A.12 Indonesia

A.12 Indonesia - Sumatra - 2009 - Overview Case study: Summary On 30th September 2009 a series of earthquakes struck West Sumatra, not far from the provincial capital of Padang. 13 out of the 19 districts in West Sumatra province were affected. Between earthquakes and landslides nearly 250,000 houses were destroyed or heavily damaged. The Government of Indonesia responded rapidly, with the assistance of the national and international humanitarian community. Whilst non-government agencies focused on emergency shelter, distributing an average of 2 tarpaulins per family, the government focused on rebuilding provincial government capacity, search and rescue and emergency relief. The emergency phase was declared over within 8 weeks. The Government of Indonesia committed to providing affected families with a community based economic stimulus package for permanent housing reconstruction, leaving the provision of emergency and transitional shelter to the humanitarian community, many of whom also focused on Disaster Risk Reduction based construction skills training.

Earthquake damage to a former 3 story government Emergency distributions of two tarpaulines per household building in Padang. were made by reponding organisations. Photo: Dave Hodgkin Photo: Dave Hodgkin Before the earthquake “Permanent” (masonry) houses, for Whilst rural housing was West Sumatra is located at the “Semi Permanent” (part masonry), commonly self-built, urban housing convergence point of four tectonic and for “Non-permanent” (timber was more commonly commercially plates and is highly prone to earth- or bamboo) houses, however, constructed with a mixture of rental quakes. A recent earthquake in limited certification (15%) along and non-rental housing. 2007 had damaged or destroyed with poor compliance and enforce- over 43,000 houses. ment had resulted in a low quality After the earthquake of general construction. The disaster caused an estimated As a result of numerous disasters, 2.3 billion USD damage to infra- both the provincial and national In West Sumatra, most homes structure and housing. Over 30% of government had significant experi- were privately owned particularly housing stock in the affected areas ence. The recently formed National in rural areas, with most inherited was destroyed, making shelter a Disaster Management Agency through matrilineal ownership priority. deployed a Technical Advisory Team systems. They were constructed in- to assist in the immediate response crementally often with the support Initially rural and semi-urban and assist in the formation of it’s of remittances from male family areas were prioritised. In these provincial equivalent. members working in the “Padang” areas, many families were living restaurants across Indonesia and in inadequate, unsafe makeshift Although established national Malaysia that the area is famous shelters, under tarpaulins within building codes, including seismic for. their plots of land, or staying in resistant construction guidelines for other people’s homes or gardens.

38 Natural disaster Shelter Projects 2010 A.12

200 USD “transitional shelter” of dissapointing quality built by an international organisation. Rural self help shelter built by earthquake affected family. Photo: Dave Hodgkin Photo: Dave Hodgkin

