Refining Policy with the Poor: Local Consultations on the Draft Comprehensive Poverty Reduction and Growth Strategy in Vietnam

In March 2001 the Government of Vietnam produced an Interim Poverty Reduction Strategy Paper (IPRSP) and announced its intention to develop a Comprehensive Poverty Reduction Strategy Paper (CPRGS) by the end of April 2002. In the IPRSP, the Government outlined its commitment to involve a broad range of stakeholders - including poor communities, local government authorities, and the enterprise sector - in drafting the CPRGS. The Ministry of Planning and Investment, who was assigned by the Government of Vietnam to lead the CPRGS drafting process, asked the World Bank and a group of international nongovernmental organizations to support them in carrying out the local consultations. The consultations took place in six rural and urban locations across Vietnam selected to represent a range of poverty situations. About 1,800 people participated in the research. This report, which is the first of three volumes documenting the local consultations, provides an account and reflection on the approach and methodology used in the consultations. It is intended this may give useful practical experience for future monitoring of the CPRGS as well as for people who are planning to carry out similar exercises in other countries. The report outlines the process that was followed from the point of developing a research outline from the IPRSP, through the fieldwork exercises, data compilation, and analysis, leading on to identification of the main policy messages made by the participants. It also describes how the findings were used to influence the final version of the CPRGS. This paper - a product of the Poverty Reduction and Economic Management Sector Unit, East Asia and Pacific Region - is part of a larger effort in the region to improve participatory processes in the formulation of policies for poverty reduction.


Preface
In May 2002, the Prime Minister of Vietnam approved the Comprehensive Poverty Reduction and Growth Strategy (CPRGS). This had been prepared over the preceding 14 months by a drafting committee of 52 government officials representing 16 agencies and ministries. The final document drew on a wide range of information sources which included analytical work produced both inside and outside Government. It was also informed by a series of consultation exercises that took place at national, sub-national and community levels.
This volume is one of a series of three reports that describe the work that took place to consult poor communities and local officials on the content and direction of the CPRGS, under the overall guidance of the Government-donor-NGO Poverty Task Force. This work was carried out in six sites across Vietnam at the request of the Ministry of Planning and Investment (MPI) by Actio nAid, Catholic Relief Services, Oxfam GB, Plan in Vietnam, Save the Children UK and the World Bank. During this exercise more than 1800 people from poor communities were asked to relate the proposals contained in the Government's strategy to their own lives and experience of poverty, to suggest improvements or revisions to the strategy and to highlight any gaps that could diminish the impact of the strategy.
The first report in this series describes how the consultations research was designed and implemented. This exercise was one of the first attempts in Vietnam to refine policy direction with poor people in such depth and on such a large scale. As such, those coordinating, managing, designing and conducting the research -collectively more than 80 people -faced a number of challenges in their work. The first volume has a very practical focus on these challenges: how do you use a government strategy as a communication device with poor households? How do you move from a broadly-phrased strategic document to a research framework that outlines questions and techniques that will make sense to people in poor, rural communities? How do you take the very detailed information gathered at a community level and translate it into policy messages? And how can you make sure the findings influence the substance of the strategy? This volume might be of interest to those planning similar work elsewhere.
The second and third reports summarize the findings of the research in different ways. The second report synthesizes the findings from the six sites by the five broad policy areas addressed in the research: • Trends in poverty; • Creating opportunities for poor people and supporting livelihoods; • Improving access to high quality basic social services; • Reducing vulnerability; and, • Institutional arrangements for delivering the poverty reduction strategy.
This report was prepared for the CPRGS drafting committee to facilitate the process of incorporating the findings into the final CPRGS. It includes a matrix of key policy measures and public actions attached as an annex.
ii The third volume compiles the 15-page site reports from each of the six research sites: • Lao Cai Province in the northern uplands; • Ha Tinh Province in the north-central coastal region; • Quang Tri Province in the central coastal region; • Vinh Long Province in the Mekong Delta; • Tra Vinh Province in the Mekong Delta; and, • Ho Chi Minh City -Vietnam's largest city.
This final volume is likely to be of most interest to an audience within Vietnam, particularly those working in areas that experience similar poverty situations as in these sites. For those working at the local level in Vietnam -particularly those involved with supporting the Government as they seek to implement the CPRGS -these site reports provide a helpful overview of the key policy issues as identified by the poor in those areas.

Volumes in the series:
Volume I

INTRODUCTION
This report is one of a three-volume set of documents that describe the process and present the findings from a research exercise in Vietnam in which poor households and communities were consulted on the content of an Interim Poverty Reduction Strategy Paper (I-PRSP). This was done at the request of the Ministry of Planning and Investment (MPI) as they worked to develop the I-PRSP into a Comprehensive Poverty Reduction and Growth Strategy. The two companion documents present the findings in different formats -one volume summarizes the findings by thematic area for policy action, the other collates the site reports and structures the findings geographically.
This report outlines the methodology used in the consultations and describes the processes followed to ensure that the findings were influential in developing the Government of Vietnam's Comprehensive Poverty Reduction and Growth Strategy (CPRGS), which grew out of the I-PRSP 1 . In part, it has been written to respond to numerous requests for information on how the consultations were carried out in Vietnam. In particular it seemed to the authors that a number of researchers are grappling with similar issues of how to bring more voices into the debates around policy choices -not just in a general sense but also in order to contribute to specific government strategies at the time that they are being formulated. By documenting the experience in Vietnam we hope to throw light on some questions that development practitioners elsewhere are asking. These include: • How can an I -PRSP, written in the (at times unclear) language of Government, be converted into a communication tool for dialogue with poor households, men and women, and local communities about policy options and areas for Government action? • What techniques might help at the community level to encourage this discussion about policy options? • How can the very detailed findings from the community level be used so they contribute in a meaningful way to the development of strategies and the choice of policy measures and public investments?
• What parts of the process have made a difference in terms of the level of influence the findings have had?
The report has been written by two people who were closely involved in the consultations from the very beginning. Although we have drawn considerably on the views and experiences of the other researchers involved in the consultations and have sought independent comment on the 1 The CPRGS -as it is referred to in Vietnam, is known as the Poverty Reduction Strategy Paper (PRSP) in most other countries. At the request of the Government, the Interim Poverty Reduction Strategy Paper (I-PRSP) was used as the basis for the local consultations as this was the official document in circulation at the time the research began. The I-PRSP was already quite comprehensive in it's policy proposals, and contained an analysis of poverty issues, which provided an adequate foundation for the consultations. Simultaneously, the Government was preparing the draft CPRGS, and the policy recommendations from the local consultations were subsequently made to this latter document.
draft, this paper primarily sets out our own reflections from within the whole process and cannot be considered an independent assessment of the value of these consultations.
We have structured the report as follows. The first section sets out the particular policy context in which these consultations took place and describes some of the previous work carried out to analyze poverty situations in different parts of the country. This reflects the authors' views that the success of the consultations can be attributed in part to work undertaken in earlier years, which was important in setting the groundwork for the consultations to be possible and influential. The second section covers the methodology in some depth and explains how the research team moved from the I-PRSP document to participatory research to analysis. The final section considers the way in which the findings have been used to influence policy direction and reflects on the main lessons and experience from the consultations.

PLANNING AND POLICYMAKING FOR POVERTY REDUCTION
This paper documents experience in trying to engage poor households and communities, as well as local government authorities, in the process of drafting the Government of Vietnam's CPRGS. Many people who have followed the formulation of the CPRGS in Vietnam believe that the community level consultations have influenced the content of the final strategy. We would argue that this is partly because earlier work had established a precedent with the Government of Vietnam for involving primary stakeholders through participatory research. It had also established certain collaborative ways of working between the Government of Vietnam, donors and NGOs that provided a mechanism to make such work more influential. Setting the CPRGS in the context of previous strategic work on poverty analysis and in the overall framework for policy-making is important in explaining the opportunities and constraints for undertaking this kind of policy-based, community consultation work.

Government's strategies and plans The Ten-year Socio-Economic Strategy and sectoral strategies
Policy formulation and public actions in Vietnam are guided by a range of strategies and plans. The period since 1999 has been a particularly active planning phase for the Government of Vietnam. During 2000 attention was focused on drafting the Socio-Economic Development Strategy (SEDS). This articulates the development vision for the coming decade and expresses a strong commitment to growth, poverty reduction and social equity. This document -prepared by the Communist Party of Vietnam and endorsed at the Ninth Party Congress in April 2001 -lays out a path of transition towards a "market economy with socialist orientation". It commits Vietnam to full openness to the global economy over the coming decade, and the creation of a level playing field between state and private sectors. It emphasizes that the transition should be "pro-poor", and notes that this will require heavier investment in rural and lagging regions, and a more gradual reform process than is often recommended by the international community. It gives strong emphasis to poverty reduction and social equity, and a more modern system of governance.
The specific actions needed to translate this vision into reality are described in the five-year plan for 2001-2005 and in a large number of sectoral ten-year strategies and five-year plans. These sectoral plans and strategies were prepared during 2000 and 2001 and most have now been approved by the Prime Minister. The strategies are often very broad and ambitious and detail a large number of targets and indicators. These targets are generally not prioritized or costed and tend to emphasize inputs and outputs rather than outcomes.
A central part of the task of formulating the I-PRSP and subsequent CPRGS was to prioritize across the wide range of goals and targets in the SEDS and the five-year plans to identify those that would most effectively capture progress in reducing poverty and promoting social equity.

The Government described the CPRGS as an "action plan that translates the Government's Ten Year Socio-Economic Development Strategy, Five Year Socio-Economic Development Plan as well as other sectoral development plans into concrete measures with well-defined roadmaps for implementation".
This involved the identification of a set of clear development goals that embody the main development challenges that lie ahead. The Government presents the linkages between the various planning documents diagrammatically in the CPRGS (Figure 1).

Figure 1. Strategic Planning Framework in Vietnam
The Poverty Task Force (PTF) and its activities This period of intensive national planning coincided with a more concerted effort on the part of the international community and the NGOs working in Vietnam to strengthen the poverty focus of national policy-making and planning processes. In 1999 the World Bank wrote to the Ministry of Planning and Investment (MPI) requesting that a group be set up to facilitate Governmentdonor-NGO collaboration on poverty issues. Initially this group -the Poverty Task Force (PTF) focused on poverty analysis, but it has since become the main mechanism for Governmentdonor-NGO interaction on strategic planning for poverty reduction.
The membership of the PTF has changed over time to reflect the workload. While the poverty analysis was being carried out, those agencies most involved with poverty diagnosis and with the greatest interest in using the results of the research were members of the PTF. As emphasis shifted to supporting the Government of Vietnam in the preparation of their I-PRSP and CPRGS membership of the PTF expanded to bring in the entire CPRGS Drafting Committee. During 2001, the PTF comprised 16 Government ministries, 6 donors (3 multilateral, 3 bilateral), 4 international NGOs and 4 local NGOs. The international NGO representatives are selected by the NGO community and also rotate regularly. Members are expected to provide active support in the form of resources or ideas.
Over recent years, the PTF has played a contributing role in introducing more debate and analysis to the process of planning and policy-making for poverty reduction. PTF members have funded and resourced analytical work that the Government could not otherwise have undertaken.

