interdisciplinary concept that allows quantitative research on distributive properties as well as qualitative research on in-group identification and out-group hostility. I see no reason why such a useful idea should be sacrificed in the everlasting search for persuasive (and self-persuasive) representations. It is crucial, however, that a perfectly acceptable intuition be identified and separated from the hyperbolic conclusion. Social polarization depends on complex preconditions, and it should not be taken as an easily predictable phenomenon. This insight, which has been spreading for the last 10 years, has no implications regarding the image of polarization. The image as a research instrument is both faultlessand sustainable. It is due to this instrument that the contingent character of the process has been identified.
A similar formal approach (cf. ‘ideal types’) is relevant in studies of income inequality. But the conceivable content is rather different. The ‘episodic’ nature of polarization (and segregation) demands a careful and detailed specification of all elements and connections in the model. One cannot neglect the specification of actors with specific attributes, or the sorting of actors to specific scenes of action, or the resulting effects in terms of structures (positions, resources, opportunities) and actors (attributes, motivations, interests, vitality, knowledge etc.). Analyses of income inequality require a less rigorous virtuality. The observation of a trend points in the direction of ‘staging’ aspects, that is, the scenes of action. If the main objective is to explain the trend, we need to know less about actors and selection mechanisms. As I have been pointing out elsewhere (Wessel 2000), we are still dealing with capable human beings, or rather human copies, but their attributes, interests and knowledge play a minor role in the creation of structural effects. Put differently, a particular result (increasing income inequality) is produced whether we represent politics as a neo-liberal agenda, dominated by austerity measures, or as a heroic social-democratic struggle against global pressures. The type of scene on which the human copies strive or become paralysed might entail diversity in terms of productivity. In this case, the logic would at least include unregulated or low-regulated wage adjustments in the consumer services. Nations that used to rely on counter-strategies are forced to defect because of relative costs: it becomes too expensive to allow service earnings to follow general wage developments in the economy, or to omit the productivity problem through direct public service provision (Esping-Andersen 1998). A decline in centralization and co-ordination among unions, as well as a general union decline, might of course enhance the market-clearing strategy. In the opposite end of the wage spectrum there is the ‘superstar effect’. This phenomenon is described by Sherwin Rosen (1981) as a matter of scale economies in sport, arts and letters, show business and certain professional services. A crucial point is that technological changes, specifically in communications and transportation, have “…expanded the potential market for all kinds of professional and information services, and allowed many of the top practitioners to operate at a national or even international scale” (Rosen 1981, p. 856). Similar ideas find expression in conversations and arguments about ‘the new economy’. Technology is generally portrayed as the motor of “…an ever-expanding choice of terrific deals, fabulous products, good investments, and great jobs for people with the right talents and skills” (Reich 2001, p. 7). One important personality in this world is ‘the geek’, the inventor who is capable of seeing new possibilities in a technical medium and who is able to deliver on these opportunities. He or she is not primarily a computer specialist, but rather ‘a dreamer, a visionary, sometimes a revolutionary’ (ibid. p. 54). The geek finds pleasure in stretching the medium as far as possible, constantly searching for new options and solving dilemmas within the medium. A second essential personality is the marketer, the trend spotter or the consultant – in short, the person who can identify possibilities in the marketplace. He or she is an expert in the wants and latent desires of consumers, often for products that do not yet exist. The competence is thus genuinely different from the conventional salesperson, whose primary obligation and gift is to persuade, to play upon peoples emotional needs by adding qualities to existing products. The new type of marketer, according to Reich, resembles a counsellor or even a psychotherapist – an empathetic rather than a charming and manipulative person. Together the skills of the geek and the marketer (‘the shrink’) form the basis of contemporary entrepreneurship. Both personalities are in great demand, all due to a technology that “…increases the value of creativity by allowing it to be spread quickly throughout an organization’s network and, ultimately, to consumers” (Reich 2001, p. 52).
