STATUS QUALITY AND THE OTHER TRADE OFF t

Urban Studies, Vol 37, No 1, 7± 35, 2000 The Donald Robertson Memorial Prizewinner 2000 Status, Quality and the Other Trade-off: Towards a New Theory of Urban Residential Location Hoang Huu Phe and Patrick Wakely [Paper ® rst received, April 1998; in ® nal form, November 1998] Summary The existing models of residential location are facing dif® culties in explaining new trends in urban developm ent such as gentri® cation and abandonmen t The mainstream approach which stresses the bid-rent formulations and the access/space trade-off seems to be at variance with the current reality of dispersal of both industry and housing in modern cities In this paper, it is proposed that the focus on the city centre(s) and distance(s) from it (or them) should be shifted to two other categories of parameter: housing status and dwelling quality A model of interaction between these parameters can be used not only to describe but also to predict various types of residential development in different urban contexts The components of a new theory of residential location are proposed Urban Residential Location: Theories and Realities Ever since von ThuÈ nen (1826/1968) gave his version of the concentric city-region, geographers, urban economists and planners have been working on theories of city structure that can both explain and predict the ways in which cities are formed and have evolved Full accounts of these efforts have been given by Hallett (1978), Hudson and Rhind (1980), Maclennan (1982), Kivell (1993) and Balchin et al (1995) As housing makes up the major part of the land of cities, the theories of city structure are, to a large extent, about residential location In a broad sense, the theories of residential location fall into two main groups: the market approach and the non-market approach The market approach is championed by the urban economists, although it has its origins in the sociological observations of the Chicago School in the 1920s (Maclennan, 1982) Three main theories are used to explain private-sector housing location: travel-cost minimisation; the travel-cost/ housing-cost trade-off; and maximum housing expenditure A detailed summary of these theories is given in Balchin et al (1995) Of the three theories, the second is the most widely accepted and, for this reason, has become the most developed theory of residential location It basically states that, given an opportunity, a perfect mobile household would move to a plot where it can satisfy its spatial requirements while paying acceptable transport costs; that is, to make the access/ Hoang Huu Phe and Patrick Wakely are in the Development Planning Unit, University College London, Endsleigh Gardens, London WC1H 0ED, UK Fax: 44-171-387-4541 Email: phebinh@cix.co.u k (or phehoang@hotmail com) and p.wakely@ucl.a c.uk 0042-0980 Print/1360-063X On-line/00/01000 7-29 Ó 2000 The Editors of Urban Studies HOANG HUU PHE AND PATRICK WAKE LY space trade-off in the way proposed by Hurd (1903), Isard (1956), Alonso (1964), Muth (1969), Evans (1973), Romanos (1976) and Thrall (1987) Although reservations have been expressed, even in its early days (Firey, 1947), the theory’ s success in replicating empirical regularities in Western cities has made it a preferred analytical tool (Basset and Short, 1980) It is now ª the dominant paradigm of urban economic researchº (Maclennan, 1982, p 7) and, as recently as in 1989, an attempt was made to extend it into a normative theory (Fujita, 1989) The non-market approach has a substantially longer history than the ® rst It dates back to the writings of the classical thinkers such as Plato and Aristotle (Roll, 1990) Basically this approach analyses the residential distribution and city structure in terms of social groups inhabiting urban areas, with some groups taking advantage or control over others In its modern form, the nonmarket approach is principally represented by sociologists, geographers and political scientists (Rex, 1968; Pahl, 1975; Harvey, 1973; Smith, 1987) Its strongest points tend to critique the market-based theories that are characterised by their ahistorical outlook They point out that instead of resulting from market competition, land and house priceÐ and, by extension, residential location patternsÐ are in fact strongly in¯ uenced, indeed manipulated, by landed capital through monopolistic rent (Harvey, 1973; Smith, 1987) If disagreement between the two approaches largely concerns their theoretical underpinning s, their application is facing no less serious a challenge Both market and non-market approaches have shown considerable discrepancies when applied to real-life situations Among the phenom ena that the market approach fails to explain, gentri® cation is a very telling example The trade-off theory maintains that the rich have a natural propensity to live in large properties in the suburbs, where the land is cheap and the environment is good, because they can afford the transport costs The poor live in the inner city because they cannot pay for high transport costs Despite a heavy degree of simpli® cation, this was more or less accepted until the 1970s and 1980s, when the rich moved into derelict areas in the inner city, renovated them and stayed, in a widespread phenomenon later called gentri® cation (Hamnett and Williams, 1980) No satisfactory explanation was offered without the risk of contradicting the fundamental assumptions of the access/space trade-off theory Similarly, the increasing incidence of abandonment in housing estates of largely decent physical quality and low rent, implying the dispersal of job and residenceÐ in other words, the everyday realities in modern citiesÐ also seems to be at variance with the access/space trade-off theory, which stresses a purely economic, rational and mechanistic behaviour of the residents It was left to the non-market urban theorists to deal with the new discrepancies, such as Smith’ s rent-gap theory for gentri® cation (Smith, 1987), and a series of post-m odern interpretations of the new urban realities (Harvey, 1991; Soja, 1996; Dear and Flusty, 1998) Although some of these theoretical lines are more successful in offering explanations for the new urban phenomena, the non-market approach’ s underlying point about monopolistic rent has never really been supported by empirical data (Kivell, 1993) Being absorbed by the role of capital, proponents of this approach have a tendency to ignore some powerful impetuses for residential location, such as life-cycles, personal preferences and taste A satisfactory theory of residential location and city structure thus still remains elusive or, as Dear and Flusty (1998) commented, a `scarce commodity’ There is a paradox in the seemingly wide acceptance