Peasant charcoal production in the Peruvian Amazon: rainforest use and economic reliance

Peasant charcoal production in the Peruvian Amazon: rainforest use and economic reliance

Forest Ecology and Management 140 (2001) 39±50 Peasant charcoal production in the Peruvian Amazon: rainforest use and economic reliance Oliver T. Coo...

124KB Sizes 4 Downloads 35 Views

Forest Ecology and Management 140 (2001) 39±50

Peasant charcoal production in the Peruvian Amazon: rainforest use and economic reliance Oliver T. Coomesa,*, Graeme J. Burtb,1 a


Department of Geography, McGill University, Montreal, Que., Canada H3A 2K6 Triton Environmental Consultants Ltd., 120-13511 Commerce Parkway, Richmond, BC, Canada V6V 2L1 Received 12 March 1999; accepted 27 December 1999

Abstract Recent studies point to the promise of rain forest extraction for more sustainable rural development in Amazonia but often overlook important differences within traditional communities in terms of relative economic reliance upon speci®c forest resources. This paper reports on a study of charcoal production among forest peasants in an Amazonian river community, near Iquitos, Peru. In-depth household interviews (nˆ36) provided information on household economic activity, demographic composition, and access to land, labor and capital as well as on the nature, role and economic importance of charcoal in the household economy. Our results indicate that peasant charcoal production Ð often cast as a rapacious, wasteful use of the forest Ð can provide signi®cant cash income for forest peoples and high returns per hectare, particularly when integrated into swidden-fallow agroforestry systems, without causing notable forest destruction. Low returns to labor, however, limit prospects for peasants to prosper by charcoal production. Variations in household output of charcoal are explained by differential access to intra- and extra-household labor. Among those households most reliant upon charcoal, two subgroups are found Ð `charcoal-dependent' households and `charcoal-specialized' households Ð both of which rely on charcoal production, but for different reasons and with distinct outcomes. These two sub-groups are divided by differences in nonmarket mediated access to local land and labor. Clearly, to be successful, initiatives aimed at promoting rain forest conservation and management among `resource-reliant' households must be informed by careful attention to the underlying conditions that give rise to differential rain forest reliance. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Charcoal; Rain forest; Non-timber forest product extraction; Swidden-fallow agroforestry; Resource reliance; Amazonia

1. Introduction In the search for alternate approaches to forest conservation and management in the humid tropics, researchers are turning increasingly to the study of *

Corresponding author. Tel.: ‡1-514-398-4943; fax: ‡1-514-398-7437. E-mail addresses: [email protected] (O.T. Coomes), [email protected] (G.J. Burt). 1 Tel.: ‡1-604-279-2093; fax: ‡1-604-279-2047.

rain forest use by traditional peoples. In Amazonia, much research focuses on two broad issues: ®rst, how indigenous and folk groups use forest resources, particularly non-timber forest products (NTFPs) (see Posey and BaleÂe, 1989; Anderson, 1990a; Redford and Padoch, 1992; Nepstad and Schwartzman, 1992); and, second, the economic value of the forest to the people who live from its resources (Peters et al., 1989; Grimes et al., 1994). Generally, such studies ®nd that Amazonian peoples draw on a wide range of forest resources which together represent a signi®cant and

0378-1127/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 2 7 ( 0 0 ) 0 0 2 7 4 - 7


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

potentially sustainable contribution to the regional economy (Fearnside, 1989a). Indeed, traditional practices Ð such as agroforestry or extraction of NTFPs Ð are seen as models for more sustainable use of the forest and numerous conservation groups currently are working throughout the Amazon basin to promote such practices as alternatives to deforestation. Among some observers, however, scepticism remains over the potential of forest product extraction to provide a viable and more sustainable path for forest conservation and poverty alleviation among rain forest peoples (see Gray, 1990; Browder, 1992; Coomes, 1995; Assies, 1997; Boot, 1997; Coomes and Barham, 1997). Central to the debate over the promise of forest product extraction in Amazonia is the issue of forest reliance among traditional peoples. Speci®cally, why do certain households (or lineages, kin groups, clans) rely more heavily upon the rain forest than others for income generation? To date, this important issue has received limited attention. Typically, researchers focus either on speci®c forest user groups (e.g., rubber tappers, acËai or babasu palm fruit gatherers) who all depend upon a particular resource, or on `generalist' peasant households who practice a mix of agriculture and forest extraction. In both cases, `within group' differences are typically overlooked. Recent studies, however, among rain forest peasants in Peru ®nd that Ð although forest peasants as a group do draw on an impressive array of rain forest products Ð much heterogeneity and differential market product specialization exists at the household level (see Coomes, 1996; Coomes et al., 1996; Kvist et al., 1998; Barham et al., 1999). As such, households can differ signi®cantly in their relative reliance on the rain forest, in general, and especially with respect to speci®c forest resources. Where households depend on distinct resources (and to varying degrees), forest management and conservation initiatives aimed broadly at `traditional forest groups' or `rain forest-reliant peoples' are likely to meet Ð at best Ð with mixed results. In this paper we report on a study of one particular rain forest use Ð charcoal production among forest peasants in the Peruvian Amazon Ð as an initial contribution to research on rain forest reliance. Seen often as a singularly rapacious, wasteful use of the tropical forest and a `last resort' of the rural poor, peasant charcoal production in Amazonia has been

