- 63 - CROP PRODUCTION AND FOOD SUPPLY TRENDS IN BOTSWANA Introductial When Botswana attained political self determination twenty years ago, the country \\IaS just emerging fran a prolonged and devastating drought. ~nty years later, the country is alJoc>st emerging fran yet another prolonged drought with equally devastating effects on agricultural production. 'Ibis time around, the effects of drought have not only been fel t in Botswana, but have affected al1lr:>st all the countries of fJUlrSaharan Africa, drawing world attention to the resultant food crises and the \\laY these countries have managedtheir ecx:nani.esduring the two decades of political independence. S\JlrSaharan Africa has been performing exceptionally poorly in food production canpared to other world regions. POor performance, together with deterioration in terms of trade in Africa's major COIIIII:Xiities(agricultural), runa\\laY debts and ba1ance-of-payments problems, unfavwrable political develqments, and high rates of pJp.1lation grOtth have created the severe food slDrtages experienced lately in the region. 'Ibis tragic situation has forced the governments of sub-Saharan Africa to re-assess and re-define in policy-terms, the role that agriculture in general, and food production in particular, should play in their national developuent efforts [see 0lW PLAN of Actial; SAIXX: Project Papers al Food and Agriculture; World Bank, 1981 Report; FAD, 1981]. - 64 - In Botswana, the food crises has not had the tragic consequences such as experienced in fot:>zambique,Ethiopia, the Sahel ca.mtries, etc. But the natural environment and the way it has been managed might suggest that this camtry's food situation is not far fran reaching tragic crisis level. For one thing, the ca.mtry experiences a regular cycle of droughts, [the 1\DStsevere occurring every twenty years or so (Botswana Society, 1979 )], and is thus seriously handicawed in food production. Ebr another, the soils are one of the world's poorest and mat unsuited for arable production. In addition, LUlI'eliablerains make it generally unecx:>nani.c to improve the productivity of the soil through the use of fertilizers (Sims, 1982: Lightfoot, 1982). Furthernore, the natural resource base for food production has deteriorated drastically C/Iler the years - particularly since independence, largely through neglect am mismanagement. Until very recently, crop production was never a priority in the .country's developnent efforts. '!his paper will review food production tren:is of the past twenty years and assess l1cM Botswana has coped with (a) drought coniitions (b) the pressures of developing a teclmologica1ly backward arable sector, (c) the CCIlpeting demarrls of an export oriented beef irxlustry and daoestic oriented crop sector, and (d) the c:x:nflicting deman'ls of econani.c grCMth and distrib.1tional equity. 'lhe major aim of this review is to assess the strengths and weaknesses of developnent efforts in order to identify areas for future intervention. Food Crq> Prod.uction and ~ I8 Food SUA:>1~ In ~ with lII:lSt develop' , lng coo.ntrles, Botswana derives the OOl.k (83 percent) f ' o her dietary energy fran crq>s and other vegetable fOods. Cereal a1 ' s one contnb.1t.e 56 percent of calories, with 49 , percent of thlS deriving fran maize and sorghum (EW), Production Year n.....1..) , • UVUII.. For the majority of the populatlon In Botswana "'--" , , baalS. Available data ' "'uuu 18 produced on a subsistence decades (e ' suggest, however, that OYer the past two xcept ln only two years) the OOOntry has never - 65 - prcrluced enough to satisfy the dietary needs of its popJl.ation. ArXl. that on average, it has at best Illanaged to prcrluce 50 percent of its cereal requirements. Even anong the farming hooseho1ds, only a tiny proportion actually ever prcrluce enough for their subeisteoce needs. Under favourable ccnditions, the majority of these farming households harvest a crop output of 250kg to 600kg per farm. But in order to meet their dietary needs, they sln.t1d produce at least 1260kg per farm (Alversal, 1979~ 01ard et a1,1982) '!be large production deficit means that, as Otzen (1981) and Grivetti (1981) observed in their village case studies, IIOStof the l1ouselx>lds are usually out of SUWly for up to six Dalths after harvesting their crops. '!he problem is further exacerbated by the fact that prcrluction is characterized by large f1uctua.tions. Since the 1960 sI drought, for instance, production of cereals has reached peaks of up to three times the 1961~5 output average. '!here was an 8 percent drop after the 1976 peak, a rise of up to 50 percent of the 1976 peak, then another plunge during the current drought to alxlut 1/6 of 1976 (Botswana Agricultural Statistics). Although these output variations can be directly attril:x.1ted to fluctuations in rainfall, they are also reflective, to sane degree, of hlJllal'l resrrmses to both environmental and socio-ecc:n:mic ccnlitions (sudl as lack of an effecti ve food policy). Botswana I s Iilysical ecology is very harsh and quite lU1Sui tad to crop farming. 