Network Cities-Creative Urban Agglomerations For The 21st Century
Network Cities-Creative Urban Agglomerations For The 21st Century
Network Cities-Creative Urban Agglomerations For The 21st Century
2, 1995 313±327
D avid F. Batten
Sum m ary. T he glob al econ om y is nurturing an innovative class of polycen tric urban
con ® guration s: netw ork cities. A netw ork city evolves w hen tw o or more previous ly independent
cities, poten tially com plem entary in function , strive to coop erate and achiev e sign i® can t scop e
econ omies aided by fast and reliab le corrid ors of tran sport and com m unication s infrastru cture.
Creative netw ork cities place a higher priority on knowledge-b ased activitie s like researc h,
education and the creative arts. The coop erative m echan ism s m ay resem ble those of inter-® rm
netw ork s in the sense that each urban player stan ds to bene® t from the synergies of interactiv e
grow th via recip rocity, know ledge exchan ge and unexp ected creativi ty. T wo case studies are
discu ssed brie¯ yÐ R andstad H ollan d and K ansai, Jap an. Since m uch of their future dynamism
m ay rely upon tran snation al hum an resou rces, it is foreseen that m ore netw ork cities w ill
transce nd nation al bord ers during the next m illen nium .
1. Introduction
M ost modern scholars of urban developm ent Transnationa l forces are having an im pact
acknow ledge that transnational processes are on more than just the largest metropolitan
having an increasingly im porta nt in¯ uence centres. Another interesting example is the
on the evolution of cities. An early observ- `transfrontier metropolis’ (Herzog, 1991).
ation was the recognition of an emerging The emergence of urban settlem ents strad-
system of world cities (Hall, 1966 and 1984), dling international frontie rs re¯ ects a selec-
a kind of urban eÂlite which is shaped in tive integration of border territory into the
part by the new internationa l divisio n of circuitry of the global econom ic system. In
labour. These cities are also thought to be certain European and American borde r re-
controlling and coordina ting global ® nance gions, transfrontie r metropolis es may eventu-
(Thrift, 1986) and produc er and business ally becom e new centres of production and
services. The view of world cities as the `key urban life.
nodes’ of the internationa l urban system is a Som e urban regions are undergoing an-
widely held one, underpinned in particular by othe r kind of metamorphos is. Ef® cient
rapid advances in the developm ent of infor- corridors of infrastructure link know ledge-
mation technology and telecom munications. intensive centres to larger metropolises.
David Batten is at the Department of Infrastructure and Planning, Royal Institute of Technology, S-100 44 Stockholm, Sweden. He
wishes to acknowledge helpful comments from participants at the Lille Conference on Cities, Enterprises and Society at the Eve of
the 21st Century, March 1994. Some of the background material for this paper was collected during a recent visiting period in the
Kansai region of Japan. The author is grateful for the support provided by the Japan Society for Promotion of Science as well as
Kyoto and Tottori Universities.
European examples include the London ± Batten, 1993; Clark and Kuypers-Linde ,
Cambridge and Stockholm ±Uppsala corri- 1994). The potential superiority of the net-
dors. Each of these high-tech corridor devel- work city model may be even more pro-
opm ents incorpo rates an international airport. nounce d among many of the still-grow ing
Since the airport±university com bination turns world cities of our time, because of the ad-
out to be one of the most synergistic factors vantages of concentrated deconcentration
currently contributing to faster and more (Hall, 1984).
prosperous urban grow th in the Sw edish con- Urban theory remains preoccupied with
text (Andersson et al., 1987), such highly the monoc entric model, Nevertheless, the
accessible places nurturing a higher level of evidence against such an oversimpli® cation
know ledge-based activities may be expected is steadily growing, This paper attempts to
to prosper in the 21st century. strengthen the case for network cities by
In these bicentric urban systems, close outlinin g som e of their general com petitive
links have been forged between places of advantages and discussing two speci® c ex-
com plem entary function, rather than simply amples. B ecause the coope rative mechanism s
on the basis of distance or dem and thres- may resemble those of inter-® rm netw orks,
holds. Relational linkages tend to be horizon- the new urban theory associated with the
tal rather than hierarchical. The resulting network city model will align itself more
urban con® gurations take the form of `corri- with the econom ic notions of differentiated
dor cities’ . produc ts and im perfectly com petitive mar-
From an American perspective, the notion kets than with the classical model of perfect
of corridor cities is far from new . The his- com petition . The paper begins with a brief
torian Robert Fishm an noted: ª By the 1980s, discussion of the dynam ics of urban hier-
even social scientists could not ignore the archies.
