Rate and Causes of Deforestation in
Indonesia:
|
| [Back to
OccPaper Top Page] [Chapter 1] [Chapter 2] [Chapter 3] [Chapter 4] [Chapter 5] [Chapter 6] [Chapter 7] [Chapter 8] Tables [Table 1] [Table 2] [Table 3] |
SMALLHOLDERS In this section, we review recent literature on the role of rural smallholders in forest cover change in Indonesia. This review covers: the "shifting cultivation-forest pioneer" continuum; smallholder tree crop production; regular transmigration; spontaneous transmigration; and the role of population density in forest cover loss.
Shifting cultivation-forest pioneer continuum "Shifting cultivation" means entirely different things to different people. This terminological confusion undermines efforts to understand the relationship between smallholder farming systems and forest cover change. Weinstock and Sunito (1989: 5) observed that "[e]ven after wading through the plethora of terminology one is confronted with a vast array of definitions of this form of agricultural production, each definition having a different view as to what types of activities should or should not be included". Some uses of the term refer only to traditional, long-fallow, sustainable systems of cultivation; this practice is often said to be associated with the conservation of primary forests. Other authors use "shifting cultivation" to mean practices of rotational agriculture that may or may not have long fallows, and that may or may not be associated with long-term conservation of forests. Still others equate "shifting cultivation" with smallholder agriculture generally in the outer islands - a use of the term that neglects the fact that many smallholders do not practise any systematic form of rotations. Weinstock and Sunito (1989: 20-21) suggest a fundamental distinction between "shifting cultivators" and "forest pioneers". [7] "Shifting cultivators" are defined as people "who practice a form of rotational agriculture with a fallow period longer than the period of cultivation... Unless faced with population pressure or other constraints, land is used only one to three years and fallowed for a relatively long period (up to twenty or more years)" (p. 20). "Forest pioneers" are defined as people "who may utilise slashing and burning of the existing vegetation but with the primary intention of establishing permanent or semi-permanent agricultural production. Although some subsistence food crops may be planted, the planting of cash crops (most often perennials) is the primary focus of attention. Land is usually not fallowed but is used continuously and is abandoned only after total or near total exhaustion of the native fertility of the soil since there is no long term plan to again return to the same site" (p. 21). In order to adequately understand the role of smallholders in forest cover change in Indonesia, it is critically important to acknowledge a continuum of farming systems running from traditional shifting cultivation (involving long fallows and long-term forest conservation) at one extreme, and "forest pioneer" cultivation (often involving long-term degradation and deforestation) at the opposite extreme. In Indonesia, an ideological polarisation exists whereby government representatives and environmental NGO representatives tend not to recognise the continuum. The government position condemns "forest pioneer" cultivation (perambahan hutan) as environmentally destructive and often does not differentiate it from shifting cultivation. If the government is serious about forest conservation, then it must support those traditional farming systems that are consistent with the aim of long-term forest conservation. [8] Conversely, environmental NGOs tend to focus their attention on traditional shifting cultivation [9] to the exclusion of less sustainable farming systems. It appears environmental activists fear that acknowledging unsustainable smallholder farming systems will undermine their claim that people at the forest margin possess wisdom (kearifan) in forest management. Acknowledgement of the full continuum of farming systems would strengthen, not weaken, the NGO position for two reasons. First, it would demonstrate recognition of the complexity of rural social change. Second, it would demonstrate concern for forest pioneers, who are every bit as much victims in the process of rapid social change as are shifting cultivators. An important milestone in better understanding the role of shifting cultivation in deforestation was achieved through Dick's (1991) critique of the World Bank (1990) and FAO (1990) studies. On the