1bo'eVer, Batswana have been producing food under these Ul1CXIIlpranisingccnditions for quite sane time. Available evidence suggests that fran the 1920's until iOOependence, dcmestic prcrluction was fairly near self-sufficiency (98 percent) except in times of drought (~, 1981~ Heisey, 1985). IiJwever, since inieperxience and the 19608 drought, self-sufficiency has never risen atove SO percent except in only two years. Botswana has therefore been ocntil'llXlUSlY receiving food aid since indeperdence. ecaercial iDpXts have - 66 - also been rising in relation to increasing demand for food. Developnents since indeperoence have tended to favour li vestock and export beef illprovements at the expense of food crcp production, even during the 1966-78 period when, by Botswana standards, the rainfall regiBes were reasonably l'X)nnal. 'Ibe crop sector has on average been attracting only a fifth of devel.opDent funds channelled into agriculture. Crop producer prices have also been fixed at 50 percent to 60 percent below world market prices (Botswana llgricultural Mem:>, 1980). By contrast, livestock (beef) production has enjoyed guaranteed and higher world market prices via the European F.cx:manic Q:mmmity, and has been generoosly suOOidized in the form of infrastructure, credit facilities, and disease oontrol services. '!be incentives and market q:JlOrtuni ties for livestock have been so lucrative that it takes little if any efficiency to make canfortable profits in this sector. 'Ibis developoent has worked against crop production in that, with the rising profitability of ccmmercialized livestock production, ownership of cattle has increasingly cane to determine the level of producti vi ty in arable sector. Whereas formerly, ...'.th the mafisa system, farmers cxuld have had indirect access to cattle: in recent times, cattle aomers are less willing to have their livestock used as mafisa, preferring instead, to breed them for sale. '!his is particularly significant considering that a tiny proportion of farmers own a disproportionate percentage of the national herd. Studies dale since the late 1970's slx:lw that families with inadequate draught resources (ie. those oming less than 40 herd of cattle or none at all) tend to plough smaller tracts of land, plCUJh and plant late as a result of later accesS to draught pcwer, and therefore, harvest lower out~ts. 'lboSe with 40 or mre cattle tend to make the IIDSt of the short rainy seascn by PloughiB} and plantiB} early and on larger tracts of land - thereby ensuriB} themse1 ves higher out~t "during normal rainfall years. Fklwever, these studies also slx:lw that as a - 67 rule, BotswanaI s large farmers have cane to eschew cr~ farming in favour of livestock breeding (Qiell, 1980; Litschauer and Kelly, 1981; Opschoor, 1981; Gulbrandsen, 1984)• '!his means that farmers with the IIDst adequate resources and the greatest capacity to significantly alter the level of food cr~ output in this country do oot participate in cr~ production. Data sl1cM that within the freehold farming sector, less than 45 percent of the 360 ccmnercial farmers engage in crops. Fbr ccmnunal land tenure areas, data are not disaggretated enough for us to determine the pr~rtion of large farmers not participating in cr~ prcrluction. The 1975 Rural Incanes Distribution Survey iooicates that the richest 90 to 99 percent of rural household do not get any inccme frClll crops. FurtheIllOre, although annual agricultural statistics do not indicate which classes of farmers participate at what level, they do suggest that anong the country's 84,600 agricultural farmholds 22 percent are purely cattle or cattle and srnallstock owners. Another six percent keep smallstock only: but these are unlikely to be rich farmers. Despite their leM participation in crop farming, the few large farmers who do engage in arable production make a disporportionately high contrirotion to total food output. Since 1980 for instance, ccmnercial farms have contriroted between four percent and 15 percent of total sorghum output fran one to three percent