Wealth differences in pastoral production systems
Wealth ranking in pastoral systems research
Informant wealth ranking in Maasailand: a case example
References
Importance de l'effet de là richesse sur la production pastorale: une méthode rapide de stratification de la richesse
Summary of discussion session 5.
Résumé des débats de la cinquième séance
Barbara E. Grandin
Anthropologist, Arid Zones (Eastern and Southern Africa) Programme, ILCA, Kenya
This paper discusses the importance of wealth differences between producers in traditional pastoral production systems. It argues that significant wealth differences exist, that these have a profound effect on production strategies and that pastoral systems research must pay attention to them at several stages from defining a target population or recommendation domain to developing and testing interventions. The paper then describes a rapid method for determining the wealth rank of producers within a given community. Such a ranking is an important tool to stratify a population of producers before sampling to ensure the representativeness of the sample along this important dimension. Alternatively, it may be used post facto for assessing the representativeness of pastoralists already interviewed.
For many years studies of traditional agricultural production systems emphasised the essential homogeneity of producers. While it was recognized that some differences existed, they were though, on the whole, not to be significant, but purely a matter of scale.
In agricultural economics, for example, earlier research results tended to be reported in terms of the average farmer, with his average family, average cropping pattern and average yields. In anthropology, to the extent that economic heterogeneity was researched, emphasis was placed on "levelling mechanisms", which by definition, functioned to counteract trends towards inequality. For example, polygamy has been discussed as having a levelling function, as more wives would mean more children and consequently the eventual fragmentation of the wealthy producers' assets.
In studies of peasant agriculture, there has been increasing recognition of the extent of wealth differences within communities and their effects on production parameters. As Hill (1972) noted in her pioneering work on inequality in a Hausa village in Northern Nigeria, "it is not merely that a few farmers operate on a much larger scale than others... but that there are many rich farmers who have entirely different economic aims from many poorer farmers." She observed that the tendency of scholars to ignore wealth differences among African peasants had hampered our ability to understand their systems of production.
Today, researchers are paying far more attention to intra-community differences between farmers. It is recognized that wealthy farmers have differential access to land, labour, animal inputs (traction and manure) and credit, to name a few important production parameters. In addition with differential savings and investment possibilities, wealthier farmers have different attitudes to risk and innovation (Cancian, 1978) than their poorer counterparts on the survival fringe.
For pastoral systems, recognition of the importance of wealth inequality within communities has lagged behind. Much of the household level research in pastoral production systems has been done (and continues to be done) by anthropologists who have paid insufficient attention to the issue of economic inequality in traditional production systems. The ideology of equality which tends to predominate. in pastoral societies as well as the apparent similarites in consumption levels of different households, (coupled with a theoretical tradition which emphasised the homogeneity of traditional communities) ,have all contributed to this apparent bias of early anthropologists in studying both peasant and pastoral societies.
Unfortunately, as Konezacki (1978) noted in his book entitled The economics of pastoralism: A ease study of sub-Saharan Africa, there are few data available on the distribution of livestock ownership within pastoral communities; the little data which are available however indicate that "the prevailing pattern of wealth, and consequently income distribution among African societies dependent on animal husbandry, is one of inequality."
Table 1 presents data from several African pastoral societies on the distribution of livestock holdings between households. These eases were chosen not because they represent extremes of inequality, but rather because they are the few eases for which data were readily available. In each instance there is a marked inequality of livestock holdings, which can be taken as a close approximation of wealth especially in a purely pastoral system where land is communally held.
Table 1. Inequality in livestock distribution: a few African examples.