that, despite all their problems, the older and more simple residential location theories are probably more workable after all (Balchin et al., 1995) At the same time, this is also a clear recognition of the limitations of existing theories Some reasons for this can be identi® ed For the market approach, perhaps the biggest obstacle that prevents it from achieving not only a close description, but also a satisfactory explanation of residential loca- A NEW THE ORY OF RESIDENTIAL LOCAT ION tion, is the excessive reliance on physical, measurable variables, the meanings of which could undergo fundamental changes during different historical times The role of the city centre and the distance to it is the clearest example It is now a truism that the centre of even the monocentric city, if it exists at all, no longer plays the overwhelmingly important role assigned to it as the only concentration of employment, nor does the distance from it act as the de® ning constraint In many countries of the West, some downtown areas could be transformed from lively shopping cores into deserted blocks if the out-of-town shopping mall developers had their way With the new trends of dispersal, the new industrial and of® ce developments (especially the high-tech ones) are no longer dependent on distanceÐ either to the marketplace or to sources of material supplyÐ for their survival, as a result of modern means of mass transit and air transport, telecommunications and computer networking The non-market approach does not suffer such an obvious obsolescence, mainly because it is based either on human relationships or on relatively stable notions such as the government and the planning organisations, etc Their underlying concepts, such as monopolistic rent and the working of institutions, however, are appealing but dif® cult to prove and quantify in reality Recognising that the neo-classical models achieved important results in replicating patterns of residential location, they call for a combined theory that can incorporate both the use value and the exchange value aspects of residential locations But so far no visible progress in this direction has been recorded (Kivell, 1993) It should be noted that most of the existing theories of residential location are the direct results of observations of Western cities, especially in the US, within a particular historical time (the ® rst half of the 20th century) There were attempts to create speci® c theories of city structure in different geographical regions such as east Asia (Rimmer, 1991; Sit and Yang, 1997), south-east Asia (McGee, 1991) and eastern Europe (French and Hamilton, 1979; Szeleny, 1983), but these efforts are being challenged by the unstoppable process of globalisation According to the concept of global urban hierarchy (Hall, 1984; Friedman, 1995; Sassen, 1995), the structure of a city depends on its position in the global urban network On the other hand, the enormous footloose capital circulating around the world can change any city almost overnight, and cities on its circuit may appear more similar to each other than to the other cities in the same country or region (a similar phenomenon was observed much earlier, during the peak of the colonial era (see Clark, 1996) Thus the traditional division of identity between cities in developed countries and those in developing countries may not be as clear-cut as in the past Therefore, it is no surprise that there are serious dif® culties in trying to apply the existing theories, based largely on historical data, to urban centres that are governed by completely new and different processes Thus the need for a new theory is selfevident In seeking to overcome the inherent dif® culties mentioned above, this paper is an attempt to reconcile the differences between the market and non-market approaches, by adopting a more ¯ exible line of reasoning It builds mainly on the social aspects of residential location, which are believed to be capable of offering a more appropriate explanation, while acknowledging the need to incorporate the methods and techniques that make it easier to quantify and describe patterns of urban residential location It should be noted that although the case study refers to Hanoi, because of access to suitable data, the proposition is by no means con® ned to cities in developing countries A New Look at the Dynam ics of Urban Residential Areas There are reasons to believe that in many real urban contexts, the logic underlying decisions in residential location is quite different from the assumptions of the access/space trade-off model Some of these reasons are new, but most others are simply ignored because they not ® t the restrictive assumptions of the trade-off models 10 HOANG HUU PHE AND PATRICK WAKE LY First, in the households’ decision-making process relating to residential location, a signi® cant role is attached to social status (Maclennan, 1982), especially in societies with a strong strati® ed structure It can be argued that social strati® cation may stem from any form of differentiation (power, wealth, knowledge, culture, etc.) and can take different forms of expression, some of the most important of which being the hom e and its location (Lawrence, 1987; Cooper, 1972) Secondly, physical distance has become less and less important with the dispersal of employment centres and increased personal mobility The information revolution of recent decades, with its computer networking and the internet, is fast shrinking the dom inance of physical distance (Harvey, 1991; Dear and Flusty, 1998) Thirdly, the demand for living space is known to be adjustable within very wide margins (Rapoport, 1977) and, in many instancesÐ especially in the more traditional societies of developing countriesÐ it is often sacri® ced for other needs, such as the desire for an extended family to live together, or the different forms of consumption (often conspicuous), including those of culture and traditions Fourthly, the historically and culturally conditioned perception of the signi® cance of the place (Bachelard, 1958; Tuan, 1961) is playing a leading role in decision-making processes The belief that a place has more than the observed physical properties is widely recognised and dates back to very early (even prehistoric) practices of ® nding appropriate locations for settlements, in which cosmological, religious prescriptions and health concerns are merged into some sort of `divine set of rules’ In many cultures with strong links with tradition, such as those in many developing countries, these factors and sentiments cannot be simply ignored in explaining residential location behaviour On the other hand, in developed countries, new trends in urban lifestyles have emerged with strong environmental and