IX. Socioeconomic constraints to the use of browse in development programmes

Study on the economic viability of browse plantations in Africa**

C. De Montgolfier-Kouèvi,

Economist, ILCA, Addis Ababa

H.N. Le Houérou

Principal Scientist, ILCA, Addis Ababa

1. Introduction

2. Spineless cacti plantations in Tunisia

3. Atriplex plantations

4. Acacia plantations

4.1 Acacia cyanophylla plantations in Tunisia

4.2 Phyllodineous acacia plantations in Senegal

4.3 Acacia senegal plantations in Senegal and Sudan

4.4 Acacia albida plantations

5. Other browse tree and shrub plantations

5.1 Prosopis plantations in the Cape Verde Islands

5.2 Leucaena plantations in Malawi

7. Conclusions

Annex I

Determination of shadow fu and DP prices

Annex II

References

1. Introduction

First of all we should take the opportunity to thank colleagues who have supplied us with the often unpublished data, used in the present study.

Specifically we would like to thank the following:
A.G. Seif el Din, IDRC expert in Nairobi;
R. Delaye, FAO expert in Tunis;
O. Hamel, Director of the CTFT in Dakar;
M. Sabra, ex-director of the FAO project in Cape Verde; as well as
J. Durkin, ILCA programmer in Addis Ababa, who was responsible for the data processing.

The data were compiled by H.N. Le Houérou, while the analysis was carried out in collaboration with C. de Montgolfier-Kouèvi, particularly as regards their economic interpretation.

Browse tree and shrub plantations have an undeniable value in countries where labour is plentiful and cheap, as is the case in Africa. In addition to supplying wood or other products such as gum arabic, they provide livestock in the arid and semi-arid zones with feed supplies to cover the bridging period and during times of drought, often with a high protein value. They help to stabilize animal production without in any way competing with food or forage crops, which they may even benefit. Since they can be planted on marginal land which cannot be used for ordinary cropping (sand dunes, steeply sloping land, rocky soil, land subject to flooding and soils with a high salt content, etc), browse trees and shrubs play an important part in protecting valley slopes and stopping erosion. Other species can be grown on land which is actually cultivated, helping to improve crop yields by supplying soil nutrients, especially nitrogen.

For these reasons the cost price of a feed unit (FU) produced from browse trees and shrubs is probably lower than one produced from most of the concentrates used as animal feed. The cheaper forms of concentrate, such as cereal bran, urea and molasses, are in any case usually only available in limited quantities in specific areas, so that transport costs render their utilization prohibitive outside the immediate production area. Their contribution to animal feeds in Africa will therefore be limited for many years to come, if not actually insignificant, whereas an enlightened policy to promote browse would probably not only improve animal nutrition but also help offset the shortage of fuel supplies (by producing firewood), at the same time as maintaining soil fertility and productivity.

On the basis of experimental cases drawn from the existing literature it therefore seemed relevant to undertake an evaluation of the feed resources which might be obtained by planting browse trees or shrubs. Our aim has been to assess the economic feasibility of such plantations and to evaluate the cost price of the animal feed which could be produced under relatively extensive conditions, in other words using a minimum of inputs.

In order to carry out this evaluation the establishment costs of browse plantations were assessed, for the most part using examples drawn from concrete experience ¹, for plants raised in the nursery and subsequently replanted in the field. These costs include soil preparation, water and soil conservation where appropriate, and maintenance (weeding and ploughing) prior to utilization. To a great extent they consist of labour costs, which often account for up to 80% of overall costs, since soil preparation has been subject to little or no mechanization.

¹ There is a large amount of data on this subject, but it is not entirely coherent. Where necessary these data have been supplemented by further calculations so as to take management factors into account.

Generally speaking mechanization does not appear to bring appreciable reductions in costs, which merely break down differently, with a greater burden on the trade balance of the country concerned. The salaries of technical staff have usually not been taken into account, nor has the purchase price of the land used.

To normal planting costs should be added those of enclosures to protect the plants. Enclosures may consist of barbed wire fences or thorn hedges. The first alternative is generally much more expensive than the second, especially in tropical Africa where, according to such estimates as were available, the cost price of barbed wire fencing is three times higher than in Tunisia ². On the other hand, hedges have to be planted 2 or 3 years before the plantation itself, so as to afford proper protection, a factor which prolongs the unproductive period of the land by this amount of time. Investment in enclosures, whether barbed wire or hedges, therefore constitutes an important entry and often a very substantial one in the overall establishment costs for a plantation. For this reason a systematic attempt has been made to isolate these costs in our calculations of cost prices and profitability.

² About US $450/ha, as against US $150 in Tunisia, for enclosures covering an area of about 10 ha.

Estimating operating costs has proved somewhat more problematic. Since very little information was available in this field we have usually had to estimate these data ourselves, basing our estimates on a number of assumptions. It has been assumed, for instance, that since this type of enterprise is primarily extensive, operating costs largely consist of fixed costs (mainly those of patrolling the plantation) which are independent of production, the latter being primarily governed by ecological factors such as soil type aridity, or socio-economic factors affecting management. Nonetheless, it was also argued that although browse plants can theoretically be used without variable operating costs since the feed can be consumed on the spot by the animals, the offtake of wood implies cutting and transport costs which are directly linked to the level of output. This approach may have led to some underestimation of operating costs, especially for plantations which were deemed to have the benefit of good management.

In most cases production was estimated on the basis of several assumptions in relations to yield, so that management differences could be taken into account and profitability thresholds determined, so long as the range of yields adopted was felt to correspond to what could theoretically be achieved under extensive conditions. Evaluating production posed intricate problems owing to the difficulty of establishing shadow prices for animal feed. The shadow prices adopted have important implications for the internal rate of return (IRR), the factor we were seeking to identify. Generally speaking the price of livestock feeds depends on their energy value (measured in forage units – FU – or any other unit directly or indirectly indicating the number of calories contained in the feed) and their protein content (measured in digestible protein – DP – per kg of feed). On the basis of these two factors and of world market prices for the main livestock feeds (cereal grains and brans, oilcakes, meat meal), the prices per FU and per kg of DP were estimated to average US $0.16/ FU and US $0.24/kg of DP during the year 1979 (see Annex 1). This structure is nevertheless unstable, since any movement in the relative prices of livestock feeds, especially between those with low and high protein contents, for example cereal brans against cakes, leads to distortion of the FU/DP price ratio (see Table 1) which may at times be considerable. Leaving aside the results for the years 1975 and 1976, which were effected by the sharp rise in cereal prices on the world market, the price per kg of DP is generally higher than the price per FU. The ration was about 1.5:1 for the period 1975–1979, and for the year 1979 itself.

Table 1. Prices per fu and per kg of dp.( in US cents )

^a Cereal bran/oilseed cakes and meat meal.
^b Wheat bran/soya cake.
^C Wheat bran/groundnut cake.

Source: FAO (1975/79) and UNCTAD (1978/79).

