4.4.1 General aspects of household economy
4.4.2 Economic comparisons among pastoral systems
4.4.3 Livestock supply to markets
4.4.4 Traditional marketing rationale
4.4.5 Marketing efficiency
4.4.6 Market evolution
4.4.7 Productivity comparisons among systems
4.4.8 Labour and the encampment
4.4.9 Milk processing
4.4.10 Dairy marketing
Gross income of household is the sum of subsistence plus marketed production as derived from the eight-cow production unit of Cossins and Upton (1987). Hence the annual gross income of an average Borana household in an average rainfall year in the early 1980s was about EB 2000. Except for trace amounts of income from sales of things like firewood, incense or handicrafts, virtually all of this income is from livestock, specifically cattle. The calculations considered the values of marketed animals (EB 552), marketed milk (EB 71), consumed milk (EB 721), six new calves (EB 600) plus other useful animal products such as hides (EB 50) for a grand total of EB 1994. Roughly 40% of gross revenue is derived from milk and the remainder from live animals or meat.
It is likely that less than 1 % of animal outputs are used as crop inputs. For example, there is apparently no effort made to manually apply the large quantities of manure that have accumulated at encampments, which is a common practice with other low-input extensive farming systems (Massey, 1987: p 93). The main interactions between crops and livestock in Borana occur when oxen or camels are sometimes used to plough fields (direct linkage), when crop residues are fed to calves in dry seasons (direct linkage) and when small amounts of cash from sales of livestock products are used to purchase seeds (indirect linkage).
In sum, this is all interpreted to show that the vast majority of the Boran still operate a largely pastoral, rather than agropastoral economy (considering the criteria of McIntire and Gryseels, 1987; R. T. Wilson, ILCA, unpublished data). Nearly all of their food and income is ultimately derived from livestock. Cash remittances from relatives outside the system were virtually nonexistent in average rainfall years during the early 1980s (Cossins and Upton, 1987: p 212). The preliminary findings of AGROTEC/CRG/SEDES Associates (1974f: pp 70-74) are similar to many reported here with other details provided. However, it is interesting to note that the mean cash income of Borana households in 1972 (EB 86) reported by AGROTEC/CRG/SEDES Associates (1974f: p 67) is only 14% of the mean of (EB 623) for 1983. Assuming the data are accurate, this may reflect changes in the prices of commodities as well as a previously lower need for market involvement (see Section 7.2: A theory of local system dynamics). To this day the majority of Boran have apparently little need or opportunity for integrating livestock and crops. Such integration is commonly considered to be slow and uncertain under semi-arid conditions (McIntire and Gryseels, 1987: p 239). Possible development opportunities to facilitate future crop/livestock integration in Borana are presented in Chapter 7: Development-intervention concepts.
There are conflicting reports as to whether cultivation has been an important aspect of the Borana economy traditionally. On one hand, Asmarom Legesse (1973) reported that the Boran during the early 1960s had only negative feeling towards farmers and farming and this implied that the Boran would never engage in cultivation unless faced with no other alternative. In contrast, Negussie Tilahun (1984: p 6) noted that cultivation has occurred in the Borana Plateau for over 40 years and that farmed acreage expanded after a drought in the mid-1970s.
The widespread potential of cultivation in the Borana Plateau was first reported by AGROTEC/CRG/SEDES Associates (1974e: pp 24, 27) where it was stated that "one third of the breeders practice cultivation, but only on a very small scale." It is likely, however, that this estimate was biased on the high side since their surveys focused on households residing near roads and towns. Even today these areas remain the dominant centres for cultivation due to favourable local environments and proximity to influences of migrant highland farmers (Hodgson, 1990; Assefa Eshete et al, 1987). A drought in the early 1970s was cited as a factor in the spread of farming on the Borana Plateau (AGROTEC/CRG/SEDES Associates, 1974e).
Assefa Eshete et al (1987: p 12) noted that it may be possible to cultivate about 30% (4600 km2) of the study area with 21 % of this taking place in valleys and the remainder above 1650 m elevation. Given this figure of 4600 km2 as the maximum possible to cultivate, about 5% of this total may have been cultivated during 1986. Assuming estimates of Cossins and Upton (1988b: pp 272-273) to be correct, the increase of 176 km2 in cropped area during the mid-1980s was actually achieved in only two years after the 1983-84 drought. If it is assumed that another 90 km2 could be brought under cultivation each year, it would still take about 48 years to use all the arable land. This is, of course, highly speculative and would vary depending on population growth (in towns as well as the rangelands), cultural commitment and the economic need to farm, uptake of animal traction which could double the rate of cropland expansion (Pingali, 1987: p 11), land use and agricultural policies, the creation of other nonagrarian economic options for people and long-term rainfall patterns.
It is also assumed in this scenario that all arable land is accessible for cultivation, which is not likely to be the case. Since crops need supervision to reduce losses from wildlife and thievery and encampments must be located within 16 km of the deep wells (Cossins and Upton,1987: p 203), arable sites outside such radii would probably not be used. While the long-term trend in the rangelands may be towards increases in cultivated area, over the short-term farming intensity would probably ebb and flow, largely depending on rainfall.
In more mesic situations where cropping carries less risk, pockets of agropastoralism have developed and will continue to expand. An example of such an area may be Dadim (10 km north of Yabelo), where a settled community of negligible numbers in 1973 had grown to over 4000 persons by 1990 (I. De Lange, Dadim Mission, personal communication). Local informants have speculated that poor pastoralists make up the majority of the new farmers that populate such villages on the edge of the rangelands today (D. L. Coppock, ILCA, personal observation).
Drought has been found to elicit at least a temporary reliance on cultivation by pastoralists until livestock productivity and numbers recover (Hogg, 1980; Toulmin, 1986: p 2; Moris, 1988: p 276). However, whether or not farming persists and is intensified may depend to a large degree on population pressure (Boserup, 1965: pp 63-64). Under recent conditions of restricted resources farming can persist in a pastoral society because a mixed system has greater proven flexibility and resilience (Campbell, 1984: p 55; Massey, 1987: p 83; Johnson et al, 1989). Droughts todate in the Borana system usually kill far more livestock than people (see Chapter 6: Effects of drought and traditional tactics for drought mitigation) and thus should accelerate a decline in pastoral production and welfare in the face of increasing population density. Opportunistic cultivation is one of the few alternatives that pastoralists have to partially compensate for such a long-term trend (Evangelou, 1984: p 258). It may be illustrative that there were 3.2 TLU/person and 10 people/km2 in semi-arid Maasailand during 1980, a time when cultivation was considered to be a growing and permanent feature in the economic diversification of that society (Evangelou, 1984: pp 17-38).
