Housing production in Brussels: the neighbourhood city to stand the test of urban densification

In recent years, housing densification in the Brussels-Capital Region has become an urban development issue and tool to cope with demographic expansion and the widening gap between supply and demand as regards social and middle income housing. The choice of location is crucial in order to meet these challenges, identified in particular in the Sustainable Regional Development Plan project. Yet, with a social housing stock of 39 000 dwellings and more than 45 000 prospective tenant households, public real estate developers do not seem to consider location as a priority criterion to decide on new housing operations in the regional territory. The present article is intended to be an updated summary of research which began approximately ten years ago and whose objective was to analyse public and private housing production since 1989 with regard to the evolution of the city project [Ananian, 2010]. A second objective has been added to this initial one, namely to compare these results to the Regional Housing Plan (2005) production and the financial framework of Alliance Habitat (2015). This synopsis highlights the necessity to direct the efforts of the public authorities towards an integrated strategy of urban planning and public housing production which takes into account the proximity of the facilities, services and amenities of daily life.

Number 107, December 19 th 2016. ISSN 2031-0293 In recent years, housing densification in the Brussels-Capital Region has become an urban development issue and tool to cope with demographic expansion and the widening gap between supply and demand as regards social and middle income housing. The choice of location is crucial in order to meet these challenges, identified in particular in the Sustainable Regional Development Plan project. Yet, with a social housing stock of 39,000 dwellings and more than 45,000 prospective tenant households, public real estate developers do not seem to consider location as a priority criterion to decide on new housing operations in the regional territory. The present article is intended to be an updated summary of research which began approximately ten years ago and whose objective was to analyse public and private housing production since 1989 with regard to the evolution of the city project [Ananian, 2010]. A second objective has been added to this initial one, namely to compare these results to the Regional Housing Plan (2005) production and the financial framework of Alliance Habitat (2015). This synopsis highlights the necessity to direct the efforts of the public authorities towards an integrated strategy of urban planning and public housing production which takes into account the proximity of the facilities, services and amenities of daily life. Priscilla Ananian is a professor of urbanism and urban studies at École des sciences de la gestion at Université du Québec à Montréal. She studied urban architecture and has a master's degree in design, a DEA in territorial planning and development, and a doctorate in the art of construction and town planning (UCL, 2010). She recently published "Bruxelles, région de l'innovation : urbanisme et intégration urbaine de l'économie de la connaissance" in Cahiers de Géographie du Québec.

Introduction
1. The Brussels-Capital Region is at a turning point in its development. It must face several challenges identified in particular in the Sustainable Regional Development Plan (PRDD, 2013), such as demographic expansion and the fight against poverty. The city already has more than 1,175,000 inhabitants (IBSA, Direction générale Statistique, 1 January 2015). And, according to forecasts by the Institut Bruxellois de Statistique et d'Analyse [IBSA, 2010], the demographic growth under way since 1996 should continue. The recent demographic perspectives of the Bureau fédéral du Plan (BFP) and the Direction générale Statistique (DGS) (2016) announce a population increase for the Brussels-Capital Region of approximately 32% by 2060 with respect to 2015, compared to 14% and 16% in Flanders and Wallonia respectively.
2. As underlined in other studies, despite the importance of migrations, demographic expansion in Brussels is greatly endogenous. It is the consequence of a high birth rate and for the most part concerns populations which are disadvantaged due to their socioeconomic conditions, as well as their difficulties in obtaining housing ; ADT États des lieux de la Région de Bruxelles-Capitale, 2011]. More recently, the short-term annual growth of the population was revised upwards following the influx of asylum seekers [Bureau fédéral du Plan, 2016].
3. Demographic expansion leads to an increase in the number of inhabitants per km 2 and in the number of dwellings in the territory. This densification of the Brussels-Capital Region influences the types of housing produced and their location [Dessouroux, Romainville, 2011].
4. Let us point out, in the case of the Brussels-Capital Region, densification and the challenges it presents in terms of housing production are not confined to the regional boundaries. This phenomenon must be seen in the context of metropolisation and urban sprawl in its hinterland (Flemish Brabant and Walloon Brabant in particular). 5. Urban densification and housing production may influence the migration of more disadvantaged populations and create new forms of social inequality at metropolitan level [Bassand, 1997]. This is due in part to the dynamics of gentrification caused by the construction of new housing in certain neighbourhoods, which could above all weaken the disadvantaged households living in Brussels [Dessouroux et al., 2016].
6. Urban densification is overseen by the public authorities via urban planning, whose land-use planning models 1 are based on the concept of the neighbourhood city.
7. This article is divided into four sections. The first presents the main schools of thought which proposed the neighbourhood city as a model to favour for the development and use of land in the process of urban densification. The second analyses the issues related to the location of public and private housing production in Brussels since 1989. The third section compares these results to the production of public housing by the Regional Housing Plan and Alliance Habitat. Finally, the fourth and last section discusses the perspectives of an integrated urban planning strategy intended to favour the neighbourhood city in the densification process in the Brussels Region.