If we turn to the equivalent African prices, based on the examples of Tunisia and Senegal in 1979/80, prices per FU and per kg of DP, worked out from cereal bran and cake prices, settle at levels quite different to those found on the world market. Protein in African countries has an extremely high value, while FU values are very low on account of low producer prices for cereal brans: 5-7 US cents per kg, as against 13.6 cents on the world market. These estimates, biased by the low prices of cereal brans as recorded at the place of production, do not appear truly representative of the real value of animal feeds in African countries. As noted above, when they are used cereal brans carry transports costs which considerably increase their cost price to the final user.

Taking into account other animal feed products with a low protein content, such as molasses, Acacia albida pods and even maize, completely preserves the FU and DP price structure (US $0.16/FU and US $0.16 per kg of DP in Senegal for 1979/80), bringing it closer to that of the world market over the same period. In short, given the unreliability of FU and DP price estimates, the structure prevailing on the world market is probably the most representative. In the end, owing to the lack of other reliable data, it was therefore adopted as the basis for the shadow FU and DP prices used to calculate the IRR on planting browse trees and shrubs in Africa. These prices are also similar to those of cereal grains on local markets. Cereals are often the only source of concentrate available to farmers and livestock producers.

Nonetheless, adopting a rather high FU shadow price is not very satisfactory, since in Africa livestock are usually fed at low cost on natural forage produced for the most part on land unfit for cultivation. However attractive from a financial point of view, browse plants at the shadow prices identified by this method would probably have no economic applicability. An attempt has therefore been made instead to estimate FU and DP cost prices directly, assuming a given IRR, i.e. an opportunity cost of capital more easily determined on the basis of interest rates. At this point further difficulties arose, first in the attempt to identify two separate prices (both per FU and kg of DP) and secondly because browse trees and shrubs can be used for several purposes, among which the production of wood at the same time as forage. For DP and FU prices a fixed structure at the level found on the world market in 1975–79 was assumed, giving a ratio of 1.5:1 between them.

As far as the production of wood alongside browse in pastoral areas is concerned, the high cost of transport from production to consumption areas is a prohibitive factor ³. In the first place it was therefore assumed that browse trees and shrubs are cultivated primarily for forage production, with wood being counted as a by-product with no market value, so that all investment and operating costs could be attributed to browse production. If this is the case the FU and DP cost prices are those which, for the IRR selected, balance the discounted cash flow of income and expenditure resulting from the production of browse alone.

³ In Senegal the cost of acacia wood was around US $50/t in urban areas in 1980. On the other hand, Pterocarpus lucens wood was priced at under US $30/t in Niono, Mali, at the same time, after the seller had devoted between 24 and 30 hours to collecting it in the bush. Taking into account the opportunity cost of his labour and other outlays (depreciation of the donkey and cart used for transport), the cost price of wood in the bush would certainly not be over US$ 15-16/t. To avoid any risk of overestimation, this was the level chosen for the shadow price of wood (acacia) used in the IRR calculations.

Secondly however, it can be assumed that browse trees and shrubs are cultivated for both wood and forage, which would mean that the various costs should be shared between the two activities. A breakdown of this kind is by no means a simple matter, especially as regards investments made for both types of enterprise. Nonetheless, the breakdown can be made to reflect the distribution of earnings from the two production alternatives, although this distribution is itself dependent on the FU and DP shadow prices which we are attempting to determine. In the end the FU and DP cost prices were calculated by breaking down investment costs according to earnings and distributing operating costs somewhat arbitrarily ⁴. Whatever the uncertainties to which it gives rise, this twofold approach allowed a range of FU and DP cost prices to be identified, which supplemented the preceding estimates already carried out for IRR. Nonetheless, since the shadow price of wood is fairly low, the second part of the approach has a fairly small impact on FU and DP cost prices.

⁴ The following procedure was followed for investments: an initial breakdown of earnings using the shadow prices adopted for the IRR calculation, followed by a second breakdown on the basis of estimated average cost prices; this enabled a second estimate to be made, which was considered as final. As regards operating costs it was decided that all fixed costs (patrolling) could be attributed to forage production, while variable costs (cutting and transport of wood), were attributed to wood production.

Assessing the viability of browse tree and shrub plantations raised one final issue, that of evaluating the opportunity cost of land. As already mentioned above, the purchase price of land was not included in the establishment costs of the plantation. In order to take this factor into account, the land-use changes which should logically result from the very establishment of the plantation have to be considered, especially as regards utilization of the grass stratum which had been previously used for feeding animals, with the result that investment and operating costs have to be analysed in terms of the incremental additional FU and DP production generated.

However, a further problem immediately arises: how to assess the value of the forage production lost, generally during the pre-development period, and specifically the shadow cost which must be applied to evaluate this output. In dry areas unsuitable for cropping the value of forage is generally thought to be very low, not merely because the opportunity cost of land is low or zero also because grass is relatively plentiful when it is available during the growing period, whereas the value of browse lies in its ability to supply an extra feed resource during periods when livestock are under-nourished. In other words, according to this theory the opportunity cost of land and the value of the unused grass stratum can be considered as nil.

This assumption is not entirely satisfactory. It fails to take into account the rising numbers of animals stocked on tropical African rangeland and the resultant growing scarcity of grass. For this reason the production of browse should not compete with that of grass but should complement it, if pastoral production systems are to survive. In the end the same shadow prices were therefore adopted for both grass and browse plants. This assumption is highly optimistic, since grass is rarely a limiting factor during the rainy season, whereas browse supplies and the protein they provide almost always are during the dry season. This decision theoretically implies underestimation of the IRR, although the impact generally appears fairly low. It is also open to the same objection as made above, that the shadow price of browse is too high given the conditions prevailing in tropical Africa, but the objection can be met with the same answer, that of estimating FU and DP cost prices on the basis of the opportunity cost of capital, assuming that the two prices are the same for grass and browse plants.

Even discounting all these uncertainties, the approach adopted is still very far from exhaustive. Simple types of management have been envisaged, although more sophisticated systems could have been imagined, for example involving deferred or rotational grazing or browsing, implying lower stocking rates in the short term but improved yields in the long run. A more dynamic approach would also have taken into account the stabilizing effect of browse forage reserves on herd numbers and productivity during extreme drought periods. However, the lack of reliable data discouraged us from undertaking such exercises, in so far as the many alternative systems which could have been put forward would have been open to accusations of irrelevance, since at present they appear to be socially inapplicable in most African countries.

2. Spineless cacti plantations in Tunisia

A bulk food with a low protein content, cacti provide an emergency feed supply and are cultivated as such in many semi-arid areas of Brazil and southern and northern Africa as part of anti-drought strategies. Cacti have high yields given the conditions under which they grow, although ecological limitations prevent them from growing in some arid zones, for example in the West African Sahel where, except for a narrow coastal strip, the air is too dry for 6 to 8 months of the year. In well managed extensive plantations yields can reach a high point of 30 to 80 t of green matter per ha (3-8 t of dry matter - DM -), at a density of 2000 plants per ha, i.e. 1500 to 4000 FU (assuming 1 kg of DM equals 0.5 FU).

