Society and Natural Resources, Volume 7, pp. 429-443 0894-1920/94 $10.00+ .00 Printed in the UK. All rights reserved. Copyright © 1994 Taylor & Francis

Soil Conservation and Poverty: Lessons from Upland

JILL M. BELSKY Department of Sociology University of Montana Missoula, Montana, USA

Soil conservation efforts in Indonesia since the Dutch colonial era have focused on in- troducing bench terraces—a costly soil conservation method for poor, upland farm- ers. Data from two villages in the Kerinci uplands of illustrate that even with state underwriting of bench terrace construction, farmers across all economic strata still resist using this method. Why the state has not pursued alternative soil conserva- tion approaches—especially ones that entail the "conservation farming " approach and that can better build upon the diversity of upland farming systems—is discussed in the context of the state's emphasis on productivist and commodity-led agricultural development and on broader geopolitical institutions and forces that perpetuate this approach. Given these constraints, state underwriting of soil conservation for poor farmers (i.e., providing "landesque capital" in Blaikie and Brookfield's 1987 termi- nology) suggests undue hope through economic remedies and the ability of the state to implement environmental and social reform, especially to benefit the poor.

Keywords Indonesia, political ecology, political economy, poverty, soil conserva- tion, terraces

Since the Dutch colonial era, soil conservation efforts in upland Indonesia have empha- sized the introduction of bench terraces. However, the long-term use of agricultural ter- racing by dryland farmers on sloping lands in Indonesia (as well as throughout Southeast Asia) has been varied and often hotly contested (Pelzer, 1945; Chapman, 1975). Assess- ments of soil conservation efforts in Indonesia have gradually come to place great stress on the role poverty (including land tenure insecurity) plays in creating obstacles to the construction and use of bench terraces by poor farmers. This growing awareness has led to the creation of national programs involving public underwriting of terrace construction costs (Ramsay & Wiersum, 1974; Departemen Kehutanan, 1985; Carson, 1989); and to proposals for the state to reorient soil conservation to focus more on "conservation farm- ing," which, among other benefits, can better build upon the diversity of upland farming systems in Indonesia (Hudson, 1988; Barbier, 1989; Arsyad, 1992). Using current and historical data from the Kerinci district in Sumatra, this article ex- amines the resistance of upland farmers to dry land terracing and shows that reasons for

Received 26 October 1992; accepted 16 November 1993. The author thanks Stephen Siebert, Frederick Buttel, E. Walter Coward, Syafruddin (RLKT/Kabupaten Kerinci), Annet van Breugel, and three anonymous reviewers for their assis- tance. The research was funded by a Mellon Grant and Collaborative Fulbright Award, which the author shared with Stephen Siebert. Address correspondence to Jill M. Belsky, Department of Sociology, University of Montana, Missoula, MT 59812-1047. 429 430 J. M. Belsky rejecting this method include, but are not limited to, constraints posed by economic fac- tors (including insecure land tenure). The article then addresses the reasons why the state in colonial to postcolonial Indonesia has promoted a soil conservation technology that creates serious constraints for the majority of poor, upland farmers. Rather than merely representing an "inappropriate" technology choice (possibly based on limited informa- tion), the article argues that the state's soil conservation policy must be viewed within a broad political economy framework, the dominance of the productivist and commodity- led agricultural development paradigm, and the geopolitical institutions and forces that perpetuate this approach. Given these constraints, state underwriting of soil conservation for poor farmers (providing "landesque capital" in Blaikie and Brookfield's 1987 termi- nology) suggests undue hope through economic remedies and state action for the benefit of the poor. Sociological theories need to be less economically reductionist and question the capacity of the state to implement environmental and social reforms.

