social forestry reconsidered - efd -...

Hyde and Köhlin Social Forestry Reconsidered

285

Social Forestry Reconsidered

William F. Hyde and Gunnar Köhlin

Hyde, W.F. & Köhlin, G. 2000. Social forestry reconsidered. Silva Fennica 34(3): 285–314.

This paper reviews the expectations for forestry’s contribution to rural development –and for its special contributions to the most disadvantaged, to women and the landlessusers of the forest commons. A growing literature challenges some of these expecta-tions; in particular, certain expectations about cultural differences and physical stocks asexplanatory factors for patterns of household behavior. This literature could also be usedto support a call for sharper definitions of deforestation, improved indicators of theeffects of forest resources on the rural poor, and improved design of forest policyinterventions. Our paper reviews the literature, suggests some unifying themes, andidentifies the critical issues that remain unanswered.

The primary contention arising from this literature is that households follow system-atic patterns of economic behavior in their consumption and production of forestresources, and that policy interventions in social forestry should be analyzed with regardto markets, policies, and institutions. Markets for forest resources generally exist insome form – although they may be thin. Successful forestry projects and policies requirecareful identification of the target populations and careful estimation of market andmarket-related effects on the household behavior of these populations. Institutionalstructures that assure secure rights for scarce forest resources are uniquely important in aforest enviornment often characterized by open access resources and weak governmentadministration. Social and community forestry, improved stoves, improved strains ofmulti-purpose trees, and even private commercial forest operations can all improve localwelfare, but only where scarcity is correctly identified and the appropriate institutionsare in place. An increasing number of observations of afforestation from developingcountries around the world is evidence that forestry activities do satisfy these conditionsin selective important cases. The critical point for policy is to identify the characteristicsof these successful cases that are predictive of other cases where new forestry activitiescan be welfare enhancing.

Keywords community forestry, deforestation, developing countries, forest resources,rural development, welfareAuthors' addresses Centre for International Forestry Research, Bogor, Indonesia; GöteborgUniversity, Göteborg, SwedenE-mail [email protected] 13 October 1999 Accepted 28 June 2000

Silva Fennica 34(3) review articles


286

1 Introduction

Social forestry generally refers to the range ofactivities associated with forest products, the ru-ral environment, and subsistence agriculturalcommunities. It often features external develop-ment assistance intended to benefit these com-munities and their environments. Over the pastten years, the prevailing wisdom on social for-estry as a welfare enhancing technology has runthe full course from optimism, even enthusiasm,to a current attitude of caution, and even scepti-cism. Meanwhile, a developing body of analyti-cal literature is beginning to provide evidence.This literature consistently supports the optimis-tic view – but only under carefully selectiveconditions.

This is a good time to examine the evidence,to identify reasonable hypotheses about the suc-cessful uses and limits of social forestry, and toidentify the important questions that are yet to beaddressed. This is our objective.

Our fundamental contention is that successfulsocial forestry activities must be assessed in termsof their contributions to human welfare. Thiscontrasts with a view that forest cover, thereforeforest protection and afforestation, is a usefulend in itself. The latter view bolsters policy deci-sions to halt deforestation or policy objectives tomaintain a fixed share of land in forest cover,“one third of the land” in the case of India (Raoand Srivastava 1992) or all land with greaterthan an eighteen percent slope in the Philippines.

Physical standards are poor measures of hu-man welfare. Welfare, including the welfare gainsdue to social forestry activities, is revealed byhousehold preferences in the context of localmarkets, institutions, and policies – not meas-ures of physical stocks. Furthermore, we antici-pate that those rural households that are affectedby social forestry follow fairly systematic pat-terns of behavior with respect to these marketsand the policies that affect them. This perspec-tive is consistent with Dewees’ (1989) earlierobservations for social forestry, with Schultz’(1964) observations for the broader experiencesof farm households, and with the general litera-ture of rural development.

Nevertheless, some important differences dodistinguish household participation in agricul-

ture from participation in social forestry. Thesource of these differences is the open accesscharacteristic of the natural forest. This charac-teristic accompanies important gender and wealthdistinctions in the classes of resource users andseasonal differences in forest use itself. It sup-ports insights to resource use (by whom, when,and for what purposes) that are different fromthe agricultural experience where households tendto have more secure rights to their lands andcapital. Questions about deforestation, sustain-able use, and reforestation each require insightsto the use of the open access forest and to theconditions that cause local farm households toshift from reliance on this resource to the second-ary forests that are beginning to appear on thefarmers’ own agricultural lands.

Our paper begins by revisiting the social for-estry concept itself and the general expectationsmany hold for its beneficial impacts on ruralcommunities and their environment. These ex-pectations lead us to the new questions, as wellas the doubts that some practitioners eventuallybegan to raise. Both the original expectationsand the new questions induce us to review theempirical evidence. We will review the evidenceon i) consumption, ii) production and supply,and iii) investment in new technologies, beforesubmitting iv) a summary characterization of thegeneral forest environment, and then v) closingwith our view of the lessons for policy analysis.

We will find that market incentives identifythe most likely households to invest in any newtechnology only when the return is good, therights to the technology and its products are se-cure, the risks are acceptable, and the policyenvironment is stable and predictable. We thinkthat some, but not all, of the critical market in-formation is well understood, and that there issome comprehension of the impacts of forestpolicies and other institutions. Nevertheless, wethink the factors that explain which regions andhouseholds actually do invest are poorly under-stood, that the opportunities for improved per-formance by local institutions are largely unex-amined, and that the crucial impact of the overallnational or regional policy environment has beenalmost entirely ignored.


287

2 Background: PrevailingWisdom

As the concept of social forestry has becomemore widely accepted, its definition has becomemore elusive. It includes local community andlocal private activities, often by subsistencehouseholds. For our purposes, it refers to theproduction and use of fuelwood, forage and fod-der, fruits and nuts, latex, gum, and various othernon-timber forest products. It includes domesticuses and local market exchange of constructiontimbers – but it does not include industrial woodproduction or domestic woodlot production forshipment beyond local markets.

The applications of social forestry have grownbeyond its original conception as seedling distri-bution, planting and technical assistance to in-corporate watershed management and the broad-er class of forest contributions to the naturalenvironment. Its practitioners recognize the soilsustaining properties of trees and good forestmanagement. They also recognize the indirectgains from substitution; in particular the gainsfrom substituting woodfuels for agricultural res-idues and the gains from introducing improvedstoves. When woodfuel replaces dung and strawthe latter remain in the fields where they addstructure and nutrients that help sustain depletedsoils. Improved stoves decrease the consump-tion of combustible material and, thereby, saveboth forests and soils.

Social forestry is seen all the more favorablybecause those who benefit from it are often themost disadvantaged: women, and the rural poor,and especially landless users of the forested com-mons. In many developing economies it is awoman’s task to collect the materials providedby the forest, and the forested commons is aresource of last resort for the poorest house-holds. Especially in times of greatest economicstress, it can become a source of both food andmarketable products.

In sum, social forestry has always had a localhousehold or community orientation and the ef-fects of activities that increase the occurrence oftrees and forests on the local landscape seem tobe uniformly positive. The additional contribu-tions of social forestry to global concerns for

reforestation and environmental sustainability areclearly positive as well.

The rural character of forestry means that manymarket transactions and a substantial amount ofconsumption from the forest never appear in anycountry’s national accounts. Nevertheless, vari-ous estimates have been calculated. Some sug-gest, for example, that hunting generates 20–50percent of cash income for forest villages indeveloping countries, and that wood constitutesfourteen percent of overall energy consumptionin developing countries in general and nearlyfifty percent of energy consumption in Africa(cited in Persson 1998). In a summary measurefrom the most complete attempt to date to incor-porate non-market environmental and resourcevalues in the national accounts, Peskin and delosAngeles’ (IRG with Edgevale Assoc. 1994) esti-mate that subsistence household use of the forestis the single largest undervaluation in the Philip-pine accounts, larger than water or air pollutionor soil depletion, and larger than subsistenceconsumption of agricultural or fisheries prod-ucts. It is a reasonable hypothesis that subsist-ence households in other developing countriesobtain comparable large relative values from theirforests as well.1

The institutional budget for forest develop-ment projects is good evidence of a policy com-mitment that is consistent with these observa-tions. The World Bank’s recent annual commit-ment has been in the neighborhood of US$250million. The Bank, through the Global Environ-mental Facility, administers another US$20–25million annually to protect biodiversity that large-ly occurs in forest environments. Bilateral aidagencies, regional development banks, and localgovernment agencies contribute additional sums.In India, for example, social forestry projectswere 25 percent of the national budget for ruraldevelopment in the late 1980s (Sharma, McGre-gor and Blyth 1991). India paid a staggering 35billion rupees for afforest 13 million hectares(Kajoor 1992). Almost all of these forest devel-opment activities have important links to thelocal human populations that live in and aroundthe forest.

Finally, the spontaneous and independent treeplanting undertaken by many local farm house-holds around the world is the best evidence of


288

the on-the-ground success of social forestry. Ourown experiences include examples in Pakistan,India, Nepal, Indonesia, the Philippines, China,Vietnam, Ethiopia, Kenya and Malawi. Oftenthe local magnitude of the tree planting activityis small but it can accumulate to real importanceacross broad regions. In Bangladesh, for exam-ple, local farm production accounted for 3/4 ofall timber and fuelwood consumption in the 1980s(Douglas 1982). It probably accounts for an evenlarger share now.

Nevertheless, despite the evidence of success,there have been many failures as well, and thesefailures have generated a healthy scepticismamong practitioners. More constructively, evi-dence of both sterling successes and stunningfailures encourage questions of “why some so-cial forestry activities succeed while others fail”or “what distinguishing characteristics of the suc-cesses would predict success in future introduc-tions of new social forestry technologies”?

We might anticipate that failures occur wherenon-forest consumption and non-forest uses ofhousehold labor are more important than forestconsumption and production. Surely it is reason-able that some poor households and some com-munities have more immediate concerns withfood than with forestry. This observation is re-lated to an income or household budgetary effectand we can anticipate greater likelihood of suc-cess where forestry occupies a larger share of thehousehold budgets of income or effort. In addi-tion, many have pointed to the importance oflocal institutions, and especially to well-estab-lished property rights for trees and forest land, asprerequisites for successful long-term forest in-vestments.2

Finally, Byron (1997) points out that the chang-es that occur in the course of normal develop-ment can be a deterrent to social forestry. Forexample, the demands for some forest productsdecline as income rise. As the values of forestproducts rise, the pace of substitution away fromthem also increases. Furthermore, the expandingcommercialization that accompanies develop-ment modifies gender roles, inter-household re-lationships, and the competition for propertyrights and, thereby, modifies production and con-sumption and the attraction of specialized socialforestry activities.

Can we build on these insights on incomeeffects, property rights, and changes in overalleconomic conditions to create a broad basis foranticipating the locus of successful social forest-ry investments? That is our task in this paper.

3 Responses to Scarcity

Both global concerns for deforestation and localconcerns for improved household welfare can beexpressed in terms of scarcity and prices. Great-er deforestation means increasingly scarce for-ests; therefore, increasingly scarce forest prod-ucts. Increased scarcity means higher prices andgreater opportunity for welfare enhancement byany of the myriad of social forestry activitiesthat decrease forest consumption or expand for-est production.

Godoy (1992) supports this argument with 21regional examples of farmers in Africa, Asia,and Latin America who responded to high forestproduct prices by planting trees. Godoy alsopoints out, however, that high prices are not asufficient condition for tree culture. Our ownobservation is that rising prices flag a potentiallyemerging opportunity for social forestry but anopportunity that is realized only when prices riseto a level sufficient to induce substitution, eithersubstitution of alternative sources of forest prod-ucts or substitution of altogether different non-forest products for the forest products.

The first part of this argument, and Godoy’sevidence, encourage us to examine the literatureon household price responses in the consump-tion and production of forest products. The latterpart of the argument – that high prices alone areinsufficient – encourages us to examine otherfactors as well as the own-price of forest prod-ucts, factors like household income and demo-graphic characteristics and factors like the pricesof substitutes.

Markets may be thin but they do exist – evenfor most minor forest products. Moreover, evensubsistence households generally participate insome markets.3 And when they do not trade inthe market for a particular product, they could –because local markets do exist. Households wouldparticipate if market prices fell sufficiently toinduce them to purchase rather than to expend


289

their own scarce labor collecting the product fortheir domestic use. When households collect fortheir own use, they are revealing that the valueof the labor they expend in collection is less thanthe market price of the product but greater thanthe wage their labor would earn in alternativeemployment.

