Household Economy Approach in Burundi

By Sonya LeJeune and Julius Holt

Sonya Le Jeune is currently a programme manager with SC UK in Liberia, working on a food security and livelihoods programme. Previously she has worked with SC UK in Burundi, where she was seconded to the WFP.

Julius Holt is a full-time partner in the Food Economy Group consultancy, with an academic background in anthropology and human nutrition. He first observed drought and hunger as a VSO relief assistant in Botswana in 1966, and became more involved with famine relief for SC-UK in Ethiopia in 1973. Since then, he has specialised in food security assessment and programme planning, with particular interest in the Horn of Africa and the Sahel.

Burundi Ngozi province. Survey to assess economic security in a village

This article outlines key elements of the Save the Children UK Household Economy Approach and Food Economy Analytical Framework, and describes WFP's successful adaptation of the methodology in Burundi. The field experiences described are based on a WFP report drafted in September 1999¹.

Since 1993, the World Food Programme (WFP) has provided food support in Burundi, with activities focused on the distribution of emergency food rations to displaced people, vulnerable group feeding and food for work projects. Since May 1995, Save the Children UK (SC UK) has seconded food security advisors to WFP-Burundi. The main role of these advisors has been to help WFP try out and adopt the Household Food Economy Assessment approach (HFEA)⁴, as a means of helping target beneficiaries within the displaced population.

Making Sense of IDP Sites

Significantly, the displaced population have accounted for about two-thirds of the WFPs food aid in Burundi. In 1999, there were roughly 750,000 internally displaced people (IDPs) in approximately 300 sites, with numbers ranging from fifty to several thousand people (almost all of rural origin) in each site. Some had fled their homes spontaneously in the face of perceived danger and had sought official protection in sites, whilst many more had been officially moved from their homes and 'regrouped' in sites for security reasons. At any given moment, while some IDPs would be newcomers, most would have been present for one or several years.

Access to land has been the single most important determinant of economic status. However cash savings, realisable assets, remittances, capacity to undertake casual employment and the offer of employment are also key determinants of economic security. The majority of people in the IDP sites were near enough to their own land to allow them convenient access for cultivation and, in some cases, those displaced could borrow nearby land. For others, however, distance or insecurity has prevented land access.

The challenge for WFP was to devise an assessment process which would inform them whether or not food aid was needed in a given site, and if so, why, for how long and whether specific groups within the site were at particular risk. Two outstanding problems had to be faced from the start. First, the sheer number of sites to be assessed limited the time that could be devoted to any one site and secondly, there was an almost complete lack of documentation regarding the sites and their immediate hinterland.

SC UK Household Economy Approach (HEA)¹

The main objective of the Household Economy Approach (HEA) is to identify the impact of a shock on the ability of households to acquire food and non-food goods. The first stage in a food economy analysis is the development of a baseline profile, which involves:

• defining the food economy/ household zones in the area of analysis
• socio-economic differentiation, defining wealth or 'access' groups within each food economy zone
• interviews to establish sources of food and income, and expenditure, for households in each wealth group.

The second stage involves collecting hazard information, such as changes in rainfall, crop production, pasture condition or market prices. Outcome analysis involves incorporating the hazard information into the baseline profile.

Both primary and secondary information is compiled, most of which is collected at a community level. Primary data are generated through PRA/RRA³, focus group interviews, key informant interviews, ranking (including pair-wise) and proportional piling. Interviews are conducted with representatives of particular wealth groups which are selfdefined by the community. The interview then refers to a 'typical' household in that group. Interview locations are usually chosen to include as much variation as possible. Interviews are highly structured and a typical interview takes about two hours. Secondary data are used to define the food economy zones.

Analysis aims to estimate the likely effect of a shock on the ability of households within a population to:

• acquire sufficient food
• maintain its non-food consumption, e.g. education, health, fuel, soap and other goods.

Food, cash income, and expenditure are converted into 'food equivalent' units. For the baseline profile, sources of food and income must add up to an average of 2100 kcals/person/day (the designated minimum food requirement for survival). The approach also assumes that there are minimum non-food requirements that need to be satisfied through income and production.

There are two steps involved in estimating whether the household faces a food deficit. The first requires calculating the likely household deficit resulting from the problem, while the second estimates the household's ability to overcome this deficit.

For example, if a household usually generates 50% of its income from food crops, a 50% fall in food crops will lead to an overall 25% fall in household income. An analysis of the ability of a household to overcome the deficit will include:

• consumption of food stocks
• expanding income from wild foods
• receiving gifts
• generating additional income through labour, sale of livestock, and borrowing from kin.

In determining the food gap and consequent requirement for food aid support, certain types of coping strategy will not be accounted for, e.g. sale of key assets, environmentally damaging activities, and illegal activities.