Urban densification and the notion of proximity
8. The process of urban densification, an inescapable phenomenon of the industrious 19th-century city which had a negative connotation at the time, has become a virtuous model of the territorial project in the paradigm of sustainable development [Declève, Ananian et al., 2009]. Nevertheless, the debate regarding urban densification and the right urban form has often been limited to environmental and economic dimensions, with the social dimension being neglected [Breheny, 1992;Jenks et al., 1996;Burton, 2000].
9. From an environmental point of view, since the energy crisis in the 1970s, researchers have contrasted the dense/compact city and the diffuse city, in search of a model of the ideal city which meets criteria regarding energy consumption, the optimisation of resources and the rationalisation of transport in a managerial approach [Newman and Kenworthy, 1989;Gordon and Richardson, 1997]. Furthermore, density and densification dealt with in their social dimension (in particular through social acceptability) are developed between the lines in the literature regarding gentrification and the right to housing [Amphoux, 1999;Smith, 2004;Lees, 2000 and2008;Bidou-Zachariasen, 2003;Van Criekingen, 2013]. Certain authors criticise more directly the social mix policies associated with urban densification and defend the notion of environmental justice, a concept specific to the field of ecology and the environment [Bonard, Thomann, 2009].
10. While densification has an impact on the configuration of urban centres and their physical and spatial forms, it also potentially restructures the social fabric and leads to new forms of social mix [Bromley et al., 2005;Rose, 2006] and the cohabitation of activities, which creates many challenges for the public authorities. These are the reasons why densification, used as an urban development tool [Da Cunha, Kaiser, 2009], questions the neighbourhood city as a desirable and positive model for the development and use of land. The concept of neighbourhood city advocates social and functional mix, access to services and facilities close to home, as well as a reduction in the length and cost of journeys.
11. The neighbourhood city is inspired by several principles and schools of thought, such as the neighbourhood unit proposed in 1920 by Clarence Perry. The neighbourhood unit is based on principles of social, physical and institutional design at neighbourhood level. In these neighbourhoods, schools, services and parks are situated within walking distance from home. This model was proposed in order to meet the need to preserve the living conditions of the neighbourhood with respect to the emergence of the car in the industrial city, and was revisited later in Europe by Léon Krier [1977] with the notion of public comfort of the neighbourhood 2 and in the United States by the forerunners of New Urbanism [Walters and Brown, 2004]. The New Urbanism movement complicates the principles of neighbourhood design (TND Traditional Neighbourhood Development) by integrating a perspective of regional planning based on public transport infrastructures (TOD Transit-oriented Development). 3 These different models all advocate the proximity of housing to services, amenities and facilities, which should be within walking distance from home [Ananian, Declève, 2010].
12. Nearly one hundred years after the emergence of the concept of the neighbourhood unit, mobility and the reduction of distances by new information and communication technologies [Huriot, 1998] call into question the notion of physical proximity behind these models.
13. The notion of proximity is also mobilised by the regional economy, which defends the idea of urban area to reinforce the attractiveness and competitiveness of territories based on polycentric development [Kloosterman, Musterd, 2001]. As regards physical and spatial development, these theories favour the multipolar city model. This model advocates development according to centres where there is a concentration of a certain number of strategic activities which are well connected to transport infrastructures.
15. This dual reference to proximity is however included in the PRDD project: Brussels must be considered as a multipolar territory where differentiated centres of activity are organised around a hierarchical structure of public transport nodes. In addition to the multipolarity developed at regional level, there is the dimension of the "neighbourhood city" which is based on the distinctive features of each municipality and each neighbourhood. [PRDD project, 2013, p. 15] 16. The PRDD project underlines the need to organise the territory according to an intricate network of nearby facilities and services: The "neighbourhood city" is one of the essential levers to strengthen the diversity of the Brussels territory and to meet the expectations of the inhabitants of Brussels as regards the quality of life: availability of sufficient housing in terms of number and quality, infrastructures and local facilities such as nurseries, schools, recreational areas and green spaces, as well as economic activities providing local employment and training. [Ibid.] 17. These two citations taken from the PRDD project raise a question which is at the heart of this article: to what extent are the neighbourhood objectives expressed in the successive city projects of the Brussels-Capital Region materialised in housing production -in particular public housing production?
18. Our hypothesis is that while urban densification in the Brussels-Capital Region favours the geographical proximity of people and activities, the processes of housing production do not consolidate the neighbourhood city due to restrictions which affect the choice of location and programming of public real estate operations.
19. Historically, in Brussels, social housing production is in keeping with two perspectives related to location. On the one hand, it involves the production of housing in an environment where buildings are dilapidated and the population is disadvantaged, i.e. by developing the historically working-class neighbourhoods (Cité Hellemans, Remparts des Moines, neighbourhood contract operations, etc.). In this perspective, the location is a determining factor and leads to a concentration of means in certain neighbourhoods. On the other hand, it involves building in places where there is still enough room for less dense housing developments and, consequently, where the value of real estate is lower (Cités jardins, such as Logis-Floréal, Bon Air, Cité Moderne, Cité Modèle, etc.). This perspective is evolving towards a programming of social mix at regional level in order to avoid a concentration of social housing in the territory. 20. These two perspectives related to location converge in the city's heritage, under the supervision of the Société du logement de la Région de Bruxelles-Capitale ( S L R B ) a n d m a n a g e d b y t h e Sociétés Immobilières de Service Public ( S I S P ) . T h i s h e r i t a g e h a s b e e n constituted over the past one hundred years, and contains very diverse housing typologies and densities (single-and multi-family dwellings, small blocks of flats and high-rise blocks). The following section will attempt to describe these two perspectives related to the location of public housing production since the regionalisation of 1989 with regard to the logic of private production.

