Các yếu tố kinh tế xã hội ảnh hưởng đến các hộ gia đình sử dụng rừng. Much research has focused on understanding the importance of forest environmental income in different communities and highlighting key socioeconomic characteristics of forest-dependent households. This paper examines the economic importance of forests among rural agriculturalists in Vietnam. Data were collected through a questionnaire survey of 104 householdsinfivestudyvillagesinHaTinhprovincein north central Vietnam surrounding the Ke Go Nature Reserve (KGNR). Variables such as migration status of the household, age, income class and landholdings were used to identify characteristics of households with high forest income in both absolute and relative terms. More than half of households reported receiving forest environmental income in cash. Socioeconomic variables were compared between forest cash income (FCI) households and non-FCI households. Non-FCI households had more alternative income sources from wage labour and livestock, while FCI households were significantly younger, tended to live closer to the forest and had larger landholdings. Contrary to other researchonforestuse,thehouseholdsderivingthemost forest income in both absolute and relative terms were notthepoorerhouseholds,butthoseinthemiddleclass. These findings highlight the need for conservation and development projects to pay attention to the specific household factors that influence forest use, rather than relying on assumptions that poverty and forests are always linked.
Environmental Conservation 35 (2): 147–159 © 2008 Foundation for Environmental Conservation doi:10.1017/S0376892908004736 Forest environmental income in Vietnam: household socioeconomic factors influencing forest use PAMELA D MCELWEE School of Global Studies, Arizona State University, PO Box 875102, Tempe AZ 85287-5102, USA Date submitted: 20 October 2007; Date accepted: 20 April 2008 SUMMARY Much research has focused on understanding the importance of forest environmental income in different communities and highlighting key socioeconomic characteristics of forest-dependent households This paper examines the economic importance of forests among rural agriculturalists in Vietnam Data were collected through a questionnaire survey of 104 households in five study villages in Ha Tinh province in north central Vietnam surrounding the Ke Go Nature Reserve (KGNR) Variables such as migration status of the household, age, income class and landholdings were used to identify characteristics of households with high forest income in both absolute and relative terms More than half of households reported receiving forest environmental income in cash Socioeconomic variables were compared between forest cash income (FCI) households and non-FCI households Non-FCI households had more alternative income sources from wage labour and livestock, while FCI households were significantly younger, tended to live closer to the forest and had larger landholdings Contrary to other research on forest use, the households deriving the most forest income in both absolute and relative terms were not the poorer households, but those in the middle class These findings highlight the need for conservation and development projects to pay attention to the specific household factors that influence forest use, rather than relying on assumptions that poverty and forests are always linked Keywords: forest environmental income, household livelihoods, non-timber forest products, poverty, Vietnam INTRODUCTION The world trade in non-timber forest products (NTFPs) (for example forest fruits, medicines, aromatics and resins) is worth billions of dollars (Iqbal 1993) Many have advocated that NTFPs and other forms of forest products be promoted to provide increased income opportunities for forest dwellers ∗ Correspondence: Dr Pamela McElwee Tel: +1 480 727 0736 Fax: +1 480 727 8292 e-mail: pamela.mcelwee@asu.edu and users (Counsell & Rice 1992; Wollenberg & Ingles 1998) However, the initial optimism that the twin goals of conservation and economic development could be linked through NTFP extraction seems to have diminished (RuizP´ rez & Arnold 1996; Arnold & Ruiz-P´ rez 2001) There have e e been naive assumptions behind many marketing plans, and historical trends in NTFP use indicate that negative outcomes are common (Gray 1990, Dove 1993) However, many millions of households continue to harvest forest products to enhance their livelihoods (Byron & Arnold 1999) Better understanding of why some households harvest forest goods while others not may help explain some of the problems encountered in NTFP promotion, such as whether the poor or rich are more likely to benefit from commercialization schemes (Neumann & Hirsch 2000; Marshall et al 2003; Belcher et al 2005) Recent research has highlighted key socioeconomic characteristics of forest-dependent households that can play roles in explaining forest use For example, in a study in the Philippines, elderly people were more likely to collect forest goods because of their more extensive knowledge of forest plants and wildlife (Lacuna-Richman 2002) Elsewhere, younger households are more dependent on wild-collected products, as they set out to start families and have lower agricultural assets than older better-established households (Coomes et al 2004; McSweeney 2004) Another key variable of interest is the relationship between income and forest use Siebert and Belsky (1985) found the households with the lowest level of rice self-sufficiency relied most on rattan harvesting for income in the Philippines Gunatilake et al (1993) found that contributions