 

A hesitant beginning

When the first food security advisor was seconded to WFP in May 1995, the HFEA methodology had only recently been formalised by SC UK and its first major application (with WFP food monitors in south Sudan) was at an early stage. Although the seconded food security advisor was aware of the technique, she had no direct experience of the field procedures. Whilst proving a useful learning experience for all involved, it did delay the WFPs particular application of the approach until the arrival of a second more experienced advisor in October 1997.

Throughout 1995, a strong commitment developed within WFP to use the methodology and a process of documentation and training began in the latter part of the year. An initial workshop was held in November 1995 to introduce the food economy framework to eleven WFP field staff. The participants then carried out assessments in a variety of settings, including amongst IDP and refugee populations. Findings from these assessments were presented during a second workshop in December 1995 where participants considered the advantages and disadvantages of the method and refined procedures accordingly. A third workshop was held in July 1996 to introduce HFEA to new field staff and assessments then continued as a periodic duty of programmes and field officers.

In early 1997, WFP established two permanent evaluation teams (three people each), who were assigned responsibility for all site assessments using HFEA. One team was based in Bujumbura, the second in Ngozi. A third team was planned in 1997 but for various administrative reasons, was not created until March 1999, again based in Bujumbura. The arrival of the new food security advisor in October 1997 marked the beginning of a second more technically secure phase in the application of HFEA to the context of IDP sites.

Donor requirement

A donor requirement attached to food aid supplied by the US was that WFP should not give food to populations it had not seen. Therefore a visit had to be made to any population identified as potentially needy. This contributed to a high workload, since it was not possible to group similar sites together for planning purposes. Also, where a population was inaccessible due to insecurity or distance from a useable road, the teams had to arrange for people to come to an accessible point, such as a nutrition centre. This congregation limited the possibility of forming representative groups and also increased expectations of a food distribution, potentially biasing answers more than usual.

The time factor

The large number of sites to be visited prevented teams spending more than one day at each site for a given exercise. Day length would be further reduced by security considerations (security clearance would only be given at 9am). In Bujumbura, signatures for forms needed to be obtained which often delayed matters up until 10am. During calm periods, vehicles had to be back by 4.30pm but when there was a 'tense' situation, this would be brought forward to 3.30pm. Given that it took an average of 1.5 hours to travel between office and site, a day in the field did not allow much time for interviews. As a result, teams generally forfeited their lunch break.

Fieldwork was a continuous activity. To avoid unthinking routine, each team could only undertake two field visits per week, with the other days devoted to complete write-up of results. During a four month period (one agricultural season), this amounted to approx. 100 site visits (not allowing for staff holidays, illness, training, being grounded for insecurity, etc.). Generally, there were more than one hundred sites identified as being potentially in need of assistance during an agricultural season. Clearly, there was a lot of pressure on teams to get enough information from just one limited day of questioning.

Population movements

HFEA depends on the target population having been present long enough to establish some identifiable pattern of life - perhaps a month or more. This was the case for most people in most sites. However, in some localities there was much movement, depending on security conditions and large numbers of recent arrivals (perhaps just coming out of hiding). New arrivals in a poor state of health automatically received full rations for several months, however the overall recommendations for the site were based on interviews with long-term residents.

First phase of HEFA (Early 1996 - October 1997)

From May 1997, the WFP carried out site evaluations on a systematic basis. Personnel identified provinces most in need and programmed a series of visits to cover all the accessible sites. In order to get through the work, teams were sometimes obliged to do two evaluations in a single day. Once the supervisor had passed all site reports and recommendations had been accepted, a summary table of distributions was prepared and agreed with the governor of the province. Distribution planning then followed, which could be several weeks after a site visit. However, for obviously urgent cases a distribution was carried out before all the site visits were completed. Sometimes food aid coincided with harvest time, or it was not given during the most food insecure months (timing of distribution was not planned to fit in with the agricultural calendar). There were occasional unscheduled visits in response to urgent requests.

The initial food assessments were based on semi-structured interviews, with questions aimed at the population as a whole rather than trying to characterise typical families. There was no wealth group breakdown. The evaluators collected information on current food and income sources and how they differed from normal (a year from before the crisis - usually 1992 when people were not displaced). The relative importance of the different food and income sources was estimated from proportions stated by informants without proportional piling, or other verification or backup calculations. Pie charts were drawn to compare current food and income sources to before the crisis. The evaluators estimated actual consumption by asking about kilos of food consumed (converted into kcalories). Based on these findings a recommendation was made for a number of monthly rations or (if no urgent need was seen) for a Food for Work (FFW) project.

