1 9674 1998 Land-Use Zoning ECONOMIC on Tropical Frontiers DEVELOPMENT Emerging Lessons from the Brazilian Amazon I NSTITUTE Dennis J. Mahar and Ce'cile E.H. Ducrot of the World Bank EDI Case Studies LAND-USE ZONING ON TROPICAL FRONTIERS: EMERGING LESSONS FROM THE BRAZILIAN AMAZON Dennis J. Mahar and Cecile E. H. Ducrot The World Bank Washington, D.C. Copyright c 1998 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. The World Bank enjoys copyright under protocol 2 ofthe Universal Copyright Convention. This material may nonetheless be copied for research, educational, or scholarly purposes only in the member countries of The World Bank. Material in this series is subject to revision. The findings, interpretations, and conclusions expressed in this document are entirely those of the author(s) and should not be attributed in any man ner to the World Bank, to its affiliated organizations, or the members of its Board of Executive Directors or the countries they represent. Stock Number: 37139 Contents Acknowledgments ....................................................v The Experience of POLONOROESTE ....................................................1 The Origins of Zoning in Rond6nia .................. ..................................3 A Brief Digression on Rond6nia State Politics .................................................... 7 Stakeholder Views on Zoning ..............8.....................................8 The PLANAFLORO Midterm Review ................................................... 12 The "Second Approximation" of Land-Use Zoning ............................................ 15 Zoning and Deforestation ................................................... 18 Conclusions and Lessons Learned ................... ................................ 21 References ................................................... 24 Boxes 1 Prescriptive versus Indicative Zoning .................................... ................4 2 Socioeconomic Trends and Deforestation in Rond6nia ............................... 19 Tables 1 Area and Prescribed Land Use under Current Zoning Plan of Rond6nia ........5 2 Stakeholders and Their Incentives to Support or Oppose Zoning .................9 3 Total Deforestation in Rond6nia, 1978, 1988-92, and 1994-96 ................... 18 Figures 1 Land-Use Zoning ...................................................6 2 Deforestation in 1989 and 1996 ................................................... 20 3 Rio Branco Indigenous Area ................................................... 22 -iii- Acknowledgments The authors wish to thank Robert Schneider, John Redwood, and Kenneth Chomitz for their insightful comments on this study. We also would like to thank David Gray for preparing the map and satellite images reproduced here. Last but not least, we thank the PLANAFLORO team and numerous stakeholders in Rond6nia for sharing their views. - v- Land-Use Zoning on Tropical Frontiers. Emerging Lessons from the Brazilian Amazon Dennis J. Mahar and Cecile E. H. Ducrot For more than 30 years, successive Brazilian governments have engaged in programs aimed at opening up the Amazon region for settlement and development. Policies implemented by military governments from the mid-1960s through the 1970s were clearly developmentalist in nature with strong geopolitical overtones. Such policies included the construction of thousands of kilometers of roads linking the previously inaccessible Amazon region with traditional population centers in the northeast and south. Attempts were made to settle migrants in planned towns along these new roads. The government also offered generous tax and credit incentives to entrepre- neurs willing to invest in the region. Although these policies served to attract settlers and investment capital to the region, resulting land-use patterns were often chaotic and inconsistent with the concept of sustainable development (Mahar 1988). Starting in the late 1980s, the Brazilian government, along with certain donors and NGOs, proposed prescriptive land-use zoning (LZ) to bring order and rationality to land use in the Amazon region. LZ is deceptively simple in its conception. Land areas are first categorized in terms of their best uses from the standpoint of sustainable development through aerial and satellite maps, soil samples, biodiversity inventories and other technical information. Using this technical information, some public au- thority (typically a state or local government) then specifies the land uses that are permitted in given areas. LZ is now being introduced in many Amazonian states. However, it was first put into place on a large scale in Rond6nia with assistance from the World Bank. As far as we are aware, Rond6nia with the neighboring state of Mato Grosso is the site of one of the largest experiments in tropical frontier LZ in the world. A decade has now passed since LZ was legally instituted by the state of Rond6nia through Decree Law 3782 of June 14, 1988. While the process of implementing LZ in Rond6nia continues, this note seeks to draw some interim lessons from the expe- rience to date. It is hoped that these lessons will contribute to improving the design and implementation of LZ not only in Brazil but in other countries of the world with similar tropical frontiers. The Experience of POLONOROESTE During the 1970s tens of thousands of migrants, mainly from the southern and northeastern regions of Brazil, converged on the state of Rond6nia in search of Dennis J. Mahar is a manager and Cecile E. H. Ducrot is a young professional with the Environment and Natural Resources Division, Economic Development Institute, World Bank. - 1 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon plots of fertile farmland that were purportedly free for the asking. The federal land settlement agency National Institute for Colonization and Land Reform (INCRA) established seven directed settlement projects in Rondonia in an ef- fort to accommodate these prospective settlers. However, the size and velocity of this new migratory flow quickly overwhelmed INCRA. After considerable analysis and debate, the World Bank in 1981 agreed to the government's re- quest for assistance to help bring order to this chaotic settlement process. The government's program was called the Northwest Brazil Integrated Develop- ment Program, or POLONOROESTE. The World Bank's decision to support POLONOROESTE was not an easy one. The Bank had to weigh the consequences of getting involved in an ongoing process over which the government had little control, against the consequences of standing by and letting the process run its course. In the latter case, it is virtually certain that the government would have allocated resources to pave the main overland artery BR-364, but it is doubtful under prevailing fiscal conditions whether adequate re- sources would have been made available for the health, nature conservation, indig- enous protection, agricultural research and extension, and other components designed to mitigate the negative effects of the frontier processes already underway. The World Bank's Board of Executive Directors approved five complementary loans, totaling approximately US$500 million in support of POLONOROESTE dur- ing 1981 and 1982. The government's share of financing totaled about US$1 billion. Subsequent implementation of POLONOROESTE was seriously unbalanced. While paving BR-364 was executed in record time, the implementation of other mitigat- ing actions, particularly the environmental and indigenous protection components, lagged. It was therefore extremely difficult for the government to contain the de- forestation and social conflict that accompanied the accelerated population growth of rural Rond6nia during the 1980s. High rates of deforestation and social conflict in Rond6nia were prevalent before the advent of POLONOROESTE. However, the Bank's association with the program became a cause celebre among environmental groups as reports of massive deforestation in Rond6nia, and elsewhere in Amazonia, began ap- pearing in the international press. Partially as a result of pressures from these groups, the World Bank decided to suspend disbursements of loans support- ing POLONOROESTE for five months during 1985. Two years later, in the wake of unrelenting external criticism of the project, the then-president of the World Bank, Barber Conable, referred to POLONOROESTE as an "environ- mentally sound effort that went wrong" in a speech to the World Resources Institute in Washington, D.C. Several factors have been identified in contemporary and ex-post reviews of POLONOROESTE that help explain the poor performance. Included among them are: (1) government economic stabilization programs that greatly reduced the avail- ability of counterpart funds and agricultural credit; (2) insufficient knowledge of Rond6nia's natural resource base and sustainable development potential; (3) highly centralized decisionmaking at the federal level and insufficient project "ownership" at the local level; and (4) weak irnplementing institutions, particularly in relation to - 2 - Dennis J. Mahar and Cecile E. H. Ducrot the size and complexity of the project. (For a more comprehensive discussion of the POLONOROESTE experience see Redwood III 1993, pp. 55-60.) The attempts to implement components aimed at protecting the fragile environment and a vulner- able indigenous population, while pioneering at the time, were especially problem- atic. As the midterm review of POLONOROESTE carried out in 1984 put it, the program's mitigating measures were "stunted by weak program coordination, insti- tutional inefficiencies, and undisguised lack of