basis of the following calculation, the World Bank (1990) asserted that shifting cultivation has by far the largest role in deforestation. The area reported to be under shifting cultivation for three provinces of Indonesia in 1990 was 14 million ha in Sumatra, 11 million ha in Kalimantan and 2 million ha in Irian Jaya (RePPProT data, as cited in World Bank 1990: 3). The total area of 27 million ha expands at the annual rate of 2% implying, according to the World Bank (1990: 3), deforestation of roughly 500,000 ha per year - by far the largest cause of deforestation. FAO's (1990) assumptions were approximately the same. Dick (1991) criticised the underlying assumptions of these studies pointing out that they lump together under the terms "smallholder conversion" or "shifting cultivation" both "traditional shifting cultivation", which he views as having relatively benign practices, and "spontaneous transmigration" which he views as having unsustainable practices and accounting for the largest share of deforestation. [10] He says that given this distinction, traditional shifting cultivators account for 21 % of total deforestation, rather than the largest share, as claimed in the World Bank and FAO studies (p. 32). He believes that even this, however, overstates the contribution of traditional shifting cultivation to deforestation because many of the forests they clear are part of a long-standing rotation on clan lands (p. 32). He points out, moreover, that traditional shifting cultivators "lack the tools necessary to convert all but the most open primary forests" (p. 27). The later World Bank assessment of deforestation, as mentioned earlier, has a much more favourable outlook on the practice of traditional shifting cultivation. The report states that shifting cultivation may be less damaging than previously thought (World Bank 1994: 19). Colfer with Dudley's (1993) research on four forest communities in East Kalimantan also encourages a more favourable outlook on traditional shifting cultivators, pointing out that they are capable of sustainable resource management and that their agricultural systems, although producing low yields of rice, actually produce a wide range of valuable goods (timber, NTFPs, foods, medicines) that tend to be overlooked in other studies. There is considerable lack of clarity, however, as to whether traditional shifting cultivators are a significant presence in the outer islands. The World Bank (1994: 19) refers to recent research in claiming that "traditional communities may be much larger than previously thought". Other observers, however, state that traditional shifting cultivators are few in number and are undergoing rapid transformation of their farming systems. Tomich and van Noordwijk (1995: 3) say that traditional shifting cultivation has "virtually disappeared" in Sumatra. Potter (1993: 109) refers to a WWF study showing that "traditional" systems have been modernising, for example with the use of chainsaws. Kartawinata et al. (1989: 603) observe that in some areas nominally under shifting cultivation in East Kalimantan, producers established perennial crops that were maintained for ten or more years. What are we to make of such widely diverging views on the characteristics of the basic characteristics of shifting cultivation - and more generally of smallholders - in the outer islands of Indonesia? It is clear that understanding the process of forest cover change in Indonesia would be well served by more careful categorisation and also characterisation of various kinds of smallholders in forest communities. Weinstock and Sunito (1989: 37-38) found that information is needed in the following areas: (1) land under shifting cultivation classified according to different land forms (slope, etc.); (2) rates of change in the area of shifting cultivation over time; (3) changes in patterns of shifting cultivation over time; and (4) data on the inflow of pioneer settlers along logging roads and the impact they have had on forest lands. These kind of data, as well as data on tree crop producers (see below), would be quite helpful. Recent research has made inroads into this area of knowledge, but much remains to be done. [11]
