POLICY RESEARCH WORKING PAPER 178 8 An Economic Analysis An economic framework and computational method for of Woodfuel Management assessing pol.cy impacts or in the Sahel the cost of woodfuei supplies. and the spatial di,tribution of biomass, in a Sahelian The Case of Chad woodland setting. Kenneth M. Chomitz Charles Griffiths The World Bank Policy Research Department Environment, Infrastructure, and Agriculture Division June 1997 | POLICY RESEARCH WORKING PAPER 1788 Summary findings The woodlands in some parts of the Sahel are effectively assessing policy impacts on the cost of woodfuel supplies, an open-access resource. Under open access, fuelwood and the spatial distribution of biomass, in a particular cutters have no incentive to allow for benefits that might Sahelian woodland setting. They use spatial data on accrue if the wooded area were managed rather than standing stock and on the costs of transport to market to mined. Those benefits include sustainable streams of model a supply curve of fuel to a fuel-consuming fuelwood, fruits and other tree products, browse for location. Given an exogenously specified demand, the cattle, and ecological services such as nitrogen fixation model simulates, period by period, the extraction, and erosion prevention. To remedy this problem, some regeneration, and transport of wood fuels. It also Sahelian areas have moved to give communities effective permits easy calculation of the dvnamic cost of woodfuel control of local woodland resources. depletion. To make it easier to analyze the economic cost of such They apply the model to evaluate the benefits and supply-side interventions, Chomitz and Griffiths present ecological impacts of varlous scenarios for woodland an economic framework and computational method for management around the city of N'Djamena in Chad. This paper - a product of the Environment, Infrastructure, and Agriculture Division, Policy Research Department - is part of a larger effort in the department to understand the causes and consequences of land use change. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Anna Marie Maranon, room N10-037, telephone 202-473-9074, fax 202-522-3230, Internet address prdeiCaworldbank.org. June 1997. (26 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center AN ECONOMIC ANALYSIS OF WOODFUEL MANAGEMENT IN THIE SAHEL: THE CASE OF CHAD Kenneth M. Chomitz Charles Griffiths Environment, Infrastructure and Agriculture Division Policy Research Departnent World Bank Conunents welcome: kchomitz@worldbank.org or to fax: 202 522 3230 AN ECONOMIC ANALYSIS OF WOODFUEL MANAGEMENT IN THE SAHEL: THE CASE OF CHAD' BACKGROUND AND OVERVIEW The issue In some parts of the Sahel, woodlands are effectively an open-access resource. Under open access, fuelwood cutters have no incentive to allow for the potential future stream of benefits that a wooded area might yield if it managed rather than 'mined'. These benefits include sustainable streams of: * fuelwood, both for sale and for autoconsumption * fruits and other tree products for direct human use * browse for cattle, which translates into both meat and nutrient-rich dung * other spillovers to agriculture such as nitrogen fixation and erosion prevention * ecological services (including the intrinsic 'existence' value of the woodland ecosystem) which may be important although difficult to monetize Undervaluation leads to overexploitation. The result is that the supply price of fuelwood rises faster than would be efficient if property rights could be costlessly enforced. At the same time, the flow of other woodland services is reduced. Furthermore, open access results in appropriation of fuelwood proceeds by a relatively small number of urban transporters rather than by the rural inhabitants of the woodlands. Two sets of remedies have been proposed for this problem, and implemented in the Niger Household Energy Project (Foley et al. 1995): Supply side: Villages can be granted the exclusive right to manage and harvest woodlands in their vicinity. With secure tenure rights, it is hoped that the villagers will be able to increase yields above the regeneration rates of degraded areas. Over the long run, it is hoped that I We are very grateful to Neil Quarmby of I.S. Ltd for providing the GIS data used in this report, to Robert van der Plas and Luis Gutierrez for a variety of demand projections and supply parameters, to Djime Adoum for agricultural data, and to Emmanuel Akpa, Elias Ayuk, Amadou Cisse, Peter Dewees, Willem Floor, Axel Martin Jensen, Andrew Millington, Azedine Ouerghi, Mead Over, Alasanne Sow, Boris Utria, Max Wilton, and Roberto Zagha for help, useful discussions, and input. However the interpretations and conclusions of this report are not to be attributed to these people or to the World Bank, and the errors are ours alone. The boundaries, colors, denominations and any other information shown on maps herein do not imply, on the part of the World Bank Group, any judgement on the legal status of any territory, or any endorsement or acceptance of such boundaries. Economic Analysis of WoodJuel Management Page 2 increased supply from these close-in villages reduces the social costs associated with transporting 'mined' wood from increasingly distant supply sources. Demand side: Demand-side interventions seek to introduce higher-efficiency stoves for charcoal, fuelwood, kerosene and LPG. An underlying assumption is that while these stoves are economically attractive to householders, market failures prevent their initial development and diffusion. For instance, they may not be effectively patentable; and skeptical consumers may be unwilling to purchase them until their performance has been credibly established. Once these hurdles are overcome, society benefits in lower effective energy costs and in lower exploitation of assumedly underpriced woodland resources. Goals of this paper The proposed remedies are quite plausible on theoretical grounds. Whether they make economic sense depends on a number of poorly-understood empirical magnitudes. For instance, if the productivity gains to woodland management are low relative to the costs of establishing tenure rights, then it may be economically more attractive simply to mine the resource -- however deplorable this would be from the standpoint of sustainability. This paper examines the most important of the empirical and methodological issues involved in the economic analysis of the supply-side interventions. 