Estimates for the entire 95000-km2 SORDU project area showed 4.1 TLU/person and 5.6 people/km2 in 1988 (Solomon Desta, nd; see Section 7.2: A theory of local system dynamics). In an aerial survey of the 15475 km2 central plateau, Milligan (1983) estimated about 2.3 TLU/person and 7.3 person/km2. The uneven distribution of water and the location of most encampments within 16 km of the deep wells (Cossins and Upton, 1987: p 203) indicates that local densities of livestock and people are considerably higher than the overall mean. If about 50% of the potential grazing area is actually used in the dry season (Milligan, 1983: p 28), then population pressure could be more on the order of 15 persons/km2. In other words, preconditions for a general shift to agropastoralism do in fact exist on the Borana Plateau today.
Drought can permanently turn poor pastoralists into farmers by the unfortunate depletion of their smaller herds; and can at least tum wealthier pastoralists into temporary or part-time farmers by reducing milk production and killing milk cows and their replacements. In such cases households would be forced to cultivate for a few years until herds recover their milk-production potential. This is the hypothesis behind the observation that several pastoral groups have abandoned cultivation a few years after a drought has ended (Hogg, 1980: Toulmin, 1986). It is believed that while cultivation used to be a sporadic response to drought for the Boran, it will now persist, even in average rainfall years. The agropastoral Gujji represent a neighbouring group that were once pure pastoralists (Asmarom Legesse, 1973).
Borana men are the heads of households and make the major decisions regarding production strategies and disposal of ruminant and equine livestock. Most of the money from the sale of these livestock apparently goes to men. Inheritance of livestock follows along male lines of descent also (Asmarom Legesse, 1973). Women, on the other hand, make the day-to-day decisions regarding milk offtake, management of calves, goat kids and lambs, and derive their cash income from sales of dairy products and possibily poultry (D. L. Coppock, ILCA, personal observation). Socio-economic patterns reported elsewhere in Africa support this generally inferior economic status of pastoral and agropastoral women but it has also been noted that these women may have more subtle rights and claims to livestock than is immediately apparent (Broch-Due et al, 1981; Wienpahl, 1984; Massey, 1987: pp 111, 171). This has not been studied in detail among the Boran to verify, so general conclusions need to be qualified. The picture is complicated by the fact that a significant proportion (perhaps 20%) of household heads among the Boran are now women. Many of the female household heads in the survey of Holden (1988) were in the poorest wealth class (S. J. Holden, ILCA, personal communication). It is reasonable to speculate that the incidence of female household heads may increase in peri-urban settings as a reflection of their frequent poverty status and thus their need to sell milk to purchase grain (Holder et al, 1991).
Some useful contrasts can be made between the Borana system and the Maasai System in Kenya because they are both in semi-arid environments, with the Maasai being considered further along a continuum of pastoral change (Evangelou, 1984: Solomon Bekure et al, 1991). Income wise the Borana cash budget is based on a 40:60 split between milk and meat, respectively, similar to the 46:54 split for Maasal families on group ranches (Evangelou, 1984: p 187). Compared to the Maasai system, however, the relative contribution of small ruminants to the household economy among the Boran seems very small (Evangelou, 1984: p 105). The Maasai have a higher population density and as a result appear to be shifting to various forms of livestock diversification and agropastoralism (Campbell, 1984; Evangelou, 1984).
Although livestock production forms the basis for the Borana economy, the opportunity to sell livestock and livestock products in order to purchase grain at favourable terms of trade is critical in providing the food that enables a higher density of people to live on the Borana Plateau. Over 35% of the annual energy requirements of households are met by purchasing nonpastoral foods (Cossins and Upton, 1987: p 213). Similarly, Upton (1986a: pp 21-22) estimated from data in Milligan (1983) that the 14.3 LSU/km2 on the central plateau yield 11154 MJ GE/year, which could support a human population density of 4.8/km2. This is only 68% of the measured density of 7/km2 in 1983. Considering information in Milligan (1983) and Upton (1986a) and assuming an annual growth rate of 2.5% in the human population (see Section 2.4.3: Human population growth) the Boran may have numbered 4.8/km2 back in 1966. Assuming that livestock productivity and density have been similar over the years of average rainfall, this suggests that the population has become dependent on imported grain for about 24 years or one generation. Maize was reported in the Borana diet in 1972 (AGROTEC/CRG/SEDES Associates, 1974f: p 73). Other population perspectives related to dietary shifts are reviewed in Section 7.2.1: Empirical modeling.
The combination of a higher density of people and increasing flows of grain at present means that dependence on grain will continue (and increase) as a permanent and growing feature of the Borana economy, depending on climate, population dynamics and related factors. Trade for grain has probably been going on in the area for centuries may have fluctuated with the fortunes and size of the human population (Wilding, 1985a). In one sense, while recent population growth may be viewed as a linear process, it is likely that there has been a high degree of fluctuation historically as a result of climate changes, migration, warfare and disease (Wilding, 1985a; see Section 7.2: A theory of local system dynamics). Various degrees of agropastoralism have probably come and gone over time. It is unclear if the present density of people in the study area has ever been equalled in the past, or whether the current trajectory of the production system is indeed unique.
Upton (1986a: pp 31-33) noted the key role of trade in providing subsistence as well as surplus in average rainfall years, and calculated that even if grain prices rose (or livestock prices fell) by up to 470%, trade would still be beneficial to the Boran. However, higher population densities of people become more vulnerable to less favourable terms of trade (Upton, 1986a: p 33). The subsidy effect of grain on pastoral population density has been noted elsewhere (Dahl and Hjort, 1976: p 155; Bates and Lees, 1977). Dahl and Hjort (1976: p 155-156) and Evangelou (1984: p 108) cited examples of a gradual increase in the proportion of agricultural produce in the Maasal diet over the past 40 years as one result of population pressure. In a 1990 survey of Borana elders, it was found that they believed regular grain consumption to be a relatively new aspect of their society (D. L. Coppock, ILCA, unpublished data).
Offtake rates (largely sales to few slaughters) were estimated to be 19% for cattle on the Borana Plateau in a normal rainfall year (Donaldson, 1986: pp 13, 57; Cossins and Upton, 1987: p 211). Cattle offtake rates were apparently higher during the drought as markets were saturated, but no reliable regional figures were obtained (see Section 6.3.1.1: Livestock dispersal and herd composition). An Offtake rate of 19% appears high in relation to some long-term estimates for other pastoral systems under conditions of average rainfall and traditional production objectives. Although the Borana system exhibits some symptoms of range degradation (see Section 3.3: Results) and the people are increasingly dependent on grain, during years of average rainfall grazing is still regarded as adequate (Cossins and Upton, 1987). Dahl and Hjort (1976: p 181) cited Morgan (1972) as saying that pastoral herds in overgrazed areas had higher Offtake of 10-12% versus those in moderately grazed areas (4-9%). Dahl and Hjort (1976: p 181) noted that such a shift in offtake does not reflect "progressiveness", but rather a reaction to low per capita productivity in the traditional system.
