Khai thác thương mại lâm sản ngoài gỗ và đời sống các hộ gia đình trong hoặc gần rừng. Nghiên cứu ở thôn Quế(Con Cuông) và thôn Ma(Tương Dương), Nghệ An, Việt Nam.In Vietnam, NTFPs has become an important source of cash income for local people living in or near the forests. The commercial collection of these products could reduce both the number of species and population of a species in the forests. In order to keep the balance between biodiversity and commercial collection of NTFPs, this paper evaluates the dependence of forest dwellers on NTFPs and identifies the relation between household characteristics and cash income generated by NTFP collection. As a result, commercial collection of NTFPs is negatively correlated to dependency ratio, poverty level and distance
SURVEY Commercial collection of NTFPs and households living in or near the forests: Case study in Que, Con Cuong and Ma, Tuong Duong, Nghe An, Vietnam Dang Viet Quang*, Tran Nam Anh Center for Agricultural Research and Ecological Studies, Hanoi Agricultural University, Trau Quy, Gia Lam, Hanoi, Vietnam ARTICLE INFO ABSTRACT Article history: Received 25 August 2005 Received in revised form 19 February 2006 Accepted 4 March 2006 Available online 27 April 2006 In Vietnam, NTFPs has become an important source of cash income for local people living in or near the forests. The commercial collection of these products could reduce both the number of species and population of a species in the forests. In order to keep the balance between biodiversity and comm ercial collection of NTFPs, this paper evaluates the dependence of forest dwellers on NTFPs and identifies the relation between household characteristics and cash income generated by NTFP collection. As a result, commercial collection of NTFPs is negatively correlated to dependency ratio, poverty level and distance to the provincial city, and positively correlated with female labors of households. Moreover, NTFPs are proved to be very important in poor households or in those who lack rice or high rate of female labor. The households who have higher dependency ratio benefit less from NTFPs sold, while those who lack rice for their own consumption or have a higher rate of female labor depend more on NTFPs. Poor households are more dependent on NTFP collection than other groups. Finally, the result of this study highly recommends that the Government ofVietnam be awareof gender balance, distanceto provincial city, poverty level and dependency ratio when applying a policy related to NTFP conservation. An appropriate incentive policy for woman and poor households might be a possible solution for less commercial collection of NTFPs. The poor households who have low dependency ratio or high female labor rate should be encouraged to participate in the forest management system to control the balance between commercial collection of NTFP and conservation. © 2006 Elsevier B.V. All rights reserved. Keywords: NTFPs Livelihood Commercial collection Dependence on NTFPs 1. Introduction Forest products, which can be classified into timber and non- timber products, are numerous and play an important role in the livelihood of people living in or near forests. People in such areas often collect forest products and use forest land for cultivation. They use forest products both for household consumption and cash income generation (Fisher, 2000). In subsistence economies, forests can provide many essential products and services for the life of local people such as “food, utensils, clothing, shelter, medicines and objects of spiritual or cultural significance” (Wong, 2000, pp. 3–4). In open economies where forest dwellers can sell or trade their products, forest products generate considerable employment and income. In ECOLOGICAL ECONOMICS 60 (2006) 65– 74 ⁎ Corresponding author. Fax: +84 4 8766642. E-mail addresses: quang74@hn.vnn.vn, quang74@cares.org.vn (D. Viet Quang). 0921-8009/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolecon.2006.03.010 available at www.sciencedirect.com www.elsevier.com/locate/ecolecon India, for example, non-timber forest products generate US $700 million annually in Madhya Pradesh and US$115 million annually in Maharashtra, while commercial wood provides US $72 million in Madhya Pradesh and US$29 million in Mahar- ashtra (Osman et al., 2000). Furthermore, forest products are the center of research on forest management, biodiversity, conservation, and poverty alleviation (Lawrence, 2003; Ambrose-Oji, 2003). Non-timber forest products (NTFPs) are diverse and consist of an assortment of products. The number of NTFPs might be larger than that of timber-based products because the NTFPs come from many different parts of the plants (Chamberlein III, 2000). Due to the diversity of NTFPs, they are extremely important to biodiversity, conservation, and forest manage- ment. NTFPs, therefore, have attracted the attention of many researchers all over the world. In 1999, at the Conference on Forest Communities in the Third Millennium: Linking Re- search, Business, and Policy toward a Sustainable Non-Timber Forest Product Sector, Iain Davidson-Hunt reviewed literature on NTFPs and divided it into seven categories, those are: “ethnographic studies, traditional ecological knowledge, eco- nomic botany, forest management and policy, biology and ecology of forests, forest products research and business organization and marketing”. The definition of NTFPs in this conference was extremely broad including foods, materials, health care, decoration, environmental products, etc. (David- son-Hunt et al., 2001, pp. 1–3). More specifically, socioeconomic research reveals that NTFPs become important in the livelihoods of many poor households who live in or near forests, especially in the tropics. Many of these studies find that the livelihoods of poor households still depend on NTFPs from fallow 1 and natural forest 2 relatively more than the other groups (Sills et al., 2003; Tickin, 2004; Osman et al., 2000; Roderick and Hirsch, 2000; Belcher and Kusters, 2004). The dependence of farmers on NTFPs mostly differs from this region to others. “In remote areas, NTFPs provide subsistence goods like food, medicines and building materials”. Near urban areas, where trading and commercial networks perform better with the environmental functions of the forests, the farmers have more