This reasoning means that market informationon forest products is valid even in subsistencecommunities – but it is insufficient because thehousehold’s consumption and production deci-sions are inseparable. That is, household produc-tion decisions affect consumption, and vice ver-sa. An economic modelling approach known asthe “new household economics” accounts forthis non-separability. It has been applied widelyin agriculture, and more recently in several for-estry cases. We will rely on this literature andevidence from surveys of rural households tosort out the factors explaining household respons-es to scarcity.

3.1 Rural household consumption of forestproducts

Table 1 summarizes the empirical evidence onrural household responses to market prices for

forest products. It features the primary forestproducts in each region of inquiry, but the pre-ponderance of evidence is on fuelwood becausefuelwood is the most widely used forest productfor rural households. For example, market fuel-wood purchases account for as much as 20 per-cent of household cash outlays in central Malawi(Hyde and Seve 1993), and market fuelwoodproduction is the largest share of the 38 percentof all household income that comes from theforest in one district of Sri Lanka (Bogahawatte1997). The FAO estimate for the annual worldvalue of fuelwood is US$42 billion, perhaps 10–15 percent of which is sold in the market (citedin Persson 1998.)

Most observers have found that the householdconsumption of forest products is price inelasticregardless of the forest product – although con-sumption is less inelastic in the arid uplands ofEthiopia for example (Mekonnen 1998), than inthe cool moist hills of Nepal (Cooke 1998a).Cooke noted that household demands are moreprice inelastic when corrected for seasonal dif-ferences in consumption patterns – which areinfluenced by seasonal differences in weather,seasonal labor availability, and household stor-age opportunities. Cooke’s households are alsomore wage responsive than price responsive, a

Table 1. Consumption responses: price.

Study Location Forest product Measure Elasticity

Cooke (1998a) Nepal’s hills fuelwood demand shadow price –0.25*forage demand shadow price –0.10*fodder demand shadow price –0.11*

Mekonnen (1998) Ethiopia fuelwood demand shadow price –0.40*dung demand shadow price –0.72*

Amacher, Hyde & Nepal’s hills fuelwood market demand price –1.47*Kanel (1998) Nepal’s tarai fuelwood market demand price –0.21

Amacher, Hyde & Nepal’s hillsJoshee (1993) low income fuelwood collection time –0.28*

high income fuelwood collection time –0.84

Heltberg, Arndt & Rajasthan, fuelwood collection time –0.11*Sekhar (1998) India

Notes: a) Cooke and AHK found that household consumption was more responsive to (implicit) wages than to prices. Mekonnen found thatconsumption was at least as wage responsive as price responsive. b) Dung is found in the fields, but also in the same open access lands thatare sources for fuelwood, forage, and fodder. c) Collection time is a proxy for labor opportunity cost in AHJ and HAS. * indicates statisticalsignificance at least at the 0.10 level.


290

condition that raises two possibilities: a) as pric-es rise, the implicit wage for collecting forestproducts for the household’s domestic use be-comes a relatively more important determinantof household consumption, or b) higher wagesare indicative of higher incomes and householdincome or wealth is more important than marketprices as a determinant of consumption.4

Amacher, Hyde and Kanel (AHK 1998) drewsimilar conclusions for fuelwood in Nepal’s lowerelevation and drier tarai region – although notfor the breadth of Nepal’s hills – and they agreedwith Cooke that wages are a more importantfactor than prices in the household consumptiondecision.

Amacher, Hyde and Joshee (AHJ 1993) andHeltberg, Arndt and Sekhar (HAS 1998) mighttake the argument about the relative importanceof wages a step further. They found either littleprice variation (AHJ) or little evidence of mar-ket purchase (HAS). Rather than price, they choseto focus on collection time, a measure of thelabor opportunity cost and a proxy for the impor-tance of the household wage. Both AHJ andHAS observed inelastic household responses tothis measure. Inelastic collection time means that,while households do respond to deforestationand increasing scarcity by decreasing consump-tion, the consumption effect is small and it isdominated by offsetting increases in collectiontime. This is not a satisfying finding for thosewho anticipate that rational household behaviorwill solve the problems of fuelwood scarcity anddeforestation. If the major response to scarcity isan increase in collection time, then increasingscarcity does not induce increased forest protec-tion.

This finding also raises questions about thesource of any increase in collection time. Is la-bor a slack variable, such that households foregolittle by extending greater effort in fuelwoodcollection? Or does an increase in collection timemean that less time is available for householdchores like agricultural activities, food prepara-tion and childcare? Many have presumed thelatter. We will turn to the empirical literature onthe question in our discussion of household pro-duction.

3.1.1 Income Effects

The observation on the relative importance ofwage effects was consistent across all of theseanalyses. This observation, together with By-ron’s suggestion that forest products may evenbe inferior goods (consumption declines as in-comes rise), encourages us to examine incomeeffects. Negative income elasticities would indi-cate inferior goods. Small positive income ef-fects would indicate necessary goods whichhouseholds consume at relatively constant levelsregardless of their wealth. Inferior goods andnecessary goods typically (not always) consumea small share of household budgets and are ofsmall consequence to many important householddecisions.

Indeed, the seven analyses reported in Table 2suggest that the effect of household income onthe consumption of forest products is generallysmall, and that some forest products are inferiorgoods in some economies. Furthermore, two ofthese analyses suggest that the household’s in-come source does not alter this finding.5 Thesource of household income suggests somethingabout the household’s lifestyle, or at least howdependent the household is on its own agricul-tural land. Mekonnen and AHJ both found that itdoes not matter whether the household relies onincome from its agricultural production, fromthe hire of its labor, or from remittances. In eachcase the household’s income level has only asmall effect on its consumption of the most im-portant forest products.

The negligible income effect is one reasonwhy many households and communities aroundthe world have not been especially receptive toexternal forestry assistance despite local defor-estation and rising prices for forest products. Wenote that this argument corresponds well withGodoy’s point that high prices are important, butthat they are an insufficient incentive for manysocial forestry activities.

3.1.2 Substitution

To find those households with sufficient incen-tives, we need evidence of substitution awayfrom the forest product or into substitutes from


291

secondary forest ”plantations”. Table 3 reviewsthe evidence for substitution in consumption.One study (Cooke) examined forage and foddersubstitution. The others feature fuelwood and itssubstitutes. Commercial fuels like kerosene orLPG seldom penetrate rural markets. Therefore,the common substitutes for fuelwood in thesemarkets are combustible agricultural residues(like straw and dung) and improved stoves, atechnological substitute.

Neither Cooke nor Mekonnen found evidenceof substitution. Indeed, both observed comple-

mentarity, between forage and fodder as animalfood in Nepal (Cooke) or between fuelwood anddung for cooking and heating in Ethiopia (Me-konnen). The latter is an especially interestingfinding because the general view holds that drieddung is a primary substitute for fuelwood andthat the combustion of dung only decreases itscontribution to soil sustainability, long-run agri-cultural productivity, and nutrition in subsist-ence agricultural communities. Alternatively, in-creasing the availability of fuelwood decreasesthe negative effects of burning dung. Mekon-

Table 2. Consumption responses: income or wealth.

Study Location Forest Measure of Elasticityproduct income/wealth (or coefficient)

Cooke (1998a) Nepal’s hills fuelwood upland land area (+)*forage upland land area (–)*fodder upland land area (+)*

Mekonnen Ethiopia fuelwood labor income 0.063*(1998) fuelwood non-labor income 0.03

dung labor income –0.02dung non-labor income –0.02

Amacher, Hyde Nepal’s hills& Joshee (1993) low income fuelwood agricultural income –0.31*

fuelwood exogenous income –0.20*high income fuelwood agricultural income 0.0005

fuelwood exogenous income 0.002low income agricultural residues agricultural income 0.36*

agricultural residues exogenous income 0.05high income agricultural residues agricultural income –0.0004

agricultural residues exogenous income –0.001*

Amacher, Hyde Nepal’s hills fuelwood farm size 0.0005& Kanel (1998) Nepal’s tarai fuelwood farm size 0.07

Heltberg, Arndt Rajasthan, fuelwood farm size –0.25*& Sekhar (1998) India

Shyamsunder & Madagascar fuelwood rice production –0.01*Kramer (1996) palm leaves rice production –0.01*

wood crab rice production 0.00

Bogahawatte Sri Lanka(1997) low income fuelwood total income < 1

medicinal plants total income < 1mushrooms total income < 1

high income fuelwood total income < 0medicinal plants total income < 0fuelwood total income < 0

Notes: a) The household’s private land area is a proxy for income or wealth for Cooke, AHK, and HAS. Household rice production is a proxyfor income for SK. b) Agricultural residues include dung, and also crop residues like straw. Dung is found everywhere that cattle roam,including in the forest. Straw, of course, is not a forest product. c) Cooke did not calculate an elasticity but we can infer elasticities of less thanor greater than one from her estimated coefficients and sample size. * indicates statistical significance at least at the 0.10 level.


292

nen’s evidence urges caution on global generali-zations of this view.

AHJ did find evidence of substitution in theirnarrower survey of two districts in Nepal’s hills,substitution between fuelwood and agriculturalresidues, and especially for low income house-holds. They also observed that low income house-holds make the technological substitution of im-proved stoves for fuelwood.6

The real importance of AHJ’s observation hasto do with target populations. Fuelwood priceswere high enough to induce substitution – butonly for lower income households. AHK con-firmed this point and they sharpened the defini-tion of the target population. In a broader surveyof districts across Nepal’s hill and tarai regions,AHK observed that as prices rise some house-holds turn from market fuelwood purchases tofuelwood collection to satisfy their own domes-tic consumption. Households that collect but donot also participate in the market (which are alsolower income households, and especially in thehill region) are statistically significant substitut-ers of improved stoves for fuelwood.

3.1.3 Summation

The consumption evidence tells a consistent sto-ry. Households do respond to higher forest prod-uct prices by decreasing their consumption. How-ever, the household income elasticities for forestproducts are generally small. This indicates thatforest products are generally necessities and alsothat we should expect only the poorest house-holds to be especially responsive to their scarci-ty. This is not encouraging for the hypothesisthat increasing scarcity will slow rates of defor-estation, or for the widespread and general intro-duction of social forestry activities. Select cases,however, offer more promise. These select casescan be identified by populations that are suffi-ciently responsive to higher prices to substitutefor their consumption of forest products. In AHJ’sexamples from Nepal, this was not an entireregional population. Rather, it was a poorer sub-set of the entire population from a region thatwas relatively more price responsive as a whole.Of course, this is precisely the population thatpublic development projects intend to assist.

Table 3. Consumption response: substitution.

Study Location Forest Measure of Elasticityproduct substitution (or coefficient)

Cooke (1998a) Nepal’s hills forage fodder shadow price (–)fodder fodder shadow price (–)

Mekonnen Ethiopia fuelwood dung shadow price –0.3*(1998) dung fuelwood shadow price –0.7*

Amacher, Hyde Nepal’s hills& Joshee (1993) low income fuelwood improved stove owner (–)*

high income fuelwood improved stove owner (+)low income ag. residues fuelwood coll’n time (+)high income ag. residues fuelwood coll’n time (+)

Amacher, Hyde Nepal’s hills& Kanel (1998) market participants fuelwood improved stove owner (+)

collectors fuelwood improved stove owner –0.33*Nepal’s taraimarket participants fuelwood improved stove owner (+)collectors fuelwood improved stove owner –0.26

Heltberg, Arndt Rajasthan, fuelwood improved stove owner (–)& Sekhar (1998) India

Where elasticities were not calculated, we can infer substitution or complementarity from the signs on the estimated coefficients and thesample sizes. This evidence on substitution must be read carefully because some coefficients refer to prices or price proxies (collection time)while some refer to quantities (wood, stoves). * indicates statistical significance at least at the 0.10 level.


293

3.2 Household Production

Household production is more complex. Therange of household options is broader, includingcollection from alternate sources of the product,labor reallocation among the household’s sub-sistence activities (including fuelwood collec-tion), and labor reallocation away from forestcollection to wage and income producing activi-ties – which then allow the full range of con-sumption alternatives. The breadth of produc-tion alternatives has led to a more diverse empir-ical literature, and our generalizations from itmust be more speculative.

