Food security and nutrition monitoring for drought ... - ScienceDirect

Food Policy 1994 19 (3) 285-300

Food security and nutrition monitoring for drought relief management The case of Botswana Krishna Belbase Project Officer for Household Food and Economic Security, UNICEF Botswana

Richard Morgan Regional Economic Adviser, UNICEF Eastern and Southern Africa Regional Office, Nairobi, Kenya Botswana presents relief management

an example

of a relatively

successful

programme

of drought

in the Southern African region. The programme has evolved over a long period and has been reasonably effective in safeguarding the welfare of vulnerable population groups, particularly during the long drought of 1982-87 as well as more recently in 1992. The key contributory factors include: a small and largely accessible national population, availability of domestic and international resources, existence of rural infrastructure, government commitment, district-level capacity and a timely and fairly comprehensive food security and nutrition monitoring system. The main focus of this paper is on the role of the food security and nutrition monitoring system in Botswana’s drought relief management, the context for its operation, and its use in policy making and programme design. The evolution of the information system and institutional coordination thereof is traced with analysis of their strengths and weaknesses. In addition, the drought relief programme is assessed in the context of persistent poverty, growing inequalities and of recent changes which aim to bring about some reduction of state subsidies. Conclusions and policy implications are drawn and their relevance for other African countries is outlined. Keywords:

drought,

early warning,

Botswana

Botswana is one of the most drought prone countries in Africa. Since the country gained independence in 1966, it has experienced drought in 13 out of 27 years. The most prolonged drought occurred during the period 1982-87. During this time, the country developed and implemented a large-scale drought relief and recovery

This article is written in the personal capacity of the authors, and does not necessarily reflect the views or policies of UNICEF or the United Nations. Grateful acknowledgement is made to G. L. Tlogelang, N. J. Manamela, T. Maribe, B. Cogill and I. Ndombi for their valuable comments and inputs on drafts of this article. Responsibility for final content rests with the authors. 0306-9192/94/030285-16

0

1994 Butterworth-Heinemann

Ltd

285

Food security and nutrition monitoring in Botswana: K. Belbase and R. Morgan

programme. The programme was generally successful in safeguarding the wellbeing and stabilizing the consumption of vulnerable population groups, although less so in remoter areas where child malnutrition rates remained relatively high. With timely relief interventions, Botswana avoided significant famine-related loss of life and forestalled any overall deterioration of the nutritional status or survival chances of young children. The relief and recovery measures also limited the impact of repeated droughts on the rural economy and helped prevent widespread abandonment of family farming. However, as will be argued, they did not effectively address pre-existing wide inequalities of income and wealth which were probably exacerbated by the droughts. Many factors have contributed to the relative success of drought relief management in Botswana since the early 1980s. These include: a small population of 1.3 million, of which three-quarters are concentrated in an easily accessible eastern belt; a strong government revenue position and substantial reserves of foreign exchange; the steady build-up and present widespread coverage of rural infrastructure and consumer market networks in all but the remoter central and western areas; continued national policy commitment to safeguarding the basic welfare of those affected by drought; considerable district-level administrative capacity and a timely and reasonably comprehensive food security and nutrition monitoring system. This paper presents a critical analysis of the drought relief and recovery programme developed by Botswana with particular focus on the role of the food security and nutrition monitoring system in policy making and programme design. It is based on a general framework for a drought relief management strategy derived from the Botswana context (not fully presented here), which assumes that:

l l l l

l

l

there are likely to be short-run effects of drought at household level on both incomes and asset-holdings; the effects of drought may persist after drought ends, and policies and programmes may be needed to support the recovery of family production; young children and mothers are likely to be among those most affected by declining access to food and clean water during a drought; needs for preventive and basic curative health care, particularly for children, are likely to increase rapidly during a drought, while the inter-action of disease and deteriorating dietary intake can lead rapidly to worsening nutritional status (UNICEF, 1990); a drought management strategy should address issues of food entitlement as well as physical access, and should seek to forestall the liquidation of savings and productive assets by poor households in order to meet current consumption needs; a drought management strategy should seek to be consistent with, as well as promotive of, long-term sustainable development. This would imply both the operation of the strategy during non-drought times, through preventive planning and vulnerability reduction programmes, and the taking of opportunities which may arise during ‘drought emergency’ to address more structural problems (Holm and Morgan, 1985).

Based on these considerations, for drought relief management: 286

the framework

identifies

the following

as key areas

Food Policy 1994 volume 19 Number 3

Food security and nutrition monitoring in Botswana: K. Belbase und R. Morgan l

l l

l

aggregated and disaggregated information systems, which reflect the effects of drought on households and individuals; are linked to decision making at various levels; and are used for event prediction, resource allocation, relief programme management and impact assessment; institutional coordination, within the public sector and between it and the non-governmental sector (broadly defined), at various levels; prior analysis of likely resource requirements for drought management (e.g. financial, managerial, transport, infrastructural) and identification of potential sources thereof; strategic planning for public sector response to drought, based on prior analysis of vulnerabilities and the likely impacts among different population groups, on natural resources and animals, and on economic conditions. Such planning should be participatory; lead to consensus and awareness among the agencies likely to be involved; and is likely to be highly cost-effective both in narrow financial terms and for the safeguarding of human lives, welfare and future productive capacity.

