Links between forests and poverty in Indonesia. What evidence? How can targeting of poverty in and near forests be improved? CESS-ODI Briefing Paper II, March 2005 Summary Available data of forests and poverty in Indonesia suffers from weak sampling in remote areas, lack of detail on the support that households derive from forests (particularly subsistence), as well as use of culturally inappropriate indicators. However, with available data, it is still possible to show that villages and households in and near forests tend to be worse off, in income- and non-income terms. Rural households below the poverty line are more dependent on forest income than those above. Poverty reduction programmes in Indonesia do not clearly differentiate approaches to tackling chronic rural poverty. The overall objectives are usually “empowerment of target groups“ or “increasing income of target groups”, but these are not quantified, and data collection on target group impact is limited. Programmes within the forestry sector, lack meaningful impact because the Ministry has neither the expertise nor the mandate for poverty reduction. Interests in conservation, and production often override the livelihood needs of local people. Interventions such as community-based forest management are designed to address all local stakeholders, so specific targeting of poor and vulnerable groups (including differentiated incentives) is often weak. Rehabilitation schemes are often weak in their understanding of how farmers can undertake tree-planting which benefits them as well as the degraded land to be reforested. Ways forward include enhancing the representation of remote areas and forest content in national data collection, poverty mapping to assist targeting at kabupaten or watershed levels, as well as greater use of participatory assessment to understand poverty and vulnerability in different contexts, and the role of forests in poor peoples’ livelihood strategies. Participatory approaches to identifying the poor will vary from area to area, depending on local culture and livelihoods systems. In some cases formal wealth ranking is appropriate (e.g. where inequality is high). In most areas, poverty and vulnerability need also to be understood in a more embedded way, in terms of secure access and control over land and natural resources. Implications for forest policy include reforms to spatial planning, differentiated management criteria and accessible decision-making structures. 1. Links between forests and poverty poor. The unit of analysis is the province. The analysis shows that there are 48.8 million a. How much envidence if there already? - people living in the national forest estate, of the Literature whom 9.5 million are poor. Meanwhile across Indonesia, only 27.1 million people live on land Many people observe an apparent link still with tree cover (much forest-estate land is between forest and poverty, but it can be treeless). Of these, 5.5 million are poor. relatively difficult to gather strong data to support this observation. Various studies also Sunderlin, et. al. (2000)2 argue that there are mention the high numbers of poor people in around 20 million people living in and near forest areas without pointing to empirical forests areas in Indonesia, of whom around 6 evidence. Studies that have tried to show a million depend on forest resources. link between poverty and forest in Indonesia include: b. Results from analysis by CESS Brown (2004)1 who analysed numbers of people living in the national forest estate in The limited availability of data showing the link Indonesia and, of these, numbers who are between poverty and forests constitutes the 1 main obstacle. This reflects lack of that villages in and near forest areas are government attention to the poor in forest worse off in terms of available infrastructure areas, as is evident from the failure to compared to those not close to forests and the accommodate poverty in forest areas within provincial average, and therefore poorer. This the National Strategy for Poverty Reduction data covers some 68,816 villages across the (SNPK). country. CESS-ODI, using formal national data, have Poor villages according to the availability of therefore attempted to analyse and infrastructure, comparing villages inside understand further the link between forests and outside forest areas, and the provincial and poverty in Indonesia (Seldadyo, 2003; average (2003) 3 Bachtiar et al, 2004) . 