Navigation – Plan du site

Poverty and Residential Environment in Nanjing

Is there a Neighbourhood Effect ?
La géographie des pauvretés à Nankin
Louise Babar et Christian Kesteloot
p. 533-549


En Chine, la transition entre deux systèmes économiques s’accompagne d’un accroissement des inégalités et de la création de nouvelles formes de pauvreté et de marginalisation sociale. Les facteurs qui en sont la cause peuvent être vus comme le résultat d’une superposition entre les réformes récentes et des structures urbaines sociales et spatiales héritées. Nous avons étudié cette question dans le contexte de Nankin, une ville de l’Est chinois affectée par ces rapides changements économiques et sociaux. Ces changements surviennent dans un contexte spatial qui porte encore les marques d’une organisation sociale et économique passée. L’article analyse le rôle de l’environnement dans les différents modes d’exclusion sociale et la façon dont les gens gèrent leurs situations. Il semble que chaque type de pauvreté appelle des réponses différentes, en fonction de la localisation dans l’espace urbain, mettant ainsi en valeur l’importance des structures spatiales héritées ainsi que le potentiel de chaque lieu.

Haut de page

Texte intégral


  • 1 Acknowledgements: We would like to thank Yuting Liu for making his data set on poor households in N (...)
  • 2 A national survey in 2002, carried out by the Ministry of Civil Affairs counted 19.38 million urban (...)

1In spite of recent rapid economic growth, the percentage of population that is impoverished stays very high in China. While rural poverty remains a large and historical problem, social inequalities and poverty are also rising in urban areas2 [Li, 2004 ; Wu, 2004].

2This paper looks at the influence of residential environment on urban poverty. After reviewing previous research on the nature and creation of new forms of marginalization in contemporary urban China, the study focuses on the case of Nanjing. The purpose is to understand how poverty varies following the location within the city. The generation of residential patterns according to socio-economic situations is examined as well as how geographic location influences coping behaviour.

3The results of data analysis indicate that spatial and social insertion of marginalized people mainly varies according to two important factors : the first one relates to the city’s historically acquired structure ; the other one relates to people background, constraints and opportunities.

1. About urban poverty in china

  • 3 From 1953 to 1978, China is ruled by a socialist planned regime. In this system, the economy is nat (...)

4The reforms that started 30 years ago removed many controls of the planning era3. Justification of such changes was mainly economic, although it carried deep political implications. The intent was to achieve greater economic efficiency by gradually replacing the centrally planned system with a market regulation allowing independence in production and consumption decisions [Li, 2004, p. 7 ; Gu & Kesteloot, 2002, p. 8].

5The reforms, as well as the on-going economic transition that followed, also changed the problem of social exclusion as new forms of poverty and exclusion appeared.

6When considering the concept of social exclusion Li reminds that it needs to be redefined before it is used in the context of contemporary China. First, he notes that social exclusion is too often understood as an economic destitution, failing to take into account politically and socially-based exclusion. In the United Nations Development Program, “social exclusion is defined as lack of recognition of basic rights, or where there is recognition, lack of access to political and legal systems necessary to realise the rights” [Burchardt et al., 2002 by Li, 2004, p. 3]. In this definition social exclusion does not only include people who are poor in terms of income, but all “those who are being denied political, social and economic rights, putting them in a very vulnerable position and under constant threats to their life and property” [Li, 2004, p. 3].

7Also, the concept of social exclusion, as shaped in developed countries, implies that the socially excluded are not the majority in terms of population. But the situation is different in developing countries. “Neglecting the difference between developed and developing countries might lead to a failure in addressing the main issues faced by large groups or even the majority of the population and only pay attention to the minority vulnerable only to a less extent.” [Li, 2004, p. 2]

8To elucidate which people fall into urban poverty the definition of poverty must be clarified. In reference to the evolution of the social context, a distinction between “traditional” poverty and “new” poverty is often used in the literature [Chen et al., 2006, p. 7 ; Wu, 2004].

9The “traditional” poor is the group known in the centrally planned system as the “3 no’s” households : no income source, no ability to work and no family to support them. The disabled, the long-term sick and the elderly who could not get support from their family members were considered, at the time, the only persons to fall outside the employment and consumption protection provided by the planned economy. They were taken in charge by the government who would provide them basic support [Chen et al., 2006, p. 7 ; Wu, 2004, p. 407].

10The “new” poor form a more heterogeneous category of people. Their experience of poverty or exclusion is the product of discrimination related to failures in the socio-economic regulation during the transition period and to new inequalities generated by the operation of the market economy.

