GaWC Research Bulletin 311

This Research Bulletin has been published in Environment and Planning A, 41 (11), (2009), 2550-2555.

doi:10.1068/a42235

Please refer to the published version when quoting the paper.

(Z)

Urban Economics in thrall to Christaller: A Misguided Search for City Hierarchies in External Urban Relations

P.J. Taylor^*

In my particular intellectual cocoon, I had imagined that the spatial organization and structure of cities was now generally agreed, crudely, to be a combination of two sets of materialist mechanisms: agglomeration processes creating economic clusters within cities and connectivity processes creating economic networks between cities. The former has been a very strong research area over recent decades contrasting with the latter, external urban relations, which has been relatively neglected. By my way of thinking, the two sets of processes are intertwined and between them generate successful, vibrant cities. But, it seems, my cocoon is not particularly encompassing. And it is the focus on networks with their inherent mutuality (horizontal inter-city relations) that appears to be far from widely accepted. For some urban scholars, inter-city relations can only be understood as hierarchical.

This intellectual antagonism was brought home to me when receiving a peer review report on a recent paper.¹ This paper sets out the network argument and shows how it relates to the much older hierarchical tradition of central place theory. The report is quite critical, suggesting changes that do not appear to be relevant to the paper's argument, and then finally, as an afterthought, the following statement appears:

On a more minor level, the authors appear to be saying that inter-city relations are essentially non-hierarchical which I find a little strange.

The whole point of the paper, the ‘more minor level' as the reader describes it, is that cities in networks share a basic mutuality. The fact that the referee seems to have realized this late on, and finds it ‘a little strange', indicates a reading of the paper from the perspective of a different paradigmatic mind-set fixated on hierarchy.

It is no coincidence that parallel paradigms exist in the less well-researched area of the external relations of cities in economies. Past neglect has left old paradigms in place decades after their research heyday. In this case the legacy of central place theory with its neat hierarchical placing of cities, so popular in geography from the late 1950s to the 1970s, appears to continue to be influential outside its original discipline.² For most geographers, such thinking is long past its sell-by date: thus for me what is ‘strange' is that non-geographers are still being seduced by the simplicity of city hierarchies.

The purpose of this short commentary is not to gripe at criticism, whether well founded or unfounded, but rather to explore the pedagogic hegemony that hierarchical relations continue to have for studying inter-city relations in some areas of urban studies. Hierarchical thinking can be deemed to be hegemonic in circumstances where scholars cannot envision a form of inter-city relation different from a central place hierarchy: cities are hierarchical – how else would they be arranged? This position is all the more remarkable because central place theory was first laid out in 1933 by Walter Christaller (1962) as a means for understanding the servicing of rural populations.³ Cities become the primary focus after the transmutation of the model to ‘national city systems' in which ‘national urban hierarchies' were studied (Berry 1961, 1964; Berry and Pred 1965). My preliminary exploration has suggested that it is this quantitative geography and the related regional science (Isard 1960), which have been the main source of hierarchical thinking in contemporary urban economics.

Paul Krugman (1997) has been the scholar who has disinterred old ‘spatial geographies' in his revival of economic geography as a new geographical economics. Thus he discovers Christaller's central place theory but initially is quite dismissive of it: he refers to it as merely a ‘German geometry' (p. 38) because of its lack of the necessary market structure for serious economic consideration. But he does not question the idea that cities form hierarchies and later concedes that ‘central place theory has been a powerful organizing principle for research' (p. 93). Therefore he saw it as the job of urban economists to produce ‘a model to exhibit at least some central-place features' because no model of ‘the spatial relationship of cities to each other' existed (p. 93). Thus has Christaller, and hierarchical inter-city relations, entered the purview of urban economists.

The purpose of this short commentary is to explore in detail how Christaller's central place hierarchy has been incorporated into urban economics. To this end I focus on the leading text in the field, Masahisa Fujita and Jacques-Francious Thisse's (2002) Economics of Agglomeration: Cities, Industrial Location, and Regional Growth (in which, of course, Krugman's ideas figure prominently). This contemporary classic is interrogated to see how and why hierarchies appear in their economic theorising. The first section of their title reflects the fact that it is urban economic research on spatial agglomeration that has advanced in recent years. However, the subtitle does indicate an attempt to move beyond this key topic to include external urban relations. Nevertheless the book title does indicate the fact that inter-city relations are not a central concern, they appear almost as an add-on, but, most importantly, they appear as hierarchy.

