ASI response paper

These are confusing times for those interested in the evolution of urban systems in advanced economies and in the changing spatial structure of urban settlements, as more evidence of devel- opments since the early 1980s becomes available. The title chosen for this chapter by the editor encapsulates the ‘paradigm’, or standard perspec- tive, that by then had become the most commonly used benchmark against which the experience of individual countries and urban regions could be compared. Reflecting urban analysts’ preoccupa- tion with ‘decades’ of change – a very unhealthy feature of urban studies, but an understandable one bearing in mind the frequency and incidence of population censuses – a picture has emerged of ‘urbanization’ predominating in the 1950s and ‘suburbanization’ accelerating in the 1960s, with the 1970s emerging as the ‘decade of counterur- banization’. The significance of the latter has remained a hotly debated topic, with past predic- tions for the 1980s and beyond variously suggesting a fuller development of centrifugal tendencies, a return to the ‘normal’ processes of metropolitan and urban concentration after the ‘anomalous 1970s’, and a period of ‘reurbaniza- tion’ associated with a natural progression through sequential stages of urban development in some form of cyclic pattern.

The aim of this chapter is to discuss the signif- icance of the more recent developments of coun- terurbanization and reurbanization in the context of the ways in which urban growth in the advanced economies has traditionally taken place. The first section focuses on urbanization and, in so doing, introduces the main approaches which have been used to make sense of the urban trends of the past half-century. Urbanization has been seen variously as the increasing concentra- tion of national populations into towns, as the increasing concentration of a country’s urban population into the largest cities, and as the increasing concentration of an urban region’s

population into its core at the expense of its surrounding ring. It is shown that in the 1950s all these three processes were operating side by side quite commonly across the developed world. The second section deals with the first main departure from these traditional patterns, namely the reversal of within-urban-region population shifts in favour of the suburban ring at the expense of the core, while the third section explains how counterurbanization represents the reversal of the concentration of the urban population in the largest cities and examines how far this process has proceeded. The fourth section explores the theoretical background to the ‘reurbanization’ hypothesis in terms of a return to traditional patterns of urban change and goes on to assess the extent to which this has occurred in recent years. It is concluded that none of these prefixed versions of ‘urbanization’ can adequately encap- sulate the developments observed since the late 1970s, raising questions about whether it is sensible to try to impose any single model.

Throughout, the primary emphasis is on urban trends as measured by data on residential popu- lations and on understanding these changes in terms of the direct demographic causes. Only limited mention is made of the environmental, economic, social, cultural and political factors that help to produce the observed changes in migration and population composition. These latter aspects are dealt with more fully in other chapters.

URBANIZATION AND THE STUDY OFURBAN CHANGE Given that this chapter is entirely about patterns of urbanization, it may seem strange to begin with a separate section on urbanization. This step is, however, essential because of the variety of

9 Urbanization, Suburbanization,Counterurbanization and Reurbanization

TONY CHAMPION

ways in which the term has been used in the liter- ature. Some have conceived of urbanization in the physical sense of the increasing area of land being developed for urban use, while others view urban- ization as a social process of people adopting the attitudes and behaviour traditionally associated with life in cities and towns, irrespective of where they might be living. Even when it is interpreted from the perspective of population geography so as to relate to the type of settlement in which people live, as it is for present purposes, the term has been applied in several different ways. According to which approach is taken, urbaniza- tion can be considered to be a continuing process, one that is long since over or one that is not currently very important but may undergo a resurgence, perhaps in cyclic fashion. The fact that the term has been elaborated by the attach- ment of various prefixes, as exemplified in the title of this chapter, suggests that our emphasis should be on ‘urban change’ rather than on ‘urbanization’ per se. The rest of this section, building on the discussion in Chapter 2, expands on this point and provides the context for the subsequent sections.

Traditionally, the most common way of meas- uring urbanization is in terms of the proportion of a national population that lives in ‘urban places’. Somewhat paradoxically, it is in this sense that urbanization appears to be a steadily contin- uing process, yet this measure is now of virtually no value for studying urban change in advanced economies. The official figures, provided by national statistical agencies and summarized for broad regional groupings by the United Nations, indicate that around three-quarters of the popu- lation of the more developed world are consid- ered to be urbanized, a proportion which has risen from two-thirds in 1970 and from just over half in 1950 and which is expected to increase further to five-sixths by 2030 (Table 9.1). Clearly

evident is the strong growth in ‘level of urbaniza- tion’ since 1950 in Eastern Europe and Southern Europe, and Japan. Even here, however, the rate of increase is now levelling off and approaching what appears to be a ceiling of 80–90 per cent – a process which was largely complete in England and Wales by 1900 and in most other parts of the more developed world by mid century. As indi- cated in Chapter 2, the anticipated future increases have more to do with the reclassification of existing rural settlements as a result of the outward spread of cities or their populations than with further large-scale movements of people from rural to urban areas.

Far more valuable in the study of contempo- rary urban change is the measure of urbanization that is based on the distribution of the population between different sizes of urban places. Its central importance was spelt out more than half a century ago in Tisdale’s (1942) description of urbanization: ‘Urbanization is a process of popu- lation concentration. It proceeds in two ways: the multiplication of the points of concentration and the increase in size of individual concentrations’ (quoted by Berry, 1976a: 17). The most extreme examples of this latter aspect relate to those coun- tries where one leading city has outstripped all the others, producing a primate rank-size settlement pattern, as in the case amongst the advanced economies for Vienna, Paris and London – the latter still six times the size of the UK’s second city, Birmingham (Champion et al., 1987). More generally, it is seen in the high ‘degree of urban concentration’ found widely across the more developed world, with cities of a million inhabi- tants or more accounting for at least a quarter of urban dwellers in most countries and for over a third in some, these latter including Australia, Greece, the USA, Austria and Portugal (Table 9.2). Formalizing this perspective, Fielding (1982) has defined urbanization as being where there

144 Handbook of Urban Studies

Table 9.1 Level of urbanization, 1950–2030

Region 1950 1970 1990 2010 2030

(%) (%) (%) (%) (%)

More Developed World 54.9 67.6 73.7 78.7 83.7

Australia/New Zealand 74.6 84.4 85.0 85.9 88.9

Northern Europe 72.7 80.4 83.0 85.5 88.8

Western Europe 67.9 76.7 80.7 84.2 87.8

North America 63.9 73.8 75.4 79.6 84.4

Japan 50.3 71.2 77.4 80.9 85.3

Southern Europe 44.2 56.6 63.1 68.1 75.2

Eastern Europe 39.3 55.8 68.1 75.7 81.3

Figures give share of population living in urban places. Regions are ranked according to share in 1950.

Source: calculated from United Nations (1998) World Urbanization Prospects. The 1996 Revision. Sales No.

E.98.XIII.6, Table A.2

exists a direct urban-system-wide relationship between the rate of net migratory growth of settlements and measures of their urban status (Figure 9.1a).

Table 9.2 Degree of urban concentration, 1965–90

1965 1990

Region and country (%) (%) Change

North America

Canada 37 39 +2

United States 49 48 –1

Eastern Europe

Bulgaria 21 19 –2

Czechoslovakia 15 11 –4

Hungary 43 33 –10

Poland 32 28 –4

Romania 21 18 –3

Northern Europe

Denmark 38 31 –7

Finland 27 34 +7

Sweden 17 23 +6

United Kingdom 33 26 –7

Southern Europe

Greece 59 55 –4

Italy 42 37 –5

Portugal 44 46 +2

Western Europe

Austria 51 47 –4

France 30 26 –4

Netherlands 18 16 –2

Australia 60 59 –1

Japan 37 36 –1

Figures refer to the proportion of the urban

population that lives in cities with at least one million

inhabitants in 1990.