spiritual contents (Lawrence, 1998), which have started to challenge the very economistic rationality on which the mainstream residential location theory is based 2.1 The Housing Status/Dwelling Quality Relationship Taking into account this apparent move towards a more diverse vision of urban life, the dynamics of change in the residential areas of cities can be conceptualised as consisting of a simultaneous shift along two dimensions; one accommodates historical transformations, whilst the other captures the relatively permanent character of the physical environment These are housing status and dwelling quality Housing status is a measure of the social desirability attached to housing in a particular locality It can represent wealth, culture, religion, environmental quality, etc., depending on the current value system of a given society and, as such, it is closely related to concrete historical conditionsÐ i.e the temporal dimension The measurement of status can be carried out, either through the estimation of a proxy, by a ranking process (with the use of focus groups, for instance) or by estimating the `implicit’ prices of attributes related to status using different regression techniques, such as hedonic analysis (Griliches, 1971; Rosen, 1974; Megbolugbe, 1986) In any case, with the computing means and methods that are now available, the quanti® cation of status is no more complex than the de® nition of such an abstract quantity as `a unit of housing service’ in the access/space trade-off model Dwelling quality includes physical, measurable characteristics such as ¯ oor area, number of bathrooms, number of stories, etc To these can be added indicators of product quality, such as durability, compatibility with a given construction technology, etc Very often these characteristics, separated from their status content, make up the bulk of housing condition statistics It can be argued that the very neutrality of dwelling quality measurements has created part of the seeming paradox of the simultaneous presence of A NEW THE ORY OF RESIDENTIAL LOCAT ION both housing shortage and ¯ oor space redundancy in many market economies This suggests that many housing units are rendered un® t for even being classi® ed as housing because they lack acceptable attributes of social status Since the physical standards of housing differ substantially from locality to locality and, likewise, the subjective criteria of housing status also differ considerably from society to society, in any particular urban context, an identi® able characteristic relationship of the two components may be found As these two componentsÐ dwelling quality and housing statusÐ can be either compatible or antagonistic to each other, a simple graphic representation of their interaction using satisfactory methods of measurement can be shown as capable of describing nearly all possible types of housing in almost any society In a very simple form, the idea behind a new theoretical model of housing dynamics can be illustrated by a series of graphs (Figure 1) In these graphs, the O± DQ axis represents dwelling quality, while the O± HS axis represents housing status At points on the O± HS axis, there is a threshold, or an acceptable dwelling quality level, below which housing can be classi® ed as sub-standard or undesirable The dotted line, connecting these points and forming an angle a with O± HS, is the threshold line between desirable and undesirable housing For any particular socioeconom ic setting, this line will have a unique position, but for simplicity of comparison, the line is drawn in its general position Examples A±F shown in Figure only re¯ ect some of the most familiar situations, and it is clear that the model is capable of depicting many more possibilities It is postulated that in real life the threshold line would be a more complex curve (Figure 2), than a straight line The hypothesis can be expanded, leading to some interesting spatial consequences of quality±status relationships In Figure 3, a hypothetical city is shown with three status poles The HS± DQ relationship as shown in Figure is expressed three-dimensionally, where the threshold lines for each of the 11 three poles together form a `threshold surface’ , above which housing units can be perceived as `desirable’ , and below as `undesirable’ Effectively, this surface divides any city into two parts: a `dual city’ expressed in spatial, three-dimensional terms It is interesting that, while the complex relationships can be formalised mathematically, their graphical representation seems to have an intuitive appeal Elements of a New Urban Residential Location Theory The plausibility of the housing situations shown in Figure suggests the possible rewards to be gained from a systematic analysis of the principles behind the formation and probable trends of the residential structure of cities, on the grounds of the relationship between housing status and dwelling quality The preliminary formulation of a new theory of urban residential location, the quality± status theory, is proposed below The residential location patterns of most cities conform to a polar structure, in which one or several poles represent the highest points of certain kinds of social status, recognised by a given proportion of the population The parameters of social status embrace such qualitatively distinctive notions such as wealth, political power, business, culture, ethnicity, education, etc The distribution of social groups is based on the following principles: (1) Residential areas in cities make up largely continuous and overlapping rings around the status pole or poles The ring pattern is the outcome of a trade-off between that desirable status and an acceptable level of dwelling quality (explained below) (2) House value for any social group consists of two components: housing status (HS) and dwelling quality (DQ) Housing status is a combination of attributes, often non-physical, that distinguish different levels of housing desirability, or status, which are accepted 12 HOANG HUU PHE AND PATRICK WAKE LY Figure Housing status (HS) and dwelling quality (DQ) in different social contexts: A Housing in a society that values social status more than the physical qualitiesÐ for example, in primitive settlements, where houses are located according to cosmological prescriptions and tribal hierarchy, where the most sought-after dwellings may not be the ones with the highest physical quality B Housing in a culture which does not of® cially recognise differences in social status Dwellings may have medium to high quality but almost no distinction in statusÐ for example, on public housing estates in (former) socialist countries, built in locations which are `rational’ in physical terms, but devoid of