fact that the whole terminology of `suburb’
and `central city’ deriving from the era of the
industrial metropolis had becom e obsoleteº
2. The Law of Rise and Fall
(Fishm an, `1990). Fishm an coined the term
`New Cities’ for those sprawling regions in Cities do rise and fall, albeit slow ly in many
which the basic unit is the `grow th corridor’ cases. They possess vinta ge properties, in the
stretching 50±100 miles. These new urban sense that even large and previou sly prospe r-
regions lack a dom inant single core and ous ones eventua lly decline in the absence of
de® nable bounda ries. They are multicentred. suf® cient renewal and revitalisation activity.
A small but grow ing num ber of modern The rise and fall of cities is inde ed well
urban agglom erations consist of an intricate documented in all parts of the world. To
web of corridor cities whose functional and catch a glim pse of this lifecycle phenom -
locationa l relationship s can provide them enon, we may turn to Europe over the last
with holistic com petitive advantages over millenniu m (see Table 1) or to the changing
som e of their monoc entric rivals. One classi- globa l hierarchy of cities over a much shorter
cal example is Randstad Holland. The Kansai period (see Table 2). M any cities which
or Kinki region of Japan is a particularly dom inated several centuries ago are rela-
innova tive case. Because of the manner in tively small today. Others, like Mexico City
which their polycentric structure has evol- and SaÄ o Paolo, have grow n remarkably in the
ved, these urban systems may be thought of space of just 50 years.
as `netw ork cities’ . Interest in them has in- It is often argued that city size fosters
creased recently, follow ing claims that cer- innovative prope nsity. In the US this became
tain network cities may enjoy greater known as the HVLT hypothesis (after Hoover,
diversity and creativity, less conge stion and Vernon, Lichtenburg and Thom pson), since
more locationa l freedom than monoc entric these authors emphasised the unique advan-
cities of com parable size (see, for exam ple, tages enjoyed by New York City in the pro-
City Population C ity Population C ity Population C ity Population C ity Population
Constantinople 450 Paris 275 C onstantinople 700 London 6480 M oscow 9000
Cordoba 450 M ilan 125 London 550 Paris 3330 Paris 8500
Seville 90 B ruges 125 Paris 530 B erlin 2424 London 7000
Palerm o 75 V enice 110 N aples 207 V ienna 1662 Leningrad 5500
K iev 45 G ranada 100 Lisbon 188 Leningrad 1439 M adrid 3000
V enice 45 G enoa 100 A msterdam 172 M anchester 1255 B erlin 3000
Regensburg 40 Prague 95 R om e 148 B irmingham 1248 R om e 2800
NETW O RK CITIES
Population Population
1950 2000
relative size of places in a local context are national level.) Each city’ s predicted
becom ing less im portant than they were in population is given by the form ula P 5 A/r,
the past. The urban-to-rural gap is widening where P 5 popula tion of the given city,
because of the declining emphasis on natural A 5 population of the largest city in the hier-
resources and the grow ing emphasis on cre- archy, and r 5 the size rank of the given city.
ative activity and accessibility to innova tive An illustration of this rank±size distribu tion
hum an resources via a coordin ated system of of cities may be found in France (Figure 2).
interdependent networks. C ollective creativ- Although the disprop ortionately large size of
ity calls for dyna mic interactions between the prim ate capital, Paris, tends to skew this
different ideas generating som e qualita tively distribu tion slightly , the rest of it has re-
new idea. In this new contact-intensive arena mained approxim ately loglinea r for more
of econom ic activity, a netw orking perspec- than 150 years.
tive is mandatory. At ® rst glance, there seems to be a remark-
able stability in this logline ar distribu tion.
Despite vast variations in the num ber and
size of French settlem ents, the rank±size rule
3. Network Cities versus Central Place
seems to hold over a 150-ye ar period. Sim i-
Cities
lar rank±size distribu tions may be found in
It must be clear by now that we need to many other parts of the world. For our pur-
address two seemingly con¯ icting models of pose, how ever, it is im porta nt to note tw o
the spatial distribution of econom ic activity: additional features of rank±size distribu tions:
the central place model and the netw ork
model. Fascinated by apparent regularities in (1) Som e individu al settlements change their
the siting, size and num bers of tow ns among rank over tim eÐ in response to changes
regions, W alter Christaller (1933) developed in their grow th rates, their principal
the central place model to predict ideal city- functions and their degree of nodality.