|
Place |
Unit |
No. of HH |
Mean |
Median |
Range |
Date |
|
1. Tuareg - A |
L.U |
14 |
30 |
34e |
0-58 |
1979 |
|
2. Tuareg - B |
Camels |
31 |
21 |
17e |
2-83 |
1982 |
|
3. Wodaabe |
Cattle |
75 |
16 |
18e |
5-45 |
1982 |
|
4. Somali |
L.U |
36 |
132 |
70 |
4-660 |
1950 |
|
5. Sebei |
Cattle |
42 |
21 |
12 |
0-100+ |
1960 |
|
6. Maasai |
Cattle |
41 |
109 |
58 |
4-499 |
1980 |
eestimated
Sources :
1. Swift (1979)
2.3. Wilson and Wagenaar (1982)
4. Lewis (1961)
5. Goldschmidt (1976)
While there are some indications in the literature of the extent of intra-community wealth differences in pastoral societies, there are few, if any, systematic explorations of the effect of wealth differences on production parameters. Intuitively, we would expect differences in wealth to affect production in important ways in a pastoral system. Wealth in the form of animal numbers both enables (and necessitates) a large family which can be garnered through polygamy, adoption, and the incorporation of poor relations as dependents. Greater wealth also means fewer cash flow problems, and increased possibilities of access to purchased inputs, including drugs and mineral supplements. Clearly a producer with 400 animals will have different management strategies and possibilities than a producer with only 4 animals. Wealthier pastoralists, particularly those with sufficient labour within the household, have more management options in terms of herd-splitting and lending of animals (both for risk avoidance and to maintain useful social networks). They would be likely to have more control over grazing and watering than poorer producers who are not able to herd independently. We would expect breeding to be affected as poor producers might not have sufficient access to a male for servicing whereas wealthy producers might have difficulty with seasonal and other control of breeding in large herds and flocks.
ILCA's research in several rangeland systems, particularly in Kenya, has focused on the effects of wealth on livestock production. Table 2 presents preliminary findings from ILCA's ongoing research in Kenya Maasailand. It shows that virtually every important production parameter varies considerably with the wealth rank of the producer. For ease of presentation, the data refer only to the rich and poor (leaving out the middle stratum), and only to one of the three group ranches under study.
Section A of Table 2 provides background information. Of forty households on the ranch, 13 were classed as poor and 12 were classed as rich. Of these' the ILCA sample included seven poor and seven rich households. The poor households owned only 5% of the livestock units on the ranch whereas the rich households owned 48% of them.
Table 2. Production unit heterogeneity: example from a Maasai group ranch .
|
A. Background |
Poor |
Rich | |
|
|
HH on ranch |
13 |
12 |
|
|
HH in sample |
7 |
7 |
|
% of ranches L.U. (total sample) |
4.9 % |
48% | |
|
per household |
0.7 |
6.5 | |
|
B. Livestock holdings |
Mean |
Mean | |
|
|
1. Value of livestock holdings |
24,154 |
234,050 |
|
|
2. No. of cattle |
31 |
302 |
|
|
3. No. of smallstock |
42 |
213 |
|
|
4. SS to cattle ratio |
1.3 |
0.7 |
|
C. Livestock: offtake |
|
| |
|
|
1. Cattle: net offtake % |
20 |
7 |
|
|
2. SS : net offtake % |
23 |
6 |
|
|
3. Slaughter per capita, Kshs. |
33 |
200 |
|
D. Livestock herding |
|
| |
|
|
1. cattle herd alone (%) |
0 |
57 |
|
|
2. SS herd alone (%) |
29 |
57 |
|
|
3. Cattle herding group size |
160 |
372 |
|
E. Labour inputs |
|
| |
|
|
1. Total workers |
6.3 |
9.8 |
|
|
2. No. of adult men |
0.9 |
1.5 |
|
|
3. No. of adult women |
1.7 |
3.4 |
|
|
4. Children in school: % |
38 |
15 |
|
|
5. Ratio of cattle/worker |
5 |
31 |
|
|
6. Hit/day to LS per worker |
3.9 |
5.2 |
|
|
7. Hit/day per livestock unit per work |
1.29 |
0.29 |
|
F. Income |
|
| |
|
|
1. Income from LS/capita (Kshs) |
657/= |
1420/= |
|
|
2. Mean value of cattle sold |
577/= |
971/= |
|
|
3. % value of commercial transactions in smallstock |
18% |
7% |
|
G. |
|
| |
|
Expenditure |
|
| |
|
|
1. Per cap. food/drink |
195/= |
465/= |
|
|
2. Per cap. HH |
165/= |
238/= |
Section B presents information on the livestock holdings of the sample households. The figures are all means for the households in that stratum. It shows that rich households have approximately ten times the value of animals, largely accounted for by their having ten times the number of cattle of poorer households. Poorer households have proportionally more smallstock, with a smallstock to cattle ratio of 1.3 as opposed to 0.7 in rich households.
The offtake data presented in Section C continue to show clear wealth differences. (de Haan has already mentioned the finding of Wilson and Wagenaar in Niger that offtake rates are negatively correlated with herd size). The same is true in Maasailand. The rich stratum has net offtake rates of 7% for cattle and 6% for sheep; for poorer households the figures are 20% for cattle and 25% for sheep. Poor Maasai (as poor pastoralists in many areas of Africa) (Wilson, 1980, Little, 1982) place far more emphasis on smallstock than wealthier producers. They own proportionately more smallstock and derive more of their income from smallstock (18% of livestock sales rather than 7%). As we would expect, there is far more slaughter per capita in rich households.