Smallholder tree crop production In addition to "shifting cultivators" and "forest pioneers", we must distinguish another basic farming system category in the outer islands: smallholder tree crop producers. Smallholder tree crop production is often understood within the general term "shifting cultivation", because many shifting cultivators produce tree crops. However, tree crop production should be seen as a distinct analytic category because, although it is strongly associated with shifting cultivation, it tends to be carried out on different kinds of lands and follows an entirely different logic of production (Dove 1993). Smallholder tree crop production has important implications for forest cover because it is often established in forest clearings and because it has grown greatly in recent years. Barlow and Tomich (1991: 32) note that about 20% of all agricultural land in Sumatra and Kalimantan is under tree crop production. In 1994, there were 8.89 million ha of the three main tree crops in Indonesia, with the following distribution: rubber (39%); coconut (41%); and oil palm (20%) (PDP 1996). The apportionment of the area of these crops among the major provinces in 1988 was: Sumatra (55%); Java (15%); Kalimantan (13%); Sulawesi (9%); and Nusa Tenggara (5%) (Dick 1991: 25). (Note that these data refer to smallholder, estate and plantation production.) Indonesia is the world's second largest producer of natural rubber (about three-quarters produced by smallholders), the second largest producer of palm oil (most from estates), the world's third largest producer of coffee (95% of production by smallholders), and the fourth largest producer of cocoa (Economist Intelligence Unit 1995a: 29-31). Almost all coconut/copra production in Indonesia is by smallholders (World Bank 1996: 164). There may be a strong association between smallholder tree crop production and deforestation. Chomitz and Griffiths (1996) found that tree crops, rather than subsistence-oriented shifting cultivation, also plays a major role in deforestation in Indonesia; among various tree crops, rubber appears to have the strongest association to deforestation. Rubber is the largest single earner of agricultural income in Indonesia (US$1.5 billion in 1994-95) and output value almost doubled in the 1984-1995 period (World Bank 1996: 141). In the 1982-1994 period, production of tree crops increased by the following rates: rubber from 900 to 1,499 tons (66%); coconut from 1,718 to 2,631 tons (60%); and coffee from 281 to 446 tons (55%) (World Bank 1996: 163). It is possible the planting of rubber trees may increase in association with their function as a property marker, in areas where land competition is strong. Dove (1993: 142) says that in the Kantu area of West Kalimantan, "[p]lanting of rubber is perhaps most important as a means to establish tenure... this tactic is used against land usurpation not only by fellow tribesmen, but also by the government". Angelsen (1995: 1724-1725) observes that in Riau province, Sumatra, the planting of rubber is a way to "obtain and secure land rights, both according to customary and national law".
Transmigration Census data from 1990 show that 108 million (60%) of Indonesia's 180 million people live on Java and Madura islands (Economist Intelligence Unit 1995a: 11), an area amounting to a mere 7% of the land surface of the country. Java has one of the densest concentrations of population anywhere in the world. There have been efforts dating back to the early 20th century to encourage families to move from Java to the outer islands in order to relieve population pressure, poverty and land degradation. There are two types of transmigrants: "regular" transmigrants receive full government assistance, whereas "spontaneous" transmigrants receive partial or no government assistance (see below). [12] There has been a tremendous acceleration in the number of transmigrant families in recent decades. In the period 1950-1979, there were an average of 6,570 regular transmigrant families each year; in the period 1980-1984, the annual average rose to 73,200 families. [13] Dick (1991: 27-29) observes there is a large divergence between two sources of information on numbers of transmigrants. In the period 1974-1989, the Ministry of Transmigration (MOT) claims there were 664,000 regular transmigrant families, whereas the Transmigration Advisory Group (TAG) claims there were 323,000 such families (Dick 1991: 28). In this same period, MOT estimates there were 455,000 spontaneous transmigrants; TAG provided no estimate. The approximate distribution of the destinations of transmigrants during REPELITA V (1989/90-1993/94) was: Sumatra (53%); Kalimantan (23.5%); Sulawesi (12.6%); Irian Jaya and Maluku (9.2%); and other locations (1.75%). [14] It is projected that in the period 1994-1999 (REPELITA VI), there will be an additional 350,000 regular and 250,000 spontaneous transmigrant families (MOF 1995: 2).