size distributions and functions. Positing that (2) The upw ard shift in the rank±size distri-
each com modity has a given threshold of bution over tim e is closely related to the
minim um dem and as well as a ® xed geo- slowly increasing mobility of travellers.
graphic al dom ain beyond which people are
unwilling to pay for it, Christaller sugge sted Both of these observations are im porta nt if
that only a certain proportion of all settle- we wish to unde rstand the respective roles
ments will offer highe r-order goods and ser- played by central place cities and network
vices. If the principal marketing function of cities over tim e.
each settlem ent is know n, a hierarchy of As activity and com munications evolve in
places can be deduced. Each node in this the course of econom ic developm ent, one
urban hierarchy has a hexagonal-shaped hin- would expect contrasting forces to emerge.
terland. W hen conceived in this way, urban The central place system is a strong and
systems will be organised into central places stable attractor in the general dynam ics of
of different order depending on their func- urbanisation, especially when competition
tional attributes. between urban econom ies may be conceived
Christaller’ s ideal model rests on assump- simply in terms of geogra phic al market ar-
tions which are not readily found in the real eas. Hence it has aided our unde rstanding of
world. Nevertheless, it is possible to ® nd how rural econom ies can support market
city-size distributions in many nations which centres of various sizes. The central place
are approxim ately loglinear, thereby con- model corresponds to the classical econom ic
form ing to the rank±size rule. (In som e model of perfect com petition in a geographi-
younge r nations such as Australia and the cal conte xt. Just as this model of econom ic
US, rank±size distribu tions may be found behaviour is less favoure d by economists
within subnational states rather than at the now adays, so it is also clear that fewer cities
Figure 2. Rank±size distribu tions of French urban agglom erations : 1831 and 1982. Source: adapted from
Pumain (1989).
and market areas of today can be draw n up Italy and the Low Countries. Such a dual
according to the central place model alone. urban core cannot be explained in terms of
W hen introdu cing their notion of a net- the traditional central place hierarchy.
work system of cities, Hohenberg and Lees The various nodes of a network city com -
(1985) emphasise the idea of nodality in a bine to form a unique yet ¯ exible exchange
netw ork of linked settlements within which environ ment. An econom ist might view a
(long-distance) traders are the prim e movers. network city as a heterogeneous product
Cities form the centres, nodes, junc tions, out- com peting in a marketplace where im perfect
posts and relays of this network. The seeds of com petition is the prevailing mechanism .
such a network economy were sown as far Their unique polyc entric structure and ¯ exi-
back as the 11th century, when safer trade bility of function breed monopoli stic advan-
route s trigge red the revival of many me- tages. Network cities exploit som e of the
dieval cities in Europe (Pirenne, 1925; Mees, opportunities arising from the diseconom ies
1975; Andersson, 1986; Batten and Thord, of urban scale in its monoc entric form. They
1994). Two major concentration s of urban may be disting uished from central place
developm ent stood out at that tim e: northe rn cities in both spatial and functiona l terms.
C entrality N odality
Size depende ncy Size neutrali ty
T endency tow ards prim acy Tendency tow ards
and subserv ience ¯ exibility and
com plem entarity
H omogeneou s goods and H eterogen eous goods
services and services
V ertical accessib ility H orizonta l accessib ility
M ainly one-w ay ¯ ow s Tw o-w ay ¯ ow s
T ransport costs Inform ation costs
Perfect com petition Im perfect com petition
over space with price discrim ination
Som e of the comparative attributes of the grow ing num ber of network city con-
tw o systems are listed in Table 3. ® gurations can com pete favoura bly with mono-
Setting aside more sophisticated economic centric cities of equiva lent populat ion size.
arguments for the tim e being, one rationale Although the pertinent calculations will not
for netw ork cities is com pellingly simple. be presented here, transport and congestion
Although som e advantages stem from the aspects serve to underpin the com parative
advent of more ef® cient long-distance com- advantages of the network city.
munication systems, this particular rationale In Christaller’ s theory of central places,
is founde d on day-to-day mobility patterns grow th potential is related to size. Thus the
(W estin and OÈ sthol, 1992). For our present case for network cities has never been ex-
purpos e, the travel behaviour of the French plored. Yet it is network cities that have
popula tion is indic ative of wider trends accounte d for an above-average share of ur-
across middle and northern Europe . Since ban grow th in today’ s Europe (see Figure 3).