options. They can find a job related toNTFPs or can produceNTFPs mainly for commercialization (Ros-Tonen and Wiersum, 2003, p. 17). Most farmers collect NTFPs both for subsistence and for trading in order to mitigate the risks as well as to diversify their income sources (Subhrendu and Sills, 2001). Nevertheless, if farmers can get high profitability from NTFPs, their harvesting could reduce both the number of species and population of a species in the forests. Roderick's work showed “a positive relation- ship between low biodiversity and high NTFP profitability” (Roderick and Hirsch, 2000, pp. 124). For those reasons, in Southern and East Africa, the management, roles and depen- dence of different groups on non-timber forest products were given the highest priority in research (Ruiz et al., 1997). In general, NTFPs play an important role both in forest conservation and improving livelihoods for forest dwellers. They can provide the incentive for participatory forest management (Ros-Tonen, 2000). Additionally, research of Wendy Cocksedge in 2001 indicated that the use of NTFPs is a possible solution to release the dependency of local people on timber. Thus, they can be a sustainable source of income for people living in or near the forests (Cocksedge, 2001). Moreover, research in South-west Cameroon shows that the value of forests will be improved by increasing the use of NTFPs, if it can incite users to enga ge with improved management and conservation strategies (Ambrose-Oji, 2003). Other crucial important roles of NTFPs are “poverty avoidance, filling gaps during periods of low income, spreading risk, and functioning as a safety net” (Belcher and Kusters, 2004, pp. 1–22). Thus, the idea integrated from this literature review is that the collection of NTFPs should be controlled in relation with forest conservation by a partici- patory management system. However, which type of house- hold should participate in this management system is still a question for both researchers and policy makers in Vietnam. This question can be answered by identifying the relation between households living in or near the forests and NTFPs collection. In Vietnam, more than 24 million people live in and near forests, accounting for 30% of country's population (Beer, 1993). Similar to other countries, there is no doubt that NTFPs can fulfill several functions. In the households which live in or near the forests of Vietnam, NTFPs can provide food, materials for construction, fodder, fuel-wood, and cash income (Beer, 1993). A case study of dependency on forest and tree products for food security in the Northern Mountainous Region of Vietnam shows that wealthier groups of forest dwellers use forest vegetables for their own consumption or buy them from the poor. The poor sell vegetables, bamboo shoots and mushrooms from the forest and use the money to buy rice. The wealthier families hunt to improve the quality of their meals while poorer families hunt to earn more income (Yen et al., 1994, cited by Rijsoort, 2000). There is little doubt that NTFPs are very important for household livelihoods and for conservation of biodiversity in the forests. Meanwhile, research on the importance of NTFPs to either of these themes is rarely done in Vietnam. On the other hand, in order to keep the balance between conservation and NTFP collection, the relation between households and commercial collection of NTFPs should be identified first. This research, therefore, aims to analyze the dependence of households on NTFPs (measured as a percentage of total cash income) in two different forest locations of Vietnam and identify the significant relations between households and commercial collection of NTFPs through a Tobit regression model. In order to achieve this goal, the following hypotheses will be tested in this study: Hypothesis 1 . The community living closer to provincial city where majority of business activities are performed will collect more commercial value of NTFPs. Hypothesis 2 . Households in research locations have scarce labor and must allocate it efficientlyin order to maximize their 1 The fallow actually is the young secondary forest with cultivated areas covering between 33% and 50% of total area (www.fao.org/gtos/tems/land_cover_classification.htm). 2 The natural forest is continuous forest cover. It can be closed capony forest, open capony forest or long fallow (www.fao.org/ gtos/tems/land_cover_classification.htm ). 66 ECOLOGICAL ECONOMICS 60 (2006) 65– 74 incomes. Thus, the cash income generated by collecting NTFPs will compete with the income generated by other activities. If the cash amount obtained from NTFP sales is high, the cash received through other income sources is low. Thus, the household characteristics related to both labor and income, such as female labor rate, dependency ratio, income from different sources and poverty level, will influence the NTFP collection of local people. 2. Methodology 2.1. Concepts and definitions All over the world, non-timber forest products have been studied in various fields such as forestry, ethno-biology, economic botany, social development, natural resource eco- nomics, conservation biology, protected area management, agro-forestry, marketing, commercial development, ecological anthropology, cultural geography and human ecology. This cross disciplinary study has led to the development of many terms and definitions for different types of forest products. For example: alternative forest products (AFP), minor forest products (MFP), non-timber forest product (NTFP), non-timber plant products (NTPP), non-timber resources and values (NTRV), non-wood forest benefit (NWFB), non-wood goods and services (NWGS), special forest products (SFP), etc. (Wong, 2000). In many publications by the Food and Agricultural Organization of the United Nations (FAO), authors use two terms extensively: non-timber forest products (NTFPs) and non-wood forest products (NWFPs), of which NWFPs are goods of biological origin, other than wood. NWFPs also include services that are related to the collection and processing of these products such as rope making and gum collecting (Hoskins, 2003). In this research, however, Non-Timber Forest Products are defined as those biological organisms, excluding timber, collected by humans for both consumptive and selling purposes found in various form of forests (Davidson-Hunt et al., 2001, cited in Berkes et al., 2002). Additionally, in this paper, the livelihood of a household living in or near the forest is understood as the cash earned through different means of living. Through this definition, contribution of NTFPs to household livelihood or dependence of a household on NTFPs is measured by the proportion of Ma Hamlet Que Hamlet Hamlets Legend Communes Vinh city River Road Research Sites in Nghe An province N S EW Fig. 1 – Research sites in Nghe An Province. 