Effects of regular transmigration The effects of regular transmigration on forest cover in the outer islands can be classified into three categories: the direct effects of forest cover removal for the transmigration site; the movement of transmigrants from their designated sites because of insufficient incomes; and the land pressures induced by transmigrants on neighbouring non-transmigrant households. 1. Direct forest cover removal. In the mid-1980s, international NGOs said the transmigration programme was responsible for a large share of deforestation in Indonesia (see especially Secrett 1986). In the aftermath of that accusation, some claimed the amount of deforestation attributable to regular transmigration was grossly exaggerated. Whitten (1987: 241-243) uses provincial data on forest cover and on land clearance for regular transmigration to show that, in 1979-84, the official programme led to the conversion of less than 1% of total forest area. [15] Whitten estimates that 30-50% of land cleared for transmigration was originally forested. This correction to past over-estimations runs the risk of going too far in the other direction. An adequate calculation of the environmental impact of transmigration must include an assessment of not only the land cleared for regular settlement, but also of land cleared by regular settlers who are not able to obtain sufficient income on lands assigned to them, and also of land cleared by "spontaneous" settlers. Whitten (1987: 242) notes that "unassisted migrants" outnumber regular ones by roughly two to one, and dismisses the importance of much of the forest conversion, saying it was land that was zoned for conversion to agriculture anyway. While this may be true, it is important to bear in mind that there has been no rational and objective basis for the designation of conversion lands (Dick 1991: 19). Dick (1991: 30) also challenges exaggerated estimates of forest clearance through transmigration. He contends past estimates of deforestation attributable to regular transmigration have been grossly exaggerated because they assume all land allocated for future transmigration use has been converted to such use, whereas in fact much of this land has not been converted. His own calculation of deforestation attributable to regular transmigration in the period 1979-1989 is 78,500 ha per year. He bases his calculation on the assumption that about 3.0 ha of forest land are cleared for each regular transmigrant family - 2.0 ha of this for agriculture and 1.0 for infrastructure (pp. 31-32). [16] Dick's calculations err in two ways. First, he equates land clearing with deforestation, yet we know that some of the land cleared for transmigration was not deforested (see above). Second, Dick neglects to follow through on his own insight (p. 29) that 2.0 ha of agricultural land for each family was insufficient, and that further deforestation will result once the occupants have exhausted the land. The first error tends to over-estimate and the second to under-estimate the amount of forest cleared as a result of regular transmigration.
2. Movement because of insufficient income. Through the mid-1980s, regular transmigration sites aimed to produce food crops (mainly rice) on 2.0 ha of land per family. This proved to be inadequate at many sites both because the area of land was insufficient, and because the relatively poor soil did not support low-input rice cultivation. In the mid-1980s it was found that 40-80% of income in regular transmigrant households was from off-farm work (World Bank 1988: xxvii). Many transmigrants are caught in a vicious cycle, whereby they search for off-site income because site-based income is inadequate, and then the potential of the site-based income is not fully realised because they must work off-site (RePPProT 1990: 41). There is little information available on what proportion of regular transmigrant families turned to shifting cultivation or became forest pioneers. [17] It has been pointed out that many transmigrants practise an unsustainable form of shifting cultivation because they lack knowledge about traditional forms of shifting cultivation and because they lack tenure security. Since the early 1980s, the transmigration programme has turned increasingly to tree crops (rubber, coconut, oil palm, etc.) in place of food crops, on the assumption that this is generally a more suitable and profitable form of land use in the outer islands. The turn to tree crops is also important because it contributes to the effort to diversify to non-oil exports (RePPProT 1990: 41). Tentative projections showed that smallholder net income would be twice as high on tree crop schemes as compared to food crop schemes once fully developed (World Bank 1988: xxxix). No national data are available showing the extent to which the trend to tree crops may have alleviated the tendency to seek income outside the transmigration site. Case studies from Riau province in Sumatra, however, present a discouraging picture (see Holden and Hvoslef 1995; Holden et al. 1995).
3. Land pressure induced by transmigration. Transmigration settlements are often established adjacent to, or sometimes in the midst of, existing communities of shifting cultivators. Settlements thus can reduce the amount of land available in traditional clan rotations for shifting agriculture. Combined with increasing indigenous population pressure, this can contribute to the problem of shortened fallows, over-use of poor soils, the turn to sedentary agriculture and speculative land acquisition. Angelsen (1995) has observed this tendency in Riau province, Sumatra.
