0044-A1

What determines household dependency on community forests? Empirical evidence from Nepal

Bhim Adhikari ¹

Abstract

The relationship between household socio-economic attributes and forest dependency has become a growing concern in issues of local level collective action. On one hand, it has been argued that poor people extract more resources from the communal forest due to their greater reliance on natural resources. On the other hand, it is claimed that, compared to the less poor, the poor may depend more on the commons in relative terms, but in absolute terms their dependency is lower. In this study I advance this argument in order to understand the relationship between household socio-economic characteristics and forest dependency in reference to community-based forest management in Nepal. The analysis is based on field data from 309 households from the mid-hills of the country. Empirical analyses suggest that household labour allocation decisions for forest product collection are dictated by various socio-economic and demographic variables. In general, it appears that household land and livestock holdings, gender, caste and average education of family members exert more influence on forest product collection and gathering activities from community forests (CFs). Based on this analysis, it can be concluded that households with less productive assets are currently facing more limited and restricted access to CFs than relatively better-off households. This is mainly due to the fact that existing management regimes for CFs are more oriented towards production of intermediate forestry products that serve the interests of "less poor" households and disregard the concerns of the "poor" who are more dependent on non-timber forest product (NTFPs) to secure their livelihoods. Policy measures that aim to promote equity of resource distribution, NTFP-oriented management regimes, and direct involvement of women and marginalized sections of the community in the overall process of community forestry, help to increase the access of poorer households to CFs where forest-based livelihoods are a pervasive feature of the rural economy.

1. Introduction

Management of the local commons is critical important in rural areas of developing countries since it contains significant biodiversity and other natural resources that contribute economic well-being of rural population and environmental conservation. The recognition of community-based resource management has led to the devolution of natural resource management from centralized government control to local people in Nepal and other South Asian countries in recent years. Devolution of forest management has been underway in Nepal since the 1990s when national forests were handed over to forest user groups (FUGs) as community forests (CF). The Government has been issuing policy initiatives to encourage participation of rural households in order to strengthen community-based institutions for the control and sustainable management of forest resources. FUGs are granted usufruct rights to forests through a legal enactment and are being encouraged to become independent and self-governing organizations. They are fully involved in preparing operational plans, harvesting, and sharing of benefits from their CF. To date more than 11,000 FUGs are managing about 848,159 hectares of CF throughout the country (HMGN/CPFD 2001). For the last two decades, FUGs have proliferated throughout the country and concept of Nepal's CF program itself has been even exported to different parts of the world (Nightangle, 2002).

Although local control over natural resources is now regarded as a win-win solution for environmental preservation and the local level economic development, the empirical evidence regarding the dependency of rural households on CF is rather mixed. Some recent studies indicate that though CF management has brought significant positive impacts on the biophysical condition of forests, equitable access to and use of forest resources within the community has not been clearly demonstrated (Branney and Yadav, 1998; Malla 2000). Researchers argue that CF programs in Nepal are so far primarily motivated by production of timber and intermediate forest products² (Richards et al., 1999) that can be utilized by households with large land and livestock holdings, consequently marginalizing poorer households whose livelihood depends on NTFPs. Significant research on the commons suggests that better-off group members are often likely to gain a larger share of benefits from a common resource (Agrawal, 2001, Adhikari, 2002). This is due to the fact that decisions with regard to biomass collection from the commons are directly related to household productive assets. Since the level of wealth of individual users affects the extent of resource exploitation and appropriation, labour allocation decisions of small holders in developing countries are largely decided their socio-economic status. Individual endowment play a complex role creating on demands on and alternative to biomass use, while also creating opportunities for more efficient or productive resource utilization for some social groups (Lele, 1997). Though there is a strand of literature on dependency of rural households on biomass resources, to my knowledge, none of these studies explicitly examine the possible relationship between socio-economic attributes and dependency on the local commons.

In this paper I seek to provide an answer to the following question: What are the socio-economic attributes of households that determine forest product collection and gathering activities from CF? In line with recent literature (Lele 1997), I hypothesised that demand for forest biomass is primarily a function of a) cultivated land holdings and crop types (driving demand for mulch and manure) b) livestock assets (driving demand for fodder and cut grass) c) the number of people in the household (for both fuel wood production and consumption) d) education, ethnicity and caste of household (caste of an individual influences cultural attitudes towards food, bathing and rituals which might drive demand for fuel wood), e) distance between forest and house (determines relative control as well as household behaviour regarding unauthorised harvesting); and e) other demographic variables. Understanding why dependence on forests differs across households is important for both forest conservation and poverty alleviation program (Fisher et al., 2002).