Cattle and sheep can consume the fleshy lobes or joints of cactus plants in amounts equivalent to up to 10% of their live-weight, provided a certain proportion (1% of live-weight) of dry roughage in the form of straw or hay is added. A cactus-based feeding regime also considerably reduces watering requirements. Cactus joints may be directly consumed by the animals, a method which has the advantage of allowing simultaneous use of the grass stratum. However, this form of management requires strict control of the herd and careful management of the plantation, to avoid overgrazing. Optimum technical conditions of this kind are unfortunately difficult to achieve in Africa, and as a result a cut-and-carry system is generally used in order to avoid the risk of overgrazing. However, the cut-and-carry method has the drawback of incurring maintenance costs for the cactus plants as well as preventing utilization of the grass stratum. The management system chosen thus has a considerable impact on cost.

Investment costs have been estimated by various authors in Tunisia. Overall they can be evaluated at some US $425/ha for planting as such, to which should be added the costs of enclosures, which are indispensable where direct grazing is concerned. For a plantation with a density of around 2000 stocks/ha, costs subdivide as follows:

1. Planting costs (soil preparation, purchase and transport of cactus joints, and planting itself), which come to approximately 60% of total investment costs and 80 man-days/ha in terms of labour. These costs were estimated during experiments on large plantations in Tunisia (Monjauze and Le Houérou, 1965). They amounted to US $104 in 1965, i.e. US $324 in 1980 prices ⁵. Barrachette (1980) estimated costs at US $430/ha the maximum for loans and subsidies granted to planters is US $265/ha. The average of these two figures (US $347) is very close to the 1965 estimate on the basis of experimental data, and as a result a figure of US $350/ha was eventually adopted as representative of planting costs in 1980.

⁵ In order to take into account the rising prices between the time of estimate and the year 1980, prices were re-evaluated using the UN index of unit values of manufactured goods exported by developed countries (UNCTAD, 1978/79), which rise from 88 in 1965 to 300 in 1980 (on the basis of 100 in 1970).

2. Maintenance costs during the 3 years prior to utilization, which can be estimated at US $75/ha/year.

3. The costs of enclosures which, as already noted, may involve either hedges or barbed wire fences. According to different estimates, the costs of barbed wire fencing can be put at US $150/ha in Tunisia on the basis of a unit cost of around US $1/metre. Costs for a double-row spiny cactus hedge reach a maximum of US $60/ha, but are incurred at least 2 years before planting can take place.

On this basis the establishment costs for a cactus plantation thus average US $0.24-0.28 per plant (US $0.21 without enclosures).

Operating costs of the fixed costs of patrolling the plantation, which can be estimated at US $12/ha/year on average and, when applicable, the costs of cutting and carrying the browse, which may be estimated at US $6/t of cactus, i.e. US $40/t of DM.

Cactus joints are primarily used as forage. Utilization can usually begin after 4 years and increase gradually until full production is reached after 6 to 7 years. It has been noted above that yields generally reach 30 to 80 t/ha when the plants are fully developed in well managed extensive plantations. However, in order to take the existence of lower management capabilities at local smallholder level into account, it has been assumed that the production of cactus joints lies between 1 and 6 t of DM/ha with a nutritive value of 0.5 FU/kg of DM and 20 g of DP/FU. Variable operating costs are directly linked with these different levels of output.

The value of production has been estimated using as shadow prices the costs per FU and per kg of DP derived from those of the main livestock feeds on the world market in 1979, namely US $0.16/FU and US $0.24/kg of DP (see Table 1). It varies between US $82 and US $495 according to assumptions with regard to yield (Table 2). It has also been assumed that the plantation was established on grazing land with an annual production of 220 FU with a DP content of 125 g/FU, i.e. having a value of US $42 on the basis of shadow FU and DP prices, and that establishing the plantation meant giving up using the grazing land throughout the period prior to utilization in the case of direct grazing (years 1 to 4, or 1 to 6 if spiny cactus hedges are introduced), and throughout the utilization period if a cut-and-carry system is used.

Table 2. Evaluation of cactus production during development period in Tunisia.

^a Based on a value of 0.5 FU/kg of DM.

^b Based on a value of 20 g of DP/FU.

^C Based on shadow prices of US $0.16/FU and US $0.24/kg of DP.

The IRR remains almost always lower than 10% when the plantations are used under a cut-and-carry system. Under these conditions a profitability level of 10% is only achieved when production reaches 3000 FU/ha year or more. It may also be noted that the IRR tends to be lower for plantations with hedges than for those with barbed wire fencing. The prolonged unproductive period of investment largely offsets the savings made by not using fencing. In short, under a cut-and-carry system it does not seem possible to produce browse at a reasonable price in Tunisia unless production is over 6 t of DM/ha/year, using an extensive system (see Tables 3 and 4).

Table 3. Internal rates of return of spineless cactus plantations in Tunisia ^{a .}                                                                                                                  -in %-

^a Project life is 20 years; shadow prices are US $0.16/FU and US $0.24/kg of

^b During development period.

Table 4. Cost prices per FU and per kg of DP of spineless cactus in Tunisia ^a.

^a Project life is 20 years; DP cost price is assumed to be 1.5 times higher than FU cost price.

^b Not significant.

^C The FU and DP cost prices are lower than shadow prices.

The result obtained when a direct grazing system is used is more promising, operating costs, being lower and the grassland being theoretically available for use during the development period. This system apparently allows an IRR of almost 10% as soon as production is over 2 t of DM/ha, a lower level than yields observed on well managed farms. As Table 4 indicates, the FU and DP cost prices, for a 10% opportunity cost of capital, tend to fall well below shadow prices, dropping to around US $0.7–0.8/FU, when cactus production reaches 4 t/ha. However, as has been seen, systems of this kind are very demanding in terms of management and do not appear widely applicable under current African conditions.

It should also be noted that these estimates relate to feed with a poor protein content which could never by itself provide adequate feed rations for animals and would have to be combined with other species with a high nitrogen content, such as Atriplex Acacia spp. or tree lucerne. IRR calculations for cactus plantations are therefore only valid in the context of diversified browse plantations which complement each other, or when other kinds of protein-rich supplementary feed are available, such as lucerne, bran, cake, urea etc, but as has been seen availability poses problems.

3. Atriplex plantations

Highly resistant to drought, Atriplex species can grow on heavy and salty or alkaline soils with an average rainfall of at least 150 mm per year. Like cacti they do not adapt well in some arid tropical zones such as the West African Sahel, but on the other hand they appear well suited to conditions found in the dry zones of eastern Africa. They provide better quality browse than cactus joints owing to their high crude protein (CP) content (15 to 25% of DM) and, unlike cactus plantations, they can be directly grazed without incurring any specific management problems, since they are little browsed when green grass is plentiful. The grass stratum can thus be used at the same time as the browse, as soon as the plantation reaches maturity. Finally, Atriplex plantations can also be used for producing wood, although its quality is rather poor and producer prices are definitely low owing to transport costs. Wood yields from Atriplex plantations are broadly similar in terms of dry weight to the leaf production used for browse. Browse yields are generally lower than for cactus, however. The dual production objective and the rich protein content of Atriplex leaves nevertheless mean that Atriplex plantations can yield financial results distinctly superior to those for cactus.