Section D shows that more rich households are able to herd alone, thus having complete autonomy in decision-making with regard to grazing. In addition, whether they herd alone or with others, they benefit from an economy of scale by having very large herd sizes for grazing, watering and dipping.
Section E presents a few data on labour. Again a clear difference emerges; rich households have more members, but because they have far more animals per worker (25 cattle per worker as opposed to 5 in poor households), they send fewer children to school and work longer hours on livestock management. Looked at from a different perspective, however, we see that they benefit both from economies of scale and also from their ability to marshal! non-household labour. Whereas each worker in poor households devotes over an hour a day per livestock unit to livestock work, workers in rich households spend only one quarter of an hour.
Section F shows that rich households have a per capita income from animal sales which is more than twice that of poor households. As rich households do not suffer the same cash flow problems of poor households, they are able to keep steers to maturity, thus fetching higher prices per animal. Again we see that poorer households engage in more smallstock sales and purchases than rich households.
Finally, Section G of Table 2 shows that higher income is reflected in higher per capita expenditures on food, drink, and general domestic goods.
Clearly, as we would expect, producers of different wealth ranks differ considerably in a variety of ways; from their access to the essential elements of production, to their herd/flock structures, offtake rates, and management strategies.
There are several points in pastoral systems research at which it would be important to take account of the wealth heterogeneity in the area under study. It would be useful in the early stages as part of the identification of the recommendation domain. Interventions which are useful to large herd owners might be unadoptable by poorer producers; in fact, they could intensify and even solidify otherwise transient economic inequalities by, for example, providing differential access to credit. Secondly, during the verification survey, it is important to be sure that the producers interviewed represent a reasonable cross-section of the population for whom interventions are to be designed.
As de Haan mentioned in his introduction, and as I have tried to demonstrate, evidence suggests that the single most important parameter for stratifying within a community is wealth rank. However, among pastoralists wealth is a very difficult parameter on which to obtain accurate data. For pure pastoralists, livestock holdings represent a close approximation of wealth. While livestock holdings can be used as a proxy for wealth in a pastoral production system, animal censuses are often difficult, if not impossible, to carry out. Producers are afraid to have their animals counted, due to fear of taxation, other government interference, or solely on the basis of a cultural taboo on the counting of animals. Even if the pastoralist can be persuaded to have his animals counted, the logistic difficulties of such an operation may be overwhelming. Animals may be scattered into different management units, animals owned and animals kept are not isomorphic categories, frequent movements and large distances add to the costs of such censuses. Some of these logistic problems may be overcome by counting animals at watering points, but young and ill animals are likely to be missed. (For agro-pastoralists or pastoralists, with significant non-pastoral occupational and investment opportunities, the situation can be far more complex).
This paper contends that livestock censuses or other objective measurements are not necessary to establish the wealth rank of producers within a community. The technique of having local informants who rank members of their community according to wealth can yield similar results at a fraction of the time and expense.
The technique of informant wealth ranking
Informant ranking was first used by Silverman in 1966 in a sociological study of prestige hierarchies. It has subsequently been used by myself and several other anthropologists to elicit wealth ranks. Informant ranking uses a card sorting technique in which the name of each producer is written on a small card and several informants are asked to place the cards in piles according to the wealth of each producer. It has three basic requirements which will be discussed in turn:
1. a list of producers to be ranked;2. a few reliable informants;
3. The elicitation of the word or phrase in the local language according to which the producers are to be ranked (in this case wealth).
List of producers
The technique is dependent upon ranking producers in relationship to each other; obviously the ranking will reflect true wealth differences in the population only to the extent that- all producers are included. This is not as overwhelming a task as it might first appear. Ways of eliciting the list of producers within the community under study include using censuses, tax roles, and/or voter registration, which can be cross-checked with chiefs, elders or other community members. If there is a known watering regime, names can be elicited at dry-season watering points. With more sedentary populations, air photographs can be used by 'walking" informants through them. If, as in Maasailand, producers have an area that they consider "home", whether or not they are present, their names can be elicited on a neighborhood by neighborhood basis. In one of the group ranches under study in Kenya, for which there was no accurate list of resident households, a neighborhood by neighborhood interview elicited the names of 206 household heads who normally use the ranch. This was done in an area of 1350 km² in less than a week.