67ECOLOGICAL ECONOMICS 60 (2006) 65– 74 cash generated by NTFP collection in total cash income. Moreover, the cash income generated by selling NTFPs is used to describe the commercial collection of NTFPs in this paper. 2.2. Study sites In December 2003, researchers from Center for Agricultural Research and Ecological Studies (CARES) and the Institute of Geography at the University of Copenhagen (IGUC) organized a research survey in two villages of two different districts: Ma hamlet in Tuong Duong district and Que hamlet in Con Cuong district. Although both of these hamlets are located in Nghe An province along the Basin of Ca River and are home to similar Thai people, they are extremely different in terms of geographical location and market access. Located in a remote area, 40 km up the Ca River from Hoa Binh Town, Ma hamlet lies in the center of a triangle that is formulated by Khe Ma and Khe Xuong streams and the Nam Non River. There is no road connecting this hamlet to other parts of the district. The only way that people in this hamlet can reach the center of district is by motorized boat. Trading activities between members of this community and outsiders can only be conducted after a 4-h commute by boat. Meanwhile, Que hamlet is located in a place which is 35 km away from the center of Con Cuong district. An inter-communal road connects this hamlet to other communes and to the district center. Even though the road is in poor condition, it is very important for local people in terms of facilitating the buying, selling and exchanging of goods. Motorbikes, cars and small trucks use this road to go to Que hamlet. Moreover, Que is closer to provincial city than Ma as depicted in Fig. 1. 2.3. Data analysis In the first part of this paper, we apply the descriptive analysis using means and Coefficient of Variation (CV) to compare the household demography, income, NTFPs and animal husbandry between the two hamlets. Additionally, T- test is applied to examine the significant differences in means of these variables between the two hamlets. At first, this analysis aims to provide the reader with a picture of the household economy in the two research locations. Secondly, it will identify the differences between households in two hamlets. In order to identify the dependence of households on NTFPs, we classified the households into four groups accord- ing to cash income: less than 5 million, 5–10 million, 10–20 million VND and more than 20 million VND. In each group, we divide the households into different categories according to percentage of cash income gained by selling NTFPs: smaller than 5%, 5–25%, 25–50% and larger than 50%. The data in this table will show the dependence of households on NTFPs by cash income level. The households who have less than 5% cash income from NTFPs are considered to be independent of NTFPs, 5–25% is we ak dependence, 25–50% is moderate dependence and the households who earn more than 50% of total cash income from NTFPs are strongly dependent on these products. The proportions of households in each category will be used to compare the dependence on NTFPs between two communities namely Que and Ma hamlets. In addition to descriptive analysis, censored regression model is employed to identify the relation between house- holds and commercial collection of NTFPs which is defined by the sold value of collected NTFPs. Another dependent variable is dependence of households on NTFPs which is measured as the percentage of cash income generated by NTFP collection. Both commercial collection and dependence of household on NTFPs are used to regress with the same group of independent variables by Tobit model because the sample data of these dependent variables are censored at zero (Greene, 2003). Regarding Hypothesis 2, the independent variables in these models should be related to both labor and income. In fact, there is labor division inside a household between children and adults, male and female. For example, children and the elderly are often doing the household work while adults are responsible for farm work; male is in charge of logging and female collects NTFPs. On the other hand, the labor allocation competes among the activities. The households who devote more labor days in raising animal will have fewer labor days for collecting NTFPs. Hence, the dependency ratio, female labor rate and number of cattle, pigs and poultries are selected to be independent variables in the econometric models. Additionally, the cash income generated by other activities could substitute the income gained from NTFPs. Other income sources such as off-farm activitie s, salary and pensio n, therefore, are also selected. Moreover, the cash income is actually important for households lacking rice for the whole Table 1 – Variables in Tobit regression model No. Variables Description Dependent M1 NTFPs value The sold value of NTFPs in each household (thousand VND). M2 NTFPs-income share The share income generated by selling NTFPs (percent age). Independent 1 Dependency ratio Number of people younger than 15 and older than 60 divided by the number of active laborers. 2 Female labor rate Number of women between 15 and 60 divided by total labor of household (smaller than one). 3 Food security Number of months that households have enough rice. 4 Off-farm income Cash income generated by off-farm activities (thousand VND). 5 Salary and pension Cash received from government budget (thousand VND). 6 Other income Cash received from fishing, relatives, selling house and social activities (thousand VND). 7 Cash income per capita Total cash income divided by total number of household members. 8 Self-evaluation of wealth Scoring: very poor 1, poor 2, moderate 3 and wealthy 4. 9 Location Que: 0; Ma: 1. 10 Cattle Number of cattle (buffaloes and cows). 