of the total land planted with sorghum. In that period, these farms accounted for less than one percent of all the farms planting sorghum. During the same period, camnercial farms made up less than one percent of farms planting maize accounted for seven percent of the maize land, rot produced 24 to 72 percent of the total maize output (Botswana Agricultural Statistics). Within the camnunal land tenure sector during the same period, farms over 10 hectares made up tlr>O to four percent of the sorghum farms: accounted for 12 to 30 percent of the sorghum 1arxi and contributed 16 percent of the total sorghumoutput. - 68 - Similarly with maize, the big farms made up t~ percent of the maize farms, five to 25 percent of the area planted with maize, and prcrluced up to 82 percent of the total maize outplt produced in the tradi tiona! sector. Together the big farms aCCOlmtfor at least 1/4 of the sorghlDlloutput and at least 60 percent of the maize output produced in the 1980s. For the earlier period, data are not available on the oontrihltions of these large farms to food output. '!he earliest available data are for 1971 when camnercial farms aCCOlmtedfor one percent of sorghlDllproduction and t~ percent of maize output. In terms of trends, it is difficult to see whether or not these farms are becaning the daninant prcrlucers of maize/sorghlIDl because of the paucity of data for the 1960s to mid 19705. Ha.1ever, Heisey (1985) suggested that there might be a general trend tcwards the big farmers daninating production. In the light of recent policy developnents on food, this Phenanenon is likely to becane DOre afPCU"ent(the National Food Strategy llCM puts greater emphasis on the developnent of the camnercial/large-scale farms to meet the food demandof the danestic market). On the wrole, although the big farms :bave considerable resources to further improve the quali ty of their crop production, their investment in arable farming has been relatively low. In an econanic analysis of crop prcrluction based on the 1975 Rural Incanes Distriootion survey (RIOO), Lucas (1981) noted, for instance, that despite the bigger larii areas allocated to large farmers (i.e larger cattle omers) there was a weak upward trend in yield per unit of land with cattle omership up to 80 herd, followed by a sharp deCline amng the biggest cattle ~rs. He thought that this decline might be explained in terms of the extensive farming techniques used by the big farmers. '1he 1980 !k)tswana Agricultural Sector MeRr:>randum also ooted that farmers generally failed to adept the technOlogical innovaticns and high yielding inputs advanced by the - 69 - goverrunent's Department of Agricultural Research. Other studies indicate that large farmers may not even incur cash costs for their cropping acti vi ties (Galbramsen, 1980; Opschoor, 1981) since they usually produce ernlgh cr~ for seeds as well as consUllption and sale. In the nore recent years, ~ver, large farmers have becane nore active in maize production where they are inproving their investment levels (e.g. roying notorized farming technology like tractors, harrows, etc). Although data are very scarce, it is nost probable that in the earlier developnent programmes for the arable sector the large farmers in the tradi tional sector were the chief beneficiaries. Even in recent years with the Arable Larrls Developnent Programme, a good percentage of the participants are large farmers with over 10 hectares and with over 40 head of cattle [See l\LDEP evaluation report]. 'Ihere is, hewever, still much roan for developnent in terms of yield rates in the large-farm sector. 'Ihe 1981-84 agricultural statistics shew that yield rates still generally decline with the increase in land area, rising a bit anong farmers over 10 hectares. In 1981 for instance the smallest farms had an average yield of 244kg per harvested hectare ccupared to 347kg anong farms over 10 hectares. 'n1e large farms I outplt superiori ty is nore a function of lard area cul ti vated than productivity per unit of land. It is estimated that it is technically possible, urrler prevailing ecological conditions in Botswana, to achieve yields of tM.O tons per hectare. But this yield does rot seem to have been achieved even in the camnercial farm sector. 