Source: World Bank (1992) World Development

Report, 1992. New York: Oxford University Press,

Table 31

The interest in this second interpretation of urbanization arises largely because there does not seem to be the same unidirectional pattern of development as there is for the simple ‘per cent urban’ measure. Table 9.2 shows a widespread tendency for a reduction in the degree of urban concentration between 1965 and 1990, with 15 of the 19 advanced countries with data available for both years being estimated to have seen a fall in the proportion of their urban population living in cities of at least one million residents. Though this tendency can arise from statistical underbounding, whereby the statistical boundaries of a city fail to keep pace with its lateral expansion, the data are

suggestive of a widespread process of population redistribution down the urban hierarchy, either through the relatively faster growth of smaller urban places or through the absolute decline of the largest cities. This development has indeed been confirmed by a number of more detailed studies which have been careful to allow for underbounding; for example Fielding’s (1982) analysis of France showing Paris’ migratory growth rate falling behind that of the next rank of cities and eventually turning negative, the studies which have shown the USA’s medium-sized and smaller metropolitan areas overtaking the growth rate of the largest ones in the 1970s (for example, Frey, 1990; Nucci and Long, 1995) and the national case studies of the present author (Champion, 1989).

This switch in the incidence of the strongest population growth away from the largest cities was termed ‘polarization reversal’ by Richardson (1977, 1980) and, along with observations of the revival of population growth in non-metropolitan America (Beale, 1975; Morrison and Wheeler, 1976), spawned the notion of ‘counterurbaniza- tion’ (Berry, 1976a). The former has been incor- porated by Geyer and Kontuly (1993) into a ‘theory of differential urbanization’, whereby patterns of gross migration alter over time and successively favour the primate city, the interme- diate-city level and ultimately the small-city level (Figure 9.2). In this approach, ‘urbanization’ is deemed to be taking place as long as the popula- tion is becoming increasingly concentrated in the primate city category, before the intermediate-size city category takes over as the fastest-growing of the three categories.

The seminal contribution to the rigorous analysis of this phenomenon was made by Fielding (1982). He recognized ‘counterurbaniza- tion’ to be taking place in a settlement system where there exists a negative relationship between migratory growth rate and urban status (Figure 9.1b). Drawing primarily on his case study of France along with statistical analyses of several other countries, Fielding visualized this new pattern as the outcome of the progressive move- ment of the strongest growth down the settlement hierarchy and anticipated its coming fully into being in Western Europe during the 1980s (Figure 9.1c). Thus presented as a revolution in settlement pattern trends linked to a societal shift from industrial to post-industrial eras, the nature and significance of this development has been widely disputed over the past two decades, as we will see below.

Besides these two usages of the term ‘urban- ization’, there is a third application which over- laps the other two to a certain extent but forms a distinctive element of an alternative way of making sense of recent trends – the ‘stages of

Urbanization 145

urban development’ model. One key feature of this approach is the idea that a city has a life cycle which takes it from a ‘youthful’ growing phase through to an ‘older’ phase of stability or decline, as the benefits of the initial investment are

progressively exhausted or the original locational advantages of the site become less relevant and are changing (Birch, 1971; Rust, 1975; for a review of this literature, see Roberts, 1991). A second feature incorporated into this approach is

146 Handbook of Urban Studies (a) (b)

(c)

Net mig

rati on

size

+

0

-

Net mig

rati on

size

+

0

-

Net mig

rati on

size

+

0

-

1950s 1960s 1970s 1980s

Figure 9.1 Relationship between migration and settlement size: (a)urbanization; (b) counterurbanization; (c) a possible sequence for WesternEurope, 1950–80 (Fielding, 1982: 8–10) Reprinted with kind permission fromElsevier Science Ltd.

Time

I II III IV V VI

Early primate city stage Intermediate primate city stage Advanced primate city stage Early intermediate city stage Advanced intermediate city stage Small city stage

U Urbanization PR Polarization reversal CU Counterurbanization

Primate city category Intermediate size city category Small size city category

I

N e t

m ig

ra ti

o n

+

0

-

II III V VIIV

U CUPR U Concentration Deconcentration

First cycle of urban development

Cleanbreak

Figure 9.2 The concept of differential urbanization (Geyer and Kontuly, 1993: 165)

the identification of the phasing of development through an examination of the internal patterning of growth, distinguishing between the main built-up ‘core’ of a city and its ‘ring’ or commuting hinterland. This approach was devised by Hall (1971) who suggested a four-stage model of metropolitan-area development, begin- ning with a period of centralization whereby people become more concentrated in the core at the expense of the ring, continuing with periods of relative and then absolute decentralization in which the core grows less rapidly than the ring and then experiences absolute loss of population to the ring, and ends up with a stage in which the metropolitan area as a whole moves into overall decline because the core’s loss becomes greater than the ring’s again. An additional feature added to this approach subsequently by Klaassen et al. (1981) is the idea of a recurring cycle; namely that, after the phase which Hall (1971: 118–19) refers to as ‘decentralization in decline’, there follows a process of reconcentration which leads on to a second cycle beginning with renewed growth overall and centralization within the core.

The main features of this cyclic model, incor- porating extra elements introduced by Berg et al. (1982, 1987), are illustrated diagrammatically in Figure 9.3. It consists of four ‘stages of urban development’ based on whether the urban region or ‘agglomeration’ (core and ring together) is experiencing overall growth or decline in popula- tion numbers and on whether it is the core or the ring that is performing the more strongly. Each of

the four stages is then subdivided into two phases based on the switching of either core or ring between gain and loss (which determines whether centralization and decentralization is absolute or merely relative), giving a sequence of eight phases altogether.

In the present context, the most important point is that the term ‘urbanization’ is restricted to the first of the four stages, referring to the situ- ation in which people are becoming more concen- trated in a single urban region while that urban region is growing overall. Beyond this, a whole urban system can be deemed to be urbanizing if the majority of its urban regions are in this ‘urbanization’ stage. The next two stages, however, are clearly meant to represent move- ments away from what might be considered to be ‘pure’ urbanization, with the ‘suburbanization’ stage involving decentralization within the urban region and the ‘disurbanization’ stage signifying decentralization beyond the urban region, a process which Robert and Randolph (1983) have termed ‘deconcentration’ to distinguish it from within-urban-region ‘decentralization’ and which equates broadly with Fielding’s definition of counterurbanization. Fourthly, the ‘reurbaniza- tion’ stage is associated with the slowing of urban-region decline, which is initiated by the core and followed by the ring and thus involves a process of renewed centralization. Finally, according to the ‘cyclic’ element of this model, this leads through into renewed urban-region growth and to a new period of ‘urbanization’.

Urbanization 147

Core

P o

p u

la ti

o n

c h

a n

g e

+

0

–

Agglomeration

Ring

Sequence Phase Stages

1 2 3 4 5 6 7 8 AC RC RD AD AD RD RC AC Urbanization Suburbanization Disurbanization Reurbanization

Figure 9.3 The stages of urban development model: classification type – A, absolute; R, relative; C, centralization; D, decentralization (after Klaassen etal., 1981: 18 and Berg et al., 1982: 36)

This formulation of urban development trends, deriving from the early 1980s, is useful in summa- rizing what was at that time generally held to be the way in which urban population patterns had been developing over the post-war period in Western Europe and, to a considerable extent, across the more developed world as a whole. The 1950s are widely regarded as being dominated by urbanization tendencies under all three defini- tions outlined above, with a shift of population from rural to urban areas, with an increasing concentration of urban population in the larger cities and (more arguably) with a majority of urban regions seeing their cores growing faster than their rings. By contrast, on average across the most urbanized countries (those in Northern and Western Europe, North America and Australasia), the 1960s are seen as the heyday of ‘suburbanization’, associated with the rapid growth of population in the commuting hinter- lands beyond the main built-up areas, or outside the ‘central city’ in US parlance, but this is also the period which saw the first significant weak- ening of Fielding’s ‘urbanization relationship’ as the growth of some of the very large cities began to falter. The 1970s are now generally considered to be the ‘decade of counterurbanization’, partly because it was during this period that the larger metropolitan areas in many countries witnessed a marked reversal in net migration flows but also partly because even before the end of the decade there were signs of urban revival, notably resi- dential gentrification, which could possibly herald the onset of ‘reurbanization’.