Evangelou (1984: p 109) quoted an average of 10% offtake for cattle by pastoralists versus 25% for commercial ranchers in Kenya. He also noted that cattle offtake for the Maasai system could vary from 10% (at times of post-drought herd building) to an average of 13 to 17% once recovery is well underway or complete. The higher rate (17%) is thus similar to estimates for Borana (19%) before the 1983-84 drought. During the 1973-74 drought in Maasailand cattle offtake rates jumped to 38% (Evangelou, 1984: p 109). Meadows and White (1979) saw this as an increased demand for cash that had to be met by the Maasai in order to buy nonpastoral food. Coppock (1992b) found that Borana elders expected livestock offtake to increase among them because of a growing need to buy grain and the changing attitudes among the younger generation towards buying consumer goods. Although sampling methods were unclear, the observation of AGROTEC/CRG/SEDES Associates (1974f: p 67) that the annual marketed offtake of livestock averaged 5.2% on the Borana Plateau in 1972 lends some credence to the stipulation that Offtake rates have gone up (see Section 7.2: A theory local system dynamics).
Negussie Tilahun (1984) found a high correlation between income and purchase of food and dry goods such as clothes. Income was derived from the sale of cattle and this increased with higher absolute holdings of animals. In one sense these findings are paradoxical: While it could be anticipated that increased wealth would result in a higher demand for some nonfood items, on the other hand wealthier families would be expected to have a higher ratio of milk cows per person and hence less of a need to purchase grain, regardless of season. Grandin and Solomon Bekure (1982) also noted that wealthier Maasai households sold a larger absolute number of animals, but that the rate of herd offtake was lower compared to poorer households. It is possible that some of the grain obtained through cattle sales by the wealthy is for impoverished relatives or friends, as redistribution of food is reportedly common among the Boran (Cossins and Upton, 1987).
Although the households studied by Negussie Tilahun (1984) on average showed a wide range of income and livestock holdings, compared to subsequent surveys (Donaldson, 1986; Holden, 1988) the sample appeared wealthier. Families were larger (over eight persons), several men had two wives and the ratio of 13.8 livestock unit/AAME is nearly three times the mean from more extensive work (Milligan, 1983). It is common that initial household studies in a pastoral society may be associated with the wealthier or more influential persons (D. L. Coppock, ILCA personal obser vation), so that if more very poor families had been included in the survey, correlation relationships may have been stronger. Negussie Tilahun (1984) speculated that offtake rate could be stimulated if the flow of appropriate consumer goods was increased to the southern rangelands. This was mentioned previously as a constraint by AGROTEC/CRG/SEDES Associates (1974f: p 70) and UNDP/RRC (1984: p 46). This topic is reviewed further in Section 7.3.3.6: Cattle marketing.
The market supply analysis confirmed the supreme role of cattle in the Borana cash economy (Dyce, 1987: pp 58-59): the number of cattle marketed was roughly double that for goats, eight times that for sheep and 80 times that for camels. The dominance of cattle was most pronounced at Negele and Agere Mariam to the north, which reflects the highland demand and the ability of cattle to be trekked long distances out of the system. Except for Moyale, where the supply of goats was greatest, all markets conformed to this species pattern. The location of Moyale on the Kenya border, where there may be a higher demand for goats as well as a "loss" of cattle to cross-border black-market trade, may best explain this pattern. In addition, the environment around Moyale may be more ecologically conducive to reefing small ruminants (Dyce, 1987: p 58).
Although the data base for the Boran only covered 35 months, there were some features that can be compared to other pastoral areas. Sandford (1983a: pp 208-210) asserted that compared to other systems, pastoral systems are more commonly characterised by high seasonal and annual fluctuations in the number of livestock marketed. Data were cited from Kenya (18 years) and Upper Volta (nine years) that showed annual coefficients of variation of 51 % and 31 %, respectively. Sandford also noted that a higher proportion of the cattle sold in pastoral systems consists of immature males and older females, and that less emphasis is placed on selling mature males ready for slaughter, although this may vary with environmental conditions. For example, if forage is abundant young stock may not be marketed. An example was shown from Herman (1983) who compared pastoral and nonpastoral systems in Upper Volta in 1976-77. Both systems were similar in that about three-fourths of all sales were male cattle, but they differed in that 46% of all sales in the pastoral system were males less than two years old. For the nonpastoral system the same group accounted for only 4% of all sales (Sandford, 1983a: p 210).
Dyce (1987: p 60) calculated an annual coefficient of variation for the Borana markets of 54%, which is in agreement with Sandford (1983a). There was no clear overall pattern suggesting that the Borana marketing flows varied by season or rainfall pattern; arguably the data set may have been too limited to address this question (Dyce, 1987: p 61). However, information from other re- searchers (Negussie Tilahun, 1984; Holden, 1988; Mulugeta Assefa, 1990), local informants (Coppock, 1992b) and government agents (Solomon Desta, TLDP economist, personal communication) and nongovemment purchasing agents (R. J. Hodgson, CARE-Ethiopia, personal communication) all indicate that the tendency is for the Boran to try to avoid selling cattle in wet periods when forage (and thus milk) production is high and to sell during dry periods of stress when they are forced to purchase grain. This supports the general observations that pastoralists principally sell animals to remedy a specific cash need that usually appears in a dry season or drought (Dahl and Hjort, 1976: p 180; Meadows and White, 1979; Cossins and Upton, 1988b: p 255).
In terms of the sex and age composition of marketed stock, Dyce (1987: p 61) noted that the most frequently marketed cattle on the Borana Plateau were mature males; this was consistent with patterns in the Sahel reported by Herman (1983) who said that the typical marketed animal was a male four to seven years old (note the contrast to the earlier citation of Herman (1983) by Sandford (1983a)). Data from a small sample collected by Evangelou (1984: pp 216, 228) showed mature males made up 56% (N = 203) and 67% (N = 177) of cattle marketed in Maasailand during 1980. Likewise except for one market at Hidi Lola, there was no indication in the southern rangelands to suggest that immature cattle played a significant role in the supply of cattle to formal markets. This is in contrast to speculation by Sandford (1983a: p 210), although sale of young calves may have been greater among producers at innermost bush markets far from town; (Cossins and Upton, 1987: p 211). Their under-representation at larger market centers may have been due to their lower ability to walk long distances. Other factors contravening the sale of immatures may include: (1) a low per kilogram price; (2) the need for milk by delaying weaning for over a year in many cases; and (3) maximising income by selling a mature animal. Nicholson (1983b: pp 27, 37) noted the highest price/kilogram was achieved for animals over 282 kg.
Wealth differences also affect livestock-marketing patterns. Informants reported that a poor herd owner does not have the choice to wait for an animal to reach optimum weight before marketing if he has to buy grain to survive (Coppock, 1992b). Thus, it is usually the poor herd owner who must sell an immature animal. If more immature cattle appear on local markets, it is assumed that this is caused by more households becoming poorer (Coppock, 1992b). As first noted by Jahnke (1982: p 92), Dyce (1987: p 61) mentioned that the low and variable market supply of immature cattle in Borana would be a major constraint to the development of an appropriately stratified beef industry in which, for example, immatures are taken off and fattened in higher potential environments (von Kaufmann, 1976: pp 273-274; Pratt and Gwynne, 1977: p 140; Jahnke, 1982: p 91; UNDP/RRC, 1984: p 46). This perspective fits the Boran especially in light of their resistance to sell younger stock (see next section).