little studied. The production of charcoal has received only passing mention as an economic activity among Peruvian forest peasants (see Padoch et al., 1985; Hiraoka, 1986; Padoch, 1987) though several recent studies address charcoal-linkages to the CarajaÂs iron ore development in Brazil (see Fearnside, 1989b; Hall, 1989; Anderson, 1990b). Elsewhere, peasant charcoal production has received more attention (see Smith, 1985; Leach and Mearns, 1988; Dewees, 1989; ClineCole et al., 1990; Ribot, 1993, 1995, 1998; Correa PeÂrez, 1998). Recent work from Africa as well as Asia indicate that ecological impacts and economic bene®ts of charcoal production vary substantially at the household level (see Smiet, 1990; Arnold and Dewees, 1995). Our study of charcoal production in the Amazonian rain forest shares this attention to household variation and is guided by three questions: 1. what are the primary sources of wood, production techniques and economic returns in the production of charcoal? 2. what factors influence the output of charcoal at the household level? 3. why do certain households rely more on charcoal production than others for income generation? We answer these questions through a case study among forest peasant households in a traditional Amazonian community in northeastern Peru. 2. Study region Our study region is located near Iquitos, the largest city in the Peruvian Amazon. Linked only by air and water to the rest of the country, Iquitos serves as the commercial and service center of the region, and draws heavily on its hinterland for resources, particularly food, construction materials, and wood-based fuels. Charcoal is used by restaurants, food counters and urban households for brazing and barbecuing meat (i.e., principally ®sh and chicken). We estimate Ð based on weekly interviews conducted with charcoal wholesales at the main ports of Iquitos Ð that approximately 5200 kg of charcoal entered city markets each day between June and August 1995, the peak period of demand, or about 156 tonnes/month. Charcoal is produced by peasants living in communities located within one day's travel by river boat or truck

O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50


from the city, principally in four areas: along the Nanay and Momon rivers, along the Iquitos±Nauta highway, and around Tamshiyacu, a river town upstream of Iquitos. Of these sources, Tamshiyacu and neighboring communities are the most important, producing at least 50% of the charcoal destined for market. The community of San JoseÂ2 was selected for study as a principal node of charcoal production in the Iquitos region. Residents refer to themselves as riberenÄos, descendants of Iberian and Amerindian people, who employ indigenous livelihood practices oriented to both subsistence production and the market economy (see Chibnik, 1991; Hiraoka, 1992). Although households in San Jose (nˆ60) practice a wide variety of economic activities Ð from agriculture to handicraft production Ð most specialize for cash income in swidden-fallow agroforestry and forest product extraction (73%, nˆ44). Other households rely more heavily upon river ®shing (22%) and various service activities, such as shop-keeping (3%) and traditional healing (2%). Traditional technology (e.g., machetes and axes) is employed with household and communal labor for agriculture and extractive activities. Decades of forest and land use around San Jose have transformed the upland primary forest into a dense patchwork of swidden ®elds, orchards and secondary forests. Situated on an old Tertiary alluvium terrace and perched some 20±30 m above the Amazon river, the upland forest soils are deep, acidic and heavily weathered with high saturation concentrations of aluminum (see Coomes and Burt, 1997, p. 30). Typical associations of the upland primary forest would include machimango (Eschweilera spp.), pashaco (Schizolobium spp.), parinarõ (Heisteria spp.) with cumala (Iryanthera spp.), quinilla (Manilkara surinamensis) and moena (Aniba spp.); tree densities are commonly 100±112 individuals/ha (ONERN, 1976, p. 81, 83). Today, however, less than 1% of the 2300 ha of land lying within community boundaries is covered by primary forest. Unlike other communities, local geographical circumstances strongly limit access to new forest land: San Jose is bounded on one side by the Amazon, and on the other sides by competing land claims of neighboring villages and state-owned low-

lands. The community has faced an increasingly acute shortage of land since the early 1980s. Households in San JoseÂ, based on our sample (see Section 3 below), are economically poor and comprised typically of six members (range: 1±13), headed by a male (95%) who is 40-years-old (median) and has received no more than a primary education. Household incomes range from US$ 416±7096/year (median: $2083), with about 50% contributed by subsistence production and the balance from market sales of agricultural and forest products. Households hold most of their wealth in land which is claimed by usufruct and transferred through inheritance and gifts rather than by market exchange. Land holdings vary widely across households, from 0.36 to 45.60 ha, with a mean area of 10.13 ha (median: 6.37 ha) (Coomes and Burt, 1997). Wage opportunities in the community are rare and remittance transfers are few. Non-land assets include farming and ®shing implements as well as small livestock, typically worth less than $500. Despite differences in land holding and other assets, signs of socio-economic differentiation (e.g., differences in dress, house size, presence of luxury goods) are few, lending the outward impression of San Jose as an egalitarian community-all households seem relatively poor and the majority make their living from the land and the forest.

2 The precise location and name of the community are not revealed to preserve respondent anonymity.

3 The last three kilns represent 42% of total production of charcoal for the period of August 1994±1995.

3. Methods In-depth interviews were conducted during 1994 and 1995 (June±August) with 36 households in San JoseÂ, representing 82% of the local population who practice agroforestry and forest extractive activities for their primary livelihood. Approximately 2±3 days were spent with each household in formal and informal interviews. Structured questionnaires solicited information on household economic activity, demographic composition, and access to land, labor and capital. Particular attention was given during interviews to charcoal production Ð data were gathered using a formal questionnaire that focused on household wood procurement, labor input, charcoal output, and economic returns for each of the last three charcoal kilns3. More general information on production