1800, mobility has increased from an average Although som e larger cities possess both net-
of 20 m to more than 30 km per person per work and central place characteristics, it is
day (see Table 4). This is a factor of 1 500. the smaller network cities that have counte r-
Nevertheless, the time spent in motion by the acted the central place trend towards prim acy
average traveller has not changed apprecia- and contribu ted to the size-neutrality of ur-
bly during this century, still averaging about ban grow th (see Robson, 1973; Hohenberg
1 hour per day. Thus it is largely the speed and Lees, 1985).
of travel which has grown signi® cantly,
fuelled by the well-known sequence of modal
4. Randstad Holland: A Classical Net-
im provem ents to transport systems. The gen-
work City
eral result has been a grow ing degree of
congestion. Although the discussion about polyce ntric
The `one-hour rule’ is a strong constraint urban developm ent is predom inantly a recent
on the penultimate size of the monoc entric one, som e network cities are rather old. A
city, given that conge stion and pollutio n are classical example is the `Ring City’ : Rand-
factors of growing im porta nce. If one hour is stad Holland. Shaped like a giant horseshoe,
the tim e limit which the average urban trav- this com plex urban agglom eration conta ins
eller is willing to spend on commuting to three major conurbations grouped around the
work and/or to othe r service facilities, then cities of Amsterdam, The Hague, Rotterdam
the increased spatial range afforded by ad- and Utrecht (Figure 4). Other smaller centres
vances in transport technology means that a such as Delft, Haarlem and Zaanstad serve as
Figure 3. G row th versus size in tw o system s of cities. Source: H ohenber g and Lees (1985).
space and attractive living conditions has more dom inant, this same dive rsity may
been accommodated, but also complemented now unlock the key to a more prospe rous
by good accessibility to nearby urban jobs future .
and services and to the internationa l world. The core cities of the Kansai region are
This ongoing princ iple of concentrated de- depicted in Figure 5. B esides Osaka, Kyoto
concentration veri® es the dedicated attitude and Kobe, strong transport links ensure that
of Dutch city and regiona l planners. It also the smaller cities of Him eji, Nara, Ohtsu and
dem onstrates the durable bene® ts of the net- W akayama are ef® ciently integrated within
work city model. easy reach of the larger trio. Form ally know n
as the Keihanshin metropolitan area, the ur-
ban core of Kansai boasts a popula tion of
over 18 million people .
5. Kansai, Japan: A Creative Network
The im mense dive rsity of the region can
City
be gauged partly from the fact that it conta ins
A fascinating network city is developing in two form er imperial capitals of Japan:
the Kansai (or Kinki) region of Japan. After Nara and Kyoto. These two delightful cities,
many years of competition with Tokyo, this with their remarkable temples and artistic
historic ally pow erful area of Japan appears to treasures, provide an incom parable cultural
have recently gained greater con® dence in its richness to complement the port cities of
ability to offer attractive alternatives to the Kobe and Osaka. Osaka is also the region’ s
national capital. Kansai is made up of the six centre of com merce and industrial activity.
prefectures of Osaka, Hyogo, Kyoto, Nara, The striking contrasts between these cultural
W akayam a and Shiga. These prefectures are and com mercial cities, com bined with their
diverse in that each has its own distinc tive willingne ss to work toge ther tow ards a
cultural and historical identity. W hilst this united vision of the future , could transform
may have caused a lack of cohesiveness the Kansai region into one of the most excit-
in the decades when Tokyo was grow ing ing network cities of the next century.
The novel netw ork city pote ntial will com- identify ing new business opportu nities. W ith
bine the econom ic and social diversity of the recent openin g of the around-the-clock
Osaka, Kyoto and Kobe with the environ- Kansai International Airport, Osaka Bay
mental and recreationa l richness of Awaji Area can develop an effective gateway to the
Island and the unique features of Shikoku internationa l busine ss com munity. The cre-
Island. If the propose d channel between ative potential of such a diverse network city
W akayam a and southern Awaji Island is re- should help Kansai to face the 21st century
alised, a com plete loop around Osaka Bay with a much greater degree of con® dence.
would be established.
Today the Osaka Bay Area has four types
of port:
6. Concludin g Remarks
(1) the tw o traditional seaports of Kobe and It has been said that the econom ic geography
Osaka; of the modern industrial world can be repre-
(2) Kansai International Airport; sented as a patchwork of dense produc tion
(3) a telecom munications port know n as the agglom erations linked toge ther by an exten-
Teleport; and sive system of interregiona l transactions
(4) a know ledge port know n as the Kansai (Scott, 1992). There is a similar patchwork of
Cultural and Scienti® c Research City. creative urban agglom erations developing at
some key locations around the world which
Businessmen emphasise the openne ss of offer a unique com bina tion of characteristics:
Kansai people to new ideas, compared with an attractive, culturally diverse environm ent,
Tokyo. They also point to their ability to advanced R&D and educationa l facilities, a
make quick decisions and their talent for ¯ exible and creative workforce, im proving
accessibility to the outside world, and a dy- transcend national borders during the next
nam ic vision of the future . century.