11 Pig Number of pigs. 12 Poultry Number of poultry (chickens and ducks). 68 ECOLOGICAL ECONOMICS 60 (2006) 65– 74 year. They use the cash earned from other activities to buy rice in the month with shortage of rice. Thus, food security should be considered in both selected models. Finally, cash income per capita and the wealth of households are chosen to represent the poverty level (Table 1). If a female labor rate in Table 1 were equal to one (i.e. the household had no male labor), women would have to do all of the work, including that a man normally does. Thus, their time for NTFP collection would be scarce. The value of commercial NTFPs in this household would be lower than others. Moreover, according to the group discussion, the collection of NTFPs is normally done by females. Thus, the female labor rate is needed to evaluate the relation between gender and NTFP collection. 3. Results 3.1. Descriptive analysis In general, the households in both Que and Ma collect NTFPs for both selling and their own consumption. The number of products solely collected for sale is very few—only one product in Ma and two products in Que. However, farmers in Ma collect NTFPs mainly for consumption while farmers in Que are more specialized in selling certain NTFPs. In Ma, although the number of NTFPs both sold and consumed is larger than those in Que, the collection of NTFPs in this area is mainly for household consumption, because as shown in Tables 2a and b, the quantity consumed is larger than quantity sold. The number of NTFPs sold in Que is less than in Ma and is also larger than the consumed quantity in Que (Tables 2a and b). Almost every household in the two hamlets collect NTFPs for either consumption or sale. The total number of family members as well as family structure plays a critical role in this activity. Table 3 illustrates household demography in Ma and Que hamlets in December 2003. Male and female laborers are people between 15 and 59 years old. In Ma, on average, each household has 4.6 members, of which 1.6 are male laborers and 1.4 are female laborers. On the other hand, the total average number of members in Que is 5.8, which is higher than in Ma at the confident level of 99%. The numbers of male and female laborers in Que are 1.5 and 1.3 respectively. There is almost no difference between the household labor in Ma and Que (<50% significant). Consequently, there are different dependency ratios between two hamlets at the confident level of 99%. Que hamlet has more members per family as well as a greater dependency ratio (Table 3). Table 2a – Purpose of NTFPs collection in both research sites Que hamlet Ma hamlet No. Products Purpose No. Products Purpose 1 Bamboo C 1 Bamboo C+S 2 Bamboo shoots C⁎ +S 2 Bamboo shoots C⁎ +S 3 Banana C 3 Banana C 4 Fish, birds, wild animals C⁎ +S 4 Fish, birds, wild animals C⁎ +S 5 Broom grass C+S⁎ 5 Broom grass C+ S 6 Firewood C 6 Firewood C 7 Honey C+S⁎ 7 Honey C+ S 8 Mushroom C 8 Mushroom C 9 Pherynium leaf C+S⁎ 9 Pherynium leaf S 10 Amomum sp. C+ S⁎ 10 Amomum sp. C⁎ +S 11 Eggplant C 11 Eggplant C+S 12 Local edible fruits C 12 Local edib le fruits C 13 Local edible vegetables C 13 Local edib le vegetables C 14 Plant used for dying cloth C+S⁎ 14 Plant used for dying cloth C⁎ +S 15 Root used to make incense S 15 Root used to make incense C+S 16 Rattan C 16 Rattan C 17 Zanthoxylum avicenniae C17“Bach bi” C⁎ +S 18 Dioscorea persimilis C18“Bach bu” C⁎ +S 19 Indian taro C 19 “Bu Bup” C⁎ +S 20 Palm leaf C⁎ +S 20 “Cay cat” C⁎ +S 21 Indian Taro C 21 “Cay nat” C⁎ +S 22 Calamus sp. C+S⁎ 22 “Chac cau” C⁎ +S 23 Artocarpus tonkiensis C+S⁎ 23 “Chac dang” C⁎ +S 24 Alpinia sp. C+S 25 Streptocaulon juventas C⁎ +S 26 Fibraurea recisa C⁎ +S 27 Schefflera octophylla C⁎ +S 28 “Pau pa” C⁎ +S 29 Zanthoxylum avicenniae C 30 “Sam cau” C⁎ +S 31 “Ta phan” C⁎ +S Sale only 1 Sale only 1 Consumption only 12 Consumption only 7 Products for both C and S⁎ 10 Products for both C⁎ and S 23 Total collected NTFPs 23 Total collected NTFPs 31 Source: survey data in 2003. “C”: consumption; “S”: sale; “⁎”: larger quantity of either sale or consumption. Names in quotation marks are local names. Table 2b – Purpose of NTFPs collection in both research sites Que hamlet Ma hamlet Purpose of collection Number of products Purpose of collection Number of products Sale only 1 Sale only 1 Consumption only 12 Consumption only 7 Products for both C and S⁎ 10 Products for both C⁎ and S 23 Total collected NTFPs 23 Total collected NTFPs 31 Source: survey data in 2003. “C”: consumption; “S”: sale; “⁎”: larger quantity of either sale or consumption Because Table 2a is too large, it is reduced into this table. 69ECOLOGICAL ECONOMICS 60 (2006) 65– 74 Animal raising is importantfor many households in thetwo hamlets. Having animals, especially cattle, is as a kind of an economic safety net for the majority of farmers. If the harvest fails, sickness occurs in the family, or any other bad thing happens, selling animals provides an economic buffer (Quang and Jakobsen, 2005). In Ma, 78.6% of the interviewed farmers owned at least one head of cattle (either cow or buffalo) with the average number of cattle per household being 5.5. Meanwhile, in Que, these numbers are 74.1% and 3.9 heads respectively. In addition, most people in the two hamlets also raise pigs at 82.1%of households for Ma and81.5% for Que, with the average number of pigs per households being 4.0 and 2.1 respectively. Poultry is the category with the greatest differ- ence between the two hamlets. In Ma, on average, each household owns 14.7 ducks and chickens while in Que, this number is tripled to 43.1 ducks and chickens, which is a significant difference (99%). In short, many people in Ma and Que raise animals and there is almost no statistically significant difference between these hamlets in terms of number of cattle and pigs owned by each household but there is a large difference in terms of poultry because in 2003 the chickens in Ma massively died due to influenza (Table 4). People do earn income from cutting wood in the forest. Yet, as timber exploitation is considered illegal in the two hamlets, interviewees were reluctant to talk about this activity. Therefore, in this paper, logging is excluded from calculations of total cash income. Income sources in the two hamlets are divided into six categories as follows: agricultural products, animal