Public and private housing production: location challenges, 1989-2013
21. Demographic growth and the increase in the gap between supply and demand in social and middle income housing nurture the recent demand for public housing production 4 in the Brussels-Capital Region. However, in as much as the administrative boundaries of the Brussels Region are a constraint to regional development, among others due to the cost of land [Vandermotten, 2014], it is essential to put into perspective the production of new housing units by the public authorities with the dynamics of housing production in the private sector. It is essential in order to understand the extent to which public housing production is influenced by private housing production and 3 how it does or does not achieve the neighbourhood objectives of the different city projects.
22. Below, we shall concentrate on operations involving 10 or more housing units. Recent housing production in Brussels (i.e. since 1989) is represented mainly by this type of operation. Since 2003, planning permission authorising 10 or more housing units represents more than 70% of the total housing production in the Brussels-Capital Region [Observatoire des permis logement, 2011, 2013].
23. Furthermore, this building operation threshold (10 or more housing units) allows an integration of a certain mix of activities in the programming (shops, services and facilities), which requires better coordination of building programming with that of the city, in particular in terms of service. Finally, more than 90% of recent regional public housing production consists in operations of at least 10 housing units.
24. In concrete terms, the research considers planning permission authorising 10 or more housing units from 1989 to 2013. The number of rooms in group housing or the number of hotel rooms are not included in the inventory. As the 2013 data are not consolidated, their use in figures 1 and 2 is restricted to the scale of statistical sectors. For subsequent analyses requiring a location of operations down to the plot level, the data used end in the year 2012. The number of dwellings authorised which are built after receiving planning permission is higher than 90%, which is why we have decided to use these data to analyse recent housing production. For more information concerning the methodology, see Observatoire des permis logement n°1.

