I. Recent trends in shelter projects

Housing conditions and political priorities vary considerably between developing countries. Yet, it is possible to detect a general trend in the approach to shelter projects. Throughout the world, it has been realized that the cost of building houses or apartments for those who cannot afford their full cost in the private market, can only be achieved by high levels of subsidy. Due to financial constraints in most developing countries these subsidies restrict the number of units provided to a small proportion of total demand. The remaining households are forced to fend for themselves in unauthorized settlements.

The emphasis of recent projects has therefore moved away from the direct provision of land, services and housing for a minority, towards a range of options that concentrate on the provision of one or two of these elements, such as land, or services. Furthermore, the residents of many recent projects are allowed to control the provision of houses according to their own priorities and resources. These changes are intended to reduce unit costs and enable scarce resources to reach a significantly higher proportion of households in need. They are also intended to enable beneficiaries to obtain the types of housing they prefer at costs they can afford, so that economic, social and environmental priorities can be achieved.

This shift to a more flexible approach can also be observed as changing attitudes towards existing forms of shelter and settlements that do not conform to official norms and standards. Whereas slum-clearance projects were the conventional response of public sector agencies during the 1960s and 1970s, it is now widely accepted that these only dealt with the symptoms and not the causes of shelter problems. Displaced households simply settled elsewhere, so that the social antagonism and economic costs offset any other benefits achieved. Recent efforts have focused on improving unauthorized settlements wherever possible, since the costs of doing so are small compared with the social and economic benefits achieved.

The case studies reviewed in this report reflect the gradual transition from conventional to more flexible and innovative approaches. They include projects for new housing and the upgrading of existing settlements by a combination of public agencies, non-governmental organizations (NGOs) and private developers. The types of shelter provided include apartments, row and core houses, prefabricated units and serviced plots. The size of projects range from 64 units on a site of one hectare to more than 10,000 units on several hundred hectares (see table 1).

The transition from conventional provision to more innovative support strategies has required considerable changes in the attitudes of professionals and in the roles of governments. These changes have not been achieved quickly or smoothly. Some of the difficulties involved are clearly evident in Turkey, where projects have tended to be small and not easily accessible to low-income groups. About 70,000 plots were allocated in sites-and-services projects between 1978 and 1989; an average of less than 6000 a year. Although this represented a considerable increase in public-sector supply, it compares poorly with the performance of informal processes that have provided about 95,000 units a year, representing 90 per cent of total low-income supply (Tokman, 1990: 10-11).

The Aktepe project in Ankara provides a good illustration of the conventional approach. In 1965, it was designated as a “squatter prevention area” by the Ministry of Reconstruction and Resettlement. By 1975 it had provided 3626 units in the form of apartments, core houses, prefabricated units and serviced plots (Tokman, 1990: 13-14). Although the project was centrally located in the city, poor project management delayed implementation and increased costs, few of which were subsequently recovered.

The Tarsus project reflects a change in that provision was based upon a housing-demand analysis, with standards and forms of housing adapted to meet the needs and resources of different sectors of demand (Tokman, 1990: 22). It also reflects a move to decentralize decision-making to local levels. Yet, decentralization of resource mobilization and allocation processes still needs to be effective. At the level of the individual project, it appears impossible to overcome this limitation. Recent projects in Turkey, such as the Bati-Kent development scheme near Ankara, suggest that the issues of scale and management efficiency are now being addressed. Few of the units in these projects, however, are affordable to low-income households, especially if transport costs to places of employment are included.

Another example of the difficulties in changing the approach to projects for new housing can be seen in the case of Colombia. For many years only a few shelter projects were undertaken. Most of these were conventional construction projects executed at a small scale and incapable of being developed into truly effective solutions. Most of these projects had little relationship to other sectoral policies and actions and rarely even fulfilled their internal objectives (Utria, 1990: 12). During the 1980s a substantial number of housing units was constructed, but these units catered only for the middle-income groups. Sites-and-services projects were introduced in an attempt to compete with illegal subdivisions. They were reasonably successful in satisfying their internal objectives of providing basic services to low-income households and security of tenure to land. Yet, the sites-and-services projects have not been free of political motivations or subsidies. Furthermore, they have attracted migrants from rural areas, drawn by the prospect of a generous solution (Utria, 1990: 30).

Some projects, however, have been undertaken at a scale that enables them to address total demand. The Kennedy City project in Bogota, for example, provided more than 10,000 dwelling units for 120,000 people in the form of apartments, row houses and plots for aided self-help construction in the early 1960s. It has since expanded to accommodate nearly 2.5 million people in what is now almost an autonomous metropolitan settlement (Utria, 1990: 80). Another project known as Bolivar City (also in Bogota) has also provided more than 10,000 units in the form of serviced plots. The project was intended to provide a legal and affordable alternative to squatting or unauthorized settlements. An added innovation was that the project also involved improvements to the largest marginal settlement in Bogota and directly or indirectly benefited 1.3 million people (Utria, 1990: 37-38).