3.2.1 Market Supply

Table 4 summarizes the three insights to house-hold supply known to us. All three observedsome degree of price responsiveness, whetherhouseholds were supplying their own domesticconsumption or the local fuelwood market. Thelatter, market suppliers, are particularly impor-tant and they are easy to overlook. Most surveysof non-timber forest products intend to be ran-dom and unbiased. Nevertheless, they tend toobserve substantial levels of market demand but

only occasional market supply. This means thatthey overlooked some market-supplying house-holds. Unfortunately, these are probably lowerincome and landless households, the householdsof our greatest general concern.

Köhlin’s (1999) evidence from Orissa in Indiasupports this contention. Köhlin found thatwomen from lower caste and lower incomehouseholds were the most likely suppliers ofmarket fuelwood. They are also more likely toobtain their market fuelwood from the open ac-cess natural forest than from their communities’managed village woodlots – although the villagewoodlots were always closer than the naturalforest in Köhlin’s region. This seems to confirmthe evidence of several (notably Jodha 2000)that open access resources are resources of finalresort for the poorest households.

AHK’s evidence largely supports Köhlin, al-though it is less complete in its description of themarket suppliers themselves. It is also consistentwith AHK’s consumption evidence (Tables 1–3)which shows regional differences in fuelwoodscarcity in Nepal and shows that, beyond somelevel of scarcity, rural households do respond toincreasing scarcity by altering their consumptionbehavior.

Finally, HAS examined consumption, not mar-

Table 4. Production responses: supply prices.

Study Location Forest product Measure Elasticity

Köhlin (1999) Orissa, Indiamarket suppliers fuelwood equivs. market price

collection shadow wage

Amacher, Hyde Nepal’s hills& Kanel (1998) market suppliers fuelwood market price 2.99*

collection shadow wage 6.81*domestic collectors fuelwood market price 1.57*

collection shadow wage 1.33*Nepal’s taraimarket suppliers fuelwood market price 0.36*

collection shadow wage 0.55domestic collectors fuelwood market price 0.71

collection shadow wage 0.09

Heltberg,Arndt Rajasthan, India fuelwood market price 0.41& Sekhar (1998) market suppliers

Köhlin’s fuelwood equivalents are measures of the combustion equivalent to fuelwood that is obtained from all burnable forest products:leaves, twigs, and larger fuelwood. * indicates statistical significance at least at the 0.10 level.


294

ket supply, but they observed that twelve percentof households in India’s Rajasthan sold to themarket. Market sales represent 26 percent of allcollection for these households and these house-holds are (insignificantly) positively price re-sponsive.

3.2.2 Household Production, Land andLabor Inputs, Gender and the Poor

Table 5 summarizes the more extensive evidenceon household production itself. Evidence from arange of cases (Nepal, Madagascar, Ethiopia,India, Malawi) supports the contention that thecollection of fuelwood and other forest productsdeclines with decreases in the available foreststock and with decreases in the accessibility ofthe remaining stock. Furthermore, under condi-tions of sufficient scarcity, private stocks do be-come substitutes for forest resources on the com-mon lands. For example, in contrasting districtsin Nepal’s hills AHJ observed that when fuel-wood becomes sufficiently scarce on the com-munity’s common lands (smaller and less acces-sible stocks and higher prices), households even-tually begin growing wood on their own privatelands. Nevertheless, all observers agree that asthe stock declines or becomes less accessible,households generally spend more time collect-ing forest products.

The evidence that collection declines with de-creasing stocks of natural forests is encouragingof the hypothesis that increasing scarcity willslow the rate of deforestation. However, it is alsoconsistent with the consumption evidence thatwage effects are more important than price ef-fects. Therefore, we might recommend closerattention to wages and collection time than toprices as evidence of the sort of increasing scar-city that will induce substitution and deter defor-estation.

Labor allocation to the collection activity hasbeen a critical question because so many forest-ry and women-in-development projects presumea unique cultural role for women. Who collectsfrom the forest? Is collection the special respon-sibility of women, or of women and children?And does an increase in time spent collectingforest products detract from other household re-

sponsibilities, in particular the food preparationand agricultural production responsibilities thatare crucial to survival in subsistence rural com-munities?

In fact, Shyamsunder and Kramer (1996), Me-konnen, and AHJ all found that labor allocationvaries between genders. Collection is not thedomain of women alone whether in Madagascar,Ethiopia, or Nepal, and on some occasions menare the primary collectors.7

Cooke (1998, 1998a) and Köhlin (1998) pro-vide the most thorough inquiries into the ques-tion of labor allocation. Cooke focused on theseasonality of collection from the forest. Köhlinfocused on labor allocation between alternativefuelwood sources. Cooke found that, in Nepal’shills, collection takes advantage of slack laborand opportunities for joint production. Womenare the most important collectors of forest prod-ucts. Nevertheless, men do collect and they in-crease their collection in off-peak seasons foragricultural labor. Women increase their collec-tion in seasons when they spend more time awayfrom the household itself and in the fields closerto the trees.8 Youths increase their collectionwhen adults, especially women, are otherwiseoccupied with peak season agricultural activi-ties. And finally, the collection of forest prod-ucts does not interfere with agricultural labor.Rather, overall increases in collection time occurduring slack seasons or originate from reducedtime for leisure activities.

Köhlin found that, in Orissa, men actually col-lect more fuelwood than women and that themarginal products of men for the collection ac-tivity are greater than the marginal products ofwomen. Men, adolescents, and higher caste wom-en do more of their collecting from village wood-lots while lower caste women collect more fromthe less accessible natural forest – whether theyare collecting for market supply, as we previous-ly discussed, or for their own household use.Köhlin’s households are sensitive to the sourceof their fuelwood. With increasing plantationstocks they tend to decrease both collection frommore distant natural forests and purchases ofcommercial fuel.

Thus, the story that began with our review ofconsumption patterns remains intact after ourreview of production patterns. That is, with suf-


295

Table 5. Production responses: land and labor inputs.

Study Location Forest Land and labor inputsproduct

Amacher, Hyde Nepal’s hills fuelwood If more plentiful supply: collect from commons;& Joshee (1993) women collect, men do not, children negative

collectors.Nepal’s hills fuelwood If more limited supply: rely more on own lands,

capital inputs significant, men and women collect butmen collect more.

Amacher, Hyde Nepal’s hills fuelwood Level of resource stock and resource access significant.& Kanel (1998) Nepal’s tarai fuelwood Levels of stock and access significant, some indication

of supply from own land and of increased supply fromlands of wealthier households.

Shyamsunder & Madagascar fuelwood Collection increases with an accessible primary forest.Kramer (1996) palm leaves Men collect more than women.

wood crab

Mekonnen Ethiopia fuelwood Resource access is important.(1998) dung On common lands: female youth are significant

collectors, children contribute negatively.From household lands: men (and women?) aresignificant collectors.

Cooke (1998a) Nepal’s hills fuelwood Evidence largely from commons.forage Collection time (CT): male and youth CT small,fodder women’s CT is large, women absorb the increase

in CT due to increasing resource scarcity.Seasonal differences in CT: Men increase CTin off-peak agricultural season. Women increase CT(as joint product) when they are in the fields forseasonal agricultural activities. Youths increase theirCT when women otherwise busy.

Cooke (1998b) Nepal’s hills fuelwood Evidence largely from commons.forage Increases in resource scarcity increase forest CT.fodder Increases in CT do not decrease time for agricultural

labor. Rather, collection increases during slack seasons.

Köhlin (1998) Orissa, India fuelwood Resource stocks from a) village woodlots (VWL,equivalents which are managed commons) and b) open access

natural forests.Women, adolescents, higher caste collect from VWL.Lower caste collect from natural forest.Men collect more in general and MP(men) >MP(women), but formation of VWLs saves womenmore time.

Heltberg, Arndt Rajasthan, fuelwood Levels of resource stock and resource access significant.& Sekhar (1998) India Some private tree substitution for open access stocks.

Only women collect. Collection isseasonal, occurringwhen energy demands are high and demands foragricultural labor are low.


296

ficient scarcity, households do respond. On theconsumption side, less-well-off households aremore likely to feel the pressure of increasingscarcity. They respond by substituting alterna-tive consumption goods and by increasing theirreliance on their own productive abilities ratherthan on market purchases. On the productionside, the literature suggests that household laborallocations are consistent with the economic re-wards – rather than with external perceptions ofabsolute cultural norms like “women collect”.Men may even collect more than women whenthe returns to male collection are greater than thereturns to women’s collection effort – althoughthis may be unusual because other male wageopportunities are generally greater yet.

Some experience suggests that a) poor house-holds respond to increasing scarcity by relyingincreasingly on more distant open access naturalforests and some suggests that b) those house-holds with land respond by growing the scarceproduct on their own private agricultural lands.The evidence is not thorough on these points andthe identifying characteristics, particularly of thelatter agroforestry households, is not clear. Ournext section on adoption and investment address-es this issue.

3.3 Investments in New Technologies

Table 6 summarizes the six household econo-metric investment analyses known to us, plusone data intensive household survey (Scherr1995) which is also presented from a perspectiveof multivariate impacts even if its method is notstatistical. The specific parameters used in theseanalyses vary too widely for meaningful com-parisons and none of them reported elasticities.Moreover, several of these analyses neglect priceand cost factors – despite the almost certain ef-fect of those factors on a household’s desire toinvest. Nevertheless, they all provide useful in-sight.

In all cases the empirical findings follow theexpectations of economic behavior. Householdsinvest when they anticipate gain in the form ofhigher prices or lower costs, and the most likelyinvestors are households that can afford to take achance on new investments of uncertain poten-

tial. These are households with “risk capital” inthe form of larger incomes or greater wealth ormore labor or land than their neighbors, perhapsenough more to allow them to chance a smallinvestment in an unproved new technology. Thepoorest households do not have the means totake this chance and the risk to them seems greaterrelative to whatever means they do have.

The investment literature tells us that better-off households invest first. Poorer householdsfollow. The consumption literature told us thatsome of these poorer households are the moreelastic respondents when they do invest.

Furthermore, the adoption of new consump-tion technologies is a more important question inurban areas where policy often encourages thesubstitution of commercial fuels as a means ofdecreasing the drain on forest resources. Sincecommercial fuels seldom penetrate rural mar-kets, it is not surprising that the only analysis ofrural adoption of a consumption technology fea-tures improved stoves. Amacher, Hyde andJoshee (AHJ 1992) observed that wealthier house-holds were the first to adopt improved stoves.They also observed that off-farm income is amore important predictor of adoption than in-come from a household’s agricultural produc-tion. We might speculate that off-farm incomeindicates a greater variety of labor experiencesand, therefore, broader household experience ingeneral and greater exposure to information aboutnew technologies like improved stoves.

3.3.1 Investments in Trees

Two of the analyses of production-enhancinginvestments also identified the importance of ac-cess to information. In these cases (Amacher,Hyde and Rafiq or AHR 1993; Thatcher, Leeand Schelhas or TLS 1997), information wasdelivered through training programs or by exten-sion foresters. AHR’s Pakistani farmers took ad-ditional notice of the personal characteristics ofthe extension agents who deliver the new infor-mation. The extension agent’s good characterdetermines the household’s openness to infor-mation delivered by the agent. To these Pakista-ni farmers, the agent’s character is even moreimportant for successful technology transfer than


297

Tabl

e 6.

Cha

ract

eris

tics

of h

ouse

hold

s an

d in

vest

men

ts in

new

tech

nolo

gies

.

Exp

lana

tory

var

iabl

es: s

ign

of e

ffec

t on

inve

stm

ent

Stud

yL

ocat

ion

Inve

stm

ent

Fore

stPr

ice

Inco

me

Subs

titut

eW

ealth

or

Lan

dSe

curi

tyH

´hol

d la

bor

Tra

inin

gpr

od.

var´

nfa

rmof

f-fa

rmla

ndqu

ality

of la

ndto

tal

mal

ean

dpr

ice

owne

rshi

pte

nure

exte

nsio

n

CO

NSU

MPT

ION

Am

ache

r, H

yde

Nep

al’s

hill

sim

prov

ed+

insi

gnif

.+

–+

NR

NR

& J

oshe

e (1

992)

stov

es

PRO

DU

CT

ION

Am

ache

r, H

yde

Paki

stan

tree

pla

ntin

g+

++

–+

++

+&

Raf

iq (

1993

)

Mek

onne

nE

thio

pia

num

ber

NE

⇐+

⇒N

E+

++

+(1

998)

of tr

ees

Tha

tche

r, L

ee &

Cos

ta R

ica

part

icip

ant:

NE

–+

NE

+–

0+

+Sc

helh

as (

1997

)re

fore

st p

gm.