Sociological Theories on the Relationship Between Soil Conservation and Poverty Beginning in the 1980s, there was an important analytical shift away from viewing agri- cultural change and soil conservation use as the result of behavioral and demographic variables to viewing the relationship as influenced by economic and structural factors. In Piers Blaikie's political economy approach to soil conservation (1985), limited access to and control over productive resources creates a "squeeze" on the poor household's ability to survive or reproduce itself, and to make investments in land capability. The "reproduc- tive squeeze" leads households to a number of strategies including diversifying economic activities, going into debt, outmigrating, and overexploiting soil resources to increase production levels (Bernstein, 1981). That poor farmers are (rightfully) more concerned about daily survival and the short- term benefits they can derive from mining soil resources, rather than the benefits they can obtain from investing in long-term conservation practices, has become a popular way to view soil conservation among social scientists from competing theoretical perspectives and even soil conservationists (for example, Watts, 1983; Redclift, 1984, 1987; Baker, 1984; Blaikie, 1985; Ashby, 1985; Blaikie & Brookfield, 1987; Chambers, 1987; Hudson, 1988; Stocking, 1988; Collins, 1987,1988; Repetto, 1985; Leonard, 1989; Barbier, 1990). While the political economy framework has been credited for bringing a powerful "sociology" back into human ecological studies (Ortner, 1984), it has also been subject to revision (even by Blaikie himself). Three issues are especially noteworthy. One, it has been criticized as overly anthropocentric and focused on social causality (Buttel & Sun- derlin, 1988), minimizing—indeed ignoring—the ways variations in particular physical environments and specific technologies affect producers' land use practices. This is dan- gerous because "ignoring ecological variation and aggregating across different technolog- ical innovations only serves ... to confound interpretation of farmer adoption behavior" (Ashby & Coward, 1980, p. 523). A second criticism is that theories focusing more or less exclusively on economic and structural factors often presume that all societies, and especially all agrarian groups, experience environmental degradation similarly as a result of their contact with capital- ism. This view insufficiently recognizes the variable ways peasants resist and alter capi- talist structures and respond to environmental disruptions. Three, the role of the state in political economy studies of soil conservation remains ambiguous. On one hand, the literature documents how state actions and policies have Soil Conservation and Poverty 431 caused great social and ecological disruptions (cf. in Africa: Anderson & Grove, 1987; Bernstein, 1981; Watts, 1983; Central and South America: Hecht & Cockburn, 1989; Faber, 1993; Asia: Poffenburger, 1990; Peluso, 1992). Various countries including In- donesia have used coercion and socially regressive policies to accomplish environmental goals (Thrupp, 1990; West & Brechin, 1991; Peluso, 1993). On the other hand, state as- sistance has been identified as critical to provide environmental assistance for the poor, such as underwriting expensive soil conservation methods (Brundtland Commission, 1987; English et al., 1984; Hudson, 1981). There is great question as to whether the state can effectively manage and centrally coordinate conservation efforts and whether it is willing or able to transform the status quo (Bryant, 1991). Moreover, it is also likely that national and/or international geopolitical and institutional forces (such as the World Bank, International Monetary Fund, nongovernmental organizations or the media) will further constrain if not shape the limits of reform. Thus, Bryant (1991, p. 165) observes, Evef a statn i e is receptive to the sustainable development argument, it may still be unable to implement reform. The literature often assumes that states are somehow omnipotent, and can easily "deliver" the ecological goods. One reform widely suggested as a means to include poor farmers in sustainable de- velopment efforts is a shift away from costly engineering models of soil conservation to less expensive, agronomically based methods. Soil conservation is typically divided into two general types: agronomic and engineering methods. Agronomic practices include ground covers, mulches, contour cultivation, strip crops, grass contour barriers or bunds, and various crop management techniques such as agroforestry systems. They seek to maximize ground cover and thereby protect the soil surface from raindrop splash as well as slowing the velocity of water runoff (Moldenhauer & Hudson, 1988; Sheng, 1981; Young, 1989). The greater attention to manipulating the land surface (i.e., maintaining dense vegetation cover) as opposed to soil and water themselves has led some specialists to emphasize land use and "conservation farming" as a key to achieving (soil) conserva- tion goals (Hudson, 1988). The conservation farming approach also gains support for so- cioeconomic reasons because it generally entails lower capital and labor costs than engi- neering methods, thereby increasing the likelihood that they will be used by farmers with limited access to land, labor, and capital, as well as insecure tenure (Hudson, 1981; Mold- enhauer & Hudson, 1988; Blaikie & Brookfield, 1987; Ostberg, 1987). In contrast, engineering approaches to soil conservation that include terraces, ditches, gulleys, and dams require relatively high capital and labor inputs for both construction and maintenance relative to agronomic methods. They also frequently require the applica- tion of fertilizers to compensate for initial crop reductions caused by the mixing of infer- tile subsoils and loss in cropping area during construction (Hudson, 1981; Siebert, 1990). These costs are particularly high for non-land-owning cultivators who take a large risk in investing capital and labor into land they may not control in the future. Agricultural terraces are one of the most commonly recommended types of engineer- ing soil conservation methods. However, terraces are not widely constructed by the poor.2 This is because terracing dry-land fields offers an unfavorable mix of costs and benefits to individual farmers (Pelzer, 1945; Chapman, 1975; Amos, 1982). In Southeast Asia, Chapman (1975, p. 131) observes, "If the constraints were not so severe we would no doubt find (dryland) terracing more widespread." Public funding for income-related incentives for soil conservation has been sup- ported for three reasons. First, the economics of soil conservation are unfavorable for the private farmer, resulting in farmers' unwillingness to invest in them. For example, the in- 432 J. M. Belsky direct or "downstream" benefits of soil conservation are frequently greater than the direct benefits derived by farmers (McCauley, 1985). Second, government intervention has been necessary to overcome market failure to regulate externality effects. Third, since soil conservation is essentially a "public good," it should be partially, if not totally, fi- nanced by public funds. Government support to poor farmers to construct terraces has also been supported as a means to "capital" for the future maintenance of land value and capability. Because ter- races have an anticipated life well beyond that of the present crop or crop cycle, Blaikie and Brookfield (1987) refer to them as providing "landesque capital." They argue that the creation of landesque capital involves substantial "saving" of labor and other inputs for future production—a special benefit for typically labor-deficient poor farmers. But poor farmers will not be able to afford such an investment without external assistance because of the high initial costs. Therefore, Blaikie and Brookfield argue that provision of lan- desque capital should be provided to these cultivators by governments, and that these ef- forts have a better chance of resulting in the introduction of soil conservation than pri- vate, voluntary strategies. Whether this assumption is empirically valid or not is not known.

Study Site and Methods After reviewing government documents and interviewing officials involved in the In- donesian upland development and conservation program (discussed later), I conducted field research in the highland valley of Kerinci (1987-1988), a 10 km wide by 80 km long valley flanked by steep uplands in central Sumatra. The Kerinci region contains the headwaters of the two largest rivers in Sumatra—the Musi and the Batang Hari—and thus is of considerable watershed management importance. It is also of major conserva- tion importance, so much so that the World Wildlife Fund for Nature and the World Bank are funding activities in Kerinci-Seblat National Park to curb high rates of deforestation, forest conversion to agriculture, soil erosion, and the loss of biological diversity. Elevations in Kerinci range from 725 to 3800 m above sea level. Unlike most of sparsely populated Sumatra (58 people/m2), Kerinci is relatively densely populated (133 people/km2) (Scholz, 1983). Forest conversion to upland farms expanded considerably in the first decade of the twentieth century as a result of Dutch colonial policies to increase production. Over the course of the last century, periods of economic and political turmoil, as well as fluctuations in world commodity prices, have led to agricultural inten- sification in upland farms. Upland soils in Kerinci are predominantly red-yellow pod- zolics (Ultisols) (Siebert, 1990), a soil type considered inferior for intensive cultivation of annual crops (Scholz, 1983). Today, soil erosion from hillside farms is a serious problem (Watanabe, 1979; Siebert, 1990). I selected two villages in Kerinci for comparative study—Sungai Ning and Koto Lebuh Tinggi. In both areas, the major agricultural activities involve lowland, irrigated farming complemented by cultivation of food and cash crops in upland farms. Both villages have participated in the Indonesian state-led upland conservation and develop- ment effort known as regreening (discussed later). The major differences between the two villages involve: (1) Geographical location—Sungai Ning is situated in the uplands along the western road to Padang approximately 4 km from the central market and administrative center of Sungai Penuh, whereas the second village, Koto Lebuh Tinggi, is lo- cated 16 km from Sungai Penuh in a more outlying area. Soil Conservation and Poverty 433

