81

The compact city and transport energyconsumption

Michael Breheny

This paper challenges an emerging conventional wisdom: that transport energyconsumption, and hence pollution, can be substantially reduced by promoting morecompact cities. Such reasoning has quickly found its way from academic studies toofficial policy in many countries. Do the likely savings from such containmentwarrant the required draconian policies? An empirical assessment of transport energyconsumption arising from decentralization is used to address this question. Two contextual reviews - of the compact city case and the strength of decentralization -precede the assessment. The conclusion is that energy savings will be minimal andthat other policies might be more fruitful.

key words United Kingdom transport energy empiricalcompact city decentralization policy

assessment

Professor of Applied Geography, Department of Geography, University of Reading, PO Box 227,Reading RG6 2AB

revised manuscript received 1 September 1994

Introduction Sustainable Development at the beginning of 1994;the Dutch government has made the compact-city

There has been a remarkable consensus in recent notion a central element of its National Physicalyears that planning for more compact cities is one of Plan (National Physical Planning Agency 1991); thethe most important ways of reducing energy con-sumption and pollution. This policy prescription has

Australian government has rejected the profligate,American-style use of land that has become an

been adopted by the European Commission and by Australian tradition (Newman 1992); and, even innational and local governments in many western the United States, it is now fashionable to deplorecountries. The consensus arises because of a limited urban sprawl and to promote growth-managementamount of academically produced evidence thatsuch a policy -is an appropriate way of reducing

policies (Wachs 1989/90; Chinitz 1990).Given the normal reluctance of governments to

environmental problems, and because governments adopt new, largely untested policy initiatives, thisare anxious to make their contribution to that most consensus is remarkable. It reflects both genuinefashionable of causes: sustainable development. concern over environmental degradation and

The European Commission (Commission of the offers - at first sight, at least - a tangible wayEuropean Communities 1990) has promoted the of contributing to international accords on the‘compact city’ on environmental and quality-of-life environment.grounds; the UK government has prompted trans- While there are many possible merits in theport and land-use planning to reduce transportenergy consumption (Department of the Environ-

urban-containment approach - reduction of landloss, the preservation of habitats and valued land-

ment 1990, 1992,1994) and has made this a central scapes (McLaren 1992; Council for the Protection ofelement of the UK Strategy for sustainable develop- Rural England 1992, 1993), the economic and socialment (Department of the Environment 1993) pre- revival of cities and, some would argue, the survivalsented to the United Nations Commission on of civilization (Bourne 1992; Robson 1994 - a

Trans Inst Br Geogr NS 20 81-101 1995 ISSN: 0020-2754 Printed in Great Britain

82 Michael Breheny

major incentive for the adoption of this policy is body of literature on the merits and occasionallythat of reducing fuel consumption and, hence, of demerits of the compact city. The intention here isreducing harmful emissions. Thus, on the under- to review this only briefly, for detailed consider-standing that fuel savings will result, governments ation is given elsewhere.’ Some coverage is requiredare introducing new - and for some of them radical here, however, because any answer to the above- policies. These may, in the long run, change the questions must draw on this research. The second,very nature of western cities. At the very least, and closely related, set of contextual material to befamiliar local patterns of development may change. reviewed concerns urban-decentralization trends. IfFor example, in the UK there is the prospect that past and continuing urban decentralization is thethere will be no additions to out-of-town retailing; villain of the piece - leading to greater car use,the dominant form of investment in retail outlets greater petroleum consumption and greater levelsduring the last 30 years. At most, the form and of pollution - then it is important to know howfunction of western cities could change dramatically. powerful those trends are. Breheny (1994) hasFor example, the cherished high land-consuming, suggested that one major question not posed in thehigh-mobility lifestyle of many North Americans compact-city debate is whether the process of urbanmay be set for a fundamental change. The aspir- decentralization can be halted, regardless of whetherations of many people fur a suburban or ex-urban it is desirable or not. Recent evidence on decentral-lifestyle - the very aspirations that in the past ization trends in the UK is presented here and isappear to have caused the problem - may be then used in addressing the questions posed above.frustrated. Likewise, the desire of many-business The approach adopted here is deliberately bothinterests to relocate or expand into such areas may generalized and specific. It is generalized becausebe thwarted. The implications of the compact-city the intention is to gain a broad, even crude, feel forpolicy are, then, profound. Potentially they involve the validity of the energy-efficiency claims of theconsiderable collective disruption, the introduction compact-city case. It is specific in that it focuses onof draconian policies and massive but unknown one aspect - transport energy consumption -redistributions of gains and benefits. In short, the within a complex debate on planning and sustain-policies had better be worth it!‘. able development. There are many aspects of the

But how valid is the evidence upon which this sustainability debate and its interpretation in anradical policy consensus is built? Although there urban context that are not covered here. It ishave been some dissenting voices, the general drift appreciated that, in considering appropriate policyof academic advice has been in favour of urban prescriptions to promote sustainable development, acontainment, and - recently at least - largely on the multitude of issues need to be considered together.grounds of transport energy savings, Given the This paper is presented on the assumption, how-importance of these issues and the relative lack of ever, that there is merit in identifying one importantsubstantiation for the policy prescriptions that are issue and investigating it in relative isolation. Theemerging, this paper asks the following questions: results of this investigation can then be used when(i) What is the magnitude of the additional energy considering these broader issues. No attempt is

consumption that results from increasingly made here to define or elaborate on the concept ofdecentralized urban forms? sustainable development (see Breheny 1994). It is

(ii) Is this magnitude - the potential saving, if sufficient for the present argument to assume thatdecentralization can be contained - such that it the reduction of energy consumption and emissionswarrants the introduction of policies that require is consistent with this concept.a reversal of the most powerful forces of urbanchange?

Evidence for the compact city

The fourth section of this paper - ‘Urban decentrali- The term ‘compact city’ is used here as shorthandzation and transport energy consumption’ - willaddress these questions directly. However,

for a variety of approaches to the planning of townsthis and cities which stress the merits of urban contain-

assessment will be facilitated by brief reviews of ment. The idea is not new. It has been promotedtwo areas of research which make up the next two over many years, with varying degrees of explicit-sections of the paper. The first concerns the concept ness, as the corollary of rural protection whichof the “compact city’. There is now a considerable remains a major motive (McLaren 1992). The

The compact city and trasnsport energy consumption

current debate adds two new motives: the transportenergy-saving possibilities and the quality-of-lifeimplications of urban compaction. At the moment itis the energy issue that, arguably, most exercisespoliticians in their promotion of urban-containmentpolicies. Certainly this is the case in the UKwhere two major government initiatives - the UKStrategy for sustainable development (Department ofthe Environment 1993) and the Planning PolicyGuidance note PPG13 (Department of the Environ-ment 1994) - invite planners to adopt a range ofmeasures to reduce the need to travel. The majorconcern is the reduction of CO, emissions. Trans-port currently accounts for only 20 per cent of UKCO, emissions but it is the fastest-rising contributorto the problem. The UK government has interna-tional obligations to reduce CO, emissions to 1990levels in the short term and to achieve furtherreductions over time.Recent arguments supporting the compact-citysolution to the urban energy-consumption problemcome from a variety of sources and adopt a varietyof lines of argument. Some are based largely onassertion. Arguably, this is the basis of the Com-mission of the European Communities‘ (1990) influ-ential Green Paper on the urban environment whichadvocates higher-density cities on the grounds ofboth environmental and quality-of-life benefits. Itgoes so far as to suggest that all developmentshould be contained within the boundaries of exist-ing urban areas. Other advocates of the compactcity derive their support from a mixture of ideologi-cal assertion and analysis. Thus, environmentalgroups such as the Council for the Protection ofRural England (Owens 1991; Jacobs 1993), Friendsof the Earth (Sherlock 1990; Elkin et a!. 1991;McLaren 1992) and English Nature (Jacobs 1992)have all argued for stronger policies of urbancontainment as part of relatively ‘deep green’approaches to sustainability.

Academic studies of compact-city ideas rangefrom idealistic prescriptions drawing on long-standing utopian notions, via mathematical model-ling of urban forms, to empirical examinations of theenergy efficiency of different urban densities andurban sizes. The idealistic line of reasoning is wellrepresented by the work of Yanarella and Levine(1992) who have searched past urban forms andcultures to identify characteristics that might be abasis for modem sustainable urban structures. Theyalight on the medieval Italian hill town as havingthe most desirable mixture of form and function.