political support for environmental and Amerindian protection" (italics added). The Origins of Zoning in Rond6nia As POLONOROESTE neared completion in 1986, both the federal and state gov- ernments began preparatory work on a proposed follow-up project with assistance from the United Nations Food and Agriculture Organization (FAO). Efforts focused on designing a prescriptive agroecological zoning scheme for the state (box 1). This tool, which had not been used previously in the Brazilian Amazon on a large scale, was intended to bring order to the chaotic and environmentally damaging land settlement process that characterized the previous decade. In principle, LZ sought to categorize areas according to their potential use, distinguishing lands with po- tential for sustainable development from those best suited for protection on envi- ronmental or social grounds. By September 1987, an initial LZ plan, based on rough maps (at a scale of 1:1 million) and soil sampling (at a scale of 1:500,000) prepared during the 1970s, was ready. As mentioned earlier, this LZ plan was legally sanc- tioned through a state decree in 1988; it was subsequently passed as a state law in December 1991 (Complementary Law 52) and included in the state constitution. The original ("first approximation") LZ plan in Rond6nia, which still remains in effect, divided the state into six macro zones. Seven subzones, recommending certain crops, were also specified under Zone 1. As table 1 indicates, LZ placed 55 percent of the state (Zones 4-6) off-limits to traditional forms of economic activity such as agri- culture and ranching. The latter was to be permitted only in areas with better soils on either side of BR-364 (see figure 1). This subregion of the state had been the focus of economic activities supported under POLONOROESTE and was the home of most of Rond6nia's rural population. It was also the part of the state that had suffered the greatest deforestation during the previous decade as new settlers cleared land for agriculture and ranching. The bulk of the land under Zones 4 through 6 was in the Guapor6 River Valley to the southwest of the main highway (see figure 1). This was a sparsely populated part of the state where deforestation had been limited to small areas near feeder roads and the Guapore River itself. Traditional economic activities in this subregion included various forms of extractivism (for example, rubber- tapping) and small-scale agriculture, logging, and mining. After extensive discussions with the federal and state governments and repre- sentatives of the local nongovernmental organization (NGO) community, the World Bank agreed to finance a large Natural Resource Management Project in Rondonia. (A similar project was also approved for the state of Mato Grosso.) A US$167 mil- lion loan in support of this project, which became better known by its Brazilian - 3 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon Box 1. Prescriptive versus Indicative Zoning Traditionally, prescriptive zoning, where land use over space is regulated by legal rules, has been applied to periurban areas to separate conflicting land uses or to guide land development. For example, prescriptive zoning has been used to keep polluting industries out of residential areas or to protect farmland from urban sprawl. In this case, zoning is the outcome of a process of land-use planning. On the other hand, indicative zoning, where the agroecological, and sometimes the socioeconomic characteristics of land over space are mapped, has been typically used as the start- ing point of a land-use planning process in rural areas where the objectives are agri- cultural production and/or environmental protection (Sombroek and Eger 1997). In such cases, zoning is mostly used as a guide for investment decisions. An example of such zoning are maps of soil suitability, which have been produced for many regions of the world. The case of Rond6nia is atypical in that prescriptive zoning has been applied to a rural area in an attempt to ensure that agricultural production is sustainable and to protect forested areas. Indicative zoning, which would normally have been used as a basis in a lengthy process of land-use planning, was rapidly turned into a pre- scriptive zoning plan, with little or no participation from the affected stakeholders. There was also little consideration for socioeconomic factors such as the identifica- tion of those areas that would rapidly come under pressure for development, such as those along roads. The criteria used for zoning were almost exclusively agroecological in nature. The objective of prescriptive zoning is to correct socially suboptimal land use caused by market failures. Zoning in Rond6nia seeks to correct two types of market failures. First, in allowing agricultural and livestock activities to take place only on those soils that are the most fertile and resilient, zoning attempts to correct a dis- crepancy between what economists call social and private rates of time preference. In other words, what leads individuals to cultivate soils in a way perceived as un- sustainable by society is the fact that individuals, particularly in a poor society, of- ten respond to immediate needs or desires, whereas society may worry more about the capacity of soils to continue producing for future generations. Second, zoning in Rond6nia attempts to correct the lack of markets for carbon sinks, biological diver- sity, and existence values provided by tropical rain forests. The recent literature sug- gests that the monetary value of these services may be high. Estimates of the value of carbon sequestered by the forest alone range from approximately US$1,000 to US$3,000 per hectare (Kengen 1997). acronym of PLANAFLORO, was approved by the World Bank's Board of Execu- tive directors in March 1992. Taking LZ as its underpinning, the project strategy called for directing all public investment and services to areas with potential for sustainable development, all of which had been settled with varying degrees of intensity before the project began. The intention was to promote the intensification of agriculture in already-settled areas, with the expectation that increasing incomes would reduce the incentive to clear more forest. Environmental and indigenous - 4 - Dennis J. Mahar and Ccile E. H. Ducrot Table 1. Area and Prescnbed Land Use under Current Zoning Plan of Rond6nia Zone Area (km') Characteristics Prescribed land use 1 61,950 Small/medium holdings; Large-scale agriculture, (28%) high deforestation; livestock, and agroforestry better soils 2 30,150 Small/medium/large Small-scale agriculture (14%) holdings; medium and agriculture mixed deforestation; medium/ with tree crops poor soils 3 5,890 Sparse population; Riverine activities, (3%) alluvial soils including floodplain agriculture and fishing 4 35,000 Mainly public lands; Extractive activities, (16%) sparse population; low including gathering of deforestation; poor soils fruits, nuts, resins, medicinal plants, etc. 5 24,350 Mainly public lands; Sustainable forest (11%) sparse population; low management deforestation; poor to good soils 6 64,000 Public lands; very sparse Permanent preservation, (29%) population; very low including conservation deforestation; fragile units and indigenous ecosystems reserves Source: State government of Rondonia 1988 data. components were also included in the project to conserve biodiversity in the rel- evant zones through the establishment and protection of conservation units, indig- enous reserves, extractive reserves, and public forests. Given the low resolution of the base maps and the haste in conducting the studies for the zoning, the project recognized explicitly that the original LZ was simply too crude to guide land use in Rond6nia in the longer term. Resources were therefore earmarked for a "second approximation" zoning exercise on a scale of 1:250,000 and, in some critical locali- ties, on a scale of 1:50,000. From the beginning, one of the greatest challenges for PLANAFLORO in general, and LZ in particular, has been to ensure effective coordination between the federal and state governments on the one hand and between the state government and the affected communities on the other. In the first case it was extremely important to have close coordination between future investments of the federal government in Rond6nia and the existing state zoning law. It was essential, for example, that zones dedared off-limits for traditional economic activities such as agriculture and ranching should not become the sites of federally financed infrastructure (particularly roads), directed settlement projects, and land titling. Because the federal government was the bor- rower under the PLANAFLORO loan, in theory this should not have been a problem. However, in reality this coordination proved more difficult than expected. The intergovernmental and intragovernmental coordination problem was perhaps most evident in the case of new INCRA settlements. INCRA had historically been a - 5 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon Figure 1. Land-Use Zoning 02E O SO Z r * v * s~ ; S E 0 Oo o < N2,, E C C 10~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~* Source: State government of Rondornia (1988). - 6- Dennis J. Mahar and Cecile E. H. Ducrot powerful force in determining land use in Rond6nia, and it was therefore important for Rond6nia and INCRA to enter into an agreement whereby the latter agreed to respect the State Zoning law.' Such an agreement was jointly signed in 1992 as a con- dition of effectiveness of the PLANAFLORO loan. The local office of INCRA, how- ever, could not