of NTFPs to incomes in Sri Lanka declined as incomes rose Similar arguments have been made elsewhere that the poor are more dependent on forest goods than better-off households (Cavendish 2000; Hegde & Enters 2000; Mahapatra & Tewari 2005), and the poor particularly rely on forest income in times of particular need (McSweeney 2002, 2004) Other studies indicate that medium-income or richer households are, in some situations, more likely to have forest income than the poor, owing to high labour requirements or elite capture of valuable resources (Godoy et al 1995; Wickramasinghe et al 1996; Ambrose-Oji 2003; de Merode et al 2004) Often the role of income class depends on what variable is measured In a meta-analysis of 54 different studies, absolute forest income increased as total household income increased and ‘was thus important not only for poor communities’, but forest income as a share of total income 148 P D McElwee decreased, indicating the poor were more dependent on forest income (Vedeld et al 2004, p xiv) This finding was echoed in Ethiopia, where wealthier households received more absolute cash income from forest produce, while poorer households were dependent on forests for a larger percentage of their income (Mamo et al 2007) When comparing forest use in South Africa, poor, average and rich households did not differ in terms of the number of NTFPs used or the proportion of households using them (Shackleton & Shackleton 2006) However, the poorer households did use more NTFPs per person in terms of volume when both income and subsistence purposes were considered Other studies have noted the importance of land tenure The landless and land-poor are often more dependent on forest product collection than the land-rich (Lacuna-Richman 2002; Pandit & Thapa 2003) For those who have no access to land for agriculture, NTFPs can provide a much needed source of support, especially when they are collected from common or open lands In Orissa (India) dependence on forest income was strongly correlated with size of land holdings, with the landless being most dependent (Fernandes & Menon (1987) Other social variables may also influence forest use In one study, NTFP exploitation was positively correlated with household debt, labour availability and male to female ratios and negatively correlated with income, education, distance to forest, involvement in non-agricultural activities and incorporation into the market (Gunatilake 1998) Factors such as the size and labour capacity of households (Mamo et al 2007), migration status (Lacuna-Richman 2006), opportunity costs of collection and the substitutions of forest products by market purchased goods (Senaratne et al 2003), and the strength of markets for forest produce (Ruiz-P´ rez et al 2004; e Bista & Webb 2006) may also be important Previous studies having highlighted heterogeneity even within smaller forestextracting communities (Coomes et al 2004; Vedeld et al 2004), more studies are needed to comprehensively account for use of forest products across a range of ecological locations and social situations In Vietnam, millions of rural people live in close proximity to forests, yet there has been little published on NTFP use, the research there predominantly focusing on ethnic minorities, who comprise around 13% of the national population and live in mountainous areas with higher rates of forest coverage (Wetterwald et al 2004; Dang Viet Quang & Tran Nam Anh 2006; Hilfiker et al 2006) There has been much less attention to forest use among ethnic Vietnamese and those in lowland areas The present study attempts to remedy this through a case study of forest extraction by Vietnamese households living in lowland and midland areas of north central Vietnam The study aimed to build on experiences garnered from previous studies on NTFP use and environmental income, and attempted to follow the ‘best practices’ of Vedeld et al (2004) First, I examined all plant and animals extracted from forests, both NTFPs and wood products, to establish their relative importance so that different sources of forest income could be clearly compared Second, I collected information on all other types of household income, both subsistence and in cash, in the study area, so that forest environmental income could be put in the context of overall household livelihoods Third, I worked with a number of households with diverse socioeconomic backgrounds, including from all income classes and with households from both the land poor and land rich, thus accounting for factors often ignored in other studies which frequently focus only on poor or landless households (Bista & Webb 2006) Fourth, I worked with migrants and local-born populations to see what effect the household’s background and history played in their forest use decisions The main objective of the project was to determine which of a number of socioeconomic factors had the strongest relationship with the use of forest produce and forest environmental income dependency METHODS Research setting The study was conducted in rural areas of Ha Tinh province, approximately 300 km south of the national capital Hanoi (Fig 1) Ha Tinh had an estimated population of 1.286 million in 1999 and is among the poorest areas of Vietnam (Department of Planning and Investment Ha Tinh 2003) The province has an area of 6055 km2 divided into 11 districts and a further 259 communes, the lowest level of state administration in Vietnam Below the commune most households group into villages, although these are not officially recognized as an administrative unit Ha Tinh is characterized by low coastal plains