The fuelwood demand reduction component of the project began with the pilot-scale dissemination of domestic stoves through the training of artisans as independent stove-producers. For reasons outlined above, this approach was abandoned as unpromising in 1986. At much the same time, we recognised the potential of an improved stove dissemination programme targetted at the institutional catering sector.
In the course of the period 1985-86 a range of fuelsaving stoves were developed by Bellerive Consultants Waclaw Micuta and Emil Haas, designed for permanent institutions catering for large numbers of people (in excess of 100 or so). The basic design (illustrated below) was that of a simple channel stove built around a cylindrical, stainless-steel pot
Pilot dissemination was an immediate success, despite being restricted, by the projects limited production capacity, to a local programme largely confined to Central Province. A pre-implementation study was commissioned, with a view to finalising the details of a national dissemination strategy.
An institutional stove
The strategy favoured initially was to develop a version of the design suitable for mass-production, and then to franchise the entire production and marketing operation to a firm in the private sector.
In the course of the pre-implementation study, an extensive market survey was undertaken by a team of Nairobi-based consultants, and the field performance of the institutional stoves was evaluated by an external technical consultant from the Woodenergy Systems Group.
The experience of pilot dissemination, together with these consultants findings, led us to revise the proposed strategy substantially, and with it the designs of the stoves to be disseminated. The dissemination programme will be retained within the Foundation, at least for the next few years. In order that key components such as after-sales-service and the training of operators in fuelwood management may be firmly entrenched before it is transferred to the private sector.
The programme now covers its operational costs, including the cost of this extension service for operators, and has secured a position as the largest single source of this type of catering equipment in Kenya, with systems which save on average more than half the fuelwood consumed by the stoves they are replacing. Given this success, we set out here to consider how the different activities involved in developing the programme were undertaken, and (perhaps more importantly) how they fitted together.
The need for accurate market data is generally accepted. But the obstacles and difficulties inherent in any method of obtaining such data seem to be less widely acknowledged. Consequently, we tend to find, both in project proposals and in the execution of ongoing projects, rather than too little market research being undertaken, too much being expected of the market research exercise.
There are certain things a market survey can tell us. And certain things it emphatically cannot. It is essential to clarify where the line should be drawn to avoid the danger of being lulled into thinking that some aspect of our strategy has been objectively determined by the market data, when in fact it has been at least partially, decided on the basis of preconceived opinion.
Recognising the limits of objectivity is essential for us to maintain an adequately flexible approach to project development.
We consider two types of research: needs-oriented and product-oriented. Needs-oriented research covers everything from the collection and analysis of data for problem-definition to the evaluation of the needs and priorities of a particular consumer group. Product-oriented research is the exercise of evaluating the nature of the market for a particular product, the relative merits of different marketing strategies for that product and so forth.
It would be tidy and convenient if there was a clear dividing line between these two activities. If there were, our course of action would be clear. We carry out needs-oriented research initially to establish what products and systems are required; we then design or identify the relevant products and systems, and subsequently carry out a product-oriented survey to establish the market size and how best to set about penetrating it. Just the sort of linear project structure the donor likes.
In the case of anything relating to fuelwood conservation, it is never this straightforward.
It is completely impossible to design and execute a totally objective needs-oriented market survey: one which is completely neutral with regard to possible interventions. The designers and implementors of the survey invariably have certain ideas in mind, and, rather than merely enumerating the consumers requirements, such surveys tend in reality to be, at least in part, assessing the probable consumer response to a particular product or strategy.
Consider the practical problem of designing a questionnaire. We have the usual conflict between open-ended, qualitative questions which usually do not yield the sort of concrete information which we need to develop a product or system design; and specific yes/no, numerical or multiple-choice questions which do yield concrete information, but only that information which the questionnaire designer thinks is relevant
And what the questionnaire designer thinks is relevant is to a large extent determined by what he or she has in mind for that particular area in the first place.
The problem is that, after the initial survey, the project management is expected to get on and do something. Few donors or government supervisory bodies would have very much patience with a project which began with a survey to establish what the problem was, and then immediately went on to commission a second survey to decide what to do about it. And action based on insufficient or vague information is as dangerous as action based on unfounded presuppositions.
But to achieve an adequate degree of detail to determine completely our subsequent strategy, a survey would have to focus on specific aspects of the situation directly relevant to a particular approach. It is simply not feasible to include, within a single, manageable survey, questions to determine what price of stove the market will bear with questions to determine to what degree fuelwood demand contributes to deforestation in the region.
An unmanageable survey
So what should we do? It seems we cannot act without relying on a number of, possibly ill-founded, presuppositions, while we cannot completely eliminate these presuppositions without becoming bogged down in academic studies. However, in any real situation, the problem is not as intractable as it might seem.
There is nothing inherently wrong with relying on presupposition or, as one might prefer to call it, on the Judgement and experience of project staff. In any case we are always obliged to do so - no objective market survey will completely determine our course of action. The important point is that we acknowledge what has been assumed, rather than pretending that everything was determined objectively.
It was assumed in the original design of the UNEP/Bellerive project, for example, that all types of fuelwood demand contributed equally to deforestation. It seemed so obvious at the time that no one recognised explicitly that this was an assumption. And until the assumption was recognised, the question of which sectors of fuelwood demand caused the most environmental damage per tonne of wood consumed, made no sense.
We soon found, as a number of others have done in recent years, that in many areas the direct impact of rural domestic fuelwood demand on deforestation is remarkably small, since a large proportion of the fuelwood used in this sector is supplied either from gathered dead wood; or from trees which would have been cut down in any case to clear land for agriculture; or from trees which died of natural causes (particularly. In this case, after the 1984-85 drought). This, together with the obvious difficulty of creating a market for improved stoves in a sector where fuelwood is predominantly non-monetised, decided us against pursuing a rural domestic stove programme along the lines originally envisaged.
Spot the cause of deforestation
At the same time, we found that the direct environmental impact of fuelwood consumption by large-scale catering institutions was substantially higher, since such consumers, needing a high-volume and regular supply, depend heavily on logged wood from trees cut down specifically for fuel. Accordingly, we focussed our attention on this sector, with much greater success.
Much may depend on the correct questions being asked, and it is almost always unrealistic to expect them all to be asked at the outset. Key questions may only be suggested by experience in the field.
Thus it is essential that needs-oriented research is not seen as a self-contained exercise to be completed at the commencement of the project, but an ongoing activity integrated into project implementation.
Likewise, project implementors must be flexible enough to accomodate radical changes in what they perceive as the needs of the project target community, since these may only emerge in the course of project implementation. Only through such flexibility can we avoid the conflict between the demand for action (as opposed to yet more research) and the demand for a purely objective basis for our action.
Quite similar conclusions emerge from our experience in product-oriented research. The Market Survey for the Kenya Institutional Stove, carried out in 1987, was essentially a product-oriented exercise. We therefore discuss these conclusions in the direct context of that component of the project.
A major component of the pre-implementation study for the national institutional fuelwood saving programme was a market survey undertaken by Mwaniki Associates Limited, a Nairobi-based team of consultants in Economics, Finance and Management.
The terms of reference for the survey oriented it specifically towards evaluating the characteristics of the market for a particular product: a version of the Bellerive Institutional Stove, designed for mass production with the national programme in mind, constructed from cast iron and (imported) metal sheet.
Even though this was not the version eventually adopted for dissemination (the decision to adopt another version being largely based on the results of the market survey) this specific product-orientation was the key to the success of the whole exercise.
Since they were dealing with a well-defined product, the consultants were in a position to quantify potential demand, make recommendations on promotional and pricing strategies and provide a series of comparative cost-benefit analyses for a range of alternative strategies. This information formed the basis of the design of the national programme.
But it would not be true to say that the market survey determined the final designs and dissemination strategy. We recognise that many of the assumptions on which our strategy is based were already in place when the survey was commissioned and formed the basis of the consultants work.
Thus in the same way that we have to rely on ungrounded assumptions in order to get the needs-oriented research exercise started, initial product-development work cannot be derived from market data. We must to some extent rely on the Judgement and experience of the experts responsible for product design. There is nothing wrong with this. The danger lies in failing to acknowledge that assumptions have been made, with a consequent loss of flexibility.
It was assumed, for example, before the Market Survey was carried out, that the appropriate stove design for a national programme was one which could be mass-produced, and that the most effective strategy would be to franchise the entire production and marketing effort to a firm in the private sector. The survey revealed that these assumptions were incorrect (or, more precisely, that they were not compatible with the projects primary objective of alleviating the problem of deforestation). The main reasons for this were:
a) Market size
The key factor determining the size of the potential market was the fact that only those institutions with readily predictable catering requirements would be interested in investing in the stoves.