Concerns over the approaching An international coordination included advice on permanent re- rainy season added to the sense of team arrived within four days of the construction. 63,000 transitional urgency. earthquake to assist the Indonesian shelter packages were provided government in coordinating over with a cost varying from 200 USD Previous experiences within 200 national and international re- to 500 USD per household. Indonesia indicated that public sponding agencies. outreach programmes on earth- Later assessment highlighted a quake resistant construction were The initial shelter strategy was lack of assistance to urban areas, important to ensure safe recon- agreed eight days after the earth- with a range of agencies then struction. quake. The strategy focused on the running clean operations in these distribution of tarpaulins and tents areas. Delays in material supplies Response capacity for the emergency phase, whilst and limited capacity saw transi- The first few weeks saw intense identifying the need for transitional tional shelter projects continuing international media attention and shelter and disaster risk reduction for over 9 months after the earth- an ensuing influx of internation- activities in the recovery phase. quake, overlapping significantly al and national funds. Over 200 with the arrival of permanent re- agencies both national and inter- Despite an overwhelming initial construction funds. national responded rapidly. Many response to the disaster there had prior experience in Yogya- remained a shortfall in funding, Government response karta earthquake and/or remnant particularly in shelter and liveli- The government of Indonesia capacity in nearby Aceh and Nias hoods. A total of 170,000 families provided grants of approximately Island from post tsunami and earth- were supplied with emergency 1,500 USD for heavily damaged quake projects. shelter within the first two and a houses, 1,000 USD for medium half months. damage (from the State Budget) However many organisations, and 100 USD for lightly damaged including the newly formed provin- Recovery shelter houses. cial disaster management agencies The Early recovery phase saw quickly found themselves over- the government focusing on the Two years after the earthquake, stretched. Many were still respond- development of permanent shelter not all funds had been released, ing to an equivalent scale earth- assistance programs, whilst non though much of the community quake in West Java less than one government agencies focused on had self funded reconstruction. month before. Many of the interna- transitional shelter needs through The 2010 earthquake in the West tional agencies soon had to relocate a range of shelter packages. Most Sumatra district of Mentawai capacity to the Haiti earthquake. assistance was in the form of cash Islands, further stretched and grants or material supply, to small expanded provincial response Emergency response community groups in line with gov- capacity. Extensive collapse of commercial ernment proposed methodology The initial government decision and government building in Padang for community built reconstruction. to focus only on permanent shelter resulted in an initial focus on search was later reviewed in light of out- and rescue with 21 teams of various Transitional shelters commonly standing transitional shelter needs, sizes being deployed. had timber frames. They were mainly clad with corrugated iron or with funds then allocated to transi- The Indonesian Government tarpaulins for roofs and tarpaulins, tional shelter in West Sumatra, and announced an end to the search plywood or timber for walls. Shelter again in Mentawai Island and other and rescue phase within weeks, packages commonly included a later responses. and allocated an initial 10 million technical advice component. Many USD to emergency relief.

www.ShelterCaseStudies.org 39 A.13 Natural disaster

A.13 Indonesia - Sumatra - 2009 - Earthquake Case study: See “A.12 - Indonesia - Sumatra - 2009 - Overview” p.38, for background

Country: Project timeline Indonesia, Sumatra, Padang Disaster: Earthquake Disaster date: September 30th 2009 5 weeks - –– Final report No. of houses damaged: 115,000 destroyed houses 135,000 damaged houses

This was a market assessment into brick production and so did not directly lead to the construction of shelters 3 weeks - –– Surveys were conducted over 3 days

Padang

September 30th Indonesia 2009- –– Earthquake

Project description This project surveyed brick production and anticipated supply and demand. It was conducted one month after the earthquake. The survey was conducted as a trial of the EMMA (Emergency Market Mapping and Analysis) methodology. The survey findings were used to inform the adopted strategy of using cash to support the construction of shelters that used both timber and bricks.

Strengths and weaknesses 99 The assessment was conducted with team 88 Surveys looked at the use of bricks but not the use members from nine different organisations. This of timber to make the bricks. process increased buy-in to the findings of the 88 The survey did not address issues of the living and assessment report, and helped to form consensus on working conditions for those in the brick kilns. the issues surrounding markets in the response. 88 The survey used human resources, meeting time 99 The bricks survey findings were used to advocate and vehicles that could otherwise have been used in for a cash based response, and for a move away from implementing the response. solid masonry buildings which potentially carried a -- It is difficult to accurately measure the impacts greater risk of causing injury in an earthquake. of this survey. Whilst it used human resources and 99 The survey came at an opportune moment after absorbed time during an emergency response, there is the earthquake. The timing of the survey needed some evidence that it helped to inform the strategies to be long enough after the earthquake that team and programmes adopted. members could be identified, access was possible and -- There are many markets that could have been those working at brick kilns could easily be found. surveyed. Bricks were chosen following experiences in Had it been any later it would not have been able to Aceh (2004) and Yogyakarta (2005). inform the strategy.