The establishment costs for a plantation of Atriplex nummularia were studied in Tunisia by Franclet and Le Houérou (1971) and in Israel by Orev (1962). Although carried out under very different experimental conditions, using manual planting in Tunisia and mechanized drilling in Israel, the two studies reach very similar results: US $80/ha in Tunisia, with A. nummularia in 1969 ⁶, i.e. US $245 in 1980 prices, and US $70/ha using A. halimus in Israel in 1962, i.e. a 1980 cost of US $220. On the other hand, Barrachette calculates investment costs at US $475/ha for plantation of 2000 stocks/ha in Tunisia in 1980, with the costs spread over a period of 2 years, counting US $400 for the first year and US $75 for the second. If enclosure costs of US $150/ha are also taken into account, total establishment costs for a plantation of Atriplex can be evaluated at over US$ 600 per ha, i.e. 31.2 cents per plant, or 23.7 cents per plant without enclosures. The latter estimates were the ones used for our own profitability calculations ⁷.

⁶ Planting manually in Tunisia (nursery, soil preparation, transport, planting out and watering) requires 40 to SO man-days/ha with a plant density of 2000 bushes per ha, i.e. 45 bushes per man-day.

⁷ In Chile, planting costs with anti-erosion banks and including both planting and enclosures have been estimated at US $200/ha by Benjamin (1980). These plantations, established in the Norte Chico region with an average rainfall of 100 to 220 mm, involved Atriplex repanda, A. nummularia, A, semibaccata, Galenia secunda and Mairena brevefolia (Kochia b. ),with browse production at between 1000 and 2000 kg of DM/ha/year.

As the Atriplex wood is marketed, the management system consists of allowing the animals to browse the leaves whenever the branches are cut. As a result operating costs primarily consist of patrolling the plantation, estimated at US $12/ha/year, and wood cutting, estimated at US $12 (4 man-days) for an output of 2.5 t. of wood.

Atriplex plantations established either with plants raised in the nursery or by direct drilling reach maturity in 2 or 3 years. If properly managed, leaf and wood production could reach 2.5 to 5 t of DM each during the development phase, between 4 and 7 years. Nevertheless output was assumed to be well below this figure, so as to take into account lower management capabilities at smallholder level and browse production levels between 1.25 and 5 t of DM/ha were adopted, i.e. between 500 and 2000 FU. On this basis wood cutting costs are estimated between US $6 and US $24/ha/year, according to the assumption with regard to production.

As in the case of cactus plantations, FU and DP shadow prices are derived from livestock feed costs on the world market in 1979, i.e. US $0.16/FU and US $0.24/kg of DP. Owing to its inferior quality the costs of Atriplex wood have been estimated at half those of acacia species, amounting to US $7.8/t (see footnote 3). Overall, the value of output varies from US $196 to US $504/ha (Table 5). It was also assumed that the plantation was established on grazing land producing about 220 FU/year, i.e. having a value of US $42 on the basis of FU shadow prices, and that the land cannot be grazed during the first 3 years of pre-development, or the first 5 years when the plantation is enclosed with thorn hedges, in other words before it reaches maturity.

Table 5. Forage and wood production in Tunisian Atriplex plantations.

^a Based on a value of 0.4 FU/kg of DM.

^b Based on a value of 300/g of DP/FU.

^C Based on shadow prices of US $0.16/FU, US $0.24/kg of DP and US $7.8/t of wood.

The IRR is over 20% when browse and wood production reach 2.5 t of DM/ha, and is still above 10% when output is equal to or above 1.25 of DM/ha (14% when no enclosure costs are included). When yields become more substantial the IRR is over 25%, and even reaches 30% without enclosures. The FU cost price at this level is only US $0.3-0.5/FU, with a 10% opportunity cost of capital, and US $0.4-0.8 for an IRR of 15%, a level well below the shadow price (Tables 6 and 7). According to the result in Table 7, the FU and DP cost prices are only higher than shadow prices when browse production is very low (500 FU) and when the opportunity cost of capital reaches 15%, whether or not enclosure costs are included. Even with IRR at 10%, cost prices remain below FU and DP shadow prices at this low level of production only when the Atriplex plantations are provided with enclosures and are not used for wood at the same time as for forage.

Table 6. Internal rates of return of Atriplex plantations in Tunisia ^a.

^a Project life is: 20 years; shadow price is US $0.16/FU.

Table 7. Cost prices per FU on Atriplex plantations in Tunisia ^a.

^a Project life is: 20 years; DP cost price is assumed to be 1.5 times higher than FU cost price.

^b FU/DP cost prices are lower than shadow prices.

To sum up, Atriplex plantations have an obvious advantage over cactus plantations, and in Tunisia at least, allow browse with a high nutritive value (good protein content) to be produced at moderate cost prices as soon as yields are over 1000 FU/ha.

4. Acacia plantations

Acacia is a genus of the Leguminosae family, probably consisting of over 900 species which vary considerably in appearance from low bush to tree, and which are distributed almost throughout the arid tropical zones of the world. Many of them adapt well to sandy environments and can be used for stabilizing sand dunes and protecting the environment, while others, such as Acacia albida, have long been grown in agropastoral systems, usually involving millet, or raised for the production of gum as in the case of Acacia senegal. About a dozen of these species are cultivated for their leaves and pods, which form an excellent quality forage, rich in protein and phosphorus although sometimes poor in glucides. Browse species of acacia can grow in very different kinds of environment, a factor which makes their classification from the point of view of economic performance difficult. In the section which follows their economic and financial performance is examined in four specific cases.

4.1 Acacia cyanophylla plantations in Tunisia

Acacia cyanophylla, as well as, to a lesser degree, other species related to the phyllodineous Australian acacias (A. salicina, A. ligulata, A. pycnantha, A. cyclops and A. pendula), have been used in forestry for 40 to 60 years on many thousand hectares in North Africa from the humid coastal strip to the arid zone, mostly for fixing coastal sand dunes but sometimes those of the interior also. The browse output from leaves, branches and pods usually lies between that of cacti and Atriplex, and may reach 6 t of DM per ha on well managed plantations (2000 FU). On the other hand, growth of the grass stratum is low and may even approach zero when the planting density is around 800 to 1000 trees hectare. Wood production is high, and its quality is better than Atriplex ⁸. The tree can be used by cutting and carrying, by trimming followed by direct grazing, or else by direct browsing with periodic pruning. The impact of the various management methods on long-term productivity is, however, little understood, and experiments to compare them have yet to be carried out.

⁸ Producing 4500–5000 Kcal/kg of dry wood.

Establishment costs for a plantation with a density of 1000 bushes per ha are about US $600/ha (US $560 for the first year and US $50 for the second), using 125 man-days. In addition, enclosure costs are about US $150/ha in Tunisia. Operating costs are fairly low when direct grazing after trimming is the management method adopted. They are estimated at US $12/ ha for patrolling, and US $12 for cutting and transporting 2.5 t of wood, the same figures as for Atriplex.