always be counted upon to submit to the directives of its main office in Brasilia. Indeed, between 1992 and mid-1995, INCRA attempted to establish seven land settlement projects in Zones 3,4, and 5 in direct conflict with the provisions of the Zoning Law. Moreover, 5 percent of the land titles issued during this period in Rond6nia were in these same zones. While a new agreement was signed between the state of Rond6nia and INCRA in August 1995, there are persistent allegations that INCRA has continued to settle people in Zones 3, 4, and 5 (Millikan 1998). A Brief Digressiori on Rondonia State Politics The politics of zoning in Rond6nia are far more complex than the impression fre- quently given in the press of local and international NGOs pressuring the World Bank to alter its policies. To appreciate fully Rond6nia's experience with land-use zoning up to the present, it is essential to understand the role of state politics dur- ing the design and subsequent implementation of LZ and PLANAFLORO. (Much of this section on LZ is drawn from the excellent political analysis in Keck 1997.) In our view, the issues associated with LZ and PLANAFLORO are more related to political factors within Rond6nia than to technical ones. During the very early stages of LZ (about 1988-89), there was no concerted at- tempt on the part of government planners to forge a broad political coalition in favor of zoning. In the words of Margaret Keck: The planners avoided negotiating with municipal and other local leaders, fearing their opposition. Groups that might have supported the plan's goals (rubber tap- pers and Amerindians, for example) were too weakly organized to be viewed as necessary interlocutors. Although a few individuals associated with NGOs or the university were involved in discussions of the zoning plan, overall there was little effort to rally political support in the state for its substance. It was in this environment that representatives of international NGOs complained to the World Bank that local communities had not participated in the design of the project. It was clear, however, that there were serious differences of opinion among the various NGOs as to which of them actually represented the affected conmmuni- ties. After a series of oftern stormy meetings held in 1990-91, turf issues among NGOs 1. Under Brazilian law, INCRA is responsible for administering land tenure matters in all areas within 150 kilometers of an international boundary. In the case of Rond6nia, this includes a large stretch of land along the Guapore River, which constitutes the boundary between Brazil and Bolivia. Until 1988, INCRA was also responsible for land tenure in areas within 100 kilometers on either side of a federal highway such as BR-364. The fact that INCRA had expropriated considerable land along BR-364 for the purposes of directed settle- ment meant that it still wielded considerable influence in the central part of Rond6nia. -7 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon were largely resolved. Upon reaching this understanding, previous opposition to PLANAFLORO on the part of the NGOs was withdrawn. The World Bank therefore decided to go ahead with the project, conditioning its financing to the participation of NGO representatives in project management. Unlike in most other states in Brazil, political parties have little bearing on the position that politicians in Rond6nia are likely to take on given issues. According to Keck, "local politics in Rond6nia is essentially pork barrel politics." This distin- guishing feature of state politics explains to a great extent why incumbents have tended to support PLANAFLORO, even those who opposed it before assuming elected office. Given the precarious state of Rond6nia's finances, the resources made available through PLANAFLORO provide politicians with practically the only opportunities for distributing patronage in the form of jobs, social services, and new infrastructure in the communities in Zones 1 and 2. It is important, however, not to confuse PLANAFLORO with LZ. In contrast to PLANAFLORO, the position of local politicians with regard to LZ has been highly ambivalent. The reason for this is not difficult to discern. Those with pow- erful economic interests in the state, who are also powerful politically, see few benefits emanating from LZ (see the next section). A similar viewpoint is held by many communities in Zones 4 through 6. State politicians have therefore been obliged to engage in a delicate balancing act. On the one hand, they must support the implementation of LZ in order to maintain the politically essential influx of outside resources through PLANAFLORO. On the other hand, by continuing to support the politically sensitive concept of land-use zoning, state politicians run the risk of antagonizing leading economic sectors and thus seriously jeopardize their chances for reelection. Stakeholder Views on Zoning The World Bank recognized the need for the involvement of local stakeholders with the design and implementation of PLANAFLORO and LZ from the outset. Indeed, this was one of the most important lessons learned from the POLONOROESTE experience. The introduction of LZ made effective stakeholder participation even more important in the case of PLANAFLORO, because it essen- tially divided the state up into areas open for traditional development and those that were not. This was a revolutionary concept in Rond6nia at the time, as the previous two decades of the state's history had been characterized by largely un- constrained frontier-type settlement and development. Many local stakeholders now believe that the existence of LZ helps to protect their interests, despite substantial skepticism at the beginning. Included in the pro- ponents of zoning are environmental activists (both local and international), repre- sentatives of indigenous communities, extractivists such as rubber-tapper groups, and, in a mixed way, some small farmers (see table 2). In the past, the economic and social viability of rubber-tapper and other extractivist communities had been threat- ened by outside intrusions because of uncertainties over land tenure. Land-use zon- ing was thus welcomed as a means to legally establish areas where only extractive -8- Dennis J. Mahar and Cecile E. H. Ducrot Table 2. Stakeholders and Their Incentives to Support or Oppose Zoning Political support for (+), or opposition Stakeholders Stakes in zoning (-) to zoning Ranchers Access to productive land Loggers Access to raw materials for production Urban population Job opportunities growth Small farmers Access to productive land Information on land aptitude + Indigenous people Territory protection + Rubber tappers Source of raw materials for livelihood + Local environmentalists Forest conservation + Rural workers Job opportunities growth Municipal government Votes Revenue base State government Votes Revenue base Federal government Votes from environmentalists (National) + Votes from population in Rond6nia International image + International community Carbon sequestration + Biodiversity conservation + Rights of native peoples + Source: Authors' research activities would be permitted. To assist these communities specifically, PLANAFLORO financed the establishment of more than 20 "extractive reserves." According to S. Schwartzman (cited in Browder 1992, p. 174), the latter are officially defined as "for- est areas inhabited by extractive populations granted long-term usufruct rights to forest resources which they collectively manage." These reserves are benefiting ex- tractive communities by providing a clear physical demarcation of traditional pro- duction areas, secured by long-term leases. As a consequence, extractivist communities have been (at least theoretically) much better able to resist pressures from ranchers and loggers seeking to clear the forest. Rubber-tapper associations, for example, have been some of the staunchest supporters of LZ in Rond6nia. While small farmers do not support the restrictive aspects of prescriptive LZ, they welcome the soil fertility information provided by LZ. This is because recent settlers, constituting most farm- ers in Rond6nia, do not benefit from the deep traditional knowledge about the agroecological environment that is passed from generation to generation in areas long settled. The mapping of soils aptitude provided by LZ can help avoid mistakes that settlers, who typically live at or below subsistence levels, cannot afford. It is important to note that some of the strongest and most influential support for zoning in Rond6nia has come from beyond the state's borders (see table 2). This is because many of the potential benefits from zoning, particularly those related to environmental protection and ensuring the integrity of tribal lands, are seen to accrue to groups outside of the state, including outside of Brazil. Brazilian environ- mental NGOs and human rights groups, for example, have closely monitored the implementation of PLANAFLORO, including its zoning component. In many cases these organizations have joined forces with similar organizations in Europe and -9- Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon North America. Bilateral donors and international financial institutions have also been important stakeholders in Rond6nia. Among other things, they share a com- mon concern for the global externalities (for example, climate change and loss of biological diversity) associated with the clearing and burning of the state's forests. International donors have also expressed concern over the impact of Rond6nia's development on vulnerable native peoples. Of course, not all segments