(introduction...)

The case studies presented in Chapter 3 include different improved processes for extracting shea nut butter and oil from sunflower seed and oil palm fruit as applied in current projects in Mali, Zambia and Togo. In the present chapter, these projects are financially analyzed in their specific local contexts, which means that all costs and prices involved were taken from recent field experience. It also means, that the results presented are not necessarily the same for other countries or for other times in the future; since, for example, margins between rawmaterial and sales prices, wages and other variables might change considerably.

For each of the case studies, alternatives were calculated, using the same costs for variables like local wages, sales price of oil and by-products, etc. In this way, the profitability of the processes described as case studies can be compared with the technical alternatives.

Finally, for all the case studies and their alternatives a sensitivity analysis has been made, the results of which indicate the profitability of each process in relation to certain critical assumptions. Keeping all others constant as originally, the variables that were modified for this analysis are in each case:

- shea butter/oil recovery of the process,
- capacity utilization of the equipment,
- local wages.

The results are illustrated in graphical form (see Figures 22, 23 and 24). As a guide to interpreting these figures, one might say that the steeper the curve, the more sensitive is the process to a modification of a specific variable.

For all calculations, the same procedure was applied and the same indicator used. The procedure is described in the “Manual for the Preparation of Industrial Feasibility Studies" by the United Nations Industrial Development Organization (UNIDO). Following this procedure, the input-data for all the cases was entered into the computer programme COMFAR (Computer Model for Feasibility Analysis and Reporting), which computes a whole set of output tables. Computations were done for a 10 year project period in each case. For reasons of space and simplicity, the output tables have been reduced to only one indicator.

As the indicator for the results of all case studies, the so-called "internal rate of return" (IRR) was chosen, which is a percentage figure and can - in simple terms- be compared with the long-term interest for a bank deposit. For any investor (i.e. in equipment for oil processing), the IRR is an indicator of whether it is more profitable to put his money in a bank account or to invest in this or that kind of technology and start production under given assumptions. An explanation on how to calculate the IRR is given as Annex 3.

One might argue that the procedure and the indicator are neither appropriate nor relevant for a traditional context where usually no money value is attached to inputs like labour, energy, etc. Nevertheless, a proper financial calculation even for the traditional technologies can introduce a scale on which all alternatives can be measured. In this way, the IRR can serve as a uniformly comparable figure which allows the analysis of profitability. However, it must be kept in mind that it is a capital-based indicator.

4.1 Shea nut processing in Mali

In Mali, the Franc CFA is the accounting currency with F CFA 50 equal to I French Franc. As the discounting rate (long-term bank deposits), 10 % was assumed.

Table 13 provides an overview of the costs and results of the improved process compared to the traditional process for shea butter extraction. The assumptions for the improved (KIT) process are based on the experiences gained so far with the GTZ/GATE/DMA project. The figures are based on field experience in Mali and can be interpreted in the following way:

Table 13: Assumptions for Shea Nut Processing in Mali (per year in F CFA, unless stated otherwise)

For the GTZ/GATE project, land is made available free of costs by the village and as with the traditional process - can not be accounted for. The building needed is a small traditional one with bricks and a roof of corrugated iron. It is to last for the whole project period but has no value afterwards.

The machinery for the KIT process, which for all major parts is expected to last for 10 years, consists of a hydraulic press, an expel stand, oven parts, a few buckets, mortar and pestles. The equipment for the traditional process includes mortar and pestles, basins, calabashes, crushing stones and other traditional material.

As current investments, the traditional equipment has to be renewed every year. For the KIT machinery, a new hydraulic jack has to be accounted for after 5 years.

Production costs are first based on the maximum capacity of the processes, assuming 300 working days per year. For both processes, shea nuts are the only rawmaterial, costing F CFA 30 per kg on average. The KIT process needs hydraulic oil (10 lifers per year at F CFA 1000 per 1), an annual technical check-up of the jack and some spare parts (rubber seals, a pump valve). Since fuel wood has little commercial value in rural areas, its price was calculated as the working time needed to collect it (for the KIT process: 2.5 working hours per production day: for the traditional process: 4 working hours per production day). The wages applied are the same as for the actual oil processing, in which 5women participate with the KIT process and 4 in the traditional one. Marketing costs have not been calculated, since the weekly visit to the market is a social event in any case.

The production programme and sales give an indication of the use of the equipment as a percentage of its maximum capacity. The actual yearly production, however, is determined by the availability of rawmaterial, which can - in the case of wildgrowing shea nut trees - be zero in one year and high in the next.