The necessity of having as complete a list as possible and having all the producers ranked, cannot be overstressed. One recent study in Kenya (White and Meadows, 1981) proposed to examine wealth differences in group and individual ranches in Kenya. The chairman of each group ranch was asked to suggest 12 members for study: four with few livestock, four with an average number of livestock and four with more than average. The researchers were pleased to find significant differences between their wealth cohorts; they assumed that their sampling technique had tapped wealth differences on the ranches. Unfortunately that was not so. One of the ranches they studied is a ranch on which ILCA has been working for two years. In checking the sample households against the population (for which we had complete livestock censuses), it was discovered that the three strata used by White and Meadows represented the middle class, upper middle class and very rich respectively. Poor households were entirely excluded.
This was not through the ignorance of the chairman as we shall see later, but perhaps was due to his assuming that researchers studying animal and especially cattle production would have no interest in households with very few cattle. Table 3 shows the differences between White and Meadows strata and those derived from a census of all ranch households. Whereas their "poor" households owned a mean of 51 cattle, the true population mean is 15. For the "average" households, they show a mean of 218 cattle whereas the population mean is 58. For the "rich" households the Meadows and White mean is 378 cattle while the true mean is 249. For the ranch, they calculate a household mean of 215 cattle, whereas the true mean is only 109. Had White and Meadows used the technique suggested here in which a complete list of households were obtained, and ranked, a much more representative sample and hence a better study would have followed.
Reliable informants
As the degree of inter-informant reliability on wealth ranking tends to be quite high, only a few informants are required. For a small community (i.e. up to 100 producers), where it is likely that everyone knows everyone else fairly well, four or five informants (preferably of different ages) should suffice. Larger communities may need to be divided into smaller sub-areas (defined by the local social structure). Although I have not yet had the opportunity to do this myself, I see no a priori reason why the results from different sub-areas should not be merged, as long as there is sufficient overlap between the rankers. For example if each of three groups of informants rank 250 producers of whom each knows 100, the overlap should be sufficient to make a composite rank.
Despite ideologies of equality, most pastoral producers recognize that there are at least temporary differences in wealth status of the members of their community. They also recognize that people in different wealth strata are in qualitatively different positions with regard to management possibilities. In the two communities where I have done informant wealth ranking, I have had no difficulty in finding a few informants willing to do the task, probably because no precise information (e.g. how many cattle, how many wives) is required.
Table 3. Distortion through poor sampling
|
Mean household cattle ownership in a Maasai group ranch | ||||
|
|
Poor |
Average |
Rich |
Ranch mean |
|
Sample |
51 |
218 |
378 |
215 |
|
Population |
15 |
58 |
249 |
109 |
Table 4. Scoring on informant's ranks.
|
Rank group |
No. of men |
Men plus dummy case |
Final score |
|
1 (high) |
13 |
13 |
6.5 |
|
2 |
15 |
16 |
21 |
|
3 |
9 |
9 |
33.5 |
|
4 (low) |
13 |
13 |
44.5 |
|
Unknown |
1 |
|
|
Local concept for wealth
In order to ensure the comparability of the data obtained through the various informants, as well as to ensure that they are ranked according to the criteria the researcher desires, it is important to spend some time, preferably with one or two good interpreters, eliciting the exact local cultural concept to be tapped during the ranking.
Most cultures have a clear concept of wealth. Once the word or phrase is elicited it is useful for the researcher to spend some time determining the various elements that are considered in determining the wealth rank of a producer. Having done this with interpreters, it is also useful to elicit this information from informants after they have completed the ranking. This ensures that the informant has understood what was required; it also can provide valuable insight into sources of wealth and differential access to local resources.
In Maasailand, the concept used was "emali" a word meaning "property" but most commonly used with specific reference to livestock holdings. Both interpreters and informants felt that the primary indication of a man's wealth was his livestock holding; subsidiary points were smallstock to cattle ratio and family size. I was told that the few Maasai who have been given individual title deeds to land would have had that property included should they have been included among the producers to be ranked.
Computation of the wealth rank
As noted, several informants are asked to rank the producers according to wealth by putting the cards on which their names are written into piles. The informant is allowed to choose the number of piles he wants. Frequently in the course of sorting, the number of piles will be increased as names appear of producers who don't quite fit into existing piles. Informants need not be literate; the names can be read to them after which they will just point to the pile where it belongs. The cards of producers whom the informant is not able to rank are just placed to one side. When the ranking is finished, the names in each pile are read to the informant. This allows for the detection of any errors; it also gives the informant a chance to review the groups he has compiled and make any necessary changes.
