28362 NUMBER 52 *E4 Precis m Operations Evaluation Department June 1993 Involuntary Resettlement Substantial human costs result from 0 Assistance with relocation and Outcomes involuntary resettlement associated with support during the transition civil works projects in developing coun- period. Incomes: Incomes were higher tries. Though the Bank has had guide- following resettlement in two of the lines in force since 1980 to protect people Assistance in re-establishing four projects featured in the impact being involuntarily resettled, neither the former living standards. evaluations. This was the case for all Bank nor borrower agencies have obtained affected households in Thailand, and enough information regarding the extent, A central criterion for judging for those eligible for resettlement in scope, and impact of involuntary resettle- the performance of projects is the last Maharashtra where the irrigation ment. This makes it difficult to ascertain element: whether resettlers' living command area experienced a lower whether the Bank is living up to its guide- standards have been restored. To incidence of poverty, in both the lines or, when evaluating proposed new reach such a judgment requires field- resettled and host villages, than aver- projects, to predict how various plans for based data, gathered both at the age for nural Maharashtra state. For resettlement will actually affect people. project identification stage, to pro- most of the resettled households in vide a "baseline," and later, after Karnataka Irrigation and for all in To help fill this gap, a new OED project implementation. Yet apprais- Ghana Kpong Hydroelectric, real study reports on the Bank's early experi- al and evaluation reports often pro- incomes fell. ence with involuntary resettlement, vide patchy information on resettle- drawing on evaluation reports for 49 ment in general and on incomes and Eligibility: The eligibility and completed projects approved since 1969.* living standards in particular. entitlement to compensation, reloca- The key sources of data for the study are tion, and resettlement were satisfacto- four impact evaluations of resettlement in Impact evaluations rily defined only in Ghana, Kpong. projects that are now well into their oper- ational phase. The study outlines factors To provide information for the Living standards: Social infra- to consider in planning and implement- study, OED undertook impact evalu- structure services-water, health, ing resettlement efforts. It finds that ations of involuntary resettlement in education, electricity, access roads- compliance with the Bank's guidelines four Bank-supported projects in were generally better in all projects requires more effective Bank and borrower which resettlement was completed than before resettlement. In Ghana, performance in planning and executing five to ten years ago-long enough to Kpong, replacement housing was resettlement. A recurring theme through- expect resettlers to have overcome provided in addition to house plots, out is the complexity of resettlement initial resettlement problems and re- successfully combining traditional efforts and consequent need for a wide established their lives and productive architecture with modem amenities. range of skills. capacity (see Box). All of these projects were planned and became ______ Information on resettlement effective before the Bank issued its first guidelines on resettlement ~ "Early Experience w'itlh Re settle- The Bank's guidelines for invol- (1980). Socioeconomic surveys un- ment", Report No. 12142, June untary resettlement in the projects it dertaken for these impact evaluations 1993. OED reports are aailable supports (Operational Directive 4.30, show how the living standards of to Bank Executive Directors and June 1990) provide for: people resettled have changed over 51aff from the Internal Documents time. They also bring out the views Unit and fromn Re'gional Inforna- * Adequate compensation for lost of resettlers on compensation, reloca- tion Serzices Centers. assets. tion, and rehabilitation. catchment area of the former villag- Projects Selected for Impact Studies es; the affected people were of the same caste (in Maharashtra); and the Thailand: Khao Laem Hydroelectric at the end of 1978 and was completed host populations sympathized with Project ($80 million, approved 1979) in 1981. the plight of the affected families (in supported construction of a reservoir Karnataka). In Ghana, relations on the Quai Noi river to provide India: Karnataka Irrigation Project started out well but deteriorated water for power generation, irrigation, ($126 million, approved 1978) was and urban water supply in Bangkok, expected to complete two dams begun followi d dipesan failr and to help reduce the area subject to in 1962, bringing irrigation to about flooding. The region is heavily forest- 100,000 ha. The 9.6 km earth-filled dam acquired for the resettlers. ed, mineral-rich, and mountainous; at Narayanpur was the subject of the only the