Limitations of the Owner-Driven Model in Post-Disaster Housing Reconstruction in Urban Settlements

Limitations of the Owner-Driven Model in Post-Disaster Housing Reconstruction in Urban Settlements Mojgan TAHERI Tafti 1 1 PhD Candidate, Dep. of planning., The University of Melbourne (3010, Victoria, Australia) E-mail: m.taheritafti@student.unimelb.edu.au This paper explores the potential shortcomings in the owner-driven model of housing reconstruction in urban settlements affected by disasters. The owner-driven model has been widely adopted in recent post-disaster housing reconstruction projects. While the advantages of this model over other models such as donor-driven or state-driven models are discussed in hazard literature, discussions over the potential shortcomings of this model are scant. Looking at the experiences of post-earthquake reconstruction in Bhuj in India and Bam in Iran, the paper sheds light on the inadequacies of the owner-driven model in housing recovery and their unintended consequences for households and the urban fabric. The paper suggests three main arguments based on the ways the owner-driven model has been conceived, formulated, and practiced in two post-earthquake housing recovery projects. The first argument is that the owner-driven model should be conceived as one component of a holistic housing recovery programme. The second argument revolves around the shortcomings of the standardized arrangements of the owner-driven model. The last argument identifies practical problems that prevailed in the two projects. While acknowledging the advantages of the owner-driven model, the paper concludes by advocating a more holistic, nuanced, and inclusive housing recovery programme in disaster-stricken urban settlements for responding the complexities of housing recovery in urban setting. Key Words : Urban disasters, owner-driven model, housing reconstruction Provision of permanent housing following disasters is often described as absorbing a considerable portion of post-disaster lending in the developing world 1) Planning for this component of post-disaster activities has decade, the advantages of the owner-driven model of housing reconstruction - particularly vis-à-vis the state-driven and the donor-driven models - has been widely acknowledged, and the model is often described in 2). However, the ways in which this model are often conceived, formulated, and implemented have implications that have been largely unexplored in the literature. The purpose of this paper, therefore, is to develop a better understanding about the limitations of this model by investigating its application in the aftermath of disasters in the developing countries. Owner-driven housing reconstruction has its roots in the pioneering work of John F.C. Turner. Following - allowed rebuilding of the damaged dwellings by their residents. Despite being acknowledged as an influential initiative in promoting mechanisms for low-income housing provision in developing countries 3), the idea of aided self-help housing provision was rarely put into practice in subsequent post-disaster reconstruction activities. During the following two decades, post-disaster housing reconstruction projects were predominantly contractor-based, managed by state or donor agencies 4). The new form of owner-driven housing reconstruction was promoted in some villages affected by the Indian Latur earthquake in 1993. The owner-driven model of housing reconstruction in its current form has become prevalent after the 2001 Gujarat earthquake 5). The main idea in this model is to enable individuals to undertake the reconstruction of their houses. Diverse mechanisms such as conditional financial assistance - usually on an instalment basis - as well as technical support and supervision are integrated into this model to ensure safe building practices. During the last decade, the model has become the default strategy in post-disaster housing recovery as a result - 347 -

of being advocated by key lending agencies including the World Bank and prominent international institutions such as the UN-Habitat e.g., 6). Funding the majority of reconstruction projects in disaster-affected developing uence post- has advised this approach as the most successful housing assistance strategy 2:93). Therefore, many disaster-stricken countries including India (after the 2001 Gujarat earthquake), Iran (the 2003 Bam earthquake), Pakistan (the 2005 earthquake), and Thailand (the 2004 Tsunami) have adopted the owner-driven model as their overall strategy for housing recovery of the affected communities. The owner-driven model has been promoted as it involves people in decisions that are directly related to them. By employing this Such involvement offers intangible benefits such as instilling a sense of place attachment and empowering people. Numerous empirical studies e.g., 7, 8-10) report that adoption of this model ensues a higher level of perceived satisfaction and higher occupancy rate compared to other models of post-disaster housing reconstruction including the donor-driven model. Some researchers also indicate that this approach is faster and more economical as owners supplement the assistance with their own assets 10, 11). A review of literature, however, reveals that apart from concerns over the safety of the reconstructed buildings