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

techniques, technologies, and risks was gathered through informal group discussions and ®rst-hand observation of all phases of the charcoal production process. 4. Results We summarize our results by providing an overview of the process of charcoal production, the primary sources of wood used and the economic returns to charcoal production. Descriptive and multivariate statistical analyses allow us to discern among households according to levels of output and relative reliance on charcoal production. 4.1. The production process Charcoal is produced in earthmound kilns, often sited along trails leading through the upland agroforest. Producers (known locally as carboneros) ®rst prepare an kiln site (hearth) nearby a ready supply of standing timber by clearing a ¯at, circular area, 5± 7 m in diameter. Trees are felled, cut into lengths of 1.5 m, and then split and stacked around the hearth site in lotes or lots. One lot contains 1.5 m3 of wood (1 m1 m1.5 m) and kilns in San Jose each consume between 3 and 13 lots. To construct the kiln, a wooden stake is driven in the ground at the center of the site, and split wood is then piled against the stake, building up a squat cone of wood that is crowned with smaller logs. Wooden stakes are driven into the ground at the base of the wood cone and a low wall of slats or palm fronds is built around the base to support subsequent layers and to provide ventilation. Fresh palm fronds are then layered over the wood pile, followed

by a large volume of soil that is packed to form the outer crust which seals the kiln. The kiln is lit from below, by a candle inserted into the core of the mound. The producer must tend the kiln for several days, ®lling holes that develop in the kiln with wood, fronds, and dirt as the wood is progressively carbonized and the kiln slowly subsides. When complete, the kiln is extinguished by removing the surrounding fence and stakes which allows the kiln to collapse onto itself. As the kiln cools, the soil crust and palm fronds of the kiln are raked off, and the charcoal is collected in rice sacks which, when ®lled, weigh about 15 kg. The `mean kiln' in San JoseÂ, which incorporated 6.3 lots (9.45 m3) of wood, produced an average of 63 sacks (range: 30±127) or 945 kg of charcoal. If kept dry, charcoal can be stored in bags for several weeks, typically under producers' huts. Charcoal production is a time-consuming activity. The mean kiln (63 sacks) requires a labor investment of about 26 days, roughly equivalent to 13 days of work for two persons. This ®gure is derived by summing the time required for each of the seven production stages: site selection and preparation (1 day), wood preparation and stacking (7 days), kiln construction (4 days), combustion supervision (4 days), charcoal harvesting and sack ®lling (5 days), sack transport to riverbank (3 days), and a return trip to Iquitos (2 days). The average price received for charcoal sold by households in the sample between August 1994 and 1995 was $1.27/sack (range: $1.00±1.59). 4.2. Wood sources and kiln yield Households in San Jose draw on three sources of wood for charcoal production: secondary forest fallows, fruit orchards, and primary forest (Table 1).

Table 1 Wood sources for charcoal production, San Jose, Perua Forest type

Secondary forest Fruit orchards Primary forest Total a

Age of forest cut (years)

Area of Forest cut (ha)

Charcoal production








No. kilns

No. sacks

8.9 13.0 ±

3.1 5.0 ±

(73) (4) ±

0.11 0.18 0.09

0.09 0.22 0.08

(65) (4) (18)

7.71 0.72 1.56 9.99

76 4 18 98

4757 297 1161 6215

Note: Data from households surveyed in August 1995, for three most recent kilns.

O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

Generally, households prefer secondary regrowth over primary forest because wood cutting is less laborintensive, though charcoal yields per lot are lower with secondary forest species which generally have lower wood densities. For most households, charcoal production is an integral part of their swidden-fallow practices. Typically, a household will open a small ®eld in secondary forest on the upland4, cultivate the plot through two crop phases Ð a swidden phase (1±2 years) of subsistence crops (principally manioc.) and a transitional fruit phases (2±6 years) of marketable fruit crops (pineapple, cashew, guava and bananas) Ð and then let the plot revert to secondary forest which is lightly managed. In some cases, the transitional ®eld may be transformed into a longer term orchard (6‡years of umarõÂ, Pouraqueiba sericea). Mean ®eld sizes of swiddens, transitional ®elds and orchards are 0.52, 0.66, and 0.81 ha, respectively. Households possess, on average, a total of nine ®elds (six in crop and three in forest fallow, range 1±26). Households tend to reclear the older forest fallows as the need for a new ®eld or cash income arises. Older forest fallows, on the acidic and highly weathered upland soils, provide higher yields of both agricultural crops and of wood for charcoal production. Households, however, are constrained by the lack of available new land and therefore must trade off their desire to leave forest fallows for long periods with the more pressing need to meet subsistence and cash needs by cutting the secondary forest, making charcoal, and cultivating staple crops. The length of time between consecutive clearings, calculated from complete ®eld histories (nˆ98), varied from 2 to 33 years with a median duration of 10 years (Coomes and Burt, 1997). The resulting cropping sequence typically consisted of 1 year of swidden cropping, 2 years of pineapple and 7 years of secondary forest regrowth. Residents deem fallows between 8 and 12 years of age as suf®ciently old for both charcoal production and agricultural cultivation Ð somewhat longer than the average duration encountered Ð and report that the removal

of wood for charcoal production does not diminish agricultural productivity. Households with only limited agroforestry holdings will ask neighbors for `surplus' wood Ð often from less desirable species Ð in their cut fallows which is then used in charcoal production. Interestingly, one large land owner manages a portion of his holdings (8 ha) through a secondary forest rotation speci®cally for charcoal wood production Ð every year about 1 ha is cut for charcoal and then left immediately in fallow; by eliminating the agricultural phase, which draws down soil nutrients, the system reportedly ensures faster regrowth of woody biomass for future charcoal wood production. Whereas most households obtain their wood for charcoal production from secondary forest fallows, some land poorer residents must cut wood for charcoal product from primary forest. Such forest is found on the lowlands along a hinterland tributary, just beyond the community boundary. Periodic ¯ooding, poor soils, and a high water table makes the land unsuitable for agriculture. Here, households harvest trees selectively Ð usually picking the largest trees of the hardest wood Ð and carry the cut wood to the drier upland where kiln construction is feasible. The speci®c tree species used in charcoal product vary depending on the location and type of forest cut (Table 2). Charcoal production consumes only a small percentage of wood generated by ®eld clearance for swidden-fallow agriculture. Typically, households clear about 0.5±1.0 ha of secondary fallow forest to begin a new ®eld but effectively draw charcoal wood only a fraction of this area; much of the remaining fallen timber in the newly opened ®eld (which could potentially serve as charcoal wood) is consumed by ®re when the ®eld is burned prior to planting. The area (AREA, hectares) and age (AGE, years) of secondary forest cut for charcoal production are found by Ordinary Least Squared (OLS) regression analysis to predict the reported yield of kilns (YIELD, sacks/kiln). ‡ 6:54AGE YIELD ˆ 15:21 ‡ 125:94AREA   …16:18†


A related agroforestry system is practiced on a Pleistocene river terrace, situated some 3±4 km inland from the community (see Coomes and Burt, 1997), but charcoal is not produced from forest fallows on these lands nor along the floodplain of the Amazon River.