On the basis of the sortings, a score is constructed for each producer for each informant. Each producer is assigned a score which is the equivalent of the midpoint value of the category in which he fell. As the number of unknown producers varies between informants a number of dummy cases equal to those of unknown producers are inserted at equidistant intervals. Table 4 show-e how the scores for one informant were calculated on a Maasai population of 51 producers. In order to arrive at a composite score for each producer, the scores for that producer are simply averaged together.
In order to test the use of informant ranking in a pastoral population (I had previously only used it with mixed farmers), I had four informants rank the 51 producers who live on a single group ranch in Kenya Maasailand. The results of that exercise are reviewed here briefly as a case study.
Table 5 shows the inter-informant rank correlations. The ranking of each informant is compared to the rank of every other informant. Informants 1 and 2 are young men of middle wealth. Informant 3 is an older, poorer man who happens also to be the chairman of the group ranch who had suggested the White and Meadows sample. Informant 4 is middle aged and the richest man in the population. The informants represented the four different neighborhoods on the ranch. Despite the differences in the informants' backgrounds, the inter-informant reliability is quite high (the Spearman's rho correlation coefficients range from 0.87 to 0.91). When the rank of each informant is compared to the composite or average rank, the correlations are even higher 0.95 to 0.98). These are all Spearman's rho and all are significant at P = 001). Clearly, despite age, wealth and neighborhood differences between the informants they were in high agreement about the relative wealth of the members of the ranch. In the Maasai ranking, scores ranged from a high of 3.7 (the richest) to a low of 49.5 (the poorest).
Table 5. Maasai wealth ranks.
|
Inter-informant rank correlation | |||||
|
|
1 |
2 |
3 |
4 |
Average |
|
1 |
- |
0.88 |
0.87 |
0.91 |
0.96 |
|
2 |
- |
- |
0.88 |
0.91 |
0.95 |
|
3 |
- |
- |
- |
0.90 |
0.96 |
|
4 |
- |
- |
- |
- |
0.98 |
|
|
|
N = 52 |
|
| |
Table 6. Informant-livestock unit correlation.
|
Livestock unit |
INFORMANT |
Average |
|
|||
|
1 |
2 |
3 |
4 |
|||
|
A. |
0.92 |
0.95 |
0.96 |
0.96 |
0.97 |
N = 22 |
|
B. |
0.95 |
0.94 |
0.97 |
0.97 |
0.99 |
N = 20 |
All correlations are significant at P = 0.001.
Table 6 shows the results of a correlation between the informant wealth rank and a ranking of producers based on an ILCA census about six months previously. It refers to the subset of 22 producers with whom the census had been undertaken. The Spearman's rho is 0.97 (Table 6A). If the two cases for which it is known that there were censusing problems (counting of more than one producer's animals), the correlation rises to 0.99 (Table 6 B) (again both with a significance level of P = 0.001). Clearly the informant results gave an almost identical ranking of producers as the census. Fig. 1 plots the producers along two axes: livestock units censused and informant wealth rank. It presents graphically what the correlation coefficient has already indicated.
The high correlations are not an artifact of a small size. Fig.2 depicts the relationship between the wealth ranks (done in late 1982) with livestock units censused in mid-1980 for the 41 households which were then resident on the ranch. Again the relationship is quite clear, although there has not been time to correlate the coefficients. Had the Kenya team stratified it's sample by informant wealth ranks, the results would have been identical to stratifying on the basis of a complete animal census. In fact, in a certain respect the stratification would have been better, as the initial census contained a number of significant errors as one would expect given pastoral resistance to censuses. Census errors are circled in Fig 2. Some of these are minor (e.g. counting a poor relation's animals with a rich man's herd) but others are significant.
Fig 1. Informant vs. census rank.
Fig 2. Informant ranking (1982) vs. Initial ILCA census (1980).
Having obtained an informant wealth ranking, the researcher can use stratified sampling to enhance his precision. Informant wealth ranking does not totally replace censusing. After stratifying and sampling it would be critical to do an animal census to know the holdings of the sample and generalize to the holdings of the community.
On the other hand, if data must be collected opportunistically (i.e. with those producers who are amenable or accessible), informant wealth ranking can still be used to gauge the representativeness of the producers interviewed or censured.
Informant ranking techniques need not be limited to wealth. Producers can be ranked on almost any variable including age, family size, management ability etc. The technique, as I hope I have shown, is rapid, simple yet accurate.