social meaning The cluster is not necessarily of low-value on O± HS C Two types of housing sharing the same high-status locationÐ for example, those located under the threshold line, which are the slums and squatters, and those located above the threshold line, which are the dwellings of the gentri® ers D Housing in a `normal’ capitalist society, with dwelling quality and housing status largely compatible with each otherÐ i.e lower-status housing has correspondingly lower dwelling quality, and high status, higher quality This type of housing alignment implies strong socio-spatial segregation in the city structure E Housing stock in a supposedly `classless’ society, which has been inherited from a status-conscious society Dwelling quality is varied, but all the houses share the same level of statusÐ for example, old quarters of the former socialist cities or housing vacated by a deposed regime which is then inhabited by residents with different values F Housing units of roughly the same standard but differing considerably in social statusÐ for example, public housing projects in capitalist societies built to the same physical standard, but located in areas with different status levels, which makes essentially similar housing units very different in their desirability A NEW THE ORY OF RESIDENTIAL LOCAT ION 13 i.e by their use value, while houses at the higher price levels are characterised more by the attributes that make them commodities and favourable investmentsÐ i.e by their exchange value Figure The smoothed threshold line, representing integrated thresholds of dwelling quality and status which is steep at the lower dwelling quality and housing status levels and more gradual, almost ¯ at, at the higher end of the dwelling quality and housing status Dwellings located below this line would be perceived as ranging from undesirable down to slums and shanties HP is the minimum quality level for a given level of status by certain social groups, sometimes irrespective of the actual physical state of the dwelling Dwelling quality embodies the physical, measurable elements that constitute the basis for the normal use of a dwelling (3) At any level of housing status, there exists an acceptable level of dwelling quality, or point, below which houses are considered as sub-standard The locus of these points forms a line called the dwelling quality threshold (Figure 5) This threshold divides the whole housing stock in question into two zones: the zone above the threshold is called `desirable’ ; the zone below it is called `undesirable’ Each housing situation (of a country or city) has a uniquely characteristic quality threshold, that can be compared with others (4) At the lower price levels, dwelling quality is the dominating component, while at the higher price levels, housing status predominates With a certain degree of simpli® cation, it can be said that housing units at the lower price levels are mainly characterised by their utility as shelterÐ On the surface, there seems to be a clear parallel between this model and the access/ space trade-off model Indeed, both re¯ ect the common concentric ring pattern of residential location in cities The differences, however, are substantial First, while the access/space trade-off model puts the physical centre at the focus of importance, the new model looks at the factors of status which make that centre important These, as mentioned before, could be wealth, political power, employment, culture, ethnicity, education, etc It is implied that if those factors are changing or shifting, the physical centre cannot stay where it is, without losing its role Thus in the quality±status theory, there is nothing strange, for example, in the simultaneous movement out to the suburbs and back to the centres: different groups are simply attracted to different status poles As the poles shift, following (or signalling) transitions in society at large, they change the spatial boundaries between desirable and undesirable zones Thus, some traditionally desirable housing areas may become less desirable, even slums, prompting the ¯ ight of the middle class, and later of other groups, ultimately leading to abandonment In the same way, it is also easy to see how some ordinary dwellings, or areas, become fashionable (see Figure 6) stimulating changes that are then magni® ed by the commercial interests of developers Indeed, it is not uncommon for developers (private or institutional) to initiate such change in the interests of uplifting the market Secondly, in the access/space trade-off model, the distance to the centre is an unambiguous, physical quantity (provided the urban boundaries are clear-cut, which they often are not) In the quality±status model, the housing status axis begins from where the status in question is lowest This point is not necessarily located at the edge of 14 HOANG HUU PHE AND PATRICK WAKE LY Figure Possible application of the model for mapping a multi-polar city the city, it is where the new sphere of in¯ uence of another pole begins (with the possibility of some overlaps) It can be de® ned by different means: the price of a particular type of housing stock, occupation density, the presence of certain social networks, etc This rather abstract distance can be calibrated in terms of physical distance in relation to the real-life delineation of the boundaries A NEW THE ORY OF RESIDENTIAL LOCAT ION 15 Figure Housing status and dwelling quality, a hypothetical model including different levels of status At the lower-quality level, a small change in status corresponds to a much larger change in dwelling quality At a higher level, the changes in dwelling quality and status are reversed At the highest-quality level, the changes in dwelling quality become negligible or impossible, while status can be changed presumably only by moving or, in extreme cases, by creating new types of status Houses in these groups have the theoretically highest-possible physical qualityÐ i.e it is impossible further to improve the quality in a rational way The status, however, can be added or createdÐ for example, by being associated with special events or personalities Figure Housing status and dwelling quality Following the logic of the quality±status theory, it can be seen that if every group is equally attracted to employment in the CBD, the two models would indeed seem very similar: status would be measured by the closeness to central-city employment, so that the physical distance from the centre would coincide with the status distance Thus, it can be argued that there is indeed a relationship between the access/space trade-off model and the quality±status model, with the former being a special case of the latter: that is, when employment in the city centre is assumed to