For the KIT process, it was assumed that 50 women form a group, in which each woman has on average 50 kg of shea kernels collected. Processing 50 kg of kernels per day, this raw- material supply would last for SO working days or 16.6 % utilization of maximum capacity. Assuming an average shea butter recovery of 38.5 %, the process would give 962.5 kg of shea butter per year.

For the traditional process, the 4 women in the group provide together 200 kg of kernels to be processed, which would last for approximately 10 (8 hour) working days or the equivalent to 20 days with 4 actual working hours. The percentage of shea butter recovery was assumed to be 25 % on average, giving 50 kg of shea butter per year. In both cases' the price of the shea butter is F CFA 300.

For the working capital requirements it has to be considered that shea nuts are stored traditionally by rural women as a kind of “bank account'´ for which, however, no interest can be calculated. For the KIT process, hydraulic oil and spare parts should be stored for a full production year.

As far as the source of finance is concerned, it has to be mentioned again that rural women not only have very little capital to invest, but also have difficulties in getting a credit. In the GTZ/GATE project, the total initial investment (building, machinery and equipment) is therefore subsidized with 50 % of its cost (= F CFA 275 000). For the feasibility calculation, however, the subsidy is normally treated as part of the equity capital and has no influence on the profitability of the project.

On the basis of the assumptions described above, which are termed “standard cases" below, the results for both processes are summarized by the calculated internal rates of return (IRR). These are for:

- KIT process 21.91 %,
- traditional process - 8.97 %.

These results mean that the traditional process, quite clearly, is not viable in financial terms, if the critical variables (shea butter recovery rate and wages for processing) are realized as assumed. In practical terms, these results mean that the-women involved in the traditional process are not able to realize the assumed wages, i.e. they work in fact for less money per hour (compare sensitivity analysis below). However, since there are very few employment opportunities for women in rural areas of Mali, even this low income is accepted.

The "standard case" of the KIT process, on the other side, produces a positive IRR of roughly 22 %. This means that the financial return, which can be expected from an investment in this process, is considerably better than a bank deposit (assumed as 10 %). The figure also means that the expected return is quite acceptable compared to what is normally expected from an investment into manufacturing industries in developing countries. Depending on the risk he takes (especially on the marketing side), a private investor would - as a rule - feel fairly safe, if his investment can be expected to produce about double the discounting rate (in this case 20 %). Since marketing of shea butter appears to pose no problems in Mali, the calculated IRR of about 22 % can be seen as well promising.

To illustrate the practical implications of the subsidy in the case of the GTZ/GATE project, it is possible (but not sound in the financial analysis of the process) to calculate the IRR only on the basis of the equity provided by the village: Subtracting the subsidy from the total investment, but still having the same cash inflow from sales, the IRR would be more than 50 %. This figure, of course, would indicate a very attractive return for the women concerned.

As shown in the summary sheet for shea nut processing in Mali (Figure 22), variations of the critical variables produce quite different results for the profitability of both processes:

Figure 22: Summary Sheet, Shea Nut Processing in Mali. Graphs are approximations.

Figure 22, A, indicates the expected IRR in relation to the rate of shea butter recovery. For the standard case of the traditional process, an average recovery rate of 25 % was assumed, which results in a negative IRR. Assuming 30 % recovery for the same process, the IRR would be just over 13 % (i.e. positive). A recovery rate of about 27% would make the IRR zero; i.e. the process would at least not produce financial losses (i.e. a negative, discounted, cummulated net cash flow).

For the KIT process, the same value (IRR = 0) starts with a recovery rate of about 25 %. Since the investment in this case is relatively big, however, so-called opportunity costs should be considered (at a discounting rate or IRR = 10). The KIT process, then, becomes financially interesting, if recovery rates of 32 % and more can be realized. At a recovery rate of 42 %, which is reportedly about the best possible, the IRR would be about 28 %.

Figure 22, B, indicates the expected IRR in relation to capacity utilization (percentage of maximum capacity at 8 working hours per day and 300 working days per year). As mentioned earlier, the bottleneck factor for the utilization of the equipment might be the availability of the rawmaterial, i.e. how many shea nuts the women can collect on average. For both processes, 50 kg of kernels per woman per year had been assumed as the standard case. However; the collected amount of nuts is reported to be higher in many cases. On the other hand, less raw material might be available in years when shea nut trees bear only little fruit. A long-term deterioration of this situation due to a shortage of shea nuts (possibly in consequence of ecological problems such as desertification) has not been considered for the analysis, but might be of importance in the future.