Although sales of goats and sheep were secondary to the Borana economy in the early 1980s, they may have become more important recently because of the increased activity of government buyers collecting animals for export starting in 1984 (Cossins and Upton, 1987: p 211). This can quadruple the price of mature male sheep, although variation in foreign demand greatly alters local marketing opportunities (D. L. Coppock, ILCA, personal observation). It has been suggested that the Boran are diversifying their holdings to include more small ruminants (Belete Dessalegn, 1985: p 26). Although this hypothesis remains untested, it would be consistent with a stress response manifest when cattle pastoralists are increasingly pauperised by high population densities and/or drought (Evangelou, 1984: p 35). The diversification into small ruminants may also indicate an interest to use the latter as saleable substitutes for cattle in light of possible increased pressure to purchase grain. The sale of small ruminants does not cut into the herd capital as deeply as would the sale of cattle, nor does it require the consensus of relatives as is the case with cattle (Coppock, 1992b).
Camels were almost negligible in the market analysis of Dyce (1987). The only reliable markets to find a few camels for sale are to the south at Moyale and to the east at Negele (Hodgson, 1990: p 123). Facilitating development of camels among the Boran is discussed in Section 7.3.3.2: Camels, donkeys and small ruminants.
The interview data used to interpret pastoral rationale are limited in that they may reveal how people think they behave rather than how they actually behave (S. Sandford, ILCA, personal communication). The results are interpreted, however, as offering important explanatory insights into herd accumulation and animal marketing behaviour that would be very difficult to obtain in a more objective fashion. That the Boran sell animals to satisfy acute cash needs, do so at sub-optimal times of the year in light of the seasonal terms of trade, and prefer to sell certain classes of cattle (older males) that tend to be different from those sold in developed animal production systems (immature males) are important findings consistent with empirical observations (Doran et al, 1979: see reviews in Sandford, 1983a; Dyce, 1987). This preference of the Boran to sell males that have completed growth explains the difficulties purchasing agents have had over the years attempting to buy younger males for ranching and stocker/feeder programmes (Solomon Desta, TLDP economist, personal communication). If more immature cattle are seen in markets in the future, it implies that the people are becoming poorer and have fewer mature males for sale (see Section 7.2: A theory of local system dynamics) and should not be erroneously interpreted as a shift towards "more progressive" marketing behaviour by the Boran consistent with Western production concepts. It also undermines concepts such as a stratified regional production systems in which immatures are taken off the rangelands and grown out in higher potential areas (von Kaufmann, 1976).
That the Boran seek to accumulate animals to promote prestige and protect themselves from perturbations have been long recognised as important elements of pastoral behaviour (Pratt and Gwynne, 1977). That the Boran attempt to avoid cattle sales by diversification into small ruminants and cultivation to help them endure increasing population pressure is another important side effect of their behaviour with implications for system transformation (see Section 7.2: A theory of local system dynamics).
The "perverse supply concept" is controversial (Doran et al, 1979; Jarvis, 1980; Low et al, 1980; Sandford, 1983a). Results suggest that perverse supply factors operate here, but vary according to the wealth class of the seller and the species of livestock. The hypothesis is thus very difficult to test using aggregated empirical data. It is also often misinterpreted (D. L. Coppock, ILCA, personal observation). Perverse supply in this context implies that marketed throughput of animals over a given period of time would be lower in response to higher prices; this would be true in situations where accumulation of animals is a traditional value and cash needs are limited. It is not to say that pastoralists will not respond in general to a market offering higher prices; in the southern rangelands the black market with Kenya thrives because the Boran seek higher prices for their cattle. The prediction, however, is that the throughput per household would be lower over a specified time interval as a result of higher prices. Marketing patterns over time are also confounded by increasing pauperisation of the B as a result of population growth and drought (see Section 7.2: A theory of local system dynamics).
This suggests that, with traditional values held relatively constant, perverse supply behaviour would occur more acutely in the future than it does today as a result of a change in the wealth structure of the society. A change in marketing values over time, however, would confound patterns. At least in the pert-urban areas, values reportedly are changing among younger herd owners towards increased monetization and market involvement and all respondents noted the increasing importance of markets to the survival of the Borana system (Coppock, 1992b). It is important to emphasise, however, that a transition to increased livestock marketing is only welcome if the desired commodities are regularly available at reasonable prices since market dependence has its own risks. In summary, this framework may offer a basis for understanding system-level variation in perverse supply responses due to culture and/or livestock wealth. Such a framework is needed to help elucidate why research on perverse supply phenomena has yielded such equivocal interpretations (Sandford, 1983a).
The idea that the Boran act like "optimistic gamblers" and delay cattle sales until they apparently have no other choice, has large implications for system stability and food security. These topics are explored in Section 6.4.4: Traditional drought mitigation tactics and Section 7.3.3.7: Mitigation of drought impact.
Cossins and Upton (1988b: pp 254-256) reviewed marketing concepts relevant to the southern rangelands. They noted that marketing strategies to increase offtake through provision of infrastructure assume that the existing marketing system is inefficient and that pastoralists will increase Offtake if prices are increased. They stated that without radical changes in the traditional system there is little scope for increasing offtake making the supply of marketable stock rather inelastic. They argued from the work of Negussie Tilahun (nd) that the structure and performance of the Borana market system to be "reasonably good" because of the ready access of primary markets and the presence of large numbers of buyers and sellers; the latter feature discounts the notion that price fixing or excessive marketing margins apply. They stated the major marketing constraint to be the decline in the terms of trade of livestock for grain during drought, which reflects problems in national policies for food production and distribution rather than deficiencies in the local characteristics of local markets.
In his review, Sandford (1983a: pp 199-229) concluded that the available evidence did not consistently support or reject the notion that pastoralists respond positively to price increases by increasing the supply of animals to market, pointing out the improbability that a single generalisation could explain response in such a diversity of pastoral situations. The second point about pastoral markets being dificient in their terms of trade during drought was contrary to evidence found for Africa by Sandford (1983a).
Evangelou (1984: p 239) cited Johnston and Clark (1982) as stating that the distinctive feature of the development process is the expansion and evolution of markets. Market systems are therefore not static but in a continuing process of change. There has not been any testing of hypotheses regarding market evolution in the southern rangelands, but looking into some of the experiences and speculation of local traders and others is useful in defining some likely elements of change.