1. How should the benefits of managed woodlands be assessed? We propose a simple framework for assessing project benefits which takes account of externalities, but does not require the explicit calculation of the shadow price of fuelwood. 2. What is the role and appropriate level of a wood tax? Taxes or 'user fees' have been proposed for a variety of reasons: as a mechanism for financing woodland management, as a device for discouraging production from unmanaged areas, and as a a Pigouvian tax on extemalities. We discuss these rationales and propose a methodology for calculating the Pigouvian tax. 3. What is the impact of management on fuel supply costs? The key issue is what happens to the spatial distribution of resources over time, because that is the chief determinant of the transport cost and therefore the supply cost of fuelwood and charcoal. Here, we combine a simple but powerful methodology with extraordinarily detailed information about the spatial distribution of biomass in order to simulate the evolution of supply costs under different assumptions. The methodology permits us to address a broader range of questions, including the effects of kerosene and gasoline taxes and subsidies on woodfuel production. 4.What is the impact of management on the level and spatial distribution of biomass? The methodology allows us to simulate policy impacts on the density and spatial distribution of woodlands. This information can in turn be used to assess the ecological and social impacts of woodland degradation. Economic Analysis of Woodfuel Management Page 3 As an example, we apply this methodology to evaluate scenarios for a system of managed woodlands in Chad2. ANALYTIC FRAMEWORK The analysis considers a project to stabilize the supply of woodfuels (wood and charcoal) to the city of N'Djamena. The project would place 320,000 ha of woodland under village management over a seven year period. Villages would be provided with technical assistance to delineate their woodlands, to manage them sustainably, and to produce charcoal with greater efficiency. Village production would be sold at authorized village markets, and subject to a village tax. Woodfuel production from unmanaged areas would be permitted, but subject to a larger tax upon entry to N'Djamena. Project impacts: changes in the level and distribution of net benefits The project changes the level, timing and distribution of various benefit streams. The conceptually simplest and most flexible way to assess the project is therefore to model the flows of those benefits and of their costs, with and without the project. This paper does so by using a simulation model to predict, over space and time, the production volumes and transport costs of fuelwood and charcoal; and the density of standing stock. These predictions form the basis for assessing economic welfare streams and ecological conditions. A detailed accounting of these flow and stocks permits three useful analyses listed below. (This paper concentrates on the first two, but provides the basis to undertake the third.) 1) Benefit/cost analysis: For those benefits which are monetizable, the aggregate benefit cost ratio can be computed as: (NPDVofflow of benefits with project - NPDVofflow of benefits without project)! (NPDVof costs) Ideally we would compute, for each year and each scenario, the consumer surplus associated with the enjoyment of wood-based energy and the producer surplus associated with its production. To this would be added the analogous welfare gains associated with fruit, forage, and woodland-enhanced agricultural output. The present discounted sum of these welfare flows are then compared with and without the project. Here we focus on the wood-based energy component, since it is presumed to be the largest source of benefits. We further assume that charcoal and fuelwood demand are price-inelastic 2 This analysis was motivated by the proposed Chad Household Energy Project. However, the assumptions and project structure used for the current paper may differ from those used for the project appraisal. Economic Analysis of Woodfuel Management Page 4 (up to a crossover point at which it is preferable to switch to kerosene). This simplifies the calculation of benefits. These reduce to difference between the present discounted cost of supplying N'Djamena with fuelwood and charcoal with and without the project. The project potentially reduces the transport costs which represent the bulk of fuel supply costs. 2) Distributional analysis: The project promises progressive income redistributional benefits in addition to efficiency gains. Taxes, for instance, will redistribute income from urban consumers to rural producers, and the establishment of property rights will transfer producers' surplus from transporters to villagers. An accounting of production, costs, and prices for managed vs. unmanaged areas allows us to examine these redistributions. 