Another example of the gradual evolution of the project approach can be seen in the case of Zimbabwe. When the populations of cities expanded rapidly, partly due to increased rural-urban migration during the war for independence, the Government implemented ultra low-cost housing projects developed by large scale contractors. Since independence, however, national policy has concentrated on seven measures: freehold tenure, higher minimum standards, aided self-help, building brigades, housing co-operatives, partnerships between public and private sectors, and full cost recovery (Mutizwa-Mangiza, 1990: 10).

The Kuwadzana project in Harare represents an example of current government approaches to addressing the housing problem. This has exerted a considerable influence on national policy through its emphasis on aided self-help for households below the median-income level (Mutizwa-Mangiza, 1990: ii). The types of housing provided represented a compromise between the desire to raise minimum initial standards, whilst at the same time recovering project costs. In this sense, the project served as an effective means of introducing changes into other developments by both public and private sectors.

Since 1980, eight shelter projects have been undertaken in Harare. Four of these have been locally funded, three by the World Bank and one by the German Agency for Technical Cooperation (GTZ) and US AID. There has been an increasing emphasis on the role of the private sector, especially building societies and employers. Building societies played a major part in the development of the Kwekwe-Gutu project, despite initial concerns regarding affordability and default rates in lending to low-income groups. The project sought to develop affordable and innovative housing solutions based on self-help practices. It also sought to achieve replicability through the establishment of domestic thrift societies and other finance mechanisms. As will be discussed below, these innovations have greatly increased the ability of public-sector agencies to address the scale and nature of housing demand in Zimbabwe.

A similar evolution has taken place in many other countries. In Indonesia, for example, government housing projects were initiated as part of the first Five Year Development Plan in 1969. These started as research and development projects, but were expanded in the second plan when Perum Perumnas, the National Urban Housing Development Corporation (NUHDC), was established. Since then, the project approach has been expanded regularly. Housing built by both public and private sectors in the period from 1984 to 1989 amounted to about 300,000 units, or an average of 60,000 a year. Yet, unauthorized settlements increased even more rapidly and the Kampung Improvement Programmes were expanded to bridge the gap. Some settlements were considered too difficult to improve, however, and the Urban Renewal Programme initiated in 1979 was directed at these. Complete settlements were demolished and rebuilt on the same sites using apartment blocks. Multi-storey housing is not, however, popular in Indonesia and many people decided to move elsewhere (Herlianto, 1990: 13).

An early example of Indonesian housing projects is located in the east of Jakarta in an area known as Klender. This was one of the first projects designated by NUHDC when it was established in 1974 and provided for a mixture of housing options from apartment blocks, duplex houses, sites-and-services with core houses and ready-to-be-built plots. The project was nominated by the Government of Indonesia as one of four demonstration projects during the International Year of Shelter for the Homeless (IYSH) in 1987. This suggests that although it was developed some years ago, it is still regarded as reflecting current thinking on housing and the role of the public sector. The project succeeded in providing housing for about 57,000 people, or nearly 10,000 households. More than 75 per cent of these, however, work for the Government or the army. Very-low-income households, or those working in the informal sector, had difficulties in participating (Herlianto, 1990: 21).

The increasing tendency to involve the private sector in housing projects affordable by low-income groups is illustrated by the Margahayu Raya project in Bandung. This was designed specifically to address low-income needs and was designed and implemented in only two years. It was frustrated by inflation, however, and ended up accommodating middle-income households (Herlianto, 1990: 33-35). In an even more innovative approach, NUHDC has initiated joint public-private-sector projects in which private developers are free to redevelop parts of inner-city slums for commercial purposes, in return for their agreement to build rental housing units for the existing low-income residents on the remainder of the site.

According to Silas (1991), the main development in the shelter sector in Indonesia during the last five years has been in improving the quality of prevailing projects, especially the Kampung Improvement Programme, which has been adapted and refined to suit local needs. In” Surabaya, it is intended to introduce an integrated urban development programme, linking the planning, implementation and management of settlement and infrastructure projects. Guided land-development projects are also envisaged.

Among all countries that have adopted an enabling strategy, Sri Lanka must be considered the most radical. In 1984, the Government initiated the Million Houses Programme (MHP). The programme consisted of six sub-programmes covering rural and urban areas, the private sector, plantation housing and major resettlement projects. The urban sub-programme consisted of about 300 housing projects in 51 local authority areas. It was launched in 1985 with three objectives:

(a) Making a substantial and lasting impact on the housing situation of the urban poor;

(b) Placing the poor at the centre of the process (by adopting a devolved approach);

(c) Ensuring that the approach was sustainable in the longer term.