83

This conclusion has a lot in common with theimplied ideal in the Commission of the EuropeanCommunities’ (1990) Green Paper. In the UnitedStates, a popular response to the compact-citydebate has been the advocacy of ‘neo-traditional’suburban developments with a small-scale, neigh-bourhood focus of daily life that will reduce travel(Handy 1992). The promotion of ‘urban villages’ inthe UK (Urban Villages Group 1992) is intended tohave similar effects. Questions of urban form, trans-port and energy consumption should also be seenwithin a broader debate about the contribution oftransport to environmental problems (Banister andButton 1993).

A rather specialist but interesting area of workhas been the modelling of urban structures withan explicit focus on the energy-consuming charac-teristics of different forms. Rickaby and associates(e.g., Rickaby 1987; Rickaby et al, 1992) havedeveloped a model of an archetypal English townand have then modelled the relative levels of energyconsumption of different forms of incrementalgrowth to that town. Their conclusions tend tofavour two types of growth: in the urban core or ina ‘decentralized concentration’ form, whereby sub-centres are developed while maintaining overallurban densities. This latter solution has receivedconsiderable support (e.g., Owens 1991, 1992;Barton 1992). This particular debate highlights thepoint that supporters of urban compaction do notnecessarily advocate the one solution. For example,in addition to the promotion of the large-citysolution and the decentralized concentrationapproach, new free-standing settlements have alsobeen proposed as part of a general containmentsolution (Breheny et al. 1993).

Although much of the academic literature focusesdirectly on the possibilities of changes in urban forminducing changed transport patterns, there is a lineof reasoning (e.g., Owens 1991, 1992) whichsuggests that such changes are unlikely to occurwith current high levels of mobility. Thus, theincreased accessibility that might arise from land-usepolicies will have little effect whilst the propensityto travel remains high; motorists will not be dis-couraged from choosing more distant destinations.The implication is that land-use policies need tobe accompanied by other policies - such as fueltaxes - which reduce overall mobility.

Empirical assessments of containment have con-centrated on the question of the degree to whichlocalities which appear to have the appropriate

84

characteristics do in fact have lower rates oftransport energy consumption, These assessmentshave focused on the two related issues of urbandensities and urban size. The hypotheses adoptedare that areas of high population density and largeurban size will have lower rates of consumption onthe grounds that (i) these areas have high levels ofaccessibility and hence require shorter journeys andthat (ii) they induce provision and use of publictransport.

Urban densities and transport energy consumptionOne of the best known studies of urban densities isthat by Newman and Kenworthy (1989a and b).They measured per capita petroleum consumptionand population densities in a range of large citiesaround the world finding a clear negative relation-ship between the two: as densities rise, fuel con-sumption falls dramatically. The villains wereidentified as US cities which had consumption ratestwice as high as those in Australian cities and fourtimes as high as those in European cities. The citieswith the highest densities were those with low carusage and high levels of provision of public trans-port. The obvious conclusion from this study wasthe need for stronger policies of urban containmentand for investment in mass transport systems.

Although Newman and Kenworthy’s work hasbeen very influential in promoting the case for thecompact city, it has generated considerable criticismon a mixture of ideological and technical grounds.Gordon and Richardson (1989) consider the workon both counts. Ideologically, they object to areliance on public intervention, particularly on thegrounds of what they see as a ‘Maoist’ planningsystem. They contrast the Australians’ finding withsome of their own in the United States wherecommuting distances and times have tended to fallin recent years because of employment decentraliz-ation and hence increased suburb-to-suburb worktrips, and where the major growth in travel arisesfrom non-work trips (see also Gordon et al. 1989,1991; Bae 1993). They argue that the developmentof polycentric cities, through market pressure, is themost effective way of dealing with the energy-consumption problem They are also sceptical aboutthe realistic prospects for massive investment inpublic transport across the world. In reviewingNewman and Kenworthy’s (1989a) sourcebook,Gomez-Ibanez (1991) has stressed the complexityof factors that determine rates of gasoline usein different cities and countries. He argues that,

Michael Breheny

in focusing on urban densities, Newman andKenworthy have largely ignored the significanteffect of household incomes and gasoline prices onconsumption, and of incomes on densities. Theiranalysis is undermined, he claims, because of theirobvious prior commitment to land-use policies ingeneral and to policies of containment in particular.Interestingly, Gomez-Ibanez (1991) makes the pointthat the costs - economic, social and cultural - ofradical containment policies have not been weighedrelative to doubtful environmental gains. This papertries to make a similar point.

Newman and Kenworthy (1992) have respondedto these criticisms in technical terms and moregenerally by summarizing the ideological questionposed by their critics: ‘is there a role for physicalplanners? They defend their focus on density as themajor determinant of gasoline consumption not ontechnical grounds but on the basis that economicdeterminants (price and incomes) are unlikely to beused seriously as policy tools by politicians. Thisleaves land-use measures, and hence the planningsystem, as the most effective means of reducingurban energy consumption. They cast doubt onsome of their critics’ data and analysis, suggestingthat a focus on US cities alone, and on Los Angelesin particular, can be misleading. If Los Angeles is anideal model, ‘then heaven help us’ is Newman andKenworthy’s (1992, 360) response. They suggestthat the Americans are confusing polycentric citieswith cities that are simply dispersing to low den-sities. They finish with a ringing defence of physicalplanning as the means to deliver a ‘less automobile-dependent urban future’ (Newman and Kenworthy1992, 360). Bourne has entered the debate on theirside, accusing Gordon et al. (1991) of ‘sweeping andsubjective generalisations’. He argues that the futureof cities cannot rest on the judgement of pro-marketeconomists, even if they can prove the economicmerits of dispersal, which he doubts. He is con-cerned that the advocacy of dispersal may become aself-fulfilling prophesy, ‘. . . contributing to the evol-ution of future urban forms that are increasinglyinefficient and socially inequitable’ (Bourne 1992,513).

A British study which provides evidence on therelationship between densities and energy consump-tion is the ECOTEC (1993) study carried out for theDepartment of the Environment. Table I is takenfrom this study. This table seems to confirm thegeneral findings of Newman and Kenworthy (1989aand b): it shows a neat inverse correlation between

The compact city and transport energy consumption 85

Table I. Distance travelled per person per week, by mode and populationdensity, 1985/6

Density (personsper hectare)

Allmodes Car Bus Rail Walk Other

Data exclude all trips less than 1·6 km and refer only to the main mode used for a tripSource: ECOTEC (1993, Table 6)

Table II. Distance travelled per person per week, by mode and urban size(km)

Car Bus Rail Walk Other Total

Source: ECOTEC (1993, Table 9)

total distances travelled per week and populationdensity. It is car travel that accounts largely for thedifference with people living at the lowest densitiestravelling approximately twice as far by car eachweek as those living at the highest densities. This isone of the findings that led ECOTEC (1993) torecommend to the British government a package ofland-use and transport policies that amount to acompact-city solution.

Urban size and transport energy consumptionIn addition to densities, much of the empiricalresearch has focused on the relationship betweenurban size and energy consumption This issue willbe elaborated here because some of the findings willbe used later to consider the energy-consumptioncharacteristics of different urban types. The majorstudies of urban size and energy consumption in the

UK are by Banister (1980, 1992) and ECOTEC(1993). Banister’s first effort focused on transportenergy consumption in villages in Oxfordshire, andhence is of little direct use in considering the fullrange of urban sizes. His 1992 study did considersuch a range but with very crude size bands. Thiswork is now being pursued further with inter-national comparisons of local transport energy con-sumption (Wood et al. 1994). The ECOTEC (1993)study overcame the problem of size bands by usinga special tabulation from the 1985/6 National TravelSurvey. Using these data, ECOTEC have produceda tabulation of urban sizes against average distancestravelled per week, by mode. The results are shownin Table II. In this table, Metropolitan Areas aretreated as one group; in the ECOTEC report, detailsare given for each of the major Metropolitan Areasin the UK.