11A first exclusion process takes roots directly in the change of production’s organisation as people who used to be integrated in the planned urban economy become excluded with the decline of public enterprises.

  • 4 Pillar-institution in the planned economy, the work-unit or danwei controls the distribution of in- (...)

12With the shift to market-led accumulation, the need for enterprises to raise their competitiveness and productivity rates leads an increasing number of people to be dismissed from their job [Kernen & Rocca, 1998, p. 6]. This necessity for enterprises to minimize the cost of labour adds to the decreasing efficiency of social insurance, which was organised at the enterprise level. Dismissed workers are either ‘laid-off’, which means that they are still attached to their work-unit4 and entitled to benefits, or ‘unemployed’, which means they are detached from their work-unit. Some workers put to early retirement also fall in this category [Li, 2004 ; Chen et al., 2006 ; Kernen & Rocca, 1998, p. 14].“Since the reform it is not rare that an employee is dismissed from his/her job getting little compensation. They are “laid off” (xiagang) workers. It means that administratively they are still attached to their employers and can enjoy the in kind benefits as they did before. However many enterprises went bankrupt since the reform. Quite often, laid-off workers lost their “work units” (danwei, workplace) and subsequently, the benefits. Even for the surviving enterprises, the market pressure prevents them from offering benefits to the laid-off workers as they did before. According to the newly established welfare schemes, individuals have to contribute to individual accounts in order to receive pension and healthcare in the future. Laid-off workers and the unemployed are not able to contribute to their own accounts. In the future, they will not be able to receive benefits.” [Sun, 2002, by Li, 2004, p. 14].

  • 5 Thus the “floating population” also includes students, soldiers, hospital patients, etc. as well as (...)
  • 6 Hukou is an administrative tool set in the central planning era and corresponds to the need to assi (...)

13Exclusion processes also result from a discrimination rooted in regional differentiation. In China, the growing economic gap between rural and urban areas generates an important migration movement. People who were integrated in the rural economy migrate to the city in search for job and better life opportunities. This group, known as the rural migrants, form the bulk of the “floating population” (liudong renkou), i.e. the population not present in its place of registration5. They face discrimination based on their administrative status. As they hold a rural hukou6, migrants are deprived from the rights granted to urban-hukou citizens and cannot enjoy the social benefits distributed via the work-unit. “While enormous numbers of rural labourers have immigrated to urban areas with the relaxation of controls on population movement, most live in poor conditions under the restriction of the household registration system and related policies of inequality” [Liu & Wu, 2006, p. 611]. As a reaction to this discrimination based on the administrative status, migrant workers develop specific social spaces, informal market niches and set up parallel networks that can become very efficient [Gu & Kesteloot, 2002 ; Zhang, 2001].

14Their location in the city is very much linked to this feature of informality. “Usually, rural workers cannot afford to buy or rent houses in the private market. Their first alternative is to rent houses in informal settlement or farmers’ houses in suburbs or even in inner city areas. The most typical example is rural workers’ “villages”. Usually residents from one village come from the same province or hometown. The second alternative is to live in the dormitories or temporary accommodations offered by employers. These dormitories are very close to work and rural workers do not have much contact with urban residents. The third alternative is to live with urban residents, in urban households as chore workers.” [Li, 2004, p. 24].

2. Analysis of a sample of nanjing’s poor population

2.1. Nanjing’s socio-spatial structure

  • 7 Nanjing was the capital of China for the first time between the 3rd and the 6th century BC. It rega (...)

15In order to analyse the population distribution of urban poor in Nanjing, it is important to remind the socio-spatial structure and historical construction of the city. As an ancient capital city7, the classic pattern of the Chinese city was applied in the urbanisation of Nanjing, having a deep and enduring impact on the spatial distribution of social groups, economic activities and administrative functions.

16The city structure has further been marked by the experience of planned economy. While the means to reshape a totally new concept of urban organisation were lacking, new urban complexes, characterised by homogeneity and stability, were developed outside the older inner city. The constructs of planned economy and urbanism are responsible for strong inertial dynamics in the city. They remain active through the institutions and even more through the created built environment – such as the work-units and their housing complexes – which linger until today. This explains, for example, why a concentration of poor household is visible in former industrial workers estates, where the industrial activity has disappeared [Wu & He, 2004 ; Wu, 2007].

  • 8 Only the use of the land has been privatised and submitted to market mechanisms, while the property (...)