have the highest desirability for everyone, and wealth is the only indication of status Instead of the static character of the 16 HOANG HUU PHE AND PATRICK WAKE LY Figure The shift of a status pole and changes in standing of different types of housing: if the pole NL shifts to N’ L’ , the boundaries of desirable and undesirable zones also change In the case shown, R1 remains within the undesirable zone; R2, however, moves from desirable to undesirable Only the position of R3 improves considerablyÐ it remains ® rmly within the desirable zone while becoming much closer to the pole access/space trade-off model (Knox, 1994), the quality±status model, through its polar mechanism, can transmit societal changes, which are making cities very different places compared with themselves a few decades ago, into the everyday urban scene Implications of the Proposed Theory 4.1 The Other Trade-off s The access/space trade-off theory is constrained by quite restrictive assumptions about the market, behaviour, preferences and taste (Maclennan, 1982) For a considerable part of the urban population, including those at the lower end of the housing market, where the burden of the transport costs is modi® ed by public facilities and travel time is not really the biggest obstacle, there must be other types of trade-off that can be made Theoretically, at any dwelling quality level there is an unlimited range of possibilities of housing status In reality, however, they are accompanied by serious conditionalities Basically, in making decisions about their housing, any household can have two types of trade-off: The trade-off with a ® xed dwelling quality level The great majority of households have very concrete ideas of what type of housing they needÐ that is to say, they have a predetermined image of what their projected residence should look like This is as common amongst start-up home-owners as it is for pensioners looking for their quiet retreats, and all those in between In this type of trade-offÐ that is, when the quality level of a housing unit is kept ® xed (Figure 7)Ð a household may choose between different locations, but all choice patterns fall into two main scenarios: Scenario A housing unit is on the left of the threshold line (point A in Figure 7) In this case, although the status of the unit is low, it nevertheless belongs to the `desirable’ zone Scenario A housing unit is on the right of the threshold line (point B in Figure 7) In this case, despite the fact that its potential status may be higher (located nearer to the status pole) this housing unit belongs to the `undesirable’ zone It can be seen that, rather than being be- A NEW THE ORY OF RESIDENTIAL LOCAT ION 21 Figure The urban development of Hanoi since 1010: A Thang Long (Hanoi) c.1010; B Thang Long (Hanoi) c.1873; C Hanoi Province c.1936; D Hanoi, 1981 boundaries Sources: National Institute for Urban and Rural Planning (1994); Vuong (1977) frastructure and the mushrooming of commercial development and construction of high-rise of® ce blocks near the centre At the same time, there has been a boom in residential construction: at any given time over the past decade, some 25 per cent of all households were involved in some form of housing development, and a functioning housing market has been established (UNDP, 1990) The built environment, however, has turned out to be relatively inert A quick look at Hanoi’ s urban form today can still identify three historically distinctive morphologic al zones: the Old Quarter, the French Quarter and the New Areas (Figure 10) In terms of population density, however, the city shows a clear pattern of concentric rings Although there are indications that new sub-centres are emerging, for the time being the city is basically monocentric in terms of population density (Figure 11) This can be explained by the perceived import- 22 HOANG HUU PHE AND PATRICK WAKE LY Figure 10 The three zones of Hanoi’ s morphology Note the difference in pattern of the main street network laid out in each zone: Zone is the Old Quarter; Zone is known as `The French Quarter’ ; Zone 3, the `New Areas’ , has a much lower density of street network coverage ance of the traditional downtown area, as a commercial quarter and especially as a cultural focus (Hoang Huu Phe and Nishimura, 1990) 5.2 Housing Conditions and the Survey The better to understand the evolution of urban housing in Vietnam, a chronological table is presented (see Table 1) Nationally, housing has always been one of the best indicators of economic progress, thanks to the traditionally high priority given to the family home as an institution, and perhaps especially because of the very low general standards of space: the average ¯ oor space in central Hanoi is only a little above sq m per capita Squatter settlements, however, are conspicuously absent by comparison with other cities in the region at the same income 1945 Independence 1946 Resistance War 1954 Geneva peace talks, country divided Low urbanisation Dislocation of population Consumption Private Detached houses Row houses Main events Urbanisation Perceived nature of housing Main mode of urban housing production Prevailing urban housing forms produced 1945±54 North Row houses High-rise communal housing estates Detached houses South Row houses Detached houses Squatters South Row houses Detached houses South Private South Private North Row houses High-rise communal housing estates Detached houses North Destruction of stock Shelter South North State subsidies South Consumption North Shelter South Refugees ¯ ocked to cities: forced urbanisation South Unplanned urbanisation Military installations Processing industries North Non-productive sector Social service North Evacuation Dispersion of settlements: zero or negative urbanisation 1965 American troops introduced to the South Destructive air war in the North North Restrained, planned urbanisation New industrial towns South Consolidation of power Repression Successive coups North Socialist industrialisation Collectivisation 1965±74 Row houses High-rise communal housing estates Detached houses ?(Slums squatters) State subsidies reduced, work-units housing production Non-productive sector Social service Income-generating means Restrained urbanisation Population relocation Economic zoning 1975 Vietnam war ended; economic embargo imposed 1976 Uni® cation 1975±84 Row houses Detached houses ?(Slums squatters) Private Consumption Income-generating means Investment Urbanisation boom 1986 Reform of economic policy introduced 1987 Promulgation of a liberal Investment Law 1985±95 Table Urban housing development in Vietnam: a chronology 1955±64 Row houses Detached houses ?