Comprehensive Poverty Reduction and Growth Strategy (CPRGS)
Annual plans

Public Investment Program
It has provided a forum where this analysis can be discussed with Government and other stakeholders and has helped to build consensus around the most pressing issues on the poverty reduction agenda. PTF members have also funded processes -consultations, workshops -that have enabled the debates on poverty reduction to be broadened, by bringing in both new topics and new voices. Figure 2 presents some of the key national planning exercises and PTF activities in the period since 1999.
As a body, the PTF has credibility with donor, Government and NGO agencies and has been able to ensure that the analysis it funds or oversees can be influential. Since 1999, two important pieces of participatory research have been coordinated from within the PTF. These are the Participatory Poverty Assessments (PPAs) in 1999 and these community level consultations on the contents of the I-PRSP (in 2001 to 2002). By design, these two exercises were closely linked and it is only possible to ascertain the influence of the consultations if we refer back to the PPAs.

Quantitative and qualitative poverty analysis in 1999
The original motivation for establishing the PTF sprang from a desire to foster more Government ownership of and more donor consensus around a high quality analysis of poverty. Because the Government was about to start drafting some of the key documents in the strategic planning framework, this was a particularly conducive period to be undertaking poverty analysis and building consensus around the findings. During 1997 and 1998 the General Statistics Office had been collecting household data under the Vietnam Living Standards Survey (VLSS) that would allow a robust assessment of trends in poverty and some social outcomes over the 1990s. In addition, four PPAs were carried out to capture the dimensions of poverty that were not well described by the quantitative data representing different regions of the country.
These PPAs were widely circulated and strongly influenced the contents of the final poverty assessment, which was a joint product of the PTF members 2 . Unusually for externally funded participatory research, much of this material has subsequently been used in policy documents or referred to in policy dialogue 3 . This is in contrast to previous attempts by many organizations, particularly international NGOs, to influence Government policy-making through micro-level participatory research. Such research had often been received with some skepticism.
Though it was not foreseen at the time, the PPAs helped to open the door for these community consultations on the I-PRSP. The consultations were held in six locations, including the four PPA sites. This has helped to establish a link between the participatory poverty analysis and the subsequent design of strategy. In these sites, the main thrust of questioning could be captured as: "Three years ago, this community told us x and y about the causes and dimensions of poverty. The Government of Vietnam proposes the following measures to address this problem. Here-in this community -will these measures provide a solution to the problem?" Looking forward, there is an opportunity for those supporting the Government of Vietnam in monitoring the implementation of the CPRGS to build further on these linkages.

Interim Poverty Reduction Strategy Paper to Comprehensive Poverty Reduction and Growth Strategy
In July 2000, the Government of Vietnam announced the intention to produce the CPRGS. Acknowledging that this exercise might take some time and require additional analytical work, stakeholders though the analysis of poverty drew partly on the findings of the PPAs. The I-PRSP marked a shift in approach to poverty reduction in Vietnam, expanding the agenda well beyond the more targeted approach outlined in the ten-year Strategy for Hunger Eradication and Poverty Reduction and beginning to draw the links between structural shifts in the economy and poverty reduction.
As soon as the I-PRSP was completed in March 2001, the Government began setting out the steps to develop the I-PRSP into a more comprehensive strategy document. The Government was keen to complete the CPRGS within a year and it was generally agreed that this would be possible because the I -PRSP was already quite advanced in terms of establishing a policy framework for growth and poverty reduction. MPI was assigned the task of drafting the CPRGS and MPI, in turn, established an inter-ministerial Drafting Committee to generate broad-based input from across Government into the drafting process (Box 1). Together with this, MPI organized consultations with a wide range of stakeholders including sub-national government authorities through regional workshops, and local communities and poor groups in these consultation exercises. The CPRGS was approved by the Prime Minister in May 2002 5 .
Box 1. Key actors involved in the work on the CPRGS • The Government of Vietnam, particularly MPI which led the work on developing a CPRGS together with members of the 16-agency CPRGS Drafting Committee (including line ministries such as the Ministries of Health; Education and Training; Transport; and, Agriculture and Rural Development); • Sub-national levels of government that interacted with the CPRGS Drafting Committee to respond to early drafts of the CPRGS and who participated in regional workshops; • The Poverty Task Force, which provides a mechanism for Government-donor-NGO collaboration on strategy formulation for poverty reduction. Other Government-donor-NGO partnership groups have also been involved in the work associated with the CPRGS in certain sectors; • The five agencies involved in the community level consultations and the Women's Union/National Council for the advancement of Women, who were involved in holding consultations with female National Assembly members and rural women's groups; and, • The Consultative Group (CG) Meeting, which over the last few years has allowed for formal dialogue between the Government of Vietnam and the donor community (including NGOs) on poverty diagnostics, strategy development and (in December 2001) target-setting for poverty reduction.

The community consultations: who was involved?
The Government of Vietnam made a commitment in the I -PRSP to ensure that primary stakeholders would be consulted as the CPRGS was developed. As MPI started drafting the document, they asked for assistance from the World Bank in coordinating work in this area. MPI was keen to involve the NGOs and eager to return to the PPA sites, but lacked the time and resources to conceptualize and manage the work. Having originally coordinated work on the PPAs, the World Bank had some experience in this sort of work and resources (financial and human) available to support.

Partner agencies, consultation sites and links to PPAs from 1999
Because MPI felt that it would be helpful to return to the PPA sites, the organizations involved in carrying out the PPAs were approached first. Save the Children UK (SCUK) agreed to return to the PPA sites in Ho Chi Minh City, Vietnam's largest city. ActionAid undertook to return to the typhoon-prone, coastal province Ha Tinh Province and Oxfam GB returned to Tra Vinh in the Mekong Delta. The Vietnam-Sweden Mountain Rural Development Program, which had been responsible for the PPA in the mountainous province of Lao Cai, was undergoing re-planning and was unable to take on the consultation. The World Bank managed the consultations in Lao Cai directly, using the same team and visiting the same sites as during the PPA. In addition to returning to sites included in the PPAs, two additional locations were selected. Plan in Vietnam managed consultations in the central coastal province of Quang Tri and Catholic Relief Services (CRS) conducted consultations in the Mekong Delta province of Vinh Long. The idea behind including these two additional sites was partly to capture a greater diversity of poverty situations (by including Quang Tri) and partly to try to throw more light on the poverty situation in the Mekong Delta (by including Vinh Long 6 ). The international NGOs working on these consultations all had long-term relationships with the authorities and communities in the areas where the research was carried out. Across the six sites, more than 80 researchers were drawn into the research and many were drawn from local communities or local Government offices and trained to undertake the research. Many of these researchers, and particularly the Vietnamese team leaders, had been directly involved in carrying out the PPAs and had a robust understanding of the local context. Their ability to draw a direct connection in the analysis between the descriptions of poverty emerging from the PPAs in 1999 and the proposed policy actions was a real strength of the exercise.
MPI stayed engaged throughout the process. They offered guidance at an early stage, indicating areas where they would find the consultations most helpful. MPI staff provided comments on the draft research framework, attended some parts of the fieldwork and the provincial review workshops that were used to validate the findings.

Motivation and reservations
The NGOs that collaborated in the consultations were motivated by a number of objectives. High on their agenda was the potential to influence national policymaking, both in content and in demonstrating new processes. The agencies previously involved in the PPAs were also keen to highlight the importance of including communities not just at the early, situation analysis stage of planning, but also uncovering the value of returning to those communities to discuss proposed actions to address the poverty they had described. Several of the agencies mentioned that the consultations have set the basis for more concerted work with provincial officials in localizing the approved CPRGS. Oxfam GB, for example, plans to work intensively with Tra Vinh Province in trying to operationalize the CPRGS -a strategy that determines priorities and outcomes at a national level -at the provincial and district level. SCUK has the intention to support similar initiatives in Ho Chi Minh City. A third motivation mentioned by the partner agencies was a desire to develop their own skills and those of their local Government partners in policy analysis work and monitoring and eva luation.
The World Bank had a number of motivations for putting resources into this work. First, there was a concern (also held by the IMF) that the process of developing the final CPRGS would not adequately reflect the views of all stakeholders in the absence of a specific exercise such as this.
Secondly, there was a interest in this work to inform the content of the Country Assistance Strategy (CAS) and the lending program that it sets out for the period to 2006. Thirdly, this exercise contributed to a broader World Bank objective of supporting the development of stronger and more evidence-based planning and policymaking processes in Vietnam. In this respect, the link to the PPAs (which the World Bank had funded and coordinated) was seen as very important.
Though retrospectively the NGO partner agencies all feel that the consultations have been an important and influential exercise, there were several reservations expressed at the outset 7 . There were some initial concerns about the level of Government ownership over a participatory process that is a World Bank and IMF requirement. There was uneasiness that they were being asked to carry out work that was legitimately the role of Government agencies. There were also worries that the World Bank would use the consultations to legitimize the CPRGS which could turn out to be strategy that some of the agencies might not want to be associated with (if the CPRGS emerged, for example, as a tokenistic product that would not turn out to be in the best interests of building strengthened national planning processes for poverty reduction). Early discussions with MPI were helpful in alleviating the former concerns. The concern over the role of the World Bank is no longer seen as an issue in the Vietnam context by the agencies who were partners in the research -even by NGOs that are choosing to disengage from these processes in other countries 8 .

Composition of research teams
The 80 researchers across the six sites came from a cross-section of organizational backgrounds. A large number were the Vietnamese staff of the INGOs, who had extensive experience in participatory research techniques, strong local knowledge of the consultation sites and good relations with the communities and authorities. A good number were also hired from academic or research institutes or were the staff of local NGOs. Several Vietnamese researchers were working as independent consultants. In all the sites the teams were supplemented with staff from local mass organizations (such as the Women's Union, the Farmer's Union or the Fatherland Front), from the District and Commune People's Committees, Village Heads and, in some sites, from the People's Councils or the local branch of the Communist Party. In many sites, members 7 The partner agencies were either interviewed or submitted written comments on their experiences of participating in the consultations. 8 As expressed in a note from ActionAid's senior management team.
of the community were also trained to carry out some of the research activities. Only a handful of foreigners were involved in the whole exercise and only one foreigner was involved in the fieldwork in any substantive way. For each of the consultations, the organizations managing the research put together a team of individuals that they felt could deliver this complex task.
Although staff from local non-governmental organizations took part in the research, no local NGOs were contracted to carry out the consultations from start to finish in any of the sites. This is a recognized weakness of the consultation that reflects a number of factors that combine to limit the role that local NGOs can play in this type of exercise. A paramount reason is that there are comparatively few independent civil society organizations in Vietnam and, constrained by the regulatory framework, the local NGO community is quite fragmented and dispersed. Advocacy work carried out by local NGOs tends to be more informal in nature and based around the strong links that the local organizations have with specific Government agencies or policymakers. Research-based policy analysis skills are also very limited among local NGOs and the few local NGOs that have staff with the research and analytical skills are extremely busy. In this respect, it is recognized that Vietnam is exceptional in the reliance on INGOs and consultants to undertake this type of work.
The hope is that there will be ways of building on the PPAs and the consultations as the CPRGS is implemented. This implies using the link that has already been established between participatory poverty analysis and participatory strategic planning in these communities to develop helpful processes of participatory monitoring and evaluation of progress on the ground. If local NGOs are going to be able to participate in this work in an influential way, then it is strategically important to start building greater capacity in the necessary skills now. Arguably, the space for credible organizations with good links with poor c ommunities to influence Government policy-making is opening more quickly than can currently be filled by organized civil society.