Botswana’s rural household

economy

Botswana has experienced high rates of economic growth since independence in 1966. Performance peaked during the 1980s with average real GDP growth reaching over 9 per cent per year. The main source of economic growth has been the diamond mining industry. Mining now contributes about half of GDP. Consistently high growth rates since the mid-1970s have supported increased investment in basic health and education, roads, housing and water supplies, and the achievement of high coverage levels of services such as primary health care, primary and junior secondary education and, to a lesser extent, clean water and sanitation. Child mortality rates have fallen rapidly, to levels among the lowest in sub-Saharan Africa (UNICEF, 1993). Botswana’s development efforts have not been so successful in reducing rural household poverty. According to a 1989 poverty datum line (PDL) study (based on 1985/86 data), 64 per cent of rural households and 30 per cent of urban households had incomes below the national PDL (CSO, 1989). A high proportion of households in the poorest strata continue to derive their income from a variety of insecure sources such as arable farming, casual employment, intra-family remittances and public transfers (MFDP, 1988). About 75 per cent of the population still live in rural areas (including the several ‘major villages’) and engage in some form of agriculture. The majority of rural households are economically disadvantaged, although in most areas have good access to basic services. Livestock and arable agriculture constitute the two main sources of rural income and employment, but raising of livestock, particularly cattle, is by far the more remunerative. The causes of rural poverty are partly structural. Cattle distribution is highly unequal. Findings from a 1987 livestock study indicate that ‘. . . cattle distribution is highly skewed with about 40 per cent of the households without cattle, while over 60 per cent of the national herd, 2.3 million, is owned by less than 10 per cent of the farming households. The severity of the drought could have further contributed to the maldistribution of cattle’ (MOA, 1991). Most poor households engage in arable agriculture, sometimes supplemented by small livestock. The former, however, is a highly insecure means of livelihood in Botswana. Rainfall tends to be erratic and prolonged drought periods are common. Opportunities for irrigation are extremely Food Policy 1994 Volume 19 Number 3

287


limited and almost all smallholder farming relies on rainfall as a sole source of moisture. Crop production is characterized by low productivity such that even during years of ‘good’ rainfall, 90 per cent of farming households are unable to meet basic food needs from their own production. A dearth of employment and income-earning opportunities in the rural areas has persisted despite very rapid growth of formal employment in the country as a whole during the 1980s (Harvey, 1992). Limited access to formal or remunerative informal employment is an important factor contributing to the persistence of structural poverty. Predisposition to poverty is exacerbated by the tendency of seasonal and casual employment opportunities (such as weeding, threshing and beer-brewing), on which low-income rural families disproportionately rely, to collapse following the failure of the rains. A more basic explanation of poverty can be found in socio-cultural and historic factors which have disadvantaged certain population groups. The phenomenon of unplanned teenage premarital pregnancy is common in Botswana (UNICEF, 1989). This leads to a high prevalence of single mothers and female-headed households which lack labour resources and productive assets more generally. Over half the rural households and over 40 per cent of urban households are femaleheaded and many of these are not economically supported by an adult male. Poverty incidence among female-headed households is disproportionately high. Nationally, cash incomes of male-headed households were in 1985/86 about 2.2 times higher, on average, than those of households headed by women (MFDP, 1988). In terms of ownership and control of resources, female-headed households are relatively as well as absolutely deprived. In the agricultural sector, such households control only about one-third of farm units. Their ownership of livestock is also low. Also among the poorest and most deprived populations tend to be historically and increasingly disadvantaged groups such as the Basurwu (so-called ‘Bushmen’), Bakgulugudi and some other minorities who often live in remote settlements and at isolated water points. Such groups are officially known as ‘remote area dwellers’ (RADs). Poverty and unemployment vary across districts and the poor are concentrated particularly in the remoter and mostly drier western, west-central and northern areas. Average incomes are higher in urban areas, but considerable poverty exists in towns and on urban fringes. The incidence of under-nutrition among children under five years varies considerably among districts, and tends to correlate closely with these areas of concentration of low incomes (although not necessarily with individual households). The high rate of population growth, of 3.5 per cent per year, aggravates the poverty situation. Average family size is four persons in the urban areas and six persons in the rural areas. The age structure is such as to lead to high dependency ratios. The most recent estimates of the Gini coefficient indicate an increase from 0.52 in 1974 to 0.56 in 1985-86, suggesting increased income inequality (MFDP, 1988; 1975). Indicators of wealth such as ownership of land, cattle, tractors, etc. also suggest that inequality is increasing. For example, during the drought years of the 1980s levels of mortality among cattle were persistently higher in the smaller herds (Morgan, 1986; MOA, various years). In this context, one of the questions regarding the response to drought in Botswana is whether the relief and recovery measures have counteracted or contributed to the apparent tendency towards greater inequality. 288

Food Policy 1994 Volume I9 Number 3


A review of Botswana’s

drought relief programme

This section starts with a brief review of the evolution of Botswana’s drought relief programme, followed by an assessment of different aspects of the relief programme, including institution coordination, information systems, relief interventions and resource mobilization. The main focus is on the information system and flows.