40 Natural disaster Shelter Projects 2010 A.13

Poorly built brick-masonry buildings were a significant cause of the damage to housing. Photo: Unknown Background Mapping and Analysis). EMMA is 3. Air drying: The bricks are laid to See “A.12 - Indonesia - Sumatra a tool designed to analyse markets dry in the sun for 5 days. Bricks - 2009 - Overview” p.38. following a disaster. EMMA uses are then stacked and air dried for background research, interviews, 30-60 days, depending upon the After the earthquake and graphic representations of weather. The earthquake in September market systems to help inform 4. Kiln drying: The dry bricks 2009 destroyed or damaged over humanitarian response options. are loosely stacked in open air 200,000 houses in West Sumatra. EMMA defines a market system kilns without chimneys. These Poorly built brick based masonry as “a web of people, businesses, kilns are rectangular or circular caused many of these buildings to structures and rules that take part in shapes. Mud is plastered around collapse. producing, trading and consuming the outside of the brick kilns to The Indonesian Building Code a product or service.” trap the heat from the fire, with space for smoke to escape and specifies that a "Permanent House" For more information on the oxygen to enter. The average means masonry, “Semi Permanent” EMMA methodology, download height of a brick kiln is 2m tall. means masonry sub walls and the EMMA Toolkit from: http://em- Bricks are typically kiln dried for timber above, whilst “non-perma- ma-toolkit.org nent” means timber or bamboo. 10 – 14 days. Brick making in Sumatra 5. Distribution: Manufacturers Experience from previous Brick making involves five steps sell their bricks directly to disasters in Aceh (2005) and Yo- and is labour intensive. masons, home owners, brick gyakarta (2006) showed that the distributors, and / or building demand for bricks for housing re- 1. Mixing: Clay, sand and water supply stores. Transportation construction quickly outstrips the are mixed together in open charges are typically 30 - 60% available supply. This often led to an pits by foot, shovels or water of the total brick price. increase in the price of bricks, and buffalos. Larger manufacturers / or periodic supply shortages that use mechanical mixers. Damage to supply delay reconstruction progress. 2. Shaping: The mix is The survey suggested that over compressed in wooden 50 million bricks were damaged in What is EMMA? frames. On average, a skilled the earthquake. This research was conducted labourer can produce 1,000– The majority of the supply was to trial EMMA (Emergency Market 1,500 bricks per day. through small scale suppliers. There

Much of the capital for small scale manufacturers Many of the bricks were made by hand. was the bricks in their kilns. Photo: Unknown Photo: Unknown

www.ShelterCaseStudies.org 41 A.13 Natural disaster were 1,800 small scale brick manufacturers, who produce an average of 15,000 bricks per month. These were the most severely affected of all brick manufacturers. The financial capital of these producers was often tied up in the number of bricks they had in their kiln, making it difficult to restart manufacture. Medium scale manufacturers (45,000 bricks per month) also suffered production losses due to The survey used teams from nine different organisations working together. the earthquake, but their stronger Photo: Unknown financial position meant that they Brick demand brick production capacity were better able to resume produc- 60% of all households inter- would likely lead to higher tion. It was estimated that it would viewed indicated that they would brick prices and delays in rural take 6–8 weeks for these manufac- re-use as many bricks as possible. A housing reconstruction. Large turers to bring new bricks to the rough estimate suggested that many brick manufacturers were likely market. households would be able to salvage to reach previous production Most of the larger scale brick 800-1200 bricks from the rubble. As capacity within two months. manufacturers were located up to an average size brick masonry house Resulting transportation cost 90km North East of Padang. Some of 10m X 12m used approximately increases could lead to a price large brick manufacturers reported 10,000 bricks, approximately 10% increase of between 100% and losing 35% of their brick produc- of this demand would come from 150% per brick. tion in the earthquake, while others recycled materials. • Small - scale brick manufacturers would be slow to resume pre- did not report significant losses. Although 67% of all households earthquake production levels interviewed said they lived in a brick Brick prices and financing without financial assistance or masonry house before the earth- Pre-earthquake brick prices favourable credit terms. Their quake, 54% of the brick masonry ranged considerably according to ability to resume production households indicated they would quality, seasonality and transport was restricted due to capital prefer to rebuild timber and brick costs. shortages, or favourable credit houses. Safety concerns were most arrangements. Following the earthquake brick often cited as the reason for this • The demand for timber and prices from suppliers for mid range preference, followed by cost consid- bricks was high, and was likely quality bricks increased by between erations. 25% and 50%. The assessment to increase. Over 60% of found that these prices were likely There was some concern raised earthquake affected households to continue to rise to 150% of their that recycled bricks would not interviewed in this survey pre-earthquake cost. perform so well as new bricks indicated that they planned because as cement mortar cannot to rebuild (or would prefer) Two years after the survey, brick bind to them so well. timber frame houses with brick prices in Pedang were between masonry infill walls over full 60% and 100% higher. Gender issues masonry construction. Concerns Women made up 40 - 60% of the Both small and medium scale over seismic safety, speed of labour force of small and medium brick manufacturers used informal construction, and lower costs scale brick manufacturers. They credit and selling arrangements were the main reasons for this were typically paid on a piecework with their customers and distribu- change in preference. basis for each brick they made. Male tors. Local supply stores typically brick labourers are likely to receive a Impacts of the survey paid small-scale manufacturers for daily wage for their work. Because the survey was bricks once they had sold them. conducted by teams from many or- As current brick production All brick manufacturers, but es- ganisations, it helped to get support for many small-scale producers is pecially small and medium scale for the findings. Although not all of affected, the ability of brick making producers, had limited storage and the recommendations were imple- women to earn wages was tempo- warehousing space. These space mented, it did help organisations rarily disrupted. limitations forced manufacturers to and coordination teams to form move their bricks to market quickly. Possible scenarios an advocacy position away from It encouraged large suppliers and The analysis suggested that: building full masonry structures, distributors to increase their prices instead promoting semi-timbered to meet speculative market demand. • Earthquake damage to regional structures with support provided in cash. 42 Natural disaster Shelter Projects 2010 A.13 ! ! n on

