The Malawi charcoal project experience and lessons

9320 INDUSTRY AND ENERGY DEPARTMENT WORKING PAPER ENERGY SERIES PAPER No. 20 The Malawi Charcoal Project Experience and Lessons January 1990 The World Bank Industry and Energy Department, PPR  THE MALAWI CHARCOAL PROJECT EXPERIENCE AND LESSONS Prepared by Witold Teplitz-Sembitzky (IENED) Gerhard Zieroth (Consultant) Industry and Energy Department Policy, Research and External Affairs The World Bank January 1990 Copyright(c) 1990 The World Bank 1818 H Street, NW Washington, DC 20433 This report is one of a series issued by the Industry and Energy Department for the information and guidance of Bank staff. The report may not be published or quoted as representing views of the Bank Group, nor does the Bank group accept responsibility for its accuracy or completeness.  ABSTRACT This paper draws together the experience the Malawi Charcoal Project which was initiated in late 1986, has gained until mid-1989. With a total installed capacity of 9,500 tons per year, the Project not only is the largest semi-industrial charcoal production program which has been implemented in SubSaharan Africa to date, but also has gathered a comprehensive set of data on various matters that similar activities would have to cope with. A unique feature of the Project is that its feedstock is provided by softwood wastes generated on large government plantations. Given this unusual resource base and its location, the Project had to seek for new solutions regarding the logistics and organizational set-up of production schemes, the choice of carbonization technologies, and the marketing of a product which, in terms of primary and secondary fuel properties, has little in common with traditional hardwood charcoal. Also, the Project had to come to terms with disadvantages at which softwood charcoal is placed due to large transport distances and handling difficulties. While softwood charcoal can be used by households for cooking, combustion trials conducted in the industrial and agroindustrial sector have shown that there is an additional market potential in non-household applications. Particularly promising are the test results from the tobacco curing industry where the use of softwood charcoal, rather than fuelwood, helps improve the quality of cured tobacco. All in all, the Project has demonstrated that softwood charcoal in many instances is a technically feasible and economically viable alternative to fuelwood and/or coal. However, additional marketing initiatives and a stronger policy support will be required to foster a wider use of softwood charcoal.  - ii - Table of Contents Page No. 1.0 Introduction .................................. 1 2.0 Charcoal Production from Waste Wood ................ 2 2.1 The Resource Base ............................. 2 2.2 Feedstock Valuation ............................. 5 2.3 The Cost-Competitiveness of Viphya Charcoal .... ........ 12 2.4 Kiln Design and Conversion Efficiency ................. 15 2.5 Fuel Properties ................................ 18 2.6 The Organizational Setup ......................... 21 2.7 Charcoal Production Costs ......................... 22 3.0 The Market Potential for Softwood Charcoal ................................... 26 3.1 The Industrial Market ............................ 26 3.2 The Tobacco Industry ............................ 28 3.3 The Household Sector ........................... 32 3.4 Medium Term Prospects .......................... 34 4.0 Transport, Packaging and Handling ................... 36 4.1 Charcoal Transport ............................. 36 4.2 Charcoal Packaging .38 4.3 Charcoal Handling .............................. 39 5.0 Marketing Trials and Publicity Campaigns ..... .......... 40 6.0 Summary and Conclusion ........................... 46 6.1 Fuelwood Crisis or Wood Surplus ....... ............. 46 6.2 Improved Charcoal Production ......... ............. 47 6.3 Potential Uses of Charcoal ........... ............. 48 6.4 Market Potential and Prospects ........ ............. 49 6.5 Institutional Issues .............................. 51 6.6 Lack of Commercialization . ........................ 52 REFERENCES ...................................... 53 ANNEX I: Welfare Implicatfons of Uniform and Discriminatory Stumpage Fees ...................... 54 ANNEX II: Performance Indicators of Charcoal Production at Viphya ......... ................... 58 ANNEX III: Industrial Fuel Consumption in Malawi ................ 62  - iii - List of Abbreviations cif = cost, insurance, freight FD = Forestry Department ha = hectare kg = kilogram mcwb = moisture content wet basis MJ = megajoule MK = Malawi Kwacha (mid 1989: 1 MK = 0.4 US$) m3 = solid cubic metre SFC = specific fuel consumption sm3 = stacked cubic metre t = metric ton tpy = (metric) tons per year  1. Introduction The Malawi Charcoal Project was initiated in late 1986 when the Malawi Government and the World Bank agreed to undertake an action-research project on the viability of charcoal production from wood wastes generated on government forest plantations. The idea underlying the project was that charcoal made from wood wastes could substitute for woodfuels supplied from customary land and, thus, reduce the gap between sustainable woodfuel supply and demand. Another role charcoal was hoped to play was that of an economically attractive substitute for imported and domestic coal. The Pilot Phase of the project which was carried out between October 1986 and January 1988 established convincing evidence on the technical feasibility of converting softwood residues into high-quality charcoal. It also identified potential market outlets in the residential, industrial, and agro-industrial sector and concluded that the prospective demand for softwood charcoal could be met in an economically viable and environmentally sound manner. What remained to be accomplished was the formidable task of commercializing all operations involved in the manufacture and distribution of the new fuel, while at the same time creating both the administrative framework and the policy instruments that would be required to effectively integrate its production and utilization into a national energy supply strategy. Therefore, the decision was made to discontinue financing the activities from the Wood Industries Restructuring Credit and embark on a commercialization phase as a special component of the Energy I Project which was launched in 1989 (World Bank, 1989). The commercialization phase which was funded out of advances for project preparation granted by the World Bank ended in August 1989, with mixed success. Of the total installed charcoal production capacity, 3,000 tpy have been contracted out to the private sector, with the balance of 6,500 tpy operating under a treasury fund account. While the tobacco curing industry is expected to purchase about 2,500 tons of charcoal in the 1989/90 curing season, and Portland Cement has agreed to take delivery of charcoal fines at an initial rate of 200 tons per month, only a small fraction of the output finds its way to household markets. Moreover, a comprehensive national solid fuel supply strategy with charcoal from plantation-derived waste wood as an integral component still needs to be developed and imp.2.liented. Thus, further efforts have to be made to consolidate ar.J upgrade the achievements of the project. However, with the Malawi Charcoal Project now being under way for almost three years, it may be the right time to draw together the project experience and findings and to pinpoint the lessons which can be learned from the activities carried out to date. There is a variety of features Malawi has in common with other Sub-Saharan countries, such as the heavy reliance on woodfuels and the environmental threats posed by a degrading natural resource base. But Malawi's situation is also unique in several respects so that the approach the Charcoal Project has followed may not be applicable to other cou,ntries. Anyhow, a great deal of experience the Project has made is in itself worth telling, and much of the  - 2 - knowledge that has been acquired in the course of implementing the Project may be relevant to similar undertakings in Sub-Saharan Africa. The present report summarized the technical performance of the Project, reconsiders the institutional/organizational arrangements it has faced or created, provides an overview of the potential niarkets for softwood charcoal and the attempts made to penetrate these markets, assesses the constraints on softwood charcoal use in the residential, industrial and agroindustrial sectors, and rex-+ews the prospects and objectives the Project has aspired to meet. Various insights gained and conclusions drawn from the success stories and failures the Project has experienced may help enhance similar activities in the area of charcoal production and marketing which are underway or envisaged in other developing countries. 2 Charcoal Production from Waste Wood 2.1 The Resource Base A unique feature of the Malawi Charcoal Project is that its feedstock comes from large, underutilized government timber plantations. Altogether, these plantations cover a total of about 90,000 ha. and are located throughout the country (see Figure 1.1). More than 80% of the publicly managed industrial forest areas are covered with softwood pine species, mainly pinus patula. By far the largest plantation is the Viphya forest (53,250 ha.), the stands of which are almost entirely composed of softwood species. Planting activities started in the early 1950s and discontinued by 1983. Due to economic constraints the initial idea of using the Viphya resources as a feedstock for pulp/paper production never materialized. The only outlet for Viphya wood is a recently constructed sawmill and plywood factory (Viply) served by a concession area of 9,000 ha. Under the premise that the entire Viphya plantation eventually can be put into productive use, government policy has -and still is - to preserve the resource base through proper silvicultural treatment. Given the age structure of the stands, such a program would be feasible for another 12 to 15 years, resulting in a steady flow of thinnings of about 130,000 tpy. Additional waste wood will be generated on the Viply concession area and from unavoidable clearfellings of overmature stands. Altogether the plantation can be expected to annually supply a minimum of 200,000 tons of non- marketable softwood over a period of at least 12 years.  - 3- Figure 1.1: Major Softwood Plantations TO DAR ES SALAAM Z~ ~ A M T A N Z A N I A Z~~~~~~~~~~~EZ ACMOBGINA 5 t30 ZMBAIMT. 3,580 N Q rMBHABE50 h RAILWAYSE2\5z |DZONZI MVAI2 300 ha A aTO NACALA |CHAMBE 5~~~~~~~~~~~~00 ka-v_n m2  - 4 - If the country's other timber plantations are taken into account, the total volume of non-marketable softwood from thinnings and sanitary clearfellings may prove as high as 390,000 tpy (see Table 2.1). To put this figure into perspective, it suffices to note that the sustained yield from customary land is estimated at 1.5 million tpy, whereas the sustained yield from existing woodfuel plantations may be around 40,000 tpy. Table 2.1: Estimated Waste Wood from Non Fuelwood Plantations (tpy) / Viphya Chambe Others Total Low Case 200,000 3,500 27,500 231,000 High Case 275,000 7,000 105,000 387,000 !