Local informants feel (D. L. Coppock, ILCA, personal observation) that markets on the Borana Plateau in 1990 are more organised and diverse compared to the early 1980s. According to these informants, government efforts were initiated in 1983 to organise large-scale purchases of sheep and cattle for export. This served to train a cadre of local traders that reportedly dominated marketing in 1990. These men are often Boran who have strong ties to local urban centres. Some reportedly attained their initial wealth dealing in sheep during years of average rainfall while others did so by buying up cheap cattle during the 1983-84 drought, securing the grazing and water to keep many of them alive and subsequently making windfall profits two years later when normal weather conditions resumed. These traders allegedly maintain large herds of mature males managed as forra animals. They purchase animals during dry seasons when prices are low and fill a seasonal marketing gap by selling animals at high prices during the rainy seasons when the traditional pastoralists are apparently less willing to do so (Coppock, 1992b). Thus, there are probably now distinct groups of sellers in the rangelands that have different seasonal strategies.
Data of Negussie Tilahun (nd) collected for 35 months prior to the 1983-84 drought, suggested that livestock prices were slowly rising. Cattle prices in 1990 for mature and immature males increased (Solomon Desta, TLDP economist, personal communication) on the order of 30% compared to those recorded by Negussie Tilahun (nd). Local informants (D. L. Coppock, ILCA, personal observation) have speculated that this increase in price may be due to a greater diversity of outside buyers (government and private operators) and local traders since 1985. This reflects higher demand from the Ethiopian highlands as well as for export (FLDP nd), a continuing strong black market trade to Kenya and the relaxation of regional controls on commodity movements within Ethiopia starting in 1990 (Ethiopian Herald, 1990).
AGROTEC/CRG/SEDES Associates (1974f: p 67) reported that in October 1972 the "farm-gate" prices on the Borana Plateau averaged from EB 30 (cull cows) to EB 70 (heifers), EB 50 (four-year-old bulls) and EB 95 (six-year-old steers). Prices for purchased commodities were not reported. It was also stated that even in 1972 about 60% of all marketed Boran cattle entered Kenyan markets. This trade remains very active today. Based on estimates from informants, Hodgson (1990: p 77) stipulated that around 17000 head of Borana cattle crossed to Kenyan markets in 1988. It is not clear to what extent the early 1970s and late 1980s are comparable in terms of climate and livestock population dynamics in relation to drought impact, which complicates direct price comparisons. However, it is likely that a general inflation has occurred and the true value of cattle has increased (R. Brokken, ILCA, personal communication).
Cossins (1985) and Cossins and Upton (1987: p 216) contrasted the secondary productivity of the Borana system with that of Kenya commercial ranches (Laikipia) and three Australian cattle stations. All were considered to have grossly similar environments in terms of rainfall and net primary production. Secondary productivity was expressed per hectare. In the pastoral systems secondary productivity was variously composed of offtake of milk, meat and offal while in the ranching systems offtake was limited to meat. It was concluded that the Borana pastoral system yielded 119 MJ GE of offtake/ha/year (based largely on 6.25 kg of meat and 21 kg of milk) and thus outperformed both the average Laikipia Ranch by 24% (96 MJ GE based on 19 kg of meat) and the cattle stations by over fivefold (22 MJ GE based on 4.3 kg of meat).
Similar perspectives were previously forwarded by de Ridder and Wagenaar (1986a,b) in their analyses of production systems in eastern Botswana. They concluded from a preliminary analysis that the traditional systems were 95% more productive than ranching systems on a per hectare basis. This calculation reflected higher stocking rates, milk offtake and the use of draft power in the traditional sector, which offset their lower cattle productivity per head compared to ranching (de Ridder and Wagenaar, 1986a). They followed this up with a more complete analysis of how efficiently forage energy (GE) and crude protein (CP) are harvested and assimilated in the two systems. For example, in terms of the net energy in animal products as a per cent of total maintenance energy, they found ranching to average 110% more efficient. This primarily occurred because of differences in the age and sex structure of herds in each system. However, the traditional systems were from 119 to 233% more efficient in terms of energy consumed as a per cent of plant energy produced, and 6 to 61 % more efficient in terms of net energy in animal products per unit of plant energy produced. Similar results were obtained for CP efficiencies (de Ridder and Wagenaar, 1986b: pp 10-11). They concluded that although ranching systems were more efficient on a per animal basis, a shift from traditional production to ranching results in serious losses of efficiency and productivity per hectare. The higher stocking rates in traditional production systems, however, could pose a higher risk of pasture degradation in years of low rainfall (de Ridder and Wagenaar, 1986b: p 14). Other crucial biological, social and economic transformations required to convert a pastoral system into a ranching system are reviewed in Behnke (1984).
Cossins and Upton (1987: p 217) also compared annual outputs of energy and cash income between a mixed crop-livestock system in the Ethiopian highlands at 845 mm of annual rainfall (Gryseels and Anderson, 1983) with that in the Borana system. Compared to the pastoral system, the mixed system yielded about 63,2 and 1.75 times more energy per hectare, household and person, respectively. The mixed system also yielded nearly 12 times more cash income per hectare than the pastoral one. However, the Borana system yields about three times more cash income per household and per person. It was concluded that although the pastoralists have an energy-deficit situation in terms of food production, they do receive more cash income. Cossins and Upton (1987: p 217) concluded that the Borana system is productive and efficient and most viable as long as livestock products can be sold to purchase energy at reasonable terms of trade.
Labour coordination is likely a major reason why the encampment-level of organisation has evolved on the Borana Plateau. Certain forms of labour strafing also are probably more important in this process than others. Helland (1980b: p 23) considered labour for watering animals from the wells as far more important than that used for herding, for example. He speculated that the increase in herd size is primarily limited by labour availability for well watering. This may imply that coordination of watering from the wells in dry seasons constitutes the major cohesive function of the encampment.
AGROTEC/CRG/SEDES Associates (1974f: p 25) reported that family heads within each encampment meet each morning at a breakfast ceremony to decide on matters of household concern (this presumably includes labour coordination). They also noted that the association of families within a Borana encampment (olla) shows a remarkable stability over time, characterised by much cooperation. Other social and production factors are obviously also important in the cohesion of the encampment throughout the year, otherwise encampment would disband in the wet season with families going in different directions. There are regular seasonal shifts of a few families in some regions that are typically caused by forage shortages and/or the need to minimise distance to water in dry periods (Hodgson, 1990: p 30). In contrast, it should be noted that extensive seasonal gathering and dispersing of families characterise some pastoral groups in more arid regions of East Africa (McCabe, 1983).
If the labour organisation for working wells is a fundamental influence on the size and composition of encampments, it might be expected that those that are found outside of well areas would differ from those within. However, this hypothesis is not substantiated by general observations of Somali encampments immediately to the east (AGROTEC/CRG/SEDES Associates, 1974g: p 21) or of Borana or Gabra encampments to the north served by the large ponds. More differences in social organisation and labour coordination between the Boran and Somali in SORDU may be expected at the level of grazing resource allocation.