ÿ 0:27AGE2  …0:15†

where R2 is equal to 0.32, F to 10.08, p0.001, n to 69 and ( ): standard error, *: p<0.10, **: p<0.05, ***: p<0.01.


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

Table 2 Tree species used in charcoal production, San JoseÂ, Perua

4.3. Household output levels

Vernacular name

Thirty-®ve of the 36 households in the sample produced charcoal between August 1994 and 1995, sending an estimated 14,673 sacks (220,095 kg) to market in Iquitos. Considerable variation exists in annual output per producer household Ð from 40 to 1300 sacks (19,500 kg) Ð with a mean for producers of 419 sacks (6285 kg) from 6.8 kilns. Charcoal output tends to be somewhat concentrated among producers: one-quarter of households produced 55% of total output. OLS regression analyses indicate that household production of charcoal depends fundamentally upon access to labor. In developing the regression model, a variety of independent variables were selected based on microeconomic theory and ®eld observations to assess their relative contribution in explaining differences among households in charcoal output (NSACKS). The ®nal model includes the number of economically active persons (between the ages of 15 and 65) living in the household (NEA), the number of dependents (under the age of 15 or over the age of 65, NDEPEND), the number of communal work days used for charcoal production (NCOMM), the number of days of hired labor for charcoal production (NHIRE), the total area of land held (LAND), and the value of capital assets possessed by the household (ASSETS).

Boa Caspi Caimitillo Canela Moena Cetico blanco Cetico shiari Chuchuhuasi Cumala amarillo Cumala rojo HuacapuÂ

Scientific name

Haploclathira cordata Chisophyllum sp. Licaria triandra Cecropia sp. Cecropia membranaceae Heisteria pallida Osteophloe platyspermum Iryanthera sp. Minquartia guianensis/ Lindackeria paludosa Huacapurana Campsiandra laurifolia Ipururo Alchornea cactanifolia Ishtapi Jacaranda copaia Machimango amarillo Eschweilera sp. Machimango negro Eschweilera sp. Paloma micuna Alchornea triplinervia Pichirina Vismia sp. Pijuaillo Bactris sp. Quinilla Manilkara surinamensis Rifari Miconia sp. Rifarillo Terminalia oblonga Sacha Uvilla Pourouma folleata Shimbillo Inga altissima TahuarõÂ Anthodiscus peruanus/ Tabebuia incana UmarõÂ Pouraqueiba sericea Yana moena Aniba perulis

Typical environment L, PF U, PF U, PF U, SF U, SF U, PF U, PF U, PF U, PF L, PF U, SF U, SF U, PF U, PF U, PF L, PF/U, SF U, PF U, PF U, SF U, SF U, SF U, SF L, SF U, SF U, PF


U: upland; L: lowland; PF: primary forest; SF: secondary forest Sources: Scientific names from EncarnacioÂn (1983); Baluarte Vasquez (1986); Bayley et al. (1992); RuõÂz Murrieta (1993).

NSACKS ˆ ÿ15:68 ‡59:24NEA ‡ 57:01NDEPEND   …102:46†



‡ 28:12NCOMM ‡ 26:55NHIRE   For a modal plot of secondary forest (i.e., 0.1 ha in area and 9 years-old), the regression model predicts a kiln yield of 65 sacks (975 kg), only slightly greater than the observed yield. This model also shows the importance of the area cut over the age of the secondary forest. Kiln yields rise at a decreasing rate with forest age, reaching a maximum at 12 years, and predicted yields for secondary forest between 8 and 12 years of age differ by only about 7%. For the 1 year period of study, charcoal-making activities were responsible for deforestation of an estimated 20 ha of secondary regrowth (or only about 5% of the total area under crop or fallow), 1±2 ha of fruit orchard (less than 4% of area under umarõÂ), and about 4 ha of primary forest.



ÿ 8:08LAND ÿ 0:03ASSETS …6:14†


where R2 is equal to 0.51, F to 4.78, p0.005, n to 35, and ( ): standard error, *: p<0.10, **: p<0.05. The model indicates that households with larger families and, to a lesser degree, better access to hired labor as well as communal labor produce signi®cantly more charcoal. Household size appears to both `push' (consumption demand) and `pull' (access to intrahousehold labor) production to higher levels. The numbers of dependents and of economic active persons strongly predict charcoal output. In young households, consumption demand increases with the addition of young children, but as the household

O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

matures, the number of dependents falls as the number of economically active members rises, increasing the pool of available labor for charcoal production. As children reach adulthood and leave home, the household labor pool shrinks and (other things being equal) production contracts. For households at the same moment in the life cycle, those with more members tend to produce more charcoal. Access to extra-household labor also in¯uences charcoal output. Households capable of harnessing greater communal labor are able to produce more charcoal. Communal labor in the region is conducted by a task-based, invitational work party. The host household is obliged to provide nourishment during the work and to reciprocate in labor to each of the participants at some time in the future. Communal labor is used for the two most labor-intensive tasks in charcoal production Ð wood cutting and charcoal transport Ð as well as in agricultural activities, particularly in clearing new ®elds from the forest. As such, communal labor serves to focus labor power at moments when cash may be needed, conferring on certain households a `labor advantage' for charcoal over the short term but which eventually would be balanced out by labor pay back over the long run. No such obligation is carried with wage labor, and households that can afford hiring labor for charcoal production do so. Wood cutters receive the equivalent of $2.30/lot which represents a day's work. Perhaps not surprisingly, the number of days of hired labor is highly correlated with the area of land holdings (Pearson's rˆ0.42, p<0.05, nˆ35 households), indicating that better off households hired more labor for charcoal production. 4.4. Economic returns Charcoal production contributed the largest share of market income of all economic activities among sample households in San Jose (Table 3). On average, households earned about $527 from charcoal sales over the 1 year period, representing 46% of their gross market income. Wide variations, however, exist in the contribution of charcoal sales, from households that sold no charcoal to those earning as much as 95% of their income from charcoal. Interestingly, the average price received for charcoal (per sack) is positively and signi®cantly correlated with the total volume pro-