of Rond6nia's society welcomed the introduction of land-use zoning (see table 2). The principal opponents of LZ have from the begin- ning been politically powerful logging and ranching interests. By and large, these groups view zoning as a camisa de forca (straight jacket) that prevents them from expanding into areas of the state now zoned for conservation or for activities with a low impact on the environment (see table 2). The residents of some rural commu- nities also opposed LZ on the grounds that the "first approximation" failed to rec- ognize the rights of small farmers who had established themselves in Zones 4 and 5 before the passage of the State Zoning Law. Many small farmers also resent the fact that LZ limits the amount of land available for settlements attributed to the LZ. As one local elite put it in a 1995 interview: The sector that has the most influence in local society is the logging industry, the sector most resistant to zoning. There are small communities located in areas of environmental protection that don't understand zoning well and are against it. The livestock sector which needs lots of land, resists zoning, particularly with regard to the demarcation of indigenous reserves (Moreira 1995, p. 62). Some local politicians also resist zoning. Municipal governments in Rond6nia with oversight on land where economic activities are restricted or prohibited have little incentive to enforce the zoning, because restrictions on settling people and on pro- duction mean forgone income for the local government and lost votes from their electorate (see table 2). In principle, the state government of Rond6nia also has little incentive to support zoning for essentially the same reasons (see table 2). The lack of broad political support for environmental protection and protection of indigenous communities has contributed to increasing pressures on the various conservation units. There is a widely held view, both in Rondonia and elsewhere in Brazil, that far too much land is held in "unproductive" indigenous reserves. Cur- rently, over 18 percent of the total area of the state is held in reserves for an indig- enous population of approximately 5,000. In this regard, one local politician is quoted as saying: "We have, today, on the strongest and richest soils of Ariquemes [located in central Rond6nia], an indigenous reserve of 1.8 million hectares to house 10 In- dians. We are prevented from carrying out economic development in this area' (Moreira 1995, p. 61). The state government itself challenged the legal status of five indigenous areas in April 1996. It did so on the basis of federal legislation (Decree 1775 of January 8,1996), which establishes administrative procedures through which nonindigenous parties can challenge indigenous peoples' claims over land. In one such case (the Mequens Indigenous Reserve), the government called for a signifi- cant reduction in the size of a reserve that had already been demarcated using - 1 0 - Dennis J. Mahar and Cecile E. H. Ducrot funds from PLANAFLORO. In the face of strong local and international pressures, the government eventually withdrew all such challenges. In theory, PLANAFLORO was designed to lessen opposition to zoning. As ex- plained earlier PLANAFLORO called for directing public investments toward sup- porting intensive production in Zones 1, 2, and 3 and environmental and indigenous protection in the other zones. Large sums were provided by the World Bank loan in support of PLANAFLORO to increase such investments. The flow of funds from the Bank loan has provided a strong financial incentive to the state to endorse zon- ing. As explained earlier, PLANAFLORO has represented a significant source of income to the state government over the past few years and has been instrumental in rallying support from the state government.2 Nevertheless, for a number of rea- sons discussed later in this paper, PLANAFLORO funds were less successful at rallying support from other stakeholder groups. To conclude on stakeholder views, local support for correcting the two types of market failures that LZ in Rondonia seeks to correct (see box 1) is mixed and weak in comparison to the incentives to counter the zoning. The local population of set- tlers tends to favor the development of information on spatial variations in pro- ductive potential. Nevertheless, it is not in the interests of cash-strapped, short-term governments to limit settlements only to the areas with the best potential. In addi- tion, while technical recommendations are useful, they do not always correspond to the individual's best interest. For example, a soil may be poor but so close to the market that it can be profitably mined for a few years before being abandoned (Schneider 1993). Therefore legal restrictions based on technical recommendations are not always welcome. As to the second type of market failure, the lack of market for environmental values provided by forested land, local political voice in sup- port of these values is weak in relation to profits accrued to the stronger local politi- cal constituencies from clearing forests. Most of the support for conserving forests originates from outside of Rond6nia. To help ensure stakeholder participation in the management of PLANAFLORO, local NGOs were given a 50 percent representation on the State Deliberative Coun- cil, a body responsible for overall guidance and policy direction for the project. NGOs were also given control over half the seats on the seven project planning councils (conselhos normativos) responsible for discussing concrete proposals with the agreed project objectives. This was one of the more innovative aspects of PLANAFLORO's original design. However, in what turned out to be a major mis- take in retrospect, local business associations, including those representing power- ful logging and ranching interests, were not given any formal role in project management. This initial oversight has been partially overcome as a result of the project midterm review process (discussed later) and subsequent project restruc- turing. However, opposition to zoning continues in Rondonia. 2. Not including the loan, receipts to the state government totaled almost US$2 billion from January 1, 1993, to December 31,1996, the period initially planned for the loan. Had the total amount of the loan been disbursed as planned it would have added approximately 8.5 percent to the receipts of the state government. -11 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon The PLANAFLORO Midterm Review The implementation of PLANAFLORO got off to a slow start. With less than six months left in the project's official five-year implementation period-the closing date for the World Bank loan was December 31, 1996-expenditures were only slightly more than half of what was expected at this advanced stage of project imple- mentation. As in the case of POLONOROESTE, the most serious delays encoun- tered were in executing critical environmental and indigenous assistance components. Furthermore, difficulties were also encountered in effectively incor- porating stakeholders' views into PLANAFLORO's decisionmaking structure; many NGOs complained that their views had not been fully incorporated. The difficulties experienced in Rondonia led in July 1995 to a request from the NGO Forum of Rond6nia that the World Bank's Independent Inspection Panel in- vestigate PLANAFLORO. Following a fact-finding mission to Brazil, the panel pre- sented its conclusions to the Bank's Board of Executive Directors in September 1995. The Bank's response to the inspection request was to acknowledge the problems with project implementation directly and to propose a time-bound action plan for dealing with them during 1996. Some of the key elements of this action plan in- cluded: intensifying institutional strengthening at the state level, concentrating project activities in 15 priority municfpios, giving highest priority to initiating work on the "second approximation" zoning exercise, ensuring the effective participa- tion of indigenous communities in PLANAFLORO affairs, establishing a govern- ment-financed health care service for them, and more closely monitoring federal and state funding and expenditures to ensure that environmental and agroforestry activities would not continue to be outpaced by infrastructure development. Based on this plan, the board subsequently concluded that a full-scale inspection of PLANAFLORO was not warranted. Through the first half of 1996, the action plan was closely monitored and some progress was achieved in project implementation. On the other hand, challenges to LZ intensified during this period, indicating that political support for zoning was still weak. It was in this environment that a comprehensive midterm review of PLANAFLORO's performance to date was planned. This review was to determine whether the objectives and design of the project remained valid. If the review con- cluded that the project objectives and design were no longer relevant, two options were open to the government: redesign the project and request an extension of the official closing date, or let the ongoing project activities wind down in preparation for a formal closing at the end of 1996. In preparation for the midterm review, the state government employed a multidisciplinary team of independent Brazilian consultants in April 1996. A two- day midterm review meeting to consider the consultants' report was scheduled for June 19-20, 1996. Representatives of all major stakeholders in the state were in- vited to the meeting. International NGOs with strong Rond6nia connections, namely Friends of the Earth and