China - Qinba Mountains Poverty Reduction Project

Document of The World Bank Report No: 32216 IMPLEMENTATION COMPLETION REPORT (IF-N0280 SCL-41870 TF-29450) ON A LOAN IN THE AMOUNT OF US$30 MILLION AND A CREDIT IN THE AMOUNT OF SDR108.5 MILLION TO THE PEOPLE'S REPUBLIC OF CHINA FOR THE QINBA MOUNTAINS POVERTY REDUCTION PROJECT June 24, 2005 Rural Development and Natural Resources Sector Unit East Asia and Pacific Region  CURRENCY EQUIVALENTS (Exchange Rate Effective May 2005) Currency Unit = Yuan Y1.0 = US$ 0.12 US$ 1.0 = Y8.27 FISCAL YEAR January 1 December 31 ABBREVIATIONS AND ACRONYMS 8-7 Plan National Seven-Year (1994-2000) Poverty Reduction Plan CAS Country Assistance Strategy CFPA China Foundation for Poverty Alleviation CWHRDC China Western Human Resources Development Center EMDP Ethnic Minorities Development Plan ERR Economic Rate of Return FCPMC Foreign Capital Project Management Center FRR Financial Rate of Return HRDC Human Resources Development Center IPMP Indigenous Peoples Mitigation Plan LGPR Leading Group for Poverty Reduction NPV Net Present Value PADO Poor Area Development Office PMO Project Management Office QAG Quality Assurance Group Qinba Project The Qinba Mountains Poverty Reduction Project QSA Quality of Supervision Assessment RSO Rural Survey Organization SAR Staff Appraisal Report SSB State Statistics Bureau SWPRP Southwest Poverty Reduction Project TVE Township and Village Enterprises Vice President: Jemal-ud-din Kassum, EAPVP Country Director David Dollar, EACCF Sector Director Mark D. Wilson, EASRD Task Team Leader: Guo Li, EASRD  CHINA QINBA MOUNTAINS POVERTY REDUCTION PROJECT CONTENTS Page No. 1. Project Data 1 2. Principal Performance Ratings 1 3. Assessment of Development Objective and Design, and of Quality at Entry 2 4. Achievement of Objective and Outputs 4 5. Major Factors Affecting Implementation and Outcome 14 6. Sustainability 16 7. Bank and Borrower Performance 17 8. Lessons Learned 19 9. Partner Comments 21 10. Additional Information 23 Annex 1. Key Performance Indicators/Log Frame Matrix 24 Annex 2. Project Costs and Financing 26 Annex 3. Economic Costs and Benefits 28 Annex 4. Bank Inputs 31 Annex 5. Ratings for Achievement of Objectives/Outputs of Components 34 Annex 6. Ratings of Bank and Borrower Performance 35 Annex 7. List of Supporting Documents 36 Annex 8. The Qinba Project in A Farmer's Eyes 37 Annex 9. Borrower's ICR Summary 40 Maps IBRD 27815  Project ID: P003590 Project Name: Qinba Mountains Poverty Reduction Project Team Leader: Guo Li TL Unit: EASRD ICR Type: Core ICR Report Date: June 24, 2005 1. Project Data Name: Qinba Mountains Poverty Reduction Project L/C/TF Number: IF-N0280; SCL-41870; TF-29450 Country/Department: CHINA Region: East Asia and Pacific Region Sector/subsector: General agriculture, fishing and forestry sector (62%); Other social services (17%); General education sector (9%); Central government administration (9%); General transportation sector (3%) Theme: Other social protection and risk management (P); Poverty strategy, analysis and monitoring (P); Participation and civic engagement (S); Other rural development (S); Other environment and natural resources management (S) KEY DATES Original Revised/Actual PCD: 03/20/1996 Effective: 10/19/1997 10/16/1997 Appraisal: 01/10/1997 MTR: 06/04/2000 06/06/2000 Approval: 06/10/1997 Closing: 01/31/2004 12/31/2004 Borrower/Implementing Agency: The People's Republic of China/Sichuan, Shaanxi, and Ningxia Provinces Other Partners: STAFF Current At Appraisal Vice President: Jemal-ud-din Kassum Jean-Michel Severino Country Director: David R. Dollar Nicholas C. Hope Sector Director: Mark D. Wilson Joseph R. Goldberg Team Leader at ICR: Guo Li Alan L. Piazza ICR Primary Author: Guo Li 2. Principal Performance Ratings (HS=Highly Satisfactory, S=Satisfactory, U=Unsatisfactory, HL=Highly Likely, L=Likely, UN=Unlikely, HUN=Highly Unlikely, HU=Highly Unsatisfactory, H=High, SU=Substantial, M=Modest, N=Negligible) Outcome: HS Sustainability: HL Institutional Development Impact: H Bank Performance: S Borrower Performance: HS QAG (if available) ICR Quality at Entry: S Project at Risk at Any Time: No  3. Assessment of Development Objective and Design, and of Quality at Entry 3.1 Original Objective: The Qinba Mountains Poverty Reduction Project (the Qinba Project) had the following development objectives: (a) to demonstrate the effectiveness of a focused multi-sectoral rural development project approach to poverty reduction; (b) to significantly reduce absolute poverty in 26 of the very poorest counties in Sichuan, Shaanxi and Ningxia provinces; (c) to facilitate increased labor mobility from the poor areas to better-off rural and rapidly growing urban areas; (d) to establish a new micro-finance mechanism to provide credit services to poor households; (e) to stabilize or even reverse upland environmental destruction; (f) to encourage greater local community participation; and (g) to upgrade poverty monitoring at the national and local levels. These objectives were carefully examined and jointly chosen by the Government and the World Bank to ensure their consistency with the Government's