At the same time, however, this version of events leaves a number of questions largely unre- solved. In the first place, how meaningful is it to talk nowadays in terms of ‘suburbanization’, as if it is something separate from urbanization? Secondly, what is the real significance of ‘coun- terurbanization’ and, in particular, does it consti- tute a fundamental reversal of traditional urban- ization processes leading to the emergence of a radically different ‘post-industrial’ settlement pattern or is it no more than a temporary stage in the life cycle of an urban region and its wider urban system? Thirdly and linked to this latter point, how much confidence can be placed in the idea of a ‘reurbanization’ stage and, even more important, in the supposition that this is not merely consolidation in decline but is the precursor of a new lease of life for the large city which introduces a new stage of ‘urbanization’ at the beginning of a second life cycle for that city (in the terminology of Berg et al., 1987) and leads to a new ‘primate city stage’ for the urban system as a whole? (in the terminology of Geyer and Kontuly, 1993). The remainder of this chapter draws on the literature that addresses these ques- tions most directly and attempts to assess the

extent to which consensus has been reached and what this means for the way in which urban areas in advanced economies are considered likely to develop in the future.

SUBURBANIZATION Suburbanization is by no means a new phenom- enon, with its origins traceable in the building of large homes by more successful entrepreneurs on the outskirts of the burgeoning centres of industry and commerce over a century and a half ago. It became significant as an urban feature during the latter half of the nineteenth century as cheap forms of mass transit loosened the ties between home and workplace for those with secure jobs and relatively ‘social’ hours of work. Subsequently, it accelerated as a result of further changes in transport and personal wealth, surging ahead especially strongly around mid century, to the extent that in the ‘stages of urban develop- ment’ model it is seen to have become the domi- nant element of urban change in the 1960s not only in the Anglo-American world but widely across Western Europe. Since then, however, the suburbs have, in one sense, been eclipsed by other developments, and at the same time have begun to undergo a revolution which appears to render the original terminology obsolete.

The term ‘suburb’ carries connotations of being something less than urbs, the city: ‘usually residential or dormitory in character, being dependent on the city for occupational, shopping and recreational facilities’ (Johnston, 1981: 331). The process has been very largely powered by the negative aspects of city cost, congestion, grime and squalor. Originally, it was dominated by the housing needs and aspirations of the family, with the emphasis very much on the male breadwinner and on healthy space in which the mother could devote her time to bringing up her children and providing for her husband. For some residents, the housing plot would be the sole concern, as everyday supplies would be delivered by firms or fetched by servants, waste would similarly be carted away and children sent off to boarding school. Even when suburban residence broadened beyond the preserve of the upper middle-class, however, the range of facilities would often be very limited: piped water (and possibly gas) supply, mains drainage in some areas (but often the cesspool), and in due course electricity, paved roads, primary school and retail outlets for everyday items (such as newsagent, grocer, post office and perhaps a garage for car maintenance), with the best services being available in strips along radial routes out of the city. The success of the suburbs in providing an escape from the

148 Handbook of Urban Studies

problems of city living as well as allowing people to avoid the limited life chances of the deep coun- tryside was, however, considered as definitely suboptimal by a number of commentators who decried the emerging compartmentalization of urban life, notably Ebenezer Howard, yet his ‘garden city’ idea – where residents would have easy access to the full range of urban facilities including jobs – had little impact until the 1950s.

Since the 1950s, suburbanization has come to be viewed more in terms of urban (or metropol- itan) decentralization, reflecting the shift in patterns of development from the mainly lateral expansion of the city’s built-up ‘core’ to much deeper penetration of house building into an extensive commuting hinterland or ‘ring’. This approach to urban analysis and official statistics was adopted in the USA for the 1960 Census in the form of the Standard Metropolitan Area (SMA), which was defined on the basis of a municipality of above a certain population – called ‘central city’ – together with contiguous counties having metropolitan character linked to the central city by a certain level of commuting – referred to as ‘suburbs’ in the majority of the literature (Shryock, 1957; US Bureau of the Budget, 1964). In England and Wales this lead was followed by Hall (1971) who identified 100 Standard Metropolitan Labour Areas, each with its ‘core’ and ‘ring’ – developments of which were used for the analysis of the 1971 Census (Spence et al., 1982) and the 1981 Census (Champion et al., 1987; Coombes et al., 1982). Similar app- roaches have been adopted by studies of urban change in Europe, most notably the ‘metropolitan areas’ defined by Hall and Hay (1980) and the ‘functional urban regions’ used by Berg et al. (1982), while in most other advanced economies some form of metropolitan area has been identi- fied for academic or official purposes, normally distinguishing between a central core and surrounding commuting zone. A common feature of these approaches to studying ‘suburbanization’ is that the ‘central city’ or core incorporates the

older suburbs, certainly those developed immedi- ately adjacent to the main urban area before the First World War and often much of the motor- based extensions of the inter-war period, so what is being studied in this way is the more recent and far-flung manifestations of this centrifugal process.

Defined as such to include the physically sepa- rate settlements in the commuting hinterland that are sometimes referred to as ‘exurbs’, suburban areas now appear to dominate life in many coun- tries. This would seem to be particularly true in the USA, where non-central-city parts of metro- politan areas accounted for 46 per cent of the nation’s population in 1990 – up from 31 per cent in 1960, 38 per cent in 1970 and 45 per cent in 1980 (Frey, 1993a). There the 1960s do seem to have represented a landmark in this process: it was at this time that the suburban population overtook that of the central cities, its share of America’s metropolitan population rising from 49 to 55 per cent on the basis of the Census defini- tions used in 1960 and 1970 respectively. At the same time, however, it can be seen that, while perhaps the major transformation from urban to suburban took place at this time, local metropol- itan decentralization is a continuing feature of population redistribution in the USA.

Similar conclusions can be drawn from studies tracing the growth of suburban populations in other countries since mid century. For France, for instance, Boudoul and Faur (1982) have docu- mented the particularly strong growth in 1968–75 of the parts of the unités urbaines lying outside the central communes of most of the largest cities. In Italy during the period 1971–81 the rings of the five major metropolitan areas recorded substan- tial population gain while their core populations declined or grew only slowly (Dematteis and Petsimeris, 1989). Data for Britain’s Functional Regions (Table 9.3) reveal a strengthening and widening of decentralization from cores since the 1950s, with greatest pressures initially on the rings (the primary commuting fields) and subsequently

Urbanization 149

Table 9.3 Great Britain: population change, 1951–91, by Functional Region zone

Rate for decade (%) Deviation from GB rate

–––––––––––––––––––––– –––––––––––––––––––––– Zone type 1951–61 1961–71 1971–81 1981–91 1951–61 1961–71 1971–81 1981–91

Great Britain 4.97 5.25 0.55 2.50

Core 3.98 0.66 –4.20 –0.09 –0.99 –4.59 –4.75 –2.59

Ring 10.47 17.83 9.11 5.89 5.50 12.58 8.56 3.39

Outer area 1.74 11.25 10.11 8.85 –3.23 6.00 9.56 6.35

Rural area –0.60 5.35 8.84 7.82 –5.57 0.10 8.29 5.32

Based on population present for 1951–81 and usual residents for 1981–91.

Source: Calculated from Population Censuses. Crown Copyright reserved

on the ‘outer areas’ (the secondary hinterlands) and the more rural labour market areas beyond.

Though comparisons between countries are bedevilled by variations in areal definitions, the differential scale and pace of decentralization tendencies in Europe can be gauged with reason- able confidence from three comparative studies. According to Hall and Hay (1980), for 15 coun- tries taken as a whole, the proportion of popula- tion living in rings rose during the 1960s after losing ground to the cores in the previous decade and grew even more sharply in the first half of the 1970s. In more detail, however, they found a clear contrast between Great Britain, where the rings’ proportion was already growing in the 1950s, and Western and Southern Europe (made up prima- rily of Belgium, France, Italy, the Netherlands and Spain) where the rings’ proportion did not begin to increase until the early 1970s. In fact, Northern Europe (Denmark, Norway and Sweden) was the only region to see the onset of above-average ring growth in the 1960s.

Meanwhile, using the ‘stages of urban devel- opment’ approach, Berg et al. (1982) have demon- strated very effectively the progress of ‘suburban- ization’ of the European urban system as a whole and of a dozen separate countries, drawn from both Western and Eastern Europe. Again, across the whole system, it is the 1960s that is definitely the decade when this process was dominant, accounting for 73 per cent of 185 Functional Urban Regions (FURs) in the study, compared with 50 per cent in the previous decade and 63 per cent for 1970–75. For individual countries, there was a profound shift in dominant stage reached in the 1950s, when only Belgium, Great Britain and Switzerland could be considered to have reached the point where the ‘suburbanization’ stage was dominant to that found in the early 1970s, when only Bulgaria, Hungary and Poland had not yet reached it fully.