Evolution of the geography of public housing production
25. The legal production of housing is analysed here, in as much as the unit analysed is authorised housing and not built housing. The gap between the number of dwellings authorised and built is approximately 6% [Ananian, 2010, p. 241]. 5 26. Figures 1 and 2 provide a more detailed view of the geography of public housing production of 10 or more housing units in Brussels, with respect to all operations of the same size. The reinforced development areas for housing have contributed to the concentration of public production in the central and surrounding neighbourhoods situated to the west of the canal. The definition of EDRL and EDRLR perimeters, in the first and second Regional Development Plan respectively, was aimed at concentrating the investments of the public authorities in the sectors where the buildings were dilapidated and where the most disadvantaged populations lived. Several mechanisms in favour of housing and urban renovation have been developed in these perimeters, such as housing renovation subsidies, neighbourhood contracts, programmes for the redevelopment of public space and the creation of local facilities.
27. During the first two periods (from 1989 to 1995 and from 1996 to 2002), public housing production has materialised in mid-sized operations which represent respectively 24% and 23% of the entire production of 10 or more housing units (Figure 1). In all of the production (Figure 2), the number of dwellings authorised is particularly significant in the Pentagon and the inner ring. This must be seen in the context of a proactive policy of the Brussels government to develop central neighbourhoods [Van Criekingen, 2013].
28. The number of dwellings authorised for the entire housing production (public and private) was much lower for the Brussels Region during the 1996-2002 period corresponding to the first PRD. Between 1989 and 2013, the lowest average annual number of authorised dwellings is observed (1,583 compared to 2,537 for the entire period, see table below). This is explained in part by economic factors related to interest rates and to a tendency for property development in Brussels to be centred on office production [Ananian, 2010].
29. The situation evolved between 2003 and 2008, with massive housing production throughout the regional area ( Figure 2). There was also the beginning of a process of deconcentration of public housing production in the western part of the outer ring. The share of public production in all of the housing production of 10 or more housing units was only 14%, i.e. a decrease of approximately 10% with respect to the previous two periods (Figure 1). 30. This period also corresponds to the implementation of the Regional Housing Plan (2005) and the Plan 1 000 logements of the City of Brussels and the Brussels CPAS (2006). Despite the efforts of the public authorities in terms of the creation of new housing units, the share of public housing production decreased due to a significant increase in housing production by private stakeholders. 31. Finally, since 2009, the process of deconcentration of public production has continued towards the eastern part of the outer ring, this time with larger operations. As regards the entire production of 10 or more housing units, it is concentrated in the canal area and the western part of the inner ring, although in a less steady manner than in the previous period (reduction in the average number of authorised dwellings from 3,400 to 2,992 units, for 2003-2008 and 2009-2013 respectively).

The contribution of housing production to the Brussels regional urban project
32. Several locations may be identified in the development projects which have been carried out in the Brussels Region: the city rings (Pentagon, inner ring created during 19th century urbanisation, outer ring created during 20th century urbanisation) and the municipalities in the first PRD, and the regional development poles and canal area in the second PRD and the PRDD project.
33. With respect to the city rings, the share of public housing production of 10 or more housing units in the Pentagon and the inner ring has decreased considerably since 2003 in favour of private housing production. Between 2009 and 2013, more than 70% of public housing production was concentrated in the outer ring, while private housing production has continued to increase in the inner ring, even though this type of production in the neighbourhoods outside the centre represents only one third of private housing production of 10 or more housing units.