National Seven-Year (1994-2000) Poverty Reduction Plan (the 8-7 Plan) and the Bank's Country Assistance Strategy (CAS). First, there was an adequate and effective involvement of the Leading Group for Poverty Reduction (LGPR) under the State Council in the whole project life; therefore, to a large extent, the Qinba Project was part of the Government's overall strategy to explore new models of poverty reduction. The development objectives, therefore, were highly consistent with the 8-7 Plan which called for innovative measures to address issues in the following key areas: rural infrastructure, farming activities, improved access of the poor to off-farm employment opportunities, human capital development, rural industrial development, and institutional building, including setting up an independent and objective poverty monitoring system. Second, these objectives were also highly consistent with the Bank's CAS, directly addressing two of the Bank's top operational priorities in China -- poverty reduction and environmental protection. Third, the selection of these objectives built upon the strength of the Southwest Poverty Reduction Project (SWPRP, Project ID P003639). The successful implementation of SWPRP confirmed that the multi-sectoral model of poverty reduction represented a significant advance over previous poverty reduction mechanisms in China and was worthy of widespread adoption. The Qinba Project focused more on the components which were most warmly received by the project's primary beneficiaries (e.g., land and farmer development, rural infrastructure, labor mobility) under SWPRP. 3.2 Revised Objective: The original objectives remained unchanged throughout the project's life. 3.3 Original Components: The project's objectives were to be achieved through the implementation of the following components: 1. Labor mobility (13% of appraisal total project cost). This component was designed to provide employment to about 155,000 upland laborers over six years and to provide an important testing ground for assisting the organized flow of rural laborers into China's fast-growing regions. The goal of the component was to be achieved through: (a) establishing a voluntary system of enhanced rural labor mobility for the upland absolute poor; (b) providing skill training through an enterprise reimbursement system; and (c) building a computerized demand-driven job placement system emphasizing local markets, monitoring of worker safety and living conditions, and reporting of abuses and grievances. This was a highly innovative component. 2. Rural infrastructure (11% of appraisal total project cost). This component was designed to improve the poor's access to basic infrastructure including rural roads, human and animal drinking water supply, small-scale irrigation, agricultural drainage works, biogas digesters and rural electrification. Rural - 2 -  infrastructure construction has been regarded as the foundation for improving the poor's basic living and production conditions. 3. Land and farmer development (54% of appraisal total project cost). This was the biggest component in terms of investment and was designed to create a substantial and sustainable increase in agricultural productivity and incomes for project households. The component also intended to stop (and in some cases to reverse) the trend of environmental degradation. A menu of crop and livestock activities was developed on a participatory basis, from which the project households were to select their favored activities. In addition, the component supported: (a) extension of improved agriculture and smallholder livestock management techniques and inputs; (b) establishment of tree crop nurseries and provision of saplings and tree crop management extension; (c) intensification of staple crop production on the modest amount of available flatland; (d) terracing for improved water and social conservation; and (e) the support of provincial and regional applied agricultural technology research. 4. Township and village enterprises (TVE) development (17% of appraisal total project cost). The component was designed to provide credit for labor-intensive commercially viable agroprocessing, mineral, service and handicraft industries that had strong backward linkages to poor households and met environmental safeguards. The enterprises were to be selected according to a set of criteria which would maximize benefits to the poor: (a) cost per job created; (b) commercial feasibility; (c) work safety and environmental risks; (d) skill training; and (e) backward linkages to products produced by project households. The selected enterprises were to be financed under sub-loans made by the project provinces on the Agricultural Bank of China's standard interest terms for similar activities. 5. Micro-finance (1% of appraisal total project cost). This was a pilot-based component with the objective of establishing a sustainable financial system to provide working capital loans (initial loans ranged from RMB400 to RMB700, and larger loans no more than RMB8000) to micro-enterprises run by poor households in project areas. The main methodology was to employ an adaptation of the Grameen Bank model (emphasizing a joint liability lending). 