(Slums squatters) Private Consumption Income-generating means Investment Status symbol Continuing urbanisation 1995 Full Integration into global economy: embargo lifted, joining ASEAN 1995 onwards A NEW THE ORY OF RESIDENTIAL LOCAT ION 23 24 HOANG HUU PHE AND PATRICK WAKE LY Figure 11 Population density in Hanoi, 1995 Source: Hoang Huu Phe and Orn (1995) level This may be attributed to both the egalitarian housing provision policies of recent decades and cultural habits Contrary to the common perception of a centrally planed economy, Hanoi has quite a high proportion of privately owned houses: more than 50 per cent in the central area and 43 per cent for the Greater Hanoi area (1989 Census) State-provided housing units make up less than 30 per cent of the total stock Although all land of® cially belongs to the state, private transactions in property were legalised in 1991, making explicit what has always been an active, if volatile, housing market Among the limitations of the ¯ edgling housing market in Hanoi, two issues stand out: the lack of appropriate hous- ing ® nance institutions and the need for enforceable planning regulations A household survey for a study on home improvement in Hanoi was conducted on a random basis in 1993 (Hoang Huu Phe, 1997) Of the 84 phuongs (city wards), were selected to represent the urban form zones mentioned above (Figure 12) Overall, a sample of 243 households was selected for detailed interviewing and observation The data set collected included variables which are divided into categories: household characteristics, dwelling characteristics, home improvement and related attitudes, locational characteristics, impact of planning, neighbourhood characteristics, and preference and taste For the purpose of this paper, A NEW THE ORY OF RESIDENTIAL LOCAT ION 25 Figure 12 The phuongs surveyed these categories were combined into two groups: housing status (HS) and dwelling quality (DQ) The key variable, house price in taels of gold (each tael equals approximately US$350), was obtained by asking the owners, and later veri® ed by the persons in charge of housing at the ward of® ces 5.3 Housing Status and Dwelling Quality in Hanoi A partial application of the proposed quality± status theory to Hanoi is re¯ ected in the conceptualisation of its housing dynamics (Figure 13) From the survey data, it was found that housing status in Hanoi is in¯ uenced as heavily by the city’ s history as by the new opportunities of a marketoriented economy Thus, in a double-log regression on house price, the traditional perception of high incomes being earned from trading activities always gives a strong boost to the price of houses with direct and easy access to the street (variable ACCESS) Likewise, the perceived status of living in the Figure 13 Housing dynamics in Hanoi 26 HOANG HUU PHE AND PATRICK WAKE LY 27 A NEW THE ORY OF RESIDENTIAL LOCAT ION `French’ Quarter (variable FRENCH), where most of the representative of® ces of the foreign and large domestic ® rms are located, greatly increases house prices Of the dwelling quality group of variables, plot area (variable LOTAREA) and number of storeys (STOREY) turned out to be the most important 5.4 The Survey Results and the Quality± Status Theory If the proposed model of residential location is correct, it is to be expected that, to some extent, the housing data on Hanoi (or of any other city) can be used to illustrate the theoretical points discussed aboveÐ namely, that the value of a house consists of two main components: housing status (HS) and dwelling quality (DQ), with the proportion of HS in the house value increasing and the proportion of DQ decreasing as house price goes up Theoretically, if a proper regression on house price including both components is selected, it can show the behaviour of each of these components The variables are selected among those which were most signi® cant in the double-log equation mentioned above A following regression equation is suggested: Price a b 1LOTAREA b 2STOREY b NEARNESS b ACCESS b FRENCH u Where, a is the intercept; b 1, b 2, b 3, b 4, b are partial regression coef® cients; u is the disturbance term; LOTAREA is the lot area in sq m; STOREY is the number of storeys; NEAR- NESS is 1/distance to city centre (in minutes); ACCESS is the presence of good access to the street; and FRENCH means being located in the French Quarter It is clear that the ® rst two continuous variables represent the dwelling quality (DQ) component The last three variablesÐ NEARNESS (continuous), ACCESS and FRENCH (both dichotomous)Ð represent housing status (HS) These are selected because they embody the elements that constitute the status, or the widely recognised desirability, which is attached to housing in the downtown area as a commercial and cultural pole Obviously, if the data set had been speci® cally designed for testing the theory, more variables could have been selected The linear functional form was deliberately chosen, for the sake of clarity of the concept, though other forms would be likely to give a better ® t The same equation was applied to both the whole sample (243 households) and to a smaller sample (73 households) that included all the houses that were newly built after 1989 (in the home improvement study, this group was found to have the strongest response to the market) The results of the regressions are presented in Tables and As the results show, the regression gives quite a high adjusted R2 for a simple, linear model For the sample of newly built houses, more than 60 per cent of house prices are explained by the chosen functional form For both samples, nearly every variable is statistically important in the equations Although the presence of a large intercept points to Table Regression results for the whole sample Variable ACCESS FRENCH LOTAREA NEARNESS STOREY (Constant) Multiple R R2 Adjusted R B 59.01984 56.63321 0.49346 251.79530 36.18439 81.00879 SE B Beta 11.46914 12.95825 0.09296 86.17825 8.52190 14.57125 0.28401 0.22763 0.27553 0.15941 0.23621 0.60743 0.36897 0.35566 T Sig T 5.146 4.370 5.308 2.922 4.246 5.559 0.0000 0.0000 0.0000 0.0038 0.0000 0.0000 28 HOANG HUU PHE AND PATRICK WAKE LY Table Regression results for newly built houses Variable ACCESS FRENCH LOTAREA NEARNESS STOREY (Constant) Multiple R R2 Adjusted R B 38.89082 11.62654 0.44986 437.01740 23.71131 62.83496 SE B Beta 9.72570 11.99403 0.13019 71.71292 7.58555 14.57123 0.31407 0.07201 0.25607 0.46777 0.25142 T Sig T 3.999 0.969 3.455 6.094 3.126 4.312 0.0002 0.3359 0.0010 0.0000 0.0026 0.0001 0.80480 0.64770 0.62141 some non-linearity of the relationship (Kain and Quigley, 1975), for the purpose of this paper, it can be ignored In Figures 14 and 15, the predicted weighted values of HS and DQ for both samples are plotted As predicted by the theory, at the lower price ranges, the increase of DQ was strong