(2) Farming practices—in part as a result of their geographical differences, the Sun- gai Ning uplands have been more intensively cultivated than the more distant Koto Lebuh Tinggi area. (3) Population dynamics—there has been substantial in-migration of Javanese farm- ers into Sungai Ning over the last half-century, which has contributed to a higher population density in Sungai Ning than elsewhere in the valley and to more in- tensive cultivation of annual crops such as cassava. These factors have led to more widespread invasion of Imperata grass and severe soil erosion in the Sungai Ning uplands than is found in the uplands near Koto Lebuh Tinggi. Data collecting involved key informant interviews, collection of oral histories, and randomized household surveys. The survey involved interviews with 25% of households from each village (n = 78). These data include a subset of households who participated in the government soil conservation program (« = 16). Additional historical material came from archival research in the Netherlands.

Regreening: Indonesia's Upland Soil Conservation and Development Program The government's approach to combatting upper watershed degradation on privately owned land in Indonesia is known as Regreening (Penghijauan). It evolved from the FAO-funded Solo Watershed project (1972-1976) on (Wiersum, 1976). This effort recognized the limitations of reforestation as a means to rehabilitate upper watersheds cultivated by poor farmers. Instead of reforestation, watershed management would focus on terraces, which would enable poor farmers to continue to grow annual crops, albeit more sustainably. However, government underwriting would be necessary because "the individual farmer would benefit materially, but he lacks the necessary capital to initiate such measures on his own; thus such programs would have to be carried out by govern- ment" (Ramsey & Wiersum, 1974, p. 8). The Regreening program subsidizes selected farmers to construct bench terraces and to adopt improved (annual) cropping practices on model farms on slopes no greater than 50%. Subsidized landowners were given Rp 100.0003 per hectare by the government to construct bench terraces, plant Setaria sp. grass on terrace risers for protection and fod- der, and cultivate soybean and peanut along with other crops of their choice on the terrace benches. Farmers were explicitly prohibited from cultivating cassava because it aggra- vates erosion. This was despite the fact that cassava was widely grown by poor farmers because of its low cost and low risk. For 2 years following the construction of bench ter- races, fertilizer and cash subsidies were provided to assist with terrace maintenance (ap- proximately Rp 50,000 or 30 days of labor per household). Regreening projects began in the village of Koto Lebuh Tinggi in 1984, and in Sungai Ning in 1986. To ascertain how poverty and government subsidization affected the use and mainte- nance of bench terraces, households in the study areas were stratified according to rice security. Rice security reflects the relative ability of households to produce sufficient amounts of rice to meet annual household consumption demand. The rice security mea- sure is also highly correlated with other traditional measures of stratification based on ac- cess to land, livestock, wealth, and overall well-being. Households were divided into three rice security categories: Households in the low stratum do not have access to irri- gated rice farming and must purchase virtually all of the rice they consume; those in the middle category produce roughly half of the rice they consume, largely through irrigated rice farming; and those in the high category produce all or a surplus of their rice.4 Survey 434 J. M. Belsky data report (and field observations confirm) that poverty is more widespread in Sungai Ning. Over half of the households in Sungai Ning were in the low stratum. These house- holds relied on crops grown in upland farms for subsistence food and on cash crops to provide income to purchase rice. They supplemented upland farming with seasonal wage labor and raising livestock. In contrast, only 10% of households in Koto Lebuh Tinggi were in the low category. The survey data further revealed that controlling for subsidies, household rice secu- rity in Sungai Ning explained almost 75% of who constructed terraces—households in the high stratum were 10 times more likely than those in the low stratum to construct ter- races (Table 1). But in Koto Lebuh Tinggi, household rice security was not a significant predictor of terrace construction (Table 2). Preexisting land uses (including cropping practices) and farmers' estimation about the relative profitability of different farming sys- tems on different soils and slopes were the major determinants of the decision to follow the Regreening strategy. Producers from Sungai Ning in the low stratum recognized ter- racing would provide protection against soil erosion, but they did not construct them be- cause they knew terraces would lower agricultural productivity. This is because farmers observed that crop yields declined following terrace construction because of the incorpo- ration of acid, infertile subsoils, and that they would lose about 30% of their cropping area to terrace risers. This "local knowledge" was confirmed by scientific experimenta- tion (Siebert & Belsky, 1990). Even with fertilizer additions, farmers were correct in as- suming that it would take years for yields to rebound following terrace construction (Hudson, 1981). Poor households in Sungai Ning would not endanger already low agri- cultural yields by building terraces, or assume additional risks involved in planting soy- bean and peanut. Though capable of fixing nitrogen, both soybean and peanut have higher nutrient demands than cassava and are more sensitive to drought and pest infesta- tion. These concerns are in addition to high labor and capital costs to construct terraces.5 Furthermore, farmers also recognized potential marketing problems and low prices when all of the participating farmers in the Regreening program harvested their soybeans and peanuts. Farmers from the high stratum in Sungai Ning, on the other hand, were more likely to construct bench terraces because they could afford to (1) hire nonhousehold labor, (2)

Table 1 Nonsubsidized construction of bench terraces by households in Sungai Ning

Household rice security stratum Number Percent

Low (n = 16) 1 6.3 Middle (n = 11) 2 18.2 High (n = 5) 3 60.0 Total (n = 32) 6 18.8