86 Michael Breheny

Table III. Primary energy consumed per person per week, by mode andurban size (megajoules (MJs))

Car Bus Rail Walk Other Total

Source: calculated from data in ECOTEC (1993)

Table II demonstrates that there is an increase inaverage distance travelled with decreasing urbansize. Outer London seems to be an exception to therule, with a high overall level of travel based onhigh car travel. The lowest level of travel is not incentral London, as might be expected, but in otherMetropolitan Areas. ECOTEC’s (1993) conclusionfrom this analysis is that large urban areas - byvirtue of higher urban densities, shorter traveldistances and mass-transit facilities - are much moresustainable. Smaller urban areas, and particularlyrural areas, they conclude, are more fuel inefficientand hence relatively unsustainable.

Surprisingly, the ECOTEC (1993) report does notgo on to produce estimates of energy consumptionby urban size. And yet, their report makes thispossible because they supply primary energy-consumption rates for different modes at differentlevels of vehicle occupancy. These rates were origi-nally derived from a European Commission study(Commission of the European Communities 1992).Using a 50 per cent occupancy consumption,Breheny (1994a) has adopted the following ratesfrom the ECOTEC study car, 1·31; bus, 0·58; rail,0·59; walk, 0·0; and other, 0·82 megajoules (MJs) perpassenger kilometre. The ‘other’ category, whichincludes two-wheeled vehicles, taxis, domesticaeroplanes, other public transport and other types ofbuses, has been given a value which is the averagefor car, bus and rail rates. The modal categories havebeen chosen to coincide with those in Table II.

These consumption rates may be multiplied bythe per capita weekly travel distances in Table II.This gives primary energy consumption from

transport per week per capita by urban size bands.The results are shown in Table III. The generalresult of this exercise to measure energy consumedis to increase the urban-rural difference relative tothe travel distances shown in Table III. Thus, ruralareas have average consumption levels 108 per centhigher than the most energy-efficient MetropolitanAreas, compared to an 87 per cent differenceon travel levels alone. This widening results, ofcourse, from the use of modes of travel to arrive atthe energy-consumption measures. The greaterdependence of rural and small-town dwellers on thecar is the major source of their greater energyconsumption.

Policy responses to evidenceAs explained earlier, despite some reservationsabout the compact-city idea2 it has gained con-siderable academic and political support. In the UKthis political support has been profound. It hasresulted in guidance being issued to local planningauthorities urging them to both adopt policies andmake development control decisions which promotestronger urban containment. This is part of aconcerted effort to promote sound environmentalpolicies (but at the same time promoting economicgrowth). The major guidance has been issued in theform of Planning Policy Guidance notes (PPGs). Thelatest detailed guidance is in PPG13 on Transport(Department of the Environment 1994). Amongstother policies, PPG13 proposes (para 1.7) thefollowing under the heading of ‘Planning for lesstravel’:

The compact city and transport energy consumption 87‘Development plans should aimtravel, especially by car, by:

to reduce the need to

- influencing the location of different types of devel-opment relative to transport provision (and vice versa);and

- fostering forms of development whichwalking, cycling and public transport use.

encourage

In order to achieve these aims, PPGl3 (paras 3.2and 3.3) proposes specific local initiatives. Onhousing, it proposes to:

- allocate the maximum amount of housing to existinglarger urban areas (market towns and above) wherethey are or can be easily accessible to facilities . . . andto a range of transport provision . . .

- . . . promote land for housing in locations capable ofbeing well served by rail or other public transport . . .

- reallocate accessible land designated for activitieswhich are not employment- or travel-intensive uses tomore intensive uses . . .

PPG13 goes on to promote policies for retailing,leisure, tourism and recreation, and education whichalso stress the intensive use of urban sites andparticularly sites served by public transport.

These policy proposals, many of them taken fromthe recommendations of the ECOTEC (1993) report,show clearly how one government has respondedradically to the environmental-sustainability prob-lem by focusing on the role of the planning systemand by adopting initiatives that, collectively, pro-mote the compact city. This line of reasoning isalso contained within the Department of theEnvironment’s (1993) Strategy for sustainabledevelopment.

- avoid any significant incremental expansion of The strength of urban decentralization in

housing in villages and small towns . . . the UK

- avoid sporadic housing development in the opencountryside, but promote appropriate developmentwithin existing communities . . .

- avoid the development of small new settlements . . .especially where they are unlikely to be well served bypublic transport and are not designed to be capable ofbeing self-contained.

- concentrate higher-density residential developmentnear public transport centres, or alongside corridorswell served by public transport . . .

- set standards to maintain existingwhere appropriate increase them . . .

densities and

- juxtapose employment and residentialfeasible, through mixed-use development.

uses, where

On employment, PPGl3 (paras 3.5 and 3.6)proposes to:

- focus the opportunities for development of travel-intensive uses (such as offices) in urban areas inlocations well served . . . by public transport.

- avoid major developments in locations not wellserved by public transport . . .

- allocate sites unlikely to be served by public trans-port solely for uses which are not employment- ortravel-intensive;

The idea of the compact city has, then, been givena considerable academic and practical considerationin many countries. The end result has been aremarkable and rapid academic and political adop-tion of the idea. However, throughout this debate,at least one fundamental question has not beenposed: given the apparent power of the forces ofurban decentralization can the idea be implemented?

This question has been addressed more fully inBreheny (1994a). It will suffice here to demonstratethat in the UK, at least, the process of urbandecentralization remains very strong; sufficientlystrong to cast doubt on the feasibility of seriouslyimplementing tougher containment or compact-citypolicies. This brief review will coven first, theempirical evidence of decentralization; secondly,the explanation of continuing decentralization; andthirdly, the policy context in which decentralizationis taking place.

Before demonstrating the strength of decentraliz-ation trends in the UK, it may be useful to clarifyterms and concepts. The term ‘decentralization’ isused here to refer to all forms of population andindustrial growth taking place away from existingurban centres. Such growth might arise fromrelocations or from indigenous change. For someyears geographers have been disputing the natureof these changes. Some argue that growth hasbeen occurring as a result of continuing suburbaniz-ation which is sometimes discontinuous - where

88

Rounding errors in columnsSource: Office of Population Censuses and Surveys (1993, 1994); components-of-change data obtaineddirectly from OPCS

development is hindered by green belts, forexample. Others argue that the process has beenone of ‘counterurbanization’. This view suggeststhat growth has been focused, not on the peripheryof existing urban areas but in the least urban areas.Thus, households and companies have been con-sciously relocating to small towns and villages in adistinct. anti-urban movement. These issues areimportant but the concern here is simply whetherthese processes are continuing and, if so, what arethe prospects that they can be reversed?

Largely Rural areas have gradually gained a largershare of net gains over the years. It is the largegains - both in absolute and percentage terms - inthe Remote, Largely Rural category that led tothe view that the changes are evidence ofgenuine ‘counterurbanization’ rather than simplysuburbanization.

Evidence of continuing decentralizationThe decentralization of population in the UK maybe most conveniently observed using the standardclassification of urban areas adopted by the Office ofPopulation Censuses and Surveys (OPCS). Theseurban types (shown in Table IV), derived from acluster analysis, generally range from large urbanareas to remote rural areas. Analysis of populationchange fur each of these urban types since 1945shows a consistent process of decentralization.Throughout all periods, the biggest losses havetended to be in traditional industrial areas: inLondon and in the other Metropolitan categories.The areas of highest population gain, however,move to successively smaller urban areas over theyears. In all periods, the New Towns have seensubstantial population growth, as have the Mixedand Accessiblé areas, both inside and outside theSouth East. Resort and Retirement and Remote,

Table IV shows details of population change bythe OPCS urban types for 1981-91. Some commen-tators (Champion 1987; Champion and Congdon1988) had anticipated a ‘return to the city’ duringthe 1980s. However, the evidence clearly showsthat the decentralization process has continued. AsTownsend (1993, 217) has stated,

Any belief that the urban-rural shift was a passingphase of British economic geography is amply dis-pelled . . . The directions of net change of the 1960sand 1970s resumed fully in the 1980s with almostlaw-like regularity.