6. Institution building and project management (3% of appraisal total project cost). This component was designed to improve the capabilities of the Foreign Capital Project Management Center (FCPMC) and project management offices (PMOs) at all levels in project design and implementation, procurement, accounting, and internal monitoring and evaluation. 7. Poverty monitoring (1% of appraisal total project cost). This component intended to use improved State Statistics Bureau (SSB) survey instruments to provide a credible and comprehensive poverty profile, analyze the accuracy of the targeting of benefits to the absolute poor, and measure and evaluate the impact of the project's individual components. There were four main elements of this component: (a) to upgrade the existing poverty monitoring system developed under SWPRP in order to adapt it to the conditions of the Qinba mountains region and more fully exploit the potential of the survey instruments for project evaluation; (b) to undertake baseline and annual follow-up surveys; (c) to process and analyze the collected survey data; and (d) to disseminate findings and information. 3.4 Revised Components: The TVE component performed poorly during the initial years of the implementation and was revised during the mid-term review, which was conducted in 2000. The size of the component was significantly reduced, from about 17% of total project cost in the Staff Appraisal Report (SAR) to around 6%. The funds released from the TVE component were used for more pro-poor programs, such as rural infrastructure and agricultural components which were most warmly welcomed by the beneficiaries. - 3 -  Many reasons explain the poor performance of the TVE component, but there are two factors worth special attention. First, there was a high degree of government intervention in the implementation of this component. The selection of enterprises, therefore, was not strictly based on the predetermined the set of criteria (see Section 3.3). Second, all the enterprises suffered from a serious lack of equity which resulted in a lack of working capital. It was impossible, therefore, for these enterprises to create job opportunities for the poor in the project area. In addition, the expected prosperous market for the locally produced raw materials did not emerge. Instead of generating tax revenue for the local governments, in many cases these enterprises became new fiscal burdens to the local governments and hence worsened local fiscal crises. All these results were contrary to the original component expectations. Useful lessons should be drawn from the implementation of the TVE component (see Section 4.2). The significant reduction in the scale of the TVE component, on the other hand, clearly demonstrated the flexibility of the project design and advantages of a multi-sectoral project. The funds released from the TVE component were mainly used for the Rural Infrastructure component and the Land and Farmer Development components, which were among the most welcomed components of the Qinba Project. For example, by the end of the project implementation, the size of the rural infrastructure component increased by about 74% (see Section 4.2), which was hailed by beneficiaries and PMOs at different levels as an unexpected huge success. 3.5 Quality at Entry: The project design predated the existence of the Quality Assurance Group (QAG) and hence there was no official assessment of the project's quality at entry. However, it is believed that the quality at entry of the project was satisfactory for the following reasons. First, there was a high degree of conformity of the project objectives with the CAS priorities and with the Government's strategy for poverty reduction. Second, the design incorporated lessons learned from previous broad-based rural development projects, particularly experience from the successful implementation of SWPRP. Third, the design paid adequate consideration to the challenges facing this project, including its high degree of complexity and number of innovations. This consideration was fully reflected by the features of the project design, including choice of project management and organizational structure. Fourth, the design was fully consistent with the Bank's safeguard policies. Finally, and perhaps of most importance, the project achieved a profound development impact by the end of the implementation period without modification of its original objectives or any major revision of the project component structure (see Section 4.1). Quality at entry, however, could have been even higher had the following issues been scrutinized and addressed more carefully. First, the probability of the assumptions employed for designing the TVE component should have been assessed more carefully. Second, a mechanism for ensuring even stronger participation from beneficiaries at the early stage of the implementation should have been established. Third, more user-friendly and appropriate procurement methods with greater community participation should have been employed at the early period of the project implementation. 