My interrogation consists of charting how understanding inter-city relations progresses through the book. I am looking for discussion of hierarchical relations, its concomitant, city competition, and especially to any reference to the alternative, city network, and its concomitant, city complementarity. Scattered amongst their dense text, I have identified six separate steps in their argument on inter-city relations.

STEP ONE: An economic landscape of no cities (The Spatial Impossibility Theorem)

This first step is about clearing the decks; it dismisses traditional ideas of economics as being antithetical to the very existence of cities.

‘For a long time, economists had problems handling complementarity effects, which can hardly be taken in account in the general competitive framework.' (p. 5)

Thus ‘constant returns and perfect competition leads to a strange “world without cities”' (p. 6); ‘such an economic space is the quintessence of self-sufficiency' (p. 7). A new theoretical start is needed for understanding cities and their external relations in an urban economics.

STEP TWO: An economic landscape of isolated cities

This step is a search for agents in ‘a market for cities' (p. 96). Two possible city developers are identified: a firm (factory town) or local government (planned town). But this step is as unsuccessful as the previous one. The first agency produces an

‘optimum city system ' [of factory towns where] ‘all cities are identical and do not trade among themselves. The city system obtained here looks like a collection of isolated cities' (p. 115)

To create trade an intervention is required to generate specialization; a city government has to be this economic agent (p. 116): ‘The role of a developer is to choose a tradable good ' (p.117, italics in original) but, it seems, city diversification eludes this particular level of theorising (p. 119).

STEP THREE: A monocentric economic landscape

Returning to inter-city relations and recognising the need to go beyond the previous step - where ‘cities are like floating islands' (p. 351) – leads the argument to central place theory. Explicit reference is made to Christaller and others who have theorised ‘a hierarchical system of urban centers' but Krugman's critique is now presented (including the geometric jibe): ‘there are no economic forces' and ‘thus it is hard to see why central places should emerge' (p. 352). To overcome this problem Thunen is brought back into the argument despite being previously criticised at the most basic level:

‘Thunen's theory left a crucial issue unexplained: Why is there a city in Thunen's isolated state?' (p. 12)

This conundrum is now overcome through a new model ‘in which city and agricultural land use are endogenously determined' (p. 353). The result is a ‘monocentric economy' in ‘a spatial equilibrium' (p. 353).

STEP FOUR: A regular polycentric economic landscape

The equilibrium of step three is undone when the population exceed a threshold. This provides an evolutionary mechanism to generate more cities:

‘as the population grows, new cities emerge when some critical population threshold is reached. When the population keeps rising, still more cities will appear, and so on. … The main result is that similar cities are created at (more or less) equal distances as the population increases continuously ' (p. 354, italics in original).

This is interpreted as a ‘formal proof of one of the key ideas of central place theory' (p. 354). Thus, unlike step three, ‘here there is no agent'. This is the crucial point:

‘City formation is the outcome of a process involving agents who do not plan a priori to create a city but rather to pursue their own interests. In other words a city appears here as a complex system whose existence is the result of a self-organizing process' (p.354).

Not quite hexagons, but geographers will recognise this result whereby cities emerge to service a rural population.

STEP FIVE: A hierarchical polycentric economic landscape

Now that central place theory has entered the argument it appears to be only natural that the next stage will be to create a city hierarchy. The previous step considered only one group of differentiated goods ‘so that all cities produce the same type of goods', which is why they all ‘have similar sizes' (p. 354). The evolutionary approach is thus extended:

‘To generate a hierarchical urban system à la Christaller' requires different groups … different transport costs … different technologies…' (p. 354).

In this way it is possible to create a hierarchical system:

‘As the population size increases … a (more or less) regular hierarchical central place system, reminiscent of Christaller's, emerges within the economy' (p. 385)

Thus, it would seem, that Christaller has finally made it into urban economic theory. But, intriguingly, this turns out not to be the case.