Cancian, F. 1978. The innovator's situation: Upper middle class conservation in agricultural communities. Stanford University Press, Stanford, California.
Goldschmidt, W. 1976. Culture and behaviour of the Sebei: A study in continuity and adaptation. University of California Press, Berkeley, California.
Hill, P. 1972. Rural Hausa: A village and setting. Cambridge University Press, Cambridge. P.S.
Konczacki, Z.A. 1978. The economics of pastoralism: A case study of sub-Saharan Africa. Frank Cass, U.K. p. 49-51.
Lewis, I.M. 1961. A pastoral democracy. Oxford University Press, London.
Little, P.D. 1982. Risk aversion, economic diversification and goat production: Some comments on the role of goats in African pastoral production systems. In: Proceedings of the Third International conference on Goat Production and Disease, Tucson, Arizona. University of Arizona, pp. 428430.
Neadows, S.J. and White, J.M. 1981. Evaluation of group and individual ranches in Kajiado district, Kenya, to economic development and pastoral production strategies. Kenya, Ministry of Livestock Development, Nairobi.
Swift, J. 1979. The economics of traditional nomadic pastoralism: The Tuareg of the Adrar N Iforas (Mali). University of Sussex. (Unpublished, Ph. D dissertation).
Wilson, R.T. 1980. Indigenous goats and livestock development in Africa. Manuscript.
Wilson, R.T. and Wagenaar, K. 1982. An introductory survey of livestock population demography and reproductive performance in the area of the Niger Range and Livestock Project. ILCA/ Mali, Arid and semi-arid zones programme, Bamako.
Résumé
Cet exposé traite de l'importance des différences de la richesse entre les producteurs dans les systèmes de production pastorale traditionnelle. Il soutient qu'il existe des différences importantes en ce qui concerne la richesse. Que celles-ci ont des effets importants sur les stratégies de production et que la recherche sur les systèmes pastoraux doit en tenir compte à divers niveaux, qu'il s'agisse de la définition de la population-cible ou de la mise au point et de l'expérimentation d'interventions.
Il décrit une méthode rapide de détermination des strates de la richesse auxquelles appartiennent les producteurs dans une collectivité donnée. Chez les pasteurs, la richesse est un paramètre sur lequel il est très difficile d'obtenir des données exactes. Pour les éleveurs purs, la possession de bétail représente une quasi-richesse. Mais les recensements des populations animales sont souvent difficiles, voire impossibles à effectuer. L'étude soutient que les recensements du bétail ou d'autres mesures objectives ne sont pas nécessaires pour établir le rang (en termes de richesse) des producteurs au sein d'une collectivité. L'utilisation d'informateurs locaux pour classer les membres de leur collectivité en fonction de leur richesse peut fournir des résultats similaires en moins de temps et à un moindre coût.
L'informateur utilise une technique de tri de cartes dans laquelle le nom de chaque producteur est écrit sur une petite carte et plusieurs informateurs sont invités à ranger les cartes les unes au dessus des autres en fonction de la richesse des producteurs. La technique requiert l'obtention d'une liste complète des producteurs opérant au sein de la collectivité.
Etant donné que le degré de fiabilité des informateurs pour le classement en fonction de la richesse tend à être élevé, seul un nombre limité de ceux-ci sera nécessaire. Pour une petite collectivité où l'on suppose que tout le monde se connaît assez bien, quatre ou cinq informateurs devraient suffire. Pour assurer la comparabilité des données obtenues par le biais de divers informateurs et pour s'assurer que le classement s'effectue selon les critères souhaités par les chercheurs, il est important de consacrer quelque temps, de préférence avec un ou deux bons interprètes, à l'étude de concepts culturels locaux à exploiter au cours de la stratification.
Sur la base du tri des cartes, des points sont attribués à chaque producteur pour chaque informateur. Chaque producteur reçoit des points qui constituent l'équivalent de la valeur moyenne de la catégorie à laquelle il appartient. Pour avoir la performance globale de tels producteurs, on fait simplement la moyenne des points obtenus par ces producteurs.
Après l'obtention de la stratification des richesses par les informateurs, le chercheur peut utiliser des échantillons stratifiés pour plus de précision. La stratification des richesses par les informateurs ne peut pas remplacer complètement les recensements. Après la stratification et l'échantillonnage, il convient de procéder à un recensement des animaux pour déterminer la population animale détenue dans l'échantillon et pour extrapoler à partir de celle-ci la structure de la propriété au sein de la collectivité.