4. Achievement of Objective and Outputs 4.1 Outcome/achievement of objective: The achievement of the objectives and outputs of the Qinba Project was highly satisfactory. The Qinba Project is an excellent example of achieving the intended development impact. First, the project generated a significant and profound policy impact. In its statement of nominating the World Bank as the recipient of the China Poverty Eradication Award on September 6, 2004, the Government made the following strongly positive evaluation of the Qinba Project: - 4 -  "... In Southwest and Qinba projects, the World Bank adopted several innovative approaches, such as village development planning and project based multi-sectoral methodology, which have generated great impact on developing China's poverty alleviation policy. Particularly, the village development planning, labor mobility, and project-based multi-sectoral approach have already been incorporated into the China Poverty Alleviation Plan (2001-2010) by the Central Government, which will have profound impact on China's poverty alleviation efforts in the twenty-first century." In addition, the World Bank became the first international organization to receive the China Poverty Eradication Award on October 17, 2004. In an article published in the Guangming Daily on March 9, 2005, the Ministry of Finance summarized the Qinba Project's policy impact in highly favorable way: "... Southwest and Qinba projects, through introducing project based multi-sectoral methodology and participatory approach, have made significant contributions in helping the Government to develop `a scientific approach to development' which is characterized by `putting the people first'." Second, the project brought about remarkable economic and social improvements. The SSB's independent monitoring of the project provides detailed information on these benefits (see Table 1). For example, the incidence of poverty was drastically reduced from 26.7% (1997) to 4.6% (2003). The current level of poverty incidence in the project area is now almost the same as the national average. However, at the beginning of the project in 1997, the incidence of poverty in the project area was five times as high as of the national average. In 2003, 100% of project villages had road access compared with only 72% in 1997. By the end of 2003, per capita income in the project area had on average increased about 64% compared with that of 1997, an annual growth rate at 8.1%. By comparison, the national average growth rate of farmers' per capita income was around 4% during the same period of time. The poor farmers' ability to deal with risks, as measured by the degree of food security and per capita cash at hand at the year-end (see Annex 1 for all key performance indicators), was also significantly strengthened. SSB's independent monitoring and evaluation also showed that, after seven years of project implementation, the project villages are evidently less vulnerable than non-project villages to natural disasters. Table 1. Selected Indicators of Economic and Social Benefits 1 2 3 4 5 6 7 8 9 Year Poverty Food p. c. Villagers Villages Labor Remittance proportion Gender incidence security* income w/ access w/ migration of meat gap** safe water access sold in (%) roads market (kg) (yuan) (%) (%) (%) (billion yuan) (%) (%) 1997 26.7 190 854 n.a. 71.8 15 n.a. 38.7 89.8 2000 17.9 331 1087 75 91.0 22 n.a. 42.7 96.7 2003 4.6 387 1398 86 100 25 1.7 55.9 98.2 * Food security is measured by per capita grain storage at the year-end. ** Gender gap is measured by the ratio of girls enrollment rate over boys' enrollment rate for the age group of 7-15. Third, poor farmers' access to off-farm employment opportunities was significantly improved. The labor mobility targets were exceeded in all project counties and the overall quality of implementation was excellent (see Section 4.2). Households which participated in labor mobility not only realized rapid and substantial increases in income from worker remittances, but also significantly expanded their knowledge and views about the outside world and strengthened their self-confidence, which will have a lasting - 5 -  development impact. Interviews and project data confirmed that the labor mobility component was widely considered to be one of the most successful and well received components of the project. Fourth, the goal of piloting micro-finance was also successfully achieved, particularly in the areas of institutional capacity building and reaching the poor. However, given the current status of the overall financial sector in China, there is still a long way to go to achieve financially sustainable micro-finance institutions. Fifth, the project also generated significant environmental and ecological benefits, which were confirmed by an independent environmental monitoring report in 2003. The forest area increased by 49%, the irrigated area by 20%, and the terraced area by 22%. The ratio of households with difficulty obtaining daily fuel materials, an indicator used to measure the change of ecological environment, decreased from 62% in 1997 to 41% in 