MORE ON SOCIAL CHOICES Looked at this way, widening income inequality approaches a ‘closed process’ (cf. Boudon 1991). Technological factors have a direct impact on institutional regulation, cultural activities and social choices, all of which contributes to a repetitive outcome. The human copies are by no means passive, irrational or incapable of learning. They are however put in situations where their areas of strategy are reduced to a minimal level (political regulation), or where the rewards and responsibilities encourage a certain course of action (social choices). In more general terms, the repetitive outcome of the process depends on a complex combination of restrictions and inducements. That complexity as well as the envisaged outcome is conceived as genuinely fragile. Introducing new components, such as supranational regulation, might easily boost the importance of agency (‘casting’).
The model presented here, with its emphasis on productivity and entrepreneurship, is not the only conceivable tactic. Danny Quah (2002) has recently suggested that several puzzles and paradoxes within the new economy dissolve once we accept the importance of human capital. The essential changes according to Quah are not productivity or supply-side improvements, but rather consumption and demand-side changes. On this view, the new economy differs because consumption occurs increasingly in goods that resemble knowledge, such as computer software, video entertainment, gene sequences and Internet-delivered goods and services. There is also a continual focus on social norms, reputation and voice mechanisms. An influential spokesman of political and sociological explanations is, astonishingly, the economist Anthony B. Atkinson. His arguments (cf. Atkinson 1999) are well worth reflecting on, although I have dared to question the empirical support (Wessel 2001).
The urban context
Agglomeration dynamics and territorial specificities are sometimes represented as ‘scale effects’ within multi-layer models (c. Burgers and Musterd 2001). This is a comprehensible and attractive way of dealing with complex interrelationships in contemporary capitalism, but it entails three dangers. First, it is easy to slip into a way of thinking where ‘processes’ operating at a global scale are mediated by welfare states and further by local actors in the worlds of politics, economy and civil life. Thus, the implication is that unknown or obscure agents are forming geographies at higher scales, whereas firms, persons, groups and organizations are actively engaged in fashioning the local community. Second, accompanying the bias towards processes is a tendency to enter differing scales in a chronological order, from top to bottom. Aside from the ontological inferences of such theorizing, there is an imminent danger that relevant propositions are left out of the analysis. Welfare state politics may be conspicuously absent at the local level and ‘social capital’ (for want of a better word) at the national level. Third, and also related to the agency gap, focusing on layers of causation tend to imply that diversity increases at lower scales. Considering sectoral shifts and technological change, the opposite may very well be the case. One should not be surprised that reports on major cities present a more clear-cut pattern of rising inequality than do OECD reports on western countries.
The approach suggested here puts agency at the centre stage. Technological, economic, political and cultural influences are, at least intentionally, transformed to frames of action, personal attributes, values and preferences, positions and incentives, selection mechanisms and aggregate effects. There are of course myriads of difficulties in terms of abstraction, symbolic representation and empirical inferences. It is obviously enormously difficult to avoid ‘black boxes’, poor deductions and insufficient measures. The best one can hope for is probably a moderately successful integration of theory and concrete research.
What then are the significant ‘urban’ effects in studies of income inequality? The answer can be split in two parts. First, we have a series of compositional effects related to uneven development, mobility and demographic composition. Economic restructuring varies by region, and so does population characteristics. There are thus numerous aggregations found in economic activities, social classes, ethnic minorities and lifestyle groups. Although politically incorrect, one might even add the composition of talented people:
“ The important practical implication is that it is monetarily advantageous to operate in a larger overall market; it is increasingly advantageous the more talented one is. (…..). The best doctors, lawyers, and professional athletes should be found more frequently in larger cities” (Rosen 2001, p. 856).