ti s ey 1 2 ey RURAL URBAN jor disrup mbol K HOUSEHOLDS HOUSEHOLDS rget group Sy Critical issue Ma Partial disruptio Colour k Ta Other type Other type

! NT

S S AL O ME CT NG CT OR & IN FORM CONTRA GOVERN CONTRA

CE CASH RI INCO P TRENDS

S AL

NT LAND ! RE

ROADS & BRIDGE E E STOR ! S STOR LABOUR

LARGE SUPPLY G NCE LOCAL SUPPLY DI NG RE IL FE BU RE

P REHOUSIN AL AND a CI WA T SO LATIONSHIPS OR R RE Sumatr DELIVERY est L TRANSP TA –W

! EN g DISTRIBU TO S adan RONM FACTORS P VI LOAN EN ! CREDIT AND Y

K nth s IC ie es

s ck Marker – mo BUFFALO BR il 0/ QUALIT th ce : kag am vi 00 on nt 0, h h /m 00 f h The Bri p me p –4 p r eir lin r 00

20 ! S r : 80 ,0 s, 0 00 port ser – in iron AL 000 le ,0 600 600 – 60 -1 600 ds ha 10 5 40 000 >1 25, 0– 0– BRICK PRODUCER tem Map: < = 0 – = TERI = = 50 50 = 1 ns, ru ctors & th = ys ket-sup ket c ket env 35 UM = = MA -s ALL BRICK PRODUCER = W infrastructure, inputs SM LARGE BRICK PRODUCER y MEDI Volume Number Price Volume Number Price RA Volume Number Price and mar Ke market a The mar The mar institutio norms & tren Market

A main tool in EMMA is the Market-System Map. This helps to visualise the difference between the markets before and after the earthquake. This map is for the brick market in Pedang following the earthquake. The black arrows show how bricks reached homeowners from the different scale suppliers, and the red lines show which supply routes were interrupted. www.ShelterCaseStudies.org 41 A.14 Natural disaster