husband- ry, NTFPs, off-farm activities, salary and pension, and others comprised of remittances and compensation from social activities. In Ma hamlet, people sometimes sell some of their agricultural products to local people or to people from the lowlands. These kinds of agricultural products include tea, vegetables, rice and bananas. However, farmers only get a small amount of money from these products with an average value of 125 thousand VND per capita. According to the statistical data from District People's Committee, Que hamlet is considered to be the poorest in Binh Chuan commune, where 38 households are classified as poor, 14 as average and 15 households are classified as well-off. On average, the poor households lack food between 4 and 7 months per year. As a consequence, they use all of their agricultural products for subsistence. The greatest income in Ma and Que comes from animal husbandry with average value of 760 thousand VND per capita and 426 thousand VND per capita, respectively. This income source comes mainly from selling cattle and pigs. Salary and pension is the category with the greatest difference between Ma and Que. In Ma, on average, people can earn 492 thousand VND per capita, meanwhile, in Que, this number is just 24 thousand VND per capita at confidence level of 98.5%. That means there are more people working for the govern- ment with higher salary in Ma than in Que (Table 5). In terms of NTFPs, people in Que earn more money from selling these kind of products than people in Ma hamlet (104 thousand VND and 22 thousand VND per capita at 99% significance) (Table 5). Fig. 2 shows the average cash income shares of households in the two hamlets. As shown, the biggest share in both hamlets comes from animal husbandry at 37% for both. NTFP selling is the second biggest share in Que at 33% while in Ma hamlet, it comprises only 5%. Income from off-farm activities such as hired labor, shop and services in Ma accounts for 33% of their total income. At the same time, in Que, this number is only 16% (Fig. 2). Table 4 – Animal husbandry (unit: heads/household) Indicators Ma hamlet (N=28) Que hamlet (N=27) T-test (0.05) Mean CV (%) Percentage of HHs Mean CV (%) Percentage of HHs P (T≤ t), two-tail Cattle 5.5 86.8 78.6 3.9 98.1 74.1 0.176 Pig 4.0 119.2 82.1 2.1 77.9 81.5 0.052 Poultry 14.7 204.0 75.0 43.1 63.9 92.6 0.001 Source: survey data in 2003. Table 5 – Cash income per capita (unit: 1000 VND/person) Indicators Ma hamlet (N=28) Que hamlet (N=27) T-test (0.05) Mean CV (%) Mean CV (%) P (T≤t), two-tail Agricultural products 125 491 –– – Animal husbandry 760 174 426 165 0.250 NTFPs 22 166 104 85 0.000 Off-farm activities 445 276 214 235 0.368 Salary and pension 492 195 24 316 0.015 Others 43 248 73 252 0.458 Total 1886 147 842 125 0.073 Source: survey data in 2003. Table 3 – Household demography (unit: number of persons/household) Indicators Ma hamlet (N=28) Que hamlet (N=27) T-test (0.05) Mean CV (%) Mean CV (%) P (T ≤ t), two-tail Total member 4.6 38.9 5.8 22.9 0.006 Male labor 1.6 63.5 1.5 55.9 0.833 Female labor 1.4 61.5 1.3 52.1 0.421 Dependency ratio 0.7 75.8 1.2 49.6 0.000 Total labor 3.0 52.1 2.8 37.8 0.538 Source: survey data in 2003. 70 ECOLOGICAL ECONOMICS 60 (2006) 65– 74 Table 6 shows the total cash income as well as the dependence level of forest dwellers on NTFPs in the two hamlets. In Que, no household can receive more than 20 million VND while 7% of households earn more than 20 million VND a year in Ma. The households, who earn less than 5 million VND per year, account for 67% of total households in Que and only 57% in Ma hamlet (Table 6). In Que, 30% of households depend considerably upon NTFPs. More than fifty percent of their total cash income is generated by NTFP collection. All of these households are in the poorest group, earning less than 5 million VND cash a year. In Ma, no household falls into this situation. In the other categories, the cash income of 15% of households in Que depends moderately on selling NTFPs while this rate in Ma is only 4%. All of these households earn less than 5 million VND per year, which can be considered to be the poorer of the two hamlets. In the group that is not dependant on NTFPs, Ma has 75% of households while Que has only 26%. It means that the number of households in Ma whose cash income depends on NTFPs is less than in Que (Table 6). 3.2. Results of Tobit estimation 3.2.1. Determinants of cash income from NTFPs This part aims to identify the relations between households and commercial collection of NTFPs by employing a Tobit regression. The models in this section, therefore, only account for the contribution of NTFPs to the cash income of house- holds. The results estimated by a Tobit regression in Table 7 shows the correlation between household characteristics and the cash income generated by NTFP collection. The results in Table 7 show that there are four independent variables significantly related to the value of NTFPs sold. These variables are dependency ratio, female labor rate, self- evaluation of wealth, and dummy variables of location, of which only female labor rate positively correlate to cash income generated by NTFP collection. The other independent variables are negatively correlated to the dependent variable. 3.2.2. Determinants of household dependence on NTFPs The most important indicator in this paper is the depen- dence of households on NTFPs. The identific ation of dependence is based on the result of Model 2 in Table 8, which depicts the significant correlation between house- holds and the share of cash income generated by selling NTFPs. This table shows that seven independent variables are significantly correlated to the dependence of households on NTFPs of which female labor rate and salary–pension are positively correlated to the dependent variable. Five other variables including food security, other income, cash income per capita, dummy location and self-evaluation of wealth are negatively correlated to the dependence of farmers on NTFPs (Table 8). 