The accessible market was thus restricted to health and educational institutions, in both government and private sectors, and large state-sector institutions such as prisons and military barracks.
Such institutions would be prepared to invest in stainless steel pots (the principal cost element in the institutional stoves disseminated by the project) on the basis of the fact that these would last for more than 10 years, while aluminium pots (the only available alternative, at one-third of the cost) would last one year at the most.
The other important group of institutional fuelwood consumers is made up of private hotels and restaurants, but the survey found that these institutions tend to require much greater flexibility in their catering equipment, to accomodate fluctuating demand, than the projects large-volume cooking systems could provide. Also, those hotels and restaurants which make extensive use of fuelwood (primarily those at the lower end of the market) tend to operate on a very short time-frame, owing to the volatility of their business.
Thus defined, the potential market for the UNEP/Bellerive projects institutional stoves was assessed at 400 to 500 units per annum. The consultants concluded that a programme aimed at supplying a substantial proportion of these institutions with improved catering equipment would not be financially attractive for a purely profit-motivated concern.
The cost of an adequate marketing and distribution infrastructure to reach the majority of potential programme beneficiaries (being principally located in rural areas) would be prohibitive (witness the fact that major oil companies are pulling out of the rural liquid petroleum gas market).
b) After sales service
The consultants found that after-sales-service would be a key factor determining the actual fuel savings achieved by the improved stoves.
Any stove needs to be adequately maintained for it to function properly and save fuel, and the majority of potential programme beneficiaries would need this maintenance to be provided by specialists (as opposed, for example, to supplying operators with a maintenance manual).
Such after-sales-service would be neglected by a commercial operation for reasons of cost, since it would have to be based on personal visits to institutions by service staff.
After sales service
c) Costs of rural-sector marketing
Since most institutions would need specialist advice on their probable catering equipment requirements prior to placing orders, it would not be feasible to sell the stoves off the shelf through established retail outlets. For a promotional campaign to be effective, it would have to rely heavily on personal contact between sales personnel and the key decision-makers in institutions.
In view of the geographical dispersal of the potential market, such a campaign would be a relatively heavy burden on the programmes cashflow: the reason why none of the established commercial producers of institutional stoves make any serious effort to market and/or distribute their products.
d) Need for concessionary credit
The number of institutions able to benefit from the dissemination programme, and thus the pace at which fuel savings could be achieved, would both increase substantially if concessionary credit facilities were available to assist programme beneficiaries to cover the initial capital outlay. In order for the majority of institutions to be able to afford it, given the high market interest rates currently prevailing in Kenya, the capital required to establish such a credit scheme would have to be provided by a donor.
If the programme were being run as a profit-making concern, then shareholders would benefit indirectly from the donor-funded credit scheme through increased market penetration. Most donors would be dissatisfied with supporting private shareholders in this way.
These conclusions, together with the findings of the external technical consultant, our own experience of pilot dissemination and input from Government of Kenya officials, served to determine the final strategy for the national programme.
The important point is that the technical design of the stoves to be disseminated was modified on the basis of these findings. The sheet-metal/cast-iron construction was abandoned in favour of an earlier construction method based on fired-clay bricks. The brick construction was rejected in 1986, since it would not have been attractive to a private company. But since the survey demonstrated that the programme was going to have to continue to be run on a non-profit basis in the short-term, a partially decentralised dissemination strategy, involving the brick construction, proved to be the least-cost approach overall (see below).
For product-oriented market research to be genuinely effective, it must be based on a precise description of the characteristics of the product to be disseminated. On the other hand, effective market research will and should suggest modifications in technical designs, dissemination strategies, and so on.
The market research exercise must be seen as an integral component of the process of product development, not as a self-contained exercise to be completed either before or after the technical design phase.
We remarked, at the beginning of this subsection, that the principal danger relating to market research is that too much is expected of it. We may, by including an extensive user-needs survey in the opening phase of the project, convince ourselves that the approach we adopt has somehow been arrived at objectively. Likewise, by commissioning a comprehensive product-oriented survey prior to implementation, we may hope to prove, equally objectively, that our approach is the correct one.
Such a view conceals the extent to which such market research exercises are themselves based on assumptions. They can and should guide the thinking of project implementors. But they cannot provide a substitute for it No survey returns can replace the experience and ideas of project staff working in the field.
And we must therefore design a fuelwood conservation project to be flexible enough for those in the field to develop their ideas, and not stifle everything in the pursuit of objectivity.
It has become commonplace to stress the importance of a realistic dissemination strategy in any project involving the introduction of improved technology. Rather than reiterating this point, we hope, in this section, to show how we may set about developing such a strategy.
The key to getting these things right is to do them in the right order.
The very importance attached to dissemination strategy often leads to it being determined at the wrong point in the project cycle, viz: at the initial project design stage.
The point is best illustrated by taking a specific example from the original (1983) UNEP/Bellerive project document. We find, mapped out in the proposal, the intended dissemination strategy for the improved domestic stoves: experts were to teach a team of instructors, who in turn would teach selected inhabitants of the project area, how to build fuelsaving stoves. The inhabitants so trained were then intended to begin building stoves either for their own families or, as an income-generating activity, for their neighbours.
Now there was nothing inherently wrong with this strategy - it had been used successfully by a number of organisations disseminating improved stoves in India and, in particular, in the Foundations own successful programme in Northern Pakistan. In any case we will not discuss here the strategy itself. The point is that it was there, specified in the original project document, before any work had been done on the ground in the project area. And we now realise that it should not have been.
With an effective promotional campaign, it is often possible to introduce a product into a community which is not necessarily tailored to the original wants of that community. Wants can change, or be changed. But we have found that to persuade the community to adopt unfamiliar methods of production and marketing is a far more difficult task. And it is the community, not the foreign expert, which is in the best position to Judge how such things should be done.
Thus it makes no sense to specify the dissemination strategy for an improved technology before an adequate, ground-level pre-implementation study has been carried out to determine what the real priorities of the target community are.
Yet even today we find donor agencies, in the name of action-oriented projects, demanding that a dissemination component should be written in to the first document they see. And the document writers feel obliged to specify the strategy to be adopted for fear of it being thought that they have not given enough attention to dissemination.
The usual result seems to be that the dissemination strategy adopted is simply that which is in vogue at the time.
Ten years ago, on the basis of experience in India, it was user-built stoves. Now, after the success of the Kenya Ceramic Jiko project, everyone is talking about the informal sector. Next...?
It is almost always unrealistic to specify the dissemination strategy for an improved technology in the initial project proposal, but this does not mean we leave out mentioning it altogether. What must be specified is how the strategy is to be developed - in particular, how the target community is to be involved in developing it.
The process of designing or identifying the optimum strategy must involve surveys of established practices (not cooking practices in this case, but production and marketing practices), comparative tests, field trials and pilot-scale implementation: all activities which are normally confined to technical product development
In any project based around the introduction of improved technology, as much attention must be paid to the design and testing of locally appropriate production and marketing systems, as is devoted to the development of the products themselves.
Where do we begin? A frequent starting point for thinking on dissemination strategy is the stove design: weve got this stove, it works, now how do we get as many as possible out into the field? This is precisely the wrong place to start. If the stove design is considered finalised before thinking on dissemination strategy has even begun it is a sure sign that we have let technical considerations gain an excessive influence over the development of the project.
An excessive influence
The problem with such a design-led approach is that the dissemination programme then tends to have the sole objective of maximising the number of stoves disseminated (or, on a slightly more sophisticated level, maximising the number of stoves in use). If project implementors are subject to evaluation-by-numbers in this way, they will understandably tend to neglect such considerations as how the stoves disseminated are operated, which may have as great an impact of fuelwood savings as the raw number in use.
The starting point for the development of a dissemination strategy should be the problem-definition exercise undertaken at the very beginning of the whole project.
It may seem paradoxical to begin considering dissemination before we have developed or identified a product to disseminate, but that is precisely what we are advocating.
It seems less paradoxical when we recall that designing improved stoves never seems to have presented very many problems, while designing effective dissemination mechanisms certainly has done.
It may, therefore, make more sense to design a dissemination mechanism and then tailor the product to fit it, rather than the other way around.