Sri Lanka’s prominence in the adoption of an enabling or support approach ensures that its experience is of particular interest and significance. Yet, it should be noted that the MHP was introduced as a national shelter strategy, which took the form of a coordinated, large-scale programme at the outset. This makes it fundamentally different from an incremental project approach adopted in most countries, where innovative projects are commonly used as experiments to test the validity of new policy options before they are adopted at a large scale. The issues of replicability and the need to generate political support had already been resolved in principle. One of the most significant issues raised by the Sri Lankan experience thus relates to the development of institutional capability to implement the new approach effectively. The scale of the programme and the commitment of the government to the approach have generated an equal commitment to monitoring and evaluating the progress of the programme. Bottlenecks can thus be identified and overcomed. Such a willingness to learn from experience is not present in all countries.

The question of subsidies is critical in assessing the possibility of expanding projects and programmes to the scale required on a long-term basis. In Sri Lanka land-ownership rights have been granted to all slum dwellers in the city of Colombo. This has been done free of cost except for a small monthly charge to cover the cost of deeds (Jayaratne, 1990: 21). This hardly acts as a deterrent to the further development of squatting and other forms of unauthorized development.

Another crucial issue in the Sri Lankan approach is the emphasis placed upon decentralization, not just of implementation, but of decision-making as well. This was also emphasized in the rural programme, where the new approach was promoted under the slogan “Minimal intervention, maximum support by the State and maximum involvement of the builder families”. The rural programme, which was initiated in 1984, did not abrogate the State’s responsibility for assisting the poor. Instead, it involved taking on a more difficult responsibility, that of providing and maintaining the necessary components of secure and affordable land, infrastructure and facilities to enable households to organize their own housing according to their means and priorities (Development Planning Unit, 1984:4).

Many other examples of innovative shelter projects exist in other countries. The Incremental Development Scheme, or Khuda ki Basti undertaken by the Hyderabad Development Authority (HDA) in Pakistan, for example, is notable in attempting to learn from the practices adopted by illegal developers, or land grabbers. The HDA even went so far as to employ one. As a result, initial standards of services provision are modest, and absolutely no standards are imposed concerning the design or construction of individual houses (van der Linden, 1989: 9). To ensure that only low-income households, or those in genuine need applied, the project agency established a reception area in which all households applying were required to live for two weeks. After this period they were allocated a plot and expected to begin construction of their house immediately (van der Linden, 1989: 10). The approach was developed by HDA’s director in secrecy, since it was feared that powerful vested interests, both inside and outside government, would seek to prevent it. Only after the scheme had been approved by the Chief Minister was it formally presented to the governing body of HDA for approval and implementation (Hasan, 1987: 17).

The willingness to address vested interests and use shelter projects as experiments as illustrated by the Khuda ki Basti case study, indicates the extent to which the project approach has evolved in recent years. Although such examples present an encouraging sign of steady progress, the majority of projects remain dependent on direct or hidden subsidies, achieve poor levels of cost recovery and fail to achieve replicability.

Approaches towards existing low-income settlements have undergone a similar transformation to that of new shelter provision. Relocation projects have been replaced by projects to improve or upgrade such settlements wherever possible. In some countries (e.g., Colombia), slum-clearance and relocation projects are still undertaken. These are less common though, and likely to be for specific reasons relating to urban renewal programmes, rather than as attempts to improve the appearance of cities, or to remove the poor. In the Chambacu project in the tourist city of Cartagena, about 1800 households were offered subsidized plots with completed houses elsewhere in the city, as compensation for vacating their informal settlement near the major tourist sites. These subsidies were intended to be generated from profits from the commercial centre developed on the site and were agreed to by residents. In the event, however, the commercial development has not taken place, suggesting that the development was, in practice, intended as a slum clearance project rather than one of urban renewal (Utria, 1990: 68).

In many other countries, upgrading projects have been undertaken at a scale sufficient to benefit the majority of low- income households. The greatest achievement in this respect must be that of Indonesia, where the Kampung Improvement Programme has benefited almost all low-income residents of informal settlements in Jakarta, Bandung and Surabaya. Although these projects concentrated on the physical upgrading of public services and adopted an engineering approach that did not attempt the full recovery of project costs (Herlianto, 1990: 18), they stimulated residents to make substantial secondary investment in shelter improvements. These dramatically improved living conditions and public health throughout each city, at density levels that would have been difficult to achieve in new development projects.

Upgrading projects place increasing emphasis on community participation in the formulation of project objectives and plans, as well as in their implementation. The Kalingalinga upgrading project in Zambia enabled 4400 low-income squatter households to obtain improved houses, services and security of tenure (Jere, 1991b). This was achieved with the active involvement of the community at all stages of the development. A more efficient land use reduced overcrowding and provided 1800 overspill plots for new houses. The Zambian objective of maximizing economic development in shelter projects was achieved by creating small companies and other commercial activities locally to produce concrete blocks and other essential elements.

One problem affecting the Kalingalinga project was that many early community leaders were members of the political party that was in opposition before Zambia became a one party State. Their access to people in positions of influence was thus limited. Oestereich (1980:31) considered the next generation of leaders to be of an inferior stature. Eventually, however, a residents’ committee was formed which was able to generate official support for the upgrading of the settlement. This suggests that the role of effective local organizations in initiating, as well as planning, implementing and maintaining a project, needs to be fully acknowledged.