4. Discussion 4.1. Descriptive analysis The farmers in Que can sell their products more easily than farmers in Ma. They just sell the products demanded by consumers in the market. In the case of Ma hamlet, farmers 37% 33% 16% 5% 9% 5% 37% 5% 33% 16% 4% Agricultural products Animal husbandry NTFPs Off-farm activites Salary & pension Others Que Hamlet Ma Hamlet Fig. 2 – Average income shares of households in Ma and Que hamlets. Table 6 – Dependence on NTFPs of households at different cash income levels (unit: percentage of households) Categories (% of total cash income from NTFPs) Total cash income (million VND) by group per year Total <5 million 5–10 million 10–20 million >20 million Ma Que Ma Que Ma Que Ma Que Ma Que <5 32 7 18 4 18 15 7 0 75 26 5–25 21 15 0 11 0 4 0 0 21 30 25–50 415 0 0 0 0 0 0 415 >50 030000000 030 Total 57 67 18 15 18 19 7 0 100 100 Source: survey data 2003 (N =55). 71ECOLOGICAL ECONOMICS 60 (2006) 65– 74 mainly sell their collected products to others in the hamlet or in the community nearby. They seldom sell to outsiders in the broad market. Consequently, in terms of NTFPs, the main difference between the two hamlets here is the value and quantity of NTFPs sold. In terms of NTFPs, people in Que earn more money from selling these kinds of products than people in Ma hamlet. There are two reasons for this phenomenon. First, in Que, as mentioned above, the average family size is larger than in Ma, therefore, there are more people who can take part in gathering NTFPs. Naturally, they might collect greater amounts of products. Second, in Que, the prices of NTFPs are normally higher than in Ma. For example, in Que, people can sell one leaf of pherynium for 100 VND, meanwhile, in Ma, the price is only 20–25 VND/leaf. Similarly, 1 kg of broom grass in Que sells for 2500 VND while in Ma it is 400–500 VND. Briefly, income from selling NTFPs in Que hamlet is greater than that in Ma while the amount of money received from salary and pension is smaller (Table 5). NTFP selling is the second biggest share in Que while in Ma hamlet it comprises only 5%. Hence, in Que NTFPs are more important for people than in Ma in terms of cash income (Fig. 2). Income from off-farm activities such as hired labor, shops and services in Ma accounts for 33% of their total income. At the same time, in Que, this number is only 16%. One of the main reasons for this is because in Ma, people sell more handicrafts (e.g. woven cloth, rattan baskets, etc.) than in Que. The least important source of income in Ma is the other category at 4%, which includes the c ompe nsa tion from participation in social activities and remittances from rela- tives. Meanwhile, in Que, salary and pension contribute the smallest part. In fact, for all the interviewed households in Que, no person received pension from the government. In conclusion, in Ma, the cash income of household comes mainly from selling animals and off-farm activities whereas in Que, the main proportion of cash income is generated by raising animals and selling NTFPs (Fig. 2). When earnings of households cannot satisfy their needs, people who live in or near the forests go to the forests to collect forest products to supplement the shortage in their house- holds. The cases in both Ma and Que hamlet show that the less cash the household earns, the greater the share of their income is generated by NTFPs (Table 6). Finally, it can be concluded that the poor who live in or near forests depend more on the NTFPs than the wealthier house- holds. At the same level of wealth, the households who live closer to provincial city can receive more cash income from selling NTFPs than those in remote areas. 4.2. Determinants of cash income from NTFPs The negative coefficient of location dummy (Ma=1) which is significant at 99% of confidence in this model indicates that the value of NTFPs sold in Que is significantly higher than in Ma (Table 7). The negative coefficient of dependency ratio implies that households who have higher dependency ratio enjoy less benefit of collected NTFPs because they lack labor for the activities of NTFP collection. In Model 1, the dependency ratio negatively correlates to the value of NTFPs sold at 95% of confidence (Table 7). Moreover, the results of the model in Table 7 show that the female labor rate is also significantly related to the value of NTFPs sold at 90% of confidence. Thus, at a confidence level of 90%, it can be concluded that the households who have more female labor or collect more NTFPs generate cash income for their households. Table 7 – Tobit estimation of cash income from NTFPs (Model 1) Independent variables Coefficients (beta) t-value Significance (P-value) Constant 1296.14⁎⁎⁎ 3.55 0.001 Dependency ratio −264.990⁎⁎ − 2.44 0.019 Female labor rate 589.224⁎ 1.96 0.056 Food security −24.368 − 0.99 0.326 Off-farm income − 0.001 − 0.04 0.970 Salary and pension 0.014 0.54 0.594 Other income − 0.067 − 0.89 0.381 Cash income per capita − 0.064 − 0.98 0.330 Self-evaluation of wealth − 162.475⁎ − 1.87 0.069 Location − 735.325⁎⁎⁎ −4.31 0.000 Cattle 23.112 1.46 0.151 Pig − 0.712 − 0.05 0.963 Poultry − 0.819 − 0.40 0.690 LR chi 2 (12) 30.23 Prob>chi 2 0.0026 Log likelihood − 402.821 Pseudo R 2 0.0362 Dependent variable: NTFPs value sold N=55 ⁎, ⁎⁎, and ⁎⁎⁎ are significance level at 90%, 95% and 99% respectively. Table 8 – Tobit estimation of dependence on NTFPs (Model 2) Independent variables Coefficients (beta) t-value Significance (P-value) Constant 105.499⁎⁎⁎ 6.16 0.000 Dependency ratio − 8.173 −1.60 0.116 Female labor rate 34.333⁎⁎ 2.44 0.019 Food security − 4.588⁎⁎⁎ − 3.99 0.000 Off-farm income 0.001 0.78 0.440 Salary and Pension 0.002⁎ 1.88 0.067 Other income − 0.007⁎ − 1.95 0.057 Cash income per capita − 0.006⁎ − 1.86 0.070 Self-evaluation of wealth − 8.724⁎⁎ − 2.14 0.038 Location −39.741⁎⁎⁎ −4.96 0.000 Cattle 0.111 0.15 0.882 Pig 0.605 0.84 0.408 Poultry − 0.104 −1.09 0.280 LR chi 2 (12) 50.29 Prob>chi 2 0.000 Log likelihood − 234.54 Pseudo R 2 0.0968 Dependent variable: NTFPs-income share N=55 ⁎, ⁎⁎ and ⁎⁎⁎ are significance level at 90%, 95% and 99% respectively. 72 ECOLOGICAL ECONOMICS 60 (2006) 65– 74 The significant coefficients of self-evaluation of wealth indicate the negative correlation between the commercial value of NTFPs collected and the poverty level of households in terms of cash income at 90% of confidence level. It suggests that wealthier households gain less commercial value of NTFPs than poorer ones. Briefly, the determinants of the value of sold NTFP, therefore, must be geographical location, dependency ratio, female labor rate, and the poverty level which is defined by wealth levels of households. 