The problem-definition exercise will yield a preliminary set of specific project objectives, which we refer to here as impact priorities, to distinguish them from the more general alms and objectives of the organisation or project as a whole. These will form the basis for initial thinking on dissemination strategy. They may change, or at least become clearer as the project develops, but it is essential to begin thinking about dissemination at this stage in order to guide the process of technical development
The types of impact priorities which may lead to the development of projects involving the dissemination of improved stoves include far more than the usually-quoted combatting deforestation. A brief list of examples might include:
· to contribute to a reduction in the rate of deforestation through the promotion of more efficient fuelwood utilisation systems and practices (the usual impact priority)
· to improve the national balance-of-payments position through the substitution of fuelwood for imported fossil fuels
· to improve the status and self image of rural women
· to reduce the risk of respiratory diseases resulting from the use of biomass fuels
· to increase the effective size of the working population available for growth-generating activities through the reduction of the burden of fuelwood collection
An improved-stove programme might have any one or a combination of these priorities (perhaps we can note in passing that a programme which tried to address them all at once would be in serious danger of losing focus). For example, the domestic stove dissemination programme currently being implemented in Kenya by Maendeleo ya Wanawake Organisation (MYWO), with technical assistance and support from the GTZ Special Energy Programme of West Germany, has as its key impact priorities the first and third examples quoted above.
In view of the importance of the status of rural women to that programme, a dissemination strategy has been developed which centres on womens groups as the primary vehicle through which the stoves are installed in rural homes. The programme organisors recognise that, from the perspective of getting the stoves out as fast as possible, other channels, such as profit-motivated artisans, might be equally effective. But that is not the point of the project
More than just a fuel-saving device
It is these impact priorities, identified through experience gained on the ground in the course of the project pre-implementation phase, which should guide the development of dissemination strategy, and not the secondary consideration of how many stoves we can sell.
Now that the institutional stove dissemination programme is fully established, we are turning our attention again to the problem of domestic fuelwood consumption. While the overall objective of these activities remains reducing the rate of deforestation, we now realise that the initial (unstated) impact priority, that of reducing domestic fuelwood demand by improving the efficiency of domestic cooking equipment, was based on two misconceptions.
First: the assumption that the technical characteristics of the equipment used is the determining factor in the overall system efficiency of a domestic kitchen. In many areas, fuelwood management may play a larger role.
Second: the assumption that domestic fuelwood consumption contributes directly to deforestation. This is probably only true for the upper-income sector, whose members rely on purchased fuelwood. The environmental impact of the collection of fuelwood by lower-income consumers is difficult to quantify, and may have been overestimated.
Two different approaches are now being explored, targeted at two different sectors of the domestic fuelwood market. For the low-income sector we have established a project, in conjunction with the Wordwide Fund for Nature (WWF) to produce training materials on energy management in the kitchen, for use in womens groups and primary schools.
It has been found that widely varying overall system efficiencies are achieved by different women using essentially the same equipment. Thus the training package alms to transfer economical fuelwood-use practices from areas which have long suffered chronic fuelwood shortages to those in which fuelwood shortages have only recently begun to be felt.
The initial series of lessons for primary schools are being piloted in Taita-Taveta and Nakuru districts, with the assistance of the District Education authorities and the Kenya Institute of Education. It is hoped that they will prove suitable for incorporation into the national Home Science curriculum. The teaching materials include an illustrated students pamphlet for each lesson, accompanied by a detailed lesson plan for the teacher.
Cooking to Conserve
It may prove appropriate to incorporate the introduction of a low-cost stove into this lesson series, but the introduction of the stove is to be seen as an extension of the training material, not as an end in itself.
The key reason we believe this extension-oriented approach will prove more successful for the low-income community than the introduction of improved technology is that the fuelwood economy among low-income consumers is predominantly non-monetised (and therefore an improved stove cannot be presented as a viable economic investment).
In the case of middle and upper income fuelwood consumers, the situation is different. Such consumers tend to purchase fuelwood for convenience, being denied alternative fuels by the limited rural distribution infrastructure for gas and kerosene. As in the case of institutions, such purchased (as opposed to gathered) firewood tends to be produced through the harvesting of whole trees: with corresponding environmental damage.
As well as having a greater negative environmental impact per tonne of wood consumed, the upper-income sector of the domestic fuelwood market is also more accessible to the introduction of improved stoves sold simply on the basis that they save fuel and improve the quality of life. Accordingly, we are introducing a range of relatively high-cost but high-performance stoves targeted at this sector.
Two production methods are being tried, tailored (with the benefit of hindsight) to established practices in Kenya.
The first method: the stoves will be mass-produced from sheet metal and sold ready-made from established retail outlets. While probably the fastest strategy, this has the disadvantage that the cost of such production contains a substantial convertible-currency component.
The second method: the design will be adapted, so that the stove can be built on site, primarily from locally-produced materials, but requiring only standard masonry skills. This bypasses the need for specifically-trained professional stove-producers which, as we noted above, was a key problem in the earlier built-on-site designs. The masonry stove, which has already aroused considerable interest in the private building sector, from small-scale building contractors to those responsible for major housing estates, is a combination of the Bellerive Pogbi design with the Nouna C stove developed for West Africa by W. F. Sulilatu of TNO Apeldoorn, the Netherlands.
The Bellerive Nouna Stove
The starting point for the development of the dissemination strategy for the institutional stove programme was the finding that the fuel consumption of any woodburning system could more than double if it is operated badly.
The difference made by operator practices is usually more marked in the case of traditional cooking systems: a badly managed Bellerive institutional stove might consume twice as much fuel as it should do, while a badly managed open fire could out-consume its well-managed counterpart by up to ten times. Even so given these findings, it seems inappropriate to speak of a fuelsaving stove, as if the stove, of itself, saves fuel A radical change in the thinking behind what we were doing was called for.
Rather than being seen as an end in itself, the introduction of improved woodstove technology must be understood to be one component only of a comprehensive training package designed to reform the overall management of firewood. The stove is a tool which cooks may use to conserve fuel.
There are many other components to the complete system, which are normally neglected by fuelwood demand reduction programmes, as they were by the UNEP/Bellerive programme until the 1986/1987 pre-implementation study drew our attention to their importance.
Accordingly, a fuelwood conservation training package was designed for application in institutional catering establishments. This package, which incorporates the introduction of the improved stoves developed by the UNEP/Bellerive project, is detailed in the Bellerive Foundation Dissemination Manual for Institutional Cooking Systems. As the title of the Manual indicates, we now believe it is no longer appropriate to speak of stove dissemination. The current programme is better described as engaged in the dissemination of improved cooking systems.
The system includes a substantial training
· firewood drying, cutting and splitting
· how to feed the fire (how much wood should be fed for a particular task)
· how to regulate the air supply (leaving the door open may make the fire blaze nicely, but actually cools the contents of the pot)
· how to cook economically (simple points such as keeping pots covered and the fact that boiling food fiercely does not cook it any faster, but simply wastes fuel)
· even how to wash up (if the soot is left to build up on the outside of the pot, as is the traditional practice in Kenya, it forms an insulating layer, which dramatically reduces overall system efficiency).
Getting such things right has as much an impact on fuel consumption as using the right equipment
The problem with introducing improved firewood-use practices into institutions (and the principal reason systems in use in institutions tend to be less efficient than domestic systems) is that the operators of institutional cooking systems have little direct personal or financial incentive to save fuel
When a domestic housewife knows that she will have to walk several kilometres to replace every twig she uses in cooking, remarkably efficient cooking practices rapidly evolve. But in an institution, the firewood bill is the bursars problem, and does not concern the catering staff. Thus the training package has to provide sufficient structure and detail to compensate for the operators lack of incentive to work out the details themselves.
A detailed operators manual was rejected as a vehicle for the training material, since a large proportion of the catering staff in Kenyan institutions are women of the often illiterate, older generations brought up before independence. Training by direct demonstration was clearly needed.
Since only low-level communication skills would be required of the trainers, the expense of a network of specialist extension staff seemed unjustified. The clear solution was that the teams of craftsmen responsible for installing and servicing the stoves would also undertake the training of operators, with the assistance of a detailed series of lesson plans, which do not require the trainers to have any previous extension experience.
The advantage of this integrated approach, incorporating the introduction of improved technologies and improved practices within a single dissemination programme, is that through the link to technology the training component may be made financially self-sustaining. It is relatively straightforward to sell a tangible product such as a stove. It is far more difficult to persuade the consumer to cover the cost of intangibles such as training of operators. By marketing stove, auxiliary equipment and training as a single package, we can recover all the operational costs of the programme, including the cost of training cooks, from the programmes beneficiaries.