A.14 Indonesia - Sumatra - 2009 - Earthquake Case study: See A.12, “Indonesia - Sumatra - 2009 - overview”, p.45 for background. Country: Indonesia, Sumatra, Padang Project timeline Disaster: Earthquake 7 months- –– Project completion and evaluation Disaster date: September 30th 2009 No. of houses damaged: 6 months- –– Cash distributions 115,000 destroyed houses 135,000 damaged houses 5.5 months- –– Market surveys No. of people affected: Approximately 1,250,000 people 5 months- –– Registration affected through total or partial loss of shelter and livelihoods 4 months- –– Project assessment Project target population: –– Project start Shelters for 750 families Household items to 30,000 3 months - –– Shelter kits, tool families kits, household and Occupancy rate on handover: hygiene items distri- Unknown bution complete Shelter size: Padang Variable Materials cost per household: Indonesia 275 USD 4 - days –– Non-food items were distributed from pre-positioned stock September 30th 2009- –– Earthquake

Project description Cash was distributed to allow 750 families to build transitional shelters. It built on the initial emergency shelter response in West Sumatra in which a package of shelter materials, toolkits, common household supplies and basic hygiene items had been supplied to 30,000 families. Each beneficiary household received approximately 275 USD and technical training on safe construction and minimum standards for shelter. A partner organisation provided technical advice on construction.

Strengths and weaknesses 99 Cash grants helped people buy what they needed and not have encouraged families to build safely. for construction. People had flexibility to build what 88 There was some resentment from those who they wanted. did not receive cash grants. There were sometimes 99 The injection of cash into the markets boosted the very slight difference between recipients and non- local economy and has assisted the self-recovery of recipients circumstances, which made it hard for some other community members, who are also starting to to understand why they had not received support. rebuild their homes. 88 Transitional shelter support should have arrived 99 Despite the amount of money being insufficient earlier. After three months of living in inadequate to complete all work required, it gave people a shelter, many households were ready to build semi- strong starting point to begin recovery. Many people permanent structures. became motivated to begin construction. 88 The half day of training provided to beneficiaries 99 Existing relationships between project staff and was insufficient. House improvements were not communities helped trainings and cash distributions covered in trainings. run smoothly, even though there was some unrest -- There are strict rules that limit logging locally. Many from those who had not received support. beneficiaries only used trees from their own land. 88 The sum of money was too small for all -- The local cost of materials did not increase. However, construction. there was a reported increase in the cost of skilled 88 Project timeframes may have rushed construction labour, which was in low supply and high demand.

44 Natural disaster Shelter Projects 2010 A.14

Background ganisation requested that these were After consulting the communi- See “A.12 - Indonesia - Sumatra - gender balanced and representative ties and other organisations working 2009 - Overview” p.38. of different age and social groups. in the sector, the Indonesian postal service (Pos Indonesia) was selected Distributions The committee’s role was entirely as the best way to distribute the The organisation initially voluntary and a Memorandum of cash grant. responded with non-food items. Understanding was signed with each This started 4 days after the initial committee to lay out clearly their A mobile post office distributed disaster. Rapid response was made roles and responsibilities. the cash grants directly to each beneficiary in their village. Other organi- possible by pre-positioned stocks Each local committee was asked sations had already used this system in Indonesia, held in the cities of to produce an initial list of ben- and its feedback was very positive. Medan, Jogjakarta and Ambon. eficiary households, whom they Since cash grants would be distribut- From October to December 2009, believed matched the targeting ed directly to each beneficiary, there shelter kits, tool kits, household and criteria. These lists were then posted was no need to establish beneficiary hygiene items were distributed to publicly. groups and train their members to 30,000 families. Project staff verified each manage the funds. Transitional shelter household recommended by the Market analysis In January 2010 the organisation committee and selected 620 names In order to monitor the impact shifted its focus to transitional shelter for the final beneficiary lists giving of the cash injection into the local through cash programming. This was priority to the most vulnerable and economy; market surveys were aimed to complement the organisa- needy, taking into account the carried out at 3 project intervals. tion’s previous work and give earth- targeting criteria. A baseline market survey was quake affected people the flexibility Implementation conducted prior to cash distribution, to purchase materials and construct The organisation distributed cash in order to establish the local avail- homes that met their needs. grants in two instalments. ability and cost of materials. This was followed by two further market The approach of providing cash An initial cash grant of 80% was surveys after the disbursement of to enable self build was encouraged followed by house by house moni- the first and second instalments of by the government, as it comple- toring to assess whether cash was the cash grant. mented its own program to distrib- being used for shelter and the com- ute larger cash grants to facilitate pliance with minimum standards. permanent construction. Technical solutions A second grant of 20% was Technical support was provided Selection of beneficiaries distributed. For both payments, through two different kinds of The selection of the community vouchers were given that were later trainings: was based on the organisation’s exchanged for cash by the mobile 1) Training facilitators existing knowledge from its initial post office. response and consideration for the Project staff received training need to have a close liaison with local Delivery mechanism: from an international organisa- authorities and key stakeholders. The organisation initially con- tion. While the training provided sidered using a bank to distribute on T-Shelter gave staff sufficient In each community, the organi- funds, but not all beneficiaries had grounding in good T-shelter con- sation presented the information a bank account or could go to the struction both for community in meetings. The communities then nearest town to collect the funds. training and monitoring, they were elected local committees. The or- not sufficiently equipped to assess