The losses are in the industrial cities, although theseare much lower than in previous decades. InnerLondon does show a significant change in itsfortunes. Substantial losses in previous periods havebeen converted to a 3 per cent gain in population.According to OPCS mid-year estimates of popu-lation (Office of Population Censuses and Surveys1994), most of this gain appears to have occurredduring the two years prior to 1991. The largestgains in Table IV are in the New Towns, Resort andRetirement, Mixed and Accessible and Remote,

Table IV. Population change, 1981-91, for OPCS urban types

Michael Breheny

Urban type

1981 1991 Change Natural Netpopulation population 81-91 Percentage change migration

(000s) (000s) (000s) change (000s) (000s)

The compact city and transport energy consumption 89

Percentage of population of England and Wales

Figure 1. Population change, 1981-91, by OPCS urban typesSource: calculated from Office of Population Censuses and Surveys (1994)

Largely Rural areas. Over a 10-year period, thesefour urban types gained nearly 1.4 million people,1.2 million by migration. The components of popu-lation change in each type are interesting. In theRemote, Largely Rural category all of the netchange resulted from migration. In the decade,nearly 0·5 million people migrated into this type ofarea. The Resort and Retirement areas also gainedpopulation largely from migration. In contrast,the rapidly growing Mixed and Accessible areasgained equally from natural increase and migration,perhaps reflecting a young age structure arisingfrom previous rounds of work-related migration.

In considering the feasibility of stronger urban-containment policies, the nature of populationchange is important. Planners can have little pros-pect of influencing the geography of population

change where gains result largely from naturalincrease. However, where in-migration is thecause of population growth, then restrictions onhousing and job increases, for example, might beused to constrain such movements. The Remote,Largely Rural areas, with large net in-migration,might be the most susceptible to constraintpolicies.

Figure 1 shows the percentage population changefigures from Table IV in graphical form. Thisdemonstrates very clearly the ongoing process ofurban decentralization in England and Wales. At itsextreme, the compact-city argument implies con-verting what is clearly a top-left to bottom-righttrend into one that will in future run from top-tightto bottom-left. Viewed in this way, the task looksformidable.

90

Explanations for decentralization

Michael Breheny

The aim here in reviewing urban decentralization inthe UK is simply to demonstrate the strength of thetrend - and hence the difficulty in reversing it - notto elaborate on explanations of the trend. Neverthe-less, some mention of possible causes may be inorder because some explanation of decentralizationprocesses will help to determine the prospects forchanging them Cross (1991) proposes alternativeexplanations for what he calls ‘counterurbanization’.He suggests five groups of theories, based on:longer-distance commuting; residential preferencefor non-metropolitan areas; economic change infavour of peripheral areas; employment decentraliz-ation; and integrated theories. He concludes that allof these explanations have some merits and hencethat an integrated theory is probably appropriate.However, the dominant causes of counterurbaniz-ation within an integrated theory are likely to beresidential preferences and economic changes whichfavour decentralized locations.

The most promising explanations, according toCross (1991), relate to the role of employmentdecentralization. It is suggested that the forcesdetermining the way that capital uses space, andhence the resultant economic geography, haveshifted dramatically in the postwar years. Arguably,two specific shifts have taken place. The first wasthe move from regional sectoral specialization, inwhich regions specialized in a particular industrialsector and held all of the operations of companiesinvolved in that sector. This, it is argued, haschanged, such that there is often a spatial separationof those operations into clear spatial divisions oflabour. This newer geography of economic activityteds to separate core functions of a company -administration, R & D, marketing - from routineproduction. The former, core functions have tendedto locate in major cities and in decentralized butaccessible locations. Production has tended to focusaway from traditional urban locations to cheaperperipheral locations, aided also by demands for newindustrial property at lower densities (Fothergill andGudgin 1982). Similar trends are also discernible inthe service sector, with the splitting of small,centralized functions from large-scale ‘back offices’in decentralized locations.

The second shift suggests that the spatial divisionof labour explanation of changing industrial geogra-phy may now itself be inadequate, as the hierarchi-cal divisions of labour developed in the postwar

period break down. The new imperative, it isargued, is flexibility. Theories, under the label of‘flexible accumulation’, suggest that pressures tocompete on the basis of production flexibility(including ‘just-in-time’ production) mean that com-panies will tend to locate functions together thatwere previously separated. Thus, new concen-trations of economic activity are being established.These are not located in the former industrial areasbut in new ‘clean’ locations, such as the fringes ofthe South East region - which tend to equate to theMixed and Accessible Rural urban type discussedearlier - that have witnessed substantial popu-lation increase in recent years. The preference ofcompanies for these types of locations has beenconfirmed in survey work by Prism Research (1991).Fielding (1991), in a study of migration to and fromthe South East of England, has attempted to relatepopulation movements to the changing locationalpreferences of industrialists.

Simple observations of employment changein England and Wales are consistent with thesetheories. Table V demonstrates a clear pattern ofjob losses in the older, industrial cities and sub-stantial gains in small town and rural locations.This table is a direct equivalent of Table IV,showing job change rather than population change.The trend in job change is more marked than that ofpopulation change: percentage losses in the largercities and gains in rural areas are higher in the caseof jobs.

The consequences of these various pressuresfor industrial and population changes have beencontinued decline in the older industrial cities -albeit at a slower rate in recent years - and furthersubstantial growth in decentralized small townand rural areas. In the recent past, then, urbandecentralization in the UK has continued. For thecontainment debate, the important question iswhether it is likely to continue. What are theprospects that the process will slow down? Whatare the prospects for slowing or reversing the trendthrough planning or other policies? There is someevidence that in the future the pressures for decen-tralization will lessen. For example, the nationalgrowth in the service sectors that fuelled somedecentralization in the 1980s has disappeared. How-ever, many of the forces determining the nature andscale of decentralization are deep-rooted. There isno evidence that the problems of the older, indus-trial cities - the ‘push’ factor - are being resolved.Indeed, some would argue that in many cases they

The compact city and transport energy consumption 91

Table V. Employment change by OPCS urban areas, 1981-91

Urban type 1981 jobs 1991 jobsChange Percentage change

1981-91 1981-91

Employees in employmentSource: NOMIS

are getting worse (Robson 1994). There is nothingto suggest that the basic desire of industrialists for‘clean’ - in terms, for example, of new sites,compliant labour and uncongested infrastructure -decentralized locations is any less now than in thepast. The recession has reduced the scale of changebut not the general geographical thrust. The PrismResearch (1991, 131) study of company moves inthe South East concluded:

Of course, the future is uncertain, but there is nothingin recent data to suggest that this underlying cycle willnot be repeated some years hence . . . a new cycle ofbusiness growth, with sufficient strength to catch upon ground lost in the recent recession, is likely tocontinue to push economic activity out of London andover the South East Regional boundary, in Westernand Northern directions.

An interesting, but neglected, question in the decen-tralization debate is the degree to which the estab-lished patterns reflect community choice, and hencethe degree to which containment policies wouldthwart that choice. Clearly, what may appear to befreedom of choice is restricted or directed bygovernment or private agencies. For example,government housing subsidies and transport policy,and housebuilders’ land-purchase policies have aconsiderable effect on changing urban structure.However, in turn, both parties would argue thatthese policies reflect consumer choice. Given thisconundrum, the assumption here is that much, if notall, of urban decentralization can be explained byhouseholders and businesses ‘voting with their feet’.

The evidence suggests that they will continue todo so.

The policy context for decentralizationWhen contemplating the scale of past and futurechange, it must be remembered that the prevailingplanning context has been and will remain one ofurban containment. A basic tenet of the Britishplanning system over the past 50 years or more hasbeen the restriction of urban encroachment into thecountryside. The scale of decentralization shown inTables IV and V occurred in a period when politicaland popular support for urban containment wasprobably stronger than at any time in the UK.Although it has been tempered to some degree bythe pro-market philosophy of governments since1979, the prevailing ethos has been one of environ-mental protection. Through the 1970s and 1980sapproximately 50 per cent of all new housing wasbuilt within existing urban areas (Breheny et al.1993). The compact-city policy would obviouslyseek to increase this surprisingly high figure stillfurther. However, some commentators suggest thatit will be difficult to maintain the 50 per cent figurein the future. For example, Grigson, reporting onwork for SERPLAN (1988), suggests that with theredevelopment of the best urban sites and newworries over ‘town cramming’, a figure of 40 percent might be more realistic in future. Recentconcerns over the contamination of urban sitesmight also decrease the prospects for urbanredevelopment.

92

Breheny et al. (1993) have estimated that, evenassuming a 50 per cent figure for urban infill, theSouth East of England will see some 300-400 000houses being built on green-field sites over the 20years from 1991. These could, of course, be built assuburbs to the largest cities and towns but thelikelihood is that market demands will dictate thatthey be built in locations further down the urbanhierarchy.