raised by few studies e.g., 12), little is known about the potential shortcomings of this model. As this model is expected to be adopted in the aftermath of future disasters, it is critical to develop our understanding about its potential drawbacks. On this basis, this paper scrutinizes the owner-driven housing reconstruction as it has been formulated and implemented in cities of Bhuj in Gujarat and Bam in Iran, affected by earthquakes in 2001 and 2003 respectively. Investigating the long-term impacts of housing recovery programmes on esearch draws upon field observations and in-depth interviews with residents of both cities, examination of policy documents and proceedings, and interviews with decision makers and some of the key informants. The paper suggests three main arguments about the unforeseen and contradictory outcomes of the adoption of owner-driven model. Each argument relies on common issues that emerged in this comparative research. The first argument is that the owner-driven model should not be conceived as the overall strategy for housing recovery aftermath of a disaster, but as one component of a holistic housing recovery programme. The second argument revolves around the shortcomings of the standardized arrangements of the owner-driven model. The last argument identifies practical problems prevailed in both projects. Before presenting these arguments, the next section provides a brief overview of the two events and their subsequent recovery activities. 1. HOUSING RECOVERY PROGRAMMES IN BHUJ AND BAM On 26 January 2001, an earthquake measuring 7.7 on the Richter scale occurred in Gujarat. In the city of Bhuj, approximately 7,000 people, mostly inhabitants of the historic part of the city, lost their lives. Nearly 50% of this part of the city was destroyed 13), with 11,036 houses completely and 27,617 houses partially collapsed. Two years later, on 26 December, 2003 a major earthquake registering 6.6 on the Richter scale devastated the city of Bam in south-east Iran. With a population of around 120,000, the death toll was 26,271. The ca- city and surrounding villages homeless 14). In the wake of the earthquake, 80% of all buildings in the city were completely destroyed. Both governments, central government in Iran and state government of Gujarat, established apex-level, central agencies to respond the disaster. These task forces adopted an owner-driven model for housing recovery in the two affected cities of Bam and Bhuj. In both cases, homeowners were provided with t financial and technical assistance, aimed at enabling these affected groups to rebuild their own houses. The financial assistance paid on an instalment basis to ensure the conformance of construction process with seismic building codes. The assistance of up to US$ 3,225 for building a 45-square-meter house in Bhuj and US$ 19,375 for building an 80-square-meter house in Bam was offered to homeowners. In the case of Bam, the ners were eligible for receiving assistance for each and every damaged house they had in their possession. In Bhuj, according the housing recovery programme, owners were given housing assistance for only one damaged residential unit in their own occupancy at the time of the earthquake. Moreover, as the reconstruction of Bhuj city was a combination of in-situ reconstruction and relocation (in contrast to the in-situ reconstruction of Bam), those landowners who opted to move to relocation sites could receive a plot of land against their old ones. Investigating the long-term recovery process of affected households eight and ten years after these earth- - 348 -

quakes provides some insights about the deficiencies in the housing recovery programmes, their formulation and delivery. In the following sections, we present three main common issues, identified in these two post-disaster housing programmes. (1) Owner-driven housing reconstruction and the role of tenure in defining access to the programme One of the concerns over the emphasis on the owner-driven model is the undermining of the housing needs of non-owner disaster-affected groups including tenants, sharers and squatters. Examination of the recovery programmes in both cases suggests that the owner-driven model was conceived as a holistic framework for post-earthquake housing recovery. Therefore, the housing recovery of the non-owner groups was treated as a marginal part of the programme. On this basis, in both cities the ownership of land or property was the determinant for 18% of households in the city, were denied housing assistance. Later, this group was offered housing assistance, provided that they could buy land within the city. In Bhuj, on the other hand, the recovery programme included some provisions for housing recovery of renters, but ignored sharers and squatters. The policy for the collapsed rental units was a payment to the owner for rebuilding the rental house, provided that she/he guaranteed the restitution of the pre- right of occupation. Based on the interviews conducted with tenants and landlords, it can be concluded that this policy did not reach its goal, mainly due to the very low level of rent price prior to the earthquake. Later, changes in programme due to this failure allowed tenants to buy a plot in relocation sites, if they could afford to do so. Moreover, the local authority invited NGOs to