Evolution of the programme The Botswana drought relief programme evolved through a process of selfevaluation wherein remedies for earlier weaknesses were sought (Holm and Morgan, 1985). When Botswana gained independence in 1966, a prolonged drought was just ending. The 1970s were generally years of good rainfall. By the mid-1970s, various consultants and local professionals began to warn that, based on past trends, the situation was likely to change for the worse (Sandford, 1977). In early 1978, the Botswana Society (a local agency based at the national museum) held a symposium which concentrated on the human implications of drought and effectively placed this issue on the government’s policy agenda (Holm and Morgan, 198.5). The implications were taken up by the (then) Rural Development Unit, a small strategic planning body in the key Ministry of Finance and Development Planning. The Unit initiated work on designing a ‘human relief’ strategy, which was soon put to the test by the drought of 1978-79. Another important event in the process of addressing the ‘human aspects’ of drought was the establishment of a nutrition monitoring system in 1978. As its coverage expanded, this system was increasingly able to provide indicative data on child nutritional status by area. Meanwhile, with support from FAO, Botswana instituted a national agro-meteorology system in the early 1980s. Also important for the implementation of drought relief was the prior existence of a rural primary school feeding programme and a clinic-based supplementary food distribution scheme for medically selected vulnerable groups. Both had been institutionalized, with support from WFP, several years before the return of drought conditions in the late 1970s. The widespread network of agricultural extension workers, created during the 197Os, was also to provide capacity for relief and recovery interventions. When the country was plunged into a severe single-season drought in 1978-79, information from the new nutritional surveillance system was used for broad targeting of feeding programmes up to the village level and for the monitoring of child nutritional status. The government programme at this stage included cash incentives for farmers to clear their fields for ploughing, provision of vaccines and food supplements for cattle, expansion of pre-existing food distribution through health facilities by relaxing selection criteria, and the supply of free seeds to arable farmers. In addition, a series of public works on a cash payment basis was initiated in the north-west region, which was simultaneously affected by a ban on cattle marketing due to foot and mouth disease, in order to cushion rural incomes. The 1979-80 relief programme had limited success, encountering serious problems in initial mobilization, targeting and timeliness of interventions, but provided experience in the management of a multi-faceted relief effort. The government conducted an evaluation of the 197940 programme soon after. The evaluation reports called for the strengthening of the food distribution system, better contingency planning, improvements in the monitoring system, increased food rations for vulnerable communities, inclusion of more groups in the feeding programme, higher wages for those employed in public works, and greater Food Policy 1994 Volume I9 Number 3

289


decentralization of projects to the district and village levels (Gooch and McDonald, 1981). Following a national workshop, many of the evaluation proposals were adopted. Staff in the Ministry of Local Government and Lands responsible for ongoing food distribution programmes were upgraded in the early 1980s and the budget and infrastructure of the retitled Department of Food Resources were expanded. This set in place the core of a national food delivery capability which could be expanded _ although not without considerable strain and partly through contracting of services from the private sector - in times of drought. Also of importance was the clarification at this juncture of national policy towards drought relief and the achievement of consensus among national and local government agencies on the main principles of drought response, as well as their respective responsibilities. The most severe drought period in Botswana since independence lasted from 1982 to 1987. The extensive relief programmes carried out during this period, supported by steady increases in government financial capacity, were evaluated in 1990. The relief programme was further modified in the response to the 1991-92 drought, based on recommendations of the 1990 evaluation (Food Studies Group, 1990). Institutional coordination A central feature of Botswana’s drought relief programme is the institutional coordination which promotes the timely flow of information and its use in drought relief. The coordination is strong at both the district and central levels. At the district level, the main coordinating body is the inter-sectoral District Drought Committee (a sub-committee of the District Development Committee; viz. Figure 1). This is chaired by the District Commissioner and has membership from the District Council, Tribal Administration, and the local representatives of central government ministries. The committees meet every two to three months in times of drought to assess the situation, coordinate the activities of various departments involved in relief, and to monitor the activities undertaken under the relief programme. They provide information and district-specific feedback on droughtrelated programmes and needs to central government. At the central level, the key coordinating body is the Interministerial Drought Committee (IMDC), established in 1978. This is chaired by the Coordinator of Rural Development within the Ministry of Finance and Development Planning (MFDP) and is composed of high ranking officials from several ministries. Placement of the coordination of this committee within the powerful finance and planning ministry has been the key to ensuring responsiveness and rapid access to resources. Meeting several times a year during drought periods, the IMDC makes recommendations to the Cabinet through the Rural Development Council on all drought related matters, including the design and scale of national relief interventions and the allocation of funds. Support is provided to the drought monitoring function of the IMDC by the Early Warning Technical Committee (EWTC) which was formally established in 1984 (Morgan, 1985). The EWTC is responsible for detailed monitoring of the drought situation and its effects. It has representatives from the Nutrition Unit, Agricultural Statistics Unit, Department of Meteorology, Department of Crop Production and Forestry, Department of Food Resources, Botswana Agricultural Marketing Board (BAMB) and the Central Statistics Office (CSO). The chairperson of the EWTC is the Chief Economist (Rural) of the Rural Development Coordination Division, who is also a member of the IMDC. Prior to an IMDC 290