STEP SIX: Towards polycentric network economic landscape?

Having introduced trade and specialization into the model, it is found that cities do more than service their hinterlands; crucially, they trade with each other. This idea that cities have a ‘two-fold function' comes, literally, on the last page of text in the book that considers urban external relations (p. 385). Thus it is found that results include

‘a more intricate pattern of trade in which horizontal relations are superimposed on the pyramidal structure of central place theory.' (p. 385).

This is interpreted as being ‘more fuzzy than in the Christaller model of central places' (p. 385) but it is, of course, much more fundamental than a mere blur in the model. This is clear from Fujita and Thisse's evaluation of the different relations: ‘horizontal relations between cities of the same order may be more important than trade with lower-order cities (pp. 385-6). Although not recognised as such, what these results produce is a network model of inter-city relations interwoven with the classic hierarchical model.⁴ The final reference in their argument is to Pred (1977) and his identification of complex inter-city relations that transcend simple hierarchies.

What to make of this sequence of economic landscapes? Three points come immediately to mind. First, Christaller and his central place theory is clearly the dominant guide to what an urban economic model of the external relations should look like. Although not actually reaching a central place theory outcome, the steps are designed towards reaching this goal. As far as external relations are concerned, central place theory appears to be hegemonic in urban economics. But, second, what the authors do reach is a pattern of inter-city relations that they identify as being described many years ago by Allan Pred (1977). His City-Systems in Advanced Economies is, perhaps, the one real classic scholarly work to come from the national urban systems school. Certainly it would have made a much better starting point for Fujita and Thisse than Christaller's classic text of a generation earlier⁵; however, as an end point it is quite disappointingly dated. Third, the modelling exercise finally stumbled upon the idea that there are two different processes in the external relations of cities: central place processes involving hinterlands and hierarchy, and inter-city trading processes that go beyond hinterlands and are networks.⁶ As a horizontal process, the latter supra-hinterland relations could have been taken directly from Pred (1977) but, better still theoretically, Jacobs (1969, 1984) offers a starting point where city complementarities are central to the argument.⁷

What has been the effect of an urban economics that slavishly adheres to the central place hierarchy dictum? For all the real advances in understanding economic agglomeration, this area of urban studies has little of nothing to say about cities in globalization. Since the latter has evolved as a world city network rather than one great central place hierarchy, using Christaller as guide has cut urban economics off from new global realities; for instance, Sassen's (1991) influential study of the ‘global city' is conspicuous by its absence from Fujita and Thisse's long list - twenty pages - of references. As well as this geographical scalar challenge, their focus on a simple hierarchical spatial structure privileges inter-urban competitive relations, which means that the inherent subtleties of urban external relations cannot be reached through their urban economics research agenda. Once it is accepted that cities are both competitive and cooperative then the question arises as to the circumstances in which one inter-city relation dominates the other. For instance, in world-systems terms, the initial thoughts are that in cyclical growth phases we might expect city relations to tend towards complementaries and for this to revert to competition in the stagnation phase (see Arrighi 1993); while in spatial terms, complementarity would tend to dominate within core processes, with competition specifically characteristic of the periphery (competitive dependencies).

There is a final strange twist in the urban economist liking for urban hierarchy. In the seminal work of Powell (1990) and Thompson (2003) market, hierarchy and network are three distinctive forms of social organization. From this perspective, Fujita and Thisse's (2002) ‘steps' laid out above are an attempt to jump from market theorising to hierarchical outcome. No wonder this is found to be difficult; that they should attempt this is testimony of the hegemonic power of central place thinking in urban external relations. Breaking free from this pedagogic power, the initial decision in trying to understand cities is whether they are considered to be primarily organised as markets, hierarchies or networks.⁸ I guess what I am saying is that the problem with urban economics is that it is economics.