Pate

l, Pi

nkne

yK

enya

num

ber

NE

–⇐

+⇒

NE

++

& J

aege

r (1

995)

(Mur

ang’

a)of

tree

s

Sche

rr (

1995

)K

enya

(Si

ega

num

ber

+–

⇐+

⇒N

E+

++

and

S.N

yang

a)of

tree

s

Shiv

ely

(199

8)Ph

ilipp

ines

num

ber

+–

––

+of

tree

s

Not

es: a

) N

R: n

ot r

elev

ant i

n th

e co

nsum

ptio

n fo

rm. N

E: t

his

criti

cal v

aria

ble

was

not

est

imat

ed. b

) A

HR

als

o no

ted

the

impo

rtan

ce o

f th

e pe

rson

al c

hara

cter

istic

s of

ext

ensi

on a

gent

s. c

) M

ekon

nen

did

not s

epar

ate

inco

me

sour

ces.

He

iden

tifie

d a

posi

tive

effe

ct f

or in

crea

sing

num

bers

of

mal

es in

the

hous

ehol

d, b

oth

yout

hs a

nd a

dults

, and

an

addi

tiona

l pos

itive

eff

ect f

or m

ale

head

s of

hou

seho

ld. d

)T

LS

also

obs

erve

d a

nega

tive

effe

ct f

or im

prov

ed la

nd q

ualit

y. T

heir

mea

sure

s of

off

-far

m in

com

e an

d la

bor

wer

e i)

off

-far

m in

com

e as

a s

hare

of

all i

ncom

e an

d ii)

labo

r/he

ctar

e. e

) PP

J us

edex

pend

iture

/cap

ita a

s a

prox

y fo

r al

l inc

ome,

and

res

tric

ted

thei

r m

easu

re o

f la

bor

to la

bor

avai

labl

e fo

r fa

rmin

g. T

hey

obse

rved

that

tree

gro

win

g is

mor

e lik

ely

for

high

er w

age

hous

ehol

ds. T

hey

antic

ipat

e th

at h

ighe

r w

ages

mea

n hi

gher

inco

mes

whi

ch s

ugge

st lo

wer

dis

coun

t rat

es. T

he la

tter

is a

n im

port

ant f

acto

r fo

r lo

ng-t

erm

inve

stm

ents

like

tree

s. f

) Sc

herr

did

not

obs

erve

pri

ces

dire

ctly

, but

she

did

obse

rve

mor

e pl

antin

g w

here

ret

urns

are

gre

ater

(pr

ices

hig

her

or c

osts

low

er).

g)

Shiv

ely’

s ob

serv

atio

ns w

ere

rest

rict

ed to

man

go tr

ees.


298

the agent’s knowledge of forestry.The literature on new production technologies

raises additional points about the income, labor,and land variables, and introduces a risk spread-ing or income diversification argument that fa-vors private investments in trees.

TLS offer a new perspective on income sourc-es and labor availability. They observe that thoseCosta Rican households that are reliant on off-farm income and employment have less laboravailable for on-farm agricultural activities. Thesehouseholds may be more receptive to tree plant-ing opportunities on their own lands because,unlike many agricultural crops, trees grow wellwith minimal labor input and that input can bescheduled for the slack seasons for off-farm em-ployment. TLS addressed these propositions byreformulating the common income and labor var-iables as i) off-farm income as a share of allincome and ii) labor available for farming.

Scherr examined gender differences in the la-bor input. She observed that while gender doesmake a difference, the relationship is complex.Investment decisions reflect distinctions in landtenure and off-farm employment and income aswell as gender – but it is difficult to separate theeffects of these characteristics on the decision toinvest in trees. For example, female heads ofhouseholds in Scherr’s regions of Kenya haveless secure tenure and males are the greater par-ticipants in off-farm employment.

Most farm households have been used to col-lecting their forest products from the natural for-est. For them, forest investments are a totallynew experience – different in their use and tim-ing of inputs and in their output markets, differ-ent from household experience with the naturalforest and different from the improvements manyhouseholds have grown to expect from new agri-cultural technologies.

Therefore, it is not surprising that Scherr ob-served that households invest only incrementallyin new agroforestry technologies, beginning at asmall scale and expanding gradually as the in-vestment demonstrates its worth. Better-offhouseholds are generally the first to invest andScherr observed that they are willing to invest inlonger rotation species and for commercial tim-ber production. When poorer households even-tually invest in agroforestry, Scherr observed

that they tend to plant short rotation fuelwoodspecies. This is consistent with Patel, Pinkneyand Jaeger’s (PPJ 1995) observation, for a dif-ferent region of Kenya, that better-off farmersplant more. PPJ speculate that the better-off have(better access to credit and) lower time prefer-ences.9 Scherr speculates that a household laborconstraint leads the poor to invest in fuelwoodplantings as a means to save time spent collect-ing fuelwood.10 In contrast, she speculates thatland is the more constraining factor for invest-ments in technologically advanced agroforestrypractices like alley cropping that are more im-portant for commercial production and more at-tractive to better-off households.

3.3.2 Risk and Uncertainty

Finally, first Dewees (1995), then Scherr andPPJ, observed the risk reducing characteristicsof agroforestry investments in their cases in EastAfrica. Dewees and Scherr consider that the re-turns from investments in trees are less volatilethan returns from agriculture. Scherr also ob-served attempts to diversify by planting a varietyof tree species. PPJ took particular notice of theimportance of tree planting as a factor in control-ling environmental risks like erosion.11

Dewees’ and Scherr’s observations on the riskreducing characteristics of many East Africaninvestments in trees are related to Shively’s(1998) observation that Philippine farmers plantmore mango trees when the variation in mangoprices is smaller. Scherr also showed that reduc-tions in price variation can be as important tofarmers as increases in product prices. All threeobservations on price variation are comparableto North American evidence that forest invest-ments are a valid tool for diversifying invest-ment portfolios. In North America, investmentsin trees yield returns that are less variable thanaverage portfolio returns, as well as returns whosecycles run slightly counter to normal businesscycles (See Binkley, Raper and Washburn 1996).Is this a general pattern for forest investments inboth developed and developing countries?


299

3.3.3 Overall Summation

The household literature explains a lot aboutthe sequence in which different classes of farmhouseholds invest and why they invest, and itreinforces the consumption and production argu-ments to carefully select target populations fortechnology transfer and other development ac-tivities. Households invest i) because forest prod-ucts are an important component of the house-hold budgets of money or (especially) time, per-haps ii) for the income protection that productdiversification offers, and iii) because relativeprices, or imperfections in labor or land markets,are sufficient to make this investment better thansome alternative expenditure of household re-sources. The question that remains is “when arethe relative forest product prices great enough orlabor costs low enough?”12 This question is es-pecially important where forestry is an altogeth-er new investment. Farmers who are entirelyinexperienced with it will be the greatest bene-factors from the advice offered in training ses-sions and from forest extension agents. Whenthe relative prices and wages are insufficient,however, no amount of additional informationcan create a good investment. Therefore, know-ing when the prices and wages are sufficient tojustify investment in a new forest technologywill predict when the various forms of technolo-gy transfer can be beneficial.

The consumption evidence suggests that pric-es and wages are sufficient when we begin toobserve substitution. We can anticipate that thesame signal is appropriate for investments innew production technologies as long as we ob-serve an important caution with respect to prop-erty rights for land and trees. Consider poor, andthen better-off households, then consider com-munity investment opportunities. Poor house-holds may substitute other goods or alternatesources of the forest resource but they will notinvest in new forest production technologies iftheir landholdings are insufficient even for morecritical household products like food. House-holds with larger landholdings may substitutefor scarce forest products but they may not in-vest in new production technologies when theperiod of land tenure is uncertain or where roadsor government policies limit their access to mar-

kets. Communities may be slow to invest if theyshare rights to common lands with a central gov-ernment agency and the system of shared rightsis uncertain or confining. This final case is com-mon because central governments retain legalownership of most countries’ forests and de-graded rural lands, and because the central gov-ernments’ rules for local uses of these lands areoften restrictive, yet the central governments donot have the resources to enforce their rules com-pletely and uniformly. We will revisit the ques-tion of private or community investment in thenext section.

4 Unifying Principles

These observations generally follow a rather ba-sic economic principle – resources (land, labor,and what limited capital these households pos-sess) are allocated to higher valued uses until themarginal gains from their use equal their mar-ginal costs. No one contends that the markets areperfect but increasing scarcity in the local mar-kets does send the anticipated signals, and ruralhouseholds do reallocate resources in the ex-pected directions. Indeed, the empirical and ana-lytical literature seems to reject the most com-mon claim for a cultural constraint on marketbehavior, i.e. women collect, in the presence ofevidence that both men and women collect, andthe question of “who collects” seems to dependon comparative economic advantages within theparticular household.

4.1 Von Thunen’s Pattern

A common pattern begins to emerge from theobservations summarized in this review, a pat-tern of rural development, deforestation and in-creasing scarcity of forest products, and eventualsubstitution with forestry investments that couldpotentially limit further deforestation. This pat-tern is an application of the concepts of econom-ic geography first proposed by von Thunen (e.g.,Dickinson 1969, Samuelson 1983). Our Fig. 1captures the basic elements of this pattern. Italso provides the key reference points necessaryfor further reflections on investment timing, in-


300

stitutional constraints, and the policies affectingsocial forestry.

Fig. 1 describes a simple landscape of agricul-ture and forests. Consider agricultural land first.The value of agricultural land is a function of thenet farmgate price of agricultural products –which is greatest when the farmgate is near thelocal market at point A. Land value declines withdecreasing access (which is closely related toincreasing distance) as described by the functionVa. Our exposition will be clearer if we separatethe costs of insuring property rights from othermanagement costs. The function Cr describesthe cost of establishing and maintaining securerights to the land. This function increases as thelevel of public infrastructure and as effectivecontrol declines as the cost of excluding tres-passers increases. Both increase with distance.Local communities may protect some lands be-yond B to a declining degree further from theirhomes – as by sending children out to managetheir grazing livestock – but eventually no numberof forest guards can fully exclude open accessusers of remote forests.

The functions explaining agricultural land val-

ue and the cost of secure property rights intersectat point B. Farmers manage land between pointsA and B for permanent agricultural activities.They use land between points B and C (whereagricultural land value declines to zero) as anopen access resource to be exploited for short-term advantage. They harvest native crops thatgrow naturally in this region, crops like fodderfor their animals, native fruits, and fugitive re-sources like wildlife. They do not invest even inmodest improvements in the region between Band C because the costs of protecting their in-vestments would be greater than the investmentsare worth. Their use of this open access region isunsustainable except in pulses of natural re-growth.

Initially, the mature natural forest at the fron-tier of agricultural development at point B has anegative value because it gets in the way ofagricultural production and its removal is costly.Settlers remove trees whenever the agriculturalvalue of converted forest plus the value of thetrees in consumption exceeds their removal costs.Eventually, the most accessible forest resourceswill have been removed. The function Vf de-

Fig. 1. The forest landscape – a new frontier.

A B C D

Decreasing access

Vf = Forest landvalue gradient

Cr = Marginal cost of obtaining andmaintaining secure property rights

Va = Agricultural landvalue gradient

Increasing land value

Sustainable agriculturalproduction

Open access,degraded forest

Mature primary forest


301

scribing forest value must now intersect the hor-izontal axis at some distance beyond B. Marketdemand for forest products may still justify theirremoval at this time, and it will continue to justi-fy their removal as the forest frontier graduallyshifts outward to some point like D. The marketprice of forest products is now just equal to thecost of their removal and delivery to the market.The in situ price at point D is zero, and the valueof forestland at D is also zero. The region ofunsustainable open access activities now extendsfrom B out to either C or D, whichever is farther.The costs of obtaining and protecting the proper-ty rights insure that this region will remain anopen access resource. Some governments pro-tect some lands past point B but they must ab-sorb the increasing protection costs – and eventhen trespass occurs. Some amount of illegallogging occurs almost everywhere in the worldand no number of well-trained and well-motivat-ed forest guards can prevent it.