This need for extensive on-site training was one of the key factors determining our decision to adopt a decentralised approach to dissemination. The strategy favoured in 1986, centralised production in Nairobi, would have maximised the rate at which improved stoves could be produced and introduced into the field. But we found that the introduction of improved stoves alone was not adequate to achieve the key impact priority of reducing fuelwood consumption.
The introduction of the other components of the cooking system, the software, required a decentralised approach. The overall cost of the programme would be minimised by integrating the hardware and software dissemination components. Thus a partially decentralised production system for the stoves themselves became perfectly acceptable, since the limiting factor determining fuel savings was not the rate at which stoves could be produced at a factory, but the rate at which they could be installed with operators fully trained in their correct use and auxiliary skills.
Consideration of two other impact priorities finally determined the dissemination strategy adopted.
These two further impact priorities emerged from discussions on the development of the programme with Government of Kenya officials in the Ministries of Energy and Planning and National Development. In the current five-year development plan, the following are explicitly stated as development objectives:
· generation of employment in rural areas, away from Nairobi
· Improvement of the national balance-of-payments position through import substitution.
Both of these priorities indicated that centralised production of complete stoves in Nairobi was inappropriate. The first, concerning rural employment, obviously implies that as large a proportion as possible of the labour involved in producing and installing the stoves should be undertaken in rural areas. This favours the production of stoves in kit form, to be assembled on site by specialised teams of craftsmen based in rural areas.
Although this is a more labour intensive production method, and the cost of maintaining teams in the field renders it a slightly higher-cost approach if the sole objective is producing stoves, the need for operator-training and on-site service and maintenance by specialists means that we have to maintain teams in the field anyway, rendering decentralised production the least-cost strategy overall.
The link between the balance-of-payments related priority and the decision to adopt decentralised production relates to raw materials. The only appropriate structural material for stoves of this size, which are to be produced in finished form in a central factory, is mild-steel sheet, which is not produced in Kenya.
On the other hand, if stoves are constructed on site, and therefore do not have to be transported in finished form it becomes feasible to use locally-produced materials such as fired-clay bricks and clay mortar to form the structure of the stove, with the minimum of metal components incorporated to give dimensional uniformity.
Through the adoption of this construction system, the convertible currency component of the cost of the units now under dissemination has been reduced to less than ten per cent.
The details of the dissemination strategy of the ongoing programme are given below in sub-section 4.5. Our point here is that impact priorities, rather than considerations of technical design, determined the final strategy adopted.
Not, ideally, the determining factor
This report is not meant to be a reference work for the woodstove designer: there are a large number of such texts available. This subsection is written for the non-technician who finds him/herself in the unenviable position of having to direct and set priorities for his/her technical staff, with neither the background nor the inclination to ascertain what, if anything, the technicians are talking about when they get onto the subject of engineering. We hope that the following will provide some ammunition.
the problem is designing one which saves fuel.
In the majority of woodfuel demand reduction projects, technical development begins from the wrong starting point, viz: the three stone fire, as opposed to the complete cooking system traditional to the project target area.
If the traditional system happens to incorporate a three stone fire, as it does in most parts of Africa, is this anything other than a verbal distinction? It is, and moreover, the assumptions implicit in taking the three stone fire alone as the starting point often lead to distortions throughout the technical development phase and subsequent components of the programme.
It has been argued, recently and forcefully, that the assumptions on which woodstove programmes are based are actually incorrect, and therefore that improved stoves are a complete waste of time. We shall argue that this conclusion is too strong:
Improved stoves do have a concrete role to play in fuelwood conservation, but this role is less universal than previously thought and in some areas, other measures may contribute as much to reducing fuelwood consumption as the introduction of new technology.
A highly simplified version of the traditional argument for the introduction of improved stoves might go as follows:
In the traditional three-stone fire, with a clay pot, operated correctly by an African woman, between 10% and 15% of the heat generated by the burning wood is transferred to the contents of the pot. In this improved stove, the rate of heat utilisation is more than 30%, with the same operator and test procedure.
Thus the open fire consumes two to three times as much fuel as the improved stove in executing the same task. Therefore, if all the households in this district were to use the improved stove, we may expect a reduction in fuelwood consumption by a factor of 2 to 3.
There are, of course, all sorts of qualifications to be made to this argument. We do not intend to attribute it to anyone in particular, but to use the general line of thought (which must be familiar) as a basis for discussion.
In recent literature, the usual response to the above naive argument is along these lines:
Woodstove designers habitually underestimate the efficiencies achieved by African women using traditional equipment. Heat utilisation rates of up to 25% have been obtained from simple three-stone fires. Since most woodstoves which are affordable for the rural poor do not claim efficiencies of more than 30%, it follows that fuel savings resulting from their use will be minimal.
One of two moves is then made:
a) It follows that more efficient stoves are required (with efficiencies over 40%) for us to make any real impact - therefore more resources should be devoted to stove design.
b) It follows woodstove programmes are a waste of time.
But neither of these moves makes sense of the fact that in many instances the introduction of improved stoves, even not particularly efficient ones, does seem to result in a reduction of total wood consumption. For example, the clay two-pot domestic stove disseminated in the initial stages of the UNEP/Bellerive project had an overall efficiency of less than 30% and yet the results of controlled cooking tests and monitoring of the frequency of fuelwood collection trips in the field indicated that the introduction of this stove reduced overall consumption by about a factor of three.
Moreover, both conclusions - that designers should try harder or that designers should give up - are still based on an acceptance of the essentials of the argument given above. Those who question the conclusion of the simple argument (that a 30% efficient stove can save fuel) do so on the basis that one of the stated premises (that the open fire is only 15% efficient) is invalid.
What is seldom questioned is the basic reasoning behind both positions, in particular the unstated assumption that stoves save fuel by increasing the percentage of the energy released by the fire which is recovered by the contents of the pot.
This assumption may seem so obviously true that it is not worth thinking about. But this is precisely the sort of assumption we should question.
Consider the situation in a little more detail. Diagram 1 below might be a rough heat balance (energy audit) for a traditional three-stone fire. None of these figures are the result of any research on our part: they are drawn from the literature simply to illustrate the point we are trying to make. Hence from the diagram, the three stone fire is said to be 15% efficient.
But where do these figures come from? They seem quite straightforward, particularly to the non-specialist who is not in the business of designing stoves but in developing implementing and/or evaluating fuelvood conservation projects. We have energy in, energy wasted and energy out. Reduce the proportion of energy wasted and we will get the same energy out for less energy in. What could be simpler?
If, however, we look carefully at some of the terms in the energy equation in diagram 1 the position seems a little less clear. Take, for example, the two most important terms - the energy released by the fire and the energy recovered by the pot contents (energy utilised in the literature).
a) Energy released by the fire
Unless it has been dried in an oven for around twenty-four hours, wood contains moisture. When it is burnt, this moisture must be heated to boiling point and evaporated away. A sample of freshly-cut green wood may be over 40% water (by mass). When such wood is burnt, over 45% of the chemical energy released through the combustion of the inflammable components of the wood is used simply to drive off this moisture.
When property air-dried wood with a moisture content (wet basis) of about 10%, is burnt, the proportion of energy needed to drive off the moisture is only about 11%.
Thus by drying out the wood sample from 45% to 10% moisture content before burning it, we increase the amount of energy effectively released when the wood is eventually burned -energy available to heat the pot, stove, chimney etc. - by up to 63%
The energy released by the fire term in the heat balance given in diagram 1, as in all calculations of the efficiencies of open fires, woodstoves etc. is calculated on the basis of effective calorific value (ECV). The ECV is a measure of the energy content of the wood after compensating for moisture content. The fact that some of the chemical energy in the wood is used to drive off the moisture is compensated for by calculating heat utilisation efficiencies as if that energy was not there in the first place. If we are using wood of 40% moisture content (w.b.) then we say the energy released in burning a sample weighing one kilogramme is 10.2 MegaJoules. Never mind that an extra 6.4 Mega-Joules could have been released from the same sample if we had air-dried it for two months before-hand: that is of no relevance to the stove designer.
The point of using ECV as the measure of energy released is to eliminate the variable of wood moisture content in order that efficiencies calculated from tests using woods of differing moisture contents may be directly compared. Any stove technician knows this. The technicians are deliberately excluding part of the picture because it is of no relevance to what they are trying to do. Yet for the programme director, whos Job it is to place the activities of the stove technicians into the wider context of fuelwood conservation, what is peripheral for them is precisely what connects their work up to rest of the project.
b) Energy recovered by the pot contents.