Clearly, while originating in the nineteenth century, suburbanization emerged as the domi- nant process of population redistribution during the post-war period, as the exodus from the older urban cores gathered pace. There is, however, ample evidence that over time the process has grown to become very different from its original nature and has more recently been manifesting itself in novel forms. One distinction has already been drawn above, in terms of the geographical scale shifting from lateral extension of the main built-up area to decentralization over a broad commuting field. A second, equally important development is the progressive disappearance of the ‘sub’ element of the process, as the arrival of residential population has since the 1950s been followed by the decentralization of industrial, commercial and retail activities and lately by the growth of office and high-tech sectors, the latter

being seen as the ‘third wave’ of suburbanization in the USA (Cervero, 1989). Over the past ten years American commentators have been coining terms suggestive of a revolution in the form of the city, as the ‘new suburbanization’ (Stanback, 1991) – with its ‘suburban downtowns’ (Hartshorn and Muller, 1986) and ‘edge cities’ (Garreau, 1991) – is increasingly being seen as challenging the central cities and threatening to turn the traditional metropolitan area inside out.

In short, the past quarter of a century has seen nothing less than the urbanization of the suburbs, or – in the far-sighted words of Birch (1975: 25) – a transformation ‘from suburb to urban place’. While this process appears to have progressed further in the USA than elsewhere (including Canada) and much debate is taking place even in America about the significance of these ‘new suburban landscapes’ (Bourne, 1989, 1994), there is no doubt that the distinction between ‘urban’ and ‘suburban’ has become increasingly blurred across the more developed world. This raises serious questions about the validity of studying ‘suburbanization’ as a distinctive phenomenon, whether merely in the guise of urban decentral- ization or as part of the ‘stages of urban develop- ment’ model, except insofar as the use of geographical boundaries identified in an earlier period can be considered useful for monitoring population shifts subsequently.

COUNTERURBANIZATION It is not only at the interface between city and suburbs that blurring has occurred, but also at the fringes of the metropolitan areas and functional urban regions, as the ‘suburban frontier’ has moved outwards to encompass wide tracts of land and embrace formerly free-standing towns and villages. While there is a remarkable degree of correspondence between the various conceptual- izations of inter-urban deconcentration, there continues to be heated discussion about whether it is really a qualitatively different phenomenon compared with suburbanization or an extension of it. Much hinges on the extent to which the growth of smaller cities and towns is the result of overspill to places that are new appendages of metropolitan territory, as opposed to arising from residents and employers seeking out these places to take advantage of their inherent characteris- tics, notably their smaller size and associated economic, social and environmental benefits. As we shall see, a central problem is that in more heavily populated regions it is very difficult, if not impossible, to distinguish between these two explanations purely by the examination of popu- lation redistribution patterns, but neither has

150 Handbook of Urban Studies

more in-depth investigation of the processes as yet managed to resolve this fundamental question to general satisfaction. This apparent impasse raises the possibility that, as with urburbaniza- tion, it may be that the traditional ways of looking at these developments are now obsolete or were even misconceived in the first place.

The degree of correspondence between the three main conceptualizations of urban change introduced earlier lies in the following: whatever the actual terminology used, urban areas are defined in functional terms to include both core and ring, the process is viewed as being system- wide and, thirdly, larger places in the urban system lose out to smaller ones. As noted above in Figure 9.1, this approach is set out most explicitly in Fielding’s (1982) definition of ‘counterurban- ization’ as the situation in which the net migratory growth rate of a place is inversely related to its size and urban status, following a period of tran- sition as the relationship switches from the posi- tive ‘urbanization’ correlation. Similarly, in Geyer and Kontuly’s (1993) concept of ‘differential urbanization’, the system goes through a transi- tional period of ‘polarization reversal’, when the intermediate-sized city category overtakes the primate city in migration growth, before reaching the ‘counterurbanization’ stage in which the small-city category in its turn becomes the strongest growing (Figure 9.2). The equivalent in the ‘stages of urban development’ model – the ‘disurbanization’ stage – is broadly the same, except that the stage reached is initially deter- mined for individual urban regions and also that in this formulation not only does the urban region as a whole have to be in decline but decentraliza- tion must also be taking place from core to ring (Figure 9.3). In this approach the settlement system as a whole is reckoned to be in the ‘disur- banization’ stage, when the latter accounts for a majority of urban regions, including no doubt virtually all the larger cities (Berg et al., 1982).

Judging by the results of the many studies of counterurbanization, there would seem to be much evidence of the emergence of this negative relationship between growth and size of urban region, with the 1970s being a relatively common period across the more developed world. It was in the early 1970s that the ‘rural population turn- around’ was identified in the USA, with the non- metropolitan growth rate moving above the national figure (Beale, 1975; Berry, 1976b), while within metropolitan America the small metropol- itan areas were growing faster than the medium ones and much faster than the large ones (Frey, 1989). Fielding’s (1982) results on Western Europe revealed counterurbanization relation- ships in the 1970s for Belgium, Denmark, France, the Netherlands, Sweden, Switzerland, the UK and West Germany. Champion et al. (1987) iden-

tified a strong negative fit between 1971–81 growth rate and urban status for groupings of Britain’s 280 Local Labour Market Areas. Geyer and Kontuly (1993) place France very firmly in the ‘small city’, i.e. counterurbanization, stage on the basis of its 1975–82 Census results.

On the other hand, other studies do not provide such a uniform picture and it would also appear that several countries were not experi- encing counterurbanization at this time. Berg et al.’s work in Europe (1982) found that, while the proportion of urban regions in the ‘disurbaniza- tion’ stage rose steadily between the 1950s and the 1970s, it was still accounting for less than one- fifth of cases by 1970–5. Moreover, their classifi- cation of European countries by dominant stage of urban development for 1970–5 put only Belgium unequivocally in the ‘disurbanization’ stage, with Switzerland, the Netherlands and Great Britain rather evenly poised between ‘suburbanization’ and ‘disurbanization’. According to Fielding (1982), Austria, Ireland, Italy, Norway, Portugal and Spain were still char- acterized by a positive or insignificant relation- ship in the 1970s. Similarly in Japan, while there was a major reduction to net migration to metro- politan areas between the 1960s and the 1970s, no clear and sustained reversal took place at this time (Tsuya and Kuroda, 1989). In related work on core/periphery differentials in migration, Cochrane and Vining (1988) identified a ‘Periphery of West Europe and Japan’ type, where differentials had merely narrowed, in contrast to a ‘Northwest Europe’ type, where periphery and core had switched over by the 1970s. A relatively cautious interpretation was also being put on events in Canada and Australia, where Bourne and Logan (1976: 136) – rather than finding clear signs of counterurbanization – concluded that ‘Urbanization . . . appears to have entered a new period’, with internal migration streams shifting away from the two dominant metropolitan areas in each country toward medium-sized cities and to smaller centres just outside the metropolitan regions.

This lack of uniformity in the 1970s can be explained in part by the fact that, because of differences in both history and geography, coun- tries are clearly not all at the same stage of urban development at any particular time, nor indeed are all the regional urban systems within a single country. Mention has already been made of the uneven progress of urbanization across the more developed world and also of the progressive diffu- sion of the suburbanization tendency across Europe, so nothing less should be expected in relation to counterurbanization. As regards within-country variations, research on the USA traditionally draws a distinction between urban patterns in the North and those in the South and

Urbanization 151

West (for example, Frey, 1989), while in Western Europe perhaps the most marked regional contrast is the one between the heavily urbanized northwest of Italy and the much more rural Mezzogiorno (Coombes et al., 1989; Dematteis and Petsimeris, 1989). It is therefore not surprising to find some evidence from more recent studies that shows the migration reversal occurring later than the 1970s. Updating of Fielding’s work on Western Europe (Champion, 1992) reveals that Austria, Ireland and Italy switched from a concentration to a deconcentra- tion relationship between the 1970s and the early 1980s, and the counterurbanization pattern was found to intensify in Belgium, France and West Germany in the first half of the 1980s (see also Kontuly and Vogelsang, 1989).