Table 3 Estimated household market income, San JoseÂ, Peru (from 1 August 1994 to 2 August 1995)a Economic activity

Market value (US$)

Household income (US$) Mean


% Range

Charcoal Agriculture NTFPsb Livestock Handicrafts Wages Remittances Total

18,983 13,598 3,348 3,265 1,348 200 166 40,908

527 378 93 91 38 6 5 1138

46 33 8 8 3 1 ± 100

(0±95) (0±77) (0±42) (0±42) (0±22) (0±7) (0±9)

a b

Nˆ36 households. Non-timber forest products.

duced (rˆ0.49, p<0.01, nˆ34 households). Larger producers tend to receive better prices for their output than small producers for several reasons: they draw their wood from older fallows which yields a somewhat higher valued (harder) charcoal; they can afford to be more selective in their use of fallow wood species; they can mobilize labor for production more readily when charcoal prices rise; and, they can afford to wait out short-term declines in prices. Returns to labor from charcoal production are modest at best, though returns to forest land are rather more generous (Table 4). For a mean kiln and an investment of 26 days of work, charcoal producers can expect a net return (at mean prices and after transportation costs) of about $2.50/day, only slightly above the local rural wage of $2.30 and well below urban wages for laborers ($4±6/day). Households with Table 4 Economic returns on charcoal production, San JoseÂ, Peru Labor returns

Annual returns Forest returns a

Mean kiln.

No. of sacks/kilna labor invested/kiln mean price/sack gross income/kiln transport costs net income/kiln return/day no. of kilns/year gross income/year net income/year no. of ha/kiln gross income/ha

63 sacks (945 kg) 26 days $1.27 $79.90 $14.52 $65.38 $2.52 6.6 $527.34 $431.51 0.1 ha $799


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

access to cash will often hire local men to cut wood for the kiln, paying the rural wage for such work and thereby freeing up their time to work in agriculture or other activities. We estimate the gross returns to secondary forest land of charcoal production to be on the order of about $800/ha. 4.5. Household reliance on charcoal production As noted above, considerable variation exists in the contribution of charcoal sales to household market income; some households appear to be highly reliant upon charcoal production whereas others are not. Generally, household reliance and output of charcoal are positively correlated (rˆ0.40, p<0.01, nˆ35 households) with ®ve of the nine most charcoal reliant households in the top 25% of charcoal producers. Contrasts across household according to reliance on charcoal, however, failed to show sharp distinctions, except with respect to the source of wood used for charcoal production. As reliance on charcoal increases, households clearly shift from drawing on wood from their forest fallows to taking wood from neighbors' fallows or from the primary forest on stateowned lowland. Nevertheless, within the most reliant

group, three of the nine most reliant households did rely on their own forest fallows. This ®nding motivated a closer examination of the top 25% most reliant households, those who earned at least 63% of gross market income from charcoal sales (nˆ9 households). By dividing the upper quartile of most reliant households according to the source of wood, two distinct sub-groups emerged: `charcoal-dependent' households (nˆ6) and `charcoal-specialized' households (nˆ3) (Table 5). Charcoal-dependent households are among the poorest and most reliant households in the sample. Unlike specialized households, they do not draw on wood for charcoal production from their own land, relying instead on state-owned primary forest or other households' forest fallows. Charcoal-dependent households are comprised of smaller, younger families with fewer economically active members. They possess considerably less land (upon which forest fallows can develop) and draw on less communal labor. Not surprisingly, their poverty in land and labor constrains charcoal as well as agricultural production. Dependent households produce signi®cantly less charcoal and have lower agricultural and total gross incomes than specialized households. Similar contrast exist between

Table 5 Mean characteristics of peasant households by reliance on charcoal production, San Jose, Peru Charcoal reliant householdsa

% Income from charcoal Charcoal output (sacks) Total income (US$) Charcoal (US$) Agriculture(US$) Livestock (US$) NTFPsb (US$) No. persons No. economically active No. dependents Age of household (years) Total landholdings (ha) Capital assets (US$) Communal labor (days) Hired labor (days) Source of woodc a

`Dependent' (nˆ6)

`Specialized' (nˆ3)

80 424 1327 1060 75 31 99 5.2 2 3.2 14.8 3.2 41 20 1.3 0

71 768 2693 1921 360 45 294 9 4.3 4.7 29.3 14.7 141 49 4 1

Households in top quartile of reliance. Non-timber forest products. c 1: own land; 0: state/other land. b

Other households (nˆ27)