OXFAM, were invited to serve as observers. Before the meeting a preliminary version of the independent consultants' report was distrib- uted to many of the groups invited. The NGOs discussed the consultants' report at - 12 - Dennis J. Mahar and Cecile E. H. Ducrot two meetings convened in the interior of the state. Using feedback from the com- munities, the NGOs prepared their own document. In addition to the NGOs, the main business association in Rond6nia, the Federation of Industries of Rond6nia (FIERO), prepared its own report to present at the meeting. The contents of the various reports are revealing with regard to the political economy of PLANAFLORO and land-use zoning and therefore merit discussion in some detail. The independent consultants' report, while noting the frustrations of many of the stakeholders, concluded that there was general support for the overall objec- tives of PLANAFLORO within Rond6nia. However, the report noted a number of problems associated with the implementation of LZ. Included among these were weak technical capacity at the state level, limited understanding of LZ at the com- munity level, existence of political pressures against the maintenance of protected areas, and failure of the "first approximation" to take into account preexisting land tenures. The report concluded that the greatest risks to the successful outcome of the "second approximation" were political pressures to redefine zones and a lack of vision on the part of decisionmakers regarding the appropriate political and institu- tional arrangements necessary for the effective implementation of LZ in the state. In order to overcome the problems encountered by LZ, the consultants' report recommended, among other things, the establishment of a community-based con- sultative body to serve as a conduit between the communities affected by zoning and the State Zoning Commission, the concentration of future zoning efforts in ar- eas of the state characterized by land conflicts, and the provision of incentives to logging and agricultural interests in support of zoning and sustainable develop- ment.3 In this regard, the report calls for a modernization of the local logging indus- try coupled with reforestation programs and more effective provision of land tenure services and agricultural credit in Zones 1 and 2. The report also recommends the provision of basic infrastructure (for example, sanitation and slum upgrading) to urban areas within Zones 1 and 2 and in Porto Velho using funds from PLANAFLORO. This recommendation responds not only to the real need for better infrastructure in the towns and cities of Rond6nia, but also to the political demands of the urban population who generally see few benefits accruing to them from land- use zoning and from PLANAFLORO-financed investments in rural areas of the state. The NGO document briefly discusses the problems experienced by PLANAFLORO and offers recommendations for necessary modification of the project. It starts out with a series of scathing criticisms of PLANAFLORO (Forum das ONGs 1996). The NGOs first refer to the state's lack of political commitment to several key features of PLANAFLORO. The implementation of LZ, the establish- ment of conservation units, and the enforcement of environmental legislation are 3. This commission was legally established in 1994 through Decree 6473 of June 29. How- ever, it was only in mid-1996 that its intemal regulations were detailed. The commission's president is the Rond6nia State Secretary of Planning and its membership includes represen- tatives of all relevant public implementing agencies, the state assembly, and main beneficiary groups and private sector associations. Its main function is to supervise the implementation of LZ in the state and to provide a forum for discussing and resolving conflicts. So far the Zoning Commission has not played an important role. - 13 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon specifically cited. The NGO document further alleges that powerful political and economic interests in the state have undermined the implementation of these mea- sures through pressures on decisionmakers. Other reasons given for shortfalls in the performance of PLANAFLORO include severe fiscal problems at the state level, a lack of effective community participation, a low level of resources actually reach- ing the communities, and the absence of a public information program. The NGOs' proposal to improve PLANAFLORO recommends the reorganiza- tion of the project around two major components, one centralized and one decen- tralized. The first would cover policies and actions related to land tenure and environmental management (including indigenous affairs) and the second increased support of community initiatives. They also requested redefinition and substantial reduction of the role of the local United Nations Development Programme Techni- cal Cooperation program and proposed reorganization of the state project coordi- nation unit into specific teams (or gerencias) for land tenure and environmental management, support to community initiatives, and administration. Finally, they proposed increased participation of NGOs and other representatives of the private sector in overall project coordination through the establishment of policy-oriented sectoral working groups (camaras setoriais) for land tenure, environmental monitor- ing, inspection and control, indigenous peoples affairs, agriculture, and industry, among other areas. The report presented by FIERO at the midterm review, while giving lip service to the concept of sustainable development, called for major expansion of PLANAFLORO (to a level of between US$600 and US$700 million) to give much greater emphasis to the state's developmental needs in industry, energy, and trans- port. The FIERO report, in fact, concludes that neither LZ nor protection of forests would be likely to survive reductions in external development capital. FIERO tends to be dominated by local logging interests who have long opposed PLANAFLORO, particularly the project's emphasis on environmental protection and LZ (FERO 1996). Two aspects of the NGO proposal deserve special mention. First, it is highly significant that the NGOs called for the formal participation of broader private sector interests in project management through representation on the proposed sectoral working groups. Although clearly far apart on their respective visions for the future of Rond6nia, the stated desire of the NGOs to dialogue with philo- sophical adversaries in the business community nonetheless represented a salu- tary first step in forging a broad political consensus on LZ and on appropriate development models for the state. A second significant feature of the NGO pro- posal is that it was largely consistent with the background report produced by the team of independent consultants. In particular, both proposals called for a simpli- fication of PLANAFLORO in both its content (number of components) and ad- ministration. Both also called for broader and more effective participation of the affected communities, including business associations in project administration. In regard to the latter, both the NGOs and the independent consultants recom- mended that PLANAFLORO become more demand-driven and based to a much greater extent on initiatives developed at the community level, rather than at the state and federal government levels. - 14 - Dennis J. Mahar and Cecile E. H. Ducrot After much negotiation among representatives of the various groups partici- pating in the midterm review, consensus was reached on a reformulation of PLANAFLORO along the lines traced in the consultants' report. Representatives of the state, the NGO Forum, and the main private sector groups ultimately signed a formal agreement stating the principles and timetable for this proposed project reformulation. Work on the reformulation was conducted over the following six months with the full participation of the NGOs and private sector. The resulting new version of PLANAFLORO maintained the initial objectives of the project, but scaled down environmental and land tenure activities, combined environmental activities into a single component, and consolidated all other activities into a de- mand-driven and decentralized Community Initiatives Program (PAIC). Subsequent to the midterm, the government of Rondonia was asked by the World Bank to take specific steps to demonstrate its firm commitment to the project's continuation. More specifically, the state was to improve the financial management of the loan; modernize its enviromnental legislation; reinforce and clarify the role of the State Zoning Commission; resolve certain land tenure issues, notably with respect to the transfer of federally owned land to the state in a number of conserva- tion units; and take concrete actions on environmental law enforcement and indig- enous health. After considerable delay, sufficient progress was made on these points to justify an extension of the World Bank loan until July 31, 1998. The "Second Approximation" of Land-Use Zoning From the beginning, it was widely held that the "first approximation" of LZ, based on maps with a scale of 1:1 million, offered only a rough temporary guide to envi- ronmentally sustainable land use in Rond6nia. While certain groups may have been unhappy with some aspects of the "first approximation," political opposition was muted by the expectation that the boundaries of the six macro zones would be al- tered in the near future. PLANAFLORO recognized the need for greater technical precision in zoning, and US$10 million of the World Bank loan was allocated