Second phase of HFEA (October 1997 onwards

The new food security advisor who arrived in October 1997 had been trained in HFEA and was seconded to WFP specifically to provide technical support to the evaluation teams. It soon became clear that there were basic comprehension problems for the teams, largely due to their incomplete initial training and lack of technical support. It was therefore decided to allow teams participate in a real HFEA exercise in a rural area, so that they could see how the various items of information are supposed to link up, to practice hypothesis testing, to understand the calculations and to appreciate the use of secondary sources. The opportunity was taken in January 1998 to conduct a full HFEA in Gitega, to coincide with a nutrition survey.

The fact that the exercise bore little relation to their usual assessment procedures served to confuse the team initially, e.g. two weeks spent in the field instead of a few hours, analysis on the basis of wealth groups, the time period being a year rather than a few months. But the teams slowly grasped the approach and the following six months were devoted to monitoring, correcting and reinforcing the teams work. During this period, a visit from a technical consultant from SC UK helped to focus the timing of evaluations and recommendations so they fitted more appropriately into the seasonal calendar.

By September 1999, WFP were prioritising site visits based on perception of greatest need, rather than systematic coverage. During harvest time there were few assessments of sites with access to land, as farmers would be busy in fields and were also unable to make accurate assessments of crop quantities. This period was therefore used by teams to arrange provincial level meetings with key informants to identify geographical areas where people had not been able to cultivate properly and where there was likely to be a need for food aid in the coming months. A calendar of evaluations was prepared and visits planned to fit in with the agricultural calendar. This process had to be flexible enough to accommodate unscheduled urgent requests.

The core framework of the HFEA was applied and adapted to the specific context in Burundi during the second phase. The chief differences from normal HFEA practice are outlined in table 1.

Conclusions

1. The adaptations made to the HFEA approach have not been fundamental and the basic field procedures, e.g. focus group interviews, have proven workable. As sites had more or less homogeneous populations, it was possible to identify typical households that represented the greatest proportion of each site.
2. The first phase of the introduction of HFEA in 1995 improved upon the previous rapid evaluations, since it added a new element of food-and-income logic to the exercise. But it was not until the second phase from late 1997, with a fuller and more rigorous application of HFEA, that greater insights were gained into the economy of sites and the economic differentiation of groups so that better reasoned estimates of need were obtained.
3. The specific constraints of the context did limit scope and depth of analysis, e.g. lack of secondary data to gain better knowledge of the physical and economic areas surrounding the sites. The sites were rather lonely places for evaluation teams and there was little extraneous information to back up what they found out.
4. Time constraints and the large number of sites to deal with limited the depth of information obtained. Teams had just enough time to get a 'convincing story' but more information, e.g. to help target within sites, would have required more time at each site.

Recommendations

First, there were some concerns about maintaining data quality. During the second phase in particular, the teams gained a strong technical background and sound work patterns developed. However, it was felt that sustaining this effort might be difficult. HFEA depends on each team member being intellectually 'alive', but repetition week in week out may encourage field workers 'to go onto automatic'. There is, therefore, a need to break routine. The authors suggest that the lengthy HFEA procedures used in revisits to sites might be curtailed (at least for the first two visits after the first full HFEA). Extended enquiry would only be necessary if some key factor had changed dramatically. Thus, two proximate sites could be visited in a day. The 'spare' time could then be devoted to outstanding questions which never get answered, for example improved understanding of access to land and its constraints, the casual employment market and the time taken for new arrivals to gain an economic foothold.

Secondly, in order for HFEA to be used to estimate the proportion of a site population requiring food aid, enquiry would need to be extended into an in-depth analysis of wealth groups. This would require more field work time.

The HFEA has proved an adaptable methodology to the IDP site situation in Burundi. It is hoped that this experience will have relevance for others, beyond the Burundian situation.

UKLiberia, email: dambudzo@mail.com, or Julius Holt, The Food Economy Group, Longview, Browns Spring, Potten End, Berkhamsted, Herts HP4 2SQ,UK. Tel: +44 (0)1442 875 709 or email: feg.uk@foodeconomy.com

¹Le Jeune S and Holt J (1999): Making sense of IDP sites. A review of the use of household food economy analysis to target food aid for internally displaced people in Burundi.World Food Programme, Bujumbura, September 1999

²The Household Economy Approach. A resource manual for practitioners. Save the Children, 2000

³ PRA/RRA: Participatory Rural Appraisal/ Rapid Rural Appraisal

⁴Since this report was written, the term Household Food Economy Analysis (HFEA) has been changed to Household Economy Analysis (HEA) within Save the Children, to reflect the fact that food is only part of the household economy.

⁵At the time of the 1999 WFP Report, HFEA calculations were based on a 1900 kcal survival ration. This has since been revised to a 2100 kcal minimum daily ration.