Much more surprising for most commentators was the discovery by the mid-1980s that many of the leading counterurbanization countries were experiencing a cutback in this process or indeed some form of reversal of the original migration turnaround. The USA provided the most notable example, with not only the return of the non- metropolitan areas to below average growth in the 1980s but also the resurgence of the largest metro- politan areas to such an extent that by the latter half of the decade their growth rate exceeded the average of those with under a million people, thus reintroducing the pattern of the 1960s (Frey, 1993b). In mainland Europe some countries which had a clear counterurbanization relation- ship in the 1970s either switched back in the early 1980s to a positive relationship between migra- tion growth and city size (Sweden) or saw a marked decline in counterurbanization (Denmark, the Netherlands), while others experi- enced a strengthening in their urbanization rela- tionship (Norway, Portugal). Moreover, as the 1980s progressed, in common with the US experi- ence, this slowdown in urban deconcentration and/or move towards greater concentration in large cities intensified in many cases (Finland, the Netherlands, Portugal, Sweden, Switzerland), and even Austria, Belgium and West Germany reverted to concentration in the second half of the decade (Champion, 1992). London, too, staged an impressive recovery from its high rates of population decline of the 1970s, but in terms of internal migration the dominant picture across Britain was of continued, though somewhat lower, rates of deconcentration (Champion, 1994).

This rather general swing away from urban deconcentration in the 1980s has presented a clear challenge to the understanding of recent urban- system trends, though at the same time it provides some opportunity for relating both upturn and subsequent downturn to possible causative factors. It is particularly problematic for those

many commentators who saw counterurbaniza- tion as heralding a shift to a new ‘post-industrial’ form of settlement based on medium-sized and smaller cities and towns and believed that they had uncovered a host of valid explanations for it (Champion, 1989: 236–7). Clearly the experience of Western Europe in the 1980s did not generally bear out Fielding’s idea that this decade would see the fulfilment of the switch from an urbanization to a counterurbanization relationship between migration and settlement size, as depicted in Figure 9.1. Even more dramatic is the way in which the events in the USA appear to refute Berry’s (1976b: 17) bold pronouncement that, ‘A turning point has been reached in the American urban experience. Counterurbanization has replaced urbanization as the dominant force shaping the nation’s settlement patterns.’

For the supporters of the counterurbanization thesis, there are at least three potential ways of making sense of the 1980s: that the process had run its course, that a chance combination of factors at that time made the decade anomalous, and that the forces of concentration and decon- centration are subject to some fairly regular cyclic rhythm.

The first interpretation does not survive the most cursory scrutiny. While it is logical that, unlike urbanization, counterurbanization is ulti- mately self-defeating because it erodes the differ- ences in settlement size which power it, the scale of deconcentration experienced during the 1970s barely dented these differences. In Britain, for instance, the seemingly large-scale loss of popu- lation sustained by the 20 Dominant Functional Regions between 1971 and 1981 reduced their share of the national population by less than 3 percentage points, from 39.9 to 37.2 per cent, while in 1981 the smallest 52 of the 280 Local Labour Market Areas averaged only 54,000 people each compared to the overall LLMA average of 194,000 and together they accounted for only one-twentieth of total population but fully one-third of national territory (Champion et al., 1987). Much more fundamental change than this had been expected from the counterurban- ization phenomenon.

By contrast, the second interpretation carries considerable weight. It is not difficult to identify features of any period that distinguish it from other periods. It has been suggested that the 1980s were characterized by various events that caused a temporary interruption to the fuller develop- ment of the new deconcentration pattern emerging in the 1970s. These include the strong growth of jobs in finance and business services in large cities, downturns in agricultural and energy prices, rationalization of both public sector and private consumer services in more rural areas, the switch of government policy towards inner city

152 Handbook of Urban Studies

rejuvenation and the ageing of the baby boom cohorts into city-loving young adults (Champion, 1992). Providing empirical support for this inter- pretation is some evidence of a resurgence of urban deconcentration in the first half of the 1990s, notably the revival of non-metropolitan growth rates in the USA (Johnson and Beale, 1994; Nucci and Long, 1995), an acceleration of the metropolitan exodus in Britain (Champion, 1996) and the ‘turnaround after the turnaround after the turnaround’ in the Netherlands (Van Dam, 1996).

The third line of reasoning – that there is some cyclic rhythm in urban change superimposed on longer term trends – represents an extension of the previous one, albeit on a more formal basis. Support for a cyclic explanation emerged almost as soon as the first signs of the further reversal appeared, notably arising from Cochrane and Vining’s (1988) monitoring of migration between core and peripheral regions. Mera (1988) suggested that, as a country reaches an advanced level of development, the pattern of migration becomes susceptible to a host of factors other than economic, citing political and technological factors as contributing to the return to clear urban concentration in Japan in the 1980s. Champion (1988) gave particular prominence to demographic factors, notably the generation- based cycles arising from baby boom and bust. Berry (1988) set the latest trends in the context of two centuries of change in the USA, demon- strating that the strength of urban growth follows a long-wave cycle that could include periods of metropolitan migration reversal. Building on this approach, Champion and Illeris (1990) recommended that research should recognize three separate groups of factors: those pulling towards concentration, those leading towards dispersal, and those that may have different geographical outcomes at different times. The latter include the economic forces of Frey’s (1989) ‘regional structuring’, while the first two together are broadly equivalent to Frey’s ‘popu- lation deconcentration’ explanation and fluctuate in their relative strengths over time and thus in their net effect on movement up or down the settlement hierarchy.

Besides these various ways of making sense of the past quarter of a century within a ‘counterur- banization’ perspective, there are two other main sets of reactions to the developments of the 1970s. One is to claim that there is, and has been, no such thing as counterurbanization, with Gordon (1979) being the first of many to criticize the idea of a ‘clean break’ from traditional urban- ization and associated urban concentration. To the extent that apparent population dispersal has been taking place, this is seen either as suburban- ization and local metropolitan decentralization

on a wider scale outpacing the redefinition of metropolitan areas and urban regions or as the embryonic stages in the growth of major new metropolitan concentrations that will come to dominate the national settlement pattern in due course. There is plenty of evidence in support of the ‘overspill’ argument, most recently and comprehensively in the US context by Nucci and Long (1995) and also relating to Australia by Hugo (1994), but so far it is not clear whether the new metropolitan areas are destined to remain relatively small and independent in daily-urban- system terms or to develop into the mega- lopolises of the future.

The other approach is conceptually the neatest, as the general switch away from urban deconcen- tration in the 1980s is exactly what is predicted in the ‘stages of urban development’ model. Rather than forming the long-term future tendency as suggested in the ‘counterurbanization’ model of Figure 9.1, the equivalent ‘disurbanization’ phenomenon is seen as being only a temporary stage in the development of any single agglomer- ation and urban region, being followed by a stage of ‘reurbanization’ before the hypothesized onset of a second cycle (Figure 9.3). Similarly, the ‘differential urbanization’ approach (Figure 9.2) anticipates the fading of the net migration gains of the small city category and the re-emergence of the primate city category as the fastest growing, ushering in a new period of urbanization. The next section trawls the literature in order to discover the extent to which there has been a recovery by the larger cities and to try and estab- lish the theoretical and longer-term significance of any such change.

REURBANIZATION The prime issue for this section to tackle concerns the nature and significance of the recovery of larger urban areas and their inner cities since the period of widespread population exodus in the 1970s. As just mentioned, one hypothesis is that it constitutes another natural stage in the ‘life’ of individual urban areas and of the urban system as a whole. Alternatively, adopting the analogy of a dying star, it may merely represent a period of consolidation, as an urban area fits itself to new circumstances, or it may lead to some form of self-reinforcing ‘implosion’ as a ‘black hole’ develops, sending shock waves across the rest of the system. As we will see below, the originators of the cyclic model were extremely cautious about the prospects for large-scale reurbanization and the evidence available so far gives only rather limited support to this conceptualization. More generally, however, a number of reasons have

Urbanization 153

been put forward to explain the recent recovery of larger cities, raising the question as to how long- lived this process is likely to be and whether it is significant enough to initiate a new round of urban growth and subsequent decentralization.