Les techniques de classification par informateur ne doivent pas nécessairement se limiter à la richesse. Les producteurs peuvent être classés sur la base de presque toutes les variables, y compris l'âge, la taille de la famille, l'aptitude à gérer, etc. La technique est rapide, simple et précise.
Chairman: Prof. Gunnar Sorbo (Norway)
Discussion led by Dr Sitta Barry (Upper Volta)
Dr Barry commented that CILSS (Comité permanent interétats de lutte contre la sécheresse dans le Sahel) was set up in 1973 by six countries (Chad, Niger, Upper Volta, Mali, Senegal and Mauritania) following the drought of 1968 to 1973. Subsequently Gambia and Cape Verde had joined CILSS, whose mission was to
- increase the awareness of the international community to drought and its effects;
- mobilise the necessary resources to combat drought and its effects;
- co-ordinate activities to develop the Sahel and to combat drought and desertification; and
- co-operate with other organisations having similar goals.
In the field of animal husbandry CILSS had established a strategy consisting of six themes:
- increasing livestock feed resources;
- increasing livestock productivity using animal health control methods; organising professional channels for livestock production;
- modernizing and improving marketing channels;
- recruitment, training and extension work;
- research.
Dr Barry reviewed the papers by Drs Cossins and Grandin and commented that demographic analysis should be carried out over homogeneous environments which may mean the regrouping of pastoralists or rural areas according to different criteria. He referred to Dr Cossins' paper which emphasised the tendency of the Afar to destock small ruminants, and said this was the opposite of the trend in West Africa. The possession of small ruminants could be more flexible and effective than that of cattle because:
- the distribution between families was less concentrated;
- the loss in value per unit was lower in the case of forced sale or loss;
- smallstock were less sensitive to weight loss in case of insufficient feed;
- the reproductive performance of smallstock was better, therefore capital could be reconsolidated more rapidly.
Dr Zulberti questioned Dr Cossins' statement that 'the pastoralist is the only one who depends on milk and not meat'. Dr Cossins pointed out that he was referring to arid and desert areas where, with the exception of irrigated areas, dairying was rare if existent. Dr Zulberti disagreed with the statement that 'the pastoralist seeks to optimist the number of people that any area is capable of supporting'. Dr Cossins replied that if pastoralists did not have the aim of optimising the number of people per unit area then they would not need to pursue pastoralism; pastoralism almost certainly evolved as a way of supporting more people per unit area than any other production mode in aride areas.
In commenting on Dr Cossins paper, Dr Diakite stressed that very different approaches had to be adopted for Australian and African pastoralists. Dr Cossins agreed and said that he hoped his paper had made that very clear.
Dr Timbo suggested that in the Sahel context, the strategies of pastoralists were governed to a large extent by the law of the greatest number in order to combat climatic or animal health problems. The primary aim of pastoralists was the production of young stock/ calves, not milk production. In the Sahel milk production had never been a production alternative beyond the traditional self-sufficiency of the pastoral family.
Dr Rhissa, in commenting on Dr Cossins' paper, said that when studying pastoralists and their production strategies the conclusions needed to be qualified depending on whose behalf one had been carrying out the study. A government's aims may be to help herdsmen organise themselves better, to provide the food security of a zone and to improve the contribution of pastoral products to the national economy. A pastoralist's aims were to provide milk, to attain his own food security and his own social security. Dr Rhissa felt that the pastoralists aims were often overlooked.
Dr Tilahun referred to Dr Grandin's paper and suggested that in the study of inequality a distinction should be made between household income and household per capita income. Dr Grandin agreed and said that studies should investigate these differences in their analysis, particularly in societies close to the subsistence level where a high proportion of expenditure was on food which increased proportionally with the number of people in the family.
Président: Prof. Gunnar Sorbo (Norvège)
Débats dirigés par le Dr Sitta Barry (Haute-Volta)
Le Dr Barry a déclaré que le CILSS (Comité permanent inter-Etats de lutte contre la sécheresse dans le Sahel) avait été créé en 1976 par six pays (le Tchad, le Niger, la Haute-Volta, le Mali, le Sénégal et la Mauritanie) à la suite de la sécheresse ]968 - 1973. Par la suite, la Gambie et le Cap-Vert avaient adhéré à l'organisation dont la mission consistait à :
- amener la communauté internationale à prendre conscience de la sécheresse et de ses effets;
- mobiliser les ressources nécessaires pour combattre la sécheresse et ses effets;
- coordonner les activités pour développer le Sahel et pour lutter contre la sécheresse et la désertification; et
- coopérer avec d'autres organisations ayant des objectifs similaires.