'lbese farms normally produce sorghum yields 2/5 greater than th>se of the large traditional farms, and maize yields 2/3 higher. However, the best average yield rate they have so far achieved has been one ton per hectare of maize (Botswana h;Jric. Statistics). - 70 - In tenns of household food availability, the proo.uction trends in the small-farm sector are particularly significant as this sector accounts for over 90% of crop farming households. Indicators suggest that although the number of farming households has doubled over the past two decades, the level of food output has not changed. Opschoor (1981) estimates that average yield rates were arot.1Irl 270kg per hectare (90kg per acre) in the 1930s and 1960s, and arot.1Irl 270kg per hectare (weighted) between 1970 and 1980. CUrrent. data suggest 250kg per hectare is still the average yield rate, rot this has drq:.ped since the deepening of the current drought. '!he implication of this stagnant level of production is that Botswana s I farmers have not improved their farming technology in line with the level of output required by their grCMing numbers and the camnercial econanic setting of IOCrlem Botswana. Yet, since the early 1960s efforts have been made by goVerrutEI1tto encourage investment in improved technology and farming practices. Up till 1974, developnent thinking and policy direction in Botswana (as in other parts of the world then) tended to focus on encouraging a minority of resource-rich subsistence farmers to adopt new and higher-yielding technologies, and to act as IIOdels for other, less progressive and lesser-endowed to emulate. Since 1974, there has been a gradual shift towards broadening access to new technology to a wider spectnun of subsistence farmers. By 1979, GoverrutEI1t had begun to initiate developnent programnes specifically geared towards ameliorating the production constraints of the resource-poor farmers who form the majority of arable producers. '!he Arable Lands Developnent Prograume (ALDEP), in particular, has been the major effort towrrds developing the small farm arable sector, with the Accelerated Rainfed Arable Project, and Drought Relief Programnes carplementing it. ALDEPwas initiated to directly benefit SCIlIe 11,000 small subsistence farmers (17 percent of Botswana's crop farms) - 71 - during its first phase of five years. 'n1e project was meant to upgrade the farming teclmologies of farms by improving their supply of draught power and providing basic implements, fencing and on-farm water supply on a graded and selective basis determined by the farmers' econanic circumstances. The project proposed no change in the basic method of land preparation and cropping patterns other than to advocate for change in the investment pattern. It was envisaged that if ploughing were done earlier as a result of this investment in on-farm water and farming inputs, planting would then be timed to make cptimum use of the scarce rain water resources. These changes were expected to bring noticeable changes in the level of total output. At one level, increases in yields per unit of land I«lUl.dbe brought about by timely ploughing and planting and, for sane, through rON-planting and use of fertilizers. At another level, increases would be gained through the expansion of cultivated land. Each farming IIOdel would be able thus to increase its output by lIVre than 50 percent so that at full-term of the Project developnent, the 11,000 farmers would have increased crcp output fran 595 tonnes during the first year to 21,581 tonnes (ALDEPAppraisal Report, 1981). Despite available data on rainfall forecasts (see 1979 Sympositunon Drought in Botswana) ALOEP's planners failed to take account of the possibility of drought, so that the project tended to be premised on unrealistically favourable rainfall patterns. In the event, the production goals set for the , 'dence of Project were unrealised largely because of the C01IlCl its implementation with the drought. '!he Project has also been overtaken by events because policy has DON shifted away fran an essentially equity-oriented thrust to a national food-security focus. 'lhe shift in policy, while still retaining the equity orientation of N.DFP has '[0# plt medium and large scale producers at the forefront of the , ct camtry's strategy for food production developnent. The 1.IIP - 72 - of this shift, and the resultant re-direction of national resources, have still to be assessed in terms of their impact on food prcductivity and security at both national and household levels. SuIImary and Qmclusion Statistical data on Botswana's food production over the first two decades of politiCal independence are not consistent and reliable enough to provide discernible trends. However, other indicators such as stagnant yields, tile persistent and high volume of food aid imports, the increasing volume of camnercial food