17Finally, the shift to the market economy triggers spatial reconfiguration as privatisation of the construction industry and creation of a pseudo land market8 reactivates urbanisation. The city is extended to new boundaries, central areas are reinvested and old parts of the city benefit from renovation. In fact, the last two operations are cause of a residential differentiation as poorer people are driven away from newly or refreshed part of the city where rents become prohibitive [Gu et Kesteloot, 2002 ; Wu & He, 2004, p. 79].

18The diagram on Error: Reference source not found summarizes the socio-spatial structure of Nanjing and the important facts of the city’s construction relevant for understanding the problem of poverty location.

Figure 1. Socio-spatial division of Nanjing

Figure 1. Socio-spatial division of Nanjing

19The pattern of the ancient city and the place of important functions are still visible today. The star symbolizing city-centre marks the cross between a north-south (N-S) and a west-east (W-E) axis. The eastern part of the city was the imperial quarter. Though the Ming palace was destroyed in 1644, this part of the city still concentrates business, administrative and educational functions. This area also counts numerous high-storey buildings ; some of them consist in work-units’ housing complexes.

  • 9 The Minimum Living Standard Program (MLSP) is a government-funded program initiated in 1993 in Shan (...)

20The historical place of high social classes is in the northwest. Today, close to the northern city wall, a second activity and business centre is developing. The socialist industrialisation extended the city further to the north, outside the traditional city walls, in the direction of the river (Chang Jiang). When these industrial activities were stopped or restructured, workers continued to live in the housing that production units provided close to the site of production. The area shows today a concentration of unemployed, which is reflected by an over-representation (up to 4 %) of beneficiaries of the Minimum Living Standard Program (MLSP)9, the government-funded helping program. Not far away, the developing business pole attracts migrants and poor urban workers in search of job opportunities, including informal ones.

21The low city lies south of the W-E axis and houses the lower classes – craftsmen and merchants and their employees – in the traditional city structure. It is a dense urban fabric, composed of traditional low-storey buildings. Within this low city, the southern city wall delimits the historical slum areas. Today they also house a high percentage of MLSP.

  • 10 Once they are swallowed by the urban fabric, these villages become “Villages in city” or Vic.

22Socialist industrialisation also developed in south and southwest directions outside the city wall. Migrant workers are located mainly outside this southern city wall, in an area of more recent expansion. This presence of rural workers relates to housing constraints migrants face when trying to live and work in Nanjing. Many of them only find a house in former rural villages10 where the former peasants, once they are included in the urban system, rent (part of) their house to rural migrants or even build additional housing on purpose [Song, 2007].

2.2. Data and methodology

  • 11 There have been other empirical researches focusing on the problem of poverty in Nanjing by Wu and (...)

23Our empirical work is based on the results of a survey led by Yuting Liu in 2002 in Nanjing. Covering nineteen of the sixty subdistricts of the city, a sample of over 400 individuals identified as poor in term of income level were interviewed about different aspects of their life : educational and professional status, economical situation, housing conditions and feelings about general living conditions. The survey results have been explored and published by Liu and Wu (2006)11. But the spatial dimension was never investigated because of the size and spatial distribution of the sample. Indeed, the sample does not reflect, for each subdistrict represented, the real proportions of each type of poor in relation to the whole population. Because of this, the survey data cannot be used to show absolute concentrations. It is not possible, for example, to draw the distribution of each category throughout the city using the proportions found in the sample. However, it is possible to extract significant information about individuals belonging to different categories of poor in relation to their environment and the geography of the city. In other words, we can compare a group of poor present in two different areas and study the differences and resemblances between these two situations.

24The choice for the method of data analysis is crucial as it influences the form of the results and their interpretation. Since we are dealing with qualitative data and searching the relations between poverty and residential environment, correspondence analysis has been privileged here.

25This method of data analysis quantifies the interrelations between columns and rows from cross tabulations and expresses them in a number of independent dimensions of decreasing importance (just like Principal Component Analysis does it with correlations between variables). The results can be visualised in scatter plots on these dimensions. The rows and columns are represented as points in a common space. In this particular case, we cross-tabulate five categories of poor with some of the key questions of the survey. We thus explore over- and under-representation of responses to these questions over the five categories of poor. The centre of a scatter plot represents the average situation. In other words, if one response is evenly distributed among the five interviewees’ categories it is projected in the centre. Similarly, if one of the categories of poor shows a relative distribution of responses that equals that of all the interviewees, it is also projected in the centre of the plane. The more a response or a category of poor diverges from this situation of equipartition, the further it will be from the centre of the plane. Categories of poor with similar responses will be appearing close together on the plane. Finally, responses and categories situated in the same direction from the centre are linked by a common over- or under-representation.