34.
A comparison of the evolution of public housing production shows that the disparities between municipalities in the inner and outer ring have decreased over time. This decrease must be seen in the context of a change of direction of public policies. During the first and second PRD, the territorial public policies for the positive discrimination of disadvantaged neighbourhoods to be renovated defined the perimeters for housing reinforcement and consolidated a territory organised in rings around the centre. But later, the Regional Housing Plan encouraged an effort to deconcentrate production in the outer ring, which we shall discuss in the third section (Berchem Sainte-Agathe, Jette, Anderlecht, Molenbeek, Neder-Over-Hembeek, Haren).
35. The two perimeters of housing reinforcement (EDRL and ERDLR) have led to a concentration of public investments in the central neighbourhoods and those outside the centre. While public housing production was carried out especially in these neighbourhoods during the first two periods analysed (1989-1995 and 1996-2002), the share of public housing production in the EDRLR (with respect to all public housing production) has decreased considerably since 2003. A reinvestment of the private sector in the inner ring may be seen, despite the fact that the average share of private housing production in the territory of the EDRLR (with respect to all private housing production) is approximately 20%, compared to 48% for public housing production ( Figure 3).  2009 concerning the elaboration of a new city project, a new model has emerged [Ananian, 2014b]. This model is based on the (former) concept of development poles, going beyond the logic of municipalities and city rings.
37. According to the terminology of the PRDD project (2013, p. 63-64), "development pole" refers to a lever of territorialised public action able to consolidate regional goals and integrate the sector-based strategies related to the different areas of public action (housing, employment, transport, etc.). In concrete terms, the development poles are identified sectors in the city which present strategic development opportunities involving the coordination of land-use planning, housing production and transport.
38. The development of poles and their surroundings by the private sector is slightly more intense than that of the public sector (Figure 3).
39. Finally, it is worthwhile to underline the evolution of housing production in the canal area -a strategic territory of regional development identified in the second PRD and in the PRDD project. For the 2009-2012 period, almost one third of private sector authorised housing is concentrated in the canal area, whereas public housing production favours the western part of the outer ring. Let us point out that a significant share of these units are part of two major building projects: the Up Site tower and Tour & Taxis. While the former was completed in 2013, the construction of the latter has not begun.
40. In the light of these results, we may conclude that the proactive policies to redevelop the central neighbourhoods, which translate into the defining of reinforced perimeters for housing, have only favoured private housing production recently. The surroundings of areas of regional interest designated as development poles in the PRDD project benefit from the existence of older and more stable private housing production.

The contribution of housing production to the neighbourhood city
41. Between the regional urban project and the neighbourhood city is the public transport network, which organises the territory and facilitates the population's access to places where activities take place. We have therefore analysed the connectivity of new housing operations to public transport networks, according to ABC 6 areas defined by the Règlement Régional d'Urbanisme (RRU, 2006). On this matter, due to a gradual appearance of public housing production in the outer ring and of private production in the inner ring, the share of public production in area C -which is not as well connected to public transport -has tended to increase since 2003. 8 6 The three areas of accessibility defined by the RRU refer to a zoning of the regional territory based on accessibility by public transport in view of defining parking standards outside public roads. For more information, see   42. The share of private housing production within walking distance to an underground station or a train station in private production as a whole is higher than that of public housing production. Furthermore, the share of housing production (private and public) is, overall, lower in the development poles identified in the PRDD project served by underground stations. This distinctive feature is explained by the fact that these poles waiting to be developed are not yet well served by public transport. 43. These observations suggest possible competition in the choice of location between functions (housing, activities, facilities) and between the types of housing (social, medium-sized or free-market housing).
44. Furthermore, the neighbourhood city postulates that each inhabitant should have access to shops, facilities and amenities within a walking distance of 5 to 10 minutes from home. The share of public housing production bordering commercial hubs defined by the Plan régional d'affectation du sol (or PRAS) -where there are local everyday shops -was approximately 70% during the 1989-1995 period. But, since 2009, it has decreased to approximately 30% of the average annual number of authorised dwellings.
45. The same observation applies to educational facilities likely to attract and retain families in the city, included in the census of the Agence de développement territorial in 2010. Approximately half of the average annual number of dwellings authorised in favour of public real estate developers are located near a nursery or a kindergarten or primary school. However, the analysis of planned development poles served by nurseries and public schools shows that the share of public production served by them is similar to the share of private production, decreasing from 50 to 25%.
46. The neighbourhood city also advocates access to public and recreational spaces such as playgrounds. The level of service (defined by the walking distance between these types of amenity and the authorised housing) is roughly the same for public and private residential productions: it is low, in as much as it is slightly less than half of the average annual number of authorised dwellings.
47. We may conclude from this that the share of the average annual number of authorised dwellings near facilities, services and amenities is slightly higher for public production than for the private sector. This observation is nevertheless mixed. Only half of public housing production really meets these conditions. This confirms that the planned development poles and their surroundings should integrate the production of new facilities in their programming to truly meet the objectives of the neighbourhood city. Finally, the share of public production near facilities tends to decrease over time, which leads us to question in the following section the perspectives regarding the location of operations for the two current regional programmes, as well as the social challenges resulting from this spatiality of housing production.