100 90 1) Correlating rural poverty with forest area 80 70 n 60 Variables used are percentage rural poverty e s r versus percentage forest cover, by province. pe 50 40 Although there is a weak overall correlation 30 between rural poverty and forest cover, the 20 correlation is most strongly seen in provinces 10 0 with high forest cover and high rural poverty Sumatera Jawa & Bali NT Kalimantan Sulawesi Maluku Papua INDONESIA Hutan (%) 62.98 40.74 66.51 76.74 58.33 76.61 90.02 60.89 Luar Hutan (%) 49.34 37.15 54.23 57.57 45.24 71.33 76.65 49.61 (0.77) [Quadrant 1], and provinces with low 52.94 37.97 56.43 65.76 48.60 74.03 85.51 53.16 Propinsi (%) forest cover and low rural poverty (0.72) 4 [Quadrant 2]. However, other data are 3) Poor households and forests: non-income needed to clarify the cause-and-effect variables relationship between poverty and forest. Here, 2003 Podes data from BPS is integrated The correlation between rural poverty and with , as well as 2003 data from the Family forest cover, by province (2003) Planning Agency (BKKBN) is used. Podes data is disaggregated according to location 60 (in/near and far from forest areas) and then 50 a integrated with BKKBN data on “pre-welfare” s Q4 Q1 40 and “welfare I” households.This data covers n di de i k 50,152,320 households across the country. 30 20 nd. mis The results of this analysis are as follows: % pe 10 Q3 Q2 • The percentage of poor households in 0 0 1020304050607080 villages in and near forests is greater that % wilayah hutan for villages far from forests. • This pattern is repeated across all areas, including Java. % % Pend. % Forest Wilayah Miskin di Coverage Hutan Desa Percentage of poor households according to % Rural 0.473 BKKBN criteria, comparing villages inside MEDIAN 35.56 19.30 Poverty (Pearson Corr) and outside forest areas, and the provincial average (2003) 80 2) Poor villages and forests: infrastructure and services 70 60 50 Here, 2003 Village Potential Data (Podes) n e s r from the Central Statistic Agency (BPS), is pe 40 used. Indicators include availability of: 30 transport, electricity, telephone, education, health, garbage disposal, markets and formal 20 credit institutions. The 2003 Podes data allows 10 0 these variable to be disaggregated and scored Sumatera Jawa & Bali NT Kalimantan Sulawesi Maluku Papua INDONESIA Hutan (%) 36.92 41.74 66.09 40.09 46.29 50.66 69.69 43.98 according to whether villages are in/ near Luar Hutan (%) 31.98 32.82 63.58 29.49 39.40 41.21 55.00 36.57 Propinsi (%) 32.94 33.90 64.15 31.89 40.89 43.19 62.27 38.09 forests, or far from forests. The analysis shows 2 4) Poor households and forests: income variables 5.00 4.50 4.00 Here Podes (2003) and 2002 National Socio- 3.50 Economic Survey (Susenas) data from BPS is 3.00 2.50 ase used. Podes data is disaggregated according t sen 2.00 Per to location (in/near and far from forest areas) 1.50 and then integrated with Susenas data (poor 1.00 households according to the provincial poverty 0.50 0.00 line). Due to a limited research budget, this NTB NTT Kalbar Sulsel Sulut Sultenggara Gorontalo Desa 2.2029 0.9999 4.6016 0.5549 0.2646 0.7175 0.4563 analysis only covers 43,294 households in Kota 0.1346 0.0000 1.1035 0.0000 0.0741 0.0743 0.0528 seven provinces: West Nusa Tenggara, East Nusa Tenggara, West Kalimantan, North Sulawesi, South Sulawesi, Southeast • In rural areas, poor household are more Sulawesi and Gorontalo. dependent on forest income that richer households; The results show that, consistent with BKKBN • However, overall, dependency on forest data, the percentage of poor households for income is apparently very limited, which is village in and near forest areas is higher likely to mainly a result of Susenas compared to those far from forest areas in sampling criteria (see Section 3). each of the seven provinces examined. Percentage household income from forests: Percentage of poor households according to poor vs. rich in seven provinces (2002) Susenas data, comparing villages inside and outside forest areas in seven provinces (2003) 5.00 4.50 100.00 4.00 90.00 3.50 3.00 80.00 ase t 2.50 rsen e 70.00 P 2.00 60.00 1.50 1.00 50.00 0.50 40.00 0.00 NTB NTT Kalbar Sulsel Sulut Sultenggara Gorontalo 30.00 Miskin 1.30 1.20 4.40 0.20 0.40 0.70 0.50 Tidak Miskin 0.00 0.30 0.80 0.00 0.00 0.10 0.00 20.00 10.00 0.00 NTB NTT Kalbar Sulut Sulsel Sultenggara Gorontalo Hutan (%) 89.34 86.66 85.03 72.04 81.73 81.91 79.17 Based on the analyses above, it can be Luar Hutan (%) 77.46 77.25 70.84 67.16 73.64 75.15 71.27 concluded that there is indeed a link between poverty and forests in Indonesia. However, 5) The contribution of forests to household tackling poverty in forest areas requires more income in-depth analysis, and stronger empirical evidence. This is needed to develop more Here, 2002 National Socio-Economic Survey effective and targeted poverty reduction (Susenas) data from BPS is used, covering policies for forest areas.