  • 12 For further explanation concerning correspondence analysis, see : Orban-Férauge, F. (1985) and Gree (...)

26In order to avoid any misinterpretation, one has to verify that the concerned rows and columns are well represented on the plane. Indeed, if a table has more than three rows or columns it is impossible to represent all the row and column deviations from equipartition in one single bi-dimensional plane. The successive dimensions embody the largest possible share of the total deviations (called inertia), but when only two of them are selected to span a plane on which rows and columns are projected, some of them can be poorly represented. In that case the position of the columns or rows concerned on that plane can be irrelevant12.

27The sample has been divided in five categories based on their relation to employment : “employed”, “retired”, “unemployed”, “laid-off” and “rural worker”. The first category gathers workers who are integrated in a work-unit. Poor employed persons are usually working in State-owned enterprise unable to pay sufficient wages (because they are competitively weak). “Retired” and “Laid-off” concern people who don’t work, who are still attached to the work-unit, but whose pension or unemployment benefits are not sufficient or unpaid for the same reason. “Unemployed” involves those who have been detached from the work-unit system. The category “rural worker” refers to rural-hukou workers, or migrants. The Error: Reference source not found shows the percentage of each group in the sample.

Figure 2. Percentage of employment groups in the sample


Total in sample

Laid-off workers




Rural workers

Total in sample







Part (%)







28Error: Reference source not found shows the presence of the different groups in the twelve subdistricts that have a representative number of interviewees.

Figure 3. Spatial distribution of the different groups of poor in the sample

Figure 3. Spatial distribution of the different groups of poor in the sample

Author : Louise Babar.
Source : Yuting Liu survey, 2002.

29Location appears to be similar for the employed and retired categories ; another similarity exists between the unemployed and laid-off categories. The category of the rural workers shows a different location. Some subdistricts have a mixed composition of laid-off workers, unemployed and rural workers. Considering this distribution it is possible to cluster subdistricts together and associate one or several dominant categories of poor.

Employed (EM) and Retired (RE)

Ruijinlu (Ru), Daguanlu (Da), Meyiuan xincun (Mx)

Unemployed (UN) and Laid-off (LO)

Nanhu (Na) and Mufushan (Mu)

Rural worker (RW)

Jianninglu (Ji), Xinglong (Xg), Xiaoshi (Xi)

Mixed districts (Unemployed ; Laid-off ; Rural worker)

Chaotiangong (Ch), Zhonghuamen (Zo), Ningnan (Ni)

30These groups of subdistricts offer different residential environments related to their position in the socio-spatial structure of the city. Since poor people of different categories can be found in them, one can unveil interrelations between poverty and residential environment by exploring differences within each category of poor present in the different groups of subdistricts. This exploration was done using correspondence analysis. We present the results of this exploration of Liu’s survey data under the form of maps, because they enhance the clarity of the interpretation and relate the results with the overall socio-spatial structure of the city.

2.3. Housing conditions and residential patterns

31First, housing appears to be a key problem. Housing conditions vary as each group is subject to different constraints. A few questions in the survey address these issues specifically : the source and type of housing, the time lived at the same address and the income. These variables are represented on maps (Error: Reference source not found to Error: Reference source not found).

  • 13 In 2002, the average income in Nanjing was 900 CYN (= 100 EUR).
  • 14 See note

32The group of migrants has an overrepresentation of young people, between 16 and 35 years old. They are also more often single. Their income levels are not as low as for the other groups, and some cases represent the highest income in the sample, around the city average13. As shown on the map on Error: Reference source not found, higher incomes among the poor migrants are found in peripheral subdistricts, where migrants of the sample concentrate. Generally speaking, migrants are out of the work-unit system and considered marginalized. They claim to rely solely on the private market, running independent business or performing informal activities. Being marginalized from the work-unit institution is reflected in the housing conditions : almost all migrants in the sample rent their dwelling. Only a couple of migrants declared living in housing provided by a work-unit. This draws attention on the particular situation of a small number of households composed of urban and rural hukou holders. This combination is a way for rural hukou holders to access urban facilities. Error: Reference source not foundError: Reference source not found also shows migrant’s housing mostly consist in the traditional low-storey buildings of high-density areas. Self-built housing is more often found in the mixed subdistricts. This form of construction happens in dense areas where urban dwellers find it lucrative to enlarge their property and rent it to migrant workers. As for living in apartments, it represents a less frequent situation in peripheral subdistricts where “Villages in Cities”14 are found.