Regional Housing Plan and Alliance Habitat: the challenges of residential deconcentration
48. Since 2005, due to the housing crisis and in response to the challenges of demographic expansion -and consequently urban densification -the Brussels-Capital Region intensified its strategies to expand its social and middle income housing stock by implementing the Regional Housing Plan, which provided for the creation of 5,000 new units, 3,500 of which are social housing and 1,500 of which are middle income housing. At the beginning of 2015, the Alliance Habitat financial framework was added to this plan, aimed at the creation of new operations by 2020 in response to the intensification of the housing crisis (Figure 3).
49. In June 2015, the Regional Housing Plan included 56 projects for a total production of 4,070 dwellings, and there were 34 projects within the Alliance Habitat financial framework for a total of 1,117 dwellings. Among these, only 8 projects were notified by the government of the Brussels-Capital Region, corresponding to 253 dwellings (Table 2). There were also an additional 10 projects overseen by SFAR (a subsidiary of Société régionale d'investissement de Bruxelles -SRIB) for a total of 848 dwellings, and 16 projects overseen by the Fonds du logement, corresponding to 399 dwellings.

Public housing production in the process of deconcentration does not contribute strategically to the regional urban project
50. The map ( Figure 5) illustrates the distribution of the operations of the two programmes according to the share attributed to social and middle income housing.
51. There is clearly a deconcentration of regional public housing production which, however, does not contribute to the emergence of new polarities. Actually, the majority of operations do not integrate a mix of functions beyond a few facilities of collective interest (community halls, offices for associations, social restaurant, nurseries).
52. As underlined in the programming of the Regional Housing Plan and as shown on the map, social housing represents more than two thirds of public production, especially in the outer ring to the west of the canal, while middle income housing exists mainly in the inner ring, in a perspective of cross-mixing (middle income housing in the neighbourhoods outside the centre and social housing in the outer ring).
53. Ten years after the Regional Housing Plan was announced, the results are mixed. Only a quarter of the units initially planned (14 projects, 1,290 dwellings) were received, mainly involving operations situated in the outer ring [data transmitted by SLRB, 2015].
54. The map ( Figure 6) shows that the majority of operations trundle through the administrative steps (obtaining planning permission, award of service contracts, award of works contracts, etc.). This observation highlights the urgent need to rationalise and accelerate the process. Pending projects are also shown, such as Dame Blanche in Woluwe Saint-Pierre, and the abandoned projects, such as the Bolivar project overseen by SFAR in the North Station neighbourhood. In addition to the obstacles faced by major building projects involving project set-up, the elaboration of planning instruments and the obtaining of planning permission, there are citizen opposition movements with respect to the programming of social mix in the affluent neighbourhoods of the region.