In the higher price ranges, DQ slowed while HS continued to rise When a trend line is added, it shows a logarithmic curve very close to the shape of the threshold line in Figure The HS± DQ relationship can also be seen clearly in Figures 16 and 17, where the values of HS and DQ as proportions of house price (HS/(PriceÐ a ) and DQ/(PriceÐ a ) are plotted Theoretically, with some quali® cations, the logarithmic curve in Figures 14 and 15 could be accepted as the threshold line It should be noted, however, that this is the threshold line for a sample that includes only owner-occup ied houses, which are normally better than the average stock in terms of physical quality In theory, each group can have a threshold line of its own The main point here, however, is that the theory can be tested by ordinary means A testing procedure should include the following steps: (1) De® ne status pole or poles (2) Delineate status zone(s) and calibrate status distance(s) (3) Run hedonic regression on house price with variables organised into two groups related to housing status (HS) and dwelling quality (DQ) for each zone (4) De® ne threshold line based on predicted HS and DQ 5.5 The Survey and the Social Structure of Residents If the regressions on house price can show the predicted relationship of housing status and dwelling quality, the socioeconomic data from the survey can also give a very good illustration of the ring pattern of the distribution of population groups as described above In Figure 18, a comparison between different groups of residents is shown These groups are de® ned by the main types of home improvem ent they undertook, namely: status-inducing, status-imitating, structural improvements, and building services improvements It is already clear by their names that the `status content’ of these types of home improvement is decreasing in this sequence Spatially, the sequence also re¯ ects the increasing distance to the downtown area Figure 18 thus shows a very explicit pattern of ring structure of groups of improvers in central Hanoi To a large extent, these groups can be treated as proxies for the city’ s social groups in general 5.6 The Need for Further Testing It is understood that for an appropriate testing of the theory, the data and statistical results of Hanoi alone are far from conclusive Moreover, the housing survey used was designed for the relatively narrow theme of a Figure 14 Dwelling quality (DQ) and housing status (HS) for the whole sample A NEW THE ORY OF RESIDENTIAL LOCAT ION 29 Figure 15 Dwelling quality (DQ) and housing status (HS) for the newly built houses 30 HOANG HUU PHE AND PATRICK WAKE LY Figure 16 Housing status (HS) and dwelling quality (DQ) as a proportion of house value for the whole sample A NEW THE ORY OF RESIDENTIAL LOCAT ION 31 Figure 17 Housing status (HS) and dwelling quality (DQ) as a proportion of house value 32 HOANG HUU PHE AND PATRICK WAKE LY A NEW THE ORY OF RESIDENTIAL LOCAT ION 33 Figure 18 Classi® cation of home-improvers by type of improvement action: characteristics of the improver groups are expressed in percentages of the value of the leading groupÐ i.e improvers corresponding to the status-inducing category of improvement Note that for all the characteristics except for length of stay (LTHSTAY ), the calculated average values of improvement expenditure (EXPENDIT), food expenses per capita as an income proxy (FOODPC) and years of schooling (EDYEAR )follow a visible pattern: the status-inducing group is the highest, followed by status-imitating, structural and housing services In terms of age (AGE), the sequence is reversed: the status-inducing improvers are the youngest different study Ideally, a series of appropriate surveys covering several cities, perhaps at different levels of income, and with different political and economic histories, needs to be undertaken to test the theory For the time being, however, the issues raised can be well served by further discussion and perhaps more testing among those who have access to more suitable data and who are interested in alternative ways of looking at residential location in cities Concluding Remarks A rede® nition and a reinterpretation of largely familiar groups of factors that in¯ uence residential locational decisions have led to the postulation of a new theoretical model which can describe and satisfactorily explain more real-life situations than the preceding models One of the most important points in which this model differs from its predecessors is the conscious avoidance of the rigid economic determinism that often assumes away many elements, some of which are indispensable in explaining the spatial behaviour of the urban households The ¯ exible de® nition of housing status, which changes quite frequently, is borne out of realistic reasoning that allows historical processes to play a role in formulating residents’ perception of housing Physical distance enters the model only as a calibrated dimension that is subject to the size of the sphere of in¯ uence exerted by social forces, rather than as the ® xed, de® ning factor, as in the conventional models Instead of being the results of the economistic access/space tradeoff, the patterns of residential location in cities have been shown to be the outcome of other kinds of trade-off, which are largely social in natureÐ the ones that are based on status and the social acceptability of dwelling quality 34 HOANG HUU PHE AND PATRICK WAKE LY At a more generalised level, the model represents a trade-off, or rather a balance, between the ideal and the possible, such as occurs in any sort of spatial activity, be it consumption or investment, production or living, working or leisure The tension created by a lack of this balance, and the constant movements within and between `desirable’ and `undesirable’ zones, are what drive the dynamics of the urban housing scene that contributes so substantially to the shaping of cities As such, this model can be adapted to analyse the spatial behaviour not only of the residents, but also of the industrial developers, investors in infrastructure and urban services In the ® nal analysis, it is never possible to model the residential structure of a city with all its complexities: there are too many variables and irregularities The case of Hanoi, however, does show some of the possibilities and limitations of the model In the analysis of urban residential development, perhaps the quality±status model allows us to go beyond the famous empirical dictum `First, location; second, location; and, third, location’ , by giving a de® nitionÐ however tentative or impreciseÐ of what is behind the notion of `location’ that makes it so powerful References A LONSO , W (1964) Location and Land Use Cambridge, MA: Harvard