imports, the rising proportion of the non-agricultural population, etc, taken tCXJether, suggest low productivity and an increasing prcduction deficit. Botswana's physical environment makes agricultural production a particularity risky venture, and developnent of the sector particularly challenging. However, there are indications that there is scope for increasing land producti vi ty and food crop output, and that this potential has yet to be realized (See Lightfoot, 1982 on the potential for traditional agricultural technology; and reports fran agricultural research projects). Governmentpolicy for realizing ~~is perceived potential has been to identify prcducers with the capacity to adapt new, ostensibly higher yielding teclmolCXJies, and to back them up with research, marketing channels, subsidies and credit facili ties. 'Ihese target prcducers have varied over the years fran "prCXJressive farmers", to "small farmers", and currently, to the multiple class categories of small-rainfed, medium and large canmercial farmers; the target group changing mainly in line with other objectives besides increased production that the policy aimed to achieve. In terms of cq>ing with the physical environment its only towards the secorxl decade of independence that policy begins systematically to makeattempts to help producers cope with the drought. Drought relief programmesand the Accelerated Rainfed - 73 - Agricultural Programmefor instance are meant to ameliorate the effects of drought and help with recovery. Research efforts have also focussed on improving tradi tional technology to increase yields. '!he mechanisms for evaluating the impact of policy on proouction at household or farm level were not quite in place during the period urrler review. Evaluations reports on ALDEP and its impact on the other hand have only IlOi begun to illprove in reliability. '!here is also a dearth of empirical data on the nature and capacity of the various classes of focXl producers to resporrl to incentives. Social Science Iiterature would suggest however, a general lack of response to new technology. But without systematic evaluation of the ecooaniC8 of such technology and the nature of the response of those participating in particular programmes, conclusions on responsE' remain largely intuitive and too general. Alverson, H. 1979 Agricultural DeVelo~nt in Botswana: Targets and Constraints rm, Gaoorone). Botswana Government 1981 mraisal Report - ALDEPPhase r Project ADB/lFADjoint Programme). Botswana Agricultural Statistics (several years) 1985 White Paper N£ 2 1985 on National Strategy. Botswana Society in Botswana (Botswana 1979 Symposium on Dr~t Society: Galx>rone• Production Year Booi': 1981 Agriculture, 'I'c:1riardsYear-2OOQ (Rane). 1982 Preparation of Project Proposals for the Establishment of an Early Warning System for Regional Food Securi ty (SALIX ca.mtries NJ:TCP/RAF/2')J)2Mission RelX'rt, RaIle). Grivetti, L.E. 1981 "GeograIitical Location, Climate and weather, ~ P-Bgis; Aspects of Agricultural Success 1n Eastern Kalahari, Botswana" in ~ Science InfrClllCl.tion, Vol 29, No.3. Gulbrandsen, 0 1984 "Access to Agricultural land and eanmercial Land Management in Eastern Botswana II (Ministry of Local Government and Lands, Galx>rone) • Lightfoot, C. 1982 "Agricul tural Research for Developnent of Small Farmers: A Closer IDok at Traditional Technology" in Botswana Notes and RecOrds Vol. 14). 1981 "lages Plan of Action for the F.cOrXJIIic Developoont of Africa, 1980 - 2000 (ClTill, II InternatiOnal Institute for labour Studies Geneva) • Odell, M. 1980 "Planning for Agriculture in Botswana: A Report on the Arable Lands survey" (UM/Ministry of Agriculture, Gaborone). otzen, u. 1981 "Nutri tiona! Needs Oriented Rural Devel<:pnent Planning for Botswana" in Ouartely Journal. of International ~icul ture Vol. 20 No 3). - 75 - Oland, K., et a1 1980 Target for Agricultural Developnent in Botswana Government Printer. Gaoorone. ~choor, J. 1981 "Crops, Class and Climate: Environmental and Econominc Contraints and Potential of Crq> Production in Botswana" in M.A. Oamen et aI, (oos) Botswana sI Economy Since Independence M;j3raw-Hi11, NewDelhi. 1984 "Proceedings of the 1984 Annual Conference, Lusaka". Sims, D. 1981 "Agroclimatological Information, Crop Requirements and Agricultural ZOnes for Botswana" Ministry of AgriCUlture. Gaoorone. Srivastava, P. & s. Matswe 1983 Report on Field visit of ALDEPr-DnitoriB;} Officers September to November 1983 N.DEP M:mitoring and Evaluation Unit, GabOrone. World Bank Acce1eratoo Developnent in Sub-Saharan 1981 Africa.