Food Policy 1994 Volume 19 Number 3


Minister

-_)

Permanent

(MFDP)

Secretary

t DEVELOPMENT

RURAL

(MFDP)

COUNCIL

Inter-Ministerial Drought Committee A

Early

Warning Technical Committee

District Drought Committee Major Nutritional Surveillance

Figure 1

Information

Sources

Agro-Meteorology

Drought related information

Crop and Livestock Assessments/ Food Security

flow in Botswana

meeting, the EWTC reviews the latest information on agro-meteorology, livestock and crop conditions, food prices and stocks, child nutrition and the relief programme itself, and prepares recommendations for the IMDC. The EWTC has been able to draw on the individual information systems built up over time by various ministries and parastatals. A further source of both forma1 and informal information for the EWTC is the ‘Drought Assessment Tour’, also instituted in the mid-1980s. Teams of IMDC/ EWTC members visit each of Botswana’s 10 districts during the single rainy season with follow-up visits during drought years. The group holds discussions with the District Drought Committees, makes field visits and meets other key informants. The rainy season tour forms the basis, when necessary, for the declaration of a drought emergency by the President, and for the initiation of detailed planning of relief programmes and supplementary resource mobilization efforts with international donors. Information

Central

sources

to drought

and flow

relief management


in Botswana

is an information

system

that 291


has evolved over the past two decades, Particular features of the system are:

(9

(ii)

(iii)

(iv)

(v)

based

on the activities

of various

sectors.

A considerable degree of decentralization. Information on key variables such as rainfall, agricultural trends and child nutrition are assessed at a large number of localities by extension cadres, as well as in aggregate at district level, and have reasonable coverage and speed of reporting. However, as discussed below, this does not extend to indicators of human health and access to clean water. The system takes some, but limited account of particularly vulnerable population groups. Children under five years attending health facilities are monitored through nutritional surveillance, while much less systematic monitoring is carried out of remote area dwellers and the destitute, through District Councils. The latter group involve those qualifying for a monthly income supplement under a regular welfare provision. There is considerable horizontal and vertical sharing of information. Both at the district and central levels, information is analysed by sectoral departments and shared with inter-sectoral groups. Information is, to some extent, used for decision making at each level. At the village and district levels, information is used for targeting of interventions, although this aspect is often weak. At the centre, it is used for policy and programme design including national and inter-district resource allocation and mobilization for drought relief. Information is also used to monitor programme performance and as a basis for evaluation. Formal information flows, within the bureaucracy, tend to be reinforced (or preceded) by a variety of other sources, including the lobbying efforts of local political representatives, a fairly active media, the growing NGO sector, and the personal experience of national decision makers (many of whom retain rural family ties and farming interests).

of the The main components Botswana are as follows:

‘formal’

drought

related

information

system

in

Agro-meteorology. Regular collection of rainfall data is integral to the drought monitoring system, whose most timely information flow is the current report of the Department of Meteorological Services (Cogill, 1990). Rainfall data are collected on a weekly basis from about 300 recording points and are analysed in terms of percentage departures from the long-term mean, both for the particular month and cumulatively for the season (Quinn et al., 1988). These data are also used to calculate a Water Satisfaction Index (WSI) for Botswana’s main cereal crops, which enables the Meteorology Unit to make timely predictions on the probability of harvest. This information is provided to the EWTC and the District Drought Committees. Agricultural statistics. Information on the agricultural sector comes from two main sources. The first is qualitative monthly reports on the status of the field crops, livestock and grazing conditions prepared by the agricultural extension agents. These reports are used by the Ministry of Agriculture to prepare a monthly ‘Agricultural Situation Report’ which makes rough crop forecasts during the growing season and grades livestock, animal watering and grazing conditions by area. The second source is the annual agricultural survey of about 2500 farms which 292



is used for medium-term analysis of trends and vulnerability. are also monitored by an annual census linked to vaccinations.