bordering the South China Sea, rising to high mountains in the Annamite chain separating Vietnam from Laos Two major nature reserves, the Vu Quang Nature Reserve and the Ke Go Nature Reserve (KGNR), have been demarcated in the past 15 years to protect what are seen as high levels of biodiversity, particularly of mammals and birds (Eames 1996; Le Trong Trai et al 1999) The KGNR was established in 1996 with > 35 000 ha, primarily to protect populations of two endemic and endangered pheasant species The Reserve was described at the time of founding as having one of the ‘largest remaining blocks of broadleaf evergreen forest in the level lowlands of central Vietnam’ (Le Trong Trai et al 1999, p vii) However, more than 75% of the forest has been classified as heavily disturbed due to past logging by state-owned timber companies, and only at higher elevations are areas of lightly disturbed forest found The topography of the Reserve is mostly low, gently sloping hills, with altitudes of 50–500 m and it supports at least 46 mammal, 270 bird and 562 plant species (Le Trong Trai et al 1999) Officially, most protected areas in Vietnam consist of a strictly protected inner core in which almost all anthropogenic activities are banned (ICEM [International Centre for Environmental Management] 2003) Within national parks and nature reserves in particular, it is prohibited to ‘log, exploit (excluding activities related to forest cleaning and Forest environmental income in Vietnam 149 Figure Map of the study area rehabilitation), hunt animals, collect specimens under any means and forms .Strict protection areas within national parks and nature preservation areas should be protected strictly Every activity that causes negative impacts to forest is not allowed’ (MARD [Ministry of Agriculture and Rural Development] 1997) Any commune sharing a border with a protected area is considered to be a ‘buffer zone’ in state law, although in reality this creates no significant restrictions on land use (Gilmour & San 1999) Many buffer zones, however, have been able to attract projects, such as integrated conservation and development projects (ICDP), in order to reduce dependence of residents on protected area resources; one such project was set up in Ke Go, funded through the IUCN (World Conservation Union), to encourage the planting of domesticated NTFPs like rattan and medicinal plants in home gardens When the KGNR was demarcated the boundaries were deliberately drawn to exclude human settlements, which fell into a buffer zone of 22 000 While no households were living inside the actual boundaries of the Reserve at the time of the present study, approximately 40 000 people lived in the buffer zone, spread over eight communes in the district of Cam Xuyen, one of the poorer districts While not all residents in these communes were involved in forest extraction activities, those in areas closer to the Reserve were often actively engaged The most accessible areas within the KGNR used by nearby villages were dominated by secondary forest and scrub growth where timber, fuelwood and a variety of NTFPs could be harvested, and buffalo and cattle were occasionally grazed as well The borders of Ke Go were only sporadically patrolled by small numbers of rangers; while in the law any exploitative extraction of goods from a nature reserve was illegal, rangers primarily focused interdiction efforts on timber, charcoal extraction, and hunting, while ignoring infractions of other NTFPs or fuelwood harvesting for the most part The typical punishment for illegal logging was confiscation of timber and any equipment used to cut it, and a monetary fine up to 500 000 VND (c US$ 34), while charcoal makers were subject to lower fines In addition to these restrictions on extraction, national law also stated that no land tenure certificates would be granted to people farming inside the official boundaries of any protected area, and any encroachment of agriculture onto the Reserve would be stopped through fines and resettlement Study sample Five villages in the KGNR buffer zone were chosen based on stratified random sampling to include those with good access and those with poorer access to the KGNR forest The approximate distance households in each village had to walk to arrive at the natural forests that make up the KGNR were used to classify villages as either close (3 km away) Once households reached the KGNR border, it was usually an additional two or more kilometres walk into Reserve areas with sufficient forest cover to collect the most common products The general topography and ecology was approximately the same for all villages; altitudes were 5.6 million VND (>US$ 386) For total cash and subsistence income, ‘low income households’ were those with incomes US$ 525) Landholdings were assessed using local land categories Agricultural lands included rice fields, both irrigated and non-irrigated, and agricultural plots for field crops such as cassava and corn Around the household compound, home gardens were usually found, and some households also had hill gardens located further away Some households’ land holdings included government-allocated forest land, which were plots of land with planted trees (mainly acacia, pine and eucalypts) or which were slated for reforestation, and which had been given to households for long-term protection rights Landholdings were measured in sao, the local land measurement unit (1 = 500 m2 or 20 = ha) Most households were clear on the number of they had land rights to as such information was listed on all household land tenure certificates, which had been recently