Many materials could be salvaged. Cash grants allowed Temporary shelter built whilst owner was awaiting labour people to pay for materials and labour according to their to complete his house. needs. Photo: Save the Children Photo: Save the Children www.ShelterCaseStudies.org 45 A.14 Natural disaster

A family who used the cash grant to purchase House for 9 people under construction in the foreground. timber beams and concrete. Photo: Save the Children Photo: Save the Children semi-permanent structures or reno- a note in a confidential complaints • A minimum of 3.5m2 covered vations to damaged homes which box. During February the monitoring living area per person. the majority of beneficiaries had and evaluation team also enabled a • A minimum of 2m from the opted for. “complaints hotline” for all sectors, ground to the eaves. so people could call or send their • The roof should provide 2) Training beneficiaries comments using text messages. adequate strength and have a Project staff held 11 two hour pitch of at least 250. workshops in the villages, to dissem- Monitoring • There should be adequate inate technical information about During the monitoring phase, ventilation. construction standards and methods the team used guidance and an • The shelter should provide among selected beneficiaries. At the agreed format to check the com- protection from rain. end of their training, beneficiaries pliance with the following cluster- • There should be at least one received vouchers to be exchanged agreed minimum standards: internal division for privacy. by cash. • Materials and construction • Building should use safe construction techniques to Complaints response should allow for 24 months of use. minimize the impact of further mechanism (CRM) natural hazards. 1) At the targeting level The committees posted the final list of names on community notice boards. At the same time, boxes were installed to collect complaints from those who had not been selected, so they had an opportunity to make their case. Three days later, boxes were collected. After analysing the messages and complaints, meetings were to be held with committees. If those who had complained qualified, they would be added to the final beneficiary list. 2) At the implementation level The community would be able to file complaints and give feedback throughout the entire duration of the project, not only during the selection phase. The communities would have the opportunity to meet directly with staff during their visits, A “renovation”: the roof and foundations were solid - the owner used materials bought with the grant to repair the shelter. approach shelter committees or drop Photo: Save the Children

46 Natural disaster Shelter Projects 2010 A.15

A.15 Indonesia - Sumatra - 2009 - Earthquake Case study: See “A.12 - Indonesia - Sumatra - 2009 - Overview” p.38 for background.

Country: Project timeline Indonesia, Sumatra, Padang Disaster: 12 months- –– Project completion Earthquake with no-cost Disaster date: extension September 30th 2009 No. of houses damaged: 115,000 destroyed houses 9½ months- –– Planned project 135,000 damaged houses completion 8½ months- (approx. 70,000 in Padang –– Second cash disbursement city) Project target population: 3,400 households (3% of overall houses destroyed) 6½ months- –– First cash Occupancy rate on handover: disbursement 66% of all shelters occupied 6 months- –– MoU signing with 12 months after the beneficiaries begins earthquake. Shelter size: Padang 5 months- –– Training for field Variable teams Materials Cost per household: Cash grants for T-shelter: Indonesia 4 months- –– First beneficiary 330 USD per unit training Government estimates for 3½ months- –– Funding proposal reconstruction of a destroyed submitted houses: 1,600 USD

September 30th 2009- –– Earthquake

Project description An international non-government organisation working through a local partner provided cash grants for shelter. Conditional cash grants were given to 3,400 families in two instalments. The local partner used six mobilisers to give technical support. Beneficiaries paid for materials and labour to build timber homes. Most shelters took 10 weeks to build. 77% of the shelters were completed within 12 months of the earthquake.