Decline of the neighbourhood city: the growing gap between public housing production and the level of service in terms of facilities
55. In June 2015, the entire production of the Regional Housing Plan and Alliance Habitat amounted to 90 operations, which we have analysed according to the same criteria as those used for the analysis of housing production since 1989 with respect to its progress. These criteria concern the contribution of operations to the four models analysed in section 2.2 (city rings, municipalities, development poles, canal area) and their level of service (public transport, borders of commercial hubs, education facilities, playgrounds).
56. The objective is to evaluate the contribution of the Regional Housing Plan and Alliance Habitat to the regional urban project and the neighbourhood city. By taking into consideration the same facilities, services, shops and amenities, we have classified the projects according to their level of service: low, moderate or high. These levels were then viewed in the context of the freedom of intervention of operations, which depends on their progress (preliminary draft, permit applications, award of contracts, construction).

10
Priscilla ANANIAN, Housing production in Brussels: the neighbourhood city to stand the test of urban densification,  57. The projects with no freedom of intervention are those which have already been built or are being built, although the creation of new facilities in the territory may still make up for certain shortcomings in terms of service. The projects with reduced freedom of intervention are pending planning permission or award of contracts. Finally, the projects with moderate freedom of intervention are upstream from the administrative processes and may possibly undergo a change of location and programming.
58. We shall concentrate on operations with moderate freedom of intervention, i.e. whose location and programming may undergo changes.
59. Due to their distribution in the territory, future projects do not necessarily consolidate the development poles identified in the PRDD project and their surroundings. Future production is concentrated in the outer ring. There is therefore a process of deconcentration of social and middle income housing which does not contribute to the configuration of new poles, which represent the structuring model of the PRDD project.
60. For access to public transport, the future projects are located mainly in area C, which is less connected to the existing public transport networks.
61. The summary map (Figure 7) links the progress of operations and the potential for public transport, services and facilities. The projects received or under way with a high level of service are located near the centre, whereas the projects with a moderate or low level of service are located between the inner and the outer ring, or at the edges of the outer ring.
62. The projects with reduced freedom of intervention with a high level of access to facilities and services are above all small projects. The projects with a moderate level of service are located mainly in the outer ring, whereas those with the lowest level of service are located on the outskirts of the city and are larger in scale.
63. The projects with moderate freedom of intervention with a high level of access to facilities and services are also located in the centre and in the neighbourhoods outside the centre. The projects with a moderate level of service are located near planned development poles, which confirms the need for alignment between the programming for residential operations and major urban projects. It is not surprising that the operations with a low level of service are located on the outskirts of the city.

12
Priscilla ANANIAN, Housing production in Brussels: the neighbourhood city to stand the test of urban densification, Brussels Studies, Number 107, 19 December 2016, www.brusselsstudies.be Figure 6. Progress of Regional Housing Plan and Alliance Habitat operations. Source: SLRB, 2015. 64. Finally, the analysis shows the decline of the neighbourhood city as a model due to an increasing gap between housing production and the level of access to services and facilities. This decline is due to several factors. Firstly, there is a lack of coordination between urban planning and urban sector-based productions, such as the production of housing, facilities and transport infrastructures. Secondly, the length of development operations associated with administrative steps and citizen opposition during the setting up of major public building projects is another factor. Lastly, it is also linked to the growing scarcity of land and its cost in the Brussels Region. This forces the authorities to work according to a logic of opportunity rather than according to a strategic vision of the development of public housing. This factor must be seen in the context of the renewed attractiveness of Brussels for private investors, which confirms that the best locations are in the hands of the market (section 2). This renewed attractiveness is justified by the rhetoric on urban densification and the zoning changes carried out in the PRAS revision, PRAS démographique. 7