issued Household expenses were calculated from informant recall, a standard tool in most household living standards surveys done in Vietnam and elsewhere (World Bank 2001) While this is an imperfect method, households often did keep receipts for many purchases, which could be checked and verified, and when exact amounts of money spent could not be verified, approximations based on total number of items purchased and current market prices were used However, because of the difficulty of recall, these numbers should be taken as relative assessments that provide approximate, not exact, amounts of expenditures Households were asked to estimate their expenses for the previous 12 months in a number of categories, including agriculture (buying seedlings, pesticides, fertilizers, irrigation, harvesting and transport), forestry, 151 house maintenance (electricity, repair and upgrade, household goods), food costs, schooling costs, health costs, agricultural and land taxes and miscellaneous travel, ceremonial and other expenses These expenses were compared with income figures to provide general estimates of household welfare Non-parametric t-tests (Mann-Whitney U tests) and nonparametric one-way analysis of variance (ANOVA) (KruskalWallis tests) were used to interrogate the variations between households in their forest use and income to determine characteristics of forest-dependent and non-forest-dependent households The non-parametric Spearman rank correlation was used to study associations among variables, and ordinary least squares (OLS) multiple regressions were used to build models of the household characteristics associated with forest income earnings, both in terms of absolute forest income and the relative share of overall household income RESULTS Household characteristics Household size was relatively even across the five study villages, with a mean size of 4.8 members (SD 1.5) Household heads on average had completed 6.6 years of schooling, and were 45 years old Sixty-three per cent of the households identified themselves as ‘migrant households’, meaning the household head had been born other than at the current place of residence, with most having moved 200 km or less The mean annual income in cash for households was 710 031 VND (US$ 325) while the absolute income including subsistence activities was 408 938 VND (US$ 442) Average landholding was 15.4 (0.77 ha), of which 62% was used for agriculture and the rest for residences and forestry Forest products collected Eighty-eight per cent of households had harvested some sort of forest product from around their villages in the past year Households on average collected 5.5 different wild species (SD 7.0) Of 10 major forest product categories, fodder was the most difficult to quantify, as several households indicated they let their animals graze in the Reserve for several months per year, but it was not possible to quantify this in terms comparable with other plant extraction activities Thus, total income figures are underestimations, as they not include fodder that was not cut and brought to the home (which could be measured and recalled) In most cases, forest products were collected for both subsistence and commercial use, with the exception of charcoal, which was produced solely for commercial sale (Table 1) The most commonly collected forest product for subsistence was fuelwood, which 76% of households obtained from natural forests, other households sourcing it from private gardens or purchase The most common commercial products were fuelwood and leaves, both sold by 35% of households The most lucrative income-generating product collected from 152 P D McElwee Table Average cash income generation from forest-based sources Forest-based activity Fuelwood Leaves Fruits Timber Rattans/bamboos Charcoal Medicinals Fodder Honey Aromatics/oils/others Animals Total number of households collecting (% of total) 79 (76%) 54 (52%) 36 (35%) 27 (26%) 27 (26%) 19 (18%) 14 (13%) (8%) (5%) (4%) (3%) Number of households selling (% of total) 36 (35%) 36 (35%) (8%) 23 (22%) (5%) 19 (18%) (7%) (3%) (1%) (1%) forests was charcoal, raising on average over million VND yr−1 (US$ 69) and accounting for 25% of the income of charcoal-making households (18% of the sample) Overall, 57% of households surveyed raised at least some cash income from collecting forest products On average, all non-wood products contributed 190 952 VND (US$ 13) in cash value per year to all surveyed households, while wood products (fuelwood, timber and charcoal) contributed 469 173 VND (US$ 32) in cash If we exclude those households who reported no cash forest income at all, the mean annual non-wood forest product income was 336 593 VND (US$ 23) and the mean wood income was 827 016 VND (US$ 57) Adding subsistence goods, the values are even greater: the mean forest environmental income was 137 649 VND (US$ 78) from all sources averaged across the sample, and was US$ 90 for only those households reporting some forest use While absolute forest income may seem small, as a percentage of total income it was a significant contributor to households in Ke Go, who reported very low total incomes overall (Table 2) Forest environmental income was the second-highest in absolute terms among all household cash income sources, and was third among all income sources Average cash obtained by selling product (VND yr−1 ) for those households with cash income from product 409 417 365 681 178 438 510 870 184 000 173 947 232 429 733 333 20 000 500 000 Average % contribution to overall household cash income for households that sold product 9.8 8.4 4.9 12.1 4.2 25.0 5.2 11.5