What do we mean by energy utilised? The VITA standard definition is straightforward: in a water boiling test, energy utilised is the energy received by the water in the pot - the sum of the energy used to heat the water to boiling point and the energy used to evaporate whatever water is boiled away. Again, from the point of view of the stove designer, this is the most appropriate quantity to measure, since what he or she is trying to do, in designing a stove, is to minimise the losses identified in the heat balance of the open fire given above. The ratio of energy utilised, so defined, to the energy released by the fire, as defined above, is the clearest and most reproducible measure of his or her success in doing this (known as percentage heat utilisation rate, or PHU). Once the energy is inside the pot, as it were, the stove designers Job is over.
But the Job of reducing the amount of fuelwood used to cook the food of the developing world is anything but over. Consider what happens to the energy after it reaches the contents of the pot during a typical cooking task: stewing beans, for example. Only a very small proportion of the energy entering the pot (less than 2%) is actually consumed in the chemical reactions which convert the beans from their uncooked to cooked state. Most of the energy is used to heat water and beans to boiling point and to evaporate away whatever water is lost as steam. If we look at the overall cooking process, the energy used to evaporate water away and even, if the hot water is thrown away afterwards, the energy used to heat the water, should be seen as system losses.
Yes the stove designer will respond. But the stove cant do anything about those losses. This is perfectly true (almost - see below). But, as in the case of losses due to high wood moisture content, it does not mean that nothing can be done: see 4.3.2 below.
Thus a slightly more comprehensive picture of the energy flow during the entire cooking process, using the open fire considered above, is shown in diagram 2.
We notice immediately that the losses which were not included in the first (simple) heat balance (diagram 1 above) and which are eliminated from the calculations of stove technicians, are precisely those which improved stoves can do nothing about. Thus if we are presented with diagram 1, and are under the impression it tells the whole story, then it seems stoves are the whole answer: every one of those losses can be substantially reduced by a properly designed stove.
Of course, the simple heat balance does not tell the whole story. Nor does diagram 2 - but a little more of it than diagram 1 does. Is it therefore coincidence that the simple heat balance picks out Just those features of the story which are relevant to stoves? Or a dark plot on the part of the woodstove designers to mislead us?
It is neither. The simple heat balance, along with the concepts of percentage heat utilisation, specific fuel consumption, and all the various test procedures, have been designed by technicians, for technicians. These concepts and procedures are specifically tailored to eliminate everything which is not relevant to stove design from the picture - which is perfectly standard scientific practice. This is the language which stove designers need to communicate with each other and pursue their work most effectively.
The danger lies in attempting to discuss the whole problem of fuelwood conservation in the same terms: when for example, the simple heat balance, which is in reality just a tool of the stove designers trade, is offered as a proof of the stove designers importance.
So the lesson for programme directors is clear when a clear-cut and obvious solution is presented, that is precisely the time to look for the preconceptions on which it is based. We have noted already that we are obliged to work on the basis of preconception. But preconceptions are only realty dangerous when they are not recognised as such.
The assumption that improved technology is the answer to woodfuel demand reduction is one of the most prevalent and insidious assumptions of all, because it is presented in the context of a complete, coherent and convincing argument, often in terms which those responsible for prioritising different programme components, as non-technicians, do not fully understand.
The stove designer provides us with a concrete solution to a genuine problem: how to direct more of the heat released by the fire into the contents of the pot. He or she does so by an impressive process of scientific reasoning, such that it is difficult for us to recall that this may not be the key problem in fuelwood conservation.
In accepting the technicians formulation we are, quite subtly, designing the problem so that our solution applies to it. In principle, at least, it might seem preferable to work in the opposite direction.
The right direction
Stoves can help to save fuel. We have, so far, played down the importance of improved technology, not because we believe it is totally unimportant, but because its importance has been generally overestimated. We are trying to redress the balance.
The only way to maintain a sense of proportion in setting priorities for a fuelwood conservation programme is to keep a clear idea of what it is we are actually trying to do; as opposed, in this instance, to what the technicians think we are trying to do. Let us suppose that the overriding priority is to combat deforestation, and that the level of fuelwood demand in the project area is actually an important factor determining the rate at which trees are cut down (this is not always the case).
The first question to ask is: is it true that fuelwood is being used inefficiently at present - and we mean here the broad sense of efficiency: could these same tasks be carried out using a substantially smaller amount of fuelwood? As opposed to the narrow sense: could the percentage heat utilisation rate of this open fire be improved?
In evaluating the overall efficiency of fuelwood use it is essential to look at all stages of the fuelwood-consumption process, from trees being cut down to the food finally being cooked (if we are focussing on the use of fuelwood in catering), and not just the specific stage in the process - the transfer of heat from fire to pot -which is normally considered when we talk of efficiency.
Interesting trends may emerge. For example, the efficiency of the equipment used in traditional institutional kitchens, narrowly defined as PHU, is approximately the same as the efficiency of traditional domestic cooking equipment. Yet we found that, in Kenya at least, the overall system efficiency of an institutional kitchen was far poorer than a domestic kitchen: much more fuel was consumed in the institution than in a typical domestic situation to carry out the same task. It was obvious that losses occurring at points in the cooking process other than those identified in the standard PHU test were much greater in the average institution than the average household.
What these losses are and which ones are the most important will depend on where we are working and how people use fuelwood in that area. In the case of Kenyan institutional catering, the two key causes of losses outside the stove (i.e. outside that stage in the cooking process to which the stove is relevant) were failure to dry wood properly before using it, and failure to regulate the power supplied to the pot.
Losses may occurr at other points in the fuelwood utilisation process
a) Wood preparation
We have discussed the implications of using wet wood in some detail above. It is easy to exhort institutions to dry firewood before using it but this is not enough. In developing the institutional fuelwood saving programme, we devised a detailed firewood preparation system with the following key elements:
· wood is cut and split before it is stored. This is important because wood dries primarily through the cut ends of the grain, not across the grain through the sides of the log. In the training package we supply the institution with a specially designed wood-cutting stool to ensure logs are cut short enough not only to dry properly, but to fit into the stove firechamber.
· wood is stacked in a triple-bay woodstore (design supplied to the institution by the programme) with open, or perforated, walls. The open sides allow air circulation to accelerate drying and prevent rot. The three-bay system is designed so that each bay accommodates one months supply. When it is exhausted, the institution begins to use the wood in the next bay and refills the first. In this way we ensure that all wood has been stored a minimum of two months before use.
· The base of the storage site is lined with ash to prevent termite infestation - an essential precaution in most tropical countries.
These are very simple points - it hardly seems appropriate to call them innovations. But if we recall that institutional catering staff have no direct incentive to save fuel, and therefore that there is no pressure for such economical practices to evolve naturally, the importance of setting these activities into a structured and detailed training package becomes clear.
b) Power regulation
It is likewise with the regulation of power supplied to the pot. Most cooking tasks in Kenya involve boiling or stewing. Many of these require that food is maintained, in water, at boiling temperature, for a certain length of time. So long as the temperature is maintained it is irrelevant how hard the water bolls. Thus once the water has been brought to the boll, the power supplied to the pot must be reduced to the level at which it only compensates for losses, no more. Energy spent in driving off steam is energy wasted.
Any woman in an area which has experienced chronic fuelwood shortages over a long period knows this. But inexperienced cooks and, in particular institutional catering staff who may not be all that interested in saving fuel, still tend to boll food as hard as possible, perhaps under the mistaken impression that this will make it cook faster. Thus the training package provided by the institutional fuelwood saving programme incorporates a detailed demonstration component in which the cooks are shown how to regulate the fire in cooking each of the main dishes used in Kenyan institutions.
As in the case of drying firewood, to exhort cooks to boll food slowly is not enough. We must provide them with a specific way of carrying out the tasks they have to undertake.
It is worth noting that using almost any improved stove does make it substantially easier for the cook to control the power supplied to the pot. This point is probably the key to the (ostensibly puzzling) fact that the use of a not-particularly-efficient stove (in terms of PHU score) can result in a reduction in fuel consumption.
For example, in the course of a comparative evaluation of the performance of various domestic stove models, carried out on behalf of UNEP by the Kenyatta University Appropriate Technology Centre, a ceramic stove from Somalia was found to be one of the least efficient of the stoves tested, in terms of PHU. In the course of controlled cooking tests, on the other hand, this same stove gave one of the lowest rates of fuel consumption - particulary for slow cooking tasks.