Dans le domaine de l'élevage, le CILSS avait mis en place une stratégie basée sur six thèmes:
- accroître les ressources fourragères;
- accroître la productivité animale en utilisant des méthodes de contrôle de la santé animale;
- organiser des circuits professionnels pour la production animale;
- moderniser et améliorer les circuits de commercialisation;
- recruter et former du personnel et effectuer des travaux de vulgarisation;
- entreprendre des activités de recherche.
Le Dr Barry a fait référence aux documents de M. Cossins et de Mlle Grandin et a déclaré que l'analyse démographique devrait être effectuée sur des environnements homogènes, ce qui pourrait signifier le regroupement des éleveurs ou des zones rurales selon divers critères. Il a fait allusion au document de M. Cossins qui a mis l'accent sur la tendance des Afars à réduire leurs troupeaux de petits ruminants et a déclaré qu'en Afrique de l'Ouest c'était la tendance contraire. La possession de petits ruminants permettait une plus grande efficacité et une plus grande flexibilité que celle des bovins, notamment parce que :
- la distribution entre les familles était moins concentrée;- la perte en valeur unitaire était plus faible en cas de vente forcée ou de perte;
- les petits ruminants étaient moins sensibles aux pertes pondérales lorsque l'affouragement est insuffisant;
- la performance du petit bétail en matière de reproduction était meilleure et permettait par conséquent de reconstituer le cheptel beaucoup plus rapidement.
Le Dr Zulberti a mis en doute l'affirmation de M. Cossins selon laquelle "l'éleveur est la seule personne dont l'existence se base sur le lait et non pas sur la viande". M. Cossins a indiqué qu'il faisait allusion aux zones arides et désertiques où (à l'exception des zones irriguées) la laiterie était rare si tant est qu'elle y ait jamais existé. Le Dr Zulberti a exprimé son accord sur l'affirmation selon "laquelle l'éleveur cherche à optimiser le nombre de personnes qu'une superficie donnée est capable d'accueillir" M. Cossins a répondu que si les éleveurs n'avaient pas comme objectif d'optimiser le nombre de personnes par unité de surface, ils n'avaient pas besoin alors de faire du pastoralisme; le pastoralisme constituait presque toujours un moyen d'accueillir plus de personnes par unité de surface que ne le permettraient les autres modes de production dans les zones arides. Dans ses réflexions sur le document de M. Cossins, le Dr Diakité a souligné que des approches très différentes devaient être adoptées pour étudier le pastoralisme en Australie et en Afrique. M. Cossins a exprimé son accord et a déclaré qu'il espérait que dans son document, cette distinction avait été clairement établie.
Le Dr Timbo a indiqué que dans le contexte du Sahel, les stratégies des éleveurs étaient régies dans une large mesure par la loi du plus grand nombre, afin de combattre les problèmes d'ordre climatique ou de santé animale. L'objectif primordial des éleveurs était la production des jeunes animaux/veaux et non la production de lait. Dans le Sahel, la production de lait n'avait jamais constitué une alternative de production au-delà de l'autosuffisance traditionnelle de la famille pastorale.
Dans les observations qu'il a formulées sur le document de M. Cossins, le Dr Rhissa a déclaré qu'en étudiant les éleveurs et leurs stratégies de production, les conclusions devaient tenir compte des objectifs du bénéficiaire de l'étude. L'objectif du gouvernement peut par exemple être d'aider les éleveurs à mieux s'organiser, d'assurer la sécurité alimentaire d'une zone et d'accroître la part des produits pastoraux dans l'économie nationale. Les objectifs de l'éleveur étaient de fournir du lait, d'assurer sa propre sécurité alimentaire et sociale. Le Dr Rhissa a déclaré qu'à son avis, les objectifs de l'éleveur étaient souvent négligés.
Le Dr Tilahun a fait allusion au document de Mlle Grandin et a indiqué que dans l'étude sur l'inégalité, une distinction devrait être faite entre le revenu du ménage et le revenu du ménage par tête. Mlle Grandin a accepté cette observation et a déclaré que les études devraient se pencher sur cette différence dans leurs analyses, notamment dans les sociétés proches du niveau de la subsistance où une forte proportion des dépenses était consacrée à l'alimentation qui augmentait proportionnellement avec le nombre de personnes dans la famille.