Livestock

numbers

Food prices. Data on food prices are collected on a monthly basis by the CSO for several towns and larger villages. The food price information is used by the EWTC to assess the possible impact of inflation and shortages in specific areas. However, since private sector food retailing in Botswana is extensive, generally competitive and heavily reliant on imports from South Africa, supply problems are uncommon except in remote areas. Price trends are not necessarily a good indicator of local and household demand, which is of major concern in a highly food-deficit country like Botswana. Food security. While the 1985 National Food Strategy, and subsequent policy statements, emphasize the importance of food security at both national and household levels (MFDP, 1985; MOA, 1991), information collection has so far been limited to food availability (or supply) at the national and district levels. The Strategic Grain Reserve Monitoring Group (SGRMG) monitors the overall availability of food in the country, while the National Food Strategy Monitoring Group provides longer-term policy guidance on food and nutrition issues. The Botswana Agricultural Marketing Board, which procures grains from domestic producers as well as foreign suppliers, coordinates closely with the SGRMG. The EWTC receives monthly reports on BAMB’s stocks of two main cereals, sorghum and maize. In recent years, direct importation of food by the private sector has been encouraged. The Food Resources Department in the Ministry of Local Government, Lands and Housing reports on the supply and distribution of food used for vulnerable group and primary school feeding programmes, for which it is responsible. Changes in numbers of vulnerable group beneficiaries reported from local levels (e.g. the numbers of mothers requesting and receiving food from health facilities) may provide some guide to the severity of drought conditions. Even with these indicators, information on household-level food security remains sporadic. Perhaps the closest available proxy indicator is data on income levels collected by the household income and expenditure surveys of the CSO. These are irregular and the most recent was in 1985/86. Lack of a more regular and comprehensive system of data collection and monitoring for household-level food security has resulted in limited ability to assess the direct impact of drought and drought relief efforts on poor and remote households. The lack of a well-targeted system of agricultural relief (see below) may be partly attributable to this weakness of the information system. The need for regular monitoring of household-level food and economic security, and its potential role in targeting and in designing area-specific programmes, has received stronger recognition since the 1990 evaluation. Establishment of such an information system could feasibly be based on data collection at sentinel sites (or villages), selected through risk mapping techniques (Cogill, 1990). It could also enable better understanding of the coping mechanisms adopted by poor families during drought and the location and characteristics of the most vulnerable households. National Nutrition Surveillance System (NNSS). The NNSS is taken as part of the formal national Early Warning System. As indicated earlier, the information is used to trigger the expansion of supplementary feeding as well as for regular Food Policy 1994 Volume 19 Number 3

293

Food security and nutrition monitoring in Botswana: K. Belbuse and R. Morgan

programme design. Comparisons between villages and districts on the basis of trends in NNSS time-series data have also been used in decision making on relief priorities and the allocation of funds for public works projects. The NNSS aims to provide information on child nutritional status for decision making at all levels (health facilities, districts and centre) and to monitor individual child growth. It is based on clinic attendances and monthly reporting of underweight prevalence and other indicators such as growth faltering. Primary data collection (age, sex, weight, guardian education) for children under five years at the local health unit is done by Family Welfare Educators. At this level, the data on weight for age are analysed (using Harvard standards) to identify severely and moderately malnourished children and those that have experienced growth failure. Both moderately and severely malnourished children are provided with supplementary rations. Severely malnourished children may be referred to clinics or hospitals for medical examination and treatment. The information summary report prepared at the health facility is forwarded to the district level, where the health officials prepare monthly summary tables. The District Health and Nutrition Education Officers (DHENOs) check the entries from the health facilities and forward the originals to the CSO. They also summarize the data by health facility and provide feedback to the reporting facilities. The District Drought Committees use NNSS data for a variety of drought relief management decisions. The information is presented by the District Health Team and is used to monitor child nutritional status across different locations and to make adjustments and interventions within available programme resources. However, due to chronic staff shortages in the DHENO cadre and other factors, use of NNSS data is liable to be weak at district level. At central level, the information submitted by districts is compiled and analysed at the CSO, which sends the results to the Nutrition Unit of the Ministry of Health as well as to other members of the IMDC. The Nutrition Unit presents an assessment of the situation to the EWTC, which analyses it together with other indicators of the district-level situation and reports to the IMDC. The information is further used to highlight the severity of drought and to make recommendations to the Rural Development Council and the Cabinet. Where necessary, the IMDC will raise specific questions or issues arising from the nutrition surveillance data with the District Drought Committees. Considerable problems of sample bias and lack of representativeness in the NNSS have been pointed out, as has its tendency to centralization (Cogill, 1990). In addition, it is pointed out that the data generated do not provide, in themselves, early warning of drought and its impact (Cogill, 1990). However, the usefulness for drought management of the NNSS time-series data lies in providing a basis, when taken in conjunction with other indicators, for broad targeting of resources during as well as for raising concern and individual or successive relief campaigns, triggering response particularly at central administrative and political levels. The creative combination of NNSS data with data from other sectors at central-level for decision making is coupled with the use of nutrition data at local levels for direct intervention, as described above. After the drought years of the 198Os, and with reintroduction of selection criteria for supplementary feeding, clinic attendance rates dropped substantially in many districts. In response to this, the potential role of community-based growth monitoring has been receiving greater attention (MFDP, 1993). Future incorporation of such an approach should strengthen the existing surveillance system, and 294