The reason for this was that the design of the stove was such that it had a very low maximum power output, and for slow cooking it is the power output, not the heat transfer efficiency, which determines fuel consumption. All the other stoves were transfering heat to their respective pots more efficiently, but since too much heat was being supplied to the pots anyway, this simply resulted in more water being boiled away, more cold water needing to be added to prevent food from burning and a lower overall system efficiency.
One response to this might be to conclude that the VITA standard test, and PHU as a measure of stove performance, are both useless, and a better test procedure and indicator should be developed. This is a completely misguided reaction.
The VITA standard water-boiling test is a specialised procedure with a very specific purpose. For that purpose, it is entirely suitable. Other tests, such as the comparative controlled cooking test, have other purposes, for which they are equally suitable. It is misleading, for example, to say that the specific fuel consumption (SFC) - the indicator which emerges from a controlled cooking test - is a better indicator of stove performance than PHU. They are measures of different aspects of the stoves performance - neither tells the whole story.
Rather than trying to devise a test which does tell us the whole story, which is probably a fruitless undertaking, we should recognise the limitations of the various tests and treat data generated by them accordingly.
Any test is an artificial situation which may or may not have any relevance to the actual performance of these stoves in carrying out the tasks they were designed for. What matters in the end is not how a stove performs in a VITA standard water-boiling test, but how using that stove affects the amount of fuelwood which the women of the project target area require to cook the food their families need.
This seems to suggest that we have to begin technical development work from scratch every time we move to a new area. There is a sense in which this is so, in that we must be prepared to find that the designs we developed for a neighbouring district will turn out, for some reason or another, to be inappropriate to that area.
But it does not mean that all work done to date is irrelevant: an engineer with experience in designing tractors who is put onto designing combine harvesters will be far better placed than one with no experience at all. Certain features, incorporated into the design of a tractor, will reduce its fuel consumption. The same features may well also reduce the consumption of the combine, even though the two are being used for different tasks and their fuel consumptions are not directly comparable.
Thus the experienced stove technician, travelling to a new region, knows that he does not carry with him the answer to reducing fuelwood demand in that region; and that he will not have the answer until he has been working on the ground, with the people of the project area, for long enough for them to make their input into whatever design is eventually developed: long enough to establish, not which stove performs best in a VITA standard test, nor even in a controlled cooking test, but which design enables the cooks of that area to use the least fuel over the year.
Rethinking the role of the technical expert in this way has a direct impact on project design. The traditional approach is for technical development to be carried out by expatriate experts in a relatively short time-frame. We have found that rather than being a short, self-contained exercise, technical development should be integrated into the overall development of the project, with the design of the technology not being considered finalised until the end of a substantial pilot dissemination phase. The role of the expert is to provide suggestions for system modifications, not to design a gadget.
Not the way to handle the technical expert
For such an approach to be adopted, a minimum level of technical expertise must be available on a continuous basis in the project area over a long period - one to two years at least.
The sort of expertise required is the ability to monitor the performance of systems in use, evaluate quickly the probable impact on performance of any design modifications which stove producers or users may come up with, adapt production systems to accommodate changes in dissemination strategy and so forth. Unless such expertise is available, the project will be tied inflexibly to the original design produced by the short-term technical consultant.
Most impact assessments, or cost-benefit analyses, of fuelwood conservation projects depend to some degree on the notion of national fuelwood demand, treated in an analogous fashion to national petroleum demand, and such like.
It is assumed implicity that the level of fuelwood demand is a cause for environmental concern in the same way that the level of petroleum demand, in a non-oil-producing country, is a cause for economic concern. It is not.
Petroleum demand within a country is linked, in a completely straightforward way, to oil imports. There is no such simple connection between fuelwood demand and the rate of deforestation. Different sectors of fuelwood demand, (such as institutional vs. domestic catering) may have a completely different environmental impact, since not all the wood that is burnt has been obtained by environmentally-damaging means. To burn trees which have already been cut down to clear land for agriculture, or which have died of natural causes, has virtually no direct impact on the rate of deforestation.
What matters is not total national woodfuel demand, but the level of demand which must be met by the harvesting of trees specifically for fuel. And a realistic figure for that is exceedingly difficult to obtain.
Several factors complicate the picture. We mentioned in sub-section 2.4.1 the problem of the proportion of fuelwood which has to be paid for in cash being naturally exaggerated by survey respondents. Thus demand-side data is inevitably suspect Moreover, in a country such as Kenya where the harvesting of trees for fuel is virtually illegal in most areas, accurate data from the supply side is equally scarce.
Even if accurate data were available, in many cases there may be no clear-cut distinction to be drawn between trees cut down for fuelwood and those cut down primarily for other reasons, but which end up being used for fuelwood.
On the one hand we have trees being cut (usually illegally) on undeveloped land or in national parks and converted into charcoal. This is a clear case of woodfuel demand doing direct environmental damage: if the demand for charcoal were not there, then these trees would have been left alone.
At the other extreme we have a housing estate developer clearing land for building, and selling the trees cut down for fuelwood. In this case it is extremely unlikely that the value of the fuelwood so obtained played any part in the developers decision to clear the plot These trees would have been cut down regardless of whether or not there was a demand for the fuelwood.
But in between these two extremes we have a great many cases in which the situation is much less clear. For example, a farmer clears an area of marginal land and sells the trees he has cut down to be made into charcoal. He plants up the plot, finds after one year that it is not viable and abandons it When we look at the price obtained for the cut trees, it turns out to be comparable to the value of the plot for agricultural purposes. Is this a case of an unsuccessful attempt at agricultural expansion, or the commercial felling of trees for fuel?
Since the latter description would almost certainly make the operation illegal. It is extremely unlikely that any survey return will give us the correct answer.
An unsuccessful attempt at agricultural expansion
Accordingly, it is impractical to consider our activities in the context of national woodfuel demand. It might be preferable to focus on a particular sector where the environmental damage per tonne of wood consumed is particularly high -such as the institutional sector in Kenya.
Our principle grounds for believing that institutions consume a higher-than-average proportion of harvested wood is that the normal unit of purchase for fuelwood among institutions is the trailer-load: one to two tonnes. While a domestic consumer might gather a back-load (30 - 40 kg) from dead branches over an area of a hectare or so, it is unlikely that the owner of a tractor and trailer would spend a proportionate amount of time gathering across 50 hectares. Whatever the legalities, systematic harvesting seems much more likely.
Given that we are focussing on the institutional sector, there are three components to an impact assessment for a project such as this one. We must evaluate:
· the average change in fuel consumption which takes place at consumption centres when the new systems are introduced
· the proportion of the total number of consumption centres in that sector into which it is estimated/projected that the project has succeeded or will succeed in introducing improved systems
· the total woodfuel consumption due to that sector which is supplied by harvested trees, as opposed to trees which are cut or fall down other reasons.
Now the first two of these figures are relatively straightforward. We illustrate using the institutional fuelwood saving programme established by this project as an example.
a) Evaluating the change in fuelwood consumption
As a component of the institutional fuelwood saving programme pre-implementation study, undertaken in 1987, a consultant from the Woodenergy Systems Group, W. F. Sulilatu of TNO Apeldoorn, the Netherlands, carried out an evaluation of the field performance of the projects improved systems compared with that of other institutional cooking systems currently available in Kenya.
The consultant found that the improved systems consumed between 40% and 75% less fuel than the alternatives, on the basis of comparative cooking tests. These tests were carried out using fuelwood with a standardised moisture content. Since the programme also provides training in fuelwood preparation, this may result in further savings which are difficult to quantify.
Field reports suggest that in the majority of cases savings observed are of the order of 70%. This may be partly due to improvements in fuelwood preparation associated with the introduction of the new systems, and partly because the systems replaced tend to be the least efficient ones. It is also possible that savings may be exaggerated by operators enthusiastic about their new equipment.
Even greater savings of woodfuel are realised when the improved stoves replace charcoal burning systems (as they often do), owing to the inefficiency of current charcoal conversion techniques in Kenya. Accordingly, we consider the consultants figures on fuelwood consumption to give a realistic (and, if anything, conservative) estimate of the probable savings achieved.
b) Evaluating actual/potential market penetration
The market survey carried out by Mwaniki Associates identified a market for institutional stoves of this type of between 400 and 500 units per annum in 1987. Total installations carried out by the programme in the course of 1988 numbered 220 units, giving us a market share of about 50%.