2. Study Site, Data Collection and Survey Methods

This study was undertaken in the two selected districts of the mid-hills of Nepal where CF programs have been implemented for the last two decades. From the lists of user groups in these two districts, Eight FUGs³ were selected for the study. A stratified sample of households was chosen by compiling a census of village households. All user households were divided into three different groups i.e. poor, middle wealth and richer households based on criteria⁴ that villagers think important while assessing an individual's socio-economic position in the village. Household questionnaires were designed to elicit forest use information from the respondent households. Pre-testing of the questionnaire was undertaken in one of the FUGs in Nasikathan VDC in Kavre district, which resides outside the sample frame. The response to pre-testing of the questionnaire resulted in the revision of sensitive questions such as gender bias in forest product extraction. Since both on-farm and off-farm agricultural activities are seasonal, care was taken to consider allocation of family labour seasonally. The data on biomass use are a combination of the survey questionnaire data and monitoring of actual use carried out by the research team in the study sites. This was elicited mainly from the structured questionnaire (recall) survey. A total of 20 per cent of households were randomly selected from each stakeholder group, with a total of 309 households belonging to three different wealth groups. Finally an econometric analysis was undertaken in order to understand the determinants of forest product collection from community forests

3. Results and Discussion

3.1 Household labour allocations

Household labour allocations in four areas included in the analytical model (firewood, tree fodder, grass fodder and leaf litter collection) are presented in Table 1. The table reveals that poorer households on average allocated a total of 116, 16, 178, and 328 hours annually for fuel wood, tree fodder, leaf litter and grass fodder collection respectively. Richer households on average allocate 144, 29, 871, and 960 hours annually in collecting these products. From the table it is evident that forest products i.e. tree and grass fodder and leaf litter are highly wealth sensitive. It appears that households with larger endowments extract more intermediate forest products than poorer households in the study area.

Table 1 Labour allocation for forest product collection (Hour/annum)

3.2 Determinants of labour allocation for forest product collection

Table 2 presents the regression analysis of various forest product collection from CF. The first column of the table 2 present determinates of fuel wood collection. First, caste is negatively associated to the fuel wood collection. This indicates that lower caste households extract less firewood from the commons. Secondly, fuel wood collection shows an interesting gender pattern as female-headed households extract less from CF than that of male-headed households. This result is similar to Kohlin (1998) who observed that males contribute significantly to fuel wood collection than females. However, this is in contrast with traditional view that producing energy is mainly a female activity. This observation is also similar to that of Amacher et al. (1993), who observed that women are not the sole collectors of fuel wood.

Table 2 Determinants of labour allocation for forest product collection from CF

Thirdly, average education of family members shows a negative relationship with fuel wood collection from the commons as increasing educational level makes fuel wood collection increasingly unprofitable. This finding is similar to Gunatilake (1998), who concludes that education level of the family is negatively related to forest dependency. Fourthly, though not significant, it appears that those households involving in various forms of decision-making activities (measured by transaction costs days spent in such activities) produce more fuel wood. This might be due to the information gained through various forms of community meetings about when to collect and where to collect fuel wood from CF. Lastly, technology used in forest products harvesting is significantly and positively associated with fuel wood production since advanced technology makes household labour more productive and marginal productivity of labour increases with advanced technology used in collection. The coefficient of distance to the forest is positive and significant, which indicates that household are willing to travel as far as possible to collect fuel wood since there are no other substitutes for fire wood.

The second column of Table 2 presents the model of leaf fodder and cut grass collection effort from CF. It appears that households with large numbers of livestock spent a lot of time on fodder collection effort. Landholding is also significantly associated with grass and tree fodder collection effort. Caste has a negative effect though not significant. This may be due to the fact lower caste households do not have more animals they extract less from CF. As discussed earlier, gender is positively associated with fodder and cut grass collection and the coefficient is highly significant. This indicates that male-headed households have higher access to tree fodder and cut grass than female-headed households. This is also contrary to traditional thinking that female members spend more time on grass collection. Women also have other household responsibilities that might affect gathering activities. With regard to the `transaction costs day' variable, households who spent a lot of time on decision-making and implementation activities often appear to spend more time on fodder collection effort. Households get correct information by engaging in the decision-making process about when and where to collect. The numbers of people in a family has positive impacts on fodder and grass collection though not significant in the model. These findings are similar to Heltberg (2001), who concluded that a larger family implies more labour available for fodder collection.