/ Available over a period of at least 12 years. Source: IPC (1988) If the available feedstock were converted into charcoal on the basis of brick kiln technologies used by the Project, the potential ciiarcoal output would range between 72,000 tpy (low case) and 120,000 tpy (high case). For comparison, current industry demand for coal works out at 40,000 tpy, while urban household charcoal consumption is estimated at 50,000 tpy. Clearly, there is no a-priori reason to carbonize the total waste wood generated on government plantations (In some locations it would be more economic to use at least part of the resources as fuelwood l). Nor would it be feasible to immediately embark on a large-scale charcoal production program. What the figures show, however, is that there exists a large, untapped, and hitherto neglected biomass resource base that could be exploited: to redress imbalances in woodfuel demand and supply, to reduce the level of coal imports, or to lower the rate at which domestic coal reserves are extracted. (At the current output of 30,000 tpy, the Kaziwiziwi/Mchenga deposits are likely to be exhausted within the next five to six years). For instance, at current costs the breakeven transport distance for Viphya charcoal varies between 100 km and 170 km, depending on the differential between the efficiency of charcoal and fuelwood use.  -5- According to governrnent policy objectives, highest priority should be given to the first option. Even though a comprehensive biomass inventory has yet to be prepared, there are reasonably accurate figures available indicating that the currently prevailing level of woodfuel demand cannot be met on a sustainable basis. As is shown in Table 2.2, the potential for sustainable wooofuel supply lies in the vicinity of 5 million m3, while overall demand is estimated to range between 8 and 10 million m3 per year. With total biomass stocks amounting to more than 300 million mn3, Malawi's woodfuel resource gap is probably not as alarming as in other African countries; but this does not relieve policy makers of the need for mitigating action. More importantly, while approximately 50% of the sustainable supply is located in the northern region, woodfuel demand is concentrated in the central and southern parts of the country. Thus, in closing the woodfuel resource gap, it becomes a .inajor task to channel surplus wood from the north to deficit areas in the south. Clearly, in order to stem a further detoriation of the country's biomass resource base, a wide range of initiatives can and should be taken (and already are underway), ranging from improved woodland management to smallholder afforestation programs. There i3, however, no immediate alternative to the option of supplying charcoal from government plantations: Waste and surplus wood generated on government timber plantations is the only additional .ource of woodfuel which is readily available; and since the lion's share of this resource base is located far away from major market outlets, its conversion into charcoal becomes an ecc,nomic imperative (which not necessarily implies that this option is economic). 2.2 Feedstock Valuation With the commissioning of the Project there arose an intense debate over the economic value that should be attached to the feedstock used for charcoal production. Different concepts and interests affected the discussions. As far as the government was concerned, pressing revenue needs were advanced in favor of high stumpage fees. But what would be the warranted level of fees, and how should they be implemented? Energy planners raised the question of whether feedstock valuation should be aligned with the replacement-cost principle. Policy makers, on the other hand, argued for valuation schemes that conform with uniform woodfuel pricing rules applied throughout the country. Agreement was finally reached on the Project's proposal to determine stumpage fees on the vasis of net-back considerations. The rationale underlying this decision is as follows: At Viphya, where the charcoal production activities are concentrated, as well as on other government plantations, the economic value of thinnings and sanitary clearfellings should be inferred from their potential use as woodfuel, since alternative uses (e.g. methanol production) are economically less attractive (IPC, 1988, Chapter 10). In fact, in the absence of the woodfuel option the price waste wood could command would be zero, if not negative (since silvicultural treatment incurs costs that cannot be recovered).  - 6 - Table 2.2: Woodfuel Potential Malawi. 1988 W' Region Source Northern Cenwral Southern Total Forest Reserves 400,000 414,000 258,000 1,072,000 Customary Land 1,640,000 730,000 250,000 2,620,000 Private Forest plantations - 105,000 153,000 258,000 Subtotal 2,040,000 1,249,000 661,000 3,950,000 Government Plantations 643,000 231,000 192,000 L.066,00 Total 2,683,000 1,480,000 853,000 5,016,000 gi m3 per annum. In the case of forest reserves, private forest plantations and customary land, the potential is equal to the estimated sustained yield. For Government plantations the potential covers thinnings and sanitory clearfellings (low case), plus the sustained yield from fuelwood and non-timber plantations (about 70,000 m3). Not included are agricultural residues and wastewood from timber processing. Source: IPC (1988a) Replacement-cost considerations, on; the other hand, have been dismissed because the plantations are designed and run as timber, plywood or woodpulp production schemes with thinnings as a by-product. The cost of replacing thinnings with plantation- grown fuelwood would only matter if the decision were made to abandon the currently prevailing management regimes. Also, the imposition of gazetted stumpage rates which the Forestry Department uniformly applies to fuelwood put into commercial uses would make little sense, because the yardstick for fuelwood production schemes, in terms of which these rates are defined, does not even remotely resemble the conditions under which fuelwood could be grown at a location like Viphya.  -7- So the Project advisors have proposed to derive the stumpage fees irom the sales revenues borne by woodfuel niarkets, i.e. to net the price charcoal commands at major market outlets back to the stump. Unfortunately, charcoal markets are not only spatially distinct; they also differ in terms of prices at which charcoal is traded and sold. In particular, prices tend to be highest in Blantyre, which not only accounts for 60% of overall urban household consumption of hardwood charcoal (and, in addition, is in the center of potential industrial charcoal users), but also is most distant from the Viphya plantation as the major source of softwood charcoal. Therefore, the Project Management has applied as pricing principle a kind of minimum welfare requirement, according to which in Blantyre the landed costs of softwood charcoal coming from Viphya should not exceed the price at which nardwood charcoal can be obtained at the wholesale level.2 Given this price cap provision, the stumpage fees levied on the wood equivalent of, say, one ton of charcoal delivered to Blantyre should be equal to the value charcoal commands on the stump after the costs of charcoal production (plus overheads and profits) and the expenditures for charcoal transport/handling are netted out. One advantage this formula has is that adjustments in the level of stumpage fees are only required if the Blantyre price cap changes while the charcoal production and transport costs remain constant, and/or there is a change in charcoal production and transport costs the price cap does not replicate. This advantage, howeve-, comes with the problem that some kind of reviewing procedure needs to be established in order to ascertain whether the level of stumpage fees fixed in the past continues to be consistent with both the costs of charcoal supply and the wholesale prices prevailing in Blantyre. In particular, there is no single wholesale price that clears the market. Rather, wholesale prices vary within some range, and this price range may widen/narrow or move upwards/downwards. Moreover, since charcoal is a potential substitute for coal (at least in industrial uses), the relationship between the landed costs of cval and (softwood) charcoal matters as well. In practice, th'ough, the proposed stumpage fee formula would have been less difficult to apply than these 2 From a statutory point of view, most of the hardwood charcoal entering urban markets is "illegal" since it evades the gazetted stumpage tax levied on wood coming from customary land. In this respect, the minimum welfare requirement implies that consumers switching from hardwood to softwood charcoal should not be worse off. For all practical purposes, however, the requirement means that softwood charcoal should not price itself out of the Blantyre market.  -8- problems suggest. (Note that stumpage fees have not yet been collected since so far the responsibility for the charcoal operations has been with the Forestry Department). As is shown in Table 2.3, in early 1988 the stumpage fees would have worked out at MK 10 per ton of Viphya charcoal (which is equivalent to MK 0.75 per sm3 of pinewood thinnings), relative to "cheap" hardwood charcoal that wholesalers could obtain at 165 MK/t. Also, with landed costs amounting to 165 MK/t, Viphya charcoal would have been competitive vis a vis (domestic) Kaziwiziwi coal. One year later, however, a dramatic rise in transport costs as well as increases in the expenditures for charcoal bagging and charcoal production would have resulted in landed costs of 280 MK/t, if the stumpage fees had been kept at the 1988-level, and if charcoal producers at Viphya had sought to recover a profit/overhead margin of 10 MK/t. This figure would ha-e exceeded the landed costs of "cheap" hardwood charcoal by 15 MK/t, but would not have eroded competitiveness of softwood charcoal vis a vis Kaziwiziwi coal.3 While the parallel increase in the landed costs of Kaziw;ziwi coal can almost exclusively be attributed to higher transport rates, hardwood charcoal became more expensive mainly because of the growing scarcity of its resource base and the threat of confiscation posed by tighter controls of the inflow of "illegal" charcoal.  - 9 - Table 2.3: Cost Structure of Viphva Charcoal Delivered to Blantye (MK/t) Landed Costs Wholesale Price Transport cost Cost of Kaziwiziwi Coal Range Hardwood Viphya Charcoal Charcoal (1) Charcoal (2) (2) Bagging early 1988 170 165-180 100 15 early 1989 290 265-290 200 25 mid 1989 290 265-290 200 15 Charcoal Profits & Stumpage Landed Costs Production Overheads Fees Viphya Charcoal Costs (1) early 1988 30 10 10 165 early 1989 35 10 10 280 mid 1989 50 10 10 285 (1) Blantyre, (2) Viphya-Blantyre by road (630 km) Thus, with unchanged stumpage fees, in early 1989 Viphya charcoal would have found it more difficult to penetrate the Blantyre charcoal market; yet this obstacle might well have created an incentive to cut production costs. In mid-1989 the delivered costs to Blantyre would have further increased to 285 Mk/t, since government raised the administered minimium wages by 100%. It is assumed, however, that this wage push could have been offset by a cut in packaging costs (see section 2.6 and 4.2).  - 10 - So there is some evidence that the stumpage fee formula proposed for Viphya charcoal, with Blantyre as the principal market outlet, could be maintain.-J even in the presence of escalating costs and prices. Two questions, though, remailk .-, be answered: What should be the stumpage fee for Viphya charcoal shipped to outlets other than the Blantyre market? And what value should be placed on thinnings from other government plantations that may also serve as i feedstock for charcoal production? As regards the first question, the principle choice is between (a) a system of uniform stumpaq fees set equal to the netback value "earned" by Viphya charcoal shipped to Blantyre with the effect that the landed costs at different market outlets would vary in direct proportion to the transport distance, and (b) a system of discriminating stumpage fees set in accordance with the final destination of Viphya charcoal so that the landed costs throughout the country would be uniformly kept at the level prevailing in the Blantyre market. At first sight the latter option appears to be most attractive since it may help maximize government revenues from charcoal production at Viphya. On second thought, however, there are at least three shortcomings that render this approach undesirable. Firstly, discriminatory stumpage fees would encourage cheating. For instance, charcoal which is supposed to go to Blantyre and, therefore, would be charged with a comparatively low stumpage rate could be sold more profitably, say, in Lilongwe. Secondly, the implementation of a scheme of market-outlet-dependent stumpage fees would place a significant administrative burden on the Forest Department. In particular. the measures and manpower required to monitor and control the flow of softwood charcoal between Viphya and different market outlets such that retrading is impeded, would certainly exceed the institutional capacities of the Government. Thirdly, in terms of the consumers' surplus, uniform stumpage fees would be economically superior to a system of discriminatory fees. If, in addition, charcoal demand were price-elastic, government revenues from uniform stumpage rates would even exceed the aggregate revenues generated by discriminatory fees. (A more detailed discussion of these issues is provided in ANNEX I.) In view of the above described drawbacks of the discriminatory stumpagc fee approach, the Project has opted for a system of uniform fees. Uniform stumpage rates that are based on the netback value of charcoal supplied to the most distant Blantyre market will eliminate incentives to retrade, are easy to collect, and tend to maximize the social x- elfare from charcoal production at Viphya.  - 11 - The stumpage fee formula recommended for Viphya thinnings is also the key to the second question, i.e. how to value wastewood that other softwood plantations generate. Should non-Viphya thinnings be converted into charcoal, the economically most compelling comparator price for feedstock valuation would be the landed costs which Viphya charcoal has at the market outlet nearest to the resource base under consideration. In fact, as can be seen from Figure 2.1, for each potential market outlet there exists a reasonably large plantation (>2,000 ha.) which, in terms of transport costs, enjoys a comparative advantage over Viphya charcoal (but also over charcoal which could come from other non-Viphya plantations). So if at each softwood plantation the stumpage fees imposed on charcoal production are brought into line with the netback value the charcoal would have relative to the costs at which Viphya charcoal can be delivered to the nearest market outlet, the following could be achieved: Firstly, there would be no market outlet to which Viphya charcoal could not be supplied competitively. Secondly, incentives to retrade would be precluded. Thirdly, government revenues would be maximized subject to the uniformity constraint imposed on stumpage fees that would be collected from Viphya charcoal hauled to different markets. So far, the only site other than Viphya where charcoal has been produced from wastewood is the Chambe Plantation located on top of the Mulanje mountain. With Blantyre as the nearest market outlet, in early 1988 the proposed feedstock valuation scheme would have led to stumpage fees of about 50 MK/t. In early 1989, though, the netback value of Chambe wastewood would have increased to MK 120 per ton of charcoal. Thus, contrary to the stumpage value of Viphya thinnings, the changes that took place with respect to the costs at which Viphya charcoal was supplied to Blantyre would have required an upward adjustment of the stumpage fee levied on Chambe wood. Finally, there is the question of how to design a mechanism for collecting site-specific stumpage fees once a concession agreement with private charcoal producers has been struck. Basically, the choice is between five different options. Government could levy (a) a specific fee (per unit of output) (b) an ad valorem fee (sales commission) (c) a profit fee (d) a per unit-of-input fee (e) a capacity fee  - 12 - What the first four options have in common is the difficulty of precisely defining and monitoring the quantitative variable relative to which the fee is calculated. Moreover, in cases a), b) and c) the imposed fee does not encourage the efficient use of the feedstock. Therefore, the Project advisors suggested a capacity fee: If the installed capacity is "taxed", the contractor will be provided with an incentive to fully utilize the installed capacity and to achieve high charcoal yields by operating the kilns in an efficient way. The installed capacity of a production camp is a well-defined basis of measurement. Since the potential annual output of a kiln is known, the capacity fee can be set at a level at which Government captures the net back value of the charcoal which the camp would produce under normal (average) operating conditions. For instance, based on a net-back value of MK 10 per ton of charcoal and assuming that the effective capacity of a standard production site is equivalent to 1,000 tpy, the stumpage fee commitment of the production camp would simply amount to MK 10,000 per annum. 2.3 The Cost-Competitiveness of Viphya Charcoal The question whether Viphya charcoal is an economically viable source of fuel depends above all on the costs involved in alternative solid fuel supply options. In the following lines we compare Viphya charcoal with the options of using domestic/imported coal, charcoal from plantation-derived fuelwood and charcoal made from wood grown on customary land. The smali size of the proven reserves, its poor (sub-bituminous) quality and logistic difficulties restrict the use of Malawi coal to domestic markets. Its economic value therefore rests on the cost of imported coal of similar quality. Because of disruptions in the supply of comparatively cheap Moatize coal from Mozambique, for some time to come Malawi has to resort to higher cost coal from Zambia and Zimbabwe which is imported by road via Lilongwe. In mid 1989, the costs of Zambia/Zimbabwe coal delivered to Lilongwe a- junted to 90 US$/t or 225 MK/t. This compares with 205 MK/t for domestic pea size, A1 from Kaziwiziwi landed in Lilongwe (transport distance to Lilongwe 450 kim; costs ex Kaziwiziwi mine: 70 MK/t). Strictly speaking, however, Malawi coal should be charged with a depletion premimum since the Kaziwiziwi deposits are likely to be exhausted within 6 years. Thus, if a 12% discount rate is applied, the premium works out at 12.7 MK/t (in mid 1989 prices) and will rise to 25 MK/t over a period of 6 years. As a consequence, the current costs of Malawi coal cif Lilongwe would have to be adjusted to 217.7 MK/t (8.7 .K/GJ), thus exceeding the landed costs of Viphya charcoal (188 MK/t or 6.3 MK/GJ) by almost 30 MK/t (2.4 MK/GJ). Farther in the south, say, in Blantyre the cost differential may I ive somewhat smaller (see Table 2.3). But it is likely to remain large enough to compensate for the higher handling and storage costs incurred by softwood charcoal and the inconvenience of having to blend charcoal with coal in industrial applications (see Section 3.1).  - 13 - As regards the option of producing charcoal from indigenous wood supplied from customary land, it is assumed that the organizational set up would resemble the conditions that characterize traditional charcoal making in earth mound kilns (charcoal conversion effieicney: 12%, labor input: 1 man-day per 32 kg bag). Currently, the stumpage fee traditional charcoal makers would have to pay amounts to MK 4.50 per sm3 of unplanted fuelwood (cut and stacked by the purchaser) which works out at MK 12.5 per ton of fuelwood. Given the earth mound kiln's conversion efficiency and shadow pricing the labor input in terms of government minimum wages (MK 2 per man-day), the econmics of traditional charcoal making work out as follows: If traditional charcoal is supplied to Blantyre, it will outprice Viphya charcoal as long as the transport radius is less than 331 km. However, if Lilongwe is the &,arket outlet, the economic transport radius for traditional hardwood charcoal reduces to 40 km. The gazetted stumpaged fees, though, which currently apply to indigenous wood from customary land do not correctly reflect the economic value of the resource in question. According to the pricing formula used by the Forestry Department, the fees would have to rise to 20 MK/sm3 (in 1989 prices) by 1996, a figure which is supposed to indicate the costs of replacing indigenous wood with plantation grown wood. However, a more plauisible assumption is that fuelwood from small farmer woodlots (rather than plantation grown wood) will substitute for indigenous species used as a source of fuel. World Bank (1989) estimates suggest that the economic value of fuelwood grown by small farmers amount to 9.2 MK/sm3 (in 1989 prices) which is equivalent to 22 MK/t. Based on this figure, the economic transport radius for traditional charcoal delivered to Blantyre proves as low as 20 km. And in the case of Lilongwe, Viphya charcoal will outbid traditional charcoal even if the latter is produced in close vicinity to the urban center. On the other hand, if plantation grown fuelwood (e.g. Ecualyptus) is used as a feedstock for charcoal production, stumpage costs of 20 MK/sm3 (53 MK/t) have to be accounted for. The organizational set up and the cost of charcoal production will by and large be similar to the conditions prevailing at Viphya. Under these assumptions, the economic transport radius of plantation-derived charcoal works out at 110 km if Blantyre is the market outlet. In the Lilongwe market, however, Viphya charcoal will outprice plantation-derived charcoal from any location. It should be borne in mind, though, that the direct use of fuelwood supplied from peri-urban plantations competes with its use as a feedstock for charcoal production. At stumpage costs of 20 MK/sm3, the selling price of plantation grown fuelwod amounts to 65 MK/t (cut and stacked, ex-collection point). Moreover, let us assume that, based on the differential in end use efficiencies and heating values, charcoal is 2.5 times more efficient than fuelwood. Then, the economic transport radius for plantation grown fuelwood that competes with Viphya charcoal in the Blantyre market, amounts to 145 km.  - 14 - This implies that, wherever the plantation is located, it will be more economic to use the resource as fuelwod rather than as charcoal. And as fuelwood the resource will have a cost advantage over Viphya charcoal, if its production site is not farther than 145 km away from Blantyre (for Lilongwe the break-even distance is 30 km, respectively). Conceming the non-Viphya woodfuel options, the results can be summarized as follows (see Table 2.4): Table 2.4: Economic Transport Distance for Non-Viphya Woodfuel i Market Charcoal from Indigenous Wood Plantation Grown Wood Outlet Case I WJ Case II Charcoal Fuelwood Blantyre 331 20 110 145 Lilongwe 40 - 30 a/ in km, defined relative to Viphya charcoal with landed costs of 290 MK/t in Blantyre and 188 MK/t in Lilongwe. Unit transport costs: 0.35 MK/t/km. b/ stumpage fee: 4.50 MK/sm3 c/ stumpage fee: 9.20 MK/sm3 For the Lilongwe area, the Viphya plantation proves the least cost source of charcoal. The same result holds, ceteris paribus, for the central region where most of Malawi's flue- cured tobacco is produced. In the Blantyre market, Viphya charcoal has a clear cost advantage over traditional charcoal made from indigenous wood as long as the resource base of the latter is properly priced (except for the negligible case where the resource base is available in the immediate vicinity of the urban center). Also, charcoal produced from plantation grown fuelwood fails to be an economically viable alternative to Viphya charcoal. Fuelwood, however, which is cultivated on plantations not too distant from the market outlets under consideration may prove less costly than Viphya charcoal. Yet this option is constrained by the availability of peri-urban land that cannot be put into uses which are more profitable than fuelwood production.  - 15 - 2.4 Kiln Design and Conversion Efficiency The principle kiln design introduced by the Project resembles that of a Half Orange Fire Brick Kiln which is widely used in Brazil. The kilns constructed at the Viphya and Chambe Plantation have a diameter of 3.5 m and their nominal (geometrical) capacity amounts to 16 m3. Under normal operating conditions the effective capacity works out at 13.3 sm3. Apart from a steel bar which is needed as a door frame, the kiln can be made entirely from ordinary fire bricks and clay soil (see Figure 2.2), resulting in investment costs about U.S.$150 (9.4$/m3). In the two and a half years since the charcoal-making operations were initiated, more than 4,000 kiln-loads of wood have been carbonized at camps in the Viphya Forest, and this operational experience has generated sufficient empirical data to ansi er conclusively the question of how efficiently brick kilns perform under field conditions. Based on the results of the projects's production monitoring program which covered well over 2,000 kiln runs, the average weight-based conversion efficiency of Half Orange kilns constructed and used by the Project amounted to 31.2%. In this context effieciency is defined as the percentage rate which expresses the ratio of the weight of the charcoal output to the weight of the wood input. The average moisture content of the feedstock was around 20% wet basis, and the average carbonization cycle lasted approx. 70 hours (see Figure 2.3). These figures apply to pine and cyprus logs with an average diameter of approx. 20 cm and a density of 250 kg/sm3. Figure 2.2 Half Orange Fire Brick Kiln ,,ne~~~S~ C_e :  - 16 - Extensive trials were also conducted with waste material from the Viply sawmills (slabs) and practically the same results were achieved. The only major difference was in the cycle time, which was approx. 10 hours shorter because on average the slabs were much thinner than the logs. The data comp..-d by the project indicate that, given the carbonization technology, the two most important determinants of conversion efficiency are the moisture content of the feedstock and the proficiency level of the kiln operators. It can be concluded that personnel must have had between 3 and 6 months of practical experience in the utilization of the technology in order to be able to regulate the carbonization process in such a way that optimal results are achieved. Operators who have just completed their on-the-job training course achieve average conversion efficiencies that are approx. 25% lower than the levels attained by more experienced workers. The data also show that the conversion efficiency is subject to seasonal fluctuations: average charcoal yields decline slightly during the cold, rainy months of the year, Figure 2.3: Production Cycle Half Orarnge Kiln laMl -ia I 8 h' X tDSToCX P.PAtULM- igI ff _.5 ft S _ nttlStURE uIEt AS1 . - _ #aparatiaaa 7 ft carbonlast.on 6.5 h a I coatn 3_ _ I | a a IfrlaaI Coolig 4 h l= - *-- I X - no tia4 l | ; ~~~~~~~~~~~~~~* *ataF nu aajcl. 'loadiung 9ah I |1Igt$thssap 8.S h I 38 V. MOISTURE unE gASI _ -->oration la h |9 S J= L = 1MH = carbaft(At'n 1168 n z z I __~~~~~~uimmm coollnq 38 n off loaditaq S b I I I 2 I ~~~~~~~~coating 4 h I . I I.Ia wgk S ou ce 3: LP (s 69 78 8e Source: IPC (1988)  - 17 - For experimental purposes, the Project has also built two Beehive Brick Kilns (see Figure 2), with an effective capacity of 35 sm3 (compared to a nominal capacity of 46 mn3). Total investment costs amounted to US$600 (13$/m3). While the kiln's conversion efficiency was in the vicinity of 30%4, with an average carbonization cycle of 84 hours, the recovered by-products were not suited as a substitute for high- value commodities such as creosote. Further processing would have been required to upgrade the condensates to marketable by-products. However, the trials have shown that externally heated kilns equipped with simple distillation facilities are capable of recovering a significant fraction of potentially harmful emissions released during the carbonization process. To sum up: The brick kiln has proved a thoroughly appropriate carbonization technology for semi-industrial charcoal production from forestry wastes in Malawi. The average yield with brick kilns is roughly 50% higher than that obtainable with steel kilns, which, moreover, would be prohibitively expensive in Malawi (specific investment costs: 5 times as high as for the Figure 2.4: Beehive Brick Kiln STtL BeLr Beehive Fire Brick Kiln with External Heating i Smoke - r Hating gIFire Chamber IT I~~ a 4 One would expect a somewhat higher conversion efficiency but since the low figure of 30% is based on a limited number of runs undertaken with unexperienced operators, generally valid conclusions cannot be drawn from the trials conducted by the Project.  - 18 - Half Orange kiln; lifetime when in continuous operation: less than one year). Based on the experience of the Malawi Charcoal Project, it can be stated that brick kilns will invariably prove the least-cost option if (a) wage rates are low enough to justify the use of labor intensive methods; (b) production operationis must be carried out on a semi- centralized basis (e.g. within plantations) and need to be closely supervised; (c) the feedstock costs are relatively high; and (d) plans call for the establishment of a medium- to-large-scale charcoal production scheme, i.e. one whose output will be large enough over time to make it feasible to recoup the rather sizable costs incurred for technology transfer and manpower training. 2.5 Fuel Properties Primary Fuel Properties Laboratory tests by Degussa AG, Frankfurt, conducted during the pilot phase of the project to determine the fuel properties of softwood charcoal yielded the following results: Lower heating value at 5% moisture content w.b. 30 MJ/kg Fixed carbon content > 85% Volatile content < 12% Ash content 3% Bulk density 200 - 230 kg/m3 Laboratory testing of the charcoal output of the various centers continued as the Project went ahead, and the figures obtained have invariably been so close to those yielded by the original Degussa tests that the above set of values may be regarded as a quality standard for pine charcoal which can be guaranteed to users. The data derived from over two years of production experience demonstrate that, given the homogeneity of the input material, even fairly significant changes in the key aspects of the carbonization process have little if any impact on the fuel properties of the output. Whenever sizeable variations have been observed, they have almost entirely been attributable to "external" factors such as increases in the feedstock moisture content. However, even when very wet wood was used, the drop in the fixed carbon  - 19 - content never exceeded 5%. And only in cases where the charcoal samples contained fairly large amounts of clay soil and sand, were ash content values higher, but even then the maximum deviation measured was only 6%. The bulk density of run-of-the-kiln charcoal, on the other hand, is primarily a function of the properties of the specific type of raw material used. For example, older wood produces charcoal with a higher bulk density than young wood, and the utilization of cypress and pinus kesiya gives lower bulk densities than those achieved with pinus patula. The combustion trials and other utilization tests carried out by the Project have demonstrated that, in terms of specific applications, the primary fuel properties of Viphya charcoal have the following characteristics: Softwood charcoal has a lower share of volatile matt-Ir than the grades of (sub) bituminous coal available in the region. This is a disadvantage in potential industrial applications (steam boilers, cement kilns), for softwood charcoal will be more difficult to ignite than the high-volatile coals that are currently being used. On the other hand, the low ash content of Viphya charcoal is beneficial in cement production; acceptable overall ash content values were achieved when charcoal was utilized in fuel blends with local coal. The fact, however, that softwood charcoal has a lower bulk density than both coal and traditional hardwood charcoal poses a problem whenever it is used in furnaces, boilers, or kilns. If such facilities are supposed to be run at their maximum thermal output, and their fuel feeding system is designed to handle denser fuels with a higher volumetric heating value (e.g. coal), charcoal cannot provide the required energy input. This is a less serious drawback in the household sector than it is in industrial and agro-industrial applications; for when softwood charcoal is substituted for hardwood charcoal in cooking applications, users can easily compensate for its low bulk density (on average 30% lower than that of the hardwood-based charcoal) by more frequent stoking or by utilizing improved charcoal stoves with vent controls (i.e. simple gates), e.g. those stoves which are being disseminated by the Wood Energy Project. Secondary Fuel Properties While the suitability of softwood charcoal in various uses is determined above all by the primary fuel properties discussed above, its friability and hygroscopicity may matter as well. Whether or not these two secondary fuel properties will in fact imperil the competitiveness of Viphya charcoal in specific applications depends less on what happens during the charcoal production process than on what is or is not done during