Accordingly, if the programme does no more than maintain its present market share, we may expect it to achieve an eventual reduction in national institutional woodfuel consumption of 20 - 37% over a six-year period (Mwanikis estimate of the average replacement cycle for institutional stoves in Kenya). If the programme can increase its market penetration to 80%, as seems possible if scale-up capital is obtained and a credit scheme introduced for programme beneficiaries, this reduction would increase to 32 - 59%.
Thus of the three components of an overall impact assessment listed above, we have reliable data on the first two, which give us an indication of the potential percentage reduction which the programme can achieve in fuelwood consumption in the institutional sector: up to 69%.
c) Evaluating total impact
Translating this into absolute figures for savings achieved is more problematic. There are three possible impact measures which a funding and/or implementing agency might be interested in: woodfuel savings in tonnes and/or hectares-of-forest-equivalent; cash savings at the individual-institution level; and economic benefits at the national level.
All of these depend on estimates of national institutional woodfuel demand, which are very unreliable. The data we have to work with comprises the market survey carried out to 1987 by Mwaniki Associates, which identified a total of about 2000 woodfuel-using institutions in Kenya; and a survey of 40 institutions for which the principle fuel was wood or charcoal, carried out by the Kenya Energy and Environment Organisation (KENGO) with the International Development Research Centre of Canada (IDRC) in 1985.
Both surveys broadly categorised institutions into school and colleges, military institutions and prisons, and hospitals and health-related institutions. On the basis of average consumption per institution within each category (from the KENGO/IDRC survey) and the total number within that category (from Mwanikis survey), we find that 2000 institutions would consume approximately 450 thousand tonnes of woodfuel per annum, either directly in the form of firewood or in the form of charcoal, taking a typical conversion rate of 5 tonnes of wood to one tonne of charcoal.
It must be acknowledged that the majority of project impact-assessment exercises would leave the statistics at this point. If we look a little closer, however, we find that the averages on which the above estimate is based are subject to standard deviations of the order of 100%, which renders any inferences from them suspect, to say the least.
A better measure, one might assume, would be woodfuel consumption per individual fed through an institutional catering establishment. This would eliminate the spread due to the variation in institutions populations in the various categories. This approach yields a total consumption of 530 thousand tonnes per annum feeding a total of 730 thousand people. But again, the standard deviations involved are still of the order of 100%, indicating the wide spread of efficiencies achieved with traditional institutional cooking systems (none of the institutions surveyed by KENGO/IDRC were using Bellerive improved systems).
The conclusion to be drawn from all this is that any estimate of national-level demand for woodfuel cannot be anything more than an order-of-magnitude indication.
For the sake of investigating the implications, we may take a national institutional woodfuel demand of 500 thousand tonnes, based on the above two figures.
If the programme increases its market share from the present 50% to 80%, then the eventual reduction in demand would be approximately 50% (taking W.F. Sulilatus performance findings and weighting according to numbers of units of each type of stove found in use by KENGO/IDRC).
On the above assumptions, this translates to an annual saving of 260 thousand tonnes of woodfuel, for which institutions would otherwise be paying a total of about 47 million Kenya shillings (2.6 million US dollars) at 1088 prices.
Given that the total cost of a fully sustainable programme of the required volume (including new stoves, servicing, regular replacement of delivery vehicles etc.) would be approximately 8 million shillings (US$450,000) per year at 1988 prices, this figure gives us an indication of the viability of the programme from the individual beneficiaries point of view. It does not give us a figure for the net economic benefit to the nation as a whole.
The average price paid by an institution per tonne of woodfuel (fuelwood or charcoal-equivalent) was found by KENGO/IDRC to be KSh 155/= in 1985 (standard deviation about 30%). After inflation, this is equivalent to approximately KSh 188/= in 1988, which is the basis of the above value of savings. But if we consider the origin, of this price, it is clear that it does not represent the true economic value of the trees cut down. This KSh 188/= is essentially the cost of the labour required to cut down the trees and transport the wood to the school, or to convert the wood into charcoal first, as the case may be.
The problem of placing an economic value on 250 thousand tonnes of woodfuel, when we are not even sure of how much of that wood was cut down specifically to provide fuel in the first place, is one which has exercised a great many authors, and no clear consensus has yet emerged as to how this should be done.
Thus: an overall cost-benefit analysis, from the point of view of individual programme beneficiaries, is possible, but only if we rely on figures for national sectoral demand which cannot be anything more than indicators of order-of-magnitude.
Attempting to produce a CBA from the point of view of national accounts (with or without some sort of compensatory terms for capital environmental resource consumption) is clearly more problematic, and may be impossible.
Following the Market Survey and Field Performance evaluation it was decided to retain the programme within the overall framework of the Bellerive Foundation, despite the fact that it is now generating enough revenue to cover its operational costs.
The current scale of operations, however (we are the largest single supplier of institutional catering equipment in Kenya with an estimated 50% market share), means that systems must be adopted similar to those operated by commercial concerns. The basic structure is as follows:
Structure of the ongoing programme:
· A network of District Focus Firewood Conservation Projects (DFFCPs) is being established as a component of the District Focus initiative of the Government of Kenya. Each DFFCP is individually funded by a donor. The DFFCP concept was piloted successfully in Taita-Taveta district with support from the Danish International Development Agency, DANIDA. Two further DFFCPs have been launched in Embu and Nakuru districts, with the assistance of the Overseas Development Administration of the U.K., and the Worldwide Fund for Nature respectively. DFFCPs are under development in conjunction with Danida and Norad (the Norwegian bilateral aid agency) in South Nyanza and Bungoma districts. (See map on page 4)
· Each of the DFFCPs incorporates or will incorporate a field officer, responsible (inter alia) for promoting the institutional fuelwood saving programme at the regional level, and a regional dissemination team (RDT): a team of craftsmen, responsible for all installation, training and service-follow-up work in their district.
· A small Programme Coordinating Unit, based in Nairobi, manages procurement of components and materials (from Kenya-based suppliers), coordination of orders, distribution of stove kits (direct to programme beneficiaries), national level promotion and all programme finances. In this way kits are delivered direct to the institution and payments remitted direct to Nairobi.
Thus the craftsmen actually installing the systems and carrying out training in the field do not have to have access to delivery vehicles nor administer the programmes cashflow (currently of the order of 4 million shillings - c.US$230.000 - per annum), all of which is essential to controlling the cost of such a decentralised approach to dissemination.
Prices charged to programme beneficiaries are currently calculated to cover
· costs of materials and components of stove kits
· delivery costs (excluding depreciation on vehicles)
· salaries of staff directly involved in the dissemination programme
· Communications costs (excluding promotion)
Our current market penetration is estimated at c. 50%: the total market was evaluated by Mwaniki Associates to be 400 - 500 units per annum and the programme installed c. 220 units in the course of 1988. Given the relatively poor performance of other systems available (on average they consume at least 100% more fuelwood than a properly managed Bellerive system), environmental considerations indicate that this penetration should be increased.
With an expanded distribution capacity and the increased marketing and installation capacity given by an eventual network of six DFFCPs in place of the present three, a market penetration of up to 80% is perfectly feasible. Once this position is achieved, it will be possible for the burden of promotion and depreciation costs to be transferred to programme beneficiaries with no increase in real unit prices.
Thus when scale-up is complete, the programme will be fully self-sustaining with an indefinite life-time. It is proposed that the scale-up should be undertaken over a five-year period, partially funded by external donors: details of the various components are given below.
This document is intended to provide ideas based on our experience to date. Accordingly, we provide only a brief indication of proposed developments.
a) The future of the Institutional Fuelwood Saving Programme.
We have explained above why we do not feel it would be appropriate at this stage to turn this programme over to the private sector. The essential training element would be neglected, and the abrupt change in pricing policy, which would be required by a private operator needing to cover depreciation on capital equipment to service a debt on scale-up costs and/or to provide a rate of return on investment typical of the Kenyan private sector, would have a negative effect on the programmes environmental impact.
Accordingly, a five-year scale-up project is proposed, during which time donor funds will be utilised to expand the programme such that, by the end of the 5-year period, economies of scale will allow the costs which are currently born by donors (principally depreciation on capital) to be transferred to beneficiaries with no change in real unit prices.
This scale-up operation comprises three main elements:
· a revolving fund to provide concessionary credit faculties to programme beneficiaries, to be administered by a leading Kenyan Bank (lack of access to capital was identified in the market survey as a major constraint on potential investors in improved stoves, despite short payback periods - average 1.5 years)
· a national promotional campaign
· capital to expand the distribution capacity of the programme: storage facilities, delivery vehicles etc.