plains along the river and at impact evaluation. It was built in a Factors explaining outcomes the foot of the mountains were suit- relatively flat broad valley bordering able for farming. Civil works started on three drought-prone districts. Al- Experience of the impact evalu- in 1980; the dam was dosed and though the dam has been completed ations consistently shows that for reservoir filling began in 1984. Forty- and filling began in 1982, the reservoir resettlement programs to succeed, one villages were affected by the is being held at the level it reached in seven factors need to be present: creation of the reservoir. Resettlement the 1986 monsoon. At full capacity it plans provided for relocation of an will inundate about 132 sq km, sub- Government/agency commitment estimated 1,800 families, or 10,800 merging 36 villages plus some or all of people. Resettlement started in 1983; their land, as well as farmland from The satisfactory outcome of planned resettlement activities were another 54 villages. Resettlement resettlement in the Khao Laem Hy- completed in 1987. began in 1972 and is not yet complete. droelectric project and in the Second Ghana: Kpong Hydroelectric Project India: Second Maharashtra Irrigation Maharashtra Irrigation project, as ($39 million, approved 1977) support- Project ($210 million, approved 1979) measured by the increased incomes ed the construction of the Kpong supported an ongoing program to of those who were eligible for reset- reservoir on the Volta River for elec- build irrigation facilities. The Dhom tlement, has much to do with the tric power generation. Civil works storage dam, on the Krishna River, was degree of government/agency com- started in 1977; all generating units completed and the reservoir was first mitment. This factor underpins the built were commissioned by 1981. filled in 1976; it submerged about 2,500 other six factors. The reservoir displaced about 1,561 ha. In total, 2,840 ha were acquired households along a 24 km stretch of from 32 villages to accommodate the Strong implementing agency river, affecting small farming and reservoir and the dam construction site. fishing villages and the larger settle- Of the 32 villages, 13 were fully sub- Power agencies have the best ment of Kpong. Resettlement started merged and 19 were partially affected. record on resettlement. Most of _______________________________________________________ those studied were already estab- lished before the project and were Settlers' perceptions: Most reset- of landlessness increased in both Thai- well staffed for resettlement, with an tlers expressed dissatisfaction with land Khao Laem Hydroelectric and in adequate budget. In the Khao Laem their compensation and resettlement. Karnataka Irrigation. project, for example, the Electricity However, Bank guidelines require Generating Authority of Thailand project performance in resettlement Women: For women, who are (EGAT) proceeded on the principle to be judged according to data on engaged in farming in all three coun- of re-establishing the livelihood of incomes, not levels of satisfaction, tries, resettlement caused no observ- the families to be displaced. Thai- which are subjective and may also be able change in employment patterns, land had no national legislation influenced by other factors besides In Ghana, Kpong, care was taken concerning compensation and reset- resettlement. The discrepancy be- to preserve the thriving women's tlement, and thus it was EGAT tween incomes and levels of satisfac- pottery industry. Better village infra- which exhorted the government to tion highlights the need for income structure and the introduction of elec- take decisive action where there was data. tricity helped reduce labor-intensive no legal precedent. Policy issues tasks traditionally done by women. were successfully dealt with because Landlessness: Landlessness de- Nevertheless, some villages now have EGAT was organized for, and capa- clined slightly in Maharashtra Irriga- poorer access to markets and fire- ble of, managing resettlement. tion II. Nevertheless, a tracer study of wood-both women's responsibilities. households from some inundated Legislative framework villages found that almost half the Hosts: In both projects in India, affected farmers were not eligible for resettlers enjoyed good relations with An appropriate legal frame- resettlement; among this group, a host populations because the new work provides established and clear fourth were landless. The incidence villages were still within the social policies and guidelines that: June 1993 * adequately define eligible rounding the Kpong rapids in Gha- equipment. Their incomes were people; na, lack of early attention to plan- slightly higher than before, but no- ning the project's agricultural corn- ticeably higher than those of farmers * assure fair compensation; ponent, including cultivation practic- who had lost both land and house, es and the amounts of land required, who spent all their compensation * establish