4.3. Determinants of household dependence on NTFPs According to results of the Tobit estimation in Table 8, both food security and location dummy negatively correlates to the dependent variable, the share of cash income generated by selling NTFPs. Both have significance at 99% of confi- dence. Obviously, the coefficient of dummy location (Que=0, Ma=1) indicates that the cash income of households in Que depends more on selling NTFPs than those in Ma. On the other hand, the coefficient of food security shows that the households who lack rice for their own consumption will depend more on NTFPs than others who have enough rice for the whole year. An important finding in this model is the female labor rate, which is calculated by the rate of female labor and total family labor. This variable positively correlates to the share of income generated by selling NTFPs at a significance level of 95%. It suggests that the households with higher rates of female labor will depend more on NTFPs than others in the same location except the case of households without male labor. 3 This result is very important because, following other research, it continues to emphasize the role of women in NTFP trading and collection (Belcher and Kusters, 2004). Additionally, the results in Model 2 show that both cash income per capita and wealth levels of households nega- tively correlate to the share income generated by NTFPs. This result suggests that the poor are more dependent on NTFPs than the wealthier households. There are several reasons behind this result. The first reason is the low capital requirement for extraction of NTFPs. Secondly, because of low income from other sources people have to collect commercial NTFPs to supplement or increase their house- hold income. The lower income from other sources also makes the income share of NTFPs higher, which in turn makes some people more dependent on these products or, in other words, more dependent on the forest. Moreover, the coefficient of other-income variable also has a negative value at a significance level of 90%. It indicates that if farmers receive more income from remittances and com- pensation from social activities their livelihood would rely less on the forest. Moreover, the positive coefficient of salary and pension indicates that the compensation from Govern- ment for wage employment and retired persons cannot help them rely less on NTFPs collection. They still collect NTFPs but not much, the coefficient is relatively low, equal to 0.002 (Table 8). Finally, it can be concluded that the determinants of a household's dependence on NTFPs are location, female labor rate, food security, other income, and the poverty level of the household. 5. Conclusions Generally, the households who live in or near the forests collect the NTFPs for both sale and consumption. However, the balance between consumption and sale of NTFPs is not the same in all locations. In Ma, the quantity of NTFPs consumed is greater than the quantity sold while the converse is true in Que. Consequently, households in Que earn more cash from these products than those in Ma. NTFPs, therefore, are more important for people in Que than in Ma in terms of cash income. That is the main difference between the two hamlets in terms of NTFPs. Additionally other aspects of household economy including labor, cattle, and income sources are also different between two hamlets. Furthermore, NTFPs are very important in poor households or in those who lack rice, have a high dependency ratio, or high rate of female labor. The households who have higher dependency ratio benefit less from NTFPs sold, while those who lack rice for their own consumption or have a higher rate of female labor depend more on NTFPs. Poor households are more dependent on NTFP collection than other groups. As a result, the determinants of household's dependence on NTFPs are geographical location, gender balance, food security, other income and poverty level which is defined by cash income per capita and wealth level of house- hold. The determinants of NTFPs sold are geographical location and dependency ratio, gender balance and po- verty level. For two communities living in or near the forests with the same ethnicity, the community living closer to provincial city will collect more value of NTFPs and the livelihoods of households in this community is more dependent upon the NTFPs. The results of this research hi ghly suggest that the Government of Vietnam should consider gender balance, location, poverty level, dependency ratio in order to keep the balance between conserving the biodiversity of NTFPs and the earnings of forest dwellers because these characteristics are significantly correlated to the dependence of farmers living in or near the forests upon NTFPs as well as cash income generated by NTFPs. Additionally, the Government of Vietnam should pay attention to the food security of forest dwellers in order to conserve the forest. Households who live in or near the forests should have enough food throughout the year. Otherwise they will collect as many NTFPs as possible. This might lead to low biodiversity. An appropriate incentive policy for woman and poor households might be a possible solution for less commercial collection of NTFPs. Generally, the poor households who have low dependency ratio or high female labor rate should be encouraged to participate in the forest management system to control the balance between commercial collection of NTFP and conservation. 3 In fact, this argument is presented in Data analysis. 