This brief review - of empirical evidence, theo-retical explanations and the policy context - sug-gests that in the UK at least, urban decentralizationremains a powerful force. The prospects are that itwill continue. The implications of this evidence areeither that the policies currently being promotedwill have little effect in reducing decentralization,or that much more draconian policies, with veryserious social and economic ramifications, will berequired.

Urban decentralization and transportenergy consumption

The logic of this paper has implied that the evidencein favour of urban compaction is likely to be validbut that the major policy required to bring it about- a reversal of urban decentralization trends - willbe difficult or impossible to implement. This latterfinding. alone is important, given the strength ofsupport for the compact-city idea and the neglecthitherto of the question of feasibility. However, theevidence reviewed in arriving at this finding allowssome analysis that questions the very validity of thebasic arguments in favour of the compact city. Thequestions posed at the beginning of this papersuggest that the evidence presented earlier on theenergy efficiency of the compact city, and generallyassumed to be valid, is suspect. To what degree hasurban decentralization - apparently the villain of thepiece - actually increased overall energy consump-tion from transport? Most of the evidence reviewedearlier concludes that it does increase transportenergy consumption but it does not explain thesize of the increase, nor whether it is significant.

Sharpe (1982) is a rare sceptic on this issue,suggesting that a tripling of the density ofMelbourne would yield only an 11 per cent trans-port energy saving. Newman and Kenworthy(1989b, 28) take a more familiar line. They havesuggested that

. . . there is a potential fuel saving of some 20 to 30percent in cities like Houston and Phoenix, if they

Michael Breheny

Table VI. A matching of OPCS urban types andECOTEC’s urban-size categories

OPCS urban types ECOTEC size categories

were to become something more like Boston orWashington, in urban structure . . . With more extremechanges in land use, such as increasing the density ofpopulation and jobs to the level of inner areas of NewYork, much higher fuel savings would seem likely.

Nothing is said about the strength of policies thatwould be required to bring such changes about.

Deriving consumption rates for urban typesAn attempt to answer questions about the energy-consumption effects of urban decentralization can bemade by putting together the information on popu-lation decentralization and energy consumptionpresented above. Unfortunately, the two sets ofinformation are not directly comparable becausethat on decentralization is presented by OPCSurban types, and that on energy consumption isavailable from ECOTEC (1993) by urban size bands.It is possible, however, to arrive at a reasonablecross-matching of these two sets of categories.OPCS do provide some guidelines on the size ofaverage settlements within each urban type, allow-ing a matching of urban size bands. Breheny (1994a)suggests the matching shown in Table VI. A finermatching might be facilitated by 1991 Census datafor urban areas not available at the time of writing.These provide Census information for settlements,rather than administrative areas.

The matching in Table VI allows an assessmentof the degree to which people are leaving areas oflow energy consumption and moving to areas ofhigh consumption, as the conventional wisdomsuggests is the case. Figure 2 plots 1991 energy-consumption rates, taken from Table III, over levels

The compact city and transport energy consumption 93

Percentage of population of England and Wales

Figure 2. Population change, 1981-91, and transport energy-consumption rates for OPCS urbantypes

of population change for each OPCS urban type,taken from Figure 1. It can be seen that there is ageneral correlation. Areas of population loss - themajor metropolitan areas - do have lower consump-tion rates, and areas of population gain - largely themore rural areas - have higher rates. Despite someinteresting exceptions such as New Towns andResort and Retirement areas, decentralization movesare generally to high energy-consuming areas. Themost interesting of the urban categories are theMixed and Accessible and Remote, Largely Ruralareas. These have been the major recipients ofdecentralization and have the highest energy-consumption rates, suggesting they are playinga major role in increased transport energyconsumption.

This simple assessment suggests, then, thatdecentralization is exacerbating energy consump-tion from transport. This finding is consistentwith most previous research assumptions andwith what is now becoming a conventional policystance. It suggests that tighter urban containment,or the compact-city solution, is an appropriatemeans of reducing energy consumption. Whatthis simple assessment does not reveal, however,is the scale of the problem. By how much hasdecentralization increased energy consumption?And is the additional energy consumptionsignificant? Does it warrant the draconianmeasures that would be required to reverse thedeep-rooted process of decentralization from ourcities?

Urban type

Two rough estimates of the scale of the addi-tional consumption that arises from decentralizedurban structures can be made. Using informationpresented above, it is possible to calculate thedegree to which:(i) absolute differences in energy consumption

between different urban types affect overallenergy consumption. This can be measured asthe difference between actual overall consump-tion and that which arises if all areas hadconsumption rates as low as the most efficienturban type. This ‘lowest rate’ approach mightdemonstrate the savings that would arise fromthe extreme policy stance of re-urbanization;and

(ii) marginal changes in decentralization - over, say,a given number of years - affect overall energyconsumption. This ‘marginal change’ approachmight demonstrate the benefits of the lessextreme - but still radical - case of haltingfurther decentralization for a given period.

The ‘lowest rate’ approach and additional energyconsumptionTable VII shows the reduction in energy consump-tion that would arise if all urban types had the sameper capita rate as the area with lowest rate: the‘Principal Cities’ of the metropolitan districts, with a

94 Michael Breheny

Table VII. Estimate of the transport energy saving if all urban types had the same consumption rateas the most efficient type, 1991

4. Lowest 6. Difference 7. Percentage2. Energy energy 5. Energy between actual differenceconsump- 3. 1991 consump- consump- and lowest betweention rate total energy tion rate tion using consump- actual and

1. 1991 (MJs per consump- (MJs per lowest rate tion levels lowestpopulation person per tion (GJs person per (GJs per (GJs per consump-

(000s) week) per week) week) week) week) tion levels

1 GJ=1000 MJs; rounding errors in columns; column 2 total calculated from column 1 and 3 totals

rate of 119.1 MJs per person per week. Columns 6and 7 show the degree to which total energyconsumption would be lower in all urban types(with the obvious exception of Principal Cities) ifthe lowest consumption rate applied. Summingcolumn 6 gives the total national saving that wouldarise from this lower consumption rate. The overalleffect would be to reduce consumption by 34 percent. This implies, then, that if it were possible toredesign the UK’s urban structure to have theconsumption characteristics of Principal Cities -higher densities, high public-transport provision,lower average journey lengths, trip mix, etc. - thetransport energy saving would be in the region ofone-third.

The adoption of the lowest consumption rate forthis analysis does not mean, of course, that inprinciple still lower levels could not be achieved.There is no reason to asume that, with more carefulland-use planning and better provision of publictransport, the average consumption rate for Princi-pal Cities should not be lower than at present. Thus,again in principle, all settlement types could alsoachieve a consumption rate below the 119·1 MJsper person per week. In practice, however, the effortrequired to bring all urban types to something likethe rate for Principal Cities would be enormous. Itwould require a long-term commitment to reverse

The compact city and transport energy consumption 95

decentralization trends; it would need drastic actionin areas such as Outer London where consumptionrates appear to be very high; it would mean, ineffect, the abandonment of many smaller ruralsettlements; and many other restraints that woulddoubtless be deeply unpopular with businesses andhouseholds alike. In reality, of course, it would beimpossible to restructure settlement patterns in sucha way as to give all urban types the characteristicsof the most energy-efficient areas. Even if it werepolitically acceptable, it would take generations toachieve through land-use change.

Given that any initiatives to achieve greaterurban containment will fall well short of the savingsthat, in principle, could be gained by this ‘lowestrate’ approach, the question arises as to whether therealistic gains are sufficient to warrant the effort. Ifthe ‘lowest rate’ approach would give gains of 34per cent, it is likely that any more realistic interven-tions - of, for example, the kinds being proposed bythe advocates of the compact city - would yieldgains well below this. A 10 per cent saving wouldrequire an average consumption rate of approxi-mately 160 MJs per person per week; a 15 per centsaving would require an average rate of about 150MJs per person per week - a rate a little lower thanthat prevailing in the Metropolitan Districts: Otherscategory.

Would advocates of the compact city be soforceful if they knew that the likely gains from theirproposals might be a modest 10 or 15 per centenergy saving achieved only after many years, andunprecedentedly tough policies? Although propo-nents of the compact city have not generallyspecified expected levels of energy saving, it is to beassumed that they have in mind levels considerablyhigher than this.