Number of trees on private land does not have negative impact on fodder and grass collection from community forest. It was expected that households that have planted a large number of fodder trees on their private land would spend less time on fodder collection in CF. It appears that most of the farmers have started tree planting only after the introduction of CF, which used to distribute free tree seedlings to the local farmers. I observed that these trees are still very young and not at a stage of providing leaf fodder. The positive sign of trees planted on private land in this production function is therefore not surprising. Technology of harvesting is still positively related to amount of tree fodder and grass collection effort. More importantly, the variable "pine forest" is negatively and significantly related to fodder and grass collection. Villagers generally perceive conifer trees to be less useful since they are not a good source of animal fodder. Average education of family members has again negative effect on tree and grass fodder collection.

The third column of Table 2 shows the estimates of household leaf litter collection from community forests. The positive signs of land and livestock holding indicate, once more, that livestock and land apparently act as major sources for residue consumption. This is particularly important in the mid-hills where leaf litter is a major source of compost. Gender and caste once again show negative and significant relationships with leaf litter collection. Household participation in various stages of meetings and decision-making through transaction costs days spent in CF enhances the amount of leaf litter collection. Technology used in leaf litter collection is positively and significantly associated with consumption of leaf litter.

Distance has a negative, albeit insignificant, impact on leaf litter collection. This implies those households residing adjoining to the forest area collect more leaf matter than those living far away from the CF. In many forest resource systems, users who live closer to the forest have a more secure and accessible supply of produce regardless of whether or not there are allocation rules in place (Varughese and Ostrom 2001). Families living close to the forest have the advantage of less time being required to reach a particular forest resource. Their links with forests are, therefore, expected to be high (Gunatilake 1998). Education has negative and significant effect on leaf litter collection. The reason is already mentioned earlier. It is evident that all labour time allocated for product collection (fuel wood, tree and grass fodder and leaf litter) is positively and significantly related to quantity of harvesting or collection.

4. Conclusions and policy implications

The relationship between socio-economic attributes of user households and dependency on local forests is getting more attention in recent years in issues of local level collective action. The main hypothesis in this paper was that access to common property forests is almost inextricably associated with household characteristics, which influence the household labour allocation decisions and nature and extent of resource appropriation and exploitation. As I hypothesised, empirical analysis suggests that labour allocation decisions for forest extraction activities are functions of various socio-economic and demographic variables like; land and livestock holding, gender, caste, average education of family members, and some other local differences. The results show that women are not sole collectors of fuel wood since access to community forestry is somehow influenced by gender of respondents. Indeed, household access to CF is different for households headed by male and female members as male-headed households have more access to CF. In general, it appears that household land and livestock holdings, household composition, gender, caste, education, leadership quality, and technology of harvesting exert more influence on household labour allocation decisions for extraction and gathering activities than other factors.

Though conservation oriented management regimes improve biophysical aspects of local commons, one can expect a very little contribution of such regimes to the income of poorer households. Policy measures to this end include forest management system that aim to produce non-timber forest products, empowerment of women and politically marginalised users, more representation of weaker sections of community in the decision-making process, and distribution of leadership roles to all community members. Since patterns of forest use differ among rural households, CF policy in this respect should be directed towards diversifying the products that meet the demand of different interest groups within the community. Arnold and Perez (2001 p. 445) point out that, `it may often be necessary in designing and implementing policy and other institutional interventions to distinguish between those who can improve their livelihoods through NTFP activities, and those who have no other option but to continue to gather NTFPs in order to survive'. Since poor people do not get substantial benefits from agricultural related forest products, forest management policy needs to be directed at increasing alternative forest products, mainly NTFPs that play a significant role in supplying livelihood needs of the poor.

Acknowledgement

The research received generous financial support from the South Asian Network for Development and Environmental Economics (SANDEE). Thanks are due to Jon Lovett, Doriana Delfino, Charles Perrings, Priya Shyamsunder, Manju Bhattarai and Claire Quinn. Comments on earlier version of this paper from participants of SANDEE's research and training workshop held in AIT Center, Bangkok is highly acknowledged.