Note. This refers to whether a household has ever con- structed bench terraces on one or more of their upland parcels without subsidization. Out of a total of 32 households surveyed, 6 households had constructed terraces (18.8%). Gamma statistic, .74359. Soil Conservation and Poverty 435

Table 2 Nonsubsidized construction of bench terraces by households in Koto Lebuh Tinggi

Household rice security stratum Number Percent

Low (n = 4) 1 25.0 Middle (« = 11) 3 30.0 High (n = 15) 5 31.3 Total (n =:32) 9 30.0

Note. This refers to whether a household has ever con- structed bench terraces on one or more of their upland parcels without subsidization. Out of a total of 30 households surveyed, 9 households had constructed terraces (30%). Gamma statistic, .07273. purchase petrochemical inputs, and (3) cultivate soybeans and peanut. Unlike the case for cassava farmers, the new cropping patterns did not represent a large departure from their preexisting cropping practices, which included high-valued market vegetables and other cash crops (such as fruit). Therefore approximately 60% of households from the high stratum followed the recommendations of the Regreening program. In Koto Lebuh Tinggi, only about one-third of farmers not receiving subsidies, and across all household rice security strata, constructed bench terraces (Table 1). The popu- larity of terracing was not high because terracing and its recommended cropping prac- tices conflicted with Koto Lebuh Tinggi farmers' preference for cultivating upland farms with cinnamon-based relay agroforestry systems.6 Importantly, cinnamon agro- forestry was preferred by households regardless of rice security status. In relay agro- forestry systems, annual crops are grown only until cinnamon trees grow large and shade out the understory (about 4 years). When this happens, annual crop production ceases and cinnamon trees are viewed as providing the necessary soil conservation. Ter- races are viewed as no longer necessary. Also at this time, farmers shift their labor to an- other parcel and begin a new cinnamon intercrop. Cinnamon-based agroforestry systems enjoy wide popularity in Koto Lebuh Tinggi, where upland lands are still extensive, not heavily degraded, and well suited to cinnamon cultivation. Moreover, cinnamon can be harvested at any time after 6-10 years (it reaches it highest quality after 20 years), re- quires little labor once seedlings are established, and can remain in the ground until the market offers a high price. To be effective, terraces require pruning of Setaria grass planted on terrace walls or risers every 7-10 days, and regular repair of risers after heavy rains or trampling by wild animals and/or people. Unmaintained terraces can result in increased soil erosion by channeling runoff in comparison to nonterraced fields where runoff is dispersed (Hudson, 1981). But terraces in Kerinci were rarely maintained longer than 3 years—the period in which subsidies for maintenance were allocated. This was particularly true among house- holds from the low stratum in Sungai Ning, and across all strata in Koto Lebuh Tinggi, regardless of whether households received government subsidies to construct terraces. There were no statistically significant differences in maintenance rates between subsi- dized and nonsubsidized terrace builders. 436 J. M. Belsky

Farmers from the low stratum in Sungai Ning did not maintain terraces because pro- tective grasses planted on terrace risers interfered with crop growth, desiccated already poor soils, and reduced overall planting area. They complained further that repeated ap- plication of fertilizers "hardens" soil, making it less, rather than more, fertile; that soy- bean and peanut crops were not thriving (due in part to pest infestation); and that market prices of soybean and peanut were low. Females from these poor households, in particu- lar, did not want to maintain terraces because they had to forfeit income previously earned from cassava-based home industries preparing and selling cassava cakes and crackers (kerupuk), and they did not relish the extra labor burden that fell to them to re- build terrace walls and weed (there were more weeds because fertilizer was added). The additional labor was especially hard on women during periods of high-labor demand in the (wet) rice cycle when men temporarily relocate near rice fields to take advantage of wage labor opportunities. In Koto Lebuh Tinggi, terraces were not maintained among cultivators across all rice security strata regardless of subsidies after the annual crop phase. During the initial phase of annual-perennial intercropping, terrace maintenance tasks overlap with annual cultiva- tion requirements. However, once trees become established, farmers return only one to two times annually to weed the parcel; they do not wish to return more frequently to care for terraces (and they point out that trees and other ground cover provide sufficient pro- tection against erosion). At this stage, they redirect their (upland) labor to beginning the intercropping cycle on another upland parcel. If the other upland parcel is located at great distance from the terraced parcel, returning to the latter for (terrace) maintenance repre- sents a labor burden. Moreover, for middle and high strata farmers who are simulta- neously irrigated rice farmers, terrace maintenance (especially during the rainy season) conflicts with irrigated rice farming. If we look back to the history of soil conservation efforts during the colonial period in Indonesia, we find striking continuity. In the following section, we see that terraces have long been the soil conservation method favored by the colonial state in Indonesia, despite its well-recognized limitations for poor hillside farmers. In the final section, the argument is made that we must understand the state's approach to soil conservation in colonial and postcolonial Indonesia in the context of its consistent goal to increase inten- sification and diversification of upland agricultural production as well as soil conserva- tion.

Soil Conservation and Upland Development in Indonesia During the Dutch Colonial Period As early as the second half of the nineteenth century, Dutch colonial agricultural advisors in Indonesia recognized the threat of soil erosion on upland farms and recommended con- struction of bench terraces on sloping lands. K. F. Holle, the agricultural advisor in the Netherlands Indies (as Indonesia was called at the time), argued that many uplands were completely exhausted after a few years because the topsoil was washed away by heavy rains. In 1873 he prepared detailed instructions for the terracing and cultivation of coffee lands throughout the country (Nederlandsch-Indische MaatschappijrHandleiding, 1873, cited in Pelzer, 1945). The following year, in Article 5 of the 1874 Ordinance on Alien- ation of Domain Forest Land the government stipulated that permission to clear govern- ment-appropriated or "domain" land that was sloping would be awarded to farmers only on condition that they terrace the slopes and control soil erosion (Ontginnings-Ordon- nantie, Staatsblad 1874, no. 79, cited in Pelzer, 1945). Soil Conservation and Poverty 437