35 364 2551 909 1051 254 218 5.2 2.8 2.4 24.9 11.8 175 32 2.2 0.65

O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

dependent households and the rest of the sample (i.e., excluding the specialized sub-group) in terms of income and access to land and labor. Although dependent household are much more reliant on charcoal sales, they do not produce signi®cantly more charcoal than less reliant households. Specialized households are similarly reliant on charcoal but they are signi®cantly older, larger and wealthier than dependent households (Table 5). On average, specialized households comprise twice the number of economically active members, draw on twice the amount of communal labor, and hold four times the area of land. Wood for charcoal production is drawn entirely from their own forest fallows and charcoal output is 81% higher than dependent households. Also unlike dependent households, specialized producers are more similar to the rest of the sample with respect to income levels, land holding, capital asset worth and wood sourcing, except that they are much large households (with more dependents and economically active members) and draw their income primarily from charcoal rather than agriculture (Table 5). 5. Discussion Our study of peasant charcoal production in a traditional peasant community of western Amazonia yields four principal ®ndings. First, charcoal production can be an intensive forest use activity which disturbs relatively small areas of forest, takes advantage of an abundance of woody biomass early in secondary succession, and is readily integrated into extant swidden-fallow agroforestry systems. Charcoal production calls upon only a fraction of the available wood when a secondary forest fallow is cleared, and removal of such wood does not, at least according to local farmers, reduce subsequent crop yields. Clearly households do not use more of the available wood for charcoal production because they are labor-constrained and must devote a signi®cant amount of time to the production of subsistence crops. Households that are land-constrained are hard pressed to incorporate forest fallows of suitable duration for charcoal production (or for sustainable agriculture, see Coomes and Burt, 1997; Coomes et al., 2000); they must rely on `surplus' wood from other households' fallows or go to the primary forest and selectively cut wood. As


such, charcoal production is not an important local cause of deforestation. Although we did not in this study assess the environmental effects of charcoal production, research elsewhere in tropical and temperate forests ®nds few long-term adverse effects, except on former hearth sites where signi®cant, long-term soil degradation and vegetation change have been observed (see Hosier, 1993; Mikan and Abrams, 1995, 1996; GarcõÂa-Montiel and Scatena, 1996). Second, charcoal production can be an important source of cash income for peasant households in `landconstrained' communities with ready access to an urban market. In San JoseÂ, the rise in importance of charcoal production accompanied the closing of the forest frontier around the community, and today virtually all village land has been incorporated into agroforestry production. Lack of access to new land has forced households, especially those with limited land holdings, to intensify their use of the poor upland soils, resulting in declines in yields of subsistence and market crops. In response, households have sought to unlock the potential bene®ts in forest fallows through charcoal production and the extraction of chambira palm ®ber (Astrocaryum chambira) for handicraft production, both of which are complementary to extant agroforestry practices. Charcoal production contributes signi®cant income, yielding relatively high returns per hectare of forest, but only modest returns on labor investment. As such, the turn to charcoal (and handicrafts) re¯ects a lack of local economic opportunities more than it suggests a promising alternate path for forest conservation and rural poverty alleviation. Third, the relative output of charcoal by households is determined primarily by access to intra- and extrahousehold labor. Households with larger families and who can draw on more communal labor and/or hired help produce signi®cantly more charcoal. Interestingly, access to extra-household labor is related to family size, so that small, young households suffer both from a lack of `in-house' labor as well as limited opportunities to hire or engage much formal communal labor; for this reason, they tend to produce less charcoal. Our ®nding of the importance of access to labor is perhaps not surprising, given that charcoal production is such a forest intensive land use and does not require much capital investment Ð labor clearly is the key input.


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

Fourth, households vary markedly in their relative reliance on charcoal for income, and two distinct subgroups exist among the most reliant households of San Jose Рcharcoal-dependent producers who are land (and labor) poor and charcoal-specialized producers who are relatively rich in labor (and land). The ®nding of two groups that both rely heavily upon charcoal production was quite unexpected. The extant literature on forest resource reliance points to the importance of the forest as a `shock absorber' whereby the poor are forced into the forest to draw on the `subsidy from nature' (or from neighbors) for their survival (see Hecht et al., 1988; Olson, 1988; Gunatilake et al., 1993; Hedge et al., 1996). Charcoal-dependent households clearly fall within this group. Without the land to invest in long cycle swidden-fallow agroforestry, which incorporates more lucrative fruit crops and more mature forest fallows, dependent households must focus agricultural production on subsistence crops. For these households, options for securing cash income are highly restricted Ð all rely upon charcoal made from wood extracted from primary forest on state land or secondary fallow forest on neighbor's land. Importantly, although poor households depend very heavily on charcoal sales, they do not produce signi®cantly more charcoal than other households; their poverty in labor actually limits their ability to produce large quantities of charcoal. Put more broadly, our case suggests that the `poorest of the poor' do not, as often argued in the literature, put the greatest pressure on the forest Ð they are simply too poor to do so. Within the charcoal-reliant group, however, we ®nd a second sub-group of households which puts considerably more pressure on forest resources and includes the largest producers of charcoal in the community. These charcoal-specialized households do not ®t the classic `reliant household' portrayal because, although they are `poor' by most standards and highly reliant on charcoal for income, they are signi®cantly wealthier than charcoal-dependent households and thus behave economically in a quite different manner. Specialized producers have considerably more land and especially labor at their disposal and choose to specialize in charcoal over other activities such as agriculture or handicraft production. They do so for three reasons: (1) their larger families have greater need for cash income at key moments when

charcoal sales can better provide income than agriculture; (2) alternate opportunities for labor are minimal because of the lack of a ®rm local labor market (i.e., beyond periodic hiring for charcoal production by the larger producers), a lack of new land that can be brought into production, the low marginal returns to intensi®cation of swidden-fallow agriculture, and the low absolute returns from handicraft production or alternate economic activity; and, (3) they hold large areas of `unused' forest fallows that can supply wood for charcoal production but for which potential agricultural yields are low. In a sense, charcoal production `soaks up' excess labor in these households and provides cash income which can be used, in part, to hire others when household labor must be diverted to agricultural production. Specialized households, thus, can clear more forest and produce considerably more charcoal than dependent households. 6. Conclusions Our ®ndings indicate the need for researchers and conservationists to look beyond classical portrayals of forest peasants toward an improved understanding of how peasant households use forest resources differentially and who speci®cally makes up the `resourcereliant'. This case study from northeastern Peru found that traditional peasants, when faced with limited access to new land, turned to charcoal production. Most households successfully incorporate charcoal production into their swidden-fallow agroforestry systems; others rely upon wood from state land or neighbors; and one farmer developed a charcoaloriented forest rotation on his land. Charcoal provides residents with signi®cant cash income from the forest without causing signi®cant deforestation-what many governmental and non-governmental organizations seek today in their efforts to promote poverty alleviation and forest conservation. Low returns to labor on charcoal production, however, promise households only continued subsistence, not prosperity. Among the group of households who rely most on charcoal production from the forest, we ®nd two distinct subgroups Ð `charcoal- dependent' households and `charcoal-specialized' households Ð which both focus on charcoal production but for very different reasons and with distinct outcomes. Such differences