Figure 4. Residential pattern of rural migrants group

Figure 4. Residential pattern of rural migrants group

33The category of employed and retired includes middle-aged to elderly people. The three subdistricts of concentration, near the city centre, have an overrepresentation of elderly people, over 60 years old, and without formal secondary and post-secondary education, confirming the presence of retired people. The same feature is also found in the subdistricts concentrating migrants.

34Generally speaking, the category of employed and retired gets low income (about half the average income in the city), but not as low as the laid-off and unemployed group does. People of this group are long-term stable inhabitants with low mobility. Their housing is either provided by the work-unit – apartments in high-storey buildings in the periphery or traditional houses in the centre – or inherited. In this last case, it normally consists in traditional low-storey houses, sometimes extended with self-constructed parts. Interestingly, the higher incomes among the poor are found in the subdistricts where housing is dominantly provided by the work unit. This reflects a closer link of employed and retirees with their danwei, offering both better housing conditions, and a better income. (Error: Reference source not found).

Figure 5. Residential pattern of employed and retired group

Figure 5. Residential pattern of employed and retired group

35However, the income level of the unemployed and laid-off category reveals a reverse geographical pattern compared to the employed and retired one. Income is, in this case, lower, or missing, in central and peripheral subdistricts where housing is still provided by the working units. One can suppose that, although their former work units can still house laid-off workers, they get lower income than the retired and employed poor. As for the unemployed, they depend entirely on their own initiatives or on the MLSP. Income, for this category, is slightly higher in the peripheral migrant subdistricts as well as in the central mixed subdistricts. These two types of residential environments also display a slight overrepresentation of women and the time lived at the same address is a little lower, less than 5 years. This could be related to better job opportunities on the private market, in small services and independent business in these areas. The same reason probably explains the better situation of migrants and unemployed urban poor in the same parts of the city.

Figure 6. Residential pattern of unemployed and laid-off group

Figure 6. Residential pattern of unemployed and laid-off group

36These findings confirm that, at least until the early years 2000, the industrial subdistricts built during the socialist development era remain strongly linked to the fate of the work-units. The inhabitants who are still employed in a work-unit or who remain linked to it by a retirement pension continue to share the same place of living as those who have been laid-off or put to unemployment ; but the income level varies among them, well above or well under half of city average. In fact, some interviewees from these categories of poor settle in the same subdistricts as the rural workers, where the expansion of the city, dense population or intense economical activity offer opportunities to earn a living on an alternative, private market.

2.4. Coping behaviour, social networks and feelings about residential environment

37After studying housing conditions, the effects of the different residential contexts on the way poor people cope with their situation are examined. The previous analysis already suggested that the central subdistricts and the migrant subdistricts (in which presumably the Villages in City are the most vital environments) offer much more heterogeneous and dynamic environments than the former industrial areas, presumably as a result of the operation of the private sector, with more opportunities for the poor to develop survival strategies.

38The available data enable further investigation of these differences ; involving questions dealing with feelings about the residential environment, desire to stay or to leave the present residential location and social acquaintances. In addition, the time people have been staying at the same address is taken as an indicator of residential stability. For the migrants, the time elapsed since leaving their place of origin is also taken into account.

39The issue is to uncover possible differences in these responses within a same category of poor, according to the residential environment in which they live. In order to generate a synthetic overview, considering the categories of poor living in mixed environments and all types of residential environments, we turn to the graphic results of the correspondence analysis.

Figure 7. Social network, feelings about residential environment and residential stability

Figure 7. Social network, feelings about residential environment and residential stability

40The graph shows the projection on the first two dimensions of the responses to these questions, cross-tabulated with the unemployed, laid-off and migrant poor, according to the subdistrict they reside in. The latter are shown on the graph using the first two letters of the subdsitrict name underlined (see Error: Reference source not found and accompanying table), followed by rw (for rural workers) or un/lo (for unemployed and laid-off) for the mixed subdistricts : Ji, Xi, Ch(rw), Ch(un/lo), etc.

  • 15 The category « very good » was also offered in the questionnaire, only one out of the 432 responden (...)