handling, transport and storage. Generally speaking, hygroscopicity appears to be a more serious problem than friability. When charcoal is stored in the open, it has a tendency to absorb a significant quantity of water from the ambient air, and in cases where it is exposed to rain the mcwb can rise as high as 67.5%, which not only significantly reduces its heating value but also makes the fuel considerably more difficult  - 20 - to ignite. For example, the utilization of very moist charcoal in tobacco curing leads to a substantial rise in specific fuel consumption levels since the net heating value of the fuel at 50% mcwb is only about 14.5 MJ/kg. On the other hand, if charcoal has been protected from the rain during storage, it will absorb comparatively little water froin the air, and its mcwb content will not rise much above the 15% level. Compared to both hardwood charcoal and coal, softwood charcoal is a relatively friable fuel, and, thus, if it is not handled carefully during the production process and in the course of transport operations, a substantial amount of fines can be generated. Obviously, the presence of small quantities of fines -although perhaps annoying will not pose any serious problems for most users. In ccrtain applications, though, the combustion efficiency of the fuel will be reduced if it contains a large share of fines. Such problems can be avoided by grading the charcoal prior to use. This, however, prompts two questions: a) What - if anything - can be done with the fines that are screened out? (depending on the size of the mesh used in the sieve, the fines "yield" will range between 15-50%); and b) where should grading be done? As it turns out, the two questions are interrelated in the Malawian context, and in answering the first one, one has in effect already answered the second one as well. As regards a), there is only one practicable option, namely the utilization of fines in the cement factory, which in any case must pulverize all fuel before feeding it into the kilns. However, as a potential large scale user, Portland Cement would clearly prefer to purchase fines from one large supplier rather from several small ones, and thus the establishment of a central grading facility located as near as possible to the charcoal production centers would be the logical solution. A special test conducted in December 1988 and January 1989 which was designed to determine the percentage of fines generated during transport and handling, showed that the bulk of all such material is produced "at the source", i.e. in the production process. Thus, if it is assumed that Portland Cement should sign a long-term purchase contract for fines, screening can be done at the charcoal camps. The project is currently looking into ways of increasing the efficiency of charcoal grading, which up until now has been performed with simple hand sieves. To recapitulate: Although the specific primary and secondary fuel properties of Viphya charcoal could cause certain problems for end-users in Malawi, these problems would not be pervasive enough to rule out the use of softwood charcoal in the various potential applications. Based on the results of the combustion and laboratory tests conducted by the Project, it can be claimed that the specific fuel properties would have the least serious consequences in the household sector, where charcoal would not be a new fuel (hardwood charcoal is already in widespread use there). From the standpoint of specific fuel properties, t'le second most promising candidate for a complete or partial switch to charcoal would be the tobacco industry, where only relatively minor changes in  - 21 - barn and furnace design and in curing procedures are required to create optimal conditions for charcoal use. 2.6 The Organizational Setup At an early stage of the Project a survey of the potential production sites helped devise a standard charcoal proJuction camp, taking into account the available resource base, its accessibility, the internal transportation costs, economies of scale and consolidation, and the level of managerial skills required. The result was a prototype production center with a catchment area of approximately 1,000 ha that consists of 14 Half Orange brick kilns of which two units serve as stand-by capacity. It also requires a charcoal store, three water tanks (eac'i with 2 m3 capacity), seven pairs of oxen and four ox-carts to be used for internal feedstock transport. The firm charcoal output of the camp would be 1,000 tpy, equivalent to an annual feedstock of 13,300 sm3. The proposed organizational set-up has worked well at the existing camps and it will also be utilized at a number of the new camps that are scheduled to be established. Nevertheless, a detailed analysis of potential future sites has shown that, if maximum cost- efficiency is to be achieved in charcoal production, a more flexible approach must be employed in planning and designing additional camps - one which makes adequate allowance for the specific set of topographical and other conditions found at each individual location. Given the topography of the Viphya Forest, site selection usually involves a tradeoff between the cost of feedstock transport and the production center's operating life which is a function of feedstock supply. Generally speaking, one can choose either a location which is linked by good, easily traversable roads to a catchment area of only limited capacity, or one which is surrounded by 'high-capacity" compartments which will yield an adequate volume of wastes over a longer period of time but which are accessible via less easily traversable roads that make feedstock transport more expensive. Based on the investigations that have been carried out by the Project, it has become obvious that a dual strategy should be pursued as charcoal production is expanded. In the areas where a supply of wood that will last for several years of operation (consecutive thinnings over a period of several years) is available close to a favorable production site, large-scale production centers can be established which utilize the standard operational set-up. However, in designing centers to process the wastes generated in areas where there are no favorable sites available for production facilities or where an adequate volume of feedstock is available for only a relatively short period of time, two modifications of the current configuration would be required in order to permit cost-efficient charcoal production:  - 22 - First, it would be necessary to build relatively large centers (capacity: approx. 2,000 tpy) just outside those areas in which the terrain is unsuitable for the establishment of kiln centers but which can be expected to yield a 'izeable volume of wastes over a relatively long period of time. This would involve the construction of a charcoal storage barn and the use of a tractor to transport the feedstocK. Second, it would be necessary to build smaller units directly in areas where large quantities of feedstock are available in the immediate vicinity, but only for a limited period of time. Such smaller units would also be set up in areas where the terrain is difficult to access, with oxen or tractor being used to haul the charcoal to a more easily accessible central store. Centers of this type would be built without a charcoal storage barn and oxcart transport would be used for feedstock supply. Charcoal transport to the central barns would have to be provided on a continuous basis with a tractor or oxen. When the feedstock supply is exhausted, the kilns would be dismantled and the bricks would be moved to another site and used to build other kilns. The first option would presumably be more suitable for an investor with a longer time horizon, i.e. one who is willing to accept a longer payback period and would be prepared to take a larger financial risk in order to get a higher return on his investment. The second option would probably be more attractive to risk-averse investors, who would require a shorter payback period and would be satisfied with a smaller, but certain, return on their investment. 2.7 Charcoal Production Costs In 1988, the total investment costs for a full-size production center (14 Htalf- Orange kilns, 4 ox-carts, 7 pairs of oxen, 1 charcoal stoi-. water tanks, tools) averaged NK 25,000. The experience gained over the past two y_ars has led to considerable improvements in construction effi_iency, and cost reducti, ns have been achieved by modifying certain design features of kilns, storage sheds a:ic other facilities. Assuming a lifetime of 5 years and applying a discount rate of 12%, tht annuitized investment costs work out at MK 6,700. The costs of maintaining a standard chlarcoal production camp numbered, on average, MK 5 per ton of firm output. This includes expenditures for kiln repair, replacement of tools, fodder and veterinary services tor oxen, etc. Labor costs, on the other hand, proved to be a much more critical variable. As is shown in Annex II, labor productivity has varied considerably depending on site- specific conditions, seasonal impacts, managerial skills, the level of supervision, and the mode of payment employed (government wages, piece rate system, bonus scheme). In fact, whenever close supervision was relaxed, the incentive structure of the reward system installed became the main determinant of the performance of the labor force.  - 23 - On a pro-forma basis, the system of regulated (low) wages that applies to the public sector appears to be the least cost solution. During the pilot phase when the Project was stuck with the government wage system the total labor costs per ton of charcoal should have been as low as MK 15, assuming that the labor force had performed efficiently. This, however, was not the case. The low wages the Project was obliged to pay exacerbated the major drawback associated with the incentive structure underlying a system of fixed wages, namely the difficulty to motivate unmonitored effort. As a consequence, the employed labor force underutilized the installed production capacity to the point that the effective labor costs soared beyond the 30 MK/t level. In fact, the Project's experience with the government wage system seems to confirm what the so-called "wage-productivity hypothesis" claims to prevail under the conditions found in many developing countries (see, for instance, Stiglitz, 1988): With the level of wages being sufficiently low, the workers' productivity tends to improve in response to rising wage rates. Stated differently, higher wages may help reduce the costs per effective unit of labor. In view of the incentive problems caused by the government wage scheme the Project decided to introduce a linear piece rate system that makes payment contingent on output. The results of this experiment, by and large, were positive. As long as there were no disruptions in the supply of essential inputs (e.g. feedstock, spare parts, packaging material) the effective labor costs went down to about 25 MK/t. Indeed, the piece rate system can be considered the first-best solution to the incentive problems that a fully commercialized production system run by the private sector would have to cope with. The conditions that prevailed at Viphya, however, were not in favor of the continued use of a piece rate reward system. As a government venture, the charcoal operations were exposed to financial constraints and a lack of supervision, resulting in frequent disruptions that were out of workers' control. Since output-based remunerations would have placed an unduly high burden on the labor force, the Project has developed a bonus scheme that reconciles income security provided by government minimum wages with incentives to perform efficiently. Essentially, the scheme guarantees the laborers a certain minimum payment which they receive independently of the amount of charcoal produced. This sets the workers free from the risk that their income drops below some predetermined floor. At the end of the month their salary will be "topped up" if capacity utilization exceeds some threshold level (e.g. 40%). The bonus schedule adopted by the Project is linear in output.5 5 Strictly speaking, the schedule presumes that the marginal supply of effort required to produce an additional unit of output is constant.  - 24 - In summary, it can be stated that in early 19?9 an average labor input of 10 man-days was required to process 1 ton of charcoal (assuming that there are uninterrupted operating conditions), resulting in labor costs of 12 MK/t. In addition, feedstock extraction and transport incurred labor costs of about 11 MK/t. Total production costs (excluding profits/overheads, stumpage fees and packaging costs) therefore amounted to MK 35 per ton of firm output. This is 10 MK/t higher than the average production costs recorded in 1987. Almost the entire increase can be ascribed to a decline in average labor productivity. It therefore seems that the significance of close supervision and incentives was underestimated in the original calculation, giving an unrealistically low specific labor cost figure. In May 1989, administered government wages were raised by 100%. This increased the labor costs (including the bonus payments) from 23 MK/t to 38 MK/t. However, as can be seen from Table 2.5, the total ex-camp costs of (ungraded) charcoal could be kept roughly at the level that prevailed in early 1989 due to a drop in the expenditures for charcoal packaging.  - 25 - Table 2.5: Charcoal Production Costs at Viphya. 1989 i/ early 1989 mid 1989 Capital Costs 7 7 Maintenance and Repair 5 5 Labor Costs 23 38 Total Production Costs 35 50 Stumpage Fee 10 10 Profits and Overheads 10 10 Packaging Costs 25 15 Total Ex-Camp Costs of Ungraded Charcoal 80 85 Grading Costs 5 Ex-Camp Production Costs of Lump Charcoal b/ (> 10 mm) 85 90 Ex-Camp Production Costs of Charcoal Fines cl (< 10 mm) 60 65 a/ in MK per ton of firm output j/ 80% of output c/ 20% of output  - 26 - 3. The Market Potential for Softwood Charcoal 3.1 The Industrial Market During the Project's pilot phase, a great deal of effort to delineate prospective markets for Viphya charcoal focused on industrial users, i.e. on the potential for replacing both imported and domestically mined coal with softwood charcoal. A market analysis which covered all relevant coal users in the country estimated both the current and future demand for solid fuels. (For details, see Annex III). Also, a series of combustion trials was carried out designed to provide practical evidence on the technical feasibility of substituting softwood charcoal for b3ituminous and subbituminous coal in key industrial applications. By and large, the results of the trials were encouraging. Test firings conducted in standard steam boilers (chaingrate stoker), for example, showed that acceptable steam production rates can be achieved if the units are run on fuel blends containing up to 50% charcoal (on a volumetric basis) and the charcoal size is similar to that of normal pea-size coal (10-35mm). The response to the Project's efforts in this area was positive, and various firms indicated that they would definitely consider the option of using a domestically produced alternative to coal. Indeed, David Whitehead Textiles, which is the country's single largest consumer of steam boiler coal, felt that the benefits of a partial fuel switch to charcoal (foreign exchange savings, increased security of supply) would be so substantial that it decided to actively support the project's activities by agreeing to purchase a sizeable quantity of Viphya charcoal. The results from combustion trials carried out at Portland Cement Limited, Malawi's only cement producer, were promising as well. Even though certain problems were encountered when charcoal was utilized in the company's outmoded attrition-mill kilns, Portland Cement is confident of several advantages from the use of coal-charcoal blends: The company reckons that charcoal use would not only lead to increased security of supply and a reduction of its current dependence on external sources, but also, by lowering the ash content of its fuel, improve cement quality and help eliminate ash- induced kiln operating problems. In view of its positive assessment of the potential financial benefits from charcoal utilization, Portland Cement decided to conduct a second combustion trial in order to ascertain whether the old attrition mill could be operated successfully with fuel blends containing 15-30% charcoal. The second trial demonstrated that no serious operating problems are encountered when a blend of 83% Kaziwiziwi coal and 17% softwood charcoal is used. The company's interest in the charcoal blend option - and its willingness to make the financial commitment required to assess its feasibility - was attributable above all to the fact that the commissioning of the new ball mill, which had originally been scheduled for 1987, has had to be postponed repeatedly: in fact, there is little likelihood that the attrition mill will be replaced in the near future. As  - 27 - long as the old mill continues to be in operation, cement production can only be expected to absorb around 3,400 tpy of softwood charcoal per year. Once the new ball mill unit comes on steam, Portland Cement is capable of absorbing as much as 10,000 tpy of softwood charcoal. In order to acquire more operational experience with softwood charcoal, the company recently decided to purchase fines from the project on a regular basis. Initially, 200 tons per month (2,400 tpy) will be supplied to Portland Cement. Despite the fact that all major industrial consumers of coal (both imported and locally mined) expressed interest in utilizing charcoal as a substitute fuel, it has not proved possible to develop this market (steam boiler operators alone could be expected to use some 13,000 tpy) as quickly and widely as was originally anticipated. The delays and obstacles that have been encountered are attributable, above all, to typical problems which are prompted by the introduction of a new fuel which, on account of its specific properties, cannot be readily utilized without a variety of modifications in operating and fuel handling procedures. With respect to steam boilers, the key factors are: (a) the necessity of providing dry fuel storage facilities (hygroscopicity); (b) the necessity of monitoring the combustion process more closely, which pushes up labor costs; and (c) the unavoidable reduction in the thermal capacity of the boilers when coaVcharcoal blends are used. Although the provision of dry storage facilities poses no major technical problems (it would be sufficient to expand and build roofs over the existing coal bunkers), extra expenditures of about MK 8 per ton of charcoal used would be required which the owners are not willing to undertake until other impediments to the utilization of coal/charcoal blends have been overcome. For example, the combustion trials conducted at the David Whitehead plant (with the assistance of a representative of John Thompson Ltd., the manufacturer of the boilers), showed that the ignition problems caused by the charcoal's low share of volatile matter could be solved either by utilizing the boilers' down-draught system or by creating positive furnace pressure at the fuel entry point. However, in order to implement these relatively minor changes in the normal firing procedure, the firm would have to provide additional training to boiler operating personnel, and, at least in the beginning, close supervision by the plant engineers would be required. Finally, when boilers are run on coal/charcoal blends, there is a decrease in the specific energy input due to the lower bulk density of charcoal. This drawback could only be overcome with the help of a boiler retrofit: the chaingrate gear would have to be modified so as to increase the operating speed of the travelling grates which  - 28 - are designed to handle the common types of steam coal, thus permitting an increased fuel throughput in volume terms. So far, however, David Whitehead and other standard boiler operators have been reluctant to carry out the requisite modifications (which would give rise to costs probably in the range of US$3,000-5,000 per boiler) mainly because the boiler manufacturer has not yet provided the company with a definitive technical lay out of the proposed retrofit. In other words, David Whitehead - and other standard boiler operators - cannot be convinced at this point that the investment required to implement the modifications would pay off. This issue will remain unclear until the results of an additional series of combustion trials have been evaluated. In sum it can be stated that the initial picture the Project had drawn of potential industrial charcoal users was overly optimistic. Partially successful combustion trials had stimulated the Project to reach the conclusion that the industrial market was easy to penetrate. It underestimated the extent to which risk aversion, imperfect technical information, uncertainty, and the need to undertake additional investments would impede entry into industrial fuel markets. While the prospects for charcoal use in industrial applications are certainly not as dismal as the limited marketing success thus far suggests, the Project has to accept that a number of problems have yet to be resolved before the industrial market potential can be fully developed. 3.2 The Tobacco Industry In 1985/86 the overall fuelwood consumption of the flue-cured tobacco growers amounted to 300,000 tons, of which not more than 40% was supplied from plantations. Even then, the tobacco industry found it increasingly difficult to obtain the hardwood needed to meet the remaining 60% of its annual fuel requirements (180,000 tons) at acceptable costs. It was becoming obvious that the days when an ample supply of firewood was available in the immediate vicinity of the tobacco estates were gone forever. Although a number of producers had begun to improve the efficiency of their curing barns and reduce specific fuelwood consumption with the assistance of the Tobacco Industry Energy Efficiency Project, there was no doubt that, over the medium-to-long term, the industry would have to chose between three options. It could either: (a) stop producing flue-cured tobacco altogether when the supply of cheap fuelwood was exhausted, (b) embark on a large-scale tree production program, or (c) switch from indigenous hardwood to alternative fuels such as coal, charcoal or plantation wood from government forests.  - 29 - Against this background, much hope was placed on combustion trials the Project initiated on several estates using softwood charcoal as a substitute for fuelwood. However, the results of the first trials that were conducted during the 1986/87 curing season with the help of General Farming, the country's largest tobacco grower (6,000 tpy), failed to be convincing. It was not feasible to efficiently use charcoal in traditional barn designs. Nonetheless, the findings indicated that refurnished barns might help improve the performance of charcoal. In a second trial that took place during the 1987/88 curing season, a total of 412 tons of charcoal was combusted in single-furnace barns equipped with internal chimneys, fire bars and redesigned furnaces. On average, the specific fuel consumption (SFC) amounted to 5.1 kg of charcoal per kg of made tobacco, a figure which was not as good as had been expected. At this rate charcoal firing would be twice as expensive as the use of "home-grown" fuelwood. In some cases, though, the SFC had been brought down to 2 kg/kg. Moreover, there was some evidence that improvements in tobacco quality could be achieved by replacing fuelwood with charcoal. Therefore, the Project and General Farming decided to undertake a third, large-scale test program. In this final field test which was conducted during the 1988/89 curing season, a total of 275 barn loads were cured using charcoal from softwood plantations. Specific fuel consumption ranged from 2kg/kg to 4kg/kg, and the average SFC for all estates worked out at 2.91 kg of charcoal per kg of made tobacco. A total of 28 cures were also conducted in wood-fired barns on 4 estates, and the average SFC for wood was 14.1 sm3 per ton of made tobacco (7.6 kg/kg). With the data base acquired it was also possible to quantify the impacts which various barn improvements will have on fuel consumption. As can be seen from Table 3.1, among the measure suited to curtail fuel consumption the switch from a standard double-furnace barn to a single-barn device will achieve the most pronounced fuel savings. On the other hand, the use of fans which is the most costly modification, would result in surprisingly small fuel savings. Taken together, the different retooling options would reduce fuel consumption by 60%.  - 30 - Table 3.1: Fuel Savings from Barn Improvements Type of Barn Investment Costs Charcoal Savings:/ Jmprovement (in MK of 1989) (%) Single Furnace 500 30 Welded Flue Pipes 55 15 Proper Furnace Door 105 15 Internal Chimney 360 5 Billie Barn Fan 1,750 10 FD Fan 600 5 / as a percentage of the charcoal consumption level associated with a standard, unmodified double- furnace barn, 20X40 ft, 5 tiers. Source: IPC (1989) The trials also revealed that management-related factors such as maintenance and supervision significantly affect the SFC. For instance, improper sealing or stoking procedures have resulted in efficiency losses in the range of 20%. Most importantly, the field tests have gathered compelling evidence of the potential for improvements in tobacco quality as a result of the use of charcoal. Based on 11 samples taken from both charcoal- and wood-fired barns loaded with tobacco from the same field and the same reaping, it could be demonstrated that the share of lower grade leaves was much higher in the wood-fired barns. The Project estimated the total revenues which the various grades in each of the samples would have yielded if they had been sold at the auction prices prevailing in 1987/88. The results of this comparison indicate that, on average, the returns from charcoal-fired barns would have exceeded those from wood-fired barns by 20%. There is a simple explanation for the observed improvement in cured tobacco quality: charcoal is a much more homogeneous fuel than wood. Thus, what is regarded as the optimal temperature curve for tobacco curing can be more closely approximated with charcoal than with fuelwood. In particular, excessive temperature fluctuations, which cause "sponging" - the most serious quality problem for tobacco since it results in the greatest monetary losses - can easily be avoided if charcoal is used.  - 31 - The findings established in the course of the third combustion trial significantly improve the outlook for charcoal use in the flue-cured tobacco industry. An assessment of the economics of charcoal firing for General Farming, whose expenses for firewood are indicative of energy cost levels in the flue-cured tobacco industry as a whole, suggests that - at current supply costs - charcoal may prove an attractive alternative to fuelwood. This result rests on the assumption that, on average, the landed costs of Viphya charcoal amount to 150 MK/t 6, whereas indigenous wood from customary land can be delivered at average costs of 16 MK/t. It is also presumed that the SFC is 4.3 kg/kg for wood and 2.16 kg/kg for charcoal. As is shown in Table 3.2, both the specific fuel (per MJ) and energy costs (per kg of made tobacco) will be higher with charcoal than with wood. As a consequence, wood firing would appear to be economically more attractive if the use of charcoal would fail to improve the tobacco's quality. However, with a 5% improvement in quality, charcoal would be on a par with wood, and if a 20% increase in tobacco quality could be achieved (as the test results indicate), charcoal clearly turns out to be the financially more attractive option. Table 3.2: Flue-curing with Wood and Charcoal - Indicative Figures on Costs and Revenues SFC Specific Energy al Average Auction Revenues Net (kg/kg) Fuel Costs Costs Price Tobacco of Energy (MK/GJ) (MK/kg) (MK/kg) Costs (MK/kg) Indigenous Wood 4.3 2.03 0.13 5.28 5.15 Charcoal: - Case I bi 2.16 5.02 0.34 5.28 4.94 - Case II ' 2.16 5.02 0.34 5.54 5.20 - Case III d/ 2.16 5.02 0.34 6.34 6.00 a/ per kg of made tobacco. Costs tor charcoal include . nnuity ot additional investments in fireboxes. Both charcoal and wood are used i.1 single furnace barns equipped with welded flues and internal chimneys. bJ no quality improvement. c/ 5% quality improvement. ~./ 20% quality improvement. 6 This figure is based on a ex-camp price of MK 80 per ton of graded charcoal, plus 55 MK/t for transport (160 km), plus 15 MK/t for storage and handling. Not included are profits and overheads since General Farming is assumed to produce the charcoal in a vertically integrated manner.  - 32 - To sum up, the Project has demonstrated that simple, low-cost barn modifications may cut the specific charcoal consumption in tobacco curing down to 2.1 kg/kg. Interestingly enough, with the same barn improvements the specific wood consumption can also be reduced considerably, from the current 18 sm3/t to a value as low as 8.3 sm3/t (equivalent to 4.3 kg/kg). The major breakthrough, however, was the observation that charcoal firing helps upgrade the tobacco quality. In fact, the potential competitive advantage which charcoal has vis a vis wood boils down to its superiority in terms of fuel properties. As a homogeneous, high quality fuel, softwood charcoal used for tobacco curing renders higher yields feasible, both qualitatively and quantitatively. These positive findings notwithstanding, a rapid, large-scale penetration of the tobacco industry with softwood charcoal is unlikely to take place in the immediate future. More likely, there will be a piecemeal transition to charcoal, depending on site- specific economic parameters as well as on the tobacco grower's willingness and ability to take the risks involved in a fuel switch. 3.3 The Household Sector Occasional disclaimers to the contrary, it appears that the picture of urban household charcoal consumption which emerged from estimates made in 1988 (IPC, 1988) is still an accurate one. Accordingly, it can be assumed that at present a total of 50,000 tons of hardwood charcoal is consumed per year in the four largest cities in the country (see Table 3.3). Table 3.3: Consumption of Hardwood Charcoal in Malawi 1989 tpy Bags per year Blantyre 30,000 900,000 Lilongwe 15,000 450,000 Zomba 3,000 90,000 Mzuzu 2,000 60,000 In view of the sharp increase in prices since 1987, per capita charcoal use has presumably declined somewhat from the level that prevailed in 1988, but given the current annual rate of population growth (approximately 3%), consumption may have remained roughly constant in absolute terms. Also, it can be assumed that the Forestry  - 33 - Department's campaign to reduce illegal charcoal production - which was pursued quite vigorously in 1987, but has for all practical purposes been discontinued in the meantime - has had little if any impact on consumption. The official FD statistics show that a total of 6,722 bags of charcoal were confiscated in the Southern Region during the fiscal year 1987/88, while the figure for the Central Region was only 647. In other words, considerably less than 1% of the country's "illegal" charcoal output was coalfiscated. The substantial rise in hardwood charcoal prices between 1986 and 1989 can be attributed to increases in supply costs, For average transport distances to urban markets increased, freight rates went up markedly, and the growing scarcity of feedstock resources in the vicinity of towns and cities forced charcoalers to exploit less easily accessible wood stocks. To some degree, the threat of confiscation may also have contributed to the trend for higher prices. Table 3.4 provides an overview of the hardwood and softwood charcoal prices recorded in early 1989 at different market outlets. Table 3.4: Average Charcoal Prices in Malawi (early 1989) MK/t MK/GJ Hardwood Charcoal - retail level Lilongwe 290 10.74 - retail level Blantyre 400 14.82 Softwood Charcoal - cif Lilongwe 165 5.50 - cif Blantyre 285 9.50 - retail level Lilongwe 278 9.26 - retail PTC/Oilcom Blantyre 650 21.67 Whether or not softwood charcoal will succeed in capturing a share of Malawi's sizeable household fuel market will be determined by three factors: (a) the degree of consumer acceptance; (b) the progress which is made in establishing and harnessing an efficient wholesale/retail network for the new fuel; and (c) the volume and price of hardwood charcoal supplied to the market.  - 34 - Based on the results of the marketing trials conducted so far, it can be conjectured that, with respect to the fuel properties of softwood charcoal, acceptance problems will be less serious than had originally been assumed. On the other hand, however, it is clear that as long as hardwood charcoal continues to be available at roughly the same price as is charged for the softwood variety, it will remain the first choice for the majority of household consumers. Nonetheless, the marketing trials have shown that the consumers' attitude toward the new fuel, which was highly skeptical at the outset, may slowly change in favor of pine charcoal. Although newspaper and radio advertisements have had - and will continue to have - a certain impact in terms of stimulating demand for softwood charcoal, it will obviously take more than PR campaigns to bring about a large-scale fuel switch in the urban household sector. Unless the government makes major strides in (a) implementing the proposed Revenue Collection Scheme (Wood Energy Project), and (b) strictly enforcing the existing laws regarding the utilization of the country's indigenous forests and woodlands (which, admittedly, is a formidable task), pine charcoal will only slowly penetrate the household fuel market. 