take a part in the process and build new housing projects for these groups. These new provisions did not provide for households with other tenure agreements which were practicing in the city before the earthquake. In addition, there were eighteen slum areas in Bhuj, some of which were badly damaged by the earthquake. These slums have not been included in the new urban development plan and their low-income residents were ignored during the recovery process. Even NGOs paid scant attention to the housing recovery of these groups following the earthquake 15). These housing reconstruction policies have had long-term ramifications for non-owner groups. These policies left tenants to the mercy of landlords and made their housing recovery dependent on the decisions of others. These groups had to wait for long periods in temporary shelters until (and if) the rental market catches up with demand. In Bhuj, these groups have been living in a site far from the city with poor access to livelihood and social amenities. A local NGO built affordable housing for some of the residents of this site and others are living in dilapidated dwellings built for temporary housing purpose. Decision-makers in both recovery programmes employed the same discourse for ignoring the housing needs We believed that they did not lose their properties and therefore they were no (Interview BM-O1-05/02/2011). This questions the very objective of allocating housing assistance which in both cases has been mainly based on the disaster-incurred material asset losses, but were represented in programmes under the guise of creating an enabling environment. Conceiving the owner-driven model as the holistic framework for housing recovery in both cases was arguably in line with a more desirable outcome of the recovery process for politicians: building houses as a tangible outcome of reconstruction 1), instead of enabling households whose housing condition has been affected by the disaster to achieve their housing recovery. The key to understand and distinguish between these Bam and Bhuj, the number of (re)constructed houses is more than the damaged ones. Many people including -earthquake housing condition, in terms of security of tenure, affordability, location, and quality of the dwelling. Housing recovery of non- dynamic nature of vulnerability, because its driving forces change over time, many earthquake-affected households in Bam could have been considered as vulnerable groups for a period of time. Among them earthquake-induced damage aggravated the precarious living condition of low-income, female-headed households. Realizing the housing needs of these households, some international institutions, including the UNDP, took part in housing reconstruction for them. Ironically, the governmental organization responsible for the reconstruction process asked these institutions to contribute to the housing recovery of these households provided that they possessed land within the city 16). This decision, in the context of Iran where women do not have equal inheritance rights, deprived many female-headed families from the contribution of these institu- - 349 -

tions. For those female-headed households who did not have a plot of land, an Iranian NGO provided a limited number of free accommodation in two complexes in the city. Holding no tenure rights, these families could stay in these, no-male-entry residences, provided the head of the family, many of them young widows, never get married again. The condition was even worse for those women who had not legally divorced, due to its cultural stigma, and based on the fieldtrip interviews constituted a considerable portion of female-headed households. This group has not been formally recognized as female-headed families and hence were deprived of receiving such helps. Around 800 female-headed families have been left with no housing recovery option. In Bhuj, despite enjoying the equal legal right of inheritance, women who were the head of their families faced difficulties in receiving housing assistance, especially because majority of them were traditionally living as sharer with their in-law families before the earthquake. Some of these interviewees were denied a share in the new reconstructed houses by their in-law families. These female-headed families with no land ownership and no family support were not offered any housing option. Some local NGOs such as Abhyan and religious-based donor organizations provided housing for a group of these households, albeit at the urban fringe where the land was available. This led to a further economic and social exclusion of these households. The issue of housing recovery of non-owner groups has not been properly addressed by research and practice in the field of post-disaster recovery as the majority of recent disasters with natural triggers in developing countries have taken place in rural settlements or small towns. After the Haitian earthquake in 2010 and due to the large number of households without formal property ownership, this issue has emerged as a concern and drew attentions. While cities accommodate a variety of households with heterogeneous tenure arrangements, the owner-driven model of housing reconstruction addresses the housing needs of just one group among the others. It can be argued that conceiving the owner-driven model