A second type of effect relates intrinsically to the local context. ‘Contextual effects’ may point towards stable attributes such as topographies, environmental qualities, infrastructure and buildings. It may also include political characteristics (e.g. local service provision, transfers, taxes, political networks and participation) and sentimental characteristics (e.g. place identity). A third sub-type, perhaps the most interesting, is derived from thresholds of population characteristics. Here, the basic point is that co-existence of individuals with similar or different features give rise to additional effects on opportunities, information, ambitions or expectations. Needless to ay, such effects are extremely difficult to document.
The Oslo context is obviously characterized by several compositional effects, most of which contributes to a higher level of income inequality than the national average. One may also safely infer that these effects have been strengthened due to deindustrialisation, service sector expansion and large-scale immigration. But are there any significant contextual effects? A tempting speculation would be that organizational changes and blinding speed have compelled harder work and made loyalty obsolete within certain professional groups. Following Robert Reich (2001), these effects may in the end have contributed to widening economic inequality. A second speculation might be that certain parts of the consumer service sector (e.g. in catering) are distinguished by a brutal business climate. The frequency of bankruptcy and the violations of labour market protections are notoriously high, with few signs of improvements. But it may also be the case that the very same services face less ferocious competition from family self-servicing than elsewhere. According to rumours, Oslo is now rapidly turning into a genuine metropolis.
- Back to Kuznets:
- Two ideal types:
1) ‘if A, then sometimes B’
2) ‘if A, then usually B’: a ‘well-tempered determinism*
ATKINSON, A. B. (1999) Is rising income inequality inevitable? A critique of the transatlantic consensus. WIDER Annual Lectures 3. UNU World Institute for Development Economic Research.
BOUDON, R. (1997) The Art of Self-Persuasion. Cambridge, UK: Polity Press.
BOUDON, R. (1991) Theories of Social Change: A Critical Appraisal. Cambridge, UK: Polity Press.
ANDERSEN, H.T. AND CLARK, E. (2001) Does welfare matter? Ghettoisation in the welfare state. Lund: Preliminary paper.
ANDERSSON, R. (2001) Spaces of Socialization and Social Networks Competition: A Study of Neighbourhood Effects in Stockholm, Sweden, in H.T. Andersen and R. van Kempen (eds.) Governing European Cities. Social fragmentation, social exclusion and urban governance. Aldershot: Ashgate, pp. 149-188.
ANGELO, L. (1995) Marching to its own drummer. Why wage trends in New York City have diverges from the nation, Urban Affairs Review, 31, pp. 104-119.
BORGEGÅRD, L.E., ANDERSSON, E. AND HJORT, S. (1998) The divided city? Socio-economic changes in Stockholm metropolitan area, 1970-74, in S. MUSTERD and W. OSTENDORF (Eds) Urban segregation and the welfare state. Inequality and exclusion in western cities, pp. 206-222. London and New York: Routledge.
BRUEGEL, I. (1996) Gendering the polarization debate: a comment on Hamnetts ‘Social polarisation economic restructuring and welfare state regimes’. Urban Studies, 33, pp. 1431-1439.
BURGERS, J. (1996) No polarisation in Dutch Cities? Inequality in a Corporatist Country, Urban Studies, 33, pp. 99-105.
BURGERS, J. AND MUSTERD S. (2001): Global and local determinants of social exclusion: Amsterdam versus Rotterdam, in H.T. Andersen and R. van Kempen (eds.) Governing European Cities. Social fragmentation, social exclusion and urban governance. Aldershot: Ashgate, pp. 125-148.
FAINSTAIN, S. (2000) Inequality in Global City Regions, in A. Scott (ed.) Global City-Regions. New York: Oxford University Press.
FREEMAN, R.B. (1997) Are Norway's solidaristic and welfare state policies viable in the modern global economy, in J.E. DØLVIK and A.H. STEEN (Eds) Making solidarity work? The Norwegian labour market model in transition, pp. 17-50. Oslo: Scandinavian University Press.