The scale-up project will be undertaken in conjunction with the establishment of further District Focus Firewood Conservation Projects (see above).
We have mentioned the ongoing technical development work being undertaken into:
· mass-producible stoves for the higher-income domestic sector
· low-cost/low-durability systems for small-scale commercial catering establishments
· training materials in Kitchen Energy Management for the low-income domestic sector
· fuelwood production units for larger institutions.
Further work is required, examining the following:
· the potential for the introduction of alternative fuels to supplement and/or replace fuelwood in the institutional sector in key regions (including agricultural wastes, sawdust, solar pre-heating of water etc.)
· mechanisms whereby other fuelwood saving systems may be disseminated through the DFFCP network (addressing baking, small-scale fuelwood-using industries etc.)
b) Beyond Kenya.
It should be clear to those who have read this report that straightforward replication of the Kenya programme in other parts of Africa is out of the question. Strategies must be developed tailored to the conditions prevailing in other countries. Recognition of every countrys unique environmental-economic situation is essential to success.
However, we believe that the experience gained in developing the Kenya programme has left us with a substantial amount to offer elsewhere, and the overall approach, combining the introduction of improved technology with an extension programme aimed at improving firewood management, should prove generally applicable.
The increasing pressure on the continents remaining reserves of standing trees was noted as a cause for grave concern by the participants of the
First African Ministerial Conference on the Environment, held in Cairo in 1985. Initiatives which result in genuine reductions in fuelwood consumption, in those sectors which are doing direct environmental damage (see above), have a clear place in the follow-up to this conference.
The approach adopted by the UNEP/Bellerive project in Kenya, whereby savings are achieved in a sustainable manner through the progressive transfer of programme costs onto project beneficiaries. Is also clearly in line with the socio-economic developmental objectives of a large number of African nations. Invitations have already been received to establish institutional fuelwood saving programmes in Tanzania and in Northern Nigeria - at the behest of the Tanzania Ministry of Energy and Minerals and the Nigerian Conservation Foundation respectively.
In each of these countries, and any further areas in which we establish operations, we aim to adopt a two-pronged approach: introducing a suitably modified version of the fuel-saving systems as developed for Kenya, while at the same time establishing a capacity to develop strategies and technologies on the ground. Thus we can aim to achieve an immediate impact while. In the long run, the Kenyan systems will be replaced by those developed entirety within the project host country.
It is certainty unusual for a non-governmental organisation to be implementing a fully-developed marketing operation, on a scale comparable to a private-sector concern, and, on the face of it, in direct competition with private business. Is it right?
There seems to be a general consensus that direct donor-assisted intervention should be restricted to such areas as research, education, extension services and the development of infrastructure. Spheres of activity which are traditionally the preserve of the private sector, such as the production and marketing of consumer durables, should be left alone. Any attempt to intervene directly in such areas, beyond such seminal activities as technical development, the creation of public awareness, and the provision of launch capital must automatically, be unsustainable.
All too often such direct intervention has taken the form of crude subsidies, and for obvious reasons programmes have tended to collapse as soon as the donor subsidy is removed. This has led to the conclusion that, once a new product, technology or system is introduced, either it will be taken up by the private sector, or it will have failed the test of the market place and should be abandoned.
The institutional stove dissemination programme is clearly no longer a pilot scale activity. It is the largest single source of this type of catering equipment in Kenya, it might be argued that, as a non-profit organisation maintaining such a significant market presence, we are having a negative effect in the long-term, since prices are thereby depressed, discouraging the establishment of independent private-sector production.
We could defend our continued direct involvement with the fact that we are using the ongoing dissemination of improved catering equipment as a means to fund an extension programme aimed at improving the management of fuelwood in institutional catering.
Thus, even though the programme is run on a non-profit basis, the prices charged to programme beneficiaries, with these training costs built in, are such that we are not undercutting a private entrepreneur who wishes to produce and market institutional stoves over a limited region.
Such small-scale production in densely populated parts of the country, is commercially viable, and at least five independent producers are currently in operation installing stoves of the Bellerive design in their own districts. Arguing that we are not undercutting the private sector does not, however, provide a positive justification of our approach. For that, we must consider the situation on a wider scale.
The 1987 market survey found that a national-scale programme, reaching those remote, poorer regions (such as the arid north of the country) which arguably have the greatest need would not be an attractive venture for a private company. Should these remote regions, which cannot support a programme run on a profit-making basis, Just wait until shortages put up the price of fuelwood sufficiently for programme beneficiaries to become prepared to pay a much higher price for fuel-saving equipment?
It is not clear whether such a non-interventionist approach would work, since the main constraint determining the present rate of dissemination is not the beneficiaries willingness to pay, but their ability to raise the necessary capital. Most target institutions have relatively rigid patterns of expenditure, and the poorer the institution, the less flexible its budget. It is difficult to see how an even higher price of fuelwood could make more capital available for the purchase of improved stoves.
Moreover, the link between fuelwood availability and fuelwood prices is far from straightforward. Fuelwood at source, is still effectively free, however scarce it may have become. An institution is paying only for delivery, not for the fuelwood itself. These delivery costs will increase as shortages become more severe and fuelwood has to be transported over greater distances, but it is unrealistic to expect a smooth, predictable relationship between price and supply for a given level of demand.
Thus, despite the obvious need for fuelwood conservation in institutions, the market cannot, at present, support a national programme unless it is operated by a non-profit-motivated organisation such as the Foundation. And this situation will not be resolved through the operation of market forces. Does this mean that we are committing ourselves to indefinite involvement, that as soon as the Foundation departs the majority of institutions will go back to using three stone fires?
the non-interventionist approach...
No. Because the project has focussed on a specific, limited sector of the fuelwood economy, the sheer scale of the ongoing programme is now having a direct, significant impact on the market situation within that sector. We are in effect, creating the conditions required for a sustainable programme.
To begin with the most obvious impact: the programme is creating awareness among institutional fuelwood consumers of the scope for fuelwood savings, and the financial benefits to be derived from the use of improved systems. Perhaps more importantly, the success of the programme has alerted the private sector to the potential of the market for improved institutional stoves. The independent producers mentioned above have established that by operating over a limited region, omitting the training component, and charging prices comparable to the programmes, they can generate an attractive return on investment.
But we have found that substantial fuelwood savings depend on the training component not being ommitted. This brings us to the fundamental impact of the ongoing programme, beyond the simple creation of awareness. At present, Bellerive institutional stoves are in use in only 10 to 15% of Kenyan institutions. The fuelwood management training package, introduced towards the end of 1987, is even more of a novelty. Consider the situation in 5 years time if the programme does no more than maintain its current share of the market.
By that time more than half the institutions in the country will be using stoves installed by the programme, and most of these will have received, at some stage, the benefit of the training package. The system will be, de facto, the national standard.
Once this position is achieved, the bulk of the work being undertaken by the programme will be the servicing and rebuilding of old Bellerive stoves, retraining cooks and so forth, as opposed to new installations.
The servicing of existing stoves will require a much lower level of working capital, and much less investment in delivery vehicles, than the installation of new ones. It will therefore be substantially easier to transfer to the small-scale independent private sector.
Once the fuelwood management techniques introduced by the programme are in general use, it may be possible for standards to be maintained by a scaled-down extension service: perhaps through the medium of are you remembering to... posters, or through the newly-established extension network of the Ministry of Energy, or through the Forest Department.
As the national standard, the improved cooking systems should no longer require an extensive promotional campaign: one of the key costs which ruled out the transfer of the programme to the private sector in 1987. Thus the Foundation will then be able to terminate direct involvement without jeopardising the programmes environmental objectives.
We may find that a new way of doing things, using a new product, for example, has clear environmental benefits, but is not economically justified. One response is to abandon the innovation, and with it the environmental benefits it was intended to yield. A second response is to defy market forces, using subsidies or regulation to impose our innovation onto the target community. The first approach will always work, but might not achieve very much. The second will work very seldom.
We must compromise: recognising that market conditions are unfavourable at the outset, but developing a strategy to guide market forces such that they eventually work towards our overall objectives. It may seem ambitious for a development project to set out to alter the market situation, but this project has shown that if we focus on a sufficiently well-defined sector of the economy, such that the scale of our intervention is significant, this approach may be the most realistic available.
...may not achieve very much