Food security and nutrition monitoring in Botswana: K. Belhase and R. Morgan

improve its capacity for nutrition promotion, health education and mobilization of parents and communities around nutritional issues. This may in turn help to address the more basic factors associated with malnutrition which are not tackled by the drought relief programme. Health and water sectors. The national information base for drought management and monitoring does not currently include epidemiological data, although, as earlier indicated, the health impact of drought can be serious. This is mainly due to the long delays which exist in the processing of health facility reports at national level, except for notifiable diseases. The lack of analysed health information may have been associated with an underestimation of the resource needs of the primary health care system in times of drought, especially during the 1980s. National data on drinking water availability are also lacking, although, as with health, information on specific indicators is often available to the District Drought Committees from the District Councils. Programme management information. This is probably the weakest element of the drought relief management system. The lack of timely information and feedback on programme delivery and impact has been a common feature of most relief programmes, and this has hampered the ability of central and local governments to adjust, fine-tune and evaluate relief efforts. Local and aggregated data on key indicators such as quantities of food delivered, numbers employed in public works and receiving agricultural subsidies have tended to be available only after long lags, reflecting both the decentralized nature of delivery systems and serious weaknesses in reporting (Simmons and Lyons, 1992). District and central planners have initially relied on data such as beneficiaries registered (rather than food received), projects approved and funds disbursed, to get some sense of programme performance. Informal reports, and information shared during the Drought Assessment Tours, have been important supplements to the limited data flows in this area. Relief interventions Drought relief interventions in Botswana are increasingly an outcome of the programme experience (and financial resources) that the country has accumulated. The most comprehensive relief effort was mounted during the 198247 period. The programme implemented in 1992-93 was on a smaller scale, following policy modifications introduced after the 1990 evaluation (Government of Botswana, 1991). There follows a brief description of the main components of the relief programme launched during the 1982-87 drought and the changes made in 1992-93. Supplementary feeding programme. Botswana has an ongoing feeding programme for malnourished pre-school children, primary school children (one meal per day on school days), TB patients and pregnant and lactating women selected on medical criteria. During the drought periods, coverage is expanded to include other vulnerable groups. The feeding programme during the 1982-87 drought was the most comprehensive. During this period, the following groups were eligible for supplementary food aid: all pre-school children, all pregnant and lactating women, all children between 6 and 10 years not in school, registered destitutes and RADs. In addition, rural primary school children who were already receiving school meals in term-time were to receive take-home rations for weekends and holidays. Coverage was impressive in aggregate terms, with around 60 per cent of the Food Policy 1994 Volume I9 Number 3

295

Food security

and nutrition

monitoring

in Botswana:

K. Be/base and R. Morgan

population reportedly receiving food rations at the height of the programme. Subsequent analysis, while estimating overall food distribution levels against beneficiary requirements of 82-85 per cent during 198647, has found considerable variation in food distribution performance between districts, and that distributive shortfalls were particularly marked in the programme component for the highlyvulnerable ‘Remote Area Dwellers’ (Simmons and Lyons, 1992). The 1992-93 feeding programme was much smaller in scale. It still aimed to cover all children under five, but excluded food aid to non-medically selected pregnant and lactating women, destitutes and the RADs. This modification reflected the government’s wish to better control the total cost of drought relief as well as to reduce the dependency of the population on government for relief which was perceived to have developed in the 1980s. The information system, including the NNSS, continued to be used for planning, managing and monitoring the child feeding programme. However, food delivery performance even for the reduced target population is recognized to have been inadequate, and partly reflected institutional difficulties in ‘gearing up’ after several non-drought years. Labour based public works. This programme aims at supporting the incomes of households most affected by drought, while, as a secondary aim, creating productive or socially useful infrastructure. During the 198247 period, it was successful in creating short-term employment and providing income support, particularly to rural women who, as seen earlier, tend to be economically disadvantaged. Roughly 70-80 per cent of the 50-70 000 temporary workers engaged each year in the programme were women. At its peak, some 20 per cent of the rural working population was involved in the public works (Buchanan-Smith, 1990). The programme has been successful in effecting timely transfers of income to rural people through the District Councils, in cushioning local economies more generally from the impact of drought, and in achieving reasonable targeting of the rural poor (through a ‘self-selecting’ standard wage rate). However, its efficiency in creating physical assets has been much more questionable. While village development committees and the district authorities were responsible for identifying small-scale projects within a range of options provided by central government, they often lacked adequate technical capacity for planning and implementation. Although considerable village-level social infrastructure was created, productivity and quality of work tended to be low. Few projects had the effect of creating future incomes or jobs. The 1992-93 drought relief programme also included labour based public works. Projects were initially selected from the current National Development Plan (1991-97) rather than by village or local authorities, for accelerated implementation mainly by central ministries at field level. Such projects were larger in scale than those of 1982437 and tended to require greater technical resources. Several districts experienced initial difficulty in implementation due to personnel shortages. Women’s participation in the programme remained high. However, since remote area populations were not provided with supplementary food (in contrast to 1982%87), one concern was whether these groups would be able to obtain adequate access to public works projects. A related question was whether the reliance on projects included in the national plan would result in an adequate geographical coverage of villages. Recognizing these concerns, and in response to local pressure, the government rapidly increased its financial allocations to districts for local initiation of public works as the 1992 relief campaign progressed. The programme eventually did provide short-term income earning opportunities to large numbers 296