Strengths and weaknesses 99 Each family was able to build according to their 88 People without land or with damaged housing did needs and wishes. This improved ownership. not get cash or any technical assistance and often 99 Families built shelters that they felt were rebuilt dangerous brick structures. permanent. Families invested and built quickly. -- Standard designs would have made quality control 99 A transparent complaints mechanism helped with much easier. However this would have curtailed the the perception that beneficiary selection was fair. freedom of the beneficiaries to build according to their 99 The project worked in remote rural remote areas needs. because people had space, owned that space and -- Donors had some concerns that permanent housing owned non-productive coconut trees. had been built with emergency funding. 88 A disaster risk reduction opportunity was missed -- The houses built might have been “safer”, but It for people with damaged housing. is a mistake to refer to them as earthquake or hazard 88 The 120 field monitors and community volunteers resistant. had only a few days technical training. It was not realistic to expect them to check the construction quality of 3,400 unique houses.

www.ShelterCaseStudies.org 47 A.15 Natural disaster

Background To be included in the project, The draft lists were posted in See “A.12 - Indonesia - Sumatra - beneficiaries had to have land for the communities along with posters 2009 - Overview” p.38. a shelter and a destroyed house. explaining the selection criteria, Selected families were in a good detailed definitions of the project, Before the earthquake position to complete their shelters an outline of a step-by-step im- In West Sumatra, most families as: plementation plan, and a hotline owned their houses before the telephone number to call or SMS earthquake. The region has a mat- • They were in less urbanised feedback, complaints or requests rilineal system with women owning areas and had previously lived for information. and inheriting land and housing. in single storey buildings. On marriage, the new husband will • They had access to timber and Senior project managers move on to the land of his wife’s experience of using it. operated the phone and were avail- family. Housing has symbolic and • They saw the transitional shelter able for office visits and had after social importance. as a permanent home, worth hour telephone numbers posted on finishing and worth investing in. the office door. Each and every case Family houses are built bit-by- was followed up on an individual bit. In rural areas people usually More than 9000 households basis with village government and paid local builders to build or some- were surveyed and given a vulner- community committees. times built houses for their own ability and eligibility score. Selection families. criteria included female and senior Implementation headed households, low-income Assessments and existing expe- Houses are not purely a finan- families, pregnant women and riences showed that communities cial investment. Remittances are children under 5. a major source of housing finance had the capacity, access to materi- and cash incomes are irregular and Feedback and complaints als, labour and community cohesion seasonal. The community feedback and to manage cash to build transitional complaints mechanisms were es- shelters. A cash approach was also Organisational capacity sential to the running of the project. promoted by the Shelter Cluster. Before the 2009 earthquake, the This system built on lessons learned Beneficiaries built according to their organisation had significant practi- from the 2005 tsunami response needs, wishes and resources. This cal emergency experience. Both the and Jogyakarta / Central Java earth- encouraged fast construction and a international organisation and its quake response programmes. sense of ownership leading to high partner understood the need for ex- perienced staff and sufficient time for community engagement. “Lots of people got jobs The organisation also had ex- as masons [because perienced senior managers and of the project]. New partners who knew the community masons were called and spoke the local languages. The ‘toukonggumpa’ local partner organisation addition- [‘earthquake masons’].” ally had good and long term relationships with the affected communities. This reduced the need for Rural community leader in Pariaman lengthy formal assessments. After the disaster The earthquake of September 2009 destroyed 115,000 houses, and damaged 135,000 houses. In Padang the government responded with assessments and the promise of compensation. Many households affected by the 2007 earthquake were only just receiving compensation at the time of the 2009 earthquake so families did not expect compensation to arrive quickly. Beneficiary selection The communities were selected because the partner organisation knew them well. The project provided cash to allow families to build what they needed. Photo: Bill Flinn