for this purpose. In contrast to the effort carried out in the late 1980s, the new zoning exer- cise was to involve statewide mapping on a scale of 1:250,000 based on new satellite imagery and extensive soil sampling, as well as better economic and social data. Preparations for work on the "second approximation" began in late 1992 when an agreement was reached between the state government and the World Bank on a timetable for preparing terms of reference for the task, as well as draft prequalification and bidding documents. However, differences of opinion between the state and the Bank over procurement methods and costs substantially delayed the bidding process. Indeed, over two years were to pass between the time that bids were let out and when the final contract was signed with the Brazilian-Dutch consortium Technosolo/DHV. Work under this contract was initiated only in April 1996 and is expected to be completed soon. It will have cost approximately US$20 million, twice the amount originally allocated to the task. The delay in the implementation of a more refined LZ plan has provoked a mounting sense of frustration among important stakeholders in Rond6nia. Many -15 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon development-oriented groups, such as business associations and farming com- munities in Zones 4 and 5, have clear expectations that the "second approxima- tion" will expand the areas of zones where their principal activities were legally permitted. In the words of one local businessman interviewed in 1995, Because these other approximations [of LZ] were not elaborated according to sched- ule, the distortions increased in size and volume, to the point where a large pro- portion of society has come to view zoning as a constraint to the state's development.. .This would not have happened if the other approximations had been elaborated. (Moreira 1995, p. 60). Frustrations over the delay of the "second approximation" have resulted in several attempts to change various aspects of the original zoning law. These pres- sures have fallen mainly on the Governor's Office and the Rond6nia State Assem- bly. Lobbying for changes in the Zoning Law increased during the first half of 1996 when it became apparent that work on the "second approximation" was about to start. It is possible that the main stakeholders increased their pressures at this time in an attempt to improve their positions in future negotiations over zoning result- ing from the "second approximation." In one case, pressures from logging interests resulted in the issuance of State Decree 7341 of February 1, 1996, which permitted logging in Zone 4 without an approved forest management plan on landholdings of up to 500 hectares. In the State Zoning Law, it is made clear that activities in this zone were to be geared to sustainable extraction of forest products, along with fishing and subsistence agri- culture. Forest clearing in Zone 4 was to be limited to 5 hectares per productive unit. Deforestation above this limit was to be permitted only upon submission of a forest management plan consistent with prevailing forestry legislation. Despite these clear legal guidelines, Decree 7341 was justified by the state government on the grounds that illegal logging was already taking place in certain parts of Zone 4 and that recognizing the status quo would generate jobs and increase tax collections (Alto Madeira 1996). Decree 7341 was subsequently challenged in the courts and is currently suspended pending a judicial ruling. In yet another attempt to change the Zoning Law, in May 1996 the Rond6nia State Assembly voted on an amendment seeking to change several areas of the state previously earmarked for extractive activities and forest reserves (Zones 4 and 5) to Zones 1 and 2. Despite the lack of supporting studies specified in the Zoning Law, and in the face of opposition from the State Technical Zoning Com- mission, the proposed amendment passed by a vote of 20 in favor, 1 against, and 1 abstention. The justification given for this amendment was that the original zoning legislation had ignored preexisting land tenure situations, including directed settle- ment projects sponsored by INCRA. The State Zoning Law, it was argued, had therefore placed an unfair burden on the local populations living in the disputed areas who, among other things, were legally prohibited from obtaining agricul- tural credit or regularizing their de facto land occupancy. - 16 - Dennis J. Mahar and Cecile E. H. Ducrot The passage of the amendment altering zoning once again placed the state gov- ernment in a difficult position. On one hand, the amendment clearly violated the Zoning Law by ignoring the stipulation that all changes in zoning would need to be supported by technical studies. For these reasons, the draft amendment was opposed by the state's own Technical Zoning Commission. On the other hand, small- scale farmers who had settled in Rondonia before the passage of the Zoning Law, and who now found their livelihoods severely jeopardized by prohibitions imposed by that law, also had a good case. Moreover, their pleas (as well as those of the logging interests who also supported the amendment) had been strongly supported by state legislators. Finally, the government feared that the amendment would con- stitute noncompliance with the legal obligations assumed by the state under the World Bank's loans for PLANAFLORO. The state governor's only alternative under the circumstances was to veto the proposed amendment to the Zoning Law. However, given increasing social tensions in areas of the state affected by the amendment, the political cost of this veto to the government was extremely high. The state therefore introduced an alternative amend- ment immediately after vetoing the original amendment. This amendment, Comple- mentary Law 152, was signed by the governor on June 24, 1996. While not changing zone categories per se, article 1 of this law essentially eliminated the restrictions on land use contemplated in the State Zoning Law in areas of Zones 4 and 5 in eight municipios. Most of these local governments are located in the Guapore Valley along two penetration roads, BR-421 and BR-429. Lands within recognized conservation units are excluded from this proviso. Due to the haste in drafting this new legislation, however, certain important loopholes remained. The law, for example, failed to specify clearly that only those persons who occupied Zones 4 and 5 before the passage of the State Zoning Law would benefit from the amendment. Thus it was possible that the amendment in its original form could be used to reward persons who had illegally settled in the area after 1991. Fortunately, this and other deficiencies of Complemen- tary Law 152 were corrected through legislation signed by the governor in late 1996. Following these attempts to modify the Zoning Law, the World Bank requested the state government to develop transparent and participative administrative rules and processes to guide any further changes in the zoning legislation before the completion of the "second approximation" LZ plan. Decree 7526, signed on Au- gust 2, 1996, defines the technical, legal, and institutional procedures to be adopted for any adjustments in the Zoning Law, including the "second approximation." The decree states that any changes should be based on studies supervised by the State Zoning Commission and conducted at a greater level of detail than the stud- ies that led to the first approximation. Proposals for changes would have to be elaborated through a participatory process involving all stakeholders. Any such proposal would then be submitted to the governor who would subsequently pro- pose a new complementary law to the state legislative assembly. Despite negative reactions to previous changes to the Zoning Law, the state as- sembly attempted once again to reduce the restrictions set by zoning on May 23, 1997. On that date, the state legislative assembly passed Complementary Law 171, - 17 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon removing restrictions on land use placed on strict conservation areas (Zone 6) in the municipality of Cerejeiras and on a small area called Capitao Silvio in the north- western part of the state. This law was subsequently vetoed by the governor. In December 1997, however, the state legislative assembly made an even stronger chal- lenge to the Zoning Law through the introduction of Complementary Law 203, leg- islation essentially giving carte blanche to INCRA to legitimize any occupation of land wherever it deemed appropriate, regardless of zoning. This legislation essen- tially annuls any restrictions imposed by the state zoning law on legalizing tenure of settled land. The governor immediately vetoed the law for being "flagrantly ille- gal and unconstitutional." Unfortunately, on April 2, 1998, his veto was overturned and the law was passed by the state legislative assembly with a vote of 19 to 0. Zoning and Deforestation The best indicators of the way in which the political battles over zoning described have played out on the ground are the rate and spatial distribution of deforesta- tion. Deforestation is not only an indicator of alterations in the forest, but also shows how agricultural and livestock activities have evolved since both these activities, as they are traditionally conducted in Rond6nia, require clearcutting the forest. To date, about 20 percent of Rond6nia's original forest cover has been cleared, an area equivalent to the total land area of Costa Rica. Total deforestation of the state has increased by approximately 19,000 square kilometers between