Timeframe and resources
The consultations were carried out to a tight deadline. The partner agencies made a commitment to participate in the research in September 2001. This was followed by a very intensive phase (completed in late November 2001) of developing the research framework and methodology based on the measures proposed in the I-PRSP. With some variation across the sites, training, fieldwork and analysis took approximately 6 weeks and about one third of this time was fieldwork. All the site reports were completed in Vietnamese and English (some in draft form) by mid-January 2002. A draft synthesis report was produced by early April, summarizing the findings from the six sites by policy area. Timing was crucial because the Government had set themselves a deadline of finalizing the CPRGS by May 2002. Though the teams felt rushed, there was an understanding that this information had to be available at the right time in order to make a difference to the strategy.
The funding arrangements varied from partner to partner. The work in Lao Cai was fully funded by the World Bank and in Vinh Long was largely supported by the World Bank. SCUK, Oxfam GB and Plan in Vietnam funded the work in their sites entirely from their own resources. Some of the additional costs associated with carrying out the work in Ha Tinh (by ActionAid) were covered by the World Bank. In addition to providing the direct financial support, the agencies involved dedicated large amounts of staff time to this exercise. With the consent of the partner agencies, the World Bank took on most of the tasks associated with coordination, pulling together the final research framework, organizing the dissemination of the site reports and synthesizing the findings from the various sites according to policy area. Three of the partner agencies (SCUK, Oxfam GB and ActionAid) have co-funded the publication of the site reports (Volume III of this series).

Why was the local consultation process influential?
In part four we set out examples of how the consultation exercise was associated with changes to the CPRGS document. The partner agencies expressed general satisfaction with the degree to which their findings had been used -not only in refining the CPRGS but also in other dimensions of policy-making and strategic decision-making. From the outset, two issues dominated those involved in leading and coordinating the research: • How to ensure that the fieldwork, findings and analysis were credible even though this was really the first work of its kind in Vietnam; and, • How to ensure that the findings were used by policy-makers to improve the content of the CPRGS.

Ensuring credibility
As the next two sections will show, methodologically this was a challenging exercise. Translating a broadly phrased I-PRSP document into a research framework and methodology that could elicit feedback from poor communities took two months of intensive work. There was real concern that the research should not leave itself exposed to criticism of poor quality fieldwork which could, in turn, undermine efforts to make the findings influential. Though the next chapters will go into some depth on how the research was conceptualized and evolved, it is worth mentioning some of the important elements of the process -rather than the methodologythat contributed to quality results.

Constancy of actors in the process.
The fact that many of the people involved in managing and leading the consultations had a previous history in implementing the PPAs was tremendously helpful. Many lessons were learned in carrying out the PPAs as to what worked well in the field and this very specific local expertise was very important in research tasks that required a high degree of flexibility on the part of researchers. Past experience with similar research exercises meant those involved in the consultations had good contacts in the (quite small) pool of Vietnamese researchers and were able to assemble strong research teams at short notices.
Building on institutional relationships, confidence and trust. Many of the local government counterparts working alongside the research teams had also been involved with the PPAs and had a clear understanding of what the consultation work might involve. Local officials, particularly those at the lower levels of Government (district and commune) were pleased to be able to participate, both as respondents in the research (focus groups were held with local officials) and as researchers in the communes and villages. Many officials operating at this level of the administration feel themselves quite alienated from decision-making processes. In five out of the six sites, local officials were supportive of the exercise and validated the findings in workshops which concluded the research in each site. In the remaining site, which had not been a PPA site, local officials were not obstructive, but were certainly less active partners in the research and were less at ease with some of the conclusions reached.
Consensus on the scope, nature and purpose of the task. Because development of the research framework was a complex task, the research agencies spent many hours together working through issues and problems. Though this required a serious commitment in terms of time, it was extremely valuable in building a shared understanding of what the research was seeking to achieve. By the time the teams went into the field to conduct the research, there was clarity on what kind of information was needed to provide feedback on specific policy measures. A reading of the six site reports demonstrates this unified approach: though there was no explicit structure proposed for the site reports beyond the need to address a range of important policy questions, all six were drafted with the same headings. This greatly facilitated the task of synthesizing the findings to address the chosen policy areas.

Ensuring the findings were influential
That the work was embedded in national policy-making processes was by its very nature a distinct advantage. The links to the key Government agencies were secured from the outset. The consultations were fully part of the workplan anticipated by the PTF that, as with the PPAs, acted as an informal governance structure for overseeing implementation of the research. This was clearly a strong starting point for an exercise oriented towards influencing the policy content of a Government strategy.
The role of partnerships. Other donors also played a valuable role in providing resources for the broader range of activities that helped to give the consultations profile and influence. Through the forum of the PTF, many donors were actively engaged in funding and organizing national and sub-national level consultations that took place at which the community consultations were presented and discussed. The support of the ADB, UNDP, GTZ and DFID was crucial in allowing these meetings to take place and their interest in helping the community consultations to be influential at these meetings was very valuable. Without the backing of the broader international community through the PTF, there is a risk that these community consultations could have been seen as a more marginal activity.
Influencing through multiple channels. Several steps were taken to enhance the influence of the findings. The main avenue was to publish and disseminate them as quickly as possible to the main audience -the CPRGS drafting committee. However, this was by no means the only way in which the findings were used. Nor, arguably, would this alone have been enough to have much impact on the final strategy. There were deliberate attempts to broaden the way in which the findings were used in the hope that this would build an acceptance of the findings and the main messages. Not all of these steps were planned at the outset and were only pursued actively as the research was already quite advanced. Earlier recognition of the multiple possibilities for using the findings could have, perhaps, influenced who was involved in the design stage. Some examples of the way in which the findings were used are as follows: • Informing a wide range of stakeholders that the consultations were happening even before work began to encourage organizations that were involved in supporting the Government in developing the CPRGS to take an interest in the work; • Presenting and disseminating the consultation findings early on in the agenda at regional consultation workshops (see below) so they could be referred to in the subsequent day-and-ahalf long discussions; • Drafting a paper which synthesized the findings by policy area so that policymakers could consider the messages thematically (mirroring the structure of the CPRGS and facilitating the process of comparing strategy against consultation findings); • Disseminating the consultation findings within line ministries, who were seeking to influence their own particular sectoral part of the CPRGS; • Encouraging the many sectoral Government/donor/NGO partnership groups to read and use the findings in their interactions with MPI on the content of the CPRGS; and • Encouraging those providing comments to MPI to use the consultation findings as evidence to support their recommendations.

THE LOCAL CONSULTATION APPROACH AND METHODS
This section provides a step-by-step account and reflective analysis of the way in which the local consultations were carried out. From inception to reporting on the results the research stretched over a period of seven months, divided into three main stages: • Preparation and grounding stage. This included developing a research framework and questions from the I -PRSP, developing the sampling approach and methodology, and selection of participants and preparation in the fieldwork locations. This took place over a period of approximately three months before the fieldwork began.
• Fieldwork facilitation stage. At each site the fieldwork included a series of focus group discussions and semi-structured interviews held with poor groups at community level and with groups of local officials, enterprise owners and other special interest groups. The fieldwork lasted 2 to 3 weeks in each site with the total fieldwork period covering 2 months for the 6 sites nationwide.
• Analysis and validation stage. This included compilation and analysis of the data to identify major policy messages and recommendations emerging from the consultations, and validation of the results and findings through feedback workshops with the local participants and at regional workshops. Around 2 months was required to compile and review the site reports and for preparation of a synthesis report from the 6 sites.
It is intended that this section may give useful practical experience to people in other countries who are planning to carry out similar consultations. To aid this, examples are given of the way in which three particular components of the I-PRSP were handled in the consultations (including basic social infrastructure, education, and local institutions). These examples are used to illustrate how the process was carried forward from the point of developing the research questions and methodology at the outset, through the fieldwork exercises, data compilation and analysis, leading on to identification of the main policy messages and r ecommendations. A similar process took place for all the 15 themes covered in the research. Box two outlines the broad research areas that emerged from this process. The full research framework is included in a fuller version of this paper, available on www.VDIC.org.vn or from the authors.

Deriving a Research Framework from the I-PRSP
The methodological challenge Policy documents are written primarily with government leaders, decision-makers and planners in mind and are not easy documents for non-professionals to digest. By their content and purpose 'comprehensive' strategic documents (such as the I-PRSP and CPRGS) assume an understanding of the multitude of sectoral and inter-sectoral concerns, interests and policies and how these fit together in the overall strategy, and prior awareness of the particular institutional context of their formulation. The question is: how does one make macro-level policy documents such as these accessible and understandable to local people, so that they can respond to them in a meaningful way?
This was the first major challenge faced by the researchers, which led to interesting areas of discussion not only with respect to how to use the I-PRSP in these local consultations -but also more broadly in terms of how does government communicate policies to the poor. It was recognized that a sizeable proportion of the community participants would not be fully literate 9 , and would certainly not be accustomed to being presented with and asked to comment on government policy in this way. In some fieldwork locations, such as in the mountainous province of Lao Cai, some ethnic minority participants (especially ethnic minority women) would not even be fully conversant in the national Vietnamese language. At the same time, it was likely that many of the local Government leaders and officials involved would not be fully familiar with the I-PRSP. It was therefore necessary to find ways of condensing and simplifying the policy document in a way that could be effectively communicated and used in the fieldwork while remaining true to its original content.
The second major methodological challenge was how to develop an approach that would yield comparable results and findings from the six fieldwork locations on the one hand, while also remaining flexible enough to pick up on locally specific concerns and responses on the other. It was intended that the research questions should be forward looking so that the focus would be on analyzing the relevance and potential for putting the policies into action, rather than providing a diagnosis of the current poverty situation. We were also concerned not only to get local people's responses to the policy actions already proposed by the Government in the I-PRSP, but also to elicit their ideas and recommendations on where there were gaps in the strategy, actions that needed strengthening and on measures necessary to ensure that commitments outlined in the CPRGS would be achieved in reality.
The first stage in preparing for the consultations involved reviewing the I -PRSP and to extrapolate from it a set of questions that could be put into a workable research framework 10 to guide the fieldwork. This demanded a considerable amount of time and discussion between the facilitation teams who met at regular intervals over a period of three months before the fieldwork began. Developing the research framework involved three key steps: • Defining the scope and content of the consultation; • Identifying the main policy actions proposed by Government for each component; and, • Developing a set of research questions related to each component Step one: Defining the scope and content of the consultation The first step was to decide which components of the I-PRSP it would be relevant to consult on at the local level. This was necessary because the scope of the I-PRSP very broad and it was clear that not everything could be covered. In consultation with MPI, it was jointly agreed to omit some of the policy measures associated with macroeconomic management and structural reforms 11 . Proposed reforms in the banking sector, for example, were not included in the research framework (although certain issues associated with access to financial services were covered). The trade reform agenda was not addressed explicitly in the research framework, partly because several of the agencies involved in the consultations (ActionAid, Oxfam GB and the World Bank) were also carrying out other research on the possible impacts of trade reform and links between globalization and poverty. Issues associated with increasing integration in global markets -such as exposure to fluctuating commodity prices and the legal framework regarding controls over exports for private companies -emerged as an important area of concern. The research did try to gather the perspectives of the poor on the proposed measures to promote the rapid growth of the private sector and on the reform of the state-owned enterprise sector. Issues associated with public expenditure management were integrated into many areas of questioning.
It was agreed to concentrate most of the research on components of the I -PRSP that local participants could directly relate and respond to including: the provision and administration of services under different sectors; measures to reduce vulnerability; local governance and local institutional aspects; and options for enterprise development and improving labor markets and employment opportunities. Accordingly, five main research areas and fifteen components were identified as related to the main sections of the I-PRSP (Box 2).