rights and responsi- was the main reason why resettled money to replace their houses. bilities, delineate powers and families received only about one duties, for relocation and rehabili- third as much farmland as they had Technical support is particularly tation for all affected persons; cultivated before resettlement. Ser- important in projects with resettle- vices to support their production ment components, since farming * provide the processes and activities were left to a post-project communities being resettled often timetables for addressing the many phase that never materialized. have no choice but to change their issues that arise in resettlement; farming systems-for example, by Development programs shifting from extensive to intensive * establish mechanisms for cultivation, from lowland irrigation handling grievances (regarding Too often in Bank-supported to upland farming, or from rainfed to cash compensation and, equally projects, the relocation of households irrigated agriculture. important, administrative deci- is regarded as the end of the resettle- sions); and ment exercise. Resettlers' dissatisfac- Community involvement tion partly reflects the lack of sound * determine priorities among development programs. In the Ma- Consultation with the affected conflicting national policies. harashtra project, which provided families can assist in both the plan- for resettlers to move into the irriga- ning and implementation process. It In Ghana, resettlement efforts tion command area, the develop- involves obtaining views on attitudes benefitted from a clear legal frame- ment program was a principal cause to evacuation, anticipated problems, work. In Karnataka, by contrast, a of satisfaction. expected assistance, and choice of resettlement bill was drafted in 1987 resettlement sites. Involving commu- with technical assistance supplied Land: In Ghana, Kpong and nities in decision making is impor- by the Bank, but is still awaiting Thailand, Khao Laem, families fac- tant for sustainability, since many of Presidential approval in 1993. Reha- ing resettlement wanted replacement the new resettlement sites have a bilitation was undertaken under the land above all else, and in both coun- more highly developed infrastructure aegis of government orders which tries, dissatisfaction with the out- that has greater needs for mainte- are not legally binding and can be come of resettlement centered on the nance. amended or revoked at any time. lack of replacement land. In the Kpong case, resettled households' Local leadership: The impact eval- If legislation cannot be enacted agricultural output was much lower uations showed that much depends to protect the livelihood of people to after the project than before, due on the calibre and foresight of local be affected by investment projects, both to the reduction in the size of community leaders, who are a vital as may be the case in countries their farms-and thus the continu- channel of communication between where the legal system is informal, ous cropping on lands which for- project authorities and affected com- then a political commitment to this merly had been left fallow for ex- munities. In the Chico project in the principle could be expressed in the tended periods-and the lack of any Philippines, community leaders form of basic entitlements for agricultural supporting services. In could not be persuaded to move to project-affected persons, agreed the Khao Laem project, farming did irrigated lands in the command area. upon early in the project planning not intensify as expected. Villagers were reluctant to leave their process. ancestral homes, especially while Irrigation: Of the four projects plenty of jobs were available around Comprehensive planning studied, only one-Maharashtra- the dam site. At impact evaluation, supplied irrigation. Irrigation was these opportunities were exhausted, Good planning helps explain the reason for the Maharashtra reset- and while the few families who had the satisfactory results in all projects tlers' higher incomes and lower pov- moved to the command area were in Thailand and in the Philippines erty levels (than before resettlement). very satisfied, leaders said they re- Magat. In all the four projects studied, sur- gretted the decision to remain be- veys showed that resettlers wanted hind. By then it was too late to ac- Often, however, planning could irrigation. In Kamataka, the low commodate more families in the have been better. Where socioeco- levels of compensation meant that command area. nomic/baseline surveys were car- some of the farmers who lost only ried out, they frequently lacked lands, not houses, invested their In Maharashtra, by contrast, local conceptual clarity. In the area sur- compensation money in irrigation leaders clearly saw resettlement as an OED Pricis opportunity to improve the commu- projects studied, most of the costs of Development programs nity's economic and social situation. compensation and resettlement The presence of fellow ethnic