2003. Sixth, participation levels of the Qinba Project's farm households and communities increased over time, particularly in the second half of the project implementation period. In the later stage of project implementation, participatory monitoring and evaluation was introduced to evaluate outcomes and impact of the project. It was evident that the targeted groups started to influence and share control over priority setting, policy-making, resource allocations and access to public goods and services. Nevertheless, the participatory work could have been better had an even more efficient mechanism been established at the early stage of the implementation. Finally, in its 2004 monitoring report, SSB concluded that, compared with the SWPRP, "...Qinba poverty monitoring component improved the sampling methods, questionnaire design, training, and field interview," and this had helped upgraded poverty monitoring at the national and local levels, and hence had an impact which went far beyond the project itself. 4.2 Outputs by components: 1. Labor mobility ($48.6 million SAR, $46.1 actual) The labor mobility component is rated highly satisfactory. This was a very challenging and highly innovative component, and the first step was the establishment of a capable implementation system with offices and staff at all levels in both originating and destination areas. The initial institution building effort was fully successful and facilitated the achievement of about 133% of the SAR planned component output. More importantly, extensive and rigorous analysis has convincingly documented that the labor mobility component was extraordinarily effective in achieving its development objective of expanding the upland poor's knowledge of and access to off-farm employment. The outflow of migrant laborers from the project's upland villages has continued to expand following project completion, and this clearly shows that the favorable output and development outcome are fully sustainable. The China Western Human Resources Development Center (CWHRDC) spearheaded the institution building at the central level, and played the crucial role in: (a) establishing the implementation organization at all levels; (b) identifying employment opportunities in the destination areas; and (c) monitoring the location and well-being of the very large number of migrant laborers spread across a wide geographic area. The Sichuan and Shaanxi Human Resource Development Centers (HRDC) worked closely with the provincial PMOs to establish the implementation organization at the county and township levels and to supervise the work of these lower level organizational units during project implementation. The county and township level units played the key role during implementation of identifying appropriate candidates for the labor mobility component and to provide them with orientation training prior to departure for their new - 6 -  jobs. This highly capable and extensive implementation organization made it possible to achieve the component's planned output targets. Nearly 200,000 upland poor found off-farm employment through the labor mobility component, including roughly 110,000 inter-provincial migrant laborers. The labor mobility monitoring system in combination with the pre-departure orientation training kept work related accidents and employment disputes to an absolute minimum. CWHRDC found that less than 0.04% of all project migrant laborers experienced serious accidents (27 reported) and disputes (40 reported). CWHRDC has conducted a careful and extensive assessment to understand the cost and benefits of the component (see Document No. 9 of Annex 7). The assessment provides estimates (based on sample surveys) of the incomes received by the labor migrants, their remittances sent home to their extended families, and the ways these incomes were utilized by the migrants and their families. These estimates show that the labor mobility component was a very powerful poverty reduction measure: total remittances are estimated to have been about RMB1.7 billion (or about RMB7300 per migrant on average), with most of these funds being used for the purchase of agricultural production materials, improved housing, family medical care, educational expenses, and debt repayment. More importantly, all of the documentation and participatory evaluation indicates that the most significant outcome has been the improved outlook on life and greater aspirations of the migrants. This was clearly the case for migrant women (comprising about one quarter of all migrant laborers), who gained greater self esteem and confidence, reduced work burdens (upon return to their home villages), and increased economic independence. The component's policy impact is also evident. Based in part on the highly successful experience of the Southwest and Qinba Mountains Poverty Reduction Projects with labor mobility, China's poverty reduction program for 2001-2010 includes a major focus on labor mobility and vocation training. The vocational training program seeks to further improve the upland poor's access of higher-paying off-farm employment. Expanding on their successful experience with labor mobility under the Qinba project, for example, the Shaanxi Poor Area Development Office (PADO) provided vocational training opportunities for some 30,000 upland poor during 2004. Based on contractual relationships with the vocational training schools, most of these upland poor were placed in better paying jobs following three to six month training periods (final payment to the vocational schools was dependent upon successful placement of the trainees in new jobs). 