The construct of Fig. 1 conforms to the com-mon description of any initial settlement. In somecases, trees actually impede agricultural devel-opment, the forest rent gradient is very low, and

point D can even be to the left of point C if netforest resource values are sufficiently low. Ap-parently, this describes the forest frontier in Coted’Ivoire (Lopez 1998) and Bolivia’s Amazonregion today (Bowles et al. 1997). In other cases,the region between B and D can be large (e.g., inNepal’s hills or India’s Rajastan) but the forestin this region is generally degraded. In the lattercase, the positive net value of the original re-source, together with the open access characterof the region, has assured removal of the bestresources. Some degraded vegetation remains inthe region and it will re-grow naturally. Thelowest wage households will continue to exploitthese resources when the scattered vegetationgrows to a minimum exploitable size or when itsfruits begin to ripen (AHJ 1993, Foster et al.1998).

As the natural forest is depleted over time, theforest margin at D will gradually extend fartherand farther from the market. Deforestation willcontinue, and the delivered costs of forest prod-ucts will continue rising. The incentives are in-sufficient to induce tree planting and any at-tempt at forest management will be unsustaina-

Fig. 2. The forest landscape – a mature frontier.

DA B C

Decreasing access

Vf = Forest landvalue gradient

Cr = Marginal cost of obtaining andmaintaining secure property rights

Va = Agricultural landvalue gradient

Increasing land value

Sustainableagriculturalproduction

Open access,degraded forest

Matureprimaryforest

B’ D’B”

Sustainableforest


302

ble. As Godoy (1992) points out, the prices offorest products may be rising, but they are notyet sufficient to induce forest management.

Eventually the margin at D extends far enough– and delivered costs and local prices becomegreat enough – to induce substitution. Substitu-tion may take the form of new consumptionalternatives to forest products (for example, ker-osene or improved stoves as substitutes for fuel-wood), or it may be production related (as inplanting and sustainable forest management onsome land closer to the market). In Latin Ameri-ca alone there are 165 million hectares of thissecondary forest, an area three-fourths the sizeof Mexico (Smith et al. 1998).

The forest rent gradient rises with the increasein delivered costs (following the arrows in Fig.2) until it intersects the agriculture rent gradientto the left of agriculture’s intersection with theproperty right cost function. We might call this a“mature” frontier of primary forest. At this timeforest product prices will be sufficient to justifythe substitution of planted and managed forestsfor the resources of the open access natural for-est. The new managed forests will occur in theregion B’B” of Fig. 2. They may take the form ofindustrial timber plantations or they may takethe forms of agroforestry or even of just a fewtrees planted around individual households. Thelatter are excluded from most measures of theforest stock but their economic importance canbe large. In Bangladesh, for example, they ac-count for 3/4 of all timber and fuelwood con-sumption (Douglas 1982). They are major sourc-es of fuelwood consumption in Malawi (Hydeand Seve 1993), timber production in Kenya(Scherr 1995), and of positive environmental ex-ternalities in northern China (Yin and Hyde 1998)or the Murang’s region of Kenya (Patel, Pinkneyand Jaeger 1995).

In all cases, removals from the mature naturalstock are concentrated around a point like D inFig. 1 – or D’ in regions characterized by thehigher forest rent gradient in Fig. 2. Mature nat-ural stocks in the region before D (or D’) wereremoved in earlier times. In most cases a maturenatural forest of no economic value exists be-yond D (or D’). Sometimes this region is small(e.g., in Ireland or Cape Verde). Sometimes thislatter region continues well beyond D (Siberia,

Alaska, northern Canada, much of the Amazon)until it can become much the largest share ofreported physical stocks. Many countries pos-sess regions of both cases because many forestproducts are either bulky or perishable and donot transport well. That is, standing natural re-serves remain in some regions while the forestsin other regions are depleted and some house-holds in the latter may have begun planting ontheir own lands.

4.2 Evidence and Qualifications

A number of empirical studies have demonstrat-ed this development pattern for commercial tim-ber.13 More recently, several have demonstratedits relevance for fuelwood and for agriculturalexpansion. In particular, Hofstad (1997) con-firmed a pattern of expanding extraction for char-coal in the vicinity of Dar es Salaam and Chom-itz and Griffiths (CG 1997) confirmed a similarpattern for charcoal from multiple populationcenters in Chad. CG also observed the substitu-tion opportunities that constrain expanding sup-ply regions and rising charcoal prices once abackstop energy source or technology price isattained. Their observation that substitution even-tually constrains deforestation is consistent withobservations from the household literature thatsubstitution constrains the consumption of highcost fuelwood.

Most everyone is careful to recognize that mar-ket access on the horizontal axis of Fig. 1 ismore complex than distance alone. It probablyincludes road quality and density and measuresof both terrain and moisture: steep and swampyareas are less accessible than flat dry areas. Sitequality can also be important for some land usesand Chomitz and Gray (1996), for example, werecareful to incorporate an element of land qualityin their assessment of agricultural expansion intothe forested regions of Belize.

Most everyone would also recognize that thereare many forest value gradients. Rather than thesingle sharply defined gradient Vf in our figures,each forest product has its own gradient. Andmost recognize that the cost function Cr varieswith the individual and institutional particularsin the local market. It almost always includes an


303

element of hazy uncertainty that is not fully ac-knowledged in our sharply drawn figures. (Wewill return to this latter point about propertyrights later.) Regardless of these adjustments,von Thunen’s pattern prevails in general.

There is a data problem, however, that impos-es exceeding difficulty for accurate measures ofthe economic forest stock, for economic meas-ures of resource scarcity, and for meaningfulestimates of rates of deforestation. The commonestimates of resource supply in forest policy anal-yses are actually measures of the standing physi-cal stock. Adjustments for the difference be-tween physical stock and economic supply areseldom incorporated in analysis and the poten-tial for error is great. Alston, Libecap and Muel-ler (1998) demonstrated the importance of infra-structure to accurate measures of secure eco-nomic supply. In particular, they demonstratedthe negative relationship between access to pop-ulation and government administrative centersand the costs of enforcing claims on land use infrontier regions of Brazil. (This is precisely thefunction Cr in our figures.) Hyde, Krutilla, Barnesand Xu (1998) observed that the standard meas-ures of physical resource stocks include a largecomponent of forest beyond point D (or D’).They corrected for the difference between physi-cal stock and economic supply with adjustmentsfor terrain, road density, and moisture. Theircorrections explained forty percent of the varia-tion in estimated physical stocks across 31regions in eight tropical countries. General poli-cy differences explained no more than twentypercent of the variation. The obvious point shouldbe that elements of infrastructure and terrain arecritical. They can dominate elements of policy,yet assessments of forest policy and forestlanduse only rarely account for them.

4.3 Further Development: the Rural Poor

So, the empirical evidence confirms the patternof our von Thunen figures, but how do the fig-ures and this evidence relate to the human popu-lations and the production and consumption ofthose products of greatest concern for social for-estry, the topic of our review? They argue thatlocal users of the forest will continue to travel

farther and farther to collect and deplete the nat-ural forest until the local market prices or theimplicit wage for their own collection time at-tains a level sufficient to induce substitution of abackstop technology. For households withoutland or without secure rights to their land, thebackstop technology can be consumption substi-tutes like kerosene or improved stoves. For thosewith secure rights to their own land, it can be aproduction substitute like private or local com-munity tree plantings. This is exactly what thehousehold literature observed with respect to pric-es and wages or collection time, substitution,and adoption.

But what about the rural poor? They are thegreatest concern of rural development, and sev-eral of the production and consumption assess-ments discussed earlier in this paper noted great-er participation of the poorest households in thearray of activities sometimes identified as “so-cial forestry”.

Foster, Rosenzweig and Behrman (FRB 1997)addressed this question with a general equilibri-um model of three sectors: agriculture, rural man-ufacturing, and open access forests. In their mod-el, land is an input to agriculture and forests, butnot to manufacturing, and forests are open ac-cess resources. FRB’s model could be interpret-ed as similar to our figures with all manufactur-ing taking place at the origin. FRB’s open accessforest is, of course, the region to the right of B inFig. 1 (or B” in Fig. 2). In FRB’s model, themarginal product of agricultural labor is its wage.The average product for forest labor falls until itequals its marginal product – in what is the stand-ard observation for open access resources – andthe wage for forest labor is below the wage foragricultural labor.

FRB confirmed their analysis with satellitedata on land use and an extensive householdsurvey from 250 villages in eleven states of In-dia. One of their conclusions is that the lowerwage for forest labor causes the removal of for-est resources to extend past the optimum forforest properties with secure property rights – asour figures also show. Perhaps more interestingfor us is the implication of their analysis that it islower wage households and individuals that par-ticipate in extraction from the forest.

We would hypothesize that since individuals


304

from lower wage households are satisfied with alower return for their marginal effort they canafford to venture further into less accessible re-gions to obtain the products of the forest. And inthe more accessible regions of depleted openaccess forests, individuals from lower wagehouseholds can afford to remove lower qualitymaterial that would not provide sufficient com-pensation for the effort of the better-off. Thepoor will renew their exploitation of the degrad-ed open access regions as the depleted resourcesin these regions recover to minimal levels. Thesedepleted resources will never attain a level suffi-cient to compensate the effort of the better-off.Of course, FRB’s analysis and this hypothesisexplain many observations around the world ofthe greater dependence of the poor and other-wise disadvantaged on the forest. Once more,they are consistent with observations from thehousehold literature, observations that poorerhouseholds are the first to respond to increasingprices and they first respond by shifting frommarket purchase to household collection fromthe open access forest.

Finally, FRB showed that technical changehas a varied effect on forest use. Technical changein the manufacturing sector raises wages andattracts labor from the other sectors. By drawinglabor from the forest, it decreases dependence onthe forest. Bluffstone (1995), with evidence fromNepal, anticipated that off-farm labor opportuni-ties in general would have this effect. On theother hand, FRB showed that technical changein agriculture has a different effect. While tech-nical change in agriculture also raises wages andattracts labor, it can deplete the forest by induc-ing land conversion to agriculture. We wouldanticipate that the new agricultural land wouldcome from the region of point B for new frontiercommunities described by Fig. 1. It would comefrom the plantation forest region B’B” for themature communities described by Fig. 2. Sun-derlin and Wunder (1998) confirm these con-trasting effects of manufacturing and agricultur-al sector growth on the forest with macroeco-nomic evidence from a broad sample of 67 coun-tries.

5 Shortcomings in Data andResearch

We have reviewed an extensive body of empiri-cal analysis. Much of it is recent, and most of itis internally consistent. This literature does haveits shortcomings, however, and it would be use-ful to identify them before drawing any conclu-sions for policy. The shortcomings fall into thegeneral categories of data, the agriculture/forest-ry interface, and macroeconomic policy impacts.Each imposes cautions on the policy conclusionswe draw in the final section of this paper andeach reminds us of research questions that areyet unanswered.14

The first set of shortcomings has to do withdata. Many of the household analyses conductedto date feature South Asian examples. A fewrely on examples from Asia outside of the Indiansub-continent and a few rely on African data.The particular dearth of Latin American exam-ples may not be a serious shortcoming becausethe majority of Latin America is not a forestscarce region, and perhaps it is not a seriousshortcoming if one accepts the general premisethat human behavior is similar everywhere. Inthis case the important range of examples forsocial forestry would be the full spectrum ofrelative scarcities for the important local forestproducts regardless of their geographic identity.Nevertheless, it would add confidence to anygeneralizations we might draw if Latin Ameri-can experiences were a greater part of the litera-ture.

A second and potentially more important datashortcoming is that most household surveysshortchange the local landless population. Thiswas not their intention but we did note in ourdiscussion of household production that mostsurveys have more evidence of household mar-ket purchases of forest products than of house-hold market sales. Since the sellers of forestproducts in subsistence economies tend to belower income and landless households, this meansmany data sets must under-represent the poorand landless households. This is a particularlyunfortunate shortcoming because the pooresthouseholds are the most important target of eco-nomic development.


305

A third data shortcoming has to do with reli-ance on physical rather than economic measuresof forest stocks and, especially, physical meas-ures that capture stocks of dense forest whetheror not these stocks are economic while entirelydisregarding the economic contribution of spars-er groups of trees in fields, along hedgerows, infarmyards, and in degraded areas. Both correc-tions can be critical. We are reminded that HK-BX’s sample from 31 regions in eight tropicalcountries found that correcting for the error oruneconomic stocks accounted for twice the vari-ation in established forest stocks due to policydifferences. Meanwhile, we are also remindedthat Douglas observed that 3/4 of economic sup-ply in Bangladesh originates from unmeasured,mostly private, stocks. Of course, the real prob-lem is that the data- and analysis-intensive effortnecessary to overcome these shortcomings wouldoverwhelm most forest policy analyses.15

The second major analytical shortcoming large-ly has to do with the agriculture-forestry inter-face. As the natural stock of forest resourcesbecomes scarcer, the economic reasoning andthe empirical evidence pretty much establish thathouseholds eventually substitute land and laboraway from activities like agricultural productionand toward forest activities. Should we be con-cerned that these poor households may be fore-going some component of their agricultural sub-sistence? They may also switch labor away fromhousehold activities like nutrition and health careto forest activities. Or, stating it as a question“How much do human health and productivitysuffer from a decline in forest cover?” In con-trast, as development proceeds, improving wagelabor opportunities may correlate with increasedforest cover. How much do health and produc-tivity gain from an increase in forest cover?