3) Ecological impacts: We may intrinsically value woodlands and the ecosystem they support. In principle, one could use contingent valuation survey techniques to assess the dollar value of these benefits to the Chadian population and to the world at large. In practice, such estimates would be difficult to produce. However, we may still be able to quantify (if not monetize) the services if we assume that they are related to the extent and condition of the standing stock. The fuelwood and charcoal market The project will alter the entire market for wood-based fuels in the N'Djamena area. To assess the benefits of the intervention, and to understand the likely impacts of proposed taxes, it is necessary to briefly sketch a stylized version of the operation of the market. All commercial woodfuel and charcoal demand is assumed to come from N'Djamena. This ignores the demand from small cities and towns. It assumes that rural dwellers satisfy their own energy demands from biomass sources which are largely (but perhaps not entirely) distinct from urban sources, such as crop residues. Supply side The wholesale prices of fuelwood and charcoal are determined in the N'Djamena market. Transport costs form a wedge between the urban wholesale price and the the 'village-gate' price of these commodities. Consequently the 'village-gate' prices declines with distance (or more precisely, transport cost) from N'Djaamena. There is some evidence from Chad and elsewhere that this is the case. Economic Analysis of Woodfuel Management Page S lItIX4 1 These relations are shown in figure 1. The line marked Village-gate 'wood' shows the price of 1 Price kg of wood as a function of distance from the urban market. The slope of the line reflects the transport kg. wood cost per kg-km. The line ,.1 kg. wood marked'charcoal' shows the price of the (approximately) 120 grams of charcoal which I kg of wood would yield. Note that, at the urban market, the price of 120g charcoal the charcoal is less than that of the original wood. This reflects energy losses in the conversion process. However, the farmgate 0 C F price of this quantity of Distance to urban Market charcoal falls off less rapidly than the bulkier wood from which it was derived. In other words, charcoal is a cheaper form in which to transport energy, because it contains three times as much energy per kg. The result is that between the urban market and crossover point C, woodcutters market their product in the form of wood; from point C to point F, wood is converted to charcoal; and beyond the frontier F it is not economically profitable to send wood to the urban market. The supply curve for fuel is the mirror image of the declining village-gate price. Figure 2 is a stylized illustration of the close correspondence between the spatial distribution of biomass and the shape of the supply curve. The left panel shows a city surrounded by a ring of low-biomass savanna, scrub, or agricultural areas. The forest starts at radius r0 from the city and extends indefinitely. Hence the lowest point P0 on the initial supply curve S0 of wood (as seen from the city) is the cost of transporting wood over the distance r,. The supply curve rises gently, since the area of economic exploitation increases with the square of the radius. If, after a period of exploitation, the forest recedes to distance rl, then the supply curve shifts upwards to S1. Economic Analysis of Woodfuel Management Page 6 igure 2 Woodfiael Price g~~~~~~~~~~~~~~~~~~~~~~~~~ 1 v / c~ity rO \so P,~ ~~~~~S Po~ ~ ~ ~~S Q i.igis A! An immediate consequence of this simple WAJoodfiuel model is that we would Price prefer to set up managed woodfuel markets as close to N'Djamena as possible. P* \ S If villages close to the city are able to enforce their property rights, they will reap large locational rents because their production costs will be about the same as more distant villages, but they will receive a substantially higher price. For instance, Po \ figure 3 shows that a wood D producer located in the _ closest woodland area Woodfuel.Quantity would earn a rent Woodfiiel Quantitr equivalent to P*-P., where P* is the market price and Economic Analysis of Woodfuel Management Page 7 P0 the transport and production cost. The farthest supplying village would have production and transport costs of P* and would gain no rent. However, close-in villages may find it more attractive to convert to agriculture than to produce fuelwood, if agricultural commodity prices are even more distance-sensitive than fuelwood. Demand side For convenience, but following convention, we assume that the demand for fuelwood and charcoal is highly inelastic to changes in the wholesale prices of these commodities. The arguments are, first, that wholesale prices are a small portion of total price; and second (and more dubiously), that cooking habits are culturally fixed and insensitive to energy price. However, consumers are thought to be price-sensitive in their choice among fuels. Thus relative prices and energy contents determine switchover points between fuelwood, charcoal, and kerosene. The justification and appropriate level of wood taxes Two kinds of taxes have been proposed for this kind of project. Here we examine the rationale, feasibility, and appropriate levels of those taxes in light of the foregoing market sketch. The rural market tax One purpose of a village-level tax is to finance village-level woodland management activities and the administrative costs of operating the market. The appropriate level of this component of the tax is straightforward to compute. It has been suggested that additional taxes be levied in order to stimulate and finance community-level development activities. If we assume that demand is almost perfectly inelastic, then this tax constitutes an income transfer from urban consumers to the rural villagers, and