O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

are particularly important in the context of `grassroots' efforts underway to promote more sustainable rain forest use among the peasantry. Initiatives aimed at the `forest reliant' without considering potential differences within this group are likely to be frustrated by the differential ability (and thus willingness) of households to participate fully in jointly furthering conservation and management aims. Clearly much closer study is needed to assess the speci®c constraints and opportunities faced by forest peasants and how household level differences in non-market mediated access to land and labor in¯uence prospects for productive use of rain forest resources. Acknowledgements The authors wish to thank the people of northeastern Peru for their kindness and hospitality throughout our stays. We are particularly grateful to the residents of San Jose who accommodated our persistent questions with patience, kindness and good humor. We also owe a debt of gratitude to Jomber Chota Inuma and Carlos Rengifo Upiachihua for their invaluable ®eld assistance. We thank Alicia Marhuenda Domenech for her assistance in Iquitos with bibliographic research on local tree species taxonomy. Brad Barham and two anonymous reviewers provided most helpful comments on an earlier draft of this paper. This research was supported by grants to the senior author from the Social Sciences and Humanities Research Council of Canada and the Fonds pour la Formation de Chercheurs et l'Aide aÁ la Recherche. References Anderson, A.B., 1990a. Alternatives to Deforestation: Steps Towards Sustainable Use of the Amazon Rain Forest. Columbia University Press, New York. Anderson, A.B., 1990b. Smokestacks in the rainforest: industrial development and deforestation in the Amazon basin. World Development 18 (9), 1191±1205. Arnold, J.E.M., Dewees, P.A., 1995. Tree Management in Farmer Strategies: Responses to Agricultural Intensification. Oxford University Press, New York, 292 pp. Assies, W., 1997. Going nuts for the rainforest. Non-Timber Forest Products, Forest Conservation and Sustainability in Amazonia. Thela Publishers, Amsterdam, 96 pp. Baluarte Vasquez, J.R., 1986. IdentificacioÂn de 20 Especies Forestales del Bosque de Jenaro Herrera, SeguÂn las Caracter-


Âõsticas MicroscoÂpicas. Faculty of Forestry, Universidad Nacional de la Amazonia Peruana, Iquitos, 176 pp. Barham, B.L., Coomes, O.T., Takasaki, Y., 1999. Rain forest livelihoods: income generation, household wealth and forest use. Unasylva 50 (198), 34±42. Bayley, P., Vasquez, P., Ghersi, F., 1992. Estudio de Impacto Ambiental del Proyecto Pacaya Samiria. The Nature Conservancy, Iquitos, 105 pp. Boot, R.G.A., 1997. Extraction of non-timber forest products from tropical rain forests. Does diversity come at a price. Netherlands J. Agric. Sci. 45, 439±450. Browder, J.O., 1992. The limits of extractivism. BioScience 42 (3), 174±182. Chibnik, M., 1991. Quasi-ethnic groups in Amazonia. Ethnology 30 (2), 167±182. Cline-Cole, R.A., Main, H.A.C., Nichol, J.E., 1990. On fuelwood consumption, population dynamics and deforestation in Africa. World Development 18, 513±527. Coomes, O.T., 1995. A century of rain forest use in western Amazonia: lessons for extraction-based conservation of tropical forest resources. For. Conserv. Hist. 39 (3), 108±120. Coomes, O.T., 1996. Income formation among Amazonian peasant households in northeastern Peru: empirical observations and implications for market-oriented conservation. Yearbook, Conf. Latin Americanist Geographers 22, 51±64. Coomes, O.T., Barham, B.L., Craig, B., 1996. Uso de recursos por los riberenÄos en la Reserva Nacional Pacaya-Samiria: Datos de una encuesta reciente e implicaciones para el manejo del aÂrea protegida, Espacio y Desarrollo, No. 8 (Centro de InvestigacioÂn en GeografõÂa Aplicada, Pontificia Universidad CatoÂlica del Peru, Lima), pp. 5±32. Coomes, O.T., Barham, B.L., 1997. Rain forest extraction and conservation in Amazonia. Geogr. J. 163 (2), 180±188. Coomes, O.T., Burt, G..J., 1997. Indigenous market-oriented agroforestry: dissecting local diversity in western Amazonia. Agroforestry Syst. 37, 27±44. Coomes, O.T., Grimard, F., Burt, G.J., 2000. Tropical forests and shifting cultivation: secondary forest fallow dynamics among traditional farmers of the Peruvian Amazon. Ecological Economics, 32, 109±124. Correa PeÂrez, G., 1998. Uso de la lenÄa y carboÂn vegetal como energeÂtico y magnitud de la reforestacioÂn. Revista GeograÂfica 124, 173±197. Dewees, P.A., 1989. The woodfuel crisis reconsidered: observations on the dynamics of abundance and scarcity. World Development 17, 1159±1172. EncarnacioÂn, F., 1983. Nomenclatura de las Especies Forestales Comunes en el Peru. Ministerio de Agricultura, Lima, 149 pp. Fearnside, P.M., 1989a. Extractive reserves in Brazilian Amazonia: an opportunity to maintain tropical forest under sustainable use. BioScience 39 (6), 387±393. Fearnside, P.M., 1989b. The charcoal of CarajaÂs: a threat to the forests of Brazil's eastern Amazon region. Ambio 18 (2), 141± 143. GarcõÂa-Montiel, D.C., Scatena, F.N., 1996. The effect of human activity on the structure and composition of a tropical forest in Puerto Rico. For. Ecol. Manage. 63, 57±78.