groups were to be met by the respective Bank policy embraces the use of in the receiving areas helped to governments. Ghana Kpong Hy- resettlement as an agent for change, smooth the transition, in that local droelectric partially met these costs including econon-c development. leaders campaigned to obtain release despite economic difficulties while All the surveys show that resettlers, of lands to give to the resettlers. several other projects failed to reach including tribal people, themselves planned targets for lack of funding. seek new opportunities. Yet project Maintenance arrangements: In the planning and implementation is Ghana, Kpong project, the Volta River Implementing Bank guidelines normally preoccupied with compen- Authority (VRA) made considerable sation and relocation, to the detri- efforts to involve the village chiefs in Among the Bank-supported ment of planning productive activi- setting up arrangements for running projects now in progress, 131 have ties and services to support them. and maintaining village services. The been identified as involving reset- chiefs' lack of interest then led VRA to tlement, affecting a total of 1.9 mil- Establishing productive activi- try to set up independent resettle- lion people. The Bank's guidelines ties to restore, or increase, former ment management committees, cover the relevant issues. But they standards of living is difficult to charged with this responsibility. But are not always applied and often go accomplish within the normal time- the chiefs opposed these committees, beyond the requirements of local frame of an investment project. One which they felt challenged their lead- legislation. A question also remains solution would be to move the reset- ership. Though the standard of vil- as to how many other projects es- tlement process upstream in the lage infrastructure was better than cape the guidelines altogether. An project cycle. Resettlement could before resettlement, the decline in ongoing Bankwide review of reset- start immediately after project initia- family incomes meant that house- tlement policy is expected to ad- tion and be completed a minimum holds were much less able to pay for dress this question. of two years before project closing to maintenance. In Thailand, the hand- help ensure that incomes are fully over of responsibility was well orga- Even in large infrastructure restored by project completion. nized; EGAT provided a revolving projects where displacement is very fund for maintenance of village infra- obvious, little information is gener- In other circumstances a sepa- structure. ally available on resettlement issues. rate development project (or sub- Together, the recommendations project, if the resettlement program Cost estimates outlined below suggest that the is quite small) to begin after people social aspects of civil works con- have been moved, designed to in- Costs of resettlement have almost struction projects need much great- crease the incomes of the affected always been underestimated. The er attention by the Bank and by families, may be the only practical most common reasons are: borrower governments. Unless solution. Compensation and reloca- these aspects are mastered, the tion could be the responsibility of * Underestimation of the scope of development enterprise itself will the agency implementing the resettlement activities, including continue to be dogged by public project, while the rehabilitation studies and planning, rehabilita- protests. project could be the responsibility of tion/transfer and maintenance, the particular line agency. development investments, and Minimizing displacement technical and social assistance. Data on living standards The importance of minimizing * Unrealistically low assumptions displacement cannot be overem- In the great majority of projects about unit prices for land acquisi- phasized. Realistic estimates of it is impossible to determine wheth- tion and social amenities. resettlement costs and borrower er living standards have been re- involvement should play an impor- stored. This is a serious omission. * Lack of detailed plans upon tant role here, since they may Not only does it give the impression which to base cost estimates. change the relation between the that the Bank is little interested in costs and benefits of alternative achieving this goal; it also deprives Another problem relates to fund- investments and may provide in- management of the information ing and disbursements. In all four centives to minin-Bze displacement. needed to take remedial measures. OED Pr&is is produced by the Operations Evaluation Department of the World Bank to help disseminate recent evaluation findings to development professionals within and outside the World Bank. The views here are those of the Operations Evaluation staff and should not beattributed totheWorld Bank or its affiliated organizations. Please address comments orenquiries to the managing editor, Rachel Weaving, E-1204, World Bank, telephone 473-1719. June 1993