as the holistic framework for housing recovery, as it has been conceptualized in Bam and Bhuj, provides benefit to a specific group and deprives others based on their tenure condition. (2) Different capability of beneficiaries Not only are cities agglomerations of variety of housing with different tenure arrangements, but also they accommodate heterogeneous communities. This paper argues that the owner-driven model, in its current shape, sets up a standardized method of payment and technical assistance which does not respond to the different capabilities of homeowner households to pursue their housing recovery. The different capability of households may stems from different cultural, social, or economic factors or may be related to the disaster-induced impacts. Looking at the long-term recovery trajectory of different groups of people depicts how different households demonstrated different capabilities in achieving their recovery. Drawing upon the interviews in the two cities, this section is to show, on a few examples, some implications of adopting such a unified -term recovery. The first example is related to land/property-owner female-headed households in Bam whose number has sharply increased due to some social issues that emerged following the disaster (some officials reported a ten-folded increase in the number of these families). The common problem these households faced was related to the fact that based on the owner-driven model the management of all administrative and construction activities is to be undertaken by owners. In the social context of this small city, many women found it difficult to interfere in this male-dominated field. Consequently, many of them had to trust and relied on the help of their male relatives or acquaintances to oversee the construction process by contractors. Many of the interviewed women reported problems with contractors, the majority of whom were non-local, and replacing them during the construction period. Another group of households with different capability were those who started the reconstruction process later than others or could not keep the same pace of rebuilding as others. An example of the first group, late-starters, is traumatized families. In Bam, where at least a quarter of the residents of the city lost their lives, traumatized households constituted a large number of population. These households or those preoccupied with the treatment-related issues of their injured family members started their housing reconstruction later than others. Given the double-digit inflation in the country, many of Bammi late-starters could not finish their housing construction as the assistance was not adequate anymore for finishing the construction of their houses. According to the Statistical Centre of Iran 17), 51 percent of households were still living in temporary housing three years after the earthquake. In the case of Gujarat earthquake, those interviewees who started late, due to problems including injuries of the family members faced with increasing construction material price and had difficulties in finishing their houses within the required time period. Moreover, some very low-income groups consumed their first in- - 350 -

stalment for their daily needs or livelihood purposes and consequently could not seek or receive their next instalments. Another problem reported by the very low-income owners was the difficulties in going through the bureaucratic process of housing reconstruction as it demanded frequent trips to the administration offices while they had to work everyday to feed their households. In general, 19,500 beneficiaries did not seek or receive their next instalments to complete their homes in Gujarat 5). The capability of households to undertake their housing reconstruction, as shown in the aforementioned examples, might be impeded by factors such as economic conditions and holding marginal assets, age, injuries or death of the head of the household, and inheritance problems. Even issues such as the level of literacy were important in accessing information and housing assistance. In both cases of Bam and Bhuj middle income educated households could manipulate the regulations and received more than housing assistance. These examples illustrate that in contrast to the equal process of housing reconstruction, assumed by the owner-driven model in housing recovery programmes of Bam and Bhuj, different homeowner households have demonstrated different capabilities in rebuilding their damaged houses. For many households, the housing reconstruction processes lasted longer than others and given the availability of financial or technical assistance only during the official period of reconstruction, they were not able to receive assistance as others did. This raises questions about the standardized strategies embedded in formulation of the owner-driven model. (3) Quality of the built environment: Practical shortcomings While the two former arguments criticize the way the owner-driven model is often conceived and formulated, the third argument is related to issues which are not inherent in the model. This section looks at common problems related to the quality of the built environment, observed in the two case studies. It is argued that these issues need to be appropriately addressed in policy and implementation arrangements