REFERENCES

Adhikari, B., 2002. Household Characteristics and Common Property Forest Use: Complementarities and Contradictions. Journal of Forestry and Livelihoods 2 (1) 3-14.

Agrawal, A., 2001. Commons Property Institutions and Sustainable Governance of Resources. World Development 29 (10): 1649-1672.

Amacher, G., W. Hyde and B. Joshee, 1993. Joint Production and Consumption in Traditional Households: Fuel Wood and Crop Residues in Two Districts in Nepal. Journal of Development Studies 30 (1): 206-225.

Arnold, J.E.M. and M.R. Perez, 2001. Can Non-Timber Forest Products Match Tropical Forest Conservation and Development Objectives? Ecological Economics 39: 437-447.

Branney P. and K. P. Yadav, 1998. Changes in Community Forests Condition and Management 1994-1998: Analysis of Information from the Forest Resource Assessment Study and Socio-Economic Study in the Koshi Hills. Project Report G/NUKCFP/32. NUKCFP, Kathmandu.

Fisher, M., G. Shively and Steven Buccola, 2002. Activity Choice, Labour Allocation, and Forest Use in Malawi. Paper Presented at the Second World Congress of Environment and Resource Economists, Monterey 24-27 June, California, USA.

Gunatilake, H., 1998. The Role of Rural Development in Protecting Tropical Rainforests: Evidence from Sri Lanka. Journal of Environmental Management 53: 273-292.

Heltberg, R., 2001. Determinants and Impact of Local Institutions for Common Resource Management. Environment and Development Economics 6: 183-208.

HMGN/CPFD Database (2001) Department of Forest, Kathmandu, Nepal.

Kohlin, G., 1998. The Value of Social Forestry in Orissa, India, Unpublished PhD Thesis.

Lele, S., 1997. Biomass Resource Management: Household Characteristics and Forest Rights: Complementarities and Contradictions. The Beijer International Institute of Ecological Economics, Research Seminar in Environmental Economics, 23-25 May 1997, Sabah.

Malla, Y.B., 2000. Impact of Community Forestry Policy on Rural Livelihood and Food Security in Nepal. Unasylva 51: 202.

Nightingale, A.J., 2002. Participating or just Sitting in? The Dynamics of Gender and Caste in Community Forestry. Journal of Forestry and Livelihood 2 (1): 17-24.

Richards, M., K. Kanel, M. Maharjan and J. Davies, 1999. Towards Participatory Economic Analysis by Forest User Groups in Nepal. ODI, Portland House, Stag Place, London, UK.

Varughese, J. and E. Ostrom, 2001. The Contested Role of Heterogeneity in Collective Action: Some Evidence from Community Forestry in Nepal. World Development 29 (5): 747-765.

¹ Environment Department, University of York, Heslington, York, YO10 5DD, England. [email protected];
Website: http://www2.york.ac.uk/depts/eeem/resource/adhikari/adhikari.htm

² Biomass input to household production system can be broadly categorised as tree woody biomass (for fuel wood, construction and fencing), grass and tree fodder biomass (fodder/forage), and shrub and tree leafy biomass (manure/mulch). Households might directly use only fuel wood and timber; all other biomass acts as inputs into the farming system. The term `intermediate product' refers to those products such as tree and grass fodder, leaf litter and other similar products, which are not directly consumed within households but they are an integral part of forestry-based farming system in Nepal. Households with different productive assets and income will have different needs of these products and different level of access to these resources. In another words, a household's use of intermediate products appears to be driven primarily by agricultural land holding and cattle ownership.

³ Saradadevi FUG, Jyala Chiti FUG, Mahavedsthan FUC and Thuli Ban FUG in Kavre districts and Gaurati FUG, Shree Chhap FUG, Janghare FUG and Karki Tar FUG in Sindhupalchowk district.

⁴ Main criteria used were the amount of land owned, the number of livestock owned, loans given and taken, and income from off-farm agricultural activities Poor households own 0 to 0.25 hectare of land, with a mean of 0.15 hectare. Middle-wealth households own between 0.28 hectare to 0.75 hectare with a mean of 0.51 hectare. Richer households own between 0.78 to 4.25 hectare, with a mean of 1.33 hectare.