This analysis is, of course, static. It does not takeinto account future increases in car ownership andoverall travel levels. Even if trend forecasts of adoubling of cars in the UK over the next 20 yearscan be reduced by government policies, the likeli-hood nevertheless remains that the kind of transportenergy savings resulting from greater urban con-tainment will be swamped by increased levels oftravel.

The ‘marginal rate’ approach and additionalenergy consumptionA second attempt can be made to measure likelysavings in transport energy consumption fromurban-containment policies. Here the aim is to look

at the additional consumption of energy that arisesfrom marginal changes in urban decentralization.Realistically, containment policy, even if appliedmuch more forcefully than in the past, is likely tomake marginal changes to urban forms rather thanrestructure them entirely. Thus, estimates of theadditional energy consumption arising from pastmarginal change may give a reasonable clue as topossible future savings if decentralization is haltedor slowed. Specifically, the question posed is: if nourban decentralization had taken place over the fast30 years, what would have been the transportenergy saving? This might give a crude guide as tothe effect that a tough - no further decentralization- policy might have over the next 30 years.

The chosen period for analysis is 1961-91.Decentralization has occurred in the UK throughoutthis period. Estimates have been made above ofenergy consumption in each OPCS area type at1991. Estimates can be made of the equivalentenergy-consumption levels on the assumption thatno decentralization occurred over the 1961-91period. The difference between the two should thenrepresent the urban energy consumption that hasoccurred because of decentralization The 1991‘non-decentralization’ population is calculated byallowing the 1961 population in each urban type togrow by the national growth rate (+ 10·8 per cent)from 1961 to 1991. This, in effect, gives a roughestimate (neglecting differential natural increases,etc.) of the 1991 population in each urban type as ifno population movement had taken place over theprevious 30 years. Calculations are shown in TableVIII. The important figures in Table VIII are given incolumns 6 and 7. Column 6 shows the absoluteenergy saving or loss in each urban type if decen-tralization had not taken place. Thus, the largercities show positive figures, reflecting higher overallenergy consumption without decentralization whilstthe more rural areas show negative figures, reflect-ing lower energy consumption without the popu-lation gains from decentralization, Column 7 givesthe appropriate percentage changes. The total incolumn 6 demonstrates that 234 899 GJs (1 GJ=1000 MJs) of energy are consumed per week intransport in England and Wales because of thedecentralization of population over the last 30years. This suggests that the decentralizationchanges over this period now mean that weeklytransport energy consumption is about 2·5 per centhigher than it would be had those changes not takenplace. It also implies that any policy of urban

96 Michael Breheny

Table VIII. Estimates of fuel consumption arising from decentralization changes, 1961-91, byurban types

5. 19912. Energy 4. 1991 total energy 6.

consumption 3. 1991 ‘non- consumption from Difference 7. Percentagerate total energy decentralized’ ‘non-decentralized’ between difference

1. 1991 (MJs per consumption population population columns 3 betweenpopulation person per (GJs since 1961 (GJs per and 5 (GJs columns 3

Urban type (000s) week) per week ) (000s)* week) per week) and 5

1 GJ=1000 MJs; rounding errors in columns*1961 population multiplied by national population change (+ 10·8 per cent) for 1961-91, i.e. an estimate of the population had therebeen no decentralization

containment implemented now for, say, the next 30 is that such policies are expected to yield marginalyears, would - assuming that other policy devices, savings relative to the more fundamental policy ofsuch as fiscal measures, remain the same - also yield containment, which we now see would itself delivertransport energy savings of this kind of magnitude. only trivial benefits.

This result is quite remarkable. Across the world, One way of putting this 2·5 per cent finding ininternational, national and local agencies are perspective is to consider other policy measures thatwedded to the notion that prevention of urban might produce similar levels of transport energydecentralization will make the major contribution to saving. These alternatives have tended to be rel-the reduction of energy consumption from trans- egated in the urban-sustainability debate in favourport. These agencies are willing to introduce the of planning-led containment policies. Three obviousdraconian measures necessary to bring this about, options are changes in vehicle technology, changeswith all that implies in terms of social and economicchange, without any serious knowledge of whetherthe desired result will ensue. The evidence pre-sented here suggests that over the last 30 yearsdecentralization has made a trivial contribution toadditional energy consumption, implying thatefforts to prevent further decentralization - ifsuccessful, which is doubtful - will also be trivial intheir effect.

It is possible that policies currently being pro-posed in addition to containment - such as highdensities at transport nodes and mixed-use develop-m e n t s - will make cities more transport energyefficient than they have been over the fast 30 years.The potential savings from such schemes isunknown. However, the assumption at the moment

in driver behaviour and fiscal measures. Bae (1993)reviews these options alongside land-use possi-bilities, assessing the merits of all approaches, bothin principle and in the specific case of reducing airpollution in southern California. She suggests thatall approaches have merits but concludes that

The automobile emissions problem is much moreamenable to technological solutions . . . than to rem-edies that rely on transportation planning and policiesbased on radical changes in travel behavior. (1993, 73)

Even assuming continued reliance on petroleum, it ispossible to envisage governments promoting yetfurther technological improvements in vehicle fuelconsumption and in other ‘tailpipe solutions’, as theAmericans call them. The range of measures aimed

The compact city and transport energy consumption 97

at affecting travel behaviour, discussed but thenrejected by Bae, such as vehicle sharing and flexibleworking hours, have hardly been discussed in theUK, let alone implemented. Technological andbehavioural measures, then, might yield modest fueland emissions savings relatively easily. Certainly2-3 per cent savings by these means do not seemunreasonable. They might also take effect relativelyquickly.

Fiscal measures may also yield similar savingsrelatively easily. The apparent confidence of theadvocates of the ‘compact city that planned urbancontainment will reap substantial energy savingshas had the effect of taking attention away from thepossible use of fiscal measures, particularly pet-roleum tax, to achieve such savings. Having saidthis, it should be pointed out that some advocates ofcompaction policies (e.g., Owens 1991, 1992) haveargued that policies will not work unless the pro-pensity to travel is also reduced, by fiscal or othermeasures But the general assumption has been thatbecause of the high price inelasticity of demand forpetroleum, fiscal devices for reducing energy con-sumption are politically unacceptable. However,there is evidence that these elasticities are not soextreme as to make the idea redundant- Bae (1993)suggests that the price elasticity of demand-for-work autotrips is between - 0·3 and - 0·6, sug-gesting that a 10 per cent increase in the price ofpetroleum would reduce demand by 3-6 per cent.MuConsult (1992) have found that in TheNetherlands, a 10 per cent price increase reducescar-kilometres by a little less than 2 per cent. Again,if we are looking for savings equivalent to the 2·5per cent discussed above, it seems that they couldbe found by relatively modest fuel price increases.Furthermore, it could be achieved virtually over-night, compared with the 30-year time periodconsidered above or the generally long timehorizons required to induce change through theplanning system.

Viewed in the context of these alternatives, theuse of urban-containment policies to achieve trans-port energy savings does not seem to be a veryfruitful route to take.

Improving the analysisPresentation of earlier versions of this paper atseminars generally elicited three responses. First,there was indignant surprise at the apparently smallscale of potential energy savings, reflecting thegeneral assumption that urban containment is a

good thing. Secondly, some respondents felt con-vinced that if only the paper covered chapter andverse on all aspects of cities and sustainability, thenthe case for compaction would remain unblemished.The reply to this had to be repeated then and maywarrant repetition now. This analysis has beendeliberately narrow and isolated, with the intentionof getting to grips with one specific issue ratherthan continuing to flounder in complexity. It mustbe stressed that there are other environmental - aswell as economic, social and cultural - motives inpressing for urban-containment policies, albeitthemselves controversial. Obvious examples are theprevention of losses of open land and destruction ofhabitats. Limiting further decentralization mighthelp to ameliorate these problems. However, by farthe strongest motivation for the compact-city pro-tagonists has been the prospects of substantiallyreduced transport energy consumption. If theanalysis presented above gives anything near to areasonable estimate of the additional energy con-sumption arising from decentralization, then thismotivation may be grossly misplaced.