The strict allocation of tasks to various age and sex groups in Borana encampments is typical of pastoral systems in general (Evangelou, 1984: pp 99-101: Fratkin, 1987). AGROTEC/CRG/SEDES Associates (1974f: pp 52-53) describing these task in 1972 noted that farming was a distinct activity of males, in contrast to observations in this study that showed both males and females participating in cultivation. This latter point is consistent with the idea that newer activities such as farming may be somewhat more flexible in relation to gender roles, as has been observed in agropastoral societies (Massey, 1987: p 65).
The relatively precise allocation of duties within the Borana household means that technical interventions to improve labour efficiency in performing certain tasks would have to be targeted at particular age and sex groups. Examples include improvements to well systems (largely to the benefit of young men), construction of water cisterns near encampments (benefit to women) and hay making for dry-season calf feed (benefit to women). These concepts are reviewed further in Chapter 7: Development-intervention concepts.
Although the important and arduous roles of women in mixed farming systems in Africa has received more recent attention (Lowe, 1986: pp 19-21; Gladwin and McMillan, 1989; Kumar, 1989; Webb, 1989), relatively less attention has been given to the study of gender issues in the pastoral sector (Broch-Due et al, 1981). However, Fratkin (1989) lists some references dating back to 1965 that deal with gender inequality in the rights and ownership of pastoral livestock.
Detailed studies dealing with the allocation of time and labour of pastoral women are rare. Massey (1987: pp 66-67) described general duties of women in agropastoral Somalia. He noted that after tending to food preparation and child care, women in the dry season spent nearly all of their time hauling water (on a daily or semi-daily basis). At other times of the year the women divided their time mostly among domestic chores, care of livestock and cultivation. While Somali men shared milking of the camels, the women always milked the cows and goats by themselves. Other domestic tasks of Somali women included managing chickens, processing milk, collecting firewood, fodder collection for young stock (shared sometimes with men), making mats and vessels and erecting and taking down the camp.
Fratkin (1989) cited Dahl (1987) as noting that pastoral women are commonly responsible for running the household as well as herding and other aspects of livestock care. Field studies of Fratkin (1989) of 39 Ariaal households in northern Kenya revealed that married women worked harder (65% of the time between 0530 and 2000 furs) than married men (43% of the time), and consequently had about two-thirds of the leisure and rest time enjoyed by men (35 vs 52%). Wealthier men and women worked less than their poorer counterparts.
In sum woman's active day of 14.714.7 was composed of domestic tasks (37%), livestock tasks (14%), manufacturing (such as weaving) tasks (14%) and others (35%).
The general findings of the above studies are consistent with patterns observed among the Boran. Women's work days are reportedly long and dominated by numerous household-support tasks and care of livestock. Verbal data also indicated significant seasonal shifts in work loads and types of activities prioritised. Alleviation of women's labour burdens has been identified as a priority for development interventions described in Section 7.1.2: Development philosophy for the Boran. Observational studies on wet and dry-season labour budgets of women in experimental formats were conducted by Coppock (1992a) in support of development concepts. One finding of these studies is that the verbally reported work load of individual women in the long dry season was probably exaggerated (for details see Section 7.3.1.1: Water-development activities and 7.3.1.3: Forage improvements).
On the Borana Plateau the processing and marketing of dairy products is under the control of the women. This has been commonly reported elsewhere in African pastoral and agropastoral systems (Dahl and Hjort, 1976: p 159; Kerven, 1987a,b; Grandin, 1988; Waters-Bayer, 1988). However, as pastoralists become more linked to pert-urban markets and milk sales become more important, it has been noted that women lose control of milk marketing activities to men. Waters-Bayer (1988) noted among settled Fulani that men began to milk animals out of concern that women would take too much from calves. Likewise Salih (1985) noted among pert-urban Sudanese pastoralists that men had usurped the traditional role of women in milk marketing leaving the women only with control over poultry production and sales.
Milk available for processing in Borana households is seasonally influenced by the total milk supply. When milk supply exceeds daily household demand such as during and soon after extended rainy periods, secondary products such as butter, ghee or long-term fermented milk are most likely to be produced. This principle also applies when considering wealth differentiation in the society. Wealthier families having a higher ratio of lactating cows/person will more commonly have a larger surplus throughout the year. This means that wealthier families could produce products like butter or ghee for a longer period each year than poorer families. Kerven (1987b: p 20) reported from work in the Sudan that poorer pastoral women processed and sold milk for about five months from when the rains started, while wealthier women were able to do so for an additional three months. It follows that wealthier households should more offen have surpluses that allow them to make and sell butter and store ghee for the dry season.
The basic patterns of milk processing observed here such as churning soured milk to make butter, dehydrating butter to make ghee and removing whey to better regulate milk fermentation are all common traditional practices in the pastoral, agropastoral and mixed farming systems of Africa (O'Mahony, 1988: p 1). In addition to these products, farmers in the Ethiopian highlands make a cottage cheese (ayib) by heating skimmed milk, which precipitates casein and some of the remaining fat. This is not produced by the Boran nor do they express interest in doing so; they stated that they don't like the taste of ayib (Tarik Kassaye and C. O'Connor, ILCA, unpublished data). Ayib may also be inappropriate for the Boran enviroment because of its short shelf-life owing to its high moisture content (O'Mahony, 1988: p 48) that precludes storage and marketing. Perhaps for some of the same reasons, making of soft cheeses by pastoralists is apparently uncommon in northern Kenya (Dahl and Hjort, 1976: p 160; Galvin, 1985) or in the Sudan (Kerven, 1987b: p 19). For a review of the biological principles involved in milk processing and their applicability to the African rural producer the reader should consult O'Mahony (1988).
Ghee provides the Boran with a high-energy food with an excellent shelf-life of seven months to one year (Tarik Kassaye and C. O'Connor, lLCA, unpublished data; D. L. Coppock, ILCA, unpublished data). However, quantities of stored ghee per household are likely small and relatively unimportant for food security during dry seasons (D. L. Coppock, ILCA, personal observation). Ghee may have been more important in the past when according to informants milk surpluses were more common as a result of a lower density of people (see below; Section 7.2: A theory of local system dynamics).
Butter has important uses in the preparation of local coffee and porridge, as a cosmetic and for cooking. The storage stability of butter, while not comparable to ghee, is still on the order of four to six weeks (D. L. Coppock, ILCA, personal observation). Hence butter has a distinct advantage over fresh milk in terms of more temporal flexibility for household use and marketing. Butter collected in Borana markets such as Yabelo, Dubluk and Mega can even be taken by traders to the southern highlands via the public transport system. Furthermore, even when slightly rancid, butter still has market value in Ethiopia. Most urban consumers all over Ethiopia and highland peasants use butter for making traditional wot is eaten with a bread-like injera made from teff. A degree of rancidity of the butter is actually desired to improve the flavour of wot (D. L. Coppock, ILCA, personal observation). In sum, it is apparent that butter has outstanding features as a marketable commodity even in semi-arid Ethiopia.