could be set almost arbitrarily. This component of the tax is also supposed to induce village solidarity. That is, villagers will have an incentive to exert social pressure to prevent neighbors from selling wood outside the official channels, because of the loss of the village's share of the tax revenue. However, an individual making a rational calculation of his personal share of this foregone village revenue would probably not be induced to interfere with a neighbor's illicit activities. This incentive is probably symbolic at best. At worst, the tax may perversely encourage village woodcutters to bypass the village market. The citygate tax Following the example of Niger, a higher tax might be imposed on charcoal and fuelwood at the gates of N'Djamena. It would have two pragmatic purposes. First, it would ensure equity between managed and unmanaged producers. If only managed areas were taxed, they would be placed at a disadvantage, and some might be priced out of the market. Second, it could finance some of the project setup costs. In addition, it could serve as a Pigouvian tax to counteract the external costs of open-access harvest. The main potential source of externality is the dynamic cost of utilization (Wiedenmann 1990): open-access exploitation shifts up the supply curve of fuelwood, increasing future costs Economic Analysis of Woodfuel Management Page 8 above what they would be in a private-property regime. Additional externalities may also come into play. If one could compute the magnitude of the external costs and dynamic costs associated with fuelwood extraction, imposition of this tax would result in the socially optimal level of extraction, as purchasers are confronted with the full social cost of the resource. (In this paper we adopt the simplifying assumption that demand is perfectly inelastic. In this case the tax would not lead to a reduction in fuelwood consumption, but the tax proceeds could in principle be used to compensate society at large for the external costs.) If Pigouvian taxes correct for the external costs of woodland depletion, why bother with setting up managed areas? Why not simply impose a tax? The answer is that the Pigouvian approach, by itself, will not lead to the introduction of improved management techniques which can potentially double yield rates per hectare. Those techniques will only be adopted if villagers feel confident in tenure rights -- which is a goal of the proposed interventions. Note also that the tax should not be viewed as an enforcement mechanism for tenure rights, as is sometimes suggested. That rationale supposes that the high citygate tax will deter transporters from removing trees without villagers' permission. This however seems unlikely. If unmanaged areas dominate supply, or are the marginal suppliers of woodfuels, then a tax on unmanaged supply will shift the supply curve up. If demand is relatively inelastic, the both the price, and the rents accruing to untaxed managed areas, will shift up almost as much as the tax. Thus the incentive for outsiders to steal wood from managed areas is little diminished. CALCULATING THE BENEFITS OF IMPROVED FUEL SUPPLY Applying the framework The most important component of hypothesized project benefits is the savings in fuel supply costs. Here we set aside (for later consideration and inclusion) the other benefits of woodlands and consider them simply as assets for producing fuel. We have two options, corresponding to the with and without project scenarios. Without the project, woodland resources are 'mined' (cut down without regard to future regeneration), starting with the areas closest to the city. As mining proceeds, the supply frontier shifts further and further from town. Since the wholesale prices of wood and charcoal consist largely of transport costs, the result is a supply curve which shifts steadily upward, and with increasing total costs of supply over time (as in figure 2). The price may eventually stabilize at a high level corresponding to a backstop technology (kerosene) or supply source (a large but distant forest which can provide many years of supply). With the project, the supply curve also shifts up initially, because managed areas are placed out of bounds for 'mining'. In the medium run, this actually increases supply costs, because marginal supply must come from farther away. In the Economic Analysis of Woodfuel Management Page 9 long run, however, the establishment of higher-yielding, sustainable supply sources stabilizes the supply curve below the long-run situation without the project. The net effect on supply costs is the difference between the net present costs of the with and without project scenario. Does this formulation take account of the full social cost of wood extraction? Recall that there are two reasons why social costs exceed private costs. First, there are spillovers, external services to agriculture. These can be reckoned separately, simply accounting for the total value of such services with and without the project. (We do not calculate them here.) Second, there are the dynamic costs of wood supply, i.e. the impact of today's extraction on tomorrow's supply. To worry about this dynamic cost is precisely to worry that overexploitation causes the supply curve to shift upwards over time, and this effect is built into the model. The simulation model Map 1: Biomass 1995 (tonslkm2) To compute the time-path of fuel o-i oo C: supply costs, we construct a 1 00-200 simple spatial simulation model of m 200-1 000 fuel supply for N'Djamena. The spatial model is apt for two reasons. First, as figure 2 illustrated, in order to draw the supply curve for fuel we need to understand the spatial distribution '~of resources, and how it evolves i : over time. Second, spatial models are the most appropriate way to represent the proj ects ecological impacts on the extent, location, and condition of habitat types. We start with 20-meter resolution data on land cover for an area of v | N'Djamna approximately 20,000 square kilometers, encompassing 7: > _ N E + TownN'Djamena's fuelwood supply Village zone (see Map 1). Applying Tl bo-d--o, d-1--no-instyon, ythe.r, -0 "P Road biomass density factors, we herei de not fLpy, nor the pwt d the World Benk Grop, angrJga tuhigs fr oco the a 1 egl -stU of n tOry.O. r en--ndoomct or so-ptono. of -uh b-di-. aggregate this unformaton into a I km grid of biomass density. For each grid cell, we estimate of the potential yield under sustainable vs. unsustainable exploitation, and the transport cost per ton to Economic Analysis of Woodfiuel Management Page 10 N'Djamena (see Map 2).3 Each year, the model satisfies exogenously-specified urban fuel demand at lowest transport cost. It starts at the closest (lowest transport cost) cell, 'mining' open access areas and harvesting from managed areas4. Tracing out the supply curve, the model exploits increasingly distant cells until total demand is satisfied. Map 2: Transport Cost Zones (50 km intervals) The model also allows for some i 0-50 road equivalent km. land transformation. Since i 50-100 road equivalent km. cultivated area in the prefecture 100-1 50 road equivalent km. l 150-200 road equivalent km. has expanded rapidly over the past l 200-250 road equivalent km. decade (see figure 4), the model l _w121 250-300 road equivalent km. ~300-350 roaLd equivalent km. converts an exogenously-specified ~~~~~......... ...area of some grassland, bushland and woodland to agriculture each year. It favors areas near N'Dj amena and areas near existing agricultural lands. It also converts wooded grassland, bushland, and woodland to managed areas at an exogenously specified rate. Supply for the following year depends on the spatial pattern of stock depletion 0 N'Djamena and regeneration this year; in 0?* Town effect, the supply curve is shifted @ Village every year. Presuming that extraction rates exceed Road regeneration rates, for instance, ~~~~~~ ~~~~~supply will come from The boundurw.r color. denomintiono end mny other roreion s.hown on mBpd prgrsivl ft areasover herein do not implytpthod"oejinoh progressovely tarteer areas over Ieq1 st*tu- f any t-ritory, or eny -ndor.en- t or *c-eptn.c cf such boundeir time. The model allows for two backstop technologies. When all woodlands in the vicinity of N'Djamena are depleted, it allows unlimited supplies of wood to be brought in from distant regions in the south and east of the country, at a high cost. (The total standing stock in Chad is estimated at 859 million tons, which if mined would serve N'Djamena for a millenium at current consumption levels). The model also allows a switchover from charcoal to kerosene when the 3On this map, N'Djamena is the cross-hatched white circle; main towns are black circles; villages are small black squares; the primary roads are shown in black, and the border with Cameroon in white. 4 Areas in Cameroon are not considered as sources of potential supply. There are however high-density stands of Sudanian woodland in Cameroon relatively close to NDjamena. cononmic Analysis of Woodfiuel Management Page II efficiency-corrected retail energy prices of the two fuels are equivalent. Figure 4 CULTIVATED AREA, CHARI-BAGUIRMI 250000 .................................................................................................................................................................................0............. 200000 inSORGHUM U MILLET 100000 U BERBERE 50000 0. m mm a, a, a, a, YEAR The present discounted social cost of meeting demand is tallied over the simulation period. This is the output of chief interest. Additional outputs include the spatial distributions of biomass over time, and the evolution of market price. It is also possible to calculate producers' surplus (locational rents) for producers in managed and unmanaged areas. [See the Appendix for more information on the data, parameters, and assumptions underlying the model.] RESULTS OF ANALYSIS Initial conditions How close in N'Djamena to sources of fuel supply? How close is the land available for conversion to managed woodlands? Table 1 shows the current distribution of land cover types Economic Analysis of Woodfuel Management Page 12 according to travel distance from N'Djamena.s 'Other' includes scrubland, inundated grassland, and settlements. The striking message of these data is that there is relatively little suitable management area close to the city. The project proposes to put about 320,000 ha under management. But within 100 km of N'Djamena there are only about 46,000 ha of acacia bushland (mean biomass, 8.5 tons/ha) and about 10,000 ha of Sudanian woodland (mean biomass, 35 tons/ha). It is not clear whether it is worthwhile managing the the 41000 ha of semi-desert wooded grassland (mean biomass, 1.7 ton/ha), but even if this area is included, it would still be necessary to go past the 125 km mark in order to meet the 320,000 ha goal. Table 1: Distribution of land cover (ha) hy travel distance from N'Djamena (km)n Distance grassland bushiand woodland agriculture other total 0-25 2243 1135 499 1430 34894 40200 26-50 8352 3113 1044 6445 96745 115700 51-75C 13499 10900 3037 17959 149604 195000 76-100 17837 31309 5576 32584 140394 227700 101-125 18628 51080 8225 34833 136634 249400 126-150 60635 296884 62882 133021 594878 1148300 Simulation results, reference runs: Assumptions for the with-project and without-project scenarios are shown in Tables 2a and 2b. Table 2a: Basic parameters: managed vs. unmanaged