At som e of these locations, network cities
are developing from the premise that nearby References
urban partners can bene® t from the dyna mic
synergies of interactive grow th via reciproc- A N DER SSON , A ;A N.E. (1986) Presidenti al address :
the four logistica l revoluti ons, Papers of the
ity, know ledge exchange and unexpected Regional Science Associati on, 59, pp. 1±12.
creativity. They can also achieve signi® cant A N DER SSON , A ;A N .E., A ND ER STIG , C . and
scope econom ies aided by fast and reliable H AÊR SMA N , B . (1987) K now ledge and com muni-
corridors of transport and com munic ations cations infrastr ucture and regional econom ic
infrastructure. The more creative ones place change . C ERU M W orking Paper 1987:25 , U ni-
versity of Um eaÊ.
a higher priority on know ledge-based activi- B AT TEN , D .F. (1993) N etw ork cities versus central
ties like research, education and the creative place cities: building a cosm o-creativ e constel-
arts. It is thus not surprising to learn that lation, in: A ;A N .E. A ND ERSSO N , D.F. B ATTEN ,
seven of the ten most creative Europe an re- K . K O BA YAS HI and K. Y OSH IKAW A (Eds) The
gions of toda y are corridor or netw ork cities C osm o-creati ve Society , pp. 137±150. H eidel-
berg: Springer.
(see Table 5). In these urban systems, close B AT TEN , D.F. and T HO R D , R. (1994) E urope’ s hier-
links have been forged between places of archical netw ork econom y, in: D .F. B AT TEN , J.
com plem entary function, rather than simply C ASTI and R . T H OR D (Eds) Networks in Action .
on the basis of distance or dem and H eidelber g: Springer. (forthco m ing).
thresholds. B AT TEN , D .F. and T OÈ R NQ VIST , G. (1990) M ulti-
level netw ork barriers : the m ethodolo gical
W ithin the above collection of creative challeng e, The Annals of Regional Science , 24,
regions, many R&D units operate in corpora- pp. 271±287.
tions whose production activitie s are based C HR IST ALLER , W . (1933, 1966) Central Places in
on know ledge engineering and a new sys- G erm any. N ew Jersey: Prentice H all.
tems architecture (rather than the factory- C LA RK , W .A .V . and K UYP ERS- L INDE , M . (1994)
C om m uting in restructu ring regions, Urban
oriented engineering of the industrial era). Studies , 31, pp. 465±483.
They are contact-intensive and may eventu- F ISHM AN , R., (1990) A merica’ s new city, The
ally form a netw ork of their own at the globa l W ilson Q uarterly , 14, pp. 24±48.
level: an international C-netw ork as de- F RISCH , R . (1965) Theory of Productio n. D or-
scribed by Andersson (1986). Since much of drecht: R eidel.
H A LL , P. (1966) The W orld C ities, 3rd edn 1984.
their future dynam ism may rely upon London: W eidenfel d and N icolson.
transnational hum an resources, it is foreseen H ER ZOG , L.A. (1991) Cross-nat ional urban struc-
that the more creative netw ork cities will ture in the era of global cities: the U S±M exico
M EES , A . (1975) The revival of cities in m edieval level of econom ic activity; P * m easures the size
Europe, Regional Science and U rban Econom - of the city’ s econom ically active populati on
ics, 5, pp. 403±425. (workers and custom ers); D m easures the size of
N OR TON , R.D. and R EE S , J. (1979) The product its external custom er netw ork; G * m easures the
cycle and the spatial decentra lization of A meri- city’ s stock of infrastru cture; a is an index of
can m anufactu ring, Regional Studies , 13, pp. technolo gical ef® ciency; b is an index of scale
141±151. of econom ic activity; z is an index of hum an
P IR ENNE , H . (1925/ 1992) M edieval Cities (English skills; and z 9 is an index of infrastru cture
transl. F.D . Halsey, 1952. Princeton , N J: quality.