Compared to the great attention which has been given to it subsequently, the originators of the ‘stages of urban development’ model proposed the ‘reurbanization’ stage in a remark- ably tentative manner. In the words of Klaassen and Scimeni (1981: 16), ‘It is not impossible that . . . the large town, with its economic and cultural potentials, will in the course of time win back the ground it has lost, that people once more flock towards the city, and that the town starts upon a new period of bloom, which may lead to a new cyclus’ (emphasis added). They expressed reservations about the term itself, pointing out that this ‘reurbanization’ stage is one of contin- uous urban-region decline and only towards the end of it is the core meant to move into absolute population growth: ‘One might wish to reserve the term “reurbanization” for the first phase of a second urban cycle’ (1981: 26). They admitted that there was so far no evidence of a second cycle, that only a few cities were as yet to be found in this fourth stage of the first cycle and that these were the wrong ones from the model’s point of view, all being in the small-size bracket rather than the largest cities in the most advanced stages of the cycle (1981: 26–7). Finally, while they offered the possibility that this stage might come about through market forces, their main hopes were pinned on purposeful government action – as was made even clearer in the policy agenda set out in the follow-up study which explicitly states: ‘Reurbanization must be brought about fast and resolutely’ (Berg et al., 1982: 44).

More recent searches for evidence of places entering the reurbanization stage have yielded what can only be described as mixed results, at best. The most systematic attempt made so far is that by Cheshire (1995), focusing on 241 FURs with at least 330,000 inhabitants in the European Union (see Figure 9.4). On the one hand, according to Cheshire (1995: 1058), among these large FURs ‘the regular onward march of decen- tralisation appears to have faltered and, in northern Europe, it has halted, even reversed’. The proportion of urban cores gaining popula- tion was found to have recovered somewhat in Northern Europe (defined as Belgium, Denmark, the Netherlands, the UK and West Germany), reaching 47 per cent in 1981–91 after slumping to 38 and 22 per cent in 1971–5 and 1975–81 respec- tively, and some recovery from the later 1970s was also evident for the France/Northern Italy group. On the other hand, these changes were not enough to produce any significant increase in the number of FURs that could be classed as being in

the reurbanization stage, even in northern Europe. In fact, here the proportion of FURs in this stage fell from 15 per cent in the later 1970s to only 6 per cent in the 1980s, and instead the overall effect of the changes was to push the urban systems back into the suburbanization stage. Moreover, it was mainly the smaller FURs – particularly those with ancient cathedrals and universities – that were participating in this stage, rather than the larger and older urban regions that were meant to lead the process. The UK’s experience was fairly typical, with this stage accounting for only four of its 36 large FURs: Glasgow, Oxford, Cambridge and Canterbury, with only Glasgow conforming to the model’s expectations.

Nevertheless, across the advanced economies there is now much case study evidence of a marked change of fortunes for large cities from the high levels of population loss sustained in the 1970s. In the case of Glasgow, for example, Lever (1993a) has shown how the whole Strathclyde region switched from substantial loss in the 1970s to small gain in the 1980s, brought about almost entirely by the reduction in Glasgow City’s annual rate of loss from 22,000 to less than 1,000. Greater London’s recovery in the 1980s was even more impressive, and Inner London’s particularly so (Champion and Congdon, 1988). The rate of population loss for all 280 of Britain’s urban cores fell from 4.2 per cent in 1971–81 to a mere 0.1 per cent for 1981–91, while the growth rate of the rings fell back substantially, particularly in comparison with the national population trend (see Table 9.3). In Scandinavia, Copenhagen saw the stabilization of its central-city population in the 1980s (Illeris, 1994), while according to Nyström (1992) the cities of Helsinki, Oslo and Stockholm experienced actual growth, though not as strongly as in the suburbs except in the case of Stockholm and even here the pattern varied during the decade and was again favouring the outer areas by the early 1990s (Borgegård et al., 1995). The overall impression conveyed by these studies is of a reduced level of variation in popu- lation growth rates between places – a pattern which is consistent with the shrinkage in regional differentials in net migration rates across Europe since the 1970s observed by Champion et al., (1996).

The picture appears similar in the New World, though trends there are superimposed on a generally more dynamic demographic situation. In the USA, as noted above, the 1980s saw the re- emergence of the larger metropolitan areas as the fastest growing element of the urban landscape, with Frey and Speare (1992: 133) referring to ‘the reurbanization of the 1980s in contrast to the 1970s pattern’. Overall, metropolitan areas with one million or more people grew by 12 per cent

154 Handbook of Urban Studies

Urbanization 155

1981–91 1975–81 1971–75 1961–71 1951–61

Northern Europe

Sequence Phase Stages

1 2 3 4 5 6 7 8 AC RC RD AD AD RD RC AC Urbanization Suburbanization Disurbanization Reurbanization

1981–91 1975–81 1971–75 1961–71 1951–61

France and Northern Italy

1981–91 1975–81 1971–75 1961–71 1951–61

Southern Europe

0 10 20 30 40 50 60 70 80 90 100

0 10 20 30 40 50 60 70 80 90 100

0 10 20 30 40 50 60 70 80 90 100

Figure 9.4 Change in frequency distributions of Functional Urban Regions bystage of urban development, 1951–91, for three European regions; see Figure9.3 for definitions of stage and phase (after Cheshire, 1995, Table 1: 1051)

in the 1980s compared with 8 per cent in the previous decade. Much of this growth was powered by the West and South, but even the North’s large metropolitan areas switched from a decline of 0.9 per cent between 1970 and 1980 to a 2.8 per cent increase in the 1980s. New York (defined on the CMSA basis to include 18 million people in 1990) was particularly impres- sive, with a switch from a 3.6 per cent loss to a 3.1 per cent gain between the two decades, and there were also strong upward shifts for Philadelphia and Boston but not for Chicago and Detroit. Within metropolitan America overall, the annual average growth rate for central cities rebounded from 0.09 to 0.64 per cent between the two decades and the suburban growth rate fell back somewhat, but in the North-East and Midwest central cities still seemed to be struggling, their average annual growth rate going up from –1.09 to +0.03 per cent in the former and up from –0.85 to –0.22 per cent in the Midwest (Forstall, 1991).

Australia and Canada also provide evidence of stengthening metropolitan-area and inner-city growth in the 1980s. In the case of the former, Melbourne appears particularly dynamic in the latter half of the decade, with an overall popula- tion increase of 190,000 in 1986–91 (not far short of the 208,000 increase for the previous ten years), Brisbane grew by 185,000 in the five years compared to 195,000 for the previous decade, and Perth 149,000 as opposed to 218,000, while Sydney and Adelaide both continued growing at close to their earlier rate (Hugo, 1994). For Canada, the 1980s witnessed a switch in the focus of population growth back to the central provinces, notably the Toronto CMA which added almost 0.9 million people, or 30 per cent, between 1981 and 1991 – twice the 1970s growth of 15 per cent and just under 0.4 million (Bourne and Olvet, 1995). In both cases, it was the outer areas that contributed most to these gains, but inner city recovery was also common. In the case of Toronto, for instance, seven- eighths of the CMA’s growth took place outside the administrative Metro area – adding over three-quarters of a million people there and almost doubling the 1981 population – and a further one-ninth was accounted for by the suburban parts of Metro Toronto, but the City of Toronto also saw a 6 per cent increase in population in the 1980s, a marked switch from its 16 per cent decline in the 1970s (Bourne and Olvet, 1995). More generally across Canada, inner cities – defined narrowly to include only the CBD and surrounding zone of mixed uses – shifted from previous decline into growth in the early 1980s (Ram et al., 1989, quoted by Bourne, 1991; Ley, 1992). A similar renaissance has been documented for Sydney, Melbourne and

Adelaide (respectively, Murphy and Watson, 1990; King, 1989; Badcock, 1991).