the introduction of zoning in 1988 and 1996, the most recent year for which deforestation data are available (see table 3 and box 2). Deforestation has been most intense in areas where there is good road access, proximity to important population settlements, or both. This observation applies even where use of land is permitted but restricted to activities that sustain forest cover (Zones 4 and 5). Satellite images of an area close to the city Ariquemes illustrate this phenomenon. In 1989, nearly all of the forested portion of Zone 4 covered by this image was still standing (figure 2). Few clearings or roads could be observed. In 1996, a number of roads plainly have been extended from Zone 1 into Zone 4 and the area has been heavily cleared and settled with almost total Table 3. Total Deforestation in Rond6nia, 1978, 1988-92, and 1994-96 Total area cleared Average annual Year (cumulative km2) deforestation (km2) 1978 4,200 1988 30,000 2,340 1989 31,800 1,430 1990 33,500 1,670 1991 34,600 1,100 1992 36,865 2,265 1994 42,055 2,595 1995 46,152 4,730 1996 48,648 2,432 Source: INPE (n.d.). - 18 - Dennis J. Mahar and C6cile E. H. Ducrot Box 2. Socioeconomic Trends and Deforestation in Rondonia Pressure on Rond6nia's forests is likely to increase in the future, according to a num- ber of socioeconomic trends. First, cattle ranching continues to spread. Cattle ranch- ing grew at the rate of approximately 25 percent per year between 1973 and 1995 (Bartholo 1997). Even during the 1980s, when the growth of cattle ranching slowed in the north and center-west regions of Brazil, rates remained high in Rond6nia. Four new slaughterhouses under construction in addition to the 18 already there demonstrate clear expectations that the beef industry will continue growing. As cattle ranching spreads, small farmers are displaced to clear new frontiers. Second, the trend toward displacement of small farmers will probably be further reinforced by the introduction of soybean cultivation in Rond6nia. In 1995 approxi- mately 10,000 tons were produced on 4,500 hectares in the south of Rond6nia. As of yet, its expansion is still speculative; however, there are signs that soybeans may play an important role in land use in the years to come. Soybean cultivation has done well in the state of Mato Grosso in the agroecological region known as the Cerrado. The fact that a significant percentage of Rond6nia's land is classified as Cerrado means that there may be a large increase in soybean cultivation. Indeed, large sums of public and private money are being invested to finance future produc- tion, marketing, and transport of soybeans (FIERO 1997) including the promotion of the Rio Madeira waterway as part of the federal Brasil em Acao program. Third, although rates of population growth in Rond6nia are no longer compa- rable to what they were in the 1970s, they are still significant. From 1991 to 1996 the total population of the state grew at an average rate of 1.7 percent per year. Unlike earlier years, population growth in Rond6nia is no longer from in-migration (in fact more people migrate out of the state than to the state) but from endogenous popula- tion growth, which is unlikely to decline in the near future (FIERO 1997). disregard to zoning. Not surprisingly, the area pictured in these satellite images areas has been subject to repeated attempts to change the zoning law to reduce legal constraints on the use of land. The incentives to clear land in zones where land use is restricted (that is, Zones 4- 6) can be strong. First, there are profits to be made in many of these areas. Even though the soils make it difficult to farm sustainably, the fact that some of these areas are close to markets can make agricultural and livestock activities more profitable than extrac- tive activities, which typically yield low returns per hectare (Schneider 1992). The ef- fect of market access provided by the existence of roads and the proximity to highly populated areas on the spatial distribution of land use is particularly pronounced in the Amazon, where transportation costs per kilometer are very high, distances are great, and most agricultural products have low value-to-bulk ratios. A second incen- tive to clear forest in zones where land use is restricted is the hope that the existence of deforestation will help convince the state government to convert these areas to zones where land use is unrestricted (that is, Zones 1 and 2). Many municipal governments in these areas are said to support new settlements, because larger populations and - 19 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon Figure 2. Deforestation in 1989 and 1996 1989 0 10 20 Kilometers 1996 Source: State government of Rondonia (1988) on zone boundaries, 1989 and 1996 Landsat scenes. - 20 - Dennis J. Mahar and Cecile E. H. Ducrot more economic activities mean more tax revenue. A third explanation is the quality of the soils. Because these soils are fragile, and illegal settlers cannot receive credit or support, most of the land is rapidly turned into pasture after a brief agricultural cycle. Poor settlers then sell their land (illegally) to land users who can afford cattle and move on to clear new areas (Schneider 1993). Finally, once an area is made eligible for the practice of extractive activities (as is the case in Zone 4) and population enters, it is exceedingly difficult to ensure that only extractive activities take place. Surprisingly, in spite of political opposition and of the pressures on forested land described, some notable success seems to have been achieved in preventing forest conversion in conservation and indigenous areas from Zone 6. One of the important contributions of PLANAFLORO, in addition to providing funding for the "second approximation" of LZ, has been to promote the legal establishment and demarcation of the various forms of conservation units under state jurisdic- tion. Considerable progress has already been made in this regard. As of April 1998, the Rond6nia State Land Institute (ITERON) had demarcated 33 conservation units, totaling over 21,000 square kilometers, or almost 10 percent of the state's land area. These include three state parks, two biological reserves, 21 extractive reserves, three ecological stations, and five state sustainable production forests (Millikan 1998). In addition, the project has financed the demarcation or boundary renewal of 17 in- digenous reserves in various parts of the state. The area covered by indigenous reserves totals approximately 45,000 square kilometers. Protecting these conserva- tion areas from intrusions by outsiders is the responsibility of the state forest mili- tary police and the federal environmental protection agency (IBAMA). A recent satellite image of Rio Branco Indigenous Area and its environs, shown as figure 3, serves to illustrate how well the forest cover can be protected in con- servation and indigenous areas. It may be noted that land clearings verge on the edge of straight line park boundaries but do not exceed them, even in the absence of natural accidents. Even local NGO leaders concede that, in spite of a number of infractions, mainly by loggers, most conservation and indigenous areas are now being reasonably well protected. This success can be explained by three factors. First, local stakeholders indicate that physical demarcation is an important factor contributing to the protection of conservation and indigenous areas because the population living around these areas, and indigenous people living inside, know precisely where the borders are. Second, the high national and international vis- ibility of park protection and, in particular, of indigenous rights issues in Brazil has also provided a strong incentive for government to ensure the physical integ- rity of these areas. Finally, the preservation of forest cover is probably much easier to enforce where all direct land-use activities and deforestation are completely prohibited than where they are permitted in a restricted fashion. Conclusions and Lessons Leamed The experience with land-use zoning in the Brazilian frontier state of Rond6nia has been decidedly mixed. The main conclusion of this paper is that zoning is a valid instrument for guiding land use on tropical frontiers, though it is far from being a panacea. Many of the difficulties of implementing LZ in Rond6nia stem from the - 21 - Land-Use Zoning on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon Figure 3. Rio Branco Indigenous Area 6 0 6 12 Kilometers Source: State government of Rond6nia (1988) on zone boundaries and 1996 Landsat scenes. state's decision to adopt a system of "prescriptive" zoning in which legal rules were established to separate conflicting land uses (for example, traditional agricul- ture as compared to forest protection or low-impact activities) in a spatial sense. These rules were imposed in a laudable attempt to reconcile the continued eco- nomic growth and development with enhanced environmental protection. How- ever, these same rules encountered strong political opposition because they prescribed land uses that often differed greatly from those that would have pre- vailed had economic agents been left to make their own decisions. Farmers, ranchers, and loggers have all, to some extent, reacted negatively to restrictions on land use because many of the environmental and social benefits resulting from leaving the forest intact (for example, carbon sequestration, preser- vation of biodiversity, ensuring the rights of tribal peoples) accrue to stakeholders outside the state and even the country. Virtually the only "compensation" received by local stakeholders