Effects of spontaneous transmigration Dating from the mid-1980s, it seems there was substantial growth in the number of spontaneous transmigrants as compared to regular transmigrants (Dick 1991: 27-28). This appears to have happened for two reasons. First, infrastructural development for the regular transmigration programme in the early 1980s attracted spontaneous migrants. Second, the regular transmigration programme was scaled back in 1986 because of declining oil revenues and because a decision was made to devote scarce funds to upgrading infrastructure and production systems at existing transmigration sites (World Bank 1988: xxii-xxiii; Thiele 1994: 191). Tomich and van Noordwijk (1995: 3-4) suggest that spontaneous transmigrants have been quick to adopt the indigenous rubber production systems in Sumatra. According to Dick (1991: 29), spontaneous transmigrants clear an average of 4.25 ha per family. On the basis of this figure, he calculates that spontaneous transmigrants are now the single most important cause of deforestation, accounting for approximately 178,500 ha per year of lost forest cover. This may be an over-estimate seeing that - as in the case of regular transmigrants - Dick assumes that all the land cleared is forest land, whereas this is not necessarily the case. Among the main questions that surface from the foregoing information are the following: (1) How has the effect of transmigration on forest cover changed with the turn away from food crops and towards tree crops? (2) Has this change increased the stability and profitability of transmigrant income? (3) If it has done so, does this in turn induce a larger stream of spontaneous transmigrants? (4) Does a growing number of transmigrants necessarily undermine traditional shifting cultivation and how is this question related to the issue of land markets, formal tenure and traditional (adat) land rights?
Population density Fraser (1996) makes the case that growth of population density is the fundamental explanation for deforestation in Indonesia. Data on population density by province in Indonesia show a strong inverse relationship with forest cover (Table 3). Several writers have observed this relationship (FAO 1990: 10; Barbier et al. 1993: 7; Fraser 1996). [18] Fraser (1996: 6, 15) states that for every 1% increase in population (the rate in the outer islands is 3%), there is an approximate decrease of forest cover of .3%. The overall annual decrease of forest cover is thus 1%, which corresponds to the annual forest cover loss of 900,000 to 1 million ha observed by FAO. There is no doubt that a growing human presence in rural Indonesia has a role in deforestation. It remains to be demonstrated conclusively, however, that growing population density is the main cause of forest cover removal, as claimed by Fraser (1996) and others. It is not clear to what extent the high inverse correlation between population density and forest cover is causal, and to what extent it is incidental. It is possible there are other variables, excluded from Fraser's bivariate measure, that contribute to high correlation. Some writers have noted that in South-east Asia - and in the humid tropics in general - there is a typical sequence whereby forests are first cleared by loggers and then occupied for agriculture (Grainger 1993: 70; Kummer and Turner 1994). If this is the case, then it is possible to argue that logging is a "cause" of deforestation and shifting cultivators merely fill a "vacuum" created by the loggers, in those areas where this sequence takes place. This line of reasoning reverses the path of causation: opening of forests leads to an increase of population density in a given area, not the other way around. [19] There is another weakness in Fraser's neo-Malthusian argument. There are various independent variables (technological change, distribution of wealth and income, demand for agricultural products, growth of infrastructure, level of women's education and participation in the labour market, etc.) that modify the influence of population on forest cover. If these variables are taken into account, the apparently fundamental role of population growth in the loss of forest cover is not so clear. It is possible, for example, for rural population to stay the same or decline (due to urbanisation), and for forest cover removal to increase because of higher per capita urban or international demand for agricultural production. Non-Malthusian explanations, however, also have their weaknesses. Angelsen (1995) uses case study data from Sumatra to make the point opposite to that being made by Fraser (1996) - namely, that increasing population is not necessarily the main cause of deforestation. He weights the factors causing annual forest clearing and finds the following: (1) the changing proportion of households opening swidden lands accounts for 70% of the total; (2) the total population of households accounts for 23% of the total; and the average size of swiddens accounts for 7% of the total (Angelsen 1995: 1724). This kind of analysis suggests that population growth explains approximately one-quarter of forest cover removal. Is it possible, however, that as in the case of Fraser's (1996) analysis, an alternative view of the linkages among relevant variables can produce a different conclusion? Is it possible, for example, that the increasing proportion of households opening swidden lands (said to be 70% of the explanation) is doing so in response to growing local population density and pressure? It is interesting to note that in Angelsen's (1995) survey, three-quarters of the farmer respondents believe that population growth explains shortages of land. Further research on the relationship of population to deforestation would be useful, but it must be done in such a way as to clearly distinguish causal and incidental relationships, clearly account for all significant paths of causation, and adequately define what forms of agency are being assumed. Perhaps the most important step towards raising the sophistication of such analyses is to ask what causes increasing population growth, population density and migration - that is to say, to treat population as an intervening variable, rather than as an independent variable. |