REFERENCES

Arrighi G (1993) The Long Twentieth Century (Verso, London)

Berry B J L 1961, “City size distributions and economic development” Economic Development and Cultural Change 9 573-588

Berry B J L 1964, “Cities as systems within systems of cities” Papers and Proceedings of the Regional Science Association 13 147-63

Berry B J L, Pred, A 1965 Central Place Studies: a Bibliography of Theory and Applications (Regional Science Research Institute, Philadelphia)

Christaller W 1966 Central Places in Southern Germany (translated from 1933 German publication) (Prentice Hall, Englewood Cliffs)

Flannery K V 1998 ‘The ground plans of archaic states' in Archaic States Eds G M Feinman, J Marcus (School of American Research Press, Sante Fe) pp.15-57

Fujita M, Thisse, J-F, 2002 Economics of Agglomeration: Cities, Industrial Location and Regional Growth (Cambridge University Press, Cambridge)

Isard W 1960 Methods of Regional Science (MIT Press, Cambridge MA)

Jacobs J 1969 Economy of Cities (Vintage, New York)

Jacobs J 1984 Cities and the Economy of Nations (Vintage, New York)

Jacobs J 1992 Systems of Survival (Vintage, New York)

Krugman P 1994 “Competitiveness: a dangerous obsession' Foreign Affairs 73 (2) 28-44

Krugman P, 1997 Development, Geography, and Economic Theory (MIT Press, Cambridge, MA)

Nissen H J 1988 The Early History of the Ancient Near East, 9000 – 2000 BC (University of Chicago Press, Chicago)

Powell W W 1990 “Neither market nor hierarchy: network forms of organization” Research in Organizational Behavior 12 295-336

Pred A 1977 City Systems in Advanced Economies (Hutchinson, London)

Renfrew C 1978 ‘Trade as action at a distance: questions of integration and communication' in Ancient Civilization and Trade Eds J A Sabloff, C C Lamber-Karlovsky (University of New Mexico Press, Albuquerque)

Sassen S 1991 The Global City: New York, London, Tokyo (Princeton University Press, Princeton, NJ)

Skinner G W (1964/5) Marketing and Social Structure in Rural China (Association of Asian Studies, Ann Arbor, MI)

Taylor P J 2001 “Specification of the world city network” Geographical Analysis 33 181-94

Taylor P J 2004 World City Network: a Global Urban Analysis (Routledge, London)

Thompson, G F 2003 Between Hierarchies and Markets: the Logic of Network Forms of Organization (Oxford University Press, Oxford)

NOTES

^* Peter J. Taylor, Department of Geography, Loughborough University, UK. Email: P.J.Taylor@lboro.ac.uk

1. The paper was not submitted to this journal; from the content of the report I have assumed the author to be an urban economist.

2. My favourite example is in archaeology: Renfrew (1975), using quantitative methods ‘inspired by geography' (p. 3), has argued that ‘a permanently functioning central place system is a feature of every civilization' (p. 12). In practical application, central place theory is used to identify the existence of a ‘civilization' from surface surveys; to qualify a region has to have at least a four-tier central place settlement hierarchy (see Nissen (1988); Flannery (1998)). However, in this commentary I discuss only the hegemony of central place hierarchy in urban economics.

3. For instance, in one of the most influential application of this theory, Skinner (1964/65) consistently refers to the model as analysis of rural marketing.

4. Fujita and Thisse (2002) do refer to ‘the urban system as a network of cities' (p. 382) but they do not recognise this as a different structure to central place theory. Indeed, they refer to 'a highly regular network of cities, as conjectured in central place theory' (p. 384), before their derivation of a city hierarchy.

5. When used as a starting point it de-privileges hierarchy and points towards network, see Taylor (2004, 18)

6. Ironically, it is this third point that was the subject matter of the peer-reviewed paper mentioned at the beginning of this commentary.

7. Jacobs (1969) does appear in the text as a contributor to the economics of agglomeration (for this she has been lauded by Krugman (1997, 5) as ‘something of a patron saint of new growth theory') but the external urban relations that are intrinsic to her theory is missing.

8. From my perspective, markets are the starting point for understanding firms (Krugman 1994), hierarchies for understanding states (Jacobs 1992), and networks for understanding cities (Taylor 2001). Of course, the real interesting points come with finding hybrids of these organizational forms.