areas Parameter Unmanaged Managed areas areas Regeneration rates (tons/ha/yr): Semi-desert wooded 0.4 0.8 grassland Acacia bushland 0.6 1.2 Dry Sudanian woodland 1.2 2.4 Charcoaling efficiency 0.13 0.20 5Travel distance is measured along the cheapest path, where secondary road travel is assumed to be twice as costly, and off-road travel four times as costly, as travel along the primary roads. The unit is the equivalent number of kilometers along a primary road. Economic Analysis of Woodfuel Management Page 13 Table 2b: other parameters Discount rate 12% Spontaneous agricultural expansion 10,000 ha/yr, first 10 years Transport cost 86.4 FCFA/ton-km Charcoal tax FCFA 15/kg (with project scenario only) Tax collection efficiency rises from 0.10 to 0.77 over 11 years Kerosene price initially $0.43, declines to $0.3 8 after two years Under our base assumptions, the deforestation frontier widens rapidly, both with and without the project. With the project, biomass is preserved within the managed areas, but is otherwise reduced to minimal levels within a more than 200 km radius of N'Djamena. A comparison of maps 3 and 4 shows the difference. Significant biomass persists in the without-project assumption only in areas which were formerly high-density Sudanian woodland -- and this because of a debatable assumption that only one-third of the standing stock is exploited by open- access cutters. rFigures 5 and 6 show the final-year supply curves FUELWOOD SUPPLY, 2016 for fuelwood and 3 0000 .................. ..................... ............................... ............. . I.-- . ------------ charcoal, both with and 25000 -..- ......._ without the project. W.U 20000 - . ...... (These supply curves are ____ ____ .___ ____ net of tax.) The area ._ _l -. P (no project) between the curves 10000 I- P (wfth project) represents the Current X oo -_ savings in fuel supply 0 _~ - costs due to the project. ° 8 8 8 CD The simulation program r4 O OD ° 8 implicitly calculates QQUANTITY (tons) these supply curves for each year and sums the discounted value of the savings. The present discounted value (PDV) of fuel supply cost savings, with the project, is about $5.9 million over a 22 year period. For comparison, a rough estimate of the present discounted value of project costs is $3.9 million. > l o ggggg gg4g q : :: i EEi22ii! D. f: -.EEEEEEEEffiEffiiii i -i ................ s g: ....................... I K lI06 O 000000000 ESi E lil iE;g :44 n U -t00 4.x. j ::j00000000ttf X400tg 0:gg:ggggg+gggg .gggXX14t tER .. 5 a zls'iCf r.vX i:i> ...........,??;g-S^; s,,qqgEg ;gkg ...........q GE!°"!!!g ggi:Xj>;g:t n i4i. : ;. .,K.tZssE, 00jg l{ jg?9 ggggggjji ., 4 .*? 00: ;,4t0 ;0i4j i;t;ttttt0 a Cc 1 Z 0)X qa2v ;uwavuvpj3 /dnfpOO fo5 s.......sfzvuv............... W4iOtt.Zigilit. Economic Analysis of Woodfuel Management Page 15 Figure 6 Charcoal supply, 2016 120000 .- ............... . ... ... ............. ........... ................... 100000 < 100000 reservoir' backstop 80 00oo0 price G 60000 c,40000 400 .P (vth project) 20000 P (nrojeci 0 0~~~~~ tos Sensitivity analyses Since the benefits may be sensitive to assumptions about paramneters whose true value is unknown, we undertook sensitivity tests (see table) as follows: Charcoal Backstop Yield PDV Supply PDV supply PDV of supply Demand supply increase cost, W/o cost, with cost savings with Elasticity source, km under project, FCFA project, FCFA project, ___________ ~~management billion billion $ million -.25 390 50% 57.9 56.6 $2.6 -.25 390 100% 57.9 55.3 $5.3 o 390 100% 62.3 59.5 $5.9 0 580 100% 63.5 59.7 $7.9 Demandaelasticity. The benefits turn out not to be very sensitive to variations in the demand elasticity in charcoal, whiich we implemented crudely via reductions in charcoal demand as a function of estimated price rises'. Cost of backstop wood supply. Benefits are somewhat sensitive to assumptions about the cost of the 'backstop' wood supply. Chad has a great deal of wood -- what is at issue is its accessibility. We have good data on supply only up to a radius of about 250 km (primary-road-travel 6 W do not allow for the small changes in consumer surplus which result when demnand is not perfectly inelastic. Economic Analysis of Woodfuel Management Page 16 equivalent) from N'djamena, and partial coverage up to about 390 km. Hence the program will show an unrealistically steep climb in supply price after the 250 km mark, which will be cutoff at the backstop.(This is evident in figure 6). However, because these distant areas are reached only after a decade or more, the discounted savings are not terribly sensitive to the shape of the supply curve at this point. We experiment with two assumptions about backstops: a) an elastic supply of charcoal at an equivalent distance of 390 km (equivalent means that cost per ton-km is being held constant -- actual distance could be larger if more efficient long-haul transport is being used; b) an elastic supply at 580 km (this is equivalent to boat transport from the Sahr region, using our standard assumption about cost per ton km). We prefer the closer backstop assumption as it compensates for our underestimate of supply in the 250-390 km range. Regeneration rates. Regeneration rates are poorly established. Our standard assumption is that managed regeneration gives twice the per-hectare yield as unmanaged areas of the same land cover class. Reducing the yield gain from 100% to 50% reduces the savings by about half, perhaps to less than the cost of implementing the project. Unit costs of transport. For the scenarios presented here, project-related savings will be directly proportional (up to a point) to the assumed unit cost of transport. (Very high unit costs of transport might result in a switchover to