73ECOLOGICAL ECONOMICS 60 (2006) 65– 74 Policies of poverty alleviation should be applied to the forest dwellers with priority to the households with low dependency ratios, high female labor rates or those that lack foods in many months of a year, to lessen their dependence on NTFPs in order to achieve efficiency in conserving biodiversity and forest management. Acknowledgements First and for most, we would like to express our sincere thanks to Professor Tran Duc Vien and Stephen Leisz who provided essential information for this research. Our special thanks go to Amanda Allbritton and Meg Hiesinger who devoted their time to edit the English of our paper. Especially, we are thankful for the Danish International Development Agency for funding the University Support to Environmental Planning and Management Project (USEPAM) to collect the data in both research sites. Finally, we would like to thank all researchers of CARES and all partners in USEPAM project who have contributed their comments to our paper in the workshops. REFERENCES Ambrose-Oji, B., 2003. The contribution of NTFPs to the livelihoods of the ‘forest poor’: evidence from the tropical forest zone of south-west Cameroon. International Forestry Review 5 (2), 106–117. Beer, Jenne H., 1993. Non-Wood Forest Products in Indochina— Focus: Vietnam. Working Paper No 0782, Food and Agriculture Organisation of United Nations, Rome. Belcher, B., Kusters, K., 2004. Non-timber forest product commer- cialization: development and conservation lessons. In: Kusters, K., Belcher, B. (Eds.), Forest Products, Livelihoods and Conser- vation, Volume 1—Asia. Center for International Forestry Research, Indonesia. 365 pp. Chapter 1, pp 1–22. Berkes, F., Davidson-Hunt, I.J., Ruta T., Sinclair, J., 2002. Scientific and First Nation Perspectives of Non-Timber Forest Products: A Case Study from the Shoal Lake Watershed, Northwestern Ontario. Project Report No. 2002-4, Natural Resources Institute, University of Manitoba. Chamberlain III, J.L., 2000. The Management of National Forests of Eastern United States for Non-Timber Forest Products. Disser- tation, the Faculty of the College of Natural Resources, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Cocksedge, Wendy, 2001. The Role of Co-operatives in the Non- Timber Forest Product Industry: Exploring Issues and Options Using the Case Study of Salal (Gaultheria shallon; Ericaceae). Occasional papers, British Columbia Institute for Co-operative Studies, University of Victoria, Cananda. Davidson-Hunt, I.J., Duchesne, L.C., Zasada, J.C. (Eds.), 2001. Non- Timber Forest Products: Local Livelihoods and Integrated Forest Management. Conference Proceedings of Forest Com- munities in the Third Millennium: Linking Research, Business, and Policy Toward a Sustainable Non-Timber Forest Product Sector, North Central Research Station, 1992 Folwell Avenue St. Paul, Minnesota 55108, pp. 1–12. Fisher, R.J., 2000. Creating incentives for conservation: non-timber forest products and poverty alleviation. Asia-Pacific Commu- nity Forestry Newsletter 13/2, 5–7. Greene, William H., 2003. Econometric Analysis. Pearson Educa- tion, Inc., Upper Saddle River, NJ. 1026 pp. Hoskins, M.W. (Ed.), 2003. Marketing Information Systems for Non- Timber Forest Products. Community Forestry Field Manuals, vol. 6. Food and Agriculture Organization of the United Nation, Rome, Italy. Lawrence, Anna, 2003. No forest without timber? International Forestry Review 5 (2), 87–96. Osman, M., Mishra, P.K., Dixit, S., Ramachandran, K., Singh, H.P., Rama Rao, C.A., Korwar, G.R., 2000. A Review of Dynamics, Management and Livelihood Contributions. Common Pool Resources Research Project Report No. 3, Natural Resource Institute (NRI) and Department for International Department (DFID), Palace Street, London, UK. Quang, Nguyen Vinh, Jakobsen, Jens, 2005. Forest and fallow products in a rotational swiddening system in North Central Vietnam. In: Duc Vien, Tran, et al. Eds. Workshop proceeding on Marketing of Agro-forestry products in the Upland. Hanoi Agricultural Publishing House, pp. 111–131. Rijsoort, J.V., 2000. Non-Timber Forest Products: Their Role in Sustainable Forest Management in the Tropics. Theme Studies Series, vol. 1. National Reference Centre for Nature Manage- ment (EC-LNV) International Agricultural Centre (IAC), Wageningen, The Netherlands. Roderick, P. Neumann, Hirsch, Eric (Eds.), 2000. Commercialization of Non-Timber Forest Products: Review and Analysis of Research. Center for International Forestry Research, Borgor, Indonesia. 173 pp. Ros-Tonen, M.A.F., 2000. The Role of Non-Timber Forest Products in Sustainable Tropical Forest Management. Holz als Roh-und Werkstoff, vol. 58, Number 3. Springer-Verlag, Heidelberg, pp. 196–201. Ros-Tonen, M.A.F., Wiersum, F.K., 2003. The Importance of Non- Timber Forest Products for Forest-Based Rural Livelihoods: An Evolving Research Agenda. International Conference on Live- lihoods and Biodiversity, Amsterdam Research Institute for Global issues and Development Studies, Bonn. Ruiz, P.M., Broekhoven, A.J., Aluma, J.R.W., Iddi, S., Lowore, J.D., Mutemwa, S.M., Odera, J.A., 1997. Research on Non-Timber Forest Products in Selected Countries in Southern and East Africa: Themes, Research Issues, Priorities and Constraints. Working Paper No. 15, Center for International Forestry Research (CIFOR), Bogor, Indonesia. Sills, E.O., Sharachchandra, L., Holmes, T.P., Pattanayak, S.K., 2003. Non-timber forest products in the rural household economy. In: Sills, O. Erin, Abt, Karen Lee (Eds.), Forests in a Market Economy. Kluwer Academic Publishers, 3300 AA Dordrecht, The Netherlands, pp. 260–281. Subhrendu, K.P., Sills, E.O., 2001. Do tropical forests provide natural insurance? The microeconomics of non-timber forest product collection in the Brazilian Amazon. Land Economics 77 (4), 595–612. Tickin, T., 2004. The ecological implications of harvesting non- timber forest products. Journal of Applied Ecology 41, 11–21. Wong, J.L.G., 2000. The Biometrics of Non-Timber Forest Product Resource Assessment—A Review of Current Methodology. Project Report of ZF0077 Forest Research Program, the Department for International Department, Palace Street, London, UK. Yen, Nguyen Thi, Nguyen Quang Duc, Vu Manh Thien, Dang Duc Phoung, B.A. Ogle, 1994. Dependency on Forest and Tree Products for Food Security, Pilot Study in Yen Huong Com- mune, Ham Yen District, Tuyen Quang Province, North Vietnam. Working paper No. 250, Swedish University of Agricultural Science, Uppsala, Sweden. 74 ECOLOGICAL ECONOMICS 60 (2006) 65– 74 . SURVEY Commercial collection of NTFPs and households living in or near the forests: Case study in Que, Con Cuong and Ma, Tuong Duong, Nghe An, Vietnam Dang Viet Quang* , Tran Nam Anh Center. one product in Ma and two products in Que. However, farmers in Ma collect NTFPs mainly for consumption while farmers in Que are more specialized in selling certain NTFPs. In Ma, although the number of NTFPs. the main difference between the two hamlets here is the value and quantity of NTFPs sold. In terms of NTFPs, people in Que earn more money from selling these kinds of products than people in Ma