For the traditional process, this variation has (almost) no influence on the IRR, because the investment for equipment, etc. is minimal. For the KIT process, however, an increased shea kernel supply of 60 kg per woman per year (equivalent to 20 % capacity utilization) would bring the IRR to about 30%; a reduced shea kernel supply of 40 kg (13% capacity utilization) would reduce the IRR to about 15%. For the 50 women in the GTZ/GATE project, the KIT process requires a minimum of about 14 kg shea kernels per woman per year to result in an IRR of 0.

Figure 22, C, indicates the expected IRR in relation to wages calculated for collecting firewood and for the actual shea nut processing. In the financial analysis of both processes, labour is the most important production cost factor. Seen from the perspective of the concerned women, the money realized from selling shea butter (wages) is a direct indicator of their available cash income and/or an indirect indicator for the additional spare time.

As can be seen from the graphs, both processes are highly sensitive to labour costs. Most interesting seems the point at which the IRR becomes 0; i.e. the maximum wages that can be realized without running into financial losses. For the traditional process, this point is reached at about F CFA 22; for the KIT process, this point is reached at about F CFA 70 (or F CFA 50 considering the opportunity costs of capital).

For the women participating in the GTZ/GATE project, this result means that their available cash income is about three and a half times the previous (traditional) level without increasing their time spent on processing the nuts. Since the process is limited mostly by the available rawmaterial, the final result for the women could also mean that they can save up to 70 % of the time previously spent on shea nut processing. To give an example: To earn F CFA 1000.- the women in the traditional process have to work for about 45 hours; with the KIT process, it takes only about 13 working hours.

4.2 Sunflower seed processing in Zambia

In Zambia, the Kwacha (K) is the accounting currency. Although the Kwacha is currently under strong pressure, the exchange rate was taken as 8 K for the US $ to calculate the investment in machinery. As the discounting rate (long-term bank deposits), 10 % was assumed. For a loan from a local bank (6 years, I year grace period), an interest rate of 20 % was seen as realistic.

Table 14 is to give an overview on costs and results for the use of the handoperated equipment developed by KIT, and as alternatives the expellers from Simon-Rosedowns and Reinartz. The figures are based on field experience in Zambia and can be interpreted in the following way:

Table 14: Assumptions for Sunflower Seed Processing in Zambia (per year in Kwacha, unless stated otherwise)

As in Mali, land in rural areas in Zambia can be assumed to be provided free of costs by a group or the village. The building would in all cases be of rather small size and is assumed to have a residual value of 50 % after the project period.

The machinery for the hand-operated process consists of a decorticator, a winnower, a roller mill, a heating oven, a spindle press and other equipment like weighing scales, buckets, etc. The two technical alternatives consist both of diesel-powered expellers, the first from Simon- Rosedowns (MINI 40), the second from Reinartz (AP VII). Both expellers (not the diesel engines) are to last for ten years and the delivery prices include spare parts for two years.

Since seasonal price fluctuations for sunflower seeds are considerable, a one-year stock of rawmaterial has to be bought at low prices just after the harvest before the first production year. A second seed stock-up of the same amount is due as a current investment within the first production year to be able to double production from the second year onwards (see production programme below).

Further current investments refer to some equipment to be renewed every year (cleaning sieve, scales, buckets, drums etc.), and machinery parts after 5 years comprising a new gear box for the decorticator and for the two technical alternatives, new diesel engines.

Production costs involve the annual expenses for the raw- material seeds (except in the 10th year of production). For the two expellers, utilities (diesel, greasing oil, etc.) have to be calculated; for the hand-operated equipment, fuel wood (as valued by the wages for collecting) applies. Wages for the hand-operated process can be calculated lower (6 men with 4.5 K per day) than those for the alternatives (2 men and 4 men respectively with 10 K per day), since the equipment demands at least semi-skilled labour. In all cases, costs for maintenance and spare parts were calculated, marketing costs refer to the common practice of door-todoor sales or running a small shop.

The production programme in all three cases is geared to 40 % of maximum capacity for the first year of production, to be doubled in the second year and afterwards kept on this level until the end of the project period. Some caution is, however, advisable in trying to apply this assumption to the real Zambian context since no thorough market investigation could be done for this publication to verify whether the whole production (especially of the Reinartz version? could be easily sold. In all cases, the oil-cake as a by-product has a significant impact on the feasibility of the projects.