Food Policy 1994 Volume

19 Number

3


of people (an estimated 90 000), but the extent of access gained families and remote communities has yet to be closely assessed.

by very poor

Agricultural relief and recovery. The agricultural relief and recovery programme has been the most expensive (in terms of finance allocation) and also most controversial component of drought relief. The initial agricultural relief programme of 198445, which provided subsidies on limited amounts of inputs and hectarages, was expanded into the Accelerated Rainfed Arable Programme (ARAP) in 1985/86. This provided an increased scale of support per farmer as well as a greater range of subsidies covering ploughing, row planting, field clearance, weeding, seed, and (under ARAP) fertilizer. These measures were, arguably, successful in enabling many small-scale farming families to retain their links with the arable sector in the face of repeated harvest failures and in promoting a marked recovery of crop production in 1987-89 when good rains returned. However, the 1990 evaluation raised strong doubts as to their targeting efficiency, particularly for the ploughing subsidies, which conferred large benefits on a small number of tractor owners (Hay, 1990; Amis, 1990; Cogill, 1990; MOH, 1992). Associated write-offs by government of agricultural loans also benefited a very limited group (ibid.). The livestock relief component provided free botulism vaccine to all cattle, subsidized livestock feed, and subsidies for rehabilitation and reticulation of boreholes. None of these measures were targeted to specific groups of farmers, and the latter components tended to benefit better-capitalized herd owners. The 1992-93 agricultural relief component was less comprehensive than those of the 1980s. The targeting aspect, however, remained very weak. Despite the call in the national plan for stricter targeting, the programme included the provision of free botulism vaccine for all cattle, although not additional subsidies on stock feed; free seed for all farmers for up to 5 hectares; and a ploughing subsidy for all farmers irrespective of size of landholding. This reflected a continued (albeit declining) financial ability to provide large subsidies and also suggested an implicit goal to maximize food production even at uneconomically high costs. On the other hand, the context for such subsidies includes the absence of a district-level information base by which to target resource-poor farmers, in addition to the management and political problems likely to be associated with selectivity. Such an information base would also assist in monitoring and evaluating the impact and efficiencies of agricultural subsidies. Human water relief. This programme provides supplementary funds to District Councils and the national Department of Water Affairs for provision of emergency human drinking water supplies (in addition to the regular non-drought water supply programme). It includes acceleration of borehole drilling and repair projects and distribution of water to remote and severely affected areas using water bowsers. This is a further example of a programme which builds on existing institutional capacities and regular development activities to address needs intensified by drought.

Conclusions Timely monitoring has been central to largely effective programmes for reducing the impact of drought on vulnerable population groups in Botswana. Strong institutional coordination has been an important contributory factor. However, it Food Policy 1994 Volume 19 Number 3

297


has been suggested that Botswana’s performance may be due as much to a ‘culture of information’ among decision makers as to the quality of the indicators used (Cogill, personal communication). Further, considerable weaknesses and gaps remain in the current monitoring system, particularly as regards indicators of household economic access to food, health status, water access, relief programme performance and the situation of very poor families especially in remote areas. With better micro-level information, targeting of both drought response and ongoing income support programmes could be improved and national resources used more equitably. The role of a food security monitoring system based on local and household level information -with possible community involvement in analysis - could be crucial in this respect. Such a system is now being officially contemplated (Government of Botswana, 1991). However, given capacity constraints, its effective implementation may require several years - particularly at local levels, where use of information remains weak in general. In the interim, there is need for assessment of existing procedures and criteria through which access is gained to support programmes such as supplementary feeding, public works, destitute grants and agricultural subsidies, from the point of view of improving targeting efficiency of those most in need, as well as other considerations. A review of this kind was recommended in the 1990 evaluation and accepted, in essence, in the government’s response thereto, as an element in the creation of a coordinated and flexible income support system, but has not yet been undertaken. When it is, particular focus will be needed on the access to relief and income support programmes of the poorest socio-economic groups and those in remote areas; and on how this can be better monitored. Closer linkage of drought relief to long-term development remains an important issue, recognized at the level of policy but not yet strongly translated into practice. Promotion of new sources of productive employment in rural areas is one of the keys, but lack of technical capacity and priority for detailed pre-planning of labour based public works oriented towards production continue to pose constraints. Botswana’s more general success during the 1980s in promoting private sector investment and employment in the towns and larger villages has gone some way to reduce the vulnerability to drought of significant numbers of rural families. There remains a large core of rural, often female-headed and/or remotely-located households with little or no direct access to employment and cattle; such families depend on sources of livelihood which are highly precarious, especially in times of drought. The climatic vulnerability of these populations has not yet been effectively addressed by development programmes, while the targeting efficiency of drought relief measures in respect of their short-run needs has been variable. While most other African countries are unlikely to be able to adopt as extensive a many Southern African countries did relief programme as that of Botswana, implement extensive measures in 1992-93 with a greater degree of donor support. Relevant lessons can be drawn from general features of Botswana’s programme, such as the importance of: a clearly stated political commitment; sustained attention to institutional coordination; reasonably decentralized management; and the building up and use of a well-coordinated multisectoral information system with strong sharing of data. More particularly, the history of drought relief in Botswana over the past 15 years has demonstrated: l

a clear desire to learn from experience, at the practical administrative level and through re-assessment and modification of policy for drought response on the basis of detailed evaluations and wide-ranging discussions;