41The responses to the survey questions are shown by the other symbols on the graph. The perceived feelings about the residential environment include five levels of appreciation (RENB, RENC, REND and RENE) and goes from good (RENB) to very bad (RENE)15. The time of residence is given by T0 to T5. T0 means less than one year, T1 = 1 to 2 years, T2 = 3 to 5 years, T3 = 6 to 10 years, T4 = 10 to 20 years and T5 = more than 20 years. The willingness to stay in the neighbourhood for a long time or not is expressed by “Stay” or “Leave”. The dominant social contacts are defined as : neighbours, colleagues, friends, family. For the migrants, the similarity in origins has also been taken into account, since the departure region is an important component of the migrants’ social networks. The time since departure (Td), for the migrants, follows the same scale as the time of residence for all respondents, but stops at 15 years.

42A crossed symbol indicates that the related response, category of poor or subdistrict, is poorly represented on the first two dimensions. These first two dimensions respectively represent 46 and 18 % of the total inertia.

43The first dimension opposes the rural migrant subdistricts to the unemployed districts. This is clearly replicated by the time of residence, and also by dissatisfaction with residential environment. Evidently, the responses concerning time since departure of the migrants, reinforces this dimension. The second dimension however, shows some differences between the migrant subdistricts and the situation of migrants in mixed districts. The former ones show a relation with short stays in the present place and both long and short times since departure and with satisfaction with residential environment, while the latter relate with somewhat longer stays and average time since departure, but less satisfaction with environment. These differences are reinforced by relations among migrants of the same origin in the former subdistricts and with immigrants from different origins in the latter (especially Chaotiangong Ch(rw)).

44Unemployed and laid-off subdistricts appear on the left side of the first dimension and are less differentiated by the second dimension. They show a relation with longer periods of stay, but less satisfaction with residential environment. Their social relations include friends, family and even more neighbours (a logic consequence of a longer period of stay). Relations with colleagues are not well represented on these dimensions, as we do not consider employed persons who meet these colleagues more frequently.

45Surprisingly, intentions to stay for a long time in the neighbourhood or to leave are not differentiated, which is expressed by the position of these points close to the centre of the graph.

46The graph also shows that social conditions and satisfaction are very different in mixed subdistricts according to the category of poor considered. Moreover, the most satisfied respondents are the rural migrants in their concentration subdistricts, especially Xiaoshi (see also Wu 2007). The mixture of short and long times since departure of the migrants, as well as a rather short presence in the subdistrict, suggest that migrants return to this area after having tried less satisfying places in the city.

47This last analysis suggests additional contrasts and dynamics to those related to the residential environment as such. The whole is summarized on Error: Reference source not found.

Figure 8. Influence of environment on feelings, behaviour and social network

Figure 8. Influence of environment on feelings, behaviour and social network

48First, the level of satisfaction toward residential environment reveals an opposition between centre and periphery. Two factors seem to have a strong influence on this feeling : residential stability and the level of income.

49Central subdistricts have an overrepresentation of settled inhabitants, urban as well as rural workers, whereas peripheral subdistricts tend host more recent population.

50From periphery to centre, the level of income decreases for unemployed, laid-off and migrants groups ; the satisfaction towards residential environment and conditions shows a similar variation. Generally speaking, migrant workers are less unsatisfied about their situation than urban poor.

51Interviews regarding social relationships indicate that socialization varies according to the category of poor. However, they also reveal that the type of social ties differs as a result of the residential environment and its geographic location.

52Relations of vicinity (with neighbours) appear as primordial in the central subdistricts. This is not only explained by high population density but also by the traditional organisation found in long-term local-based community. Apparently that kind of community feeling is still alive in some old parts of the city.

53Migrants claim to socialize mostly with other migrants. In the peripheral “migrants-districts”, relationships occur predominantly among persons from the same region of origin, while central subdistricts offers possibilities of more diversified relations.

54In contrast, among urban unemployed and laid-off, family represents a major social sphere in the south-western concentrating subdistrict while relationships between colleagues are more important in the northern subdistricts. This can be related to the fact that people from a same work-unit also often live in the same building creating a specific form of socialization.


55The location of the poor in different residential environments of Nanjing impacts on their poverty in various ways. This relation can intensify exclusion or act as source of support by generating alternatives to cope with marginalization. The results indicate that several factors affect the way marginalized groups face poverty in different residential environment : the type of housing they can access, the work and income earning opportunities and the social network.

56A constraint hanging over rural-hukou holders is their exclusion from housing provided by public agencies and work-units. Reduced housing opportunities are therefore a determining factor in the residential location of migrants in the city. But dependence on cheap rented housing is not just a characteristic of rural workers. An increasing number of people, excluded from central locations by urban restructuring or detached from the public work-unit organisation, look in the periphery for alternative housing options.