48 Natural disaster Shelter Projects 2010 A.15

Monitoring safety of the structures was very challenging given that each family had the freedom to build according to their needs. Photos: Bill Flinn completion rates and additional trainings on construction and how Participants interviewed during investment by beneficiaries. This to use salvage materials. Better con- the final evaluation stated that they was despite their low and irregular struction was promoted through had spent between 500 USD and incomes. minimum construction standards; 1,000 USD of their private funds in training for field staff, beneficiaries completing the shelters, and that The amount of cash was agreed and masons; production of posters the grant served as an “injection of with other agencies. It was enough and pictures; and weekly technical motivation to a traumatised popu- to build a shelter if supplemented monitoring visits for all recipients of lation”.This resulted in variations in by salvage and available resources. the cash. final shelters with many exceeding The cash was given in two instal- the minimum quality standards. ments (3 million rupiah or 330 Logistics and materials USD). People could only get the Outsourcing material procure- It is difficult to evaluate impacts second amount if they built a safer ment and cash distributions was on a local economy (especially house. decided to be more effective than without baseline data) but new jobs as “earthquake masons” and as Grants were delivered via the In- using the organisation’s internal “chainsaw masons” were created donesian post office in two stages. and limited capacity. by the project. The injection of cash First the participants received 75% Good roads for material supplies and short time frame for building of the funds to complete 85% of and spare local capacity for labour- briefly inflated the prices of some the construction. In the second ers and suppliers to start up helped labour and some materials. Cash phase, the remaining 25% of the the project. also appeared to have pushed grant was disbursed. It was possible that more some new businesses to open (e.g. At the outset of the project, remote communities might have a hardware store). families had to sign a Memoran- to pay higher prices for transport Completed homes were likely dum of Understanding that com- and labour. However, it turned out to be “safer” than the construc- mitted them to spend the money on that people further from roads paid tion practices that have become timber framed transitional shelter only slightly higher prices. The fixed prevalent over the past 30 years but and not on a permanent house or cash grant for all families was seen cannot be described as earthquake repairing an original house. as fair. or hazard resistant. The freedom Technical Impact which was a strength also lead to a wide variation in quality and diver- Four models of shelter were Twelve months after the earth- gence from design principles. designed, but beneficiaries were quake: 77% (2,603) of the transi- free to build according to minimum tional shelters were complete, 11% standards. (369) of the shelters were incom- A 60-strong team of mobilisers plete but in progress, 8% (265) was established to motivate benefi- of the shelters were incomplete ciaries to build to an agreed quality and without sufficient progress to and on time, over 10 weeks. receive the second cash instalment, and less than 5% (163) had not Participants received technical been built.

www.ShelterCaseStudies.org 49 A.15 Natural disaster

Guidance used for a feedback and complaints handling mechanism • Ensure that simple complaints and feedback mechanisms are written into project strategy and budget. • Ensure that ‘complaints handling’ is written into job descriptions of staff at all levels of the organization, and that staff are adequately prepared and trained in handling complaints. • Consult communities and select context appropriate means of communication and technology to receive feedback and complaints and provide a response (e.g. phone or email systems, visiting hours, feedback boxes). • Define the process for complaints handling including timeframes, appeal process and explain the complaints you can and cannot handle. • Ensure the mechanisms are safe, non-threatening and accessible to all. • Inform communities about the complaints process, explain it is a right and encourage communities to use it. • As much as possible, involve local community members, leaders and authorities in the handling of registered complaints. • Provide communities with relevant and timely information about project criteria and parameters to use the feedback and complaints mechanisms, and of improvements and changes made to the project (or why changes are not possible). • Ensure sufficient time and flexibility of implementation to respond to complaints. • Keep records of incoming feedback and complaints, and evidence of follow-up to allow senior management supervision and external evaluation. • Ensure mechanisms are in place for serious complaints, like allegations of sexual abuse, fraud or other sensitive issues.

Various types of structure were built during the project. 50 Photos: Bill Flinn