2. Rural infrastructure ($40.5 million SAR, $70.6 million actual) This component is rated highly satisfactory. The rural infrastructure component was one of the most warmly welcomed components. In response to the farmers' demand, the size of the component was increased significantly to $70.6 million actual, an increase of 74% over the estimated cost at appraisal. The funding for increasing this component mainly came from funds released from the TVE component, increased counterpart funding, and the original unallocated funds. Most of the increased funding was used for constructing access roads, rural energy supply, and small irrigation systems. Implementation of the component significantly improved the project beneficiaries' basic living and production conditions. Project information from Sichuan and Shaanxi showed that 2,258 km of roads were improved and constructed to Class IV Road standards, as compared with 892 km of new and improved roads envisaged at project appraisal. In addition, 7,493 km of village access roads were constructed. Road access coverage has now been extended to about 1.9 million people, more than three times the originally planned number of beneficiaries. Village access roads, generally varying from several km to 15 km in length and accessible to tractors and light vehicles, have been effective in linking many natural villages and hamlets to existing rural road networks, drawing them for the first time into the local main stream of rural economic activities. Project households demonstrated great enthusiasm for this component, particularly for road construction, - 7 -  with a very active participation. Beneficiaries' contributed labor far exceeded in value the amount of cash injection by the project. Construction of drinking water facilities was also one of the most welcomed activities. By the end of the implementation, the project has successfully solved the drinking water problem for about 650,000 people, compared with the SAR target of 466,000. These are low cost, simple facilities which have effectively relieved tens of thousands of project households from the daily hard labor of fetching drinking water from remote sources, thereby setting free numerous labor-days to be profitably utilized for food production and income generation activities. In addition, small irrigation systems constructed through the project were very effective in stabilizing and increasing production of paddy and other food crops for home consumption and animal feed where there is surplus. One useful lesson from the component implementation is that infrastructure construction, particularly road access, should be implemented first so that the benefits of activities under other components (e.g., agricultural component and labor mobility component) can be maximized due to improved access to input supplies, services, information, and markets. 3. Land and farmer development ($ 194.8 million SAR, $232 million actual) This component is rated highly satisfactory. This is the largest component in terms of investment and focused on solving the food security and cash income generation problems. As shown by the SSB's independent monitoring report, both of the two major goals were fully achieved through activities such as farming, raising livestock, improving land quality, and applied agricultural research. For example, average per capita income increased by about RMB544 (from RMB854 in 1997 to RMB1398 in 2003), which exceeded SAR target (RMB400). The per capita amount of stored grain at the year-end was more than doubled from 190 kg/person in 1997 to 387 kg/person in 2003 (see Table 1, Section 4.1). Project area field visits consistently confirmed that the food security is no longer a problem for almost all farm households. In addition, the general living standard of the poor households has been significantly raised. Another powerful indicator is the consumption structure, as measured by the Engel coefficient, which was significantly improved. By the end of 2003, the project villages' Engel coefficient--the ratio of expenditures on food consumption out of total expenditures--declined from 64% to about 53%, and this decline for the project villages was greater than that of non-project villages. The poor households' ability to deal with risks and degree of market participation were also significantly strengthened (see Annex 1, Key Performance Indicators). Technical quality of project implementation has generally met the project's technical standards. In response to the farmers' demand, the size of the component was increased from $195 million to $232 million, an increase of about 19%. The funding for increasing this component mainly