Box 2. Research areas and components of the I-PRSP used in the consultations (see also Annex 1: Research Framework)
• Research Theme 1: Trends in poverty and targets for poverty reduction 1. Current situation, recent poverty trends, and the overall Government strategy. 2. For urban areas: addressing urban poverty. 3. For rural areas: ways to stabilize and raise the living standards of ethnic minorities.
• Research Theme 2: Creating opportunities for poor households and supporting livelihoods 4. Improving basic social infrastructure. 5. Intensifying and diversifying agricultural production. 6. Measures to help the poor improve their participation in the market place. 7. Small and medium sized enterprise (SMEs) and household enterprise development. 8. Strengthening the ability of the poor, especially women, to access credit. 9. Improving vocational training opportunities and helping the poor to learn how to do business.
• Research Theme 3: Improving access to bas ic social services 10. Improving access to basic education. 11. Improving the health of poor people.

• Research Theme 5: Institutional issues and opportunities
14. Institutional aspects for implementation. 15. Local development plans and budgets.
Step two: identifying the main policy actions proposed for each component The second step in developing the research framework was to make a summary of both the general orientation and specific policy actions proposed by the Government under each of these components. In this section we give examples of how two components were developed (basic infrastructure and local institutions) to illustrate the process. Because the I-PRSP wo uld be new to a majority of participants, it was necessary to convey the general orientation as a means to introducing the purpose and importance of the I -PRSP before moving on to the detailed proposals in each thematic area. For some components this was relatively straightforward, since the I-PRSP already proposed a fairly clear and concrete set of actions. This was the case, for instance, with respect to improving basic social infrastructure (see Box 3). Here the Government already had a well-established and articulated strategy for concentrating public investment on developing basic infrastructure in poor and remote areas of the country to improve people's access to services and markets.

Orientation:
The Government is planning to develop infrastructure in poor and remote areas to improve access to services, markets and opportunities. The government would like households and local communities to be involved in the planning and management of infrastructure investments. It plans to achieve this through the following actions:

Policy Actions:
• Invest in developing various categories of social infrastructure, with special attention to rural, remote and isolated areas. Attach special importance to the development of roads to poor communes; • Encourage the poor to participate in building these basic infrastructure projects, considering that a means to create jobs for them and improve their incomes; • Combine the task of building infrastructure projects with preventive measures to combat floods and other natural disasters; • Create more opportunities for localities where poor communes and areas are seen to take initiative in managing the development and maintenance of their own rural transport infrastructure; and • Encourage people to take part in developing their own rural infrastructure, especially rural electricity, safe water, schools, health stations, commune cluster centres, markets, etc. and to operate and manage these works by themselves.
Other components of the I -PRSP were less clearly defined at the outset and so required a different approach to developing the guiding questions. This was particularly the case with respect to the institutional aspects (see Box 4) on which few specific details were given in the I-PRSP. In this case it was necessary to go through the I-PRSP and pull together quite scattered references to institutional and governance-related measures. In some instances where the I-PRSP was particula rly unclear or broadly-phrased it also required going back to other, supporting Government strategy documents (particularly some of the sectoral ten year strategies) to clarify the specific messages that the I-PRSP was seeking to deliver.

Orientation:
In the I-PRSP no separate section was devoted to the local institutional aspects of implementing the strategy, or to the role of local institutions, local organizations and associations. However, there is mention of the role of NGOs, Community Based Organizations and private organizations under various chapters. Relevant paragraphs referring to institutional issues include: Step three: developing a set of research questions related to each component This third stage involved formulating and agreeing on set of key questions to address under each component that would be used as a basis for the fieldwork. Obviously, with a limited fieldwork period it would not be possible to go into all aspects of the I-PRSP in equal depth or detail. Each of the facilitation teams and agencies had prior experience working in practical development settings and some had previously been involved in the PPAs. Both of these factors provided a basis on which to identify the most pressing issues and questions to address in these local consultations.
Continuing with the example of improving basic i nfrastructure from Box 3, for instance, new proposals in the I-PRSP included the intention of the Government to achieve greater involvement of households and commune authorities in planning and managing infrastructure investments, as well as to provide greater local employment opportunities through these schemes. It was decided therefore to concentrate the guiding questions on these particular issues of how to maximize wage labor opportunities for the poor, local participation in the planning and management of works, and how to develop affordable and effective operation and maintenance systems that do not place too heavy a burden on the poor (see Box 5).

Main Questions for Discussion with Participants in the Local Consultations
Objective: to find out how to maximize the opportunities for creating wage labor opportunities for local people in infrastructure development.
• Are people aware that it is the priority of the Government to create wage labor opportunities and are they informed about wage labor opportunities for commune works? • Are many people in the commune/village involved in building local infrastructure on a wage labor basis? • When contractors come to work in the area, do they usually employ local labor from the commune or do they usually employ labor from other places? • What specific problems do local people face in getting wage labor from contractors? What solutions do they propose to overcome these problems?
Objective: to find out how to develop affordable and effective operation and maintenance systems for local infrastructure that do not place too heavy burden on the poor • Are people consulted and/or involved in developing local level plans for infrastructure? What mechanisms are needed so that people can improve their input into the planning process? • Does the commune have special Boards established for management and maintenance of local infrastructure? • What specific problems or difficulties do communes/villages face in managing and maintaining infrastructure that is built under government programs)? • Are local people nowadays expected to contribute to maintenance of local infrastructure (in labor or cash contributions) than before? Does this place a heavy burden on some households?
Taking the local institutions component as an example again, the research team decided to attempt to gain more in-depth understanding of which organizations poor people think are important at the local level in delivering key components of the strategy (Box 6). A second line of questioning related to how well-informed local people are about local government plans and budgets and how they could have a greater and more effective input and participation in the future.

Box 6. Local Institutional Aspects: Main Questions for Discussion with Participants in the Local Consultations
Objective: to find out which organizations poor people think are important at the village/commune level in delivering key components of the GoVN strategy and establish how can they be more inclusive and accountable to the poor?
• Which local organizations can help you and your family have a better life? Are there any ways in which these organizations should change so that they serve the needs of your community more effectively? • How can you affect what your local organizations do in your community? Could the Government do something to make this easier for you? Have you got any ideas for anything else like this that should happen? • When you have to rely on people outside of your household for support (for example when you are sick and cannot work), which organizations help you? How can these organizations change so they can do a better job of helping your household when you face difficulties?
Objective: to find out how well informed local pe ople are about commune / ward development plans and budgets and to gather suggestions about how they could have a greater and more effective input.
• The Government has decided that plans and budgets at commune level should be posted for everyone to see. Is this happening in your village? If not, what should change in order to make it happen? If so, how was this information made available? Is it sufficient? What more information would you need to help the commune/ward authorities make plans and budgets in l ine with the needs of poor households? • When you have community meetings, what do you talk about mainly? Has your opinion been listened to? Did you receive a reply? How can people, especially women, be more involved in contributing ideas and comments for preparation of the plans? • The Government has suggested that local organizations should have better access to legal information, and that you should also have improved access to the law. If this is important to you, how should it be done? • What information should the Government gather in order to learn whether its services are being delivered effectively? Who should they ask?
In developing the fieldwork methodology from this research framework, each of the facilitation teams needed to refine these questions further to make them more appropriate to the situation of different participant groups. This was particularly important for the poor groups and community participants in order to develop a line of questioning and dialogue that would stimulate people's interest and understanding of the purpose of the consultations.

The participants and sampling approach
The local consultations took place in Ho Chi Minh City and 5 predominantly rural Provinces representing delta, coastal, midland and mountainous regions of the country (Table 1). This included 8 administrative Districts and 13 rural Communes or urban Wards. In total, around 1800 people participated in the consultations, of which approximately 40% were female. This included around 250 District and Commune / Ward officials and services providers, the remainder being local people, poor households, special interest groups, enterprise owners and household entrepreneurs. The consultations involved over 150 focus group discussions ranging between 5 and 20 people in size.

Triangulation of information and information sources
An important guiding princ iple behind the consultation methodology was to carefully triangulate the ideas and opinions (both within and between) four main stakeholder groups. These included: (i) District level government leaders, officials and service providers; (ii) rural Commune and urban Ward leaders and service providers; (iii) the private sector including SME owners and household entrepreneurs; and (iv) local people, poor households and special interest groups ( Figure 3).

Figure 3. Main Stakeholder Groups
An essentially similar set of topics and questions was covered with each of these participant groups based on the research framework, in order to compare and contrast their differing viewpoints on the proposed policies and actions set out in the I-PRSP. A short example from the consultation in Ha Tinh Province can be given at the outset to illustrate this principle:

PRIVATE SECTOR
Small and medium scale enterprises, household enterprises and entrepreneurs.
by the women's groups, which gave highest priority to upgrading the qualifications of the teachers.
The need for effective tria ngulation of data and data sources is a universally recognized element of participative research methods. However, it is often not brought to the surface as a primary facet in the analysis of results. In local consultations such as these, this was considered essential in order to systematically deal with the multitude of viewpoints and opinions given on such a broad and diverse set of Government policies. This was in order to identify where the main areas of consensus and difference of opinion lay.