Princeton U niversity Press. Equation (1) de® nes a producti on function for
P R IEM US , H. (1994) Planning the Randstad: be- our city of interest. D is intended to re¯ ect
tween econom ic grow th and sustaina bility, U r- the role (if any) of the city as an internat ional
ban Studies , 31, pp. 509±534. market centre. The size of its custom er netw ork
P UU , T . (1985) A sim pli® ed m odel of spatiotem - should be m easured on the relative scale of
poral populati on dynam ics, Environm ent and internati onal links in com parison with local and
Planning A, 17, pp. 1263±1269. regional links. Thus it might be w eighted accord-
R O B SON , B .T . (1973) Urban G rowth: An Ap- ing to the ratio of internati onal to intra-na tional
proach . London: M ethuen. econom ic activity.
S AB EL , C .F. (1989) Flexible specialis ation and the Econom ic activity X reaches its maxim um
re-em ergence of regional econom ies, in: P. when both m arginal producti vities are zero, w hich
H IR ST and J. Z EITLIN (Eds) Reversing Industri al occurs w hen
Decline , pp. 17±70. Oxford: Berg.
S C OTT , A.J. (1992) The collectiv e order of ¯ exible P5 G5 4b/3.
producti on agglom erations: lessons for local
The technica lly optim al scale is achieved w hen
econom ic develop ment policy and strategic
marginal and average producti vities are equal, i.e.
choice, Econom ic G eograph y, 68, pp. 219±233.
when
S HA C HAR , A . (1994) R andstad H olland: a `world
city’ ? Urban Studies , 31, pp. 381±400. (P 3 1 G 3)/(P 1 G )2 5 b/2.
T H RIFT , N. (1986) Globalisa tion, regulati on, ur-
banisatio n: the case of the Netherlan ds, U rban The term 2PG account s for synergis tic inter-
Studies , 31, pp. 365±380.
U NITED N ATION S (1980) Patterns of urban and
rural populati on growth . Populatio n Studies actions betw een the local populat ion, external
No. 68, U N, N ew York. custom ers and infrastru cture.
W ES TIN , L. and OÈ ST HOL , A . (1992) City network s It is illustrati ve to consider a situation in w hich
and the search for regiona l potentia l. CERU M G 5 0. This might corresp ond to a sim ple agrarian
W orking Paper 1992:13 , U niversity of Um eaÊ. econom y lacking any significa nt infrastru cture
Y OSH IKAW A , K. (1993) C reative renaissa nce of the links to the outside w orld. Then the m odel re-
Osaka B ay area: tow ards a cosm o-creati ve re- duces to the follow ing:
gion in the 21st century, in: A .E . A ND ER SSON , X5 a{bP 2 2 P 3} (2)
D.F. B A TTEN , K . K OB AYA SHI and K. Y OS HIKA W A
(Eds) The C osm o-creati ve Society. pp. 281± This function is plotted in Figure 6. It satisfies the
294. H eidelber g: Springer- Verlag. `regular ultra passum law ’ proposed by Frisch
(1965), in so m uch as it contains a zone of no infrastru cture links to the outside world, som e
increasi ng returns to scale and a zone of decreas- interesti ng findings em erged:
ing returns to scale. Puu (1985) propose d this
function al form as a m odificati on to Hotelling’ s (1) W hen P , b/2, a netw ork city possessi ng
populati on grow th model. nodes of different sizes is more producti ve
Such a sim plified m odel w as the basis of a than one w ith nodes of equal size.
com parative analysis of the three different urban (2) W hen P 5 2b/3, the bicentric corridor city
configu rations depicted in Figure 1 (see B atten, is as producti ve as a m onocent ric city of
1993). Each city’ s stage of developm ent in the equivale nt size.
urban lifecycle w as defined by its position along (3) W hen P . 2b/3, a netw ork city is more
the X±P curve in Figure 6. In this sim plified product ive than its monocent ric counterp art,
econom y, the technica lly maxim al scale occurs and one with nodes of equal sizes out-
w hen P 5 2b/3. T hree zones of producti vity m ay perform s one where the nodal sizes differ.
be identifie d:
It should be stressed that the analytic al fram ework
(a) increasi ng returns to scale: P , b/2; outlined above is unaccep tably sim ple. Am enity
(b) decreas ing returns to scale: b/2 , P , 2b/3; attractiv eness, com m uting distance s and conges-
and tion effects should also be consider ed. In its dy-
(c) decreas ing output: P . 2b/3. nam ic form , the model visualise d would also
include addition al interacti ve term s to cater for
Even in the case of a rudim entary econom y with migration , trade and other netw orking activitie s.