Production, mainly involving browse and wood, can generally begin after 3 years, building up gradually to full development towards the sixth or seventh year. Expressed in terms of DM, wood and leaf production are identical, such that 3 kg of wood are produced for each FU. The feed value of the leaves is around 0.33 FU/kg of DM, with 200 g of DP/FU. A properly managed plantation above the 200 mm isohyet annually produces 3000 kg of DM in phyllodes and 3000 kg of wood per ha. However, four different assumptions as to production have been chosen, so as to represent management systems of varying efficiency (Table 8). On this basis, and taking into account the shadow prices of acacia browse and wood, the value of output varies between US $127 and US $510 per ha, with 20% attributable to wood and 80% to browse. It has also been assumed that the grass stratum would be unable to grow at this plant density, so that the value of the original grazing is lost throughout the development period, amounting to US $42/ha on the basis of an output of 220 FU with a DP content of 125 g/FU.

Table 8. Production of wood and browse on Acacia cyanophylla plantations in Tunisia.

^a Based on a value of 0.33 FU/kg of DM.

^b Based on a value of 250 g of DP/FU.

^C Shadow prices of US $0.16/FU and US $15.6/t of wood.

At equal production levels, and despite a higher shadow price of wood, IRR levels are lower than for Atriplex plantations (Table 9). They are below 5% when browse production is 500 FU/ha, and are close to or slightly above 10% when it reaches 1000 FU (3 t of DM/ha), a level roughly equivalent to that achieved on properly managed plantations. The IRR settles at around 20% when production reaches about 6 t of DM per ha (2000 FU). The FU and DP cost prices at this level are fairly low for a 10% opportunity cost of capital, especially if wood is produced at the same time as browse.

Table 9. Internal rates of return of Acacia cyanophylla plantations in Tunisia ^a.

^a Project life is 20 years; shadow prices are US $0.16/FU, US $0.24/kg of DP and $15.6/t of wood.

When browse production is around 1000 FU/ha the FU and DP cost prices are only lower than shadow prices when certain favourable conditions are also fulfilled, namely when IRR is 10% and no enclosures are used, or when IRR is 10% and both wood and forage are produced if the plantations are enclosed (Table 10). Finally, when production is very low (500 FU/ha), FU and DP cost prices are 2.5 to 6 times higher than shadow prices. Cost prices are thus highly sensitive to variations in yields, when these are low.

Table 10. Cost price per fu on Acacia cyanophylla plantations in Tunisia^a

^a Project life is 20 years; DP cost price is assumed to be 1.5 times higher than FU cost price.

^b FU and DP cost prices are lower than shadow prices.

4.2 Phyllodineous acacia plantations in Senegal

Hamel has carried out estimates on the Senegalese production of phyllodineous acacias from Australia (A. holosericea and A. linaroides), species which seem to have acclimatized successfully so far, although they have not been tested in the more demanding zones.

In this country the establishment costs of a plantation with a density of 1000 bushes per ha have been estimated at around US $500 to 600⁹, a figure slightly below those for Acacia cyanophylla plantations in Tunisia. However, the costs of barbed wire fencing are much higher. Various evaluations indicate that the cost of barbed wire fences averages US $450/ ha (on the basis of a unit cost of US $3/metre), a figure three times as high as for Tunisia. Hedges, on the other hand, work out at a cost probably very close to that estimated for thorn hedges in Tunisia, about US $150/ha for a double-row thorn hedge.

⁹ These costs are distributed as follows:

Operating costs have not been specified, but it can be estimated that they must be close of those for Atriplex and Acacia cyanophylla plantations in Tunisia, namely US $12/ha for patrolling and US $12 for cutting and transporting 2.5 t of wood, i.e. US $20 per year for an output of 4 t of wood.

According to Hamel's estimates, browse production can begin in the third year, with output at 1250 kg of DM/ha, while wood production at 2 t reaches 2000 kg during the fourth year and 2500 kg in the fifth year. As indicated above, FU and DP shadow prices can be estimated respectively at US $0.16 and US $0.24. On this basis the value of output at full development is US $332/ha. A lower production assumption than Hamel's has nevertheless been adopted, according to which browse and wood production are sustained at 1250 kg of DM and 2 t of wood per ha throughout the development period (Table 11). It has also been anticipated that non-utilization of the grass stratum during the pre-development period results in a primary production loss of 200 FU and 25 kg of DP with a value of US $38/ha.

Table 11. Production of wood and browse on phyllodineous acacia plantations in Senegal.

^a Based on a value of 0.4 FU/kg of DM.

^b Based on a value of 200 g of DP/FU.

^C Based on shadow prices of US $0.16/FU, US $0.24/kg of DP and US $15.6/t of wood.

Although investment costs for barbed wire are very high, the IRR is over 15% for plantations with enclosures of this kind, while browse production is 2500 t of DM/ha (Table 12). It even reaches 29% when no enclosures at all are involved. As observed for cactus plantations in Tunisia, thorn hedges, although three times cheaper than barbed wire fencing, only marginally improve the IRR. When browse production is low (500 FU/ha) the IRR is below 10% for plantations with enclosures, and FU cost prices reach US $0.25 to US $0.35 in the least favourable cases (Table 13). On the other hand, when enclosures are not used the IRR reaches 16% for this level of production. However, in this case the FU and DP cost prices are above shadow prices if, for a 15% opportunity cost of capital, wood is not produced together with forage (Table 13). A similar situation is found — IRR above 15%, cost prices higher than shadow prices — when, given a 15% opportunity cost of capital and production at 1000 FU/ha, the plantations are equipped with barbed wire fences or thorn hedges. In most cases where FU cost prices are lower than shadow prices, these tend to lie at about US $0.10 to 0.15, only falling to below US $0.05/FU under the most favourable conditions, namely with IRR at 10%, no enclosures and production at 1000 FU/ ha.

Table 12. Internal rates of return of phyllodineous acacia plantations in Senegal^a                                                                 -in%-

^a Project life is 20 years; shadow prices are US $0.16/FU, US $0.24/kg of DM and US $15.6/t of wood.

Table 13. Cost prices per FU on phyllodineous acacia plantations in Senegal^a                                                                              -in US$-

^a Project life is 20 years; DP cost price is assumed to be 1.5 times higher than FU cost price.

^b FU and DP cost prices are lower than shadow prices.

4.3 Acacia senegal plantations in Senegal and Sudan

The main production objective in plantations of Acacia Senegal is gum arabic. Forage output is low, the tree only being available for browse outside the gum collection period, in other words at the beginning of the rainy and dry seasons, giving only a marginal value to the forage as a supplementary source of feed supplies during drought periods. Two plantations have been subject to particular study: both are found in the Sahel zone in areas receiving some 450 mm of rain per year, one in Senegal and the other in Sudan.