Focus group sampling approach
There was a deliberate attempt to try to capture the diversity of viewpoints that could prevail at the community level and focus groups were structured to reflect this. In the 5 rural locations in which the consultation was undertaken the facilitation teams worked with a range of participant groups, as follows: • Women and men's groups (all sites) • Groups of poorest category households (all sites) • SME owners and household entrepreneurs (some sites) • Commune and Village leaders and service providers (all sites) • Old people and children's groups (some sites) • Poor landless villagers (some sites).
Sampling for participatory research in the rural areas of Vietnam is in some respects comparatively straightforward. This is because the Commune is a well-defined and wellestablished administrative unit, below which rural people generally have a strong sense of individual identity to their village or hamlet as a settlement and home area of the family. The lesser degree of deeply engrained social stratification found in rural Vietnam, as compared to some other countries, also makes it easier in a practical sense to bring villagers together in village meetings and focus group discussions for instance.
Ethnic composition requires attention in order to devise inclusive approaches to the study. Two of the rural locations contained a significant or majority population of ethnic minorities. The participants in Lao Cai included representatives from Hmong, Dao, Kinh, Tay, Phu La, Han, Nung and Giay ethnic groups, with ethnic minorities making up approximately 79% of the respondents. In Tra Vinh Province the poverty rate amongst the Khmer ethnic minority people is much higher than that of the Kinh majority in the district and communes where the consultation took place and special attention was given to incorporating their viewpoints.
Sampling in high-density urban areas, such as in Ho Chi Minh City, is more complex because of the over-lapping livelihood networks and 'senses of community' that are created by and enabled for different categories of long-term and short-term residents. The sampling approach used by SCUK in the local consultation in Ho Chi Minh City was of focus group discussions with participants divided by gender, age, and social status. In each urban Ward, eight focus group discussions were held, including seven groups of local people and one group of local officials. These included: • Unregistered poor female migrants • Unregistered poor male migrants • Local poor women • Local poor men • Young people (mixed male and female groups) aged 18-26 years • Children aged 12-16 years • SME owners and entrepreneurs • Ward and District officials.
Each of these urban groups was allocated 3 to 5 discussion themes from the research framework so that every theme could receive sufficient opinions from women, men, the poor, adults, young people, business people and officials at both ward and district levels. Ethnicity was also a concern in these urban districts since the proportion of ethnic Chinese in some selected wards ranged from 50% to 80% of the population.
In all urban and rural locations, focus group discussions were also held with District and Commune/Ward leaders and officials, representatives from concerned sector services and mass associations. These generally took the form of one-day workshops, in which the participants were divided into thematic or sectoral groups relating to different components of the Strategy. For instance, the District participants in Lao Cai Province were divided into three thematic discussion groups as follows: •

Factors influencing participation
A range of factors that influenced representation and participation in the consultation meetings. In some cases it was not possible to get adequate representation from the intended participant groups, while in others there was over representation of certain groups. In one of the consultation sites where only limited participatory research of this type had been done before, it was found that the inexperience of local government cadres in practicing participation had a detrimental impact on the fieldwork. This was manifest, for example, in red tape at local level which require people to obtain special permission to attend the meetings. In addition, entrenched cultural norms regarding gender roles in the village limit the potential for women to participate in informational and decision-making meetings at village level, which led in some cases to a low level of confidence in the focus group discussions for this research. As one woman who did participate here commented: "This is the first time I have attended a meeting with so many new things. Meetings in the local area are usually not informational at all."

Preparation in the fieldwork locations
Thorough and timely preparation is critical in order to make sure the fieldwork runs smoothly. In particular, it is necessary to ground the consultations carefully with local Government leaders and officials so they fully understand the purpose, content and approach that will be used and what local resources will be required. Based on our experience, the steps involved in preparing for the fieldwork should ideally include the following.

Preparatory visits to the fieldwork locations
These should be made at least 2 weeks in advance of the main fieldwork period. Visits should be made not only to the province or district capital, but also to the actual fieldwork sites and communities. The purpose of these visits is to brief local leaders at each level on the purpose of the consultations and to make logistical arrangements such as identifying the personnel who will be required to help organize and support the fieldwork. At this stage it is also essential to agree on the schedule and intended participants.

Distribute the official policy document in advance
It was found that local government officials at all levels and in all sectors were not previously well informed about the I -PRSP. As part of the consultation exercise the facilitation teams therefore had to explain the process involved in developing the I-PRSP and its contents. All teams underestimated the time required for this.
Distributing the official I-PRSP document -as printed and published by the Government under official seal -gave legitimacy and authority to the consultation exercise. In some locations this was distributed together with the research framework to District and Commune staff so they could read these documents in advance. In situations where this material was not distributed beforehand, it proved a more difficult and lengthy process to introduce the I -PRSP at the beginning of each session in the fieldwork. For instance, in Ho Chi Minh City, some District officials questioned the research team why they had not sent the questions beforehand so that they could prepare and speak effectively, although this complaint was not heard from any of the community groups.

Conduct group training / orientation sessions
In each location, key district staff and officials, commune/ward and in some cases village leaders were brought together with the facilitation team members for a 1 or 2 day training session immediately prior to the fieldwork. The purpose of this was to run through the methodology and schedule of the consultation exercise. Here again, such group orientation sessions with the involvement of local leaders and officials added legitimacy to the exercise.

Carefully explain the purpose, approach and methodology to local officials
At all stages, it was necessary to be aware of the potential sensitivity and novelty of carrying out this type of consultative study. Asking local people to directly comment on and be critical about government policies can easily be perceived as a threat by local leaders and officials. Similarly, some local Government staff initially expressed skepticism about the ability of the poor to discuss and contribute to policies.
Careful explanation of the purpose and approach of the consultations is therefore required in order to allay these potential fears, to get the full support and backing of local leaders, to ensure the meetings are open for people to attend and are inclusive of different interest groups. All the teams were involved in raising awareness to emphasize that if correctly handled the consultations could yield meaningful results. It was found that when the purpose of the consultations was well understood, local officials were generally appreciative of the approach. As expressed by one commune official: "Before there were some consultations with our commune staff. But

The consultation techniques
This section describes and comments on the use (and limitations) of the main techniques used in the consultations. As indicated above, the research questions were designed to be forward looking so the focus would be on discussing the relevance of the proposed policies, and factors that would influence and improve their application. The methodology also had to yield comparable results and findings between these participant groups and study locations, while also picking up on locally specific issues and concerns.
Accordingly, the same topics and questions were covered in all 6 locations and with the different participant groups (but with urban and ethnic minority issues only covered in relevant areas). An essentially similar range of participatory techniques were used by the facilitation teams, although it was left to the individual teams to decide which techniques were most appropriate for which sets of questions and participant groups. These techniques included: • Focus group discussions and semi-structured group interviews; • Sorting and ranking on cards and flipcharts; and • Household / individual interviews and case studies.

Focus group discussions and semi-structured group interviews
Focus group discussions and semi-structured group interviews formed the core of the consultations in all locations. These were organized in different ways, including men's and women's groups and special interest groups at community level and sectoral or thematic groups with local officials and service providers. An example of how the village level discussion groups were organized over a two-day fieldwork period in one rural location, and the topics covered by each group, is given in Table 2. The focus group discussions generally began by introducing the overall objectives of the exercise, while seeking questions and clarifications from the participants about this. Following this the particular component(s) of the I-PRSP for discussion were introduced, including the general orientation and main policy actions proposed by the Government. This was used as the platform for the discussions.
In guiding the focus groups, experienced and strong facilitation is required in order to ensure the discussions result in a clear set of policy related outputs and recommendatio ns. For this the facilitators need to adopt both a general line and a specific line of questioning. On the general level, it is necessary to follow a clear sequence in order to ensure the discussion remains on track in the limited time available, as follows: • Are the proposed policy actions relevant to needs and conditions of the group?
• Do people agree with them? (If yes, why? If not, why not?) • Are the policy actions realistic and achievable or not?
• If the policy actions are to be implemented in the area how can they be achieved?
• From the proposed policy actions, which are most important? (Priority ranking) • From the proposed policy actions, what is missing?
• Do the participants have other priorities?
• Would they recommend adding anything to the policies?
The specific line of questioning concerns the particular issues and questions relating to each component of the I -PRSP. For instance, the questions used to guide the community level discussion groups in Vinh Long Province on the topic of how to improve poor people's access to education were as follows: • What groups of children in the commune are illiterate or not able to go to school? • What are the reasons for this? • What could be done to get them to school? • What specific problems do girls have in completing primary school and what can be done to address these problems? • What difficulties might the Government face in trying to teach the goal of universalizing primary and lower secondary education? • What changes would be necessary to make sure Government achieves the goal of all children completing primary school in the commune? (Probe about access and quality) • What are the problems with implementing the Government policy to exempt the poor from school fees and to lend money and books to students? • What can be done to resolve these problems? • How do you know whether education/teaching is of high quality or not? What information do you rely on?

Sorting and ranking on cards and flipcharts
The focus group discussions were combined with more structured tools and exercises in order to obtain concrete responses from the participants on the proposed policy actions of the Government. This was done mainly through sorting and ranking of people's ideas and opinions on cards and flipcharts. This particular technique was used in a number of ways in order to: (i) capture the range of responses with respect to the perceived relevance and applicability of the policy actions ; (ii) to rank and/or score the priority given to different policy actions; (iii) to assess the potential bene fits of the policy actions to different poor groups from an equity perspective; and (iv) to collect and prioritize people's ideas on policy actions that need to be added and recommendations regarding implementation of the policies.
One way of doing this is to use written responses on colored cards. After introducing the actions proposed by the Government under a particular component, the participants were asked to write down their recommendations on the cards that were then sorted into groups. Alternatively, the cards could be used to rank the policy actions (for instance, using a scoring range of 1 to 3 denoting 'highly important policy action', 'medium importance', 'less / not important'). This was done individually, or in groups of two or three which may facilitate discussion and make the exercise both easier and more interesting.
There are a number of advantages in using the card technique. Firstly, it is a way of systematically capturing and dealing with a diversity of ideas and opinions from a large number of participants. It gives everybody a more equal chance to contribute, as well as to put forward ideas and opinions they may not be willing to share in public. As succinctly stated by the Chairman of the Fatherland Front in one commune in Lao Cai Province: In Ho Chi Minh City it was found the card technique was a particularly useful way of opening up discussion and stimulating ideas on the Government policies from children and young people's groups (see Box 7).

Box 7. Making policies interesting for children and youth in Ho Chi Minh City
With the children and young people's groups, it was both necessary and easy to apply more participatory methods. Some games were played with the children at the start of the group meetings, such as singing and 'snowball' speaking -meaning that the participants took turns in telling observations or their understanding of a poverty-related issue. Children were especially excited with the card writing exercise, and in some cases they were asked to work in pairs, as this helped a lot in terms of participation for those whose education is low and so felt shy to work alone. Other similar techniques were also applied such as ranking of the priority options. When the facilitator collected all the ideas she would ask one or two children in the group to join her reading out the results. This really made the little participants concentrate as their ideas were brought up to the whole group.
There are, however, drawbacks with the card technique. It assumes the participants have good prior understanding of the policies under discussion and so are in a position to give concrete ideas and suggestions. In some locations, it was found difficult to use this technique 'up-front' at the beginning of the session. Rather, it was more useful as a way of summarizing the results of a more open-ended introduction to and discussion on the policy actions.
The written card technique also assumes a degree of literacy and that the participants are confident in writing down their ideas and so is most useful with groups of local officials and agency staff (see footnote 15 on literacy rates). The generally high levels of adult literacy in many parts of Vietnam also makes this technique easier to use here than in some other contexts.
In the discussion groups held at community level and with poor groups, it was found that oral presentation of people's ideas and opinions was most appropriate. These were listed and sorted by the facilitators on flipchart paper and then ranked by the group members according to priority. As part of the data analysis, the responses and ranking given on these cards or flipcharts can be compiled at different levels of aggregation (focus group, participant group, location and area) to obtain an overall assessment of the main policy messages and recommendations.

Individual and household interviews and case studies
In some locations interviews were also held with individuals and households to complement the focus group discussions and semi-structured group interviews. These were an additional means of eliciting the ideas of special interest groups and of poor people who were less confident in contributing to the group discussions, and allowed for a further degree of triangulation between participants groups. The value and interpretation of these interviews and case studies in contributing to the policy analysis is discussed further in Section 3.5.