57Nanjing is expanding beyond its walls toward the north and southwest creating new poles of activity. Alternatives for housing and income seem to be easier to find in these areas of the city’s expansion as they seem to attract certain categories of poor who find there opportunities to earn a living ; for example people excluded from public institutions find an alternative in the private market. This is supported by the fact that (except for the group of employed and retired people who are still strongly integrated in the work-unit organisation) migrant workers and unemployed urban poor have lower income in the central subdistricts than in subdistricts located on these areas.

58Finally, in a context of transition between two modes of production, the complexity of situations generated by the rise of poverty unquestionably participates in the socio-spatial reconfiguration of the city. As marginalized people invest the available space of the city and make use of its potential, it produces unique and opportunistic forms of social and economical organisation.

Haut de page


BURCHARDT T. et al. (2002), « Introduction », in J. Hills et al. (eds.), Understanding Social Exclusion, Oxford University Press, pp. 1-12.

CHAN K. W. (1994), Cities with invisible walls. Reinterpreting urbanization in post-1949 China, Oxford University Press.

CHEN G., GU C., WU F. (2006), Urban poverty in the transitional economy : a case of Nanjing, China, Habitat International, vol. 30, pp. 1-26.

GREENACRE, BLASIUS (eds.) (1994), Correspondence Analysis in the Social Science, Academic Press, London.

GU C., KESTELOOT C. (2002), « Beijing’s socio-spatial structure in transition », in I. Zchnell, W. Ostendorf. (eds.), Studies in segregation and desegregation, Ashgate, Aldershot, pp. 285-311.

KERNEN A., ROCCA J.-L. (1998), La réforme des entreprises publiques en Chine et sa gestion sociale. Le cas de Shenyang et du Liaoning, Les Études du CNRI, vol. 37, pp. 1-37.

KHAN A. (1996), The impact of recent macroeconomic and sectoral changes on the poor and women in China, International Labor Organization, New Delhi.

KHAN A., RISKIN C. (2001), Inequality and poverty in China in the age of globalization, New York, Oxford University Press.

LI B. (2004), Urban social exclusion in transitional China, Centre for Analysis of Social Exclusion, London School of Economics.

LIU Y., WU F. (2006), Urban poverty neighbourhoods : typology and spatial concentration under China’s market transition, a case study of Nanjing, Geoforum, vol. 37, pp. 610-626.

ORBAN-FÉRAUGE F. (1985), Apport structurant de l’analyse factorielle des correspondances dans le traitement de données pour lesquelles on ne veut, au départ, privilégier aucun descripteur, Acta Geographica Lovaniensia, n° 26, pp 79-104.

PARK A., WANG S. (2001), China’s poverty statistics, China Economic Review, vol. 12, pp. 384-398.

SONG Y. (2007), Housing Rural Migrants in China’s Urbanizing Villages, Land Lines, vol. 19, n° 3, pp. 2-7.

SUN B. (2002), Social exclusion and marginalisation in transition – A case study of the laid-off workers in mainland China, Discussion Paper, n° 15, Sociology Research Institute, China Social Science Academy.

WU F. (2004), Urban poverty and marginalization under market transition : The case of Chinese Cities, International Journal of Urban and Regional Research, vol. 28, n° 2, pp. 401-423.

WU F., HE, S. (2005), Changes in traditional urban areas and impacts of urban redevelopment : a case study of three neighbourhoods in Nanjing, China, Tijdschrift voor Economische en Sociale Geografie, vol. 96, n° 1, pp. 75-95.

WU F. (2007), The poverty of transition : from industrial subdistrict to poor neighbourhood in the city of Nanjing, China, Urban Studies, vol. 44, n° 13, pp. 2673-2694.

XU Y. (2008), Social Assistance in China : the Minimim Living Standard Guarantee Scheme, The Institute of Social development and Public Policy, Beijing Normal University, 13p. (

ZHANG L. (2001), Migration and privatization of space and power in late socialist China, American Ethnologist, vol. 28, n° 1, pp. 179-205.

Haut de page


1 Acknowledgements: We would like to thank Yuting Liu for making his data set on poor households in Nanjing available for further analysis and Katherine Graham for her involvement in the editing and for the rapidity and precision of her corrections.

2 A national survey in 2002, carried out by the Ministry of Civil Affairs counted 19.38 million urban poor (using locally defined poverty lines). The most recent estimates, based on the Minimum Living Standard Guarantee Programme recipients (see further), consider between 5.5 and 6% of urban (non-agricultural) poor in the period 2002-2007 (about 22 million people) [Xu, 2008].