in the owner-driven housing reconstruction. The first problem, also raised by other studies 12, 18), is related to physical vulnerability of the reconstructed buildings. In both cities of Bam and Bhuj, the reconstruction process was predominantly executed by contractors or skilled labours hired by homeowners. Supervising the process of construction, engineers appointed by the government had the responsibility of controlling the conformity of construction with the seismic code at each stage and issuing permit for disbursement of the next instalment of the grant to owners. Although several programmes for training skilled labours and contractors were carried out in Gujarat, interviewed academics expressed their concerns over the physical fragility of the rebuilt residential units. In Bam, where the earthquake destroyed 80 percent of buildings including many newly built structures, people became very conscious about adopting safety measures in construction of their houses. Over time and partly because of a very slow reconstruction process, this concern lost its priority to the economic hardship. Consequently, some low-income groups returned to their pre-earthquake unsafe construction practices. Furthermore, with the official termination of the reconstruction before a considerable part of construction activities finished and given inadequate technical knowledge transfer to builders, enduring safe construction practices became less probable 19). These examples illustrate that inadequacies in safe construction practices are partly resulted from an insufficient technical knowledge transfer to builders. The training components of the owner-driven housing programmes are normally evaluated through the quantitative data about the number of trainees, instead of looking at how safety practices are communicated with skilled labours in the ways tailored to their literacy level and construction knowledge. The second problem is related to the difficulties in reconstruction of multi-unit buildings with mixed use and mixed tenure. In these two small cities, the issue of multiple unit buildings with mixed tenure did not appear to be a problem, because the new building regulation limited the height of buildings to double story. However, the reconstruction of buildings with mixed use was treated differently. In Bhuj, the housing assistance was exclusively allocated for residential purposes. This led to the disappearance of a prevalent building typology, i.e. buildings with shops on their ground floor from the old neighbourhoods of the city. This was not the case in Bam where a more flexible housing recovery programme allowed construction of such buildings. Adequate flexibility which accommodates the requirements of diverse arrangements is often a neglected aspect of the owner-driven housing reconstruction which impinges upon the applicability of this model in urban settlements. The problem is not specific to the owner-driven model but is accentuated partly as this model implies more conceived autonomy of owners in building their homes. In both cities, the historic fabric of the city, where housing - 351 -

comprised considerable portion of the cultural landscape, was dramatically transformed through the reconstruction process. Housing reconstruction offered a unique opportunity to people to build their dream houses, and in the absence of legal guidelines and prescriptions, the result was a hodgepodge of diverse building forms and materials, and hence loss of distinctive characteristics of these historic cities. Beyond addressing the short-term housing demands of the affected communities, it is essential that the owner-driven model incorporates considerations about the quality of the built environment. Integrating guidelines corresponding different aspects of the physical setting including physical resilience, function, and place identity into housing recovery programme can contribute to the improvement of the quality of the built environment. 2. CONCLUSION This paper has scrutinized the owner-driven model of housing reconstruction based on the ways it has been conceived, formulated, and practiced in two post-earthquake housing recovery projects. It has been argued that the owner-driven housing reconstruction responds to the housing needs of just one group of affected communities and therefore, cannot be conceptualized as a holistic framework for housing recovery. In fact, conceiving the owner-driven model as the main framework for housing recovery implies the exclusion of non-owner groups from housing recovery options. This means that an indispensible part of the housing recovery process is left, without a clear plan, to the non-public sector, i.e., housing market or NGOs. NGOs, as depicted, might not be able to or willing to appropriately cover all the left groups. Also, literature shows that housing market is not often able to provide affordable housing in the aftermath of disasters 20). Moreover, the formulation of the owner-driven model implies an incorrect assumption about equal capability of households to manage their housing reconstruction. Different factors including ex-ante vulnerabilities, disaster impacts, or contextual issues might curtail the capability of households to achieve their recovery. Even more problematic, the most vulnerable groups received the least