The M'Bidi plantation in Senegal covers several tens of hectares, at a density of 620 trees/ha (spacing of 4 by 4 metres). According to Koné (personal communication), establishment costs were probably about US $330/ha and were spread over 3 years (US $240 during the first year and US $30 during the 3 following years). The costs do not include enclosures, which we estimate at US $450/ha or US $150/ha according to whether barbed wire or hedges are used. The El Obeid plantation in Sudan has, according to Seif el Din (personal communication), a density of 610 to 630 trees per ha, similar to the figure for M'Bidi in Senegal. However, establishment costs for this plantation were estimated at under US $200/ha (as against US $330 in Senegal), spread over 2 years at US $160-180 for the first year and US $20 for the seconds ¹⁰ to this figure should be added enclosure costs on the same basis as those outlined above, namely US $450/ha for barbed wire and US $150 for hedges.

¹⁰ First year:
Production and transport of young plants
(US $0.046 to 0.062/plant)                                                       $28 - 39/ha
Soil preparation, bush clearance, burning off
(5 man-days)                                                                           $20
Planting, digging of seed holes, making a watering
basin round each plant (10-12 man-days)                              $40 -45
First weeding, with a hoe (12 man-days)                                $48
Second weeding, with a hoe if necessary
(3.5 man-days)                                                                        $20

Second year:
Selective weeding, by hand or tractor on sandy
soil 15 to 20 cm in depth (US $15/ha)                                      $20
                                 Total                                                   $176 - 192

Apart from patrolling costs estimated at US $12/ha, operating costs in both cases primarily concern the harvesting and transporting of gum. They are estimated to correspond to 5.6 man-days per ha for harvesting the gum, an average of US $66/ha.

Gum production is estimated at 250 g/tree (155 kg/ha) from the fifth year onwards and over a period of 35 years, at the end of which the plantation has to be entirely renewed. The trees are therefore cut down and the wood is sold in bulk at the end of the production period. Browse production, consisting of leaves, can be estimated at 200 FU/ha with a DP content of 200 g/FU. Every tree thus gives a yearly total of 1 kg of DM of leaves with a feed value of 0.33 FU/kg of DM. In short, the annual value of production at full development lies at US $163/ha, with the salvage value of wood when the plantation is cut down at the end of the production period standing at some US $290/ha (Table 14). Also taken into account is the fact that the grazing land taken up by the plantation cannot be used throughout the pre-development period of 5 years. The forage production lost has been estimated at 150 FU/year with a protein content of 125 g of DP/FU.

Table 14. Browse, wood and gum production on acacia senegal plantations in Senegal.

^a Salvage value on the basis of 30 kg/tree at end of project life.

^b In kg of DM.

^C Based on a value of 0.33 FU/kg of DM.

^d Based on a value of 200 g of DP/FU.

^e Based on shadow prices of US $0.16/FU, US $0.24/kg of DP, US $0.8/kg of gum, and US $15.6/t of wood.

Table 15. Internal rates of return of Acacia senegal plantations in Senegal and Sudan^a                                                 -in%- 

^a Project life is 20 years.

The IRR, which is slightly higher for the El Obeid than for the M'Bidi plantation owing to lower investment costs, is under 10% when the plantations are equipped with enclosures and between 10 and 13% when no enclosures are used. Acacia senegal plantations are primarily intended for gum production, and consequently it has been assumed that in contrast to the former cases browse production could take place at the same time as gum production, or at least be treated as a by-product of the latter. In this case the FU and DP cost price, designated the marginal cost price (see Table 16), is the one enabling the discounted cash flow from net expenditure to be balanced, when profits from the sale of gum are deducted.

Table 16. Cost price per FU on Acacia senegal plantations in Senegal and Sudan^a                                                                                                    -inUS $-

^a Project is 20 years DP cost price is assumed to be 1.5 times higher than FU cost price.

^b Simultaneous production of gum and browse.

^C Browse treated as a by-product of gum.

^d The cash flow from FU and DP losses is higher than from browse production.

^e Gum production alone is enough to ensure financial equilibrium at an IRR of 10%, so that the FU cost price can be rated at zero.

^f The FU cost price is lower than the shadow price.

According to the results in Table 16 the average FU and DP cost prices are always higher than the shadow prices, even in the most favourable cases when no enclosure costs have to be included and when the opportunity cost of capital is 10%. The discounted cash flow resulting from FU gains and losses becomes negative when the plantations are hedged (Table 17), since hedges prolong the unproductive period. Treating browse as a by-product of gum improves the situation when the plantations are not enclosed and the opportunity cost of capital is 10%. The marginal FU cost price falls in this case to US $0.10 and even to zero in Sudan, where gum production by itself is enough to ensure financial equilibrium this level of IRR. Acacia senegal plantations thus appear to offer only a very marginal interest in terms of browse production.

Table 17. Current value of the FU and DP (kg) lost and gained.

4.4 Acacia albida plantations

Acacia albida is very widespread between sea level and 1800 m in the semi-arid and subhumid zones of tropical Africa. Unlike most deciduous tropical trees it loses its leaves during the rainy season and keeps them throughout the dry season. It is usually planted in areas with an annual rainfall of 400 to 800 mm, in millet, cowpea and groundnut fields, forming part of a traditional agricultural system used by several African peasant civilizations. Millet is cultivated beneath and between the trees. The interception of light by the foliage during the millet growing season is negligible, so that shading has no effect on the crops growing at the foot of the trees. Also, as the leaves fall at approximately the same time as ploughing begins, most of them are ploughed into the soil and act as fertilizer, improving crop yields. It has been estimated that millet yields can be multiplied by 2 to 2.5 in fields planted with acacia, in comparison with yields in open fields without fertilizer. Under the conditions found in the semi-arid zones of West Africa this means that millet yields can increase from 500–800 kg/ha to around 1000–1500 kg/ha. In addition, pod production, reaching an average of 200 to 600 kg/ha/year, provides livestock with a protein-rich feed to supplement their roughage. Acacia albida pods were being sold at 45 CFA francs (US $0.21) per kg in Senegal in 1980, around US $0.27/FU (1 kg represents 0.7 FU).

Acacia albida generally lives to about 80 years. The tree becomes productive only after 15 years and is adult at the age of 25. Fruit production is therefore zero between 0 and 15 years, then gradually increases. The fertilizer effect, which also increases gradually, does not begin to show before the trees reach 10 to 15 years, and probably takes about 10 years thereafter to reach its full extent, at about 25 years when the tree is mature. The final plant density of Acacia albida is usually 10 to 50 trees per ha, but given the length of the production cycle it is preferable to have a higher density at the outset in order to meet losses, with subsequent elimination of some of the trees as they become adult.

Planting costs can be estimated at US $1 per plant, i.e. US $0.55 for planting as such and US $0.03/year for protecting the trees until they become productive, that is until aged 15 years. These figures add up to a total of US $100 for an initial planting density of 100 trees/ha, US $55 for the first year and US $3 for the next 14 years, to which should be added any enclosure costs incurred (US $450/ha and US $150/ha respectively, according to whether barbed wire or hedges are used). No further costs are incurred, apart from collecting the pods and cutting the branches when pruning, but these were considered negligible.