kerosene rather than exploitation of the fuelwood 'reservoir'.) The project appraisal team used an estimate of 86.4 FCFA/ton-km, which is 20% lower than the mean of a small informal sample of transporters taken in February 1996. It is considerably lower than the costs used by a recent World Bank transport study, but the lower figures are reasonable in light of the size and age of the vehicles involved in charcoal transport. We have taken the 86.4 FCFA figure to represent transport costs along primary roads, and assume that costs are twice as high on secondary roads and average four times as high on tracks and off-road. Growth of agriculture. The scenarios reported above assume an exogenous growth in agricultural land of 10,000 ha/year for ten years, roughly continuing past trends. We reran the base scenario with alternative assumptions of a) no agricultural expansion; and b) expansion of 20,000 ha/year for ten years. The results were surprisingly insensitive to these assumptions. With no agricultural expansion, the savings decreased very slightly, while with more rapid agricultural expansion, they slightly increased. This is because, with more rapid displacement of the fuelwood frontier, there are greater returns to establishing close-in sources of managed supply. Efficient charcoal conversion The analysis assumes that managed areas adopt higher efficiency charcoal to wood conversion techniques. Failure of this technology to diffuse would result in lower charcoal production in managed areas, and lower savings. Dynamic costs of woodfuel extraction One advantage of the methodology presented here is that it facilitates calculation of the dynamic Economic Analysis of Woodfuel Management Page 17 costs of woodfuel depletion. This external cost is defined as: dynamic costs = a (PDV of all future fuel supply costs)/O (quantity extracted) - current unit cost of supply In other words, the current costs of supplying an extra ton of wood are less than the total cost of supply, because an extra unit of extraction today forces all future supply to travel a slightly greater distance. We compute the marginal change in the total discounted value of fuel supply costs (over the next twenty years) of a marginal increase in production this year. Under our base assumptions, this dynamic cost exceeds the current costs by about 4600 FCFA/ton. It would therefore justify a Pigouvian tax of at least that level. Distributional impacts The project has potentially significant redistributive effects. These occur because managed woodlands close to town get large locational rents (ie. producers's surplus) which we assume was formerly appropriated by the transporters. In addition, since marginal supply always comes from nonmanaged areas, the tax differential on nonmanaged production boosts the price received by the managed areas, resulting in an additional rent transfer. .ieumi XThis is illustrated in Figure 7, which is Pnice Demad schematic. The supply PRentt received bymnamged as curve (without tax) is related to the distribution ... ...... =. ............ . of supply sources by \ maiaged distance from N'Djamena. tWIthta,x) ) hl The further from the city is a plot of woodland, the higher is its delivered price of fuelwood or charcoal in the city; the supply curve consists numaed : tax 10 cm diameter) ranging from 3.0 m3/ha in the north to 5.4 in the south. This compares favorably with areas classed as wooded grassland in Chad. However, the simulation program assumes that there is no production of fuel in agricultural areas. The program takes as parameters, annual regeneration rates in tons/ha for each class of woodland (wooded grassland, bushland, Sudanian woodland, managed woodlands). We have roughly mapped these woodland classes onto different rainfall categories in order to suggest regeneration rates. These class-specific rates are applied regardless of current standing stock, subject to standing stock being less than pre-specified maxima for each class. Spontaneous agricultural conversion. According to official reports, agricultural area devoted to staple crops has been rapidly expanding in Chari-Baguirmi prefecture, which encompasses the fuelwood supply area. Area devoted to sorghum, millet and berbere increased from about 45,000 ha in 1983 to 205,000 in 1995 (see figure 4). Presumably this growth reflects in part the growth of the urban market. Economic Analysis of Woodfuel Management Page 25 As noted earlier, the IS study (Quarmby et al. 1996) appears to detect some conversion of woodlands to agricultural cultivation. If agriculture yields substantially higher returns than woodfuel production, such conversion is presumably efficient and should not be seen as posing a fuelwood 'problem' -- unless it is unsustainable and itself results in land degradation. We thought, a priori, that it was important to take account of agricultural expansion, because conversion of woodlands to agriculture will potentially affect the time path of fuel supply in both the with-project and without-project scenarios. While we lack the data to predict expansion, the simulation program allows exogenous specification of the annual number of hectares to be converted. This demand is fulfilled starting near the city by assuming that the rate of conversion within a cell is proportional to the existing extent of agriculture within the cell. This simple heuristic captures the ideas that a) current agricultural locations are proxies for areas with suitable agroclimatic conditions; b) controlling for agricultural suitability, areas near town are the most attractive candidates for conversion. The model does not allow conversion of inundated grassland, scrub, or managed woodlands. Designation of managed woodlands The program takes as