Interpreting differing viewpoints and conflicts that may arise
In carrying out this type of research, it is essential for the facilitators to quickly pick up on the level of cohesion or dissention in the particular local community or participant group. This is in order to be aware of the extent to which local political factors may influence people's responses to the policy proposals under discussion. This is critically important both in the facilitation of the fieldwork and in the analysis and interpretation of results. For this the facilitators require an unusually high degree of understanding of the local situation and the nua nces of what lies behind the viewpoints of different participants.
It is also necessary for the facilitators to anticipate the local conflicts that may arise during the course of the discussions. If this occurs, then it is necessary to quickly understand and respond to the situation in an appropriate way to guide the discussion towards a successful outcome. For example, when the topic of improving health care services for the poor was introduced in a men's group discussion in one village in Lao Cai Province, it quickly became apparent that there was a considerable amount of dissatisfaction in the village about the management of the Commune Health Center. People had lost confidence in the commune health staff as there were instances of health staff i nforming patients that the clinic had run out of medicines, but these medicines were available for sale at their home although with the service and medicines sold informally at a higher rate.
The facilitators can do several things in this type of situation. First is to remind the participants that the purpose of the consultations is not to solve local conflicts. At the same time, it is necessary to recognize that the conflict exists and to propose ways of dealing with it in the meeting. This can be done by asking the participants to propose 'positive' solutions in the form of suggestions for improving the Government's policies. In this particular case, the men's group in Lao Cai came up with a very useful set of recommendations, including: • Increasing the wages of commune health staff; • Upgrading clinics for better accommodation for health staff and patients; • Making sure that there is at least one health worker in each hamlet; • Ensuring a regular and sufficient medicine supply; • Prohibiting the free sale of medicines at markets by persons lacking expertise; • Banning medical workers from selling medicines at home; and • Establishing official medicine stalls at local markets managed by district authorities.
From a policy perspective, it is necessary for the facilitators to ascertain to what extent this type of conflict is a one-off situation, or whether it is indicative of a wider issue and/or lack of clarity in the policy and regulatory framework; in this case, for instance, regarding the regulatory framework for the sale of medicines and administration of the commune health services.

Note taking in the focus group discussions
Even though an attempt was made to record people's responses systematically through the card technique or on flipcharts, it was inevitable that many ideas and lines of discussion would not be adequately captured through these means. It was found therefore that good note taking by one of the facilitators was an essential source of back-up information for the subsequent analysis of the results. In some locations this was done better than in others. One important lesson, however, is that insufficient attention was given in the preliminary training sessions to note taking skills and what the facilitator should concentrate on in making these notes. One important function of the notes should be to capture the main areas and points of consensus and discrepancy between the participants, as well as to record of particularly interesting or novel ideas put forward by the participants.

Dealing with expectations
In any type of participatory assessment it is essential, at the outset, to articulate clearly to the participants what the expected outcomes and follow-up of the assessment will be. This is in order to avoid raising people's expectations (for instance with regard to possible follow up support) as well as to give people an idea of why they should spend their valuable time on the discussions. In this type of consultation on Government policies, it is especially important to explain the purpose of the study to local leaders and officials. Even so, in some locations, it was found that difficulties arose because local staff and participants began discussing project related activities and inputs. In this type of situation it is important for the facilitators to re-emphasize that the consultations are not intended to be directly related to project / program activities, but are a way of informing Government policies that will in turn influence the way in which projects and programs are carried out in the future. This may seem an obvious point to make, but in practice, it is essential to have clarity on this with the participants so the discussions are focused on the task in hand.

Aggregating and analyzing the data and results
The local consultations yielded large data sets of numerous viewpoints and ideas, priorities and recommendations put forward by the local participants. The main steps involved in processing this information included: (i) compiling and aggregating the data from the focus group meetings and participant groups in each location; (ii) comparing the results within and between these participant groups in order to identify the most important policy messages and recommendations; and (iii) preparation of draft summary reports from the individual study locations. This section describes how this was done using specific examples to highlight various aspects in the interpretation of the results.

Compilation and aggregation of the data
The basis of the analysis was to aggregate the frequency and/or ranking given to particular responses or recommendations as made by the participant groups in each locality. The degree to which this could be done systematically depended on the way in which the information was recorded in the field. An example of the type of output emerging from the community level groups is given in Table 3 from Lao Cai Province. This is from one Womens' group discussion on the topic of improving education. While this group recognizes that improvements have been made in the provision of primary schooling in recent years, they recommend the need to reduce the side costs of attending school and to improve material conditions in the village schools. In addition, this women's group identifies the need to strengthen the Government policy with respect to: (i) improving access to secondary schooling for poor ethnic minority pupils, (ii) the provision of vocational and job training for school leavers, and (iii) improving education opportunities for adult and illiterate women. Compilation of this type of information rests on assessing the extent to which similar viewpoints and recommendations emerge from other participant groups.
A more structured approach to data aggregation can be made with the outputs from the discussion groups in which the card technique was used. An example of this is given in Table 4, showing the aggregated responses to education questions from one commune level discussions group (this example is also from Lao Cai from the same commune as the village example given in Table 3 above).

Response Frequency
Parents are not aware of the importance of children's schooling and think they cannot be accepted to state-owned companies nor get employment after leaving school 9 Many households are too poor to afford their children's education. Poor children still have to pay tuition fees and pay for textbooks and textbooks are not cheap enough for the local people to buy 9 It takes time and effort to travel to schools, especially in flood seasons 5 Some of the children have to do the housework like babysitting, buffalo-feeding 3 The family is so big that parents can not afford their children's schooling 2 The children's rights to education have not been of proper concern from different branches and agencies 2 Most illiterate children are girls 2 Question 3a/2: What should the Government change to reach the objective of having all children going to school?

Response Frequency
Fees and textbooks given free of charge to children from poor households from kindergarten up to high schools. Reduce all the contributions to schools.

10
Branches and agencies should encourage parents to let children go to school 3 Put the laws on primary school popularization into effect 3 Teachers from the locality should be trained and teachers from other areas should not be employed like now.

1
The Government should build more boarding schools for children in upland areas 1 Provide electricity to the villages 1 Give insurance assistance to pupils of primary schools 1

Identifying the main areas of commonality, consensus and discrepancy
Based on this type of aggregation it is possible to begin analyzing patterns in the data according to the frequency with which certain responses occur. This is where triangulation of the responses and recommendations made by different participant groups becomes important. The purpose of this is to identify areas of commonality in which there is a high degree of consensus and also areas in which there are major differences of opinion between one or more groups. An example of this is given in Box 8 regarding the measures proposed by different groups on how to improve the support given to commercial shrimp production in Tra Vinh Province in the Mekong Delta.

Box 8. Policy recommendations on improving assistance to shrimp farmers in Tra Vinh
Shrimp production has been expanding in Duyen Hai District for nearly ten years with strong support from the Government authorities. However, sustainable methods of shrimp farming that are suited to local environmental and socio-economic conditions have not been fully developed. Shrimps have made many households rich, but have also made some households bankrupt. Some poor households say they are continuing to save money, to obtain loans and to rent land for raising shrimps with the dream of having a new life -as one successful harvest can make up for two lost harvests and even make some profit. The policy measures proposed by different participant groups on how to improve the assistance given to shrimp farmers share common points, but differs in the level of priority attached to different measures: Local authorities -this group recommends that the government invest more in the canal and dyke system for appropriate planning of shrimp-raising areas and mitigating environmental pollution; provide support for local production of good breeding stock; assist in training technicians (in the form of establishing an aquaculture skills training school) in the locality.
Better-off farmers -this group recommends that the government have preferential measures and that banks provide larger loans with longer terms (3-5 years) for semi-industrial and industrial shrimp raising; that the government assists in finding markets to ensure an appropriate selling price for shrimp raisers; and to carry out effective quarantining of shrimp breed, feed, and disease prevention.
Poor farmers -this group recommends that the government postpone payment of previous debts and provides new loans in a timely way for raising shrimps; providing technical guidance in shrimp raising; carrying out effective quarantining of shrimp breed, feed and disease prevention.
Women's group -this group recognizes that shrimp raising has high risks and can have strong negative effects on their livelihoods ("We are too afraid of shrimp"); and they want to get loans through the mass associations (Women's Union) for diversifying income sources (raising other livestock, growing cash crops, doing supplementary work).
As noted in this example, the policy measures proposed by different participants for effective development of shrimp farming share many common points, including: the need to improve the quality of the shrimp breed; the selling price; capital sources and the timing of capital inputs; disease prevention and treatment techniques; environmental management (for instance, water flow regulation); and feed availability and cost. At the same time, there are differences in the level of priority attached to each measure.
Iterative analysis of the data and in-depth knowledge of the local situation is required to interpret and assess the relevance and policy implications of this type of information. Differences of opinion within and between participant groups may only be indicative of local conflicts arising over resource management or the provision of services. Alternatively, they may be more broadly indicative of key areas of the policy and regulatory framework that need to be improved.
This was found to be the case, for instance, with respect to the prohibitive side costs of education that emerged as a major concern in the consultations. L ocal government officials frequently stated that primary school tuition is free of charge and there are exemption policies already in place to assist poor households in sending their children to school (for instance, text book cost exemptions). From the perspective of many poor households, it was found that the school tuition fees themselves are not the main problem -it is the many contributions and supplementary costs which make the financial burden unmanageable. Similar statements about the need to reduce the financial burden of educating young children were repeated in every site, demonstrating that this problem is not simply a localized issue.
• The Ho Chi Minh City report states that: "the official school fee is just token, but other costs and contributions are considerable, such as uniforms, textbooks, facilities, computer costs, extra classes, and even unreasonable e.g. contribution to flood victims. Required contributions can hurt and humiliate poor people. One woman reported that she was shouted at by the headmaster when she submitted a request for reduction of some contributions for her child".
• In Lao Cai, one women's group gives an account of some of the additional costs associated with education: "(a) for kindergarten -tuition fees, payment for toys, and cost of a meal if this is not included in the tuition fee; (b) for primary school -contribution for school building, text books, uniforms, photographs for pupil's cards; (c) for junior secondary school -contribution for school building, guards, bicycle park, school yard, and tuition fees." • In Vinh Long, one mother said that because she was not able to find work, she could not pay the contributions: "My children cried and did not dare to go to school" . Another mother reported that: "the school did not allow my son to go to school or take examination unless I made the payment" , which she could not so he dropped out. Commune officials appeared reluctant to admit the existence of the school maintenance fee burden, one saying only that "sometimes, people have to contribute to school maintenance" , and another that "local people don't have to make contributions to maintenance of the schools" .

Identifying the main policy messages and recommendations
Through this type of iterative analysis an assessment can be made of the overall priority attached to the proposed policy actions in order to identify the main policy messages and recommendations emerging from the consultations. An example of the priority given to the proposed policy actions on developing social infrastructure from Vinh Long Province is given in Table 5 (see also Box 1 & Box 4 for the original policy actions and research questions). This priority ranking also includes policy actions proposed by the local participants but not included in the draft poverty reduction strategy. Reducing the burden of paying for infrastructure maintenance (RES).
Construction of priority projects, including roads (GOV/RES).
Ensuring high quality of infrastructure construction works (RES).
Increasing local involvement in planning, management and maintenance of local infrastructure (GOV).
Improving the quality of operation and maintenance (RES).
Increasing the potential for income generation through employment on construction projects (GOV).
Combining infrastructure provision with natural disaster prevention measures (GOV).
Reorganizing the population into cluster communities to facilitate the development of social infrastructure (GOV).
GOV: Policy actions proposed by the Government in the I-PRSP RES: Additional policy actions proposed by the participants.
This same principle of assessing the degree of commonality and priority attached to different areas of the policy was used in the overall synthesis of results from the six consultation sites. An example of this higher order of aggregating the results and findings is given in Table 6, summarizing the main policy messages regarding improvement of basic social infrastructure in poor areas. Here again, the synthesis and interpretation of information at this level requires good understanding of both the national context and regional differences, and referral back to the understanding of the fie ldwork teams, in order to draw out the most important policy messages. The majority of participants agree that investment in infrastructure in poor areas should be given high priority in the I-PRSP.