Two somewhat contradictory sets of conclu- sions would seem to flow from these empirical observations. On the one hand, there are wide- spread signs of renewed growth or reduced decline for larger metropolitan areas and urban regions and also of a population recovery for urban cores and their inner areas. On the other, however, there appears to be no evidence of suburban ring areas losing out to core areas – not even relatively, let alone in absolute terms whereby core growth takes place alongside suburban decline, as required of the later phase of reurbanization in the ‘stages of urban develop- ment’ model (see Figure 9.3). The New World experience of the 1980s and early 1990s would seem to echo Cheshire’s (1995) findings for Europe, namely that during this period there has been backtracking towards the pattern associated with the first phase of the suburbanization stage, involving renewed growth of both urban region and core but also with relative decentralization from core to ring.

Given the cyclic approach’s apparent failure to caricature recent events, it is necessary to examine the individual lines of explanation which have been put forward to account for them. The litera- ture falls into two rather separate camps, one dealing with the relative performance of urban regions as individual entities and the other focusing quite specifically on inner-city revitaliza- tion including gentrification. The former has already been introduced in terms of the coun- terurbanization debate, itself commonly being compartmentalized between the population deconcentration perspective and the effects of economic restructuring. In relation to the latter, much emphasis in the 1980s was placed on identi- fying ‘winners’ and ‘losers’ in the new global competition between cities. Studies of the USA (for example, Frey, 1987; Frey and Speare, 1992; Noyelle and Stanback, 1984) have shown that the greatest benefits have accrued to the larger metro- politan areas that serve as centres of finance, corporate headquarters and advanced services and to places specializing in information, high tech- nology or knowledge-based industries, together with certain consumer-oriented areas with tourism and retirement functions. In Europe, too, various studies have pointed to the growing concentration of corporate power in particular cities, demon- strated the growing distinctiveness of cities by functional type and measured ‘urban success’ for different levels of the settlement hierarchy (respec- tively Meijer, 1993; Kunzmann and Wegener, 1991; Parkinson, 1991; see also Illeris, 1992; Lever, 1993b). Similarly, according to Bourne and Olvet, ‘Canadian cities are becoming more unlike in terms of their socio-economic attributes’, leading

156 Handbook of Urban Studies

them to wonder whether there is still ‘such an object as a “Canadian” city’ (1995: 47).

As regards inner-city revitalization, there is now a voluminous literature on gentrification and other aspects of redevelopment (Smith and Herod, 1992). Perhaps the most fundamental point made there is that gentrification, as origi- nally defined in terms of the social upgrading of a neighbourhood through the rehabilitation of the existing housing stock, is only one element – and by the 1980s only a minor one – of a much broader process. As set out by Bourne (1992), residential revitalization on the ground comprises four main components: intensification through the construction of infill housing or the high- density redevelopment of older neighbourhoods, implantation through housing being inserted into existing high density commercial and insti- tutional districts, conversion of older non-resi- dential structures such as unused warehouses, and extension through the penetration of residential uses into formerly non-residential areas such as vacant railway, port and industrial sites. This last component has been responsible for some of the most impressive changes in inner city areas, particularly where associated with waterfront redevelopment schemes such as for New York, Toronto, Vancouver and London Docklands (see, respectively, Fainstein, 1994; Bourne and Olvet, 1995; Ley, 1981; Church, 1988).

There seems to be wide agreement about the types of reasons for the residential revitalization of inner urban areas in the 1980s (see, for instance, reviews by Bourne, 1992; Badcock, 1995; Lever and Champion, 1996). As envisaged originally by Klaassen et al. (1981) and Berg et al. (1982), in most cases government action played a key role in encouraging and facilitating the process, involving the winding up of previous dispersal policies and the disbursement of often huge amounts of public money into programmes ranging from grants for improving individual homes and for local economic development incentives to massive investment in land reclama- tion and upgrading of infrastructure. But market forces are also seen to have played a very impor- tant role. Chief among these, as for the recovery of the whole metropolitan area, is economic restructuring, with inner-city deindustrialization running its course and with the rapid growth at this time of advanced producer services that seem less prone to decentralization than manufacturing and personal services. The greater availability and falling relative cost of space in inner areas has been cited frequently, both in terms of the ‘rent- gap’ thesis in explanations of the onset of gentri- fication and more generally in recognition of the vacant and underused land and buildings left by previous decades of population exodus and busi- ness closures and relocations, contrasting with

rising land costs and congestion in more suburban areas. Higher energy costs since the mid-1970s have also been cited as the cause of more high income people choosing to live close to CBD-based (central business district) jobs in large cities in preference to a long commute.

Demographic developments and related social changes have also been widely quoted in explana- tions of residential revitalization in the inner areas of large cities. Increased immigration, whether arising from growing labour market demands or as a result of the huge rise in asylum- seeking in the 1980s and 1990s, has not only been focused on the largest urban regions of destina- tion countries but has normally impacted most strongly on their more central residential areas. Here the lower-skill migrants stand the best chances of finding older cheaper housing released by the previous exodus as well as ready access to the biggest concentrations of menial jobs created by the latest round of the economic restructuring, while high-skill international migrants – often on short-term postings for their companies – can take advantage of the cultural and entertainment attractions in the down town area. Secondly, and related to the latter, was the emergence of young, upwardly mobile professionals (‘yuppies’) as a significant component of society and urban geography, as these singles or partners – usually in the pre-child stage of the family life cycle and working within the financial and producer-serv- ices sectors – took over from the original ‘Bohemian’ gentrifiers of artists and academics as the main force in the social upgrading of inner areas in the 1980s. Thirdly, has been the much more general process of household change, involving the rapid increase in lone-parent fami- lies, in households without children and particu- larly in individuals living on their own as a result of population ageing, rising rates of divorce and separation and the growing tendency for young adults to leave the parental home for higher education, work, independence or escape from abuse. By and large, these groups prefer – or, because of limited financial resources, have to make do with – smaller housing units in less sought-after areas compared to the stereotypical suburban family.

What the literature is far less in agreement about, however, is the significance of the repopu- lation of inner city areas as a phenomenon. Part of this debate is related to whether a fundamental change is taking place in the nature of inner city populations – attitudes to which seem to vary considerably according to particular countries and cities and to the criteria adopted. In an analysis of Canadian cities, for instance, Bourne (1993a) was unable to detect any clear signs of a reversal of the traditional direct relationship between social class and distance from city

Urbanization 157

centres, but Filion (1987) and Ley (1992) are much more upbeat about the achievements of residential revitalization in Canada, particularly by comparison with the changes occurring in the USA’s largest cities. In the Australian context, Badcock (1993, 1995) reckons that the progress of revitalization has proceeded even more strongly than in Canada, while in continental Europe it is generally concluded that inner city areas were not as severely affected by depopulation in the 1970s as in the New World and have benefited consider- ably from the insertion of new residents and housing (see, for instance, Harding et al., 1994). Nevertheless, in both Old World and New, the arrival of wealthier residents has often gone hand-in-hand with the impoverishment of the remaining residents, with the emergence of the ‘dual city’ (Fainstein et al., 1992; Mollenkopf and Castells, 1991). Only time will tell whether the up- market housing investment will constitute more than ‘islands of renewal in seas of decay’ (Berry, 1985: 69).

The likely longevity of inner city revitalization, however, also appears to be a highly contentious issue (Lees and Bondi, 1995). For instance, Bourne (1993b, 1993c) argues that the conditions producing gentrification are waning and in some cases being reversed, ushering in a ‘post-gentrifi- cation era’. He cites the effects of the mid-1960s ‘baby bust’ on the pool of potential young gentri- fiers in the 1990s, allied with the shake-out in employment in financial services which began in the late 1980s and was reinforced by the early 1990s recession, and he also observes that newer immigrant groups are tending to bypass the inner city and go directly to the suburbs. By contrast, Badcock (1993, 1995) reckons that residential revitalization ‘shows no signs of abating during the 1990s’ (1993: 194), emphasizing the supply- side aspects relating to growing market confi- dence in inner city areas and the long-term nature of the trends towards advanced services employ- ment growth, increasing household fission and greater women’s labour market participation.