Most, not all, data sets and household analysesto date have restricted themselves to open accessforest resources and to forest-related householdactivities. We cannot expect to understand socialforestry and its welfare impacts without improv-ing our understanding of the opportunities foralternate allocations of the household’s most im-portant productive inputs, its labor and its agri-cultural land.16

Finally, our questions have all been framed inthe perspective of what scarcity means for sub-

stitution. We might also inquire what improve-ments in agricultural productivity have meantfor the availability of land and labor for forestproduction. Does increasing agricultural produc-tivity effectively, if indirectly, eliminate scarcityin forest products?

The third major shortcoming has to do withpolicy, specifically policy that induces either im-migration or emigration to the rural and forestedinterior and alters the composition of the popula-tion of rural poor. Because neither immigrantsnor emigrants are likely to be perfectly repre-sentative of the rest of the population, the popu-lation subsequent to imposition of the policymay have different consumption and productionpatterns. Our ability to predict the performanceof the new population with respect to social for-estry activities or anything else will suffer untilwe learn more about the policies that inducerural migration and the behavior of the popula-tions that most commonly migrate. This is animportant shortcoming because so many govern-ment policies and so much successful develop-ment is accompanied by migration either to orfrom the forested rural interior.

6 Final Policy Observations

Despite these research shortcomings, we havelearned a tremendous amount about rural house-hold responses to the increasing scarcity of for-est resources and about opportunities to intro-duce the technologies generally identified withsocial forestry. Despite forestry’s unusual char-acteristic as a standing mature open access re-source that is almost always available as a sub-stitute for managed production, we have evi-dence that rural households follow systematicpatterns of economic behavior in their produc-tion and consumption of forest resources. Wealso have evidence that managed and sustainablesocial forestry activities are successful in selecteconomic situations within select regions andfor select populations. Furthermore, their suc-cess often occurs among those poorest house-holds who are the intended targets of most de-velopment activity. We can identify these selectsituations. The remaining questions have to dowith policy. Where should we look for policy


306

opportunity? And what cautions does the evi-dence recommend for policy makers?

Our figures can help reflect on these two ques-tions. All three functions in the figures can beaffected by policy, and shifts in any of the threeaffect the management of either sustainably man-aged forests, or open access natural forests, orboth. We will consider each in turn.

But first, let’s consider government investmentsin infrastructure, in roads in particular. Theseinvestments affect the overall form of the vonThunen model, improving access for all landuses and shifting the intersection points for theagriculture and forest land use gradients (C andD or D’, respectively) to the right. The net effecton sustainable forest crops (region B’B” in Fig.2) is uncertain, but road improvements also clear-ly extend the region of consumption from theopen access natural forest. Therefore, road de-sign is clearly important to local forest use andgovernments would be well-advised to designtheir road improvement programs to minimizeaccess to those forested areas that are critical fortheir non-market values.

6.1 Spillovers from Agricultural Policy

Agricultural incentives shift the agricultural rentgradient upward and outward. Where forestry isa sustainable activity and managed forests repre-sented by the region B’B” in Fig. 2 do occur,then agricultural incentives make agriculture rel-atively more attractive than managed tree cover.They induce land conversion away from sustain-able forestry and toward agriculture.

Agricultural incentives have a very differenteffect in regions that are still reliant on the openaccess natural forest. These regions are betterdescribed by Fig. 1. Increasing the agriculturalrent gradient in this case induces land conver-sion as well, but from either the mature forest atnew frontiers where forest removal is still a de-terrent to agricultural expansion (the neighbor-hood of point C), or the degraded forest in longerestablished communities (between C and D).Agricultural incentives in cases characterized byFig. 1 also raise the price threshold or backstopthat must be overcome before forestry can be-come sustainable.

Therefore, we anticipate that agricultural in-centives will always work against sustainableforestry activities, either currently or by delay-ing their application in the future. Revising ill-conceived agricultural incentives can only pro-mote sustainable forestry, either now or oncelocal forest product prices have risen sufficientlyto justify sustainable activity.

6.2 Direct Forest Policies

Direct forest policies include incentives like costsharing, free seedlings, technical assistance, in-vestments in public lands, and public silvicultur-al research, as well as disincentives like taxes orrestrictions on harvesting and shipment. Incen-tives raise the forest rent gradient and disincen-tives lower the gradient. The latter, especiallydisincentives on shipment, are common in de-veloping countries. They include export bansand domestic processing requirements, anynumber of licensing requirements, and restric-tions on the mills or regions within a country towhich raw material shipment can be made. Theirobjectives are generally to assist the local indus-try or the environment – although their effect isnot always favorable for either one.

The impacts of either incentives or disincen-tives depend, once more, on whether the affect-ed region is a region of sustainable forestry (de-scribed by Fig. 2) or a region that still relies onan open access natural forest (described by Fig.1), and some forest policy instruments affect oneregion but not the other. For example, simplysubsidizing the prices of forest products (or re-ducing taxes on them) raises the forest rent gra-dient and increases the area under sustainableforest management in Fig. 2. It also extends thearea of open access forest depletion. Restrictionson shipments do the opposite. They decrease thenumber of potential competitive purchasers and,therefore, decrease the prices of forest productsand shift the rent gradient downward, decreasingthe area of sustainable management but also de-creasing the area of open access forest depletion.

Seedling distribution and technical assistance,on the other hand, decrease forest managementcosts and raise the net value forest gradient inregions characterized by sustainable forestry.


307

They have no effect on regions characterized byopen access extraction because prices in theseregions are insufficient to justify planting andhouseholds in these regions obtain no economicadvantage from the free seedlings or the man-agement advice offered by technical assistance.

These forest policy observations imply thatsiting can be an important aspect of social forest-ry. Köhlin’s (1998) work with households in 23villages in Orissa illustrates this point. (We dis-cussed this work in our sections on consumptionand production.) Köhlin observed that some vil-lages gain from the introduction of new villagewoodlots while others do not. Location is anessential variable for the villages that do gain.The greater the distance between new woodlotsand the natural forest, the greater the gain.

Location is especially important in subsist-ence economies because the regional infrastruc-ture tends to be poorly developed. Shipment ofbulky and perishable products is restricted andmarkets are geographically limited. Since forestproducts are generally one or the other, bulky orperishable, a weak regional infrastructure can beparticularly restrictive. Where the markets areall local, social forestry projects and target popu-lations must also have a very local orientation.Geographically broader forestry projects risk in-corporating both regions characterized by sus-tainable forestry and other regions characterizedby newer frontiers and open access forest con-sumption – which may ensure project failure inone region or the other.

Fortunately, many social forestry developmentprojects are local. Also fortunately, governmentextension programs provide an instrument that iscapable of addressing local differences. Exten-sion’s role in technology transfer may be partic-ularly appropriate where forestry is a new “crop”with which local households have little priormanagement. If extension services can identifythose communities where forest product pricesare about to attain the threshold that justifiessubstitution, and if they can identify the appro-priate technology for the conditions of land andlabor availability that characterize the communi-ty, then their role as transfer agents can be vitalfor improving both the forest condition and localwelfare.

No empirical literature known to us fully de-scribes how to predict when prices will reach thesubstitution threshold, or when some householdsand regions will be ready for the technologytransfer. This is an unsolved problem worthy ofreal consideration. Once this problem is solved,the literature on adoption suggests that the poor-est households are not the best initial targets fornew technologies. Households with more resourc-es, households that can afford to take some risk,are more likely initial adopters. The poor rapidlyfollow the leadership of successful adopters, how-ever, and they may receive the greatest eventualbenefit from many social forestry activities. Ob-taining these benefits, is generally contingent onpossessing secure land use rights for the proper-ties on which households might plant and man-age trees. This brings us to the final category ofpolicy opportunities.

6.3 Secure Rights to Forest Properties

The third function in our figures reflects thecosts of establishing and maintaining secure prop-erty rights. Decreasing these costs shifts the func-tion downward and converts some of the degrad-ed open access area to the right of B or B” in thefigures into sustainable use – either for agricul-ture or for forestry. In regions where forestry isnot yet sustainable, decreasing these costs reduc-es the price threshold that sustainable forestrymust eventually overcome. Some promising pol-icy options for social forestry; e.g., joint forestmanagement and community forestry; can bedescribed in terms of shifts in this third function.

The function captures two categories of activi-ties, direct costs and uncertain rights. Direct costsare costs like title or contract registration andfencing. Uncertain rights refer to either the un-certainty the user has for use of the land or itsforest resources, or the uncertainty the user hasabout the overall policy environment. An uncer-tain policy environment affects long-term ex-pectations for returns on investments in the landand, therefore, willingness to make long-terminvestments like forestry.


308

6.3.1 Direct Costs

The important prospects for reducing direct costslie in improvements in the regional infrastruc-ture that could make title registrations easier informal economies. Simply opening local officesfor registration or reducing the bureaucratic for-mality for registration are two examples. Thecounterpart to these in less formal economies isimproved access to the village headman or otherenforcer of local rules.17 Of course, the right toeasy transfer must accompany possession of se-cure rights. Otherwise, higher use values willeventually replace some historical values andcreate land use conflict and uncertainty even forthose with formal rights to the land.

Formal rights can be uncertain even in thepresent. The classic case in forestry features agovernment ministry which holds the formalrights and which may attempt to enforce them,but which does not have the means to enforcefull compliance with all of the ministry’s rulesfor forest use. The many examples of this casehave led to joint (ministry and local community)forest management (JFM) in India and to argu-ments for both privatization of natural forestsand land transfers to community-based forestryin the Philippines, Indonesia, VietNam, Colom-bia, Zimbabwe, and numerous other countries.

6.3.2 JFM, Community Forestry, andPrivatization

JFM provides some rights for local users butreserves other rights (usually for mature timber)for the government. Kant (1996) showed thattwo conditions are necessary for successful JFM:heavy local demands on the forest and a popula-tion that is homogeneous in these demands. Witha lower level of dependence, the local communi-ty has little incentive to invest its own resourcesin forest management and remaining open ac-cess stocks may be sufficient. Heterogenous de-mands suggest competing interests and difficul-ties in arriving at and respecting community de-cisions. Even where his two conditions are met,Kant observed that local demands on the forestvary across communities. Therefore, he arguedthat JFM contracts must be community specific.

Furthermore, most of the household productionliterature shows that local demands are not ho-mogenous. Women and men, landed and land-less households, farmers and those with off-farmemployment opportunities, the poor and the bet-ter-off all have different demands on the localforests. Therefore, communities characterized byboth of Kant’s conditions may be uncommonand successful applications of JFM may be spe-cial cases.

Finally, Johnson (1998) reminds us that sub-optimal rewards are obtained when the residualclaimant is not the party with greater ability toaffect forest use. In Johnson’s example fromHonduras as well as India’s JFM, the govern-ment claims the residual timber but local usershave greater impact on the forest. In Honduras,local resin tapping delays timber growth to ma-turity and delays the financial returns expectedby the government.18 This is an additional rea-son to anticipate that successful joint public agen-cy and local community management operationsmay occur only infrequently.

Some recommendations for community for-estry, and even for private forest management,overcome these problems. They recommend ei-ther permanent transfers or long-term agreementswith private contractors or local citizen groups –who have full rights rather than partial rightsshared with the government agency, as in JFM.The argument for the transfer of rights to com-munities is that governments, as absentee land-lords, have difficulty enforcing their rules or intaking advantage of locally specific land use op-portunities. Local managers are better at both.Furthermore, if local groups managing the forestas a commons have a cost advantage over pri-vate individuals it is in protecting the forest forpreferred common uses. (Their cost advantage isthe equivalent of a reduction in the propertyrights cost function in our figures. The cost re-duction extends the area of sustainable agricul-ture or plantation forest management and de-creases the region of open access depleted natu-ral forest.)