The third response was to argue that if only theanalysis presented here had taken account of variousdetailed factors, the results would be very different.This criticism is the most direct and hence deservessome further consideration. As explained at theoutset, the approach adopted here has been con-sciously generalized and aggregate. However, it ispossible to conceive of improvements to the analy-sis in principle. In practice, they would prove verydifficult to implement, Nevertheless, there is merit inrehearsing what form such improvements mighttake and, importantly, assessing the degree to whichthey might change - and in what direction - theresults obtained above. Would improvements givegreater negative percentage figures, thus implyingthat decentralization has had a greater effect ontransport energy consumption than suggested here?Or would improvements lower the figures, suggest-ing that the estimates made here overstate theenergy-consuming effects of decentralization? Orwould improvements have no clear effect on theanalysis? A set of possible, but by no meansexhaustive, improvements will be discussed brieflyand an indication given of their likely effect on theresults of the ‘lowest rate’ and ‘marginal’ analysespresented above.

Taking account of social structure: This additionalanalysis concerns the question of the degree to

98 Michael Breheny

which different rates of travel in urban types issolely a function of accessibility. It is possible thatsocial structure will also be a determinant of relativetravel rates. If, on average, higher social groups arewell represented in small town and remoter areas,their higher incomes, higher car ownership andspecific lifestyles are likely to increase average ratesof travel in those areas. Controlling for socialstructure would thus reduce the differences inenergy-consumption rates between urban and ruralareas, Decentralization to rural areas would thushave a lower overall effect on consumption than inthe analysis presented above.

lnclusion of short trips in the analysis: The ECOTEC(1993) data upon which the above analysis is basedomit short trips of less than 1·6 km. It can beassumed that many of these trips would be by footor bicycle. The effect of including such trips in theECOTEC data would be to increase the lengthof total trips per person per week. This would bethe case most especially in larger cities and towns.Thus, the gap between the shortest and longesttotal trip lengths in Table If would be narrower.However, this would have no effect on relativeenergy consumption between urban types aswalking/cycling trips have a zero consumption rateas shown in Table III. The shorter distancestravelled in larger cities and towns reflects, to someextent,cycling

the greaterjourneys.

opportunities for walking and

Assuming differential population growth in urban andrural areas: It might be argued that the assumptionadopted in the ‘marginal’ analysis above of aconstant 10·8 per cent population growth in allurban types from 1961 to 1991 is unrealistic. Itmight, therefore, be more reasonable to adoptnatural increase rates appropriate to each urban type(they do differ considerably, as Table IV has shown),so generating a lower population growth in urbanareas than the 10·8 per cent and a higher rate inmore rural areas; an ‘unstoppable decentralization’effect. With regard to the ‘marginal analysis, thiswould have the effect of less energy consumption inurban areas and more in rural areas over the 30 yearperiod. However, as Table IV shows, there is noobviuus pattern in rates of natural increase, at leastduring the 1981-91 period It is not obvious thatsuch changes would tend to increase or decrease the- 2·5 per cent figure.

Changing consumption rates for urban types: Thecurrent levels of transport energy consumption inurban types relative to 30 years ago is a function ofboth changing population movements - decentral-ization - and changing rates of consumption in eachtype. Ideally, the effects of each of these changes onconsumption would be separated out. Thus, the‘pure’ decentralization effects would be available.Unfortunately, consumption rates for 1961 are notavailable and hence the effects cannot be separated.In reality, of course, it would be difficult to separatethe two because they are interrelated. Decentraliz-ation is likely to have changed the relative con-sumption rates of urban types by, for example,changing their social composition.

Although it is not possible to account for adifferent set of consumption rates in 1961, it ispossible to speculate about the implications of suchrates having been different from those in 1991. Inparticular, it is possible to consider the effects if thevariation in the 1961 rates between the mostefficient and most profligate urban types had beengreater or less than the variation at 1991. For thepurposes of considering this question, it is assumedthat the relative profile across urban types wouldhave been similar in 1961 to that in 1991; that is,large urban areas had the lowest per capita rates andremoter rural areas had the highest rates. Thequestion is whether the variation between theextremes would have been greater or lesser.

If the 1961 rates had shown a lesser variation,then a larger proportion of the overall change inconsumption from 1961 to 1991 would have beendue to changes in rates rather than decentralization.Thus, the percentage figures calculated here over-state the case. In other words, if the 1961 rates hadbeen used in this analysis, they would have given alower percentage result, indicating yet lower levelsof additional consumption arising from decentraliz-ation. The opposite case, then, also applies. If the1961 rates had shown a greater range than the 1991rates, then the implication is that the analysis hereunderstates the extra energy consumption that hasarisen because of urban decentralization.

The logic is clear but is it likely that 1961consumption rates would have shown more or lessvariation than 7991 rates? Some possiblereasons why the 1961 rates would have shownmore variation are as follows:

The social structure of decentralizing households,as mentioned above, is such as to increase travelin remoter areas. It is the case that many

The compact city and transport energy consumption

migrants to remoter rural areas are more affluent,and hence travel more, than the indigenouspopulation of those areas. Decentralization isthus likely to have increased per capita travelrates in remoter areas. Thus, energy consumptionrates in 1991 are likely to be higher - relative toother urban types - than they would have beenin 1961.

Urban decentralization has encouraged govern-ment to upgrade and build new roads in decen-tralized locations, thus facilitating more travel insuch areas. There is little doubt that this has beenthe case over the last 30 years. Rural andsemi-rural areas have been the recipients of majorroad investment. If the same road pattern existednow as 30 years ago, presumably people in ruralareas would now travel less (and, of course, someof them might not have migrated).

Effects that imply less variation in consumptionrates in 1991 relative to 1961 include:

The decentralization of some work and servicefacilities has increased the average mileage ofurban residents. Urban rates might thus havebeen relatively low in 1961, suggesting widervariation in rates at that time.Car ownership rates, although lower overall in1961, are likely to have shown a greater urban/rural variation in 1961 than in 1991, withconsiderably higher rates of ownership inrural areas because of the necessity of privatetravel.

If these additional factors were built into the quan-titative analysis of the lowest rate’ and ‘marginal’cases presented here, it is not obvious that theywould have any significant effect on the results.Indeed, the minor variations on the results that eachwould give might well cancel each other out,leaving the results more or less intact. This suggeststhat the crude analysis presented here may also be arobust analysis.

Conclusion

This paper has examined the remarkable politicalconsensus in a number of countries in favour ofurban-containment policies - the compact city - asa way of reducing transport energy consumptionand hence of contributing to the achievement ofsustainable development. Some of the academicevidence to support this view has been reviewed.The general drift of this evidence supports thepolicy stance. The implications of this policy

99

approach are profound. Major changes in the formand function of cities, towns and villages areimplied. Inevitably such changes will have signifi-cant social and economic consequences. Giventhat many people and businesses have consciouslychosen decentralized environments - creating thevery problem now being confronted - these conse-quences will be unpopular. In order for theseconsequences to be acceptable, containment policieswith need to be demonstrably successful in reducingtransport energy consumption.

Politicians, it seems, are expecting great thingsfrom the compact city. But what are their expecta-tions? What have they been led to expect by theacademics? The answer seems to be that there areno explicit expectations. Certainly only very rarelyhave academics promised levels of energy saving orreductions in emissions. But the implicit assumptionof both politicians and academics must be that-thebenefits are significant. Otherwise, why would radi-cal policies - and in the context of all that has gonebefore, they are very radical - be so fashionable inso many places?

It has been suggested here that, despite consid-erable attention, the compact-city debate has failedto address two fundamental questions: can urbandecentralization be halted or significantly slowed?And, more importantly, does urban decentralizationreally increase transport energy consumption to thedegree that is supposed? The first question has beendealt with only briefly here, having received fullerattention elsewhere (Breheny 1994a). The conclu-sion, however, is that decentralization remains apowerful force, at least in the UK, and may bedifficult to contain beyond existing levels ofrestraint. The second question - the major focus ofthe paper - has been-considered through a crude,but probably robust, empirical analysis. Thissuggests that energy savings from urban contain-ment are likely to be disappointingly low. Indeed,even modest savings could only be achievedthrough draconian policies of containment. Giventhe implicit expectations of politicians about thebenefits of containment, realistic savings from thisapproach are likely to be trivial.