In conclusion, reurbanization as defined by Klaassen et al. (1981) has not yet emerged as a significant feature in the urban system of advanced economies and there is considerable disagreement about the extent to which the inner city residential revitalization which took place in the 1980s will be able to continue and lead to a fundamental change in the form of the Western city. This is a question that is taken up in more depth in other chapters, so suffice it here to say that much depends on the scale of the challenge in terms of differences between cities in the current disparities between their inner and outer areas. Also crucial, given the importance of public funds in underwriting much of the costs of inner renewal in the 1980s and 1990s, is the extent

to which housing providers and buyers can be convinced that the appropriate range of private and public services will be available in these areas in the future and that new investment will not be undermined by the further growth of ‘underclass’ neighbourhoods. A final, critical issue is that of densification and urban consolidation, whereby increased house building on empty or derelict land and potential green spaces may lead to increased traffic congestion and less attractive environments for residents, representing a diffi- cult balance to be struck.

CONCLUDING COMMENTS If there is one certainty to be drawn from this review of urban population trends, it is that nothing is the same as it was in those apparently straightforward days when ‘urbanization’ meant greater population concentration in any geographical framework used. First, most parts of the developed world have reached the situa- tion where there is little scope for further increases in the proportion of people classified as living in urban places. Secondly, local decen- tralization, while perhaps the most enduring feature of the past few decades, is now very different from the dormitory style suburbaniza- tion of the early post-war period, now involving the veritable ‘urbanization of the suburbs’ and some withdrawal of ‘urban’ facilities from the traditional cities through a form of ‘deurbaniza- tion’ process. Thirdly, in the light of the apparent rise and fall of counterurbanization and of the rather patchy and limited progress made so far by reurbanization, it is tempting to conclude that urban systems are currently passing through some sort of transitional period, but a transition for which no clear outcome is in sight – or, at least, none that can be presented in as simple terms as the conceptualizations discussed in this chapter.

Fundamentally, the advanced economies have moved to a situation ‘beyond urbanization’. Not only is the overall concept of ‘urbanization’ now of very limited value – indeed one could readily argue that its continued use would be obstructive rather than merely neutral – but in retrospect it also seems not to have been very helpful to conceptualize subsequent events as variants or opposites of traditional ‘urbanization’ by adding prefixes to the term. If this conclusion is anywhere near correct, then maybe it is a mistake to rationalize the latest developments as ‘back to the future’, though this phrase does neatly encap- sulate the idea of the resumption of previous trends within a new and evolving context. Perhaps it is more realistic to recognize that the

158 Handbook of Urban Studies

present-day urban systems that form the basis for the future are very different from those which existed 50 years ago, that these are being affected by a range of individual and only partially linked processes rather than some unidirectional form of progression, and that the overall result on the ground is a mosaic of growth and decline, repre- senting the outcome of the success and failure of separate spatial groupings of people to adjust to the opportunities and weaknesses inherent in their various circumstances. Merely to talk in terms of ‘systems’, however, reflects the desire to find some order out of apparent chaos and to have some certainties to build on, so the contin- uing talk of ‘Blue Bananas’, ‘growth corridors’, ‘edge cities’, ‘multinodal regions’, even ‘return to the city’, may possibly serve to influence the future course of events.

REFERENCES Badcock, B. (1991) ‘Neighbourhood change in inner

Adelaide: an update’, Urban Studies, 28 (4): 553–8.

Badcock, B. (1993) ‘Notwithstanding the exaggerated

claims, residential revitalisation really is changing the

form of some western cities: a response to Bourne’,

Urban Studies, 30 (1): 191–5.

Badcock, B. (1995) ‘Building upon the foundations of

gentrification: inner-city housing development in

Australia in the 1990s’, Urban Geography, 16 (1):

70–90.

Beale, C.L. (1975) The Revival of Population Growth in

Non-metropolitan America. Washington, DC:

Economic Research Service, US Department of

Agriculture.

Berg, L. van der, Burns, L.S. and Klaassen, L.H. (eds)

(1987) Spatial Cycles. Aldershot: Gower.

Berg, L. van der, Drewett, R., Klaassen, L.H., Rossi, A.

and Vijverberg, C.H.T. (1982) Urban Europe, vol. 1:

A Study of Growth and Decline. Oxford: Pergamon.

Berry, B.J.L. (ed.) (1976a) Urbanization and Counterur-

banization. Beverly Hills, CA: Sage.

Berry, B.J.L. (1976b) ‘The counterurbanization process:

urban America since 1970’, in B.J.L. Berry (ed.),

Urbanization and Counterurbanization, Beverly Hills,

CA: Sage. pp. 17–30.

Berry, B.J.L. (1985) ‘Islands of renewal in seas of decay’,

in P. Petersen (ed.), The New Urban Reality. Wash-

ington, DC: The Brookings Institution. pp. 69–96.

Berry, B.J.L. (1988) ‘Migration reversals in perspective:

the long-wave evidence’, International Regional

Science Review, 11 (3): 245–52.

Birch, D.L. (1971) ‘Towards a stage theory of urban

growth’, Journal of the American Institute of Plan-

ners, 37 (2): 78–87.

Birch, D.L. (1975) ‘From suburb to urban place’, Annals

of the American Academy of Political and Social

Science, 422: 25–35.

Borgegård, L.-E. Håkansson, J. and Malmberg, G.

(1995) ‘Population redistribution in Sweden: long-

term trends and contemporary tendencies’,

Geografiska Annaler, 77B (10: 31–45.

Boudoul, J. and Faur, J.-P. (1982) ‘Renaissance des

communes rurales ou nouvelle forme d’urbanisa-

tion?’ Economie et Statistiques, 149: I–XVI.

Bourne, L.S. (1989) ‘Are new urban forms emerging?

Empirical tests for Canadian urban areas’, Canadian

Geographer, 33 (4): 312–28.

Bourne, L.S. (1991) ‘Recycling urban systems and

metropolitan areas: a geographical agenda for the

1990s’, Economic Geography, 67 (3): 185–209.

Bourne, L.S. (1992) ‘Population turnaround in the

Canadian inner city: contextual factors and social

consequences’, Canadian Journal of Urban Research,

1 (1): 69–92.

Bourne, L.S. (1993a) ‘Close together and worlds apart: an

analysis of changes in the ecology of income in Cana-

dian cities’, Urban Studies, 30 (8): 1293–317.

Bourne, L.S. (1993b) ‘The demise of gentrification? A

commentary and prospective view’, Urban Geog-

raphy, 14 (1): 95–107.

Bourne, L.S. (1993c) ‘The myth and reality of gentrifi-

cation: a commentary on emerging urban forms’,

Urban Studies, 30 (1): 183–9.

Bourne, L.S. (1994) Urban Growth and Population

Distribution in North America. Report prepared for

IIED London and UN Centre for Human Settle-

ments. Toronto: Centre for Urban and Community

Studies, University of Toronto.

Bourne, L.S. and Logan, M.I. (1976) ‘Changing urban-

isation patterns at the margin: the examples of

Australia and Canada’, in B.J.L. Berry (ed.), Urban-

ization and Counterurbanization, Beverly Hills, CA:

Sage. pp. 111–43.

Bourne, L.S. and Olvet, A.E. (1995) ‘New urban and

regional geographies in Canada: 1986–91 and

beyond’, Major Report, 33, Centre for Urban and

Community Studies, University of Toronto.

Cervero, R. (1989) America’s Suburban Centers: The Land

Use–Transportation Link. Boston: Unwin Hyman.

Champion, A.G. (1988) ‘The reversal of the migration

turnaround: resumption of traditional trends?’ Inter-

national Regional Science Review, 11 (3): 253–60.

Champion, A.G. (ed.) (1989) Counterurbanisation: The

Changing Pace and Nature of Population Deconcen-

tration. London: Edward Arnold.

Champion, A.G. (1992) ‘Urban and regional demo-

graphic trends in the developed world’, Urban

Studies, 29 (3/4): 461–82.

Champion, A.G. (1994) ‘Population change and migra-

tion in Britain since 1981: evidence for continuing

deconcentration’, Environment and Planning A, 26

(10): 1501–20.

Champion, A.G. (1996) ‘Migration to, from and within

the United Kingdom’, Population Trends, 83: 5–16.

Champion, A.G. and Congdon, P.D. (1988) ‘Recent

population trends for Greater London’, Population

Trends, 53: 11–17.

Urbanization 159