There are also important cautions for expecta-tions of successful community forestry. First,the community management solution may not bepermanent. As development proceeds, local rel-ative values may be modified and some other


309

lower cost management arrangement may re-place the original community management. Thisis not a problem as long as the forest ministryexpects and permits modifications in the con-tractual arrangements over time.

Second, some forest ministries impose seriousrestrictions on community forest activities. Eachrestriction only offsets the cost savings that trans-fers to local use were designed to create. Philip-pine plans for Community Based Forest Man-agement (CBFM), for example, once requiredlocal hire of a professional forest manager, aforest inventory and plan, several steps of publicand government agency oversight, restrictionson shipments of forest products beyond provin-cial boundaries, and payment of 44 percent ofgross forest revenues to the ministry (Hyde et al.1997). The Philippines’ more recent success withCBFM has only begun after many of these re-strictions were simplified or eliminated. Currentdiscussions of transfers from the ministry to com-munity management in Colombia, Nepal, andZambia all include similar restrictions. We mustrecall that one reason for the transfers to com-munity management is the ministry’s prior in-ability to enforce rules like these. The ministriesmust minimize their restrictions on communitymanagement and focus their own efforts on en-forcing simple rules for the remaining ministryobjectives.

6.3.3 The Long-term Policy Environment

We initially identified a second category of un-certainty, an uncertain policy environment. Con-flict in the contract between the formal owner(generally the ministry) and the community groupwith use rights is not the only source of uncer-tainty. Uncertainty can also arise from anynumber of external sources that raise doubts aboutcontinued land use. Consider a couple examples.

Deacon (1994), with data from 120 countries,observed that rates of deforestation increase inunsettled political environments; for example, inregions characterized by substantial political tur-moil, military governments, or guerrilla activity.One explanation for Deacon’s observations wouldbe that, in unsettled political environments, thosewith rights to commercial forests perceive a risk

of loosing their claims on these forest resources.Therefore, they claim the market reward to theirrights while the rights are still in their posses-sion. Those without rights may also feel thatunsettled times provide them with greater oppor-tunity for trespass and theft with impunity.

Yin and Newman (1997, updated and recon-firmed by Zhang, Uusivuori and Kuuluvainen1999) provided confirming evidence of short-term maximizing behavior for forest manage-ment under political uncertainty. They examinedtimber harvests and reforestation in two regionsof China after individual farmers in communeswere given long-term land use contracts. In oneregion the new contractual rights were adminis-tered consistently and the lands were fully redis-tributed within six years. In this case, farmersharvested timber and invested in reforestation.Authorities in the second region first distributed,then recalled, contracts for some forestlands.Despite a 15-year upward trend in land underlong-term private contracts and despite risingtimber prices, farmers who held contracts in thissecond region perceived the uncertainty in thelocal authorities’ decision making, and they har-vested without reforesting.

In our own experience, land managers in thesouthern Philippine island of Mindanao have beenslow to invest in forestry: i) despite rapid growthin long-term capital investment in other sectorsof the local and national economies (indicatinggood macroeconomic conditions for forestry in-vestment), and ii) despite evidence of forest plan-tations in neighboring Indonesia (indicating goodbiological and external market conditions for for-est investment). We would hypothesize that thedifference (until recently) has been a stable poli-cy environment in both Indonesia and in othersectors of the Philippine economy but a Philip-pine policy of ever-changing environmental rulesfor the forest industry.

In each of these cases uncertainty made a riskyprospect of land use and the rights to the stand-ing forest resource or to new investments in for-estry. The manager’s uncertainty could be due touncertainty in the broader economy, an unpre-dictable local political environment, or uncer-tainty specific to the prevailing rules for forest-ry.19 In all cases, the source of the uncertainty isexternal to the behavior of the land manager, and


310

in all cases it induces short-term unsustainabledecisions characterized by resource depletion anddeforestation without reinvestment. This behav-ior is identical to our expectations for behavioron open access forestland without any legitimateand enforceable claims. Removing the basis forthe uncertainty effectively secures tenure for moreland and forest resources and permits the landmanager to obtain the rewards from investmentsin lands and forests. Therefore, removing theuncertainty lowers the cost curve for propertyrights and increases the likelihood of sustainableactivities.

7 Conclusions

In conclusion, forestry, forest depletion, and thescarcity of fuelwood and other forest productsconsumed by rural households are not the prob-lems many anticipated when Eckholm (1976)called attention to these issues 25 years ago. Wenow have good empirical evidence that peopleadjust to changes in forest cover. Rural house-holds all over the world display a large amountof flexibility in their consumption and produc-tion decisions for forest products and services.

Rural households eventually take advantageof opportunities to substitute for their consump-tion of forest products; for example, substitutingagricultural residues and improved stoves forsome fuelwood. In the event of real scarcity,however, their initial adjustments tend to occuron the production side, and household labor – orthe time households devote to collecting forestproducts – is as important as the physical scarci-ty of the forest resource. In particular, the oppor-tunities households have to collect while jointlypursuing other household activities or to collectduring seasons of slack labor use provide flexi-bility for many forest production activities.

As scarcity increases yet further, farm house-holds do respond by planting trees. They treattrees as another crop and many of the experienc-es of subsistence agriculture become relevant forforestry. There are constraints on agroforestryactivities, however. Land tenure and governmentpolicy can be critical deterrents. The pooresthouseholds often have no land to plant trees.And those households with land are reluctant to

plant unless their tenure is certain for a period atleast as long as the productive life of a tree crop.Where land and its tenure pose no problem, thengovernment policy often acts as a deterrent totree planting and management, sometimes creat-ing an uncertain overall environment for invest-ment of any sort, sometimes encouraging agri-culture or some other land use in preference toforestry, and sometimes directly discouragingforestry itself by restricting the price or move-ment of forest products. Uncertain tenure anddetrimental public policies reduce household ex-pectations for forest product values, decrease thevalue of land in forest cover, and decrease thearea of sustainable forestry.

We must recognize that even when the condi-tions of land tenure and government policy areentirely favorable, sustainable forestry will neverbe a universal phenomenon. The very best poli-cy cannot make it so because the costs of estab-lishing and maintaining secure property rightswill always exceed the value of the standingforest resource on some frontier lands. Efficienthousehold behavior will always treat lands de-scribed by this condition as open access resourc-es and exploit the forests on these lands unsus-tainably. And for lands without pervasive posi-tive externalities, efficient government policy willnot alter this land allocation.

Nevertheless, the depletion of forest cover andthe increasing relative prices of forest productsthat we observe in numerous regions around theworld create an expanding opportunity for socialforesty. Households and communities in a broadrange of geographic and demographic situationshave responded by planting and managing theirown trees. The forest management responses ofmany farmers add up and the total impact on acountry’s forest cover can be great, despite thelocal limitations of many markets for forest prod-ucts. Markets are often local because forest prod-ucts are generally bulky and perishable and alsobecause roads and other regional infrastructureare not well-developed.

If we accept this description of rural house-holds and the rural environment, then forestershave a critical role. They can learn i) where thelocal relative prices are great enough to inducethe introduction of social forestry technologiesand ii) where the local experience with these


311

technologies is limited. Foresters have a role asagents of technology transfer in these locations.The regions, target populations within a region,and even the sites and institutional arrangementsfor forestry assistance require careful selectionfor the foresters to be effective. Forest ministriesand international donors can support this tech-nology transfer role. They have two additionalroles: i) identifying detrimental tenurial arrange-ments and unfavorable government policies andadvocating for their change, and ii) identifyingand protecting areas of significant external so-cial value within the regions of open access andunsustainable forest depletion. Altogether, thetree planting and management activities of ruralhouseholds, technology transfer activities of for-esters, and policy and protection activities ofministries and donors can improve the socialwelfare of rural households and increase the re-gions of sustainable land use. They must be se-lective and well-designed, however, because scar-city of forests and forest products is distinctlynot a universal phenomenon.

Notes

1 Environmental accounting incorporates the value of non-market environmental and natural resource services in acountry’s system of national accounts. See Hyde andAmacher (1996) for a review of forest values in environ-mental accounts.

2 Dan Bromley and many others with connections to theUniversity of Wisconsin were early leaders in the discus-sion of this issue. Jeff Romm and more recently JeffCampbell of the Ford Foundation have encouraged theidea of formal local participation in what traditionallyhave been public or state forest properties. Many goodscholars have developed the original themes of the Wis-consin and Ford insights.

3 Indeed, the initial conceptualizers of the dual economywere clear on this point. See Boeke (1948, 1953) andFurnival (1939, 1948), or see Ginsberg (1973) for areview. Boeke and Furnival featured households andpopulations that are predominantly subsistence oriented,but which always include at least occasional marketparticipation. Neither Boeke’s and Furnival’s householdsnor their communities relied totally on subsistence pro-duction without any market exchange.

4 Wages appear in the demand functions of householdswhose production and consumption decisions are non-separable (Singh, Squire and Strauss 1986). Some of thehousehold forestry analyses calculate a shadow wagefrom the marginal product of labor in fuelwood collec-tion.

Wage responses that are more elastic than price responseswould indicate greater flexibility in household labordecisions than in household consumption decisions. Thisis a reasonable finding in regions with remaining openaccess forest resources, seasonally slack labor, and forforest products that can be stored. The relative effects ofwages and prices are a theme that reappears throughoutour paper.

5 The income effect is small in all six studies that estimateits magnitude. Whether it becomes larger or smaller withan increase in income remains uncertain. Bogahawatteand AHJ observe a switch from a small negative to a smallpositive income effect for various forest products asincomes rise – although their statistical reliability is notgreat. A declining income effect, indicating an inferiorgood, would suggest that forestry activities will becomeless important as regions develop.

6 AHJ’s evidence contrasts with prior observations of manydevelopment practitioners that improved stoves and in-creased fuelwood consumption are correlates. In fact,stoves and consumption are correlates even with AHJ’sdata, but multi-variate techniques that separate the effectsof price, income, and improved stove on fuelwood con-sumption reveal that stoves are technological substitutesfor fuelwood.

7 In contrast, HAS report that only women and childrencollect. It is not clear whether their survey inclluded amale collection response. Collection did increase with thenumber of men in HAS’s households in Rajasthan.

8 Mekonnen and AHJ reinforce Cooke’s observation onjoint production with another example. They both ob-served that children spend time collecting forest productsbut their contribution to total collection is negative. Thissuggests that when women take children to the forest, theyexpect to produce something more than forest products.Perhaps instruction and child rearing are joint productswith collection from the natural forest. HAS added a thirdjoint production possibility: collection while tending graz-ing livestock.

9 Sajise (1998), with evidence from the Philippines, isexamining whether the income distributive incentives fornew investments are more complex than yet recognized.The issues of income source and level will remain impor-tant in our review. We will revisit them yet again in ourdiscussions of deforestation, property rights, and policy.

10 This observation is consistent with previous observationsthat fuelwood consumption and production are morewage than price responsive; in particular with AHK’sobservation that poor households in Nepal are the mostwage responsive in fuelwood production.

11 Dewees and PPJ disagree in one important respect. Dew-ees believes farmers invest in trees because imperfectmarkets limit their other opportunities. Relying on evi-dence from the same region of Kenya, PPJ argue that oncethe erosion control contributions of trees are accountedfor, then the local markets appear efficient.

12 Sajise is also examining the rate and timing of adoption,critical questions both for endangered environments andfor estimates of the economic potential of long-termtechnology transfers. He points out that while the litera-ture we have reviewed focuses on the decision of whetherto invest, the investment timing decision can be evenmore important. That is, even if prices are high enough tojustify investment, three characteristics of investments


312

can explain landowner decisions to delay: i) their uncer-tain nature, ii) the sunk cost and, therefore, irreversiblenature of investments in forestry, and iii) the fact thatinvestments can always be postponed. The many exam-ples of government policies that restrict private harvestsor require licensing or other permits before transportationof privately produced forest products only enhance thesecond characteristic and reinforce the option to delay.

13 See Berck (1979) or Johnson and Libecap (1980) forexamples from US history. Stone’s (1998) assessment forBrazil’s Paragominas is an especially clear example for adeveloping country and a modern context.

14 A reviewer pointed out that more evidence on the effi-ciency of local markets is always welcome. The revieweris correct, of course. Most of the literature we havereviewed, however, seems to suggest that the existingmarkets are more instructive than often anticipated. There-fore, we prefer to highlight these other three shortcomingsin the literature.

15 Chomitz’ work is all the more important because itsintensive use of GIS data does help overcome the problem(Chomitz and Gray 1996, Chomitz and Griffiths 1997).