3.4 Medium Term Prospects The following Table 3.5 summarizes the market potential for softwood charcoal by sector. While maximum potential demand is as high as 90,000 tpy, a market share of 15,000 tpy can be considered a realistic target over the next 5 years. In the first three market segments listed in Table 3.5, the development of effective demand between now and 1995 will be determined primarily by the prices charged for competing fuels, above all coal.  - 35 - Table 3.5: Potential Markets for Softwo )d Charcoal 1990 -1995 (tpy) Market Segment Maximum Possible Targeted Cement Production 10,000 5,000 2,500 Steam boiler operators 10,000 3,000 1,000 Small industries 5,000 2,000 5(0 Flue-cured tobacco growers 20,000 15,000 6,000 Households 45,000 30,000 5,000 Total 90,000 55,000 15,000 At present, Kaziwiziwi coal, which is mined some 100 km north of the Viphya Forest, is sold for MK 70 per ton (ex mine). As a result of the sluccessful rehabilitation programme that was recently carried out at the mine, output has once again reached a level of 36,000 tpy. However, if production is maintained at this level, the existing reserves will be exhausted within the next 5 to 7 years. It appears that the Malawi Government will continue to promote the development of other mines in the country. These new mines, however, will eventually have to compete with coal from Mozambique, where the Moatize coal mine is currently being rehabilitated. Moatize coal is most likely to become the least-cost option for industrial consumers located in the southern part of the country. In any case, the medium-term prospects for industrial charcoal demand will be coal-constrained. Of the targeted 4,000 toy, the 2,500 tpy envisaged for Portland Cement Ltd. are most likely to materialize. In the tobacco curing industr-. however, the picture is different. A fuel switch from firewood to charcoal can be ex- ected to improve the quality of the output, thus increasing revenues. On this score, softwood charcoal will not necessarily have to be the cheapest source of energy available in order to capture part of the market. So it would not come as much of a surprise, if in the mid-1990s softwood charcoal succeeded in covering a market share of about 6,000 tpy. Finally, the development of effective demand in the household market will be strongly determined by government policies in the woodfuel subsector. If the specific measures that have been recommended by the Wood Energy Project and those which are called for by the Forestry Act will gradually be implemented, and if continued marketing efforts are made to convince households of the benefits inherent to an improved  - 36 - stove/softwood charcoal package, it should be feasible to pave the way for annual sales of about 5,000 tons. 4. Transport. Packaging and Handling 4.1 Charcoal Transport In 1986 and 1987, Viphya charcoal benefited from favorable overall conditions in the road transport sector. On the one hand, with construction work on the Viply sawmill in full swing, additional back haul capacity was available in the Northern Region. On the other hand, the Kaziwiziwi coal mine was producing at a rate of only about 10,000 tpy and, thus, absorbed only a small fraction of the available transport capacity. Moreover, the private trucking firms serving the Viply project preferred to utilize their back haul capacity for charcoal transport not only because loading times were shorter than with coal, but also because they could not always be sure that there wold actually be freight to haul at the mine, which is in any case located more than 100km further north than the charcoal camps. Indeed, the project was able to negotiate freight rates with the transport firms that operate in the northern part of Malawi which were calculated on a per-ton-hauled rather than a carrying-capacity basis. In 1987, between MK 100 and MK 115 per ton were paid for charcoal hauled to Blantyre, which is 650 km from the Viphya forest. Thus, the specific costs worked out at MK 0.16/t/km, which was even lower than the rates set forth in the ADMARC tariffs generally considered to be the benchmark rates for road transport in Malawi. In 1988, as the output of the Malawi Charcoal Project increased, construction activity at the Viply plant slowed down and the output of the Kaziwiziwi mine expanded, the demand for road transport capacity in the north exceeded the available supply and, as a result, rates went up sharply. In addition, a sizeable new source of demand for road haulage services - one with substantial purchasing power - emerged, namely the aid programs which are responsible for supplying food, clothes and other necessities to the 600,000 Mozambican refugees in Malawi. As more and more of the available road transport capacity was absorbed by the organizations involved in these aid programs, the transport market, which had previously been quite stable, rapidly became supply-constrained. Freight rates rose by more than 100% in 1988. Although the aid organizations brought in a certain number of vehicles of their own to move the relief supplies, thus reducing the truck shortage somewhat, the resulting increase in capacity was far too small to significantly redress the growing imbalances. For the Malawi Charcoal Project, this meant that road transport costs now have to be calculated on a round-trip basis. Accordingly, specific transport costs (Viphya Forest - Blantyre) are currently in the range of 0.35-0.4MK/t/km.  - 37 - In view of the sizeable increases in the rates charged by road carriers, haulage by rail (which had to be ruled out in 1987 for economic reasons) can now be regarded as a potentially viable transport option for softwood charcoal. Unlike private trucking firms, though, Malawi Railways does not regularly adjust its rates in accordance with changes in market conditions. Therefore, the cost advantage which the road/rail option currently seems to enjoy (see Table 4.1) may be fallacious. Should the Project enter into negotiations with Malawi Railways, different rates might be set that could well place the rail/road option at a disadvantage. Table 4.1: Transport Costs for Softwood Charcoal 1987 and 1989 (MK/ton) Destination Distance 1987 Road/ Distance 1989 Road/ (km) Road Rail ' (Iam) Y Road Rail Lilongwe 280 50 - 280 98 - Zomba 560 95 124 560 195 160 Blantyre 630 110 132 580 200 170 a/ Road to Lilongwe and rail from Lilongwe to Zomba or Blantyre. b/ Transport distance to Blantyre reduced owing to construction of new, shorter road link. In summary, there are three major lessons to be learned from the Project's experience with charcoal transport. Firstly, the fact that Viphya charcoal had to be hauled over distances of more than 600 km did not severely impair the viability of the charcoal operations. However, since transport costs placed a heavy burden on the economic performance of the Project, considerable efforts were needed to efficiently manage the transport activities. Secondly, road haulage which was organized on the basis of contractual agreements with private carriers, proved the most reliable mode of transport. It also kept the administrative requirements at a comparatively low level. Thirdly, given the low bulk density of softwood charcoal, cost efficient road transport should resort to vehicles with a high ratio of bed-surface area to payload capacity (m2/t). In fact, the optimal solution would be the use of 7-15 t long-bedded lorries in combination with trailers. The Project, however, has to cope with trucks designed for heavy loads (20t-35t). Therefore, the policy was to bargain transport rates based on the  - 38 - proposition that the carrier would fully use the rated payload capacity (which, technically, was not feasible). Moreover, the rates were determined on a per bag basis. As a consequence, the Project, by-and-large, avoided to pay for carrying capacity (t) that, due to the low bulk density of pine charcoal, could not be utilized, while the transport companies had an incentive to haul as many bags as possible. 4.2 Charcoal Packaging During the initial stage of the Project, when the combined output of the charcoal centers was still relatively small, charcoal packaging posed no major problems. For approximately 6 months, an adequate supply of second-hand bags (polypropylene) was available at a reasonable price. By mid-1987, however, it became clear that (a) the quantity of second-hand bags that could be obtained would not be sufficient to package the growing charcoal output; (b) polypropylene packaging material was generally becoming scarce in Malawi (owing to problems encountered by the bag manufacturer, Blantyre Netting, in the importation of such material), and, as a result, increasingly expensive as well; and (c) the type of polypropylene that is commonly used for bags in Malawi is not an optimal packaging material for charcoal, especially if the fuel has to be stored outdoors: owing to its sensitivity to ultraviolet radiation, it deteriorates quickly when exposed to sunlight for an extended period of time. The maximum recycling factor for new bags was 3. Table 4.2: Development of Packaging Costs 1986-1989 Price per bag Recycling Factor Packaging Costs a/ MK MK/ton 1986 0.6 1.5 8.6 1987 1.76 3.0 15.8 1988 2.48 3.0 22.3 1989 i/ 2.80 3.0 25.2 af at 36 bags per ton kJ November/December c/ January/February  - 39 - As can be seen from the Table 4.2, charcoal packaging costs went up by approximately 200% between late 1986 and early 1989. By comparison, the overall cost of living in Malawi rose by only about 40% during the same period. It should be noted in this context that the price increase for packaging material has affected both softwood and hardwood charcoal. At present, hardwood charcoal traders are attempting to maximize the recycling factor, and thus keep outlays for packaging material to a minimum, by requiring customers to turn in an empty bag if they wish to purchase bagged hardwood charcoal. In the meantime, the project has also adopted this practice: customers who have already purchased charcoal must return their empty bags if they wish to purchase additional quantities of the fuel. W.ile a certain percentage of the returned bags are no longer serviceable and must be replaced, many can be re-used several times and it now appears that the project is able to bring average specific packaging costs for 1989 back down to roughly the 1987 level (MK 15/t). 4.3 Charcoal Handling Grading trials conducted by the Project demonstrated that charcoal fines (< 10 mm) are produced mainly during the carbonization process. Contrary to what had been expected, however, transport and handling generate only a comparatively small amount of fines: charcoal that had been screened prior to transport with a 10-mm mesh was screened again after it arrived in Lilongwe using the same size mesh. The resulting fines share was less than 10% - an acceptable value for most applications. A series of screening tests was carried out with ungraded charcoal delivered to tobacco estates in the Central Region. Screens with two different mesh sizes - 10mm and 15 mm - were used and sieving losses varied between 20% and 36%. (see Figure 4.1). Figure 4.1: Screening Test -18 !!68 128-