came from the funds released from decreasing the TVE component and unallocated fund. This component also included supporting activities of applied research (total about $0.5 million). About 75% of applied research investment was in Shaanxi Province. The applied research program was a notable success for three principal reasons: (a) it was carefully focused on specific production bottlenecks in the farming system; (b) the research was done at the beginning of the project so that the results could be applied through the extension activities in later years; and (c) many of the research results was also applied to other parts of the project provinces. As a result, the applied research program had a much wider impact than originally anticipated. The coverage rate (i.e., the proportion of project households which participated in the Land and Farmer Development Component) was basically 100% in all three provinces, with an average investment intensity of RMB1988 per household. In addition to the investment, the project also provided technical training to - 8 -  the project households. Women made up about half of those who received training; this represents significant progress compared with the beginning of the project implementation when women were only about 20% of the total. There were many factors which contributed to the success of the component's implementation. First, there was a synergy between the development of diversified farm structures and labor mobility cash income diversification, since both helped to reduce income risk and reduce households' vulnerability to falling back to poverty. Second, there was a strong participation in terms of activity selection by project households, which led to increased profitability of component activities and increased repayment rates. Third, the provision of series small support (ranged from RMB500 to RMB1200 per year) enabled farm households to build their asset base gradually over time with minimal risk and debt exposure. Fourth, the development of farmer human resource capacity through training programs and the building of farmer self-confidence had a major impact and provided the human capital base for long-term economic development in the project areas. Finally, there was high degree of integration of rural infrastructure with land and farmer, and labor mobility activities, which was a major catalyst to economic development through providing access to inputs, markets, services and information. There were some factors which constrained the component from achieving an even higher degree of success. First, tree crop activities did not fully achieve the expected benefits due to the long period (usually 4-5 years) before production and harvesting began, often weak training and extension, variable seedling quality, market changes, and seedling vulnerability to drought. In future poverty projects, tree crops should generally be included as part of an integrated, diversified farm investment approach rather than as single project investments for poor households. Second, an inappropriate choice of procurement methodology (i.e., time-consuming distribution process and a shrinking price difference between domestic and international markets of international competitive bidding procured fertilizer, chemicals, and plastic mulch, etc.) at the early stage of the project implementation caused delay of the activities. 4. TVE Development ($62.2 million SAR, $24.5 million actual) This component is rated unsatisfactory. Although the design of this component incorporated the lessons learned from initial implementation of the similar component in SWPRP, it turned out that the component was still too ambitious and underestimated the difficulties associated with the TVE sector as well as the rapid progress China was achieving in transforming its economy from a central-planned one to a market-oriented one. Following the mid-term adjustment, the total amount of investment for the component was cut down from US$62 million to US$25 million (i.e., from 17% to around 6% of the total project investment). The greatest reduction was in Sichuan. Investment in Shaanxi for this component was moderately adjusted down from US$28 million to US$20 million. The released funding was reallocated to the activities which were warmly welcomed by the beneficiaries (such as the rural infrastructure component, land and farmer development component). At the time of appraisal, the Qinba Project was expected to support 150 enterprises. At the time of the project completion, only 22 enterprises had been completed. Of these, 12 were operational, 3 were in trial production and 7 were either not operational yet or had stopped operation, mostly because of a lack of working capital for operation. All of the enterprises which were not-operational are in Shaanxi Province. Of 23 small farmers market, 20 were operational, a much higher proportion than that for the enterprises. It should be pointed out that, while most of the medium and larger TVEs supported by the project were eventually found not to be financially viable or to have had significant poverty reduction impacts, small scale enterprises