for restricting land use has been the salaries and investments in Rond6nia financed by the national government and by international donors like the World Bank. This cornpensation, however, has mainly benefited the state gov- ernment, agricultural communities in Zones 1 and 2, and tribal peoples living in demarcated indigenous areas. In contrast, politically powerful urban interests, as well as logging and ranching interests, believe that they have received little or no direct benefits from zoning. Overall, the compensation has been too small in rela- tion to profits generated by activities that disrespect the zoning. - 22 - Dennis J. Mahar and Cecile E. H. Ducrot The Rondonia experience underscores the fact that LZ is an inherently political process that will not work in the absence of broad public support. In our opinion, land-use planners frequently overemphasize the technical aspects of LZ, as in the case of the "first approximation" zoning exercise. While technical aspects are cer- tainly important, they should not be allowed to overshadow and obscure the poli- tics of zoning. By its very nature, prescriptive LZ tends to polarize societies into "winners" and "losers." Typically, success in implementing a prescriptive zoning plan depends on the strength of the land users' incentives to challenge zoning rela- tive to the capacity and incentive of the government to enforce and maintain zon- ing laws (Schneider 1992). It is therefore important to identify beforehand the various groups that have a stake in zoning and to evaluate how the strength of their incen- tives to oppose various restrictions on land use varies over space (Verissimo and others 1998). Stakeholder analyses conducted at an early stage can help to avoid attempts to implement unrealistic zoning plans and to design policies to lessen incentives to counter the zoning, in particular where those incentives are strongest. It is also imperative to include the full spectrum of stakeholders (including nonlocal) in the planning and implementation of any zoning plan in order to reach agree- ment on common goals and means of attaining them. This participatory process, moreover, should be continuous and not a one-off event. By and large, the state has attempted to control challenges to conservation units through traditional command-and-control enforcement activities. However, the task of ensuring the integrity of conservation units must be supported ultimately by the community.4 In essence, the communities surrounding or within conservation ar- eas must be provided with incentives to undertake only environmentally sustain- able economic activities. The redesigned PLANAFLORO is, in a broad sense, based on this principle in that its new community-based component (PAIC) concentrates most resources in areas close to conservation units: incentives are given to explic- itly environmental protection projects; other projects are only allowed on the con- dition that they do not harm the environment. More specifically, communities are required to contribute only 10 percent of their own funds for environmental pro- tection projects while 20 percent is required for other projects. So far PAIC funds have been requested overwhelmingly for projects that have either productive goals or aim to improve basic infrastructure, health, and education. Hardly any funds have been requested for projects explicitly aiming to protect the environment. This is a strong indication of the low interest in environmental protection in relation to other perceived needs. Land-use zoning in Rond6nia is currently entering a critical juncture. With the imminent termination of World Bank financing in support of PLANAFLORO, it will be far more difficult for the state government to defend the restrictions on land use imposed by LZ. This difficulty is likely to worsen over time judging by the fact that socioeconomic trends indicate pressure on Rond6nia's forests will probably continue to increase. Over the past year, the state government, with financial support from the 4. For information on the international experience with community involvement with protected area management, see Wells and Brandon (1992). - 23 - Land-Use ZoninS on Tropical Frontiers: Emerging Lessons from the Brazilian Amazon donor community, has engaged in an innovative program known as Projeto Humidas, which seeks to forge a political consensus on Rond6nia's future development trajec- tory. This is an important step that may ultimately attract additional external fund- ing to the state. Without further inflows of external resources to help finance the state bureaucracy and basic infrastructure such as roads, powerful ranching and logging interests, which have opposed zoning from the very beginning, can be expected to intensify their efforts to overturn LZ legislation. Nineteen ninety-eight is an election year and many politicians will undoubtedly find it expedient to campaign against any "second approximation" zoning plan that maintains large areas of land under forest cover. Under such circumstances, it might make sense for the federal government to assume a more active role in state zoning exercises. This is already happening in an indirect way. The federal government has recently offered to reduce the proportion of each property that must be kept forested in those states that adopt a federally approved zoning plan. in other words, if Rond6nia's second approximation of the zoning is approved by the federal government, landholders in the state will no longer be obliged to keep 80 percent of their property forested, but rather 50 per- cent. Other incentives to support the zoning should also be considered, for example, providing subsidies and/or special revenue sharing to local governments in con- servation areas. Other states in Brazil such as Parana and Minas Gerais have had a positive experience with the so-called "ICMS ecol6gico." In this scheme, munici- palities that implement environmental protection policies are rewarded by receiv- ing from the state a higher proportion of the revenues from the locally raised value-added tax than they would have in other circumstances. Similarly, federal revenue-sharing with the states could be made conditional on the enforcement of land-use zoning. Creating markets for environmental services is yet another pos- sible way to provide incentives to conserve forest in designated areas. At present, international markets for carbon offsets offer the most promising option. References Alto Madeira. 1996. Porto Velho, Rond6nia, March 27, p. 6. Bartholo, R. S., Jr. 1997. "Diagn6stico Setorial da Economia do Estado," second preliminary version. Brasilia: ltmidas Project. Browder, John 0. 1992. "The Limits of Extractivisim." BioScience 42(3):174. Federac,o das Ind6strias do Estado de Rond6nia (FIERO). 1996. "Propostas para um Novo PLANAFLORO. Porto Velho: FIERO. - 1997. "Rond6nia, Perfil Socioecon6mico e Industrial." Porto Velho, Rond6nia: FIERO. Forum das ONGs e Movimentos Sociais que Atuam em Rond6nia. 1996. "Observac6es Criticas e Propostas de Reformulacao do PLANAFLORO: Resoluc6es da Assembleia." Porto Velho. INPE. N.d. Deforestation 1995-1997 Amaz6nia. Pamphlet. - 24 - Dennis J. Mahar and Ceile E. H. Ducrot Keck, M. E. 1997. "PLANAFLORO in Rondonia: The Limits of Leverage." In J. Fox and L. D. Brown, eds., The Struggle for Accountability: The World Bank, NGOs, and Grassroots Movements. Cambridge: MIT Press. Kengen, S. 1997. Forest Valuationfor Decision-Making. Rome: United Nations Food and Agricultural Organization. Mahar, D. J. 1988. Government Policies and Deforestation in Brazil's Amazon Region. Washington, D.C.: World Bank. Millikan, B. H. 1998. "Politicas Piublicas e Desenvolvimento Sustentavel em Rondonia: Situacao Atual e Abordagem para um Planejamento Participativo de Estrategias para o Estado." Unpublished. Porto Velho: Relatorio de Consultoria, Governo do Estado de Rond6nia/SEPLAN/PNUD. Moreira, C. 1995. "Las Perspectivas del Desarollo Sustentable en Rond6nia." Unpublished. Brasilia: United Nations Development Programme. Redwood, J. III. 1993. World Bank Approaches to the Environment in Brazil: A Re- view of Selected Projects. Washington, D.C.: World Bank Operations Evalua- tion Department. Schneider, R. 1992. Brazil: An Analysis of Environmental Problems in the Amazon. Washington, D.C.: World Bank. ___. 1993. Land Abandonment, Property Rights, and Agricultural Sustainability in the Amazon. Washington, D.C.: World Bank. Sombroek, W., and H. Eger. 1997. "What Do We Understand by Land Use Plan- ning: A State-of-the-Art Report." Agriculture and Rural Development, Rome. Verissimo, A., C. Souza, Jr., S. Stone, and C. Uhl. 1998. "Zoning of Timber Ex- traction in the Brazilian Amazon: A Test Case Using Para State." Conserva- tion Biology 12(1):128. Wells, M., and K. Brandon. 1992. Parks and People: Linking Protected Area Manage- ment with Local Communities. Washington, D.C.: World Bank/World Wildlife Fund/U.S. Agency for International Development. -25 - The Economic Development Institute (EDI) was established by the World Bank in 1955 to help promote international development. EDI conducts learning programs for officials from developing countries to assist them in planning and managing their investments more productively. The Institute produces and disseminates publications and electronic information products that support these objectives. For information on EDI publications write to: EDI Publications Economic Development InstitLite The World Bank 1818 H Street, N.W. Washington, D.C. 20433 Tel: (202) 473-6349 Fax: (202) 522-1492 Visit EDI's home page on the World Wide Web at: http://www.worldbank.org/edi