FRIEDRICHS, J. (1998) Social inequality, segregation and urban conflict: the case of Hamburg. in S. MUSTERD and W. OSTENDORF (Eds) Urban segregation and the welfare state. Inequality and exclusion in western cities, pp. 168-190. London and New York: Routledge.
HAMNETT, C. (1996) Social polarisation, economic restructuring and welfare state regimes, Urban Studies, 33, pp. 1407-1430.
HAMNETT, C. AND CROSS, D. (1998) Social polariation and inequality in London: the earnings evidence, 1979-95, Environment and Planning C: Government and Policy, 10, pp. 659-680.
KESTELOOT, C. (1998) The geography of deprivation in Brussels and local development strategies, in S. MUSTERD and W. OSTENDORF (Eds) Urban segregation and the welfare state. Inequality and exclusion in western cities, pp. 126-147. London and New York: Routledge.
KORPI, W. (1983) The Democratic Class Struggle. London: Routledge.
LEHTO, J. (2000) Different cities in different welfare states, in A. BAGNASCO and P. LE GALÈS (Eds) Cities in contemporary Europe, pp. 112-130. Cambridge: Cambridge University Press.
MARCUSE, P. AND VAN KEMPEN, R. (2000) Conclusion. A Changed Spatial Order, in P. MARCUSE AND R. VAN KEMPEN (Eds) Globalizing Cities. A New Spatial Order. Oxford UK: Blackwell Publishers, pp. 276-301.
MCCALL, L. (2000) Explaining Levels of Within-Group Wage Inequality in U.S. Labour Markets. Demography, 37, 4, pp. 415-430.
MURDIE, R. A. (1998) The welfare state, economic restructuring and immigrant flows: impacts on socio-spatial segregation in Greater Toronto, in S. MUSTERD and W. OSTENDORF (Eds) Urban segregation and the welfare state. Inequality and exclusion in western cities, pp. 64-93. London and New York: Routledge.
NIELSEN, F. (1994) Income inequality and industrial development: dualism revisited. American Sociological Review, 59, pp. 654-677.
PINCH, S. (1993) Social Polarization – A comparison of evidence from Britain and the United States. Environment and Planning A, 25, 6, pp. 779-795.
QUAH, D. (2002) Technology dissemination and economic growth: some lessons from the new economy. London: Centre for economic policy research. London School of Economics. Discussion paper no. 3207.
REICH, R. (2001) The Future of Success. New York: Alfred A. Knopf.
ROSEN, S. (1981) The Economics of Superstars, The American Economic Review, 71, pp. 845-858.
SASSEN, S. (2000) Cities in a world economy. Thousand Oaks m.fl: Pine Forge Press.
SASSEN, S. (2001) Global Cities and Developmental States: How to Derail What Could Be an Interesting Debate: A Response to Hill and Kim, Urban Studies, 38, pp. 2537-2540.
SIBLEY,D. (2001) The Binary City, Urban Studies, 38, pp. 239-250.
SIMPSON, M. (1990) Political Rights and Income Inequality. A Cross-National Test. American Sociological Review, 55, pp. 682-93.
SOJA, E.W. (2000) Postmetropolis. Critical Studies of Cities and Regions. Oxford UK: Blackwell Publishers Ltd.
STORPER, M. (1997) The Regional World. New York: The Guilford Press.
VAN KEMPEN, E. (1994) The Dual City and the Poor: Social Polarisation, Social Segregation and Life Chances. Urban Studies, 31, pp. 995-1015.
WALDINGER, R. (1996) Still the Promised City. Afro-Americans and New Immigrants in Post-industrial New York. Cambridge MA: Harvard University Press.
WESSEL, T. (2000) Social polarisation and socioeconomic segregation in a welfare state: the case of Oslo, Urban Studies, 37, pp. 1947-1967.
WESSEL, T. (2001) Losing Control? Inequality and Social Divisions in Oslo. European Planning Studies, 9, pp. 889-906.