But Holle's instructions and the regulations of the Ordinance of 1874 were criticized by other administrators and often ignored by farmers. Therefore, in 1889, the colonial government commissioned a forester named Berkhout to evaluate the use of terracing in upland fields. Forester Berkhout's report summarized the arguments at the time for and against terracing of dry fields. In addition to commonly known conservation benefits, it is noteworthy that Forester Berkhout included the following arguments in favor of terracing (cited in Pelzer, 1945): (1) "The Treasury finds it easier to collect taxes from a sedentary, stable population than from a people that is constantly moving." (2) "It considerably reduces the land requirements of the native population in the mountains. It is therefore easier to expand private plantation agriculture and gov- ernmental coffee cultivation." Berkhout listed the arguments against terraces on dry fields as the following: (1) "The people are not used to terracing dry land." (2) "Terracing demands more labor." (3) "It reduces productivity in the first year if the land is rather steep and the humus layer thin." (4) "It ties the native more to the soil; so that he can less easily escape the Treasury and the control of native officials. (This is a disadvantage for the individual but an advantage for the government)." The historical record reveals terracing upland farms was promoted by officials in Kerinci in the early 1900s, but with limited success. A Kerinci upland farmer shared with me a letter written in 1916 that gave permission to his father (who was then living in one of the hamlets of Sungai Penuh) to open up a coffee farm in Sungai Ning. Permission would be granted only on the condition that the farmer construct terraces (bertingkat tingkat, literally "to make steps") before planting coffee (Karim, 1987). The permission to cultivate the land would be withdrawn if terracing practices (as well as other stipula- tions) were not followed. He recalled in an interview what he remembered about the ter- races his father built on his upland farm: My father built the terraces because the Dutch told him to. I don't remember the terraces being maintained for a long period. My father spread cow manure to rebuild the soil after building terraces. The major problem with maintaining the terraces, the informant remembered, was that Imperata (or alang alang) grass kept returning. He remembered his father complain- ing about these weeds and the time it took to take care of the terraces, and finally that the walls broke apart. He did not care to maintain them. In 1929 Nooteboom, the chief Dutch colonial officer in charge of Kerinci from 1925 to 1929, advocated terracing in Kerinci's uplands. Nooteboom was concerned mostly about increasing production of coffee, which was mandated by colonial policy and, as a result, became the major hillside farm crop and an important source of revenue for the colonial cache. In his report to his successor, Nooteboom cites reluctance on the part of Kerinci farmers to terrace the slopes against erosion (1929). Bouman, the assistant resi- dent between 1930 and 1936, included exactly the same recommendation made by Van Aken 20 years earlier in his final report. Importantly, he recalled very little progress in convincing local fanners to construct terraces in the 20-year interim period (Bouman, 1936). 438 J. M. Belsky

Additional written and oral evidence suggests that terracing was not widely practiced during the later years of the colonial period in Kerinci. Huitema (1935, p. 152) writes that "ground tillage (annual cropping) as well as terracing do not occur in most of the area." A former mendapot (Dutch name for a local leader) recalled that very few farmers built ter- races during the colonial period, despite strong encouragement from the Dutch. Those that did, he claimed, did not use them after the Dutch left. They did not maintain terraces for two main reasons: First, terraces required too much work to maintain, especially in light of the fact that most people also cultivated one or more upland parcels plus lowland rice paddy; and second, most people did not rate the conservation value of terracing to be higher than their negative impact on productivity. As already discussed, these limitations to terracing resurfaced in contemporary efforts to promote terracing of upland fields.