Apart from the stock-up on rawmaterial, working capital is required for utilities and spare parts (bearings, scrapers, leather belts for the first case; choke rings, rings, screws for the second; and scrapers, parts of the screw, cage-bars, etc. for the third). The finished products, which are usually sold only once every two weeks, have to be stored for 7 days on average.

In all cases, a local loan will probably be necessary in addition to equity capital to cover the investments. As can be seen from the figures in the above table, the total volume of the investment, at least for the Reinartz expeller, is not the smallest scale any more and could probably only be managed by a single investor or a well organized formal cooperative.

The profitability of the three versions of equipment for sunflower seed processing is again given by the internal rate of return (IRR) as the major indicator for the financial evaluation. The following figures were calculated:

- hand-operated equipment 22.47 %,
- Simon-Rosedown expeller 7.54 %,
- Reinartz expeller 26.55 %.

Under the assumptions described above (standard cases) for Zambia, the Simon Rosedowns expeller cannot be seen as a feasible enterprise, partly due to expensive spare parts. The hand-operated equipment undoubtedly has its strength for small-scale processing with little capital to invest and to risk. The Reinartz expeller requires a sizable market outlet, considerable volumes of rawmaterial supply and investment and management skills in general, but has its advantages due to a relatively good rate of oil recovery.

As for Mali, the standard cases in Zambia have been further analyzed by modifying some critical input data. The results of this sensitivity analysis are illustrated in Figure 23 and can be summarized as follows:

Figure 23: Summary Sheet, Sunflower Seed Processing in Zambia. Graphs are approximations.

Assumed standard wages per day:

- 100 % = 4,5 K for handoperated process (unskilled labour)
- 100 % = 10 K for expeller (semi-skilled)

Figure 23, A, indicates the expected IRR in relation to the rates of sunflower oil and oilcake recovery. (Recovery rates of oil and cake are seen as complementary; i.e. if in the modifications the oil recovery rate was reduced by 5 %, the oilcake recovery would be increased by 5 % and vice versa. The difference is, of course, the sales price of the two products.) Depending on the quality (oil content) of the seeds and the condition of the equipment, the oil recovery varies over a considerable range. As illustrated by the graphs, all three processes are financially very sensitive to varying recovery rates.

For the hand-operated equipment, an oil recovery rate of 20 % was assumed as the standard case, for which about 22 % IRR can be expected. The same process would produce more than 39 % financial return with 25 % oil recovery; below about 13 % oil recovery, the project would run into financial losses.

The Simon-Rosedowns equipment, which is just barely feasible to operate with standard case assumptions (20 % oil recovery) would make a financial loss with oil recovery rates below 17%. The process, however, gives good results (IRR more than 20 %), if the recovery rate can be brought up to 25 %.

The process with the Reinartz expeller shows an IRR = 0 point at a minimum recovery rate of about 15%. Within a more realistic range (23 % to 33 % oil recovery, taking 28 % as the standard case) the process gives very attractive financial returns of between 14 % and 37 % (27 % for the standard case).

Figure 23, B, indicates that all three processes are less sensitive to varying rates of capacity utilization which can be explained by the high proportion of variable costs in the production. As for the standard cases, all the variations illustrated by the graphs assume that for the first production year actual production is only half of what is produced from the second year onwards.

At full production (100 % capacity utilization from 2nd year), the process with the Reinartz expeller gives the best financial returns (34 %), followed closely by the hand-operated process (29 %). The Simon-Rosedowns equipment performs still rather poorly (IRR = ll.5 %) under these optimal conditions.

Figure 23, C, shows the relation of the IRR to the wages calculated for collecting fuel wood and the actual processing. The hand-operated equipment requires, of course, the most labour-intensive process and is therefore the most sensitive one in terms of labour costs (or wages realized). In this process, maximum wages of about 13.5 Kwacha per day could be paid before the project would run into financial losses. For the process using the Simon Rosedowns expeller, this point would be reached at around 14.5 Kwacha per day. The Reinartz process is so capital intensive that wages play only a minor role for its profitability.

4.3 Oil palm fruit processing in Togo

As in Mali' the Franc CFA is the accounting currency in Togo. A 10 % discounting rate and loan conditions of 18 % (6 years, I year grace period) were assumed.