• • • • • •
Information on local infrastructure works, including budgets and expenditure, is not easily available to local people.

• • • •
Local people are keen to have more paid labor opportunities and stress priority should be given to local labor force.

• • • •
Mechanisms are needed to ensure that contractors fulfill the Government's intention of providing more local labor opportunities.

• • • •
More attention should be given to the quality of construction and should be implemented by local contractors.

• • • •
Supervision boards have been established in the study locations but it appears that their authority and specific responsibilities are not clear.

• • •
Local people think that if they have a greater role in providing labor and supervision the quality of construction will be ensured.

• • •
People are worried that their contributions to infrastructure mainly in terms of labor or in kind payments are rather high.

• • • • •
It is recommended that the strategy should be more detailed with respect to the mechanisms for implementation.

Identifying important policy linkages
Consulting on the individual policy components (health, education, employment, agriculture and so on) does not fully expose the way in which different sets of policy actions need to come together to create an effective poverty reduction strategy. In this respect the individual / household interviews provided an invaluable source of additional information and understanding. The interviews can be used to reveal the interdependency of factors and forces that contribute to individual / household poverty situations -and so the way in which policy actions to need to be integrated and harmonized. The key to interpreting such interview material from the policy perspective lies in identifying critical policy linkages that are likely to yield the greatest benefits to poor people.
In Ho Chi Minh City, for instance, the interviews were used to understand and illustrate changes in urban livelihoods under the impact of economic growth and poverty reduction policies that have been implemented so far. These case studies covered a range of situations including both successful and failing business enterprises, poor households getting out of poverty and better-off households falling into poverty due to negative i mpacts of the national or regional economic situation, and the stories of unregistered male and female migrants, young job-seekers and working children. The example given in Box 9 is of a young man seeking stable employment in Ho Chi Minh City.
An important proposal contained in the I -PRSP is to help create maximum employment opportunities for the poor by various means including development of the private sector through small and medium enterprises. The proposal in the I-PRSP is that the urban poor will be assisted in accessing the newly created jobs through establishment of Employment Service Centers. This example shows that while these Employment Centers have an important function, for the poor they can present insecurity due to the introduction of fees that are required to access the service, while obtaining a job is not guaranteed. The Ho Chi Minh City report concludes that dispensation should be made to help the poor access these advisory services and that, even then, this measure will only be useful t o the poor as part of a broader range of measures to create access to better employment opportunities.

Box 9. Employment Service Centers are not the only solution to urban unemployment
Tuan is 20 years old. He grew up as one of 5 children in a poor family in Ward 12 in Ho Chi Minh City. His father works as a motorcycle taxi driver and his mother sells groceries. Due to his family's difficult circumstances, Tuan dropped out from school in Grade 8 and worked to help take care of his younger brothers and sisters. His first job was to sell mooncakes in Ba Chieu Market. This job was only seasonal so the income is not stable. After that, he worked as a shop assistant in the market. He worked full-time so his income was a little better. He thought he had found a suitable stable job, but after three months the boss fired him due to his health problems. Finding it too difficult to apply for a job, Tuan began to work as a street vendor in the Eastbound Coach Station where he had many difficulties at first. But when he had just got used to the work, he had to return home to help take care of his ill mother. It has been over one year since then. Tuan has applied for a job in many places but has failed to get one due to his poor health and lack of skills. He has been to the Employment Service Centre several times. He had to pay 40,000 VND for registration each time, but no job has been found for him so far. The local authorities do not cause him any problems when certifying his papers and the Employment Service Centers welcome him because he brings them money. He has joined the activities of the Youth Union but has no information on the program supporting young people in choosing a career. Tuan has heard a lot about labor export, but he thinks that this is very difficult because it requires skills and mortgage and he has taken no vocational training courses. In order to help young people in getting jobs, Tuan thinks that the government should restrict immigration. At present migrants are preferred because they accept lower wages. The government should have institutions offering free vocational training to young people and young people should be able to receive loans for doing business. Tuan himself is going to a supplementary high school to earn a diploma, which he hopes will help him find a job.
In Quang Tri Province, household interviews were undertaken that reveal the range of strategies employed by poor rural households to diversify their income sources in both the short and longterm perspective. The example given in Box 10 is of one household that is engaged in production of a major cash crop (pepper) that is promoted by the Government extension service in this locality, as well as in sideline marketing of small produce and occasional off-farm labor. At the same time, the 'long term plan' and economic prospects of the household are dependent on the parents' strategy of securing children's education and training.

Box 10. Household diversification strategies and long term plans in Quang Tri
Trung Thi Ba and her husband frequently discuss whether to grow more pepper trees or to try planting other crops. The price of pepper drops every year and they have begun growing lemon and grapefruit trees, in addition to owning two fishponds. For extra income, both of them work as occasional laborers and Ba sells fruit during the rainy season. As Ba says "It's all part of our 50 year plan so we can invest in our kids' education." They have four children; the eldest has graduated from high school and works to help his parents. The second eldest is attending vocational training outside their village. Despite leaving the area for a few years, they returned to Tan Xuan Hai village and are determined to improve their living standards and keep their children in school.
One of the ma jor policy recommendations to emerge from these local consultations is the need to ensure that enough flexibility exists in rural development interventions at the local level to balance building economies of scale (i.e. diversification on a regional scale to boost cash crop production for regional and export markets) and promoting local income diversification (i.e. to reduce poor farm household vulnerability). This example from Quang Tri illustrates how diversification of the rural economy needs to be understood and promoted on different levels simultaneously. The economic strategy of many poor rural households is clearly towards maximizing diversification according to local market, employment (and education) opportunities -to spread risk and to maximize different income sources both now and in the future. The diversification strategy of the Government, on the other hand, has conventionally been towards developing concentration areas of cash and commodity crop production for national and export markets. This can be a highly risky venture for poor households in the event of price fluctuations.
As both these examples -from rural Quang Tri and urban Ho Chi Minh City -illustrate, poor people give high priority to obtaining new skills and education. It was found that there is a huge expressed demand amongst poor people in all six consultation sites for more and better vocational and skills training opportunities. This demand appears to be indicative of a growing awareness amongst local people that education and skills are essential for making one's way in the modern world. And the linkages between improving education and vocational training opportunities and other components of the poverty reduction strategy emerged as one of the most important areas of policy l inkage. As one villager from the upland area of Lao Cai Province commented: "only a few people in the upland areas finish primary school and it seems impossible to set up many new businesses due to the shortage of trained workers". And in Ho Chi Minh City 'having no skills' and 'poor education' are commonly cited by the poor and others as the biggest barrier for the poor in accessing employment and developing businesses.

Validating the Results and Findings
However carefully the research framework, questions and exercises are developed, the fieldwork and, particularly, the analysis of findings are processes fraught with biases. The reality of trying to organize community responses and use them to make a meaningful contribution to policy formulation -as distinct from using them descriptively in a poverty assessment -often requires some extrapolation and interpretation of original findings. The following questions arose in connection with discussions about measures to deal with movements in agricultural commodity prices, but similar dilemmas were a feature in several parts of the research.
• Do repeated statements by poor households that they need subsidized agricultural inputs or guaranteed selling prices really reflect the best approach to dealing with a problem of fluctuating agricultural commodity prices?
• Or are respondents proposing this because this is a way that agricultural extension services have behaved in the past and this is what lies within their frame of reference as being possible?
• Is it also possible that other policy solutions might be more appropriate and sustainable (a focus on diversification of livelihoods rather than concentration on particular cash crops, and availability of a more varied range of financial services, including facilities for cash savings), but do not come to mind so easily because they are outside familiar territory?
• How should researchers and analysts deal with proposals from the community that either are known to have failed badly in other parts of the country, are clearly unsustainable, or that are likely to be unaffordable for Government?
Policy solutions identified by poor households often differ from the solutions that the researchers themselves would develop to address the same problems and this was particularly true of solutions that involved input subsidies. Unlike the PPAs, where findings were intended to be descriptive of the causes and dimensions of poverty, collaborative researcher-participant policy analysis requires more probing and results are strongly guided by the direction of this probing. In these circumstances, it is extremely important to verify the findings and the main conclusions with participants and other stakeholders to ensure that the analysis has not somehow drifted away from what people were trying to say.

Feedback workshops with participants and local authorities
Feedback sessions are important in all forms of participatory research so that conclusions can be checked with those who have contributed to the research. These sessions are especially valuable when the post-fieldwork analysis requires condensing the very detailed findings into suggested policy responses. In all the sites, feedback sessions were held at the community level to verify the findings. In addition, workshops were held for l ocal Government officials (at district and provincial levels) to ensure that the findings were rooted in local policy-making processes. This was extremely important in making the research more rounded and realistic. The consultations covered territory that was quite sensitive in nature and it was important that authorities felt that there was a shared understanding of the issues and that the consultations were balancedreflecting both the expectations of the communities and the constraints faced by local officials. From Lao Cai, for example, a key aspect of the province feedback workshop was to get consensus and a "green light" on some potentially contentious issues. For instance, disagreements over the management of commune clinics and local infrastructure works had emerged as problem areas that were discussed at length at the provincial workshop.
This consensus was important for two reasons. Firstly, if -as is hoped -the PPAs and the CPRGS consultations will be built upon in the future to play a role i n the monitoring and evaluation of the CPRGS, then having a shared analysis to refer back to is a very helpful starting point. Secondly, these consultations were presented at the regional meetings to serve as a reference point in the discussions and debates between national and subnational government on how to modify the draft. It was important in terms of ensuring the credibility of the research for national government officials to see that the consultation findings were endorsed by the local levels.

Regional workshops
Approximately 500 officials from most of Vietnam's 61 provinces participated in regional workshops to discuss the draft CPRGS. Officials from the Districts where the consultations had been carried out made presentations on the findings from their area. For example, a workshop was held in Tuyen Quang to bring together officials from across the northern uplands regions. Officials came from Lao Cai to discuss the consultation findings with other provinces. This was very important in demonstrating to the meeting that the consultation findings had the backing of local authorities. The Vice Chairman from one of the Districts made comments on the process and findings from the consultations, which represents a powerful expression of buy-in from the level of the administration closest to the communities (even though many of the constraints identified to successful policy implementation were at this level).
The findings and recommendations from the community consultations were presented early on in the two-day regional meetings. This encouraged the subsequent breakout groups -which tackled sectoral and cross-sectoral themes in the CPRGS -to use the findings as a resource in their discussions.