In contrast to butter, fresh milk may become rancid within one day. Consequently, it can only be sold by families within a close proximity to markets. These are usually poor families that cannot store enough sour milk to make butter. Tittarelli (1990: pp 54-57), in his study of dairy marketing in the Ethiopian highlands at Selale, also found an effect of marketing distance on the sale of various types of dairy products. He found that smallholders within 5 km of collection centres or roads sold fresh milk more often, while those further away processed and sold more butter.
Besides being easier to sell in terms of spoilage risk, butter also commands a higher price than milk per kilogram throughout the year (see appendix 4 in Holden, 1988). Averaged for two markets, butter varied from EB 6.26/kg (wet season) to EB 7.36/kg (transition) and EB 9.00/kg (dry season). In contrast, fresh milk varied from EB 0.68/kg (wet season), to EB 1.04/kg (transition) and EB 1.11/kg (dry season). All seasonal means within each commodity were significantly (P£ 0.05) different (Holder, 1988). However, Holden (1988) also noted that on an energy basis fresh milk (3.34 MJ GE/kg) was a bit more expensive than butter (29.8 MJ GE/kg) with annual means from EB 0.25/MJ GE (butter) to EB 0.28/MJ GE (fresh milk).
As practiced by the Boran and others, the souring of milk offers several practical advantages over fresh milk. The acidity retards the growth of undesirable microbes and to a certain point improves efficiency of milk churning (O'Mahony, 1988: pp 19,46). Smoking of containers prior to use for milk storage likely provides important bacteriostatic (cleansing) effects (Ephraim Bekele and Tarik Kassaye, 1987). It is not known to what extent the smoke of wood chips from different tree species affects bacteria, or if species are selected merely on the basis of the flavors they impart to the milk. Both factors are probably important. Informants reported that A. nilotica chips are the most commonly used material for smoking gorfa. When a gorfa is smoked with A. nilotica chips milk reportedly takes a bit longer to sour and has a superior taste (D. L. Coppock, ILCA, personal observation). This suggest that the volatile compounds present in these chips are effective in killing bacteria.
The production of long-term fermented milk with a shelf life of up to 60 days has stimulated particular research interest in its microbiological and chemical processes (Tarik Kassaye, 1990). However, it is important to note that household studies of milk processing and allocation indicate that in terms of annual volume it is virtually insignificant.
On the Borana Plateau cattle provide most of the milk for processing, followed distantly by small ruminants. The inability of camel's milk to yield an apparently suitable butter (even after vigorous churning) may be an important species difference that could affect the choice, given limited management resources, of whether a cattle-owning household should also acquire camels (see Section 7.3.3.3: Dairy processing and marketing). Even though the evidence regarding the suitability of camel's milk for processing is equivocal, Dahl and Hjort (1976: p 185) stated that camel's milk sours quickly. They also mentioned that some camel-keeping groups process camel's milk by heating it and preparing the curds into storable cheeses. The Gabra have demonstrated to ILCA staff how they make a waxy butter from camel's milk using a process that involved heating and extended agitation (D. L. Coppock, ILCA, personal observation). But Dahl and Hjort (1976) did also cite a variety of other conflicting references regarding the suitability of camel's milk for processing. Other perspectives on the subject are provided in Rao et al (1970) and Galvin and Waweru (1987).
Results from milk processing studies that evaluated the efficiency of butter making are presented in Section 7.3.3.3: Dairy processing and marketing including a discussion of the feasibility of implementing dairy technology interventions. As is discussed below, the presumed steady decline in the ratios of livestock to people on the plateau may preclude much optimism regarding the implementation of milk processing interventions, since it is anticipated that surpluses will become smaller. The practice of milk sharing within encampments may further reduce available surpluses even though the actual degree of milk sharing may be exaggerated (Coppock et al, in press).
4.4.10.1 Dairy marketing and welfare of humans and calves
4.4.10.2 Dairy marketing in a wider perspective
Detailed discussions of Borana dairy marketing may be found in Holden (1988) and Holden et al (1991) while the main points are summarised here:
Market access is a critical factor in the participation of pastoral women in dairy marketing. Households closer to market are able to sell more frequently and this was a reflection of the opportunity cost of women's travel time to market, which influences net returns (Askari and Cummings, 1976). Effects of distance to market varied with household wealth, and wealth has been found elsewhere to be a critical factor in pastoral dairy marketing (Kerven, 1987b; Grandin, 1988). Very low levels of milk supply in poor households during dry seasons precluded their ability to increase marketable output in response to a given reduction in distance. In contrast, during the wet season poor households had a higher milk supply and a greater marketing response over distance. The pattern was somewhat similar for other wealth classes, but they had greater flexibility in supply and could be more responsive over distance regardless of season.
Compared to poor ones wealthy households sold greater absolute quantities of dairy products, but on average retained four times more milk for household and calf consumption than poor households. The poor had only about 25% as much milk per person as the wealthy. It also appeared that, when milk supply per household increased, either because of effects of wealth or season, people first increased their consumption of dairy products in preference to maximising dairy sales. This seems in contrast to highland smallholders with cross-bred cattle who may first use milk for sale (Wagenaar Brouwer, nd).
Consumption was a priority with wealthy households as sales of surpluses are used primarily for nonessential purchases (Nester, 1985: Kerven, 1987b: Waters-gayer, 1988). By contrast, the poor appeared compelled to buy grain because their levels of milk offtak were below subsistence requirements. Again, one litre of milk (3.3 MJ GE) sold bought 3.5 kg of grain (52.5 MJ GE) in the dry season, providing nearly a 14-fold increase in energy. This illustrates that having surplus milk is not a prerequisite for dairy marketing to be important. Indeed, the reverse is more true for a pastoral society.
Despite their lower absolute volume of dairy sales, income from dairy sales provided 37% of the annual income of poor households when close to market; for the wealthy this was 22%. The poor, with few animals to sell without endangering their herd capital (Swift et al, 1984; Behnke, 1987), had no viable alternatives to selling milk in order to get money. As long as the lives of nursing calves are not endangered, dairy marketing would generally contribute to the food security of poor households. That is from the direct effect of providing cash income and the indirect effect of delaying sales of animals for some other crisis in the future. Also that compared to animal sales, dairy sales permit purchasing of quantities of grain that are more convenient to handle by the household. For example, Nestel (1985) noted that the Maasai who purchased large volumes of grain from an animal sale often lost a significant amount to neighbours and relatives who came to request hand-outs.
Field data supported some, but not all, elements of the hypothesis that poorer households closer to market would be relatively more affected by trade in dairy products. Improved access to market only appeared to reduce the proportion of milk allocated to calves in poorer households. As there was no clear effect of proximity to market on intake of dairy products by women or young children, the increase in milk offtake by poor households closer to town is probably best attributed to increased daily sales.