On the basis of the yields estimated above and of shadow prices of US $0.16/FU and US $0.24 per kg of DP, somewhat lower than the price for pods quoted on local markets in 1980, the value of pod production is US $73/ha between years 15 and 24 and US $146/ha from year 25 onwards. Wood production is very low, consisting of the sale of the 50% of trees eliminated after 20 years. On the other hand, the impact on crop yields is thought to be substantial, with at least twofold increases taking place within 10 years, between years 15 and 25 (Table 18).

Table 18. Impact of Acacia albida plantations on browse, wood and crop production.

^a Based on production of 5 kg of DM per tree between years 16 and 20 (100 trees), 10 kg between years 20 and 25 (50 trees) and 20 kg thereafter.

^b Based on a doubling of yields in 10 years. `

^c At a rate of 100 kg per tree eliminated.

^d Based on a nutritive value of the pods of 0.70 FU/kg of DM. `

^e Based on a DP content of 200 g/FU.

^f Based on shadow prices of US $0.16/FU, US $0.24/kg of DP, US $0.20/kg of millet/sorghum and US $15.5/t of wood.

Despite the length of the production cycle and the delay before Acacia albida plantations become productive, the IRR is generally higher than 10%, except when the plantations are supplied with barbed wire fencing (Table 19). It is even close to 20% when no enclosures are used. This result is astonishing, if the very long-term impact on crop and animal production is taken into account, although this positive aspect has hardly any impact on IRR when the depreciation period is extended to 80 years. The reasons why the Sultan of Zinder, in Niger a century or so ago, decreed that anyone cutting down an Acacia albida tree would himself be decapitated can easily be understood.

Table 19. Internal rates of return of Acacia albida plantations ^a  -in%-

^a Based on shadow prices of US $0.16/FU, US $0.24/kg of DP, US $0.20/kg of millet/sorghum and US $15.6/t of wood.

^b Project life.

Since the value of a cash flow often becomes negligible after 30 years, the FU cost price calculation has been limited to a period of this length, although the actual life of Acacia albida is 80 years. Over 30 years, browse production alone is enough to guarantee financial equilibrium, even with an IRR of 15%, when no investment is made in enclosures. Taking into account the improvement in agricultural productivity lowers the FU and DP cost prices to below the shadow price when plantations are provided with thorny hedges (but not barbed wire fences) and the opportunity cost of capital is 10% (Table 20). In all other cases the FU and DP cost prices are well above the shadow prices adopted for the IRR calculation. In the most favourable cases the FU cost prices are US $0.03 and US $0.06, subtracting the improvement in crop yields.

Table 20. Cost prices per FU and per kg of DP on Acacia albida plantations^a                                                                         -in US$-

^a Project life is 30 years.

^f FU and DP cost prices are lower than shadow prices.

5. Other browse tree and shrub plantations

Although their characteristics have been more widely studied and are therefore better understood, cacti, Atriplex and acacia are not the only species used for livestock feeds in Africa. Other examples can be cited, especially Prosopis and Leucaena, which have characteristics not fundamentally different from those analysed above.

5.1 Prosopis plantations in the Cape Verde Islands

Prosopis species closely resemble acacias in terms of the nutritive value of their leaves, branches and pods, although they tend to be less rich in protein (approximately 100 g of DP/ FU). Several species have been grown for browse in various arid zones throughout the world, especially Prosopis chilensis and P. juliflora in the arid tropics of Africa and the Middle East, P. cineraria in Asia and P. tamarugo in Chile. In the latter country some 30,000 ha of P. tamarugo have been planted in the northern Atacama desert, with a density usually around 100 to 120 trees per ha. According to Robertson (1980), the establishment and maintenance costs of a plantation during the pre-development period are around US $0.32 per plant, i .e. US $32/ha for a density of 100 trees. Again according to Robertson, browse yields were 7 t/ha/year.

According to Sabra (personal communication) establishing a plantation of Prosopis juliflora and Parkinsonia aculeata in an area of the Cape Verde Islands receiving 100 to 300 mm of rain per year costs on average about US $0.526 per plant for 500,000 trees, in other words US $210/ha for an average density of 400 trees per ha (plant spacing of 5 by 5 cm), which is thus far higher than the Chilean examples just cited. The costs are spread over 2 years, at US $150 for the first year and US $60 for the second. They include raising the plants in the nursery and transporting them to the planting site, soil preparation, digging of holes, water conservation measures, and finally, watering after planting out. Obviously no investment costs for enclosures are included.

The minimum production estimated in preliminary studies (measurement of primary production and evaluation of secondary production) was around 2 t of DM, of which 1 t was actually consumable, i.e. 350 FU/ha or 0.35 FU/kg of DM between the sixth and twentieth year of the project (Le Houérou, 1980; Lepape, 1980), allowing a stocking rate of 1 sheep/ha/year. Wood production can be estimated at 400 kg of DM/ ha/year, i.e. 13 kg per tree (Table 21).

Table 21. Production of wood and browse on Prosopis and Parkinsonia plantations in Cape Verde islands.

^a Based on a value of 0.35 FU/kg of DM.

^b Based on a value of 100 g of DP/kg of DM.

^C Based on shadow prices of US $0.16/FU, US $0.24/kg of DP and US $15.6/t of wood; the development period is from year 6 onwards.

Operating costs can be estimated at US $5/ha/year for patrolling and US $12 for cutting 2.5 t of wood, i.e. US $5.5/ha during the development period. It has also been assumed that the production losses due to non-utilization of the grass stratum during the pre-development period amount to 100 FU/ ha/year with a DP content of 125 g/FU.

The IRR is higher than 20% when the plantation is without enclosures, falling to somewhat less than 10% when barbed wire fences are used and 12% for thorn hedges (Table 22). Given the relatively low production levels it is possible to achieve in the arid zone, installing enclosures, especially barbed wire, would remove some of the financial attractiveness from this operation. Average FU and DP cost prices, often higher than the shadow prices, nevertheless fall to under US $0.10/FU (or even US $0.05) when plantations are left without enclosures and the opportunity cost of capital is 10% (Table 23).

Table 22. Internal rates of return on Prosopis plantations ^a      -in%-

^a Project life is 20 years shadow prices are US $0.16/FU, US $0.24/kg of DP and US $15.6/t of wood.

Table 23. Cost prices per FU and per kg of DP on Prosopis plantations in Cape Verde islands a                                            - in US $/FU -

^a DP cost price is assumed to be 1.5 times greater than FU cost price.

^b FU and DP cost prices are lower than shadow prices.

5.2 Leucaena plantations in Malawi

Leucaena leucocephala is a leguminous tree of central American origin, found only in the subhumid and humid tropics between 600 and 1800 mm of rainfall. It can be used for producing both wood and browse. It gives not only a browse product of high quality, with a protein content similar to that of lucerne (25% CP), but also a high yield, reaching 20 t/ha/year, Leucaena nonetheless contains a toxic amino acid, mimosine, which limits the proportion it is advisable to include in the feed ration of ruminants to a maximum of 30%, expressed in DM.