It has been suggested that the levels of energysavings likely to result from even quite toughcompact-city proposals could be achieved in other,much simpler and relatively immediate ways. Thepromotion of improved vehicle technology and theraising of fuel costs have been suggested as twosuch ways.

100 Michael Breheny

This paper is not intended as the last word on the Bourne L 1992 Self-fulfilling prophecies? Decentraliz-merits of urban containment as a device for reducing ation, inner city decline, and the quality of urban

energy consumption. It is intended to raise doubts life Journal of the American Planning Association 58

about the validity of a political and academic move- 509-13

ment that has now gained considerable momentum. Breheny M 1992a The contradictions of the compact

There are, of course, environmental benefits ofcity: a review in Breheny M ed. Sustainable development

containment, other than energy efficiency. Indeed,and urban form Pion, London 138-59

there may be much broader economic, social andBreheny M 1992b The compact city: an introduction

Built Environment 18 241-46cultural benefits. These are important and need to be Breheny M 1994a Counterurbanisation and sustainableincluded in the debate. Nevertheless, the focus of urban forms in Brotchie J Batty M Hall P andthe environmental debate has been on energy Newton P eds Cities in competition: the emergence ofconsumption, and it is on this fundamental issue productive and sustainable cities for the 21st century

that the proponents of the compact city must Longman Cheshire, Melbourne

demonstrate - rather than simply assert - its merits. Breheny M 1994b Planning for environmental capacity: the

The statement was made earlier that, given the case of historic towns. Paper presented to The Inter-

radical and probably unpopular nature of the con-national Symposium on Urban Planning and the

tainment policies now being proposed, ‘they hadEnvironment, Seattle, Washington, March 1994

better be worth it’. The evidence presented hereBreheny M Gent T and Lock D 1993 Alternative

suggests that they will not be.development patterns: new, settlements HMSO, London

Champion A 1987 Momentous revival in London’spopulation Town and Country Planning 56 80-2

AcknowledgementsChampion A and Congdon P 1988 An analysis of the

recovery of London’s population change rate BuiltThanks are due to Roger Lee, three anonymousreferees and Richard Bate of Green Balance, allof whom offered valuable comments on earlierversions of this paper. Alan Townsend, of DurhamUniversity, gave useful advice on NOMISretrievals.

Notes

1. Newman and Kenworthy (1989a); Breheny (1992aand b); McLaren (1992).

2. See, for example Gordon and Richardson (1989);Breheny (1992a); Blowers (1993).

References

Council for the Protection of Rural England 1993 Theregional lost land Council for the Protection of RuralEngland, London

Bae C-H 1993 Air quality and travel behavior - untyingthe knot Journal of the American Planning Association 5965-74

Cross D 1991 Counterurbanisation in England and WalesAvebury, Aldershot

Banister D 1980 Transport mobility and deprivation in innerurban areas Saxon House, Farnborough

Banister D 1992 Energy use, transport and settlementpatterns in Breheny M ed. Sustainable development andurban form Pion, London 160-81

Banister D and Button K eds 1993 Transport, theenvironment and sustainable development E and FN Spon,London

Department of the Environment 1990 This commoninheritance HMSO, London

Department of the Environment 1992 Developmentplans and regional planning guidance PPG12 HMSO,London

Department of the Environment 1993 Sustainabledevelopment: the UK strategy Cmnd 2426 HMSO,London

Department of the Environment 1994 Transport PPG13HMSO, London

Barton H 1992 City transport: strategies for sustain- ECOTEC 1993 Reducing transport emissions throughability in Breheny M ed. Sustainable development and planning HMSO, Londonurban form Pion, London 197-216 Elkin T McLaren D and Hillman M 1991 Reviving the

Blowers A ed 1993 Planning for a sustainable environment city; towards sustainable urban development Friends of theLondon, Earthscan Earth, London

Environment 13 193-211Chinitz B 1990 Growth management: good for the town,

bad for the nation? Journal of the American PlanningAssociation 56 3-8

Commission of the European Communi‘ties 1990 Greenpaper on the urban environment European Commission,Brussels

Commission of the European Communities 1992 Theimpact of transport on the urban environment EuropeanCommission, Brussels

Council for the Protection of Rural England 1992 Thelost land Council for the Protection of Rural England,London

The compact city and transport energy consumption

Fielding A 1991 Migration to and from South EastEngland Report to Department of the Environment,London

Fothergill S and Gudgin G 1982 Unequal growth: urbanand regional employment change in the UK HeinemannEducational, London

Gomez-Ibanez J 1991 A global view of automobiledependence - review of P Newman and J KenworthyCities and automobile dependence: a sourcebook Journal ofthe American Planning Association 57 376-9

Gordon P Kumar A and Richardson H 1989 Conges-tion, changing metropolitan structure, and city size inthe United States International Regional Science Review12 45-56

Gordon P and Richardson H 1989 Gasoline consump-tion and cities - a reply Journal of the American PlanningAssociation 55 342-S

Gordon P Richardson H and Jun M 1991 The commut-ing paradox - evidence from the top twenty Journal ofthe American Planning Association 57 416-20

Handy S 1992 Regional versus local accessibility BuiltEnvironment 18 301-13

Jacobs M 1992 Strategic planning and sustainable develop-ment English Nature, London

Jacobs M 1993 Sense and sustainability: land use planningand environmentally sustainable development Council forthe Protection of Rural England, London

McLaren D 1992 Compact or dispersed? - dilution is nosolution Built Environment 18 268-84

MuConsult 1992 Evaluatie tussenbalans: effecten van deverandering in vervoerskosten op de mobiliteit (Evaluationof the mid-term review: effects of the change intransport costs on mobility) MuConsult, Utrecht

National Physical Planning Agency 1991 Summary ofthe fourth report extra on physical planning Ministry ofHousing, Physical Planning and Environment, TheHague

Newman P 1992 The compact city - an Australianperspective Built Environment 18 285-300

Newman P and Kenworthy J 1989a Cities and automobiledependence: a sourcebook Gower, Aldershot andBrookfield, Victoria

Newman P and Kenworthy J 1989b Gasoline consump-tion and cities - a comparison of UK cities with aglobal survey Journal of the American Planning Associ-ation 55 24-37

Newman P and Kenworthy J 1992 Is there a role forphysical planners? ]ournal of the American PlanningAssociation 58 353-62

101

Office of Population Censuses and Surveys 1993OPCS monitor PP1 93/2 final mid-1991 population esti-mates for England and Wales and costituent local andhealth authorities based on 1991 census results Office ofPopulation Censuses and Surveys, London

Office of Population Censuses and Surveys 1994OPCS monitor PP1 94/1 revised population estimate for1981-1990 Office of Population Censuses and Surveys,London

Owens S 1991 Energy conscious planning Council for theProtection of Rural England, London

Owens S 1992 Energy, environmental sustainability andland-use planning in Breheny M ed. Sustainabledevelopment and urban form Pion, London 79-105

Prism Research Ltd 1991 Mobility of industry fromSouth East England Report for Department of theEnvironment, London

Rickaby P 1987 Six settlement patterns comparedEnviromnent and Planning B, Planning and Design 14193-223

Rickaby P Steadman P and Barrett M 1992 Patterns ofland use in English towns: implications for energy useand carbon dioxide emissions in Breheny M ed.Sustainable development and urban form Pion, London182-96

Robson B 1994 No city, no civilization Transactions of theInstitute of British Geographers 19 131-41

SERPLAN 1988 Housing provision in the South East: reportby WS Grigson RPC 1230 SERPLAN, London

Sharpe R 1982 Energy efficiency and equity of variousurban land use patterns Urban Ecology 7 1-18

Sherlock H 1990 Cities are good for us Transport 2000,London

Townsend A 1993 The urban-rural cycle in the Thatchergrowth years Transactions of the Institute of BritishGeographers 18 207-21

Urban Villages Group 1992 Urban villages: a concept forcreating mixed-use urban developments on a sustainablescale Urban Villages Group, London

Wachs M 1989/90 Regulating traffic by controlling landuse: the southern California experience Transportation16 241-56

Wood C Watson S and Banister D 1994 The relationshipbetween energy use in transport and urban form Paperpresented to Universities’ Transport Studies Group,Leeds, January 1994

Yanarella E and Levine R 1992 The sustainable citiesmanifesto: pretext, text and post-text Built Environment18 301-13