Discussion and Conclusion The Indonesian state's approach to soil conservation is not the result of a misguided pol- icy based on limited information. Rather, it reflects highly constrained decisions in the context of prevailing institutional factors. Rather than a function of its ignorance, the state has continued to favor bench terraces because they are understood by agriculture and development planners, then as well as now, Indonesian as well as foreign, as a promising means to increase the intensification and diversification of upland agricultural production and soil conservation.7 Over the last century, the Indonesian state has promoted bench terraces as a means to settle and regulate shifting cultivators for both environmental and political-economic rea- sons. Encouraging shifting cultivators to adopt sedentary farming practices was intended to curtail deforestation and invasion of obnoxious grasses (especially Imperata cylindrica) associated with repeated burning of swidden fields. Transforming shifting cultivators into sedentary, upland farmers, or into laborers on state plantations, was also pursued to increase crop production for domestic consumption and/or foreign sale, both necessary for a fledgling economy. Supporting agronomic or "conservation farming" techniques as the basis for agricultural conservation and development policies would have meant embracing a decentralized and unregulated model of development (i.e., in terms of crop choices, marketing outcomes, and surplus extraction). Such a model would not be consistent with the top-down, bureaucratically authoritarian approach to develop- ment pursued in Indonesia, especially since the New Order regime came to power under the leadership of Suharto.8 Major features of the New Order have been the steady concen- tration of political and economic power in the armed forces and growth of the state bu- reaucracy (Hart et al., 1989). The military has been of foremost importance in suppress- ing opposition, and especially for preventing the resurgence of agrarian mobilization that could be used to challenge power at the center. As Hart (1986, p. 29) notes, "Maintaining tight control over the rural sector has remained a constant theme of New Order policy, even though the instruments of control have become more diverse." A third feature of the New Order government has been a concern to maintain food availability and security. Indonesia's history has been one of chronic food production deficits (Mears, 1984). Severe food scarcities (especially in urban areas) have been sug- gested to be the most important cause of the collapse of the Sukarno regime (Husken & White, 1989). Consequently, the need to ensure national food security at relatively stable prices in the cities became an overriding imperative for the survival of the New Order government, along with maintaining political control in outlying rural areas. Despite peri- odic outbursts of protest, the bureaucratic-authoritarian model of government continues Soil Conservation and Poverty 439 in Indonesia and continues to support conventional, production-led approaches to eco- nomic development. The collapse of oil prices in 1981-1982, and more disastrous falls in early 1986, dev- astated the Indonesian economy, which was heavily financed by revenues from oil ex- ports and taxes on oil and natural gas. As a result, economic policies shifted from an em- phasis on import substitution to a reorientation of the economy toward producing tradeable goods. A major focus of Indonesian trade policies since this time has been the diversification of agricultural exports. Production of food crops grew an impressive 4.3% a year between 1978 and 1988, largely as a result of favorable government pricing, re- search and extension efforts, and investment policies toward rice and, to a lesser extent, some other crops (Gonzales et al., 1993). Over the last decade, Indonesia has tried to ex- tend its successful rice intensification strategy to nonrice crops by encouraging develop- ment of marginal regions (such as nonirrigated, upland areas) bypassed by the rice-based program (KEPAS, 1984). A focus on these regions is largely due to the recognition that significant gains in the productivity of lowland agricultural systems will be more difficult to achieve than in the past. The uplands are one such marginal environment targeted for "sustainable develop- ment." In the upland context, sustainable development means increasing production of nonrice (i.e., "secondary" or palawija crops and estate crops). Because of characteristic sloping fields and poor soils, soil conservation is mandatory to protect the stability and sustainability of upland farms. The huge investments necessary to establish engineering conservation methods are thought to encourage cultivation of high-value crops, which can raise upland farmers' low incomes and ensure that they have sufficient funds to main- tain conservation works (Barbier, 1989).9 Terraces are also viewed as critical to protect lowland cropping systems, especially those planted with the new rice technology. Hydro- logical disturbances could turn back the clock on the substantial gains Indonesia has achieved in meeting rice production and rice self-sufficiency targets. In this light, engi- neering models of conservation are viewed as necessary, given their recognized ability to reduce runoff and erosion and to increase agricultural productivity by keeping fertilizers and other modern inputs where they have been applied. Agricultural policies driven by highly centralized production targets have translated into a reliance on monocultures. Production and area-planted targets for most crops are set by Ministry of Agriculture central planners and, by definition, cannot be adapted to local ecological and socioeconomic conditions. Given the diversity of agro-ecological systems that characterize Indonesia's marginal lands, it is questionable how a centralized, commodity-based approach could ever be suitable (Barbier, 1989). Furthermore, the fail- ure to consider farming and cropping systems as the basis for agricultural development strategies means that many traditional agroforestry systems are not being adequately de- veloped (Tarrant et al., 1987). This has certainly been the case in the village of Koto Lebuh Tinggi where the Regreening program attempts to convert cinnamon-based agro- forestry systems to annual crop-based farms with bench terraces. The Indonesian case briefly examined here suggests both confirmation and the need for revisions in our conventional wisdom on the relationship between soil conservation and poverty. For example, poor producers (i.e., those lacking access to irrigated rice fields) from the village of Sungai Ning represent a classic example of the downward, in- tertwining spiral of poverty and ecological degradation. However, high rice security (and high incomes) was not sufficient to ensure maintenance of bench terraces, nor was gov- ernment underwriting of construction costs (what Blaikie and Brookfield term the "lan- desque capital"). This was especially the case in the village of Koto Lebuh Tinggi where 440 J. M. Belsky neither government subsidies nor the food and income security of high-stratum house- holds led to continued use of bench terraces. A widespread preference for traditional, cin- namon-based relay agroforestry systems, and their complementary labor regime with irri- gated rice farming, precluded sustained use of bench terraces throughout that village. Environmental sociological theories require revision. The benefits from a broadly de- fined political economy/political ecology framework have been suggested. These include a more critical stance toward theories that are economically reductionist and do not give sufficient attention to ecological factors, that assume all problems of soil erosion result from capitalist development rather than specific state policies, and that fail to appreciate the variable ways those without access to productive resources and formal power react to and upon social and environmental systems. Lastly, there needs to be more locally spe- cific analyses of the capacity (as well as receptivity) of the state to promote environmen- tal and social reform. Recognition of the ambiguities of, and variations in, state power is critical to ascertaining why in some locales state action hinders, helps, or does nothing to alter the situation of the persistently poor, even in the good name of conservation and sus- tainable development.

Notes 1. Many of the above concerns have been incorporated into Blaikie and Brookfield's (1987) "regional political ecology" approach. The name signals greater attention to regions and their unique agroecologies, as well as location-specific development histories and relationship to the world system. "The phrase 'political ecology' combines the concerns of ecology and a broadly de- fined political economy. Together this encompasses the constantly shifting dialectic between soci- ety and land-based resources, and also within classes and groups within society itself (Blaikie & Brookfield, 1987, p. 17). Their approach also puts more emphasis on political relations as opposed to being economically deterministic—another charge leveled at structural theorists. A more com- prehensive framework for "political ecology" has recently been suggested by Bryant (1992). 2. Terraces are widely used in rocky areas and in places with very high population densities, such as in areas of the Himalayas and Andes. 3. In 1987, 1640 Rp equaled $1.00. 4. In Kerinci, the tradition is for people to cultivate their own rice, even if they can afford to purchase it. Households who can meet their own rice consumption needs through rice production— chiefly in irrigated rice fields—are those with the highest food security and are considered by local standards to be the most "well off." The survey confirmed that there is a positive correlation be- tween irrigated rice-producing households and those with high social status and income, the latter earned largely through cash cropping in upland farms. 5. Terrace construction required about 100-150 worker days per hectare on moderate slope. At the going wage rate of Rp 2000.00/day, the total cost was between Rp 200,000.00 and Rp 300,000.00/ha. When the total annual incomes for low-stratum households in both villages were calculated, the cost of terracing represented from 25 to 45% of farmers' annual income, or calcu- lated in labor time, one-third of their total labor. 6. There were only four low rice security status households in Koto Lebuh Tinggi. Their land use practices were characterized as the following: One practiced continuous annual monocropping, mostly corn and cassava; two practiced relay agroforestry based on cinnamon; and one maintained a (cinnamon) tree farm. 7. In a similar vein, Dove (1983) writes about the "political economy of ignorance" regarding shifting cultivation in Indonesia. He argues that various myths about shifting cultivation have per- sisted despite empirical inaccuracies because they facilitated the extension of external administra- tion and exploitation into the territories of swidden agriculturalists. 8. Following the demise of Sukarno and of the PKI (i.e., Communist party), the rulers of the so-called "New Order" pursued economic strategies that would consolidate the interests of the Soil Conservation and Poverty 441 dominant political and social forces, while at the same time addressing the fiscal crisis. These strategies differed from the older ones in their virulent anti-Communism, pragmatic commitment to rehabilitating the political and economic infrastructure of the country, and favorable disposition to Western aid and foreign assistance (see e.g., Anderson, 1983; Robison, 1986, 1987; Hart, 1986; Husken & White, 1989). 9. Various analysts have discussed the high economic and ecological costs of this approach (Barbier, 1990; Tarrant et al., 1987). In particular, the cost of subsidizing fertilizer, pesticides, and labor in upland areas has imposed a potentially unsustainable financial burden on the government. Subsidizing high rates of petrochemical input-use has also led to considerable external costs in terms of agricultural pollution, waste, and resource depletion.