A project involving small plantations managed by peasants was launched in Malawi during the 1970s in order to boost the national production of feed bran. According to Beale (1980), investment costs can be estimated at US $272318/ha, of which US $93 to 151 were devoted to the enclosures deemed necessary to protect the plants from being damaged by goats. However, these costs would probably have been considerably reduced (perhaps by 50%) if large-scale plantations had been involved, allowing substantial economies of scale, especially for enclosures. Operating costs were estimated at US $66/ha/year for a production of around 2 to 3 t of DM per ha. The leaves are purchased from producers at US $0.08/ kg of DM, while the wood is not sold commercially, being used for drying the leaves. The sale of the latter shows a return of US $160 to 240/ha/year for the smallholdings financed by the project.

As a cash crop Leucaena leaves were considered the best production alternative, allowing small farmers to make a net profit of US $153/ha, whereas under local conditions profits from mixed farming are barely over US $120/ha.

7. Conclusions

It is difficult to throw very much light on the economic value of browse tree and bush plantations owing to the unreliable nature of the data used for the IRR calculations. The most consistent estimates are those relating to operating costs, but this is primarily because they were usually carried out by ourselves owing to the lack of concrete information. In general terms it was assumed that these costs were fairly low owing to the extensive nature of the plantations studied, the fact that labour requirements to protect them and ensure their productive output are fairly modest, and that labour costs in Africa are in any case low. It is possible that operating costs have been underestimated in several cases, a factor which will have overvalued profitability.

Estimates for investment costs, largely based on concrete examples, are more widely dispersed. In general terms the cost of establishing a plantation in Africa appears to stand at between US $200 and 500/ha excluding enclosures, with some exceptionally high value (Acacia cyanophylla, US $610) and others which are very low (A. albida, US $55; Leucaena leucocephala, US $150). Unit costs nevertheless appear less widely dispersed, usually varying around US $0.50 per plant, with some values appreciably lower (US $0.20-0.25) when the planting density is high, as for cactus and Atriplex, or when planting is carried out by direct drilling and not by transplanting young plants raised in the nursery (Leucaena leucocephala and A. semibaccata). Direct drilling, which is very much less expensive, cannot be used for the time being in the semiarid and arid zones of tropical Africa, where it is too risky. In areas where rainfall is unreliable even transplanting nursery plants is not without risk of failure, and so requires special care (the plants must be watered) to ensure survival. As a result tree and shrub plantations in these areas are fairly expensive. A second factor further increases the establishment cost of plantations in Africa, namely the cost price of barbed wire fencing. An alternative solution can be used, consisting of planting thorn hedges. However, this solution has the drawback of prolonging the non-productive pre-development period. In short, it may be estimated that in Africa a browse tree or bush plantation with a density of 1000 stocks per ha costs on average US $500/ha without expenditure on enclosures, and that establishment costs can double when the plantation is equipped with barbed wire fencing.

Evaluating production also posed a number of problems. The yields of browse trees and shrubs vary enormously under different ecological and management conditions, so that sometimes we have been led to assume several different levels of production in order to cover the range of yields possible at given levels of technology. To this initial cause of uncertainty may be added the further difficulty of estimating forage shadow prices. The criterion adopted, prices on the world market, has certain drawbacks since the resulting shadow prices are fairly high given animal feed conditions in Africa. Naturally the prices adopted have some impact on the profitability of the operation under study. In any case these operations can probably only be justified if browse plants are treated as a supplementary feed supply to replace concentrates during a drought period, which is after all their most important function.

A related problem arises in the form of non-utilization of the grass stratum, the value of which has been estimated on the basis of the same shadow prices as those used to assess browse production, a questionable assumption since the grass stratum can never fulfill the same function as browse plants. As has been seen, this assumption was based mainly on the idea that grass is becoming scarce in tropical Africa, so that browse should not be produced to the detriment of grass but rather as a complement to it, enabling livestock to be supplied with forage throughout the year at reasonably low cost. Estimating cost prices on the basis of the more easily determined opportunity cost of capital not only obviated the need to rely on shadow prices which were more or less insignificant, but also provided a range of prices, which turned out to be a more useful approach. Uncertainties regarding investment costs and the level of browse yields have also been used to analyse the sensitivity of browse cost prices, with the different figures being considered as variables illustrating profitability thresholds, which can thus be determined.

The IRR thus appears extremely sensitive to variations in investment costs and browse yields, especially when these are low. This emerges clearly from Table 24, which reproduces variations in average IRR for different plantations as a function of investment and yield variations. According to these results the IRR falls by 0.6% when investment costs rise by 1%. IRR variations in relation to browse yields are even stronger, especially when these are low. The IRR thus doubles when browse production moves from 500 to 1000 FU/ha, and increase by almost 70% when production rises from 1000 to 2000 FU/ha. If a level of 10 or even 15% is identified as the minimum IRR, it appears that on the basis of the FU and DP shadow prices adopted the profitability of browse plantations is assured at all production levels (except in the case of cacti and Acacia cyanophylla) when investment costs do not include enclosures, and that this threshold generally seems to rise to 1000 FU when investments include barbed wire fences. Using hedges instead helps to improve IRR slightly, but even with them it usually remains below 10% (Table 25).

Table 24. Impact of investment costs and yields on IRR for browse tree and shrub plantations ^a                                                                           - in % 

^a Project life is 20 years (for A. albida, 30 years); shadow prices are US $0.16/FU and US $0.24/kg of DP.
^b Average of IRRs.

Table 25. Impact of forage yields and investment costs on IRR of browse tree and shrub plantations in Africa ^a                         -in%-)

^a Project life is 20 years (for A. albida, 30 years); shadow prices are US $0.16/FU and US $0.24/kg of DP.

^b Barbed wire fences/spiny cactus hedges.

^c Barbed wire fences/no enclosures.

^d Barbed wire fences/hedges/no enclosures.

Cactus plantations raise several important points, amongst which the problem of the complementary production of wood or other products, the impact of the management method, whether direct grazing or cut-and-carry, and the utilization of the grass stratum. Among the various examples discussed cactus plantations appear to have the lowest IRR for a given level of browse production. The IRR for directly grazed plantations ¹¹ is 17 to 19% in Tunisia, where output is 2000 FU/ year, fairly close to the IRR on Acacia cyanophylla plantations where output is 1500 FU and on Atriplex plantations where production is 1000 FU. Not only is cactus the only one of the species studied here to produce a rather poor quality feed owing to its low DP content, but also adopting a cut-and-carry system reduces net output of forage while incurring far heavier operating costs (US $0.08/FU produced). Profitability is thus severely handicapped, whatever the production level. For these reasons, unless rather high production levels are achieved for the extensive nature of the plantations and a strictly controlled management method is adopted, namely direct grazing, cactus plantations do not appear to produce browse at sufficiently cost to justify the level of investment required.

¹¹ Assuming that the technical constraint mentioned previously (risk of overgrazing) is removed.

The situation appears better for most of the other browse species. Nonetheless, the extremely high cost of barbed wire fencing in tropical Africa and the prolonged unproductive period caused by introducing hedges handicap the financial viability of the management systems analysed here, given the limits on yields imposed by their extensive nature. As may be seen from Table 26, in which the FU and DP cost prices for an IRR of 10% are outlined, browse production at under 500 FU/ year in no case justifies