16 Indeed, in some cases these relationships become verycomplex. For example, increasing the forest cover nearstationary water can improve the habitat for mosquitosand increase the prevalence of malaria. On the other hand,one species, Azadirachta indic, or the neem tree, producesseeds and a chemical discharge that may be mosquitodeterrents. Rural households in parts of East Africa andIndia plant this species near the household for that pur-pose.

17 For agriculture, Feder et al. (1988), with an example fromThailand, measured the productivity gains from titleregistration and showed that these are sufficient to coverthe costs of the government titling activity. For sub-Saharan economies with a briefer formal history, Migot-Adholla et al. (1991) concluded that the informal systemworks well and there are few gains from establishingformal title.

18 This is reminiscent of taungya, an agroforestry systemthat originated in Burma’s hills a century ago. Taungyacombines government agency objectives for a high-val-ued timber plantation crop with subsistence agriculturalobjectives. In exchange for protection of newly estab-lished forest plantations, the agency allows the localcommunity the rights to produce agricultural crops on theplantation until the young tree growth shades out theagricultural crops (Nair 1991). Of course, taungya oftenfails because the community quickly perceives its ownincentives to maintain the agricultural rights by deterringthe forest growth.

19 An unstable macroeconomy has the same effect becausemacroeconomic instability makes long-term rewards un-certain and, therefore, deters long-term investments. Wecan observe relevant examples in the Russian forestrysector since the early 1990s and in Indonesia since thebeginning of the East Asian financial crisis in July 1997.

References

Alston, L., Libecap, G. & Mueller, B. 1998. Titles,conflict, and land use: the development of proper-ty rights and land reform on the Brazilian frontier.Ann Arbor: University of Michigan Press.

Amacher, G., Hyde, W. & Joshee, B. 1992. The adop-tion of consumption technologies under uncer-tainty: a case of improved stoves. Journal of Eco-nomic Development 17(2): 93–105.

— , Hyde, W. & Joshee, B. 1993. Joint productionand consumption in traditional households: fuel-wood and crop residues in two districts of Nepal.Journal of Development Studies 30(1): 206–225.

— , Hyde, W. & Kanel, K. 1999. Nepali fuelwoodproduction and consumption: regional and house-hold distinctions, substitution, and successful in-tervention. Journal of Development Studies 35(4):138–163.

— , Hyde, W. & Rafiq, M. 1993. Local adoption ofnew forestry technologies: an example from Paki-stan’s Northwest Frontier Province. World Devel-opment 21(3): 445–453.

Berck, P. 1979. The economics of timber: a renewableresource in the long run. Bell Journal of Econom-ics 10(2): 447–462.

Binkley, C., Raper, C. & Washburn, C. 1996. Institu-tional ownership of US timberland. Journal ofForestry 94(9): 21–28.

Bluffstone, R. 1995. The effect of labor markets on de-forestation in developing countries under open ac-cess: an example from rural Nepal. Journal of En-vironmental Economics and Management 29(1):42–63.

Boeke, J. 1948. The interests of the voiceless Far East.Leiden: Universitare Pers.

— 1953. An economic policy of dual societies. NewYork: Insitute of Pacific Relations.

Bogahawatte, C. 1997. Non-timber forest products andrural economy in conservation of wet zone forestsin Sri Lanka. Unpublished report to Economy andEnvironment Program for Southeast Asia. Singa-pore.

Bowles, A., Rice, R., Mittermier, R. & Fonseca, G.1998. Science 280: 1899–1900.

Byron, N. 1997. Income generation through commu-nity forestry. Unpubl. mss. Centre for Internation-al Forestry Research, Bogor, Indonesia.

Chomitz, K. & Gray, D. 1996. Roads, land use, and: aspatial model applied to Belize. World Bank Eco-


313

nomic Review 10(3): 487–512.— & Griffiths, C. 1997. An economic analysis of

woodfuel management in the Sahel: the case ofChad. World Bank policy research working paper1788.

Cooke, P. 1998. The effect of environmental goodscarcity on own-farm labor allocation: the case ofagricultural households in rural Nepal. Environ-ment and Development Economics 3: 443–469.

— 1998a. Intrahousehold labor allocation responsesto environmental good scarcity: a case study fromthe hills of Nepal. Economic Development andCultural Change 46: 807–830.

Deacon, R. 1994. Deforestation and the rule of law ina cross section of countries. Land Economics 70(4):414–440.

Dewees, P. 1989. The woodfuel crises reconsidered:observations on the dynamics of abundance andscarcity. World Development 17: 1159–1172.

— 1995. Trees on farms in Malawi: private invest-ment, public policy, and farmer choice. WorldDevelopment 23: 1085–1102.

— & Saxena, N. 1995. Wood products markets asincentives for farmer tree growing. In: Arnold, J.& Dewees, P. (eds.). Tree management in farmerstrategies. Oxford University Press.

Dickinson, H. 1969. Von Thunen’s economics. Econ-omic Journal 79: 894–902.

Douglas, J.J. 1982. Consumption and supply of woodand bamboo in Bangladesh. Field document 2,UNDP/FAO project BGD/78/010, BangladeshPlanning Commission, Dhaka, Bangladesh.

Eckholm, E.P. 1976. Losing ground: environmentalstress and world food prospects. W.W. Norton &Company.

Feder, G., Onchan, T., Chalamwong, Y. & Hongla-darom, C. 1988. Land policies and farm produc-tivity in Thailand. Baltimore: Johns Hopkins.

Foster, A., Rosenzweig, M. & Behrman, J. 1997. Pop-ulation and deforestation: management of villagecommon land in India. Draft mss., Department ofEconomics, University of Pennsylvania.

Furnival, J. 1939. Netherlands India: a study of ruraleconomy. Cambridge University Press.

— 1948. Colonial policy and practice: a comparativestudy of Burma and Netherlands India. CambridgeUniversity Press.

Ginsberg, N. 1973. From colonialism to national de-velopment: geographical perspectives on patternsand policies. Annals of the Association of Geo-

graphers 63(1): 1–21.Godoy, R. 1992. Determinants of smallholder com-

mercial tree cultivation. World Development 20(5):713–725.

Heltberg, R., Arndt, C. & Sekhar, N. 1998. Fuelwoodconsumption and forest degradation: a householdmodel for domestic energy consumption in ruralIndia. Draft mss., Institute of Economics, Univer-sity of Copenhagen.

Hofstad, O. 1997. Deforestation by charcoal supply toDar es Salaam. Journal of Environmental Eco-nomics and Management 33(1): 17–32.

Hyde, W. & Seve, J. 1993. The economic role ofwood products in tropical deforestation: the se-vere experience of Malawi. Forest Ecology andManagement 57(2): 283–300.

— & Amacher, G. 1996. Applications of environ-mental accounting and the new household econ-omics: new technical economic issues with a com-mon theme in forestry. Forest Ecology and Man-agement 83(3): 137–148.

— , Harker, B., Guiang, E., Dalmacio, M. & Harker,B. 1997. Forest charges and trusts: shared benefitswith a clear definition of responsibilities. Journalof Philippine Development XXIV(2): 223–256.

— , Krutilla, K., Barnes, D. & Xu, J. 1998. Theimportance of economic corrections for physicalstocks: a forestry example. Draft mss. ForestryDepartment, Virginia Tech.

IRG with Edgevale Associates. 1994. The Philippineenvironment and natural resource accountingproject (ENRAP phase II). IRG, Inc. Washington.

Jodha, N. 2000. Common property resources and thedynamics of rural poverty: field evidence from thedry regions of India. In: Hyde, W., Amacher, G.& colleagues. Forestry and rural development: anempirical introduction from Asia. Ann Arbor: Uni-versity of Michigan Press.

Johnson, R. 1988. Multiple products, community for-estry and contract design: the case of timber har-vesting and resin tapping in Honduras. Journal ofForest Economics 4(2): 127–145.

— Libecap, G. 1980. Efficient markets and GreatLakes timber: a conservation issue reexamined.Explorations of Economic History 17(3): 372–385.

Kant, S. 1996. The economic welfare of local commu-nities and optimal resource regimes for sustain-able forest management. Unpublished PhD thesis.University of Toronto.


314

Kapoor, R.P. 1992. Investment in afforestation: Trendsand prospects. In: Agarval, A. (ed). The price offorests: Proceedings of a Seminar on the Econom-ics of the Sustainable use of Forest Resources.New Delhi: Centre for Science and Environment.

Köhlin, G. 1998. The value of social forestry in Oris-sa, India. Unpublished PhD thesis. Department ofEconomics, Göteborg University.

— 1999. Supply of fuelwood in rural India: evidencefrom Orissa. Draft manuscript, Department ofEconomics, Göteborg University.

Lopez, R. 1998. The tragedy of the commons in Coted”Ivoire agriculture: empirical evidence and im-plications for evaluating trade policies. World BankEconomic Review 12(1): 105–132.

Mekonnen, A. 1998. Rural energy and afforestation:case studies from Ethiopia. Unpublished PhD the-sis, Department of Economics, Göteborg Univer-sity.

Migot-Adholla, S., Hazell, P., Barel, B. & Place, F.1991. Indigenous land rights systems in sub-Saha-ran Africa: a constraint on productivity? WorldBank Economic Review 5(1): 155–175.

Nair, P. 1991. An introduction to agroforestry. Dor-drecht: Kluwer.

Patel, S., Pinkney, T. & Jaeger, W. 1995. Smallholderwood production and population pressure in EastAfrica: evidence of an environmental Kuznet’scurve? Land Economics 71(4): 516–530.

Persson, R. 1998. Fuelwood: crisis or balance. Un-published draft mss, Centre for International For-estry Research, Bogor, Indonesia.

Rao, D.N. & Srivastava, M.K. 1992. Role of forests inthe energy, economy relationship. In: Agarvel, A.(ed.). The price of forests. New Delhi: Centre forScience and the Environment.

Sajise, A. 1998. Information, labor constraints and thetiming and intensity of agroforestry adoption. Draftmanuscript. Agriculture and Resource EconomicsDepartment, University of California, Berkeley.

Samuelson, P.A. 1983. Thunen at two hundred. Jour-nal of Economic Literature 21(4): 1468–1488.

Scherr, S. 1995. Economic factors in farmer adoptionof agroforestry: patterns observed in westernKenya. World Development 23(5): 787–804.

Schultz, T. 1964. Transforming traditional agricul-ture. University of Chicago Press.

Sharma, R., McGregor, M. & Blyth, J. 1991. Thesocioeconomic evaluation of social forestry in Oris-sa, India. The International Tree Crop Journal 7:41–56.

Shively, G. 1998. Economic policies and the environ-ment: tree planting on low income farms in thePhilippines. Environment and Development Econ-omics 3(1): 83–104.

Shyamsunder, P. & Kramer, R. 1996. Forest farmersin the tropics: An agricultural household model offorest resource use. Unpublished draft mss.

Singh, I., Squire, L. & Strauss, J. 1986. The basicmodel: theory, empirical results, and policy con-clusions. In: Singh, I., Squire, L. & Strauss, J.(eds.). Agricultural households models. Johns Hop-kins University Press.

Smith, J., van de Kop, P., Reategui, K., Lombardi, I.,Sabofal, C. & Diaz, A. 1998. Dynamics of sec-ondary forests in slash and burn farming: interac-tions among land use types in the Peruvian Ama-zon. Draft manuscript, Centre for InternationalForestry Research, Bogor, Indonesia.

Stone, D. 1998. The timber industry along an agingfrontier: the case of Paragominas (1990–1995).World Development 26(3): 443–448.

Sunderlin, W. & Wunder, S. 1998. Mineral exports,Dutch disease, agriculture, and forests: new in-sights on the variability of tropical deforestation.Draft mss., Centre for International Forestry Re-search, Bogor, Indonesia.

Thatcher, T., Lee, D. & Schelhas, J. 1997. Farmerparticipation in reforestation incentive programsin Costa Rica. Agroforestry Systems 35: 269–289.

Yin, R. & Newman, D. 1997. Rural reforms: the caseof the Chinese forestry sector. Environment andDevelopment Economics 2(3): 289–304.

— & Hyde, W. 1998. Trees as an agriculture sustain-ing resource. Agroforestry Systems (forthcoming).

Zhang, Y., Uusivuori, J. & Kuuluvainen, J. 1999.Impacts of economic reforms on rural forestry inChina. Draft manuscript, Forestry faculty, Uni-versity of Helsinki.

Total of 66 references

social forestry reconsidered - efd -...

Documents