References Amos, M. J. 1982. Economics of soil conservation. MSc. thesis, National College of Agricultural Engineering, Silsoe, Bedfordshire, UK. Anderson, B. 1983. Old state, new society: Indonesia's New Order in historical perspective. Jour- nal of Asian Studies 42:477-96. Anderson, D., and R. Grove. 1987. Conservation in Africa: People, policies and practice. New York: Cambridge University Press. Arsyad, S., I. Amien, T. Sheng, and W. Moldenhauer. 1992. Conservation policies for sustainable hillslope farming. Ankeny, IA: Soil and Water Conservation Society. Ashby, J. A. 1985. The social ecology of soil erosion in a Colombian farming system. Rural Sociol- ogy 50(3):377-396. Ashby, J., and E. W. Coward, Jr. 1980. Putting agriculture back into the study of farm practice in- novation. Rural Sociology 45:520-523. Baker, R. 1984. Protecting the environment against the poor: the historical roots of the soil erosion orthodoxy in the third world. Ecologist 14(2):53-60. Barbier, E. B. 1989. Cash crops, food crops and sustainability: The case of Indonesia. World Devel- opment 17(6):879-895. Barbier, E. B. 1990. The farm-level economics of soil conservation: The uplands of Java. Land Economics 66(2):199-211. Bernstein, H. 1981. Concepts for the analysis of contemporary peasantries. In The political econ- omy of rural development, ed. R. Galli, pp. 3-24. Albany, NY: State University of New York Press. Blaikie, P. 1985. The political economy of soil erosion in developing countries. London: Longman Press. Blaikie, P., and H. Brookfield. 1987. Land degradation and society. London: Methuen. Bouman, S. 1936. Nota betreffende het district Korintji, Algemeen Rijks Archief, Den Haag. Kolo- nial Archief no. 261. Brundtland Commission. 1987. Food 2000: Global policies for sustainable agriculture. London: Zed. Bryant, R. L. 1991. Putting politics first: The political ecology of sustainable development. Global Ecology and Biogeography 1:164-166. Bryant, R. L. 1992. Political ecology: An emerging research agenda in Third World studies. Politi- cal Geography 11(1);12-36. Buttel, F. H., and W. Sunderlin. 1988. Integrating political economy and political ecology: An as- sessment of theories of agricultural and extractive industry development in Latin America. Paper prepared for presentation at the 46th International Conference of Americanists, Amster- dam, 4-8 July. Carson, B. 1989. Soil conservation strategies for upland areas of Indonesia. Occasional Papers of the East-West Environment and Policy Institute, Paper No. 9. Chambers, R. 1987. Sustainable livelihoods, environment, and development: Putting poor rural people first. Institute of Development Studies, University of Sussex, Brighton, UK, Discussion Paper 240. 442 J. M. Belsky