assistance, compared to the elites who often expand their capabilities. These arguments indicate a structural incongruity in housing recovery programmes which define their goal as creating an enabling environment and at the same time deprive those who need the most contribution to be able to achieve their recovery. Finally, the paper has underpinned the importance of a better formulation of the owner-driven model by relying on a more nuanced understanding about the complexities of housing recovery in urban setting. While knowledge transfer about best practices in post-disaster housing recovery is essential for informing decision-makers in cases of future disasters, an appropriate communication of the limitations in the previous experiences can result in a better preparation and a more efficient and equitable use of the scarce financial resources. In this sense, the owner-driven housing reconstruction model requires a revisiting in conceptualization, its formulation and practical arrangements with a view to create an inclusive and enabling environment for post-disaster housing recovery. ACKNOWLEDGMENT: I wish to thank Prof. Richard Tomlison, the head of planning department at the University of Melbourne, for his comments and contribution. REFERENCES 1) Freeman, P.K.: Allocation of post-disaster reconstruction financing to housing, Building Research and Information, Vol. 32 (5), pp. 427-437, 2004. 2) Jha, A.K., J.D. Barenstein, P.M. Phelps, D. Pittet, and S. Sena. Safer Homes, Stronger Communities: A Handbook for Reconstructing after Natural Disasters.. Washington D.C. : The World Bank, 2010. 3) Keivani, R. and E. Werna. Refocusing the housing debate in developing countries from a pluralist perspective, Habitat International, Vol. 25 (2), pp. 191-208, 2001. 4) elter?, Environmental Hazards, Vol. 10 (3-4), pp. 193-212, 2011. 5) World Bank. Implementation completion and results report for Gujarat emergency earthquake reconstruction project, 2009, Available at: http://www-wds.worldbank.org/external/default/wdscontentserver/wdsp/ib/2009/06/08/000333038_20090608014529/rende red/pdf/icr6380p074018101official0use0only1.pdf - 352 -

6) UN-Habitat. Building back better in Pakistan, 2007, Available at: http://www.unhabitat.org/downloads/docs/4627_75789_gc%2021%20financing%20field%20report%20pakistan.pdf 7) Karunasena, G.: Post-disaster housing reconstruction: Comparative study of donor vs owner-driven approaches, International Journal of Disaster Resilience in the Built Environment, Vol. 1 (2), pp. 173-191, 2010. 8) Baresnstein, J.D. and S. Iyengar. India: From a culture of housing to a phylosophy of reconstruction. in Building Back Better: Delivering people-centred housing reconstruction at scale. M. Lyons and T. Schilderman, Editors. Warwickshire: Practical Action. 2010. 9) Barenstein, J.D.: Who governs reconsrtuction? Changes and continuity in policies, practices and outcomes. in Rebuilding after disasters: From emergency to sustainability. G. Lizarralde, C. Johnson, and C. Davidson, Editors. New York: Spon Press. pp. 149-176, 2010. 10) Lyons, M.: Building back better: the large-scale impact of small-scale approaches to reconstruction, World Development, Vol. 37 (2), pp. 358-398, 2009. 11) Ingirige, B., R. Haigh, C. Malalgoda, and R. Palliyaguru. Exploring good practice knowledge transfer related to post tsunami housing re-construction in Sri Lanka, Journal of construction in developing countries, Vol. 13 (2), pp. 21-42, 2008. 12) Barakat, S. Housing reconstruction after conflict and disaster, 2003, Available at: http://www.odihpn.org/documents/networkpaper043.pdf 13) Balachandran, B.R. The Reconstruction of Bhuj Case Study: Integration of Disaster Mitigation into Planning and Financing Urban Infrastructure after an Earthquake, 2005, Available at: http://info.worldbank.org/etools/docs/library/114715/istanbul03/docs/istanbul03/07bala3-n%5b1%5d.pdf 14) World Bank. Technical annex for a proposed loan for Bam earthquake emergency reconstruction project, 2004, Available at: http://www-wds.worldbank.org/servlet/wdscontentserver/wdsp/ib/2004/10/12/000160016_20041012100319/rendered/pdf/ T7637.pdf 15) Mukherji, A.: Negotiating housing recovery: why some communities recovered while others struggled to rebuild in post-earthquake urban Kutch, India. Department of Architecture, University of California at Berkeley, Doctor of Philosophy, 2008. 16) UNDP. Final review: sustainable housing reconstruction programme in Bam through community mobilization and participation. 2008 December 2011]; Available from: erc.undp.org/evaluationadmin/downloaddocument.html?docid=1838. 17) The Statistical Centre of Iran. National Population and Housing Census in 2006, 2006. 18) Todd, D. and h. Todd. Natural Disaster Response Lessons from Evaluations of the World Bank and Others (Evaluation Brief 16), 2011, Available at: http://lnweb90.worldbank.org/oed/oeddoclib.nsf/docunidviewforjavasearch/6e05abfae2ed2cf58525794400774eae/$fil e/eval_brief_nat_disaster_response.pdf 19) Gharaati-Kopaei, M.: Knowledge transfer in post-disaster reconstruction The Problem of Post-post-disaster Reconstruction. School of Architecture, McGill University, Montreal. Doctor of Philosophy in Architecture, 2009. 20) Comerio, M.C.: Disaster hits home : new policy for urban housing recovery. Berkeley : University of California Press, 1998. - 353 -