Perspectives of an integrated urban planning strategy
65. Our research shows that the recent production of public housing does not necessarily consolidate the Sustainable Regional Development Plan project, when it could be a powerful lever in favour of the city project elaborated by the Brussels Region.
66. This observation highlights the need for the development of the regional housing and urban planning policy to be integrated in public action. The regional housing policy is a dynamic and integrated system, whereby it is not enough to produce housing units to fill an urgent gap between supply and demand.
67. In order to achieve such integration, according to our analyses, it would be necessary to develop a true strategy for the location of social and middle income housing in the regional territory, and to stop operating on a case-by-case basis according to the opportunities. Such a location strategy must include principles which contribute to consolidating the regional urban project based on development poles and their surroundings as well as the neighbourhood city at neighbourhood level. This cannot be done without accompanying the dynamics generated by private housing production, which seems to consolidate the model of the neighbourhood city for those who are better off.

13
7 PRAS démographique aims to adapt the reference plan for land-use planning constituted by PRAS to the rapid population growth in the Brussels Region, which is particularly noticeable in the central neighbourhoods and the canal area. For more information: https://urbanisme.irisnet.be/pdf/pras-demographique Priscilla ANANIAN, Housing production in Brussels: the neighbourhood city to stand the test of urban densification, Brussels Studies, Number 107, 19 December 2016, www.brusselsstudies.be Figure 7. Level of access and freedom of intervention of Regional Housing Plan and Alliance Habitat 68. There are three main principles of this strategy. The first concerns the coordination between public housing production, urbanism and public transport. The restructuring of transport networks must take place upstream in order to restrict urban densification. Furthermore, as the inhabitants of social housing are more dependent on public transport, the location of new units should favour connectivity with transport networks, and should preferably be within walking distance to underground stations, pre-metro stations, tram stops and train stations, in particular the urban train stations included in the RER project, such as Germoir in Ixelles.
69. The second principle consists in aligning residential and urban programming in terms of facilities and amenities. The results of our analyses show that only 50% of the average annual number of dwellings authorised in favour of public real estate developers are served by nurseries, schools and public playgrounds. The difficulty therefore lies in project set up, as housing production and the creation of public facilities are a matter for different stakeholders and financing programmes.
70. The third principle consists in reinforcing the role of social and middle income housing in the development poles and their surroundings. Admittedly, the major urban projects in strategic areas integrate social and middle income housing in their programming thanks to master plans in particular, but these are not only incentive mechanisms. It is important to reinforce the normative mechanisms intended to favour the creation of social and middle income housing in the poles recommended by the PRDD project, to complement the planning permission charges 8 imposed with town planning permits authorising housing operations of 1,000 m 2 or more [Kestemont, De Muynck, 2015].
71. Furthermore, especially since 2003, there is more and more private housing production in the surroundings of the planned development poles. This private production is more stable in time than its public counterpart due to land management and its capacity to implement large-scale projects more quickly. Once again, this raises the question as to the use of mechanisms according to the logic of the EDRLR, i.e. incentive for small interventions -renovations, enhanced value of housing above shops or management of private housing by the public authorities, via Agences Immobilières Sociales (AIS). These mechanisms may contribute to reinforcing the role of social and middle income housing in the surroundings of the poles, with the creation of new housing operations.
72. The definition of development poles in the PRDD project reinforces the will of the Brussels Region to ensure the coherence of a series of projects at different stages of progress [Ananian 2014, a and b]. Thus, the multipolar city echoes a strategy which goes back to the regionalisation of 1989 and which has evolved with the first two PRDs, to be consolidated finally in the PRDD project.
73. This representation of the city project has evolved from a municipal logic and one based on the rings (from the centre to the outskirts), towards a regional logic based on spaces devoid of centrality, with most planned development poles being port, railway or industrial wastelands.
74. The Brussels Region has few levers to achieve this model (which is nevertheless that of the PRDD project) and to reinforce the role of social and middle income housing among property developers. Yet the municipalities may count on historical structures of housing production and development such as the SISP and the ring-based approach has benefited over the past 25 years from positive discrimination territorial policies for neighbourhoods (EDRL, EDRLR, neighbourhood contracts, ERDF area etc.).