3 From 1953 to 1978, China is ruled by a socialist planned regime. In this system, the economy is nationalized and controlled by the central government through a plan. Rapid industrialisation is the main goal of such a planned economy. Based on the extension of industry at the cost of other economic sectors’ development, the system will lead to deep economical disequilibrium. From 1978 on, the introduction of reforms progressively lifts the controls on the economical and social organization, leading to the liberalization of the economy [Chan, 1996].

4 Pillar-institution in the planned economy, the work-unit or danwei controls the distribution of in-kind benefits (housing, education, health insurance, pension) that insures the worker the satisfaction of his basic needs.

5 Thus the “floating population” also includes students, soldiers, hospital patients, etc. as well as people visiting other places for several reasons.

6 Hukou is an administrative tool set in the central planning era and corresponds to the need to assign people a place of work and living in order to maintain planning coherence. In practice it has been a powerful tool to avoid over-urbanisation in China. The hukou, being either set as urban or rural, determines the right of urban citizen to be integrated in a work-unit and to enjoy social benefits. In urban areas, holding an urban hukou thus gives right to social services such as health insurance, housing, education and pension, provided through the work unit. In rural areas, it was assumed that the rural communities covered the basic needs.

7 Nanjing was the capital of China for the first time between the 3rd and the 6th century BC. It regained the status of capital later for limited periods – between 937 and 975 BC and between 1368 and 1421. It continued to serve as a capital intermittently until the 1912.

8 Only the use of the land has been privatised and submitted to market mechanisms, while the property of the land remains in the hands of the state.

9 The Minimum Living Standard Program (MLSP) is a government-funded program initiated in 1993 in Shanghai and extended to all cities in 1999. It provides cash assistance for urban households falling under a poverty line. This standard, fixed by local cities and counties governments, differs among regions. The program is not able to cover all the people in need. (Xu, 2008)

10 Once they are swallowed by the urban fabric, these villages become “Villages in city” or Vic.

11 There have been other empirical researches focusing on the problem of poverty in Nanjing by Wu and He in 2004, by Chen, Gu and Wu in 2006, and by Wu in 2007.

12 For further explanation concerning correspondence analysis, see : Orban-Férauge, F. (1985) and Greenacre & Blasius (1994).

13 In 2002, the average income in Nanjing was 900 CYN (= 100 EUR).

14 See note

15 The category « very good » was also offered in the questionnaire, only one out of the 432 respondents used that response.

Haut de page

Table des illustrations

Titre Figure 1. Socio-spatial division of Nanjing
Fichier image/jpeg, 544k
Titre Figure 3. Spatial distribution of the different groups of poor in the sample
Crédits Author : Louise Babar.Source : Yuting Liu survey, 2002.
Fichier image/jpeg, 408k
Titre Figure 4. Residential pattern of rural migrants group
Fichier image/jpeg, 568k
Titre Figure 5. Residential pattern of employed and retired group
Fichier image/jpeg, 584k
Titre Figure 6. Residential pattern of unemployed and laid-off group
Fichier image/jpeg, 552k
Titre Figure 7. Social network, feelings about residential environment and residential stability
Fichier image/jpeg, 320k
Titre Figure 8. Influence of environment on feelings, behaviour and social network
Fichier image/jpeg, 471k
Haut de page

Pour citer cet article

Référence papier

Louise Babar et Christian Kesteloot, « Poverty and Residential Environment in Nanjing », Espace populations sociétés, 2009/3 | 2009, 533-549.

Référence électronique

Louise Babar et Christian Kesteloot, « Poverty and Residential Environment in Nanjing », Espace populations sociétés [En ligne], 2009/3 | 2009, mis en ligne le 01 décembre 2011, consulté le 22 octobre 2019. URL : ; DOI : 10.4000/eps.3854

Haut de page


Louise Babar

Université Libre de Bruxelles
CP 130/02
Avenue F.D. Roosevelt
B. 1050 Bruxelles

Christian Kesteloot

Katholieke Universiteit Leuven
Celestijnenlaan 200E
B. 3001 Leuven-Heverlee

Haut de page

Droits d’auteur

Licence Creative Commons
Espace Populations Sociétés est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Haut de page
  • Logo Université de Lille 1 - Sciences et technologies
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo DOAJ - Directory of Open Access Journals
  • OpenEdition Journals