Table 15 gives an overview on costs and results for the use of the hand-operated equipment developed by KIT, and as alternatives the process developed by TCC, Ghana, and the CALTECH expeller. The figures are based on field experience in Togo and can be interpreted in the following way:

Table 15: Assumption for Oil Palm Fruit Processing in Togo (per year in F CFA, unless stated otherwise)

As for the other case studies, land in rural areas in Togo is assumed to be available free of costs. The building in all three alternatives consists of a simple, large (60 m²) shed, which would have no residual value after 10 years.

The machinery, for the hand-operated process consists of a UNATA spindle press, cooking kettles, buckets and other equipment like oil containers and pounding sticks, which are to be renewed every year. For the TCC process, one TCC pounding machine, two TCC screw presses, one large cooking kettle and some buckets are required. The engine of the pounding machine is to be renewed after 5 years. The CALTECH process only includes as expenses for machinery the expeller and some drums. The expeller is driven by a petrol engine, which probably has to be renewed in the fourth and seventh year of production. A new screw for the expeller should be accounted for after 5 years.

Production costs at maximum capacity (250 working days per year) would involve loose oil palm fruits as the only raw- material in the above given quantities and for a price of F CFA 35 per kg. For the motor-driven alternatives, costs for utilities are for diesel/petrol and lubricants. The costs for fuel wood are again calculated on the basis of the wages for collecting. Wages for the hand-operated and the TCC process cover 8 women at F CFA 400 per day; with the CALTECH version, 2 women at F CFA 400 per day and 2 semi-skilled labour (i.e. usually men) at F CFA 800 per day get employed. Whereas maintenance is mainly for the engines and minor repairs of the equipment, spare parts include drums, kettles, oil filters, etc. In all three cases, costs for administration and marketing have been accounted as one working day once a fortnight.

The production programme and sales are, in all three standard cases' based on the assumption that the maximum capacity is only used to 10 % in the first year (= 25 working days) and to 20 % from the second year onwards (= 50 working days per year). The reason for this limitation is, in the local context, the availability of rawmaterial. Palm oil and palm kernels are sold at prices fixed by the Government. Palm kernels are not processed further because this involves either a very labour or very capital intensive procedure which is financially not attractive in comparison to the sales price of the kernels.

As working capital, spare parts and a reserve of petrol/diesel fuel is required for the motor-driven alternatives. Finished products are usually taken on stock for a month.

As source of finance, a loan from a local bank is probably required in addition to equity capital to cover the investments.

For the the financial evaluation of the three versions of equipment for oil palm fruit processing, again, the internal rate of return (IRR) has been chosen as the major indicator. The following figures were calculated:

- hand-operated equipment 38.35 % (30.0 % oil recovery),
- TCC process 14.91% (26.5 % oil recovery), - CALTECH expeller 22.00 % (28.4 % oil recovery).

Under the assumptions made for the standard cases in Togo (for oil recovery rates, see figures in brackets above), all three processes would give satisfactory returns on invested capital, the handoperated equipment even producing a very good rate of return. The sensitivity analysis which is illustrated in Figure 24, however, indicates that the financial returns for all three processes react rather sensitively to changing oil recovery rates and wages and very sensitively to changing capacity utilization.

Figure 24: Summary Sheet, Oil Palm Fruit Processing in Togo. Graphs are approximations.

Figure 24, A, indicates that for the handoperated equipment the expected IRR would still be very satisfactory (just below 30 %), if the the oil recovery rate were 2 % lower than the assumed standard case. The same reduction in oil recovery for the TCC equipment would, on the other side, produce an IRR of less than 6 %, which is less than the interest on a bank deposit and therefore not an attractive investment.

Figure 24, B, indicates that capacity utilization (i.e. rawmaterial supply) is a very critical factor for all three processes. For the standard cases, 50 working days per year (or 20 % utilization of maximum capacity) had been assumed. If this is reduced to 25 working days, the TCC process would already run into financial losses. At 75 working days per year, on the other side, the TCC and the CALTECH equipment would show very good results (IRR = 27 % and 35 % respectively.); the hand-operated equipment would even allow financial returns of more than 56%.

Figure 24, C, indicates that the handoperated process is, naturally, most sensitive to changing labour costs, but could financially still tolerate maximum wages (unskilled workers) of about F CFA 1,200 per day. A project using the TCC equipment, on the other side, would go bankrupt, if wages of more than F CFA 800 per day were paid. The CALTECH equipment is less sensitive in this respect.