298



the importance of a relatively democratic and ‘open’ socio-political context, to which regular elections (with a largely rural electorate) and an active private press are central, in prompting political concern and official response to drought; and of the existence of a reasonably efficient civil service, significant administrative capability at local government level, and a healthy budgetary position allowing rapid initial commitment of national resources for response; the potential effectiveness of a food security and nutrition monitoring system, particularly within the favourable context of the factors above, in providing at least broad guidance for drought management and as a basis - in conjunction with other forma1 and informal information sources - for monitoring (Davies et al., 1991).

References Amis, P (1990) Financial Efficiencies in Drought Relief, Study Paper 3, Food Studies Group, Oxford Buchanan-Smith, M (1990) Drought, Income Transfers and the Rural Household Economy Study Paper 2, Food Studies Group, Oxford Central Statistics Office (CSO) (1989) A Poverty Datum Line for Botswana, Government of Botswana, Gaboronc Cogill, B (1990) ‘Information systems and the drought relief programme’ Study Paper 5, Food Studies Group, Oxford Davies, S, Buchanan-Smith. M and Lambert, R (1991) Early Warning in the Sahel and Horn of Africa: the State of the Art: A Review of the Literature, 1, Institute of Development Studies, University of Sussex Food Studies Group (1990) University of Oxford, Report on the Evaluation of the Drought Relief and Recovery Programme, 1982-90, 6. Prepared for Government of the Republic of Botswana, Government Printer, Gaborone Gooch, T and McDonald, J (1981) Evaluation of 1979180 Drought Relief Programme, 1981 and Evaluation of Labour Related Projects in Drought Relief and Development Gaborone, Botswana Government of Botswana (1991) Government Response to the Report on the Evaluation of the Drought Relief and Recovery Programme 1982-90, Government White Paper, Government Printer, Gaborone Harvey, C (1992) Botswana: Is the Economic Miracle Over. 7, Discussion Paper, Institute of Development Studies, University of Sussex Hay, R (1990) Major Conclusions and Recommendations Study Report. Food Studies Group, Oxford Helm, J D and Morgan, R (1985) ‘Coping with drought in Botswana: an African success’, J. of Modern African Studies, 23, 3, pp 463-482 Ministry of Agriculture (MOA) Botswana Agricultural Statistics Government Printer, Gaborone (various years) Ministry of Agriculture (MOA) (1991) Botswana’s Agricultural Policy: Critical Sectoral Issues and Future Strategy for Development, Government Printer, Gaborone Ministry of Finance and Development Planning (MFDP) (1975) Rural Income Distribution Survey, Government Printer, Gaborone Ministry of Finance and Development Planning (MFDP) (1985) Report on the National Food Strategy, Rural Development Council, Gaborone Ministry of Finance and Development Planning (MFDP) (1988) Household Income and Expenditure Survey 1985186, Central Statistics Office, Government Printer, Gaborone Ministry of Finance and Development Planning (MFDP) (1991) National Development Plan 7: IYYI-1997, Government Printer, Gaborone Ministry of Finance and Development Planning (MFDP) (1993) In the Best Interests of the Child: A National Programme of Action for the Children of Botswana, Gaborone, (forthcoming) Ministry of Health (MOH) (1992) Report on the Food and Nutrition Situation in Botswana (prepared for the International Conference on Nutrition), Gaborone Morgan, R (1985) ‘The development and application of a drought early warning system in Botswana’, Disasters 9, 1 Morgan, R (1986) ‘From drought relief to post-disaster recovery: the case of Botswana’ Disasters, 10, 1 Quinn, V, Cohen, M, Mason, J and Kgosidintsi, B N (1988) ‘Crisis-proofing the economy: the response of Botswana to economic recession and drought’ in Cornia, G A, Jolly, R and Stewart, F (eds) Adjustment With a Human Face: Ten Country Case Studies, UNICEF, Clarendon Press, Oxford Food Policy I994 Volume 19 Number 3

299

Food security and nutrition monitoring in Botswana: K. Belbase and R. Morgan Sandford, S (1977) Dealing With Drought and Livestock in Botswana, Gaborone, Botswana. While commissioned to look at the livestock sector, this report stressed the need for a drought relief strategy addressing the effects on both humans and animals Simmons, C and Lyons S (1992) ‘Rhetoric and reality: the management of Botswana’s 1982-88 drought relief programme’ J. of International Development 4, 6 UNICEF/Government of Botswana (1989) Children, Women and Development in Botswana: A Situation Analysis, Gaborone United Nations Children’s Fund (UNICEF) (1990) Strategy for Improved Nutrition of Children and Women in Developing Countries, New York United Nations Children’s Fund (UNICEF) (1993) State of the World’s Children, New York

300

Food Policy 1994 Volume I9 Number 3