Chapman, E. C. 1975. Shifting agriculture in the tropical forest areas of Southeast Asia. In The use of ecological guidelines for development in tropical forest areas of South East Asia, Session II, Paper 6. Papers and proceedings of the regional meeting held at Bandung, Indonesia. Inter- national Union for Conservation of Native and Natural Resources, Switzerland. Collins, J. 1987. Labor scarcity and ecological change. In Lands at risks, eds. P. Little and M. M. Horowitz, pp. 19-37. Boulder, CO: Westview. Collins, J. 1988. Unseasonal migrations. Princeton, NJ: Princeton University Press. Departemen Kehutanan. 1985. Pedoman Penyusunan Rancangan Unit Percontohan Usaha Pelestar- ian Sumberdaya Alam (UP-UPSA). Direktorat Jenderal Reboisasi dan Rehabilitasi Lahan, , Indonesia. Dove, M. R. 1983. Theories of swidden agriculture, and the political economy of ignorance. Agro- forestry Systems 1:85-99. English, B. C , J. A. Maetzold, B. R. Holding, and E. O. Heady, eds. 1984. Future agricultural technology and resource conservation. Ames, IA: Iowa State University Press. Faber, D. 1993. Environment under fire. New York: Monthly Review Press. Gonzales, L. A., F. Kasryno, N. D. Perez, and M. W. Rosegrant. 1993. Economic incentives and comparative advantage in Indonesian food crop production. Research Report 93. Washington, DC: International Food Policy Research Institute. Hart, G. 1986. Power, labor and livelihood. Berkeley, CA: University of California Press. Hart, G., A. Turton, and B. White, eds. 1989. Agrarian transformations: Local processes and the state in Southeast Asia. Berkeley, CA: Berkeley University Press. Hecht, S., and A. Cockburn. 1989. The fate of the forest: Developers, destroyers and defenders of the Amazon. New York: Verso Press. Hudson, N. 1981. Soil conservation. Ithaca, NY: Cornell University Press. Hudson, N. 1988. Tilting at windmills or fighting real battles. In Conservation farming on steep lands, eds. W. C. Moldenhauer and N. W. Hudson, pp. 3-8. Ankeny, IA: Soil and Water Con- servation Society. Huitema, W. K. 1935. De Bevolkingskoffiecultuur op Sumatra: Met een Inleiding tot hare geschiedenis op Java en Sumatra. Ph.D dissertation, Wageningen Agricultural University, The Netherlands. Husken, F., and B. White. 1989. Java: Social differentiation, food production, and agrarian control. In Agrarian transformations: Local processes and the state in Southeast Asia, eds. G. Hart, A. Turton, and B. White, pp. 235-265. Berkeley, CA: Berkeley University Press. Karim, A. Interview by Jill M. Belsky. Dusun Baru, Sungai Penuh, Indonesia, 9 Nov. 1987. KEPAS (Kelompok Penelitan Agro-Ekosistem). 1984. The sustainability of agricultural intensifi- cation in Indonesia: A report of two workshops of the research group on agro-ecosystems. Agency for Agricultural Research and Development, Ministry of Agriculture. Jakarta, In- donesia. Leonard, H. J., ed. 1989. Environment and the poor: Development strategies for a common agenda. U.S.-Third World Policy Perspectives, No. 11, Overseas Development Council. New Brunswick, NJ: Transaction Pub. McCauley, D. S. 1985. Upland cultivation systems in densely populated watersheds of the humid tropics—Opportunities and constraints relating to soil conservation: A case from Java. Unpub- lished paper, East-West Center, Honolulu, HI. Mears, L. 1984. Rice and self-sufficiency in Indonesia. Bulletin of Indonesian Economic Studies XX(2):122-138. Moldenhauer, W. C , and N. W. Hudson, eds. 1988. Conservation farming on steep lands. Ankeny, IA: Soil and Water Conservation Society. Nooteboom, S. G. 1929. Nota betraffende de onderafdeeling Korintji. Algemeen Rijks Archief, Den Haag, Koloniaal Archief M.v.O, no. 279, p. 14. Ortner, S. B. 1984. Theory in anthropology since the sixties. Society for Comparative Study of Soci- ety and History 26(1):126-166. Soil Conservation and Poverty 443

Ostberg, W. 1987. Ramblings on soil conservation: An essay from Kenya. Stockholm: Swedish In- ternational Development Authority. Peluso, N. L. 1992. Rich forests, poor people: Resource control and resistance in Java. Berkeley, CA: University of California Press. Peluso, N. L. 1993. Coercing conservation?: The politics of state resource control. Global Environ- mental Change. June: 199-217. Pelzer, K. J. 1945. Pioneer settlement in the Asiatic tropics, pp. 67-68. New York: American Geo- graphical Society. Poffenberger, M., ed. 1990. Keepers of the forest. West Hartford, CT: Kumarian Press. Ramsey, D. M., and K. F. Wiersum. 1974. Problems of watershed management and development in the Upper Solo River Basin. Conference on ecologic guidelines for forest, land and water re- sources development in Indonesia, Bandung, Indonesia. Repetto, R. 1985. Soil loss and population pressure on Java. Ambio 15:14-18. Redclift, M. 1984. Development and the environmental crisis: Red or green alternatives. London: Methuen. Redclift, M. 1987. Sustainable development: Exploring the contradictions. New York: Methuen. Robison, R. 1986. Indonesia: The rise of capital. Sydney: Allen & Unwin. Robison, R. 1987. After the gold rush: The politics of economic restructuring in Indonesia in the 1980s. In Southeast Asia in the 1980s: The politics of economic crisis, eds. R. Robison, K. Hewison, and R. Higgott, pp. 16-51. Boston: Allen & Unwin. Scholz, U. 1983. The natural regions of Sumatra and their agricultural production pattern: A re- gional analysis. Central Research Institute for Food Crops, Padang, Indonesia. Sheng, T. C. 1981. The need for soil conservation structures for steep cultivated slopes in the humid tropics. In Tropical agricultural hydrology, eds. R. Lal and E. W. Russell, pp. 357-372. New York: Wiley. Siebert, S. F. 1990. Soil erosion and conservation farming in Kerinci, Sumatra. PhD dissertation, Cornell University, Ithaca, NY. Siebert, S. F., and J. M. Belsky. 1990. Bench terracing in the Kerinci uplands of Sumatra, Indone- sia. Soil and Water Conservation 45(5):5559-5562. Stocking, M. A. 1988. Socioeconomics of soil conservation in developing countries. Journal of Soil and Water Conservation 43:381-385. Tarrant, J., E. Barbier, R. J. Greenberg, M. L. Higgins, S. F. Lintner, C. Mackie, L. Murphy, and H. Van Veldhuizen. 1987. Natural resources and environmental management in Indonesia: An overview. United States Agency for International Development, Jakarta, Indonesia. Thrupp, L.-A. 1990. Politics of the sustainable development fad: Do the prevailing "panda-hug- gers" think about poor people? Unpublished manuscript, Energy and Resources Group, Uni- versity of California, Berkeley, Berkeley, CA. West, P. C, and S. R. Brechin, eds. 1991. Resident peoples and national parks. Tucson, AZ: Uni- versity of Arizona Press. Watanabe, M. 1979. Report on the erosion problems in Kerinci district. Directorate of Rivers, Ker- inci, Indonesia. Watts, M. 1983. Silent violence. Berkeley, CA: University of California Press. .Wiersum, K F. 1976. Protective forest and erosion control policy in Java: A comprehensive review of ideas and practices. Upper Solo Watershed Management and Upland Development Project, Jakarta, Indonesia. Young, A. 1989. Agroforestry for soil conservation. Oxon, UK: CAB International.