exogenous the annual number of hectares to be converted to management. It does not convert agricultural areas, or those covered in scrub. (These exclusions may be too conservative.) Each year, conversion demand is satisfied starting with the closest available bushland to N'Djamena. This maximizes both the efficiency gains and capture of locational rents associated with the institution of managed woodlands. Discount rate The current flow of project benefits may be negative in the short term; it is only in the medium to long term that positive benefits are felt. Hence the social discount rate will be an important determinant of project viability. We have applied a discount rate of 12%. This can be easily varied. Demandfor fuelwood and charcoal An ideal model would include fully-specified demand function for fuelwood, charcoal, kerosene, and LPG, incorporating cross-price elasticities and stove choices. This is well beyond the scope of the current exercise. We take as exogenously specified for each year the demand for charcoal, fuelwood, kerosene, and LPG. These demands were taken from van der Plas and Gutierrez (1996). We assume completely price-inelastic demand within these categories -- except that we allow for a switchover from charcoal to kerosene if the efficiency-adjusted price of energy content is the same. (The price of kerosene over time is an input to the model.) Thus kerosene is a backstop technology, placing a ceiling on the price of charcoal and therefore on the extent of deforestation. The model fulfills all fuelwood demand as close as N'Djamena as possible before starting to fufill charcoal demand. This reflects the well-known economic dominance of fuelwood at short distances. Economic Analysis of Woodfuel Management Page 26 The model allows differential efficiency in wood-to-charcoal conversion between managed and unmanaged areas. Since achievement of higher efficiency requires higher labor costs, the high efficiency techniques is used only where either the input price is high or where the available stock of wood is limited. Neither condition obtain in unmanaged lands; imposition of a charcoal tax will not affect the production incentives in these areas. Villages in managed areas would not be influenced by a wood tax (since they self produce), but might have an incentive to use higher efficiency techniques because of limitations on their annual wood production. This incentive will depend on whether the increase in yield, multiplied by the village-gate price, exceeds the extra production costs. Villages distant from N'Djamena might not have an incentive to adopt the more efficient techniques. Backstop technologies Two backstops limit the potential rise of fuel prices. The first, mentioned above, is kerosene. Since kerosene deliver twice the effective energy per kilogram, consumers become indifferent between kerosene and charcoal when the per-kg price of the latter rises above half the price of kerosene. (This abstracts from the costs of stove acquisition, which would delay switchover. Consumer preferences for the convenience of kerosene vs. the alleged better taste of charcoal cooking will also affect the switchover point ). The second backstop is the availability of woodfuels from distant but very large 'reservoirs'. Open access Finally, the model presumes that all areas not deliberately and painstakingly placed under village management are open access. In fact, a variety of tree tenure institutions are found throughout the Sahel, so it would be worth verifying the true extent to which urban woodcutters have unimpeded access to trees and woodlands. Policy Research Working Paper Series Contact Titie Author Date for paper WPS1773 The Costs and Benefits of J. Luis Guasch June 1997 J. Troncoso Regulation: Implications for Robert W. Hahn 38606 Developing Countries WPS1774 The Demand for Base Money Valeriano F. Garcia June 1997 J. Forgues and the Sustainability of Public 39774 Debt WPS1775 Can High-Inflation Developing Martin Ravallion June 1997 P. Sader Countries Escape Absolute Poverty? 33902 WPS1776 From Prices to Incomes: Agricultural John Baffes June 1997 P. Kokila Subsidization Without Protection? Jacob Meerman 33716 WPS1777 Aid, Policies, and Growth Craig Burnside June 1997 K. Labrie David Dollar 31001 WPS1 778 How Government Policies Affect Szczepan Figiel June 1997 J.Jacobson the Relationship between Polish Tom Scott 33710 and World Wheat Prices Panos Varangis WPS1779 Water Allocation Mechanisms: Ariel Dinar June 1997 M. Rigaud Principles and Examples Mark W. Rosegrant 30344 Ruth Meinzen-Dick WPS1780 High-Level Rent-Seeking and Jacqueline Coolidge June 1997 N. Busjeet Corruption in African Regimes: Susan Rose-Ackerman 33997 Theory and Cases WPS1781 Technology Accumulation and Pier Carlo Padoan June 1997 J. Ngaine Diffusion: Is There a Regional 37947 Dimension? WPS1782 Regional Integration and the Prices L. Alan Winters June 1997 J. Ngaine of Imports: An Ernpirical Won Chang 37947 investigation WPS1783 Trade Policy Options for the Glenn W. Harrison June 1997 J. Ngaine Chilean Government: A Quantitative Thomas F. Rutherford 37947 Evaiuation David G. Tarr WPS1784 Analyzing the Sustainability of Fiscal John T. Cuddington June 1997 S. King-Watson Deficits in Developing Countries 31047 WPS1785 The Causes of Government and tne Simon Commander June 1997 E. Witte Consequences for Growth and Hamid R. Davoodi 85637 Weil-Being Une J. Lee WPS1786 The Economics of Customs Unions Constantine Michalopoulos June 1997 M. Pateffa in the Commonwealth of David Tarr 39515 Independent States Policy Research Working Paper Series Contact Title Author Date for paper WPS1787 Trading Arrangements and Diego Puga June 1997 J. Ngaine Industrial Development Anthony J. Venables 37947