O.T. Coomes, G.J. Burt / Forest Ecology and Management 140 (2001) 39±50

Gray, A., 1990. Indigenous peoples and the marketing of the rainforest. Ecologist 20 (6), 223±237. Grimes, A., Loomis, S., Jahnige, P., Burnham, M., Onthank, K., AlarcoÂn, R., Palacios Cuenca, W., CeroÂn MartõÂnez, C., Neill, D., Balick, M., Bennett, B., Mendelsohn, R., 1994. Valuing the rain forest: the economic value of nontimber forest products in Ecuador. Ambio 23 (7), 405±410. Gunatilake, H.M., Senaratne, D.M.A.H., Abeygunawardena, P., 1993. Role of non-timber forest products in the economy of peripheral communities of Knuckles National Wilderness area of Sri Lanka: a farming systems approach. Econ. Bot. 47, 275±281. Hall, A., 1989. Developing Amazonia: Deforestation and Social Conflict in Brazil's CarajaÂs Programme. Manchester University Press, New York, 295 pp. Hecht, S.B., Anderson, A.B., May, P., 1988. The subsidy from nature: Shifting cultivation, successional palm forests, and rural development. Human Organization 47 (1), 25±35. Hedge, R., Suryaprakash, S., Achoth, L., Bawa, K.S., 1996. Extraction of non-timber forest products in the forests of Biligiri Rangan Hills, India. 1. Contribution to rural income. Econ. Bot. 50, 243±251. Hiraoka, M., 1986. Zonation of mestizo riverine farming systems in northeast Peru. Natl Geographic Res. 2 (3), 354±371. Hiraoka, M., 1992. Caboclo and riberenÄo resource management in Amazonia: a review. In: Redford, K.H., Padoch, C. (Eds.), Conservation of Neotropical Forests: Working from Traditional Resource Use. Columbia University Press, New York, pp. 134± 157. Hosier, R.H., 1993. Charcoal production and environmental degradation: environmental history, selective harvesting and post-harvest management. Energy Policy 21, 491±509. Kvist, L.P., Gram, S., CaÂcares, A., Ore, I., 1998. Socio-economy of villagers in the Peruvian Amazon with a particular focus at extraction: a comparison of seven floodplain communities along the lower Ucayali and Maranon rivers. (mimeo). Leach, G., Mearns, R., 1988. Beyond the Woodfuel Crisis. Earthscan Publications, London, 309 pp. Mikan, C.J., Abrams, M.D., 1995. Altered forest composition and soil properties of historic charcoal hearths in southeastern Pennsylvania. Can. J. For. Res. 25, 687±696. Mikan, C.J., Abrams, M.D., 1996. Mechanisms inhibiting the forest development of historic charcoal hearths in southeastern Pennsylvania. Can. J. For. Res. 26, 1893±1898.

Nepstad, D.C., Schwartzman, S. (Eds.), 1992. Non-Timber Products from Tropical Forests. Evaluation of a Conservation and Development Strategy. Advances in Economic Botany 9, New York Botanical Garden, The Bronx, 164 pp. Olson, S., 1988. Environments as shock absorbers, examples from Madagascar. Environ. Rev. 12 (4), 61±80. ONERN, 1976. Inventario, EvaluacioÂn e IntegracioÂn de los Recursos Naturales de la Zona Iquitos, Nauta, Requena y Colonia Angamos. Oficina Nacional de EvaluacioÂn de Recursos Naturales, Lima. Padoch, C., 1987. The economic importance and marketing of forest and fallow products in the Iquitos region. In: Denevan, W.M., Padoch, C. (Eds.), Swidden-Fallow Agroforestry in the Peruvian Amazon. Advances in Economic Botany, Vol. 5. The New York Botanical Garden, The Bronx, pp. 74±89. Padoch, C., Chota Inuma, J., de Jong, W., Unruh, J., 1985. Amazonian agroforestry: a market-oriented system in Peru. Agroforestry Syst. 3 (1), 47±58. Peters, C.M., Gentry, A.H., Mendelsohn, R.O., 1989. Valuation of an Amazonian rainforest. Nature 339, 655±656. Posey, D.A., BaleÂe, W. (Eds.), 1989. Resource Management in Amazonia: Indigenous and Folk Strategies. Advances in Economic Botany 7, New York Botanical Garden, The Bronx, 287 pp. Redford, K.H., Padoch, C. (Eds.), 1992. Conservation of Neotropical Forests: Working from Traditional Resource Use. Columbia University Press, New York, 475 pp. Ribot, J.C., 1993. Forestry policy and charcoal production in Senegal. Energy Policy 21, 559±585. Ribot, J.C., 1995. From exclusion to participation: turning Senegal's forestry policy around. World Development 23 (9), 1587±1599. Ribot, J.C., 1998. Theorizing access: forest profits along Senegal's charcoal commodity chain. Development Change 29, 307± 341. RuõÂz Murrieta, J., 1993. Alimentos del Bosque AmazoÂnico: Una Alternativa Para la ProteccioÂn de los Bosques Tropicales. UNESCO/ORCYT, Montevideo, 226 pp. Smiet, A.C., 1990. Agro-forestry and fuel-wood in Java. Environ. Conserv. 17, 235±238. Smith, A.E., 1985. An analytical approach to the economics of small-scale charcoal production in developing countries. Trop. Sci. 25, 29±39.