University Press B ACHELARD, G (1958) La poetique de l’ espace Paris: Presses Universitaires de France B ALCHIN, P., K IEVE, J and B ULL , G (1995) Urban Land Economics London: Macmillan B ASSETT , K and S HORT , J (1980) Housing and Residential Structure: Alternative Approaches London: Routledge & Kegan Paul C LARK, D (1996) Urban World/Global City London: Routledge C OOPER, C (1972) The house as a symbol, Design & Environment from Design & Environment, 3(3) D EAR, M and F LUSTY, S (1998); Post modern urbanism, Annals of the Association of American Geographers, 88, pp 50±72 E VANS, A (1973) The Economics of Residential Location London: MacMillan F IREY , W (1947) Land Use in Central Boston Boston, MA: Harvard University Press F RENCH , R A and H AMILTON, F E I (Eds) (1979) The Socialist City: Spatial Structure and Urban Policy Chichester: John Wiley F RIEDMAN , J (1995) Where we stand: a decade of world city research, in: P Knox and P Taylor (Eds) World Cities in a World-system, pp 21± 47 Cambridge: Cambridge University Press F UJITA , M (1989) Urban Economic Theory, Land Use and City Size Cambridge: Cambridge University Press G RILICHES, Z (Ed.) (1971) Price Indexes and Quality Change Cambridge, MA: Harvard University Press H ALL , P (1984) The World Cities, 3rd edn London: Weidenfeld & Nicholson H ALLETT , G (1978) Urban Land Economics London: Macmillan H AMNETT , C and W ILLIAMS , P R (1980) Social change in London: a study of gentri® cation, Urban Affairs Quarterly, 15, pp 469±487 H ARVEY , D (1973) Social Justice and the City London: Edwards Arnold H ARVEY , D (1991) The Condition of Postmodernity Oxford: Blackwell H OANG H UU P HE (1997) Housing and urban form in Vietnam: a study of home improvement in central Hanoi Unpublished PhD thesis, University College London H OANG H UU P HE and N ISHIMURA , Y (1990) The historical environment and housing conditions in the `36 Old Streets’ quarter of Hanoi Research Paper No 23, Human Settlements Development Division, Asian Institute of Technology, Bangkok, Thailand H OANG H UU P HE and O RN , H (Eds) (1995) The phuongs of Hanoi: Vietnam urban transport assistance project (VUTAP) Swedish International Development Authority, Gothenburg and Hanoi H UDSON , R and R HIND , D (1980) Land Use London: Methuen H URD, R M (1903) Principles of City Land Value New York: Record and Guide ISARD, W (1956) Location and Space Economy Cambridge, MA: MIT Press K AIN, J and Q UIGLEY J (1975) Housing Market and Race Discrimination: A Microeconomic Analysis New York: National Bureau of Economic Research K IVELL , P (1993) Land and City: Patterns and Processes of Urban Change London: Routledge K NOX , P (1994) Urbanization: An Introduction to Urban Geography New Jersey: Prentice Hall K NOX , P (1995) Urban Social Geography Harlow: Longman Scienti® c & Technical L AWRENCE , R (1987) Housing, Dwellings and Homes: Design Theory, Research and Practice Chichester: John Wiley L AWRENCE , L (1998) Smoke and Mirror, Far A NEW THE ORY OF RESIDENTIAL LOCAT ION Eastern Economic Review, 22 October, pp 60± 61 M ACLENNAN, D (1982) Housing Economics Harlow: Longman Group Ltd M CG EE , T G (1991) The emergence of desakota regions in Asia: expanding a hypothesis, in: N G INSBURG , B K OPPEL , and T G M CG EE (Eds) The Extended Metropolis: Settlement Transition in Asia, pp 3±25 Honolulu: University of Hawaii Press M EGBOLUGBE , I (1986) Econometric analysis of housing trait prices in a Third World city, Journal of Regional Science, 26, pp 533±547 M UTH , R (1969) Cities and Housing Chicago, IL: University of Chicago Press N ATIONAL INSTITUTE FOR U RBAN AND R URAL P LANNING (1994) Data Book Hanoi City, Project VIE/88/P02 Ministry of Construction, Hanoi P AHL, R E (1975) Whose City? And Further Essays on Urban Society London: Penguin R APOPORT A (1977) The Human Aspect of Urban Form: Toward a Man Environment Approach to Urban Form and Design Oxford: Pergamon Press R EX , J (1968) The sociology of a zone in transition, in: R E P AHL (Ed.) Readings in Urban Sociology, pp 211±231 Oxford: Pergamon Press R IMMER , P J (1991) Exporting cities to the western Paci® c Rim: the art of the Japanese package, in: J B ROTCHIE , M BATTY , P H ALL and P N EWTON (Eds) Cities of the 21st Century: New Technologies and Spatial Systems, pp 243±261 Melbourne: Longman Cheshire R OLL , E (1990) History of Economic Thoughts London: Faber and Faber R OMANOS , M C (1976) Residential Spatial Structure Lexington, MA: D C Heath R OSEN , S (1974) Hedonic price and implicit markets: product differentiation in pure compe- 35 tition, Journal of Political Economy, 82, pp 34±55 S ASSEN , S (1995) On concentration and centrality in the global city, in: P K NOX, and P Taylor (Eds) World Cities in a World-system, pp 63± 75 Cambridge: Cambridge University Press S IT , V F S and Y ANG , C (1997) Foreign-investment-induced exo-urbanisation in the Pearl River Delta, China, Urban Studies, 34, pp 647± 677 S MITH , N (1987) Gentri® cation and the rent gap, Annals of the Association of American Geographers, 77, pp 462±465 S OJA , E (1996) Los Angeles 1965±1992: the six geographies of urban restructuring, in: A J S COTT and E S OJA (Eds) The City: Los Angeles and Urban Theory at the End of the Twentieth Century, pp 426±462 Los Angeles: University of California Press S ZELENY, I (1983) Urban Inequality under State Socialism Oxford: Oxford University Press T HRALL , G I (1987) Land Use and Urban Form New York and London: Methuen T HUÈ NEN , J H VON (1826/1968) Der Isolierte Staat in Beziehung auf Landw irtschaft und National Ekonomie Hamburg P.G H ALL (Ed.) (1968) Von ThuÈ nen’ s Isolated State, trans by C M W ARTENBURG Oxford: Pergamon Press T UAN, Y F (1961) Topophilia, or sudden encounter with landscape, Landscape, 11, pp 28±32 UNDP (UNITED N ATIONS D EVELOPMENT P ROGRAMME) (1990) Report on the Economy of Vietnam Hanoi: UNDP/SPC V UONG, T Q (1977) Hanoi from pre-history to the XIX century Vietnamese Studies Vol 48 Hanoi: Foreign Literature Publishing House W HEATLEY, P (1983) Nagara and Commandery: Origins of the South East Asia Urban Traditions Research Paper Nos 207±208, Department of Geography, University of Chicago ... quality? ?status theory is able to describe satisfactorily the intricate relationships between status, quality and value It has been shown that if the residential location decision is a trade- off, then... has its own threshold line, then the trade- offs will not be easy to escape even for the rich, albeit at quite different quality levels An important point is that the proposed quality? ?status theory... structure as the outcome of the trade- off between a desirable status and an acceptable level of dwelling quality: that is, the trade- off between the ideal and the acceptable Applied to historical