The nutritional consequences of trading dairy product for grain by poorer families can be illustrated by assuming: (1) such households to have 3.5 AAME and 3.8 lactating cows (evenly distributed among the three productivity classes); (2) a nutrient content of milk of 3.3% crude protein (CP) and 3.7 MJ GE/kg (Roy, 1980; Nicholson, 1983a); (3) maize grain having 15 MJ GE and 7.4% CP per kg on a dry-matter basis (Cossins and Upton, 1987: p 213; ILCA Nutrition Unit, unpublished data); and (4) the nutritional guidelines for a 55-kg average adult male (Cossins and Upton, 1987) as recommended in FAO (1973) and NAS (1974).
If this family resides far from town and sells no milk, the daily milk offtake rate of 48% (1.8 litre) found among the modal poor family could provide about 39% and 18% of the daily total requirements for CP (154 g) and GE (37 MJ), respectively. This implies that the family must either reduce demand for food and/or receive more food from their social network, gathering bush foods or relief grain. If the family increases milk offtake to 63% (2.31), then about 49% and 23% of its daily requirements for CP and GE, respectively, are provided, but this is still short of meeting needs. However, if the increment of 15 percentage points of increased offtake (0.5 litre) is sold at a ratio of 3.5:1 that means for 1.75 kg of maize about 125% and 87% of the CP and GE requirements, respectively, will be met (Holder, 1988). This shows that proximity to market and favourable terms of trade are especially important for poor families. A chronic pressure to trade even more milk could compromise the nutritional status of people for vitamins or amino acids foregone in milk and not found in grain (Nester, 1985: pp 158-163; Shrimpton, 1985); and nutritional balance of grain and milk is desirable (Nester, 1985: p 158).
It was not confirmed by the researchers that poorer families closer to town would have more grain in their diets. Informants, however, reported that it is "common knowledge" that poor people close to town sell milk daily and live on purchased grain, often to the detriment of their children (D. L. Coppock, ILCA, personal observation). A significant effect of wealth on intake of dairy products by women and children was observed, unlike patterns observed for Maasailand where all wealth classes had similar intakes (Grandin, 1988). This discrepancy may have been due to the extreme poverty of many of the Boran spreading the differences among all wealth classes. It could also be because of the lingering effect of the 1983-84 drought on herd-level milk production (Holder et al, 1991).
The overall calf mortality rate of 18% and milk offtake of 41% recorded by Holden et al were consistent with other findings for an average rainfall year in the study area (see below). Despite the more restricted milk intake of calves in poor households overall, this did not appear to affect the mortality rate at the time. Households were able to minimise acute effects on their calves (Wagenaar et al, 1986).
Paradoxically, calf morbidity rates were reportedly high in both wealthy and poor households, despite their differences in milk offtake One hypothesis is that morbidity of calves in poor households was due more to nutritional stress from milk restriction, while that for calves in wealthier households was probably due more to health management problems that arose from more calves being concentrated in a small area and hence getting less individual management attention (Mulugeta Assefa, 1990: p 26).
The increase of 15 percentage points in milk offtake for cows of poorer households closer to market may be equivalent to 150 ml/calf/day in the dry season. This is 28% of the calculated intake for calves of poor families at the time and represented a substantial decrement that could reduce vigour and increase the loss of calves to nutrition-related diseases. Informants reported that this is a common pattern among the poorer families nearer to market that must sell milk to survive (Coppock et al, 1992).
Considering the conceptual model of economic stress responses in pastoral societies addressed thus far, some of the literature on pastoralism suggests that dairy marketing is symptomatic of increasing poverty (Dahl and Hjort, 1976: p 181; Toulmin, 1983; Waters-Bayer, 1988). This is illuminated further by the contention that the Borana used to have taboos against selling dairy products but have probably been forced to do so more recently (Dahl and Hjort, 1976: p 181).
If pastoral poverty is defined as a decline in per capita milk production and livestock holdings, then it is anticipated that the long-term trend would be for most of the Boran to become poorer. This would result from a steady increase in the human population, low rates of emigration and economic diversification and limits on the livestock that can be carried by the system (see Section 7.2: A theory of local system dynamics). Effects of drought (such as killing a high percentage of cattle and largely sparing people; see Chapter 6: Effects of drought and traditional tactics of drought mitigation) would be in addition to these pressures. This means that drought could periodically exacerbate the poverty among the Boran. If it is assumed that the primary reason for selling dairy products is to buy grain because the milk supply is insufficient for survival in terms of energy, this suggests that an increasing number of families would be regularly involved in dairy marketing over the long term (at least from the pool of encampments within reach of a marketing town), and that a big upsurge in dairy marketing activity could be expected during the early stages of a drought and during the post-drought recovery period when there is sufficient milk to sell, but not enough to sustain households. The degree of dairy marketing therefore is likely to be variable from year to year.
Lack of a standardised time series for the Boran precludes a test of this hypothesis and existing data often lumps dairy sales with those of other commodities. For example, AGROTEC/CRG/SEDES Associates (1974f: p 71) noted that about 22% of the annual household income of EB 86 in 1972 was derived from the sale of "animal products" other than live animals and presumably some of this was milk and butter. Negussie Tilahun (1984) reported that an average of 5% of the annual income of EB 830 in 1983 was from the sale of "animal products" that included butter and milk and that dairy income was important for the purchase of dry goods (another indicator that these families were wealthier than average; see Section 4.4.1: Genreal aspects of household economy). Donaldson (1986: pp 50-51) found that 30% of the household budget of EB 384 during five months at the height of the drought in 1983-84 was from sales of dairy products, hides, handicrafts and other mis- cellaneous items. Holden and Coppock (1992), whose pert-urban sampling approach probably provided the most accurate range of data, figured that 20% of the average annual income of EB 440 was derived from dairy products. Holden (1988: p 42) found that about 30 of 105 households reported that the income from the sale of dairy products was more important to them than consuming the milk itself. However, the fact that the study reported in Holden (1988) and Holden and Coppock (1992) was carried out in 1987, only three years after the end of the 1983-84 drought, may imply that a period of particularly high levels of dairy marketing was witnessed in view of the drought-recovery scenario proposed above.
Like livestock, formal dairy markets were perhaps small and less well organised prior to the 1980s. Urban populations were smaller and the number of milk or butter traders was limited. In contrast for 1987, S. J. Holden (ILCA, personal communication) noted the complexity of trading interactions in the towns of Mega and Yabelo (with populations of 3000 and 7000, respectively); many pastoralists had private arrangements to supply dairy products directly to urban families and restaurants in addition to regular market-day activities. Informants reported, that milk was sold prior to 1980, but that the Boran used this income more commonly to purchase items like coffee and sugar, not grain (D. L. Coppock, ILCA, unpublished data). At least in the small town of Dubluk, butter traders reported in 1987 that competition to buy from producers was far less in the early 1980s than in the late 1980s.
In sum, these studied observations provide useful focal points for system monitoring and tests of hypotheses. They also point to important dynamic constraints in the consideration of dairy processing or marketing interventions. This perspective, as well as a comprehensive theory of pastoral development and change that takes into account long-term trends and short-term cycles, is presented in Chapter 7: Development-intervention concepts.