and small scale farmers' markets were typically far more successful. These enterprises, which were often privately owned, provided goods and services for the local market and relied on simple technologies. Most of these enterprises were found to be financially viable, to have had favorable poverty - 9 -  reduction and local fiscal impacts, and to have provided marketing channels for farm and other products produced by the poor. Another indicator of success of the farmers' markets construction is that Sichuan Province, while significantly reducing the size of the component as a whole, doubled the number of markets constructed from 6 to 12. The implementation of the TVE component has provided the following lessons. First, the experience of the component in Shaanxi indicates that it was unwise for the project to invest in large-scale enterprises, given the natural and economic conditions and the level of government intervention in these counties. Moreover, the project should have mitigated the risks by not focusing the investment in the six enterprises of the three project counties. The investment in large-scale enterprises suffered from moral hazard problems when the enterprise owners did not have sufficient capital funds, as no individuals are finally responsible for the loan repayment. Second, the time-consuming process of selecting enterprises and the relatively high rates of interest charged to them may have an adverse effect on the performance of the component: those enterprises which are able to access commercial loans (interest rates were around 5.5 per cent per annum) would not be interested in the project loans. This meant that it was the enterprises without commercial sources of finance which were most likely to apply for the project's funding. Third, the small plants and small farmers' markets performed better than the large-scale enterprises. This is because these activities require less capital investment and working capital, and less complicated technology and hence take less time to prepare. All these fit the poor areas better. Finally, agroprocessing plants do not necessarily add value to the farm products produced in the project areas. A money-losing processing enterprise could have an adverse impact on the incomes and welfares of poor farmers in the project areas, as the local governments may have to collect more taxes and fees to make up the loss and/or persuade farmers to grow the raw materials for the plant which may not be in the best interest of farmers. 5. Micro-finance ($4.1 million SAR, $6.8 million actual) The component is rated marginally satisfactory. In terms of the number of borrowers and the total amount of loans disbursed, the program has attained its targets and has so far provided almost RMB60 million of micro-loans to more than 40,000 borrowers in two project provinces. By the end of June 2004, the program had a loan portfolio of over RMB10 million in Langzhong and Tongjiang counties of Sichuan (the program in Shaanxi was completed in September 2002). The program targeted the poor in the project counties and both the micro-loans and the training program provided by China Foundation for Poverty Alleviation (CFPA) to the program beneficiaries have impacted positively on the poor in the project counties. Of all the micro-loans, more than 70 per cent has been used by the project beneficiaries for cropping and animal production and the rest mainly for transportation and agroprocessing. The program was responsible for a number of innovations, including group lending, dynamic incentives and emphasis on women's participation. Project beneficiaries decided on loan uses, and loans were provided in cash rather than in kind. In addition, the program established basic loan disbursement and repayment procedures, an accounting and reporting system, and an internal control system. The program is more transparent than other government channels in fund use and able to detect and trace the funds diverted by the county and township governments from the component. Consequently, most micro-finance funds did reach the poor households in the project areas, and the program has achieved a much higher overall loan repayment rate compared with the agricultural component. The micro-finance program in the project counties, however, will have a difficult time achieving financial sustainability, because (a) the program has no capital funds and almost all the funds for the micro-finance - 10 -  operation are from the Bank loan that need to be repaid; (b) the rate of interest charged by the program (7% nominal and about 11% effective) barely cover the operational costs and the cost of funds (around 3.2 per cent per annum); and (c) the start-up cost is higher and is not covered by the grant funds. Moreover, the county governments in Ankang and Langzhong have been reluctant to use the project funds for the micro-finance component, and this led to frequent conflicts between the county governments and CFPA, the implementation agency in Beijing. The program's major loan quality and financial indicators are shown in Table 2. Table 2. Key Performance Indicators for the Micro-finance Component