REFINING POLICY WITH THE POOR
This final section looks at how the results and findings from the consultations were used in formulation of the CPRGS. Because the two companion volumes describe and synthesize the field results in some detail, we do not present a rounded picture of all the main policy messages emerging from the research here. Rather, this section concentrates on some of the main aspects in which the local consultations are seen to have influenced and contributed to modifying and refining the CPRGS.
This section also reflects on the main lessons and experience gained from the exercise. This is both in terms of the requirements for successfully linking community level consultations to the overall policy-making process, and the practicalities of actually facilitating the fieldwork on the ground. Finally -taking a look towards the future -we suggest ways in which the capacity that has already been built up through the PPAs and these consultations may be developed in future monitoring of the CPRGS implementation.

The policies and implementation of the policies
One of the clearest messages that resonated from these consultations was the concern at community level (in all sites) about the gap between policy statements and reality on the ground. There was often a high level of expressed support for the stated policy direction in the I-PRSP, but coupled with a list of reasons why the policy objectives and actions may not actually be achieved in practice. As an example, there was little disagreement that the Government should seek to reduce the costs of curative healthcare for poor households. But this was qualified by reservations that (among others): • The criteria and regulations surrounding exemptions from health costs would be so poorly publicized and explained that those who were most in need of this exemption would be unable to claim it; • That health cards might be distributed, but the financing would not reach the point of service delivery so charges would have to be levied anyway; and • That without changes in the incentives structure and accountability mechanisms within the health system as a whole, there would be no improvement for the poor in access to affordable health services.
The challenges to implementing the CPRGS are formidable. One of the most helpful features of the consultations was to shine light on some of the institutional and other constraints that would confront implementation on the ground. The following extract from the Vinh Long consultation is an example of the guiding principles for implementation that emerged from the meetings: "One major challenge of implementing the CPRGS once it is finalized, therefore, will be coordination of actions at the policy, planning, design, implementation, and monitoring stages to maximize their complementarity. Respondents did, however, suggest several conditions under which the CPRGS is most likely to work in favor of the poor. While again, most of these are not new ideas, their importance means they should be emphasized in the document: • If attention is paid to special groups such as the aged (homes for the elderly) and disabled children (special schools); • If the quality of poverty reduction work is improved: higher-quality education (through improved teacher quality and a better curriculum) and infrastructure works (through better construction and maintenance); • If investments are made in institutional strengthening : more staff (doctors in rural areas, agriculture extension workers) with better conditions (salary, incentives to work in rural areas); adequate budgets for Government offices; • If ne w local organizations are established to help prevent risk and provide spiritual support; • If self-reliance is emphasized: better personal healthcare, making best use of aid provided, family support for students; • If appropriate policies are issued on key topics such as minimum wages, hiring of local labor, and minimum required maintenance periods; • If the process of poverty reduction work is improved through better planning, management, implementation and monitoring of development projects and policies: agricultural and industrial production planning, implementation of Government policies ("treat first, pay later"; the grassroots democracy decree), monitoring of the actions of middlemen, the implementation of Government policies, the number of out-of-school children; • If procedures are simplified for employment in SMEs, for registration of SMEs, for access to health services for health insurance card holders; • If participation in local planning is improved through many specific measures mentioned above; and • If mo re aid is provided through expanding the number of service providers in the fields of credit, training." The final version of the CPRGS tackles some of these implementation issues quite specificallyfor example in the request by the Prime Minister in the Preface to the document that agencies incorporate the CPRGS into their activities: "in order to successfully implement this strategy, I call upon ministries, sectors, government agencies, local authorities and mass organizations… to reflect the objectives and institutional arrangements of this document in your annual implementation plans." The CPRGS also sets out the institutional mechanisms that will facilitate cross-agency coordination as the strategy is implemented. These mechanisms have now been established in a Decree issued by the Prime Minister.
Converting the commitments in the CPRGS into actions on the ground will require that these priorities are transmitted into local-level planning processes and it is at this, local level that the constraints identified in the consultations will be most pressing. The 61 provinces in Vietnam have important policymaking and resource allocation powers and national leaders recognize that the achievement of nationally-defined objectives depends on local-level decisions. In three consultation sites further work is now (or will shortly be) underway to support the local authorities to "localize" the CPRGS. The background PPA and consultation work provides a basis for assisting local agencies with evidence-based planning.

Areas in which the consultations had an impact on the policy revision
The I-PRSP -that was used as a basis for the consultations -underwent profound revisions before being approved by the Prime Minister as the CPRGS. Five draft versions of the CPRGS were translated into English, but there were many more drafts in Vietnamese. The community level consultations were by no means the only source of information that MPI and the Drafting Committee used as the iterations took place. National consultations, sub-national consultations, submissions and meetings with line ministries, submissions from donors, NGOs and partnership groups, background papers produced by consultants and researchers and reviews of existing data and information all played a role in influencing the content of the final document. Nor was this the first time that many of these issues were raised -the position of migrants and the affordability of basic social services, both used as examples below, were not raised for the first time in the consultations. The value appears in some cases to lie less in the novelty of the messages and more in the timing and the way in which they were raised.
Attributing any particular change between draft and final versions of the document to the community cons ultations is problematic and depends to some degree on intuition and individual perceptions. Nevertheless, there are many areas in which the researchers and managers most involved with the consultations believe they can trace a direct impact and six examp les are presented here (though there were several others that were described by the participants in the research who have been tracking the impact of the consultations quite closely). These include examples on different kinds of impact: where the text of the CPRGS was changed, where the lending program of one donor has been influenced, and where there has been an impact on the way the CPRGS may be implemented.

Example One -Tackling the social exclusion of migrants in urban areas
There is strong support for the notion that the combined efforts of the PPAs and the consultations were instrumental in putting the rights of urban migrants squarely on the policy agenda. The CPRGS has one of the most eloquent commitments made by Government to date on the need to promote more inclusive service delivery in urban areas: "Solve the problem of urban poverty of special characteristics with regard to employment, income and housing. Ensure the urban poor have equal access to resources, public services and basic social services. Improve the access of migrants, especially their children, to these resources and services" Review labor migration policy and household registration policy to make it easy for people to move to better-paying jobs." This paragraph was not in the earliest drafts of the CPRGS. It is likely that the consultations influenced this change of direction, both directly and indirectly by providing the Urban Forum Sub-group with evidence to make an influential submission to MPI. This commitment has also been picked up in the new World Bank Country Assistance Strategy (September 2002) where it has been highlighted as a "trigger" for moving to a high case lending scenario.

Example Two -Addressing the in affordability of basic social services
A clear message emerged from the consultations that the Government would not meet the ambitious targets it was setting for the universalisation of primary and lower secondary education if the many fees and charges associated with educating children were not reduced. This was the case in every consultation site and was just as important an issue in Vietnam's biggest city as it was in the remote ethnic minority communities in the northern uplands. Like the findings on the urban migrants, this was not the first time that the private costs of educating children or of accessing curative healthcare had been documented. But this message was picked up from the consultations and reinforced by a number of stakeholders in their interactions with MPI as the CPRGS was re-drafted. The World Bank, for example, used this evidence from the consultations in its correspondence with MPI and suggested that reducing the costs of education for poor households was one of the most important changes that could be made to the CPRGS. Later drafts of the CPRGS included the statement: "Construct the package of exemptions and full support mechanisms in primary education for children from poor households, covering school fees, cost of textbooks, contribution fees, cost of meals, lodging costs, transport costs." Analytical work is now underway to enable the Government of Vietnam to assess the resource implications of reducing these costs.

Example Three -Ensuring greater local participation in infrastructure development
The consultations provoked discussion about the Government's intention to use infrastructure development as a means of employment creation. There was a view that this would not happen if the planning and implementation of infrastructure projects remained unchanged. Without more local participation in the choice and planning of projects, without more training and without a concerted effort to use local construction companies, participants felt that they were unlikely to get maximum benefit from either the end product (because it may be either unsuitable for local needs or of low quality) or from the process of construction (through employment). Later drafts of the CPRGS included several specific commitments to change this situation: "Continue the mechanism of the "State and people will do it together" to develop infrastructure in rural areas. The Government will provide support to train and coach staff who are in charge of management, operation, and maintenance of basic infrastructure constructions. Formulate regulations on utilizing basic infrastructure constructions at the commune level.
Promote the use of local labor for implementing infrastructure projects.

Strengthen the participation of local people in planning, implementation and maintenance of local infrastructure projects."
The provision of basic infrastructure through a program targeted at the communes defined as being in "especially difficult circumstances" is a key element of Government's expenditure on poverty reduction. Improving transparency and participation in the implementation of this program through some of the measures suggested in the consultations is likely to lead to a strengthened poverty impact from the substantial investment that goes into this program.

Example Four -Improving transparency and accountability at local levels of Government
Researchers were surprised that it was easier to generate debate on governance issues, particularly those relating to participation in decision-making, than had been the case when the PPAs were carried out in 1999. Participants had a range of suggestions that could help the Government achieve their stated objectives of improved grassroots democracy. As indicated above, strategies for achieving this objective had not been well specified in the I-PRSP. The policy matrix in the synthesis report that was submitted to MPI included more specific proposals (Box 11).
Box 11. Proposals in the Synthesis Policy Matrix for Improved Participation in Decision-making • Establish strong mechanisms for cross-agency coordination and consistency for the CPRGS at all levels of Government; • Identify clear, transparent, public action plans from the strategic framework set by the CPRGS and assign the roles and responsibilities of various agencies and officials in implementing the plan. Ensure communities are fully informed; • Implement the Grassroots Democracy decree nationwide; • Establish mechanisms for giving poor communities a louder voice in determining how local resources are used, which will involve: o Much more information: especially the need for greater transparency of plans, budgets and entitlements under targeted Government programs; o Improving participation: enabling communities to influence decisions about resource allocation and public actions at a local level; o Ensuring accountability: especially improving downward accountability of local officials to the communities, but also horizontal accountability to the People's Councils; • Improve the quality, quantity and integrity of monitoring; • Introduce a strong campaign against corruption at local levels and ensure that communities know how they can play their role in attacking corruption; • Disseminate information about legal rights and entitlements to poor communities, clarifying where and when people must go and what they should do if they need assistance; and • Encourage National Assembly members to develop more interactive communications with their constituents.
Again, many of these points were reinforced through other mechanisms that were operating to influence the CPRGS. Collectively, the efforts made to ensure that these issues were taken seriously in the final draft of the CPRGS were quite successful. There is a section entitled "Implementation of the Public Administration Reform, Legal Reform and Pro-poor Good Governance" that covers many of these areas of concern and some clear statements in the policy matrix that tackle some of these issues. The final document also has sections setting out measures to "Enhance Grassroots Democracy and Strengthen Dialogue between Local Governments and Poor Communities" and to improve the "Provision of Legal Support to the Poor".

Example Five -Upholding labor standards
The I -PRSP set out a strategy for economic growth that was dependent on continued development of the private sector. There is strong evidence that employment opportunities created by a growing private sector (particularly in small and medium enterprises) will be central to poverty reduction over coming years. The consultations confirmed that poor people see this as an important means of moving away from dependence on low productivity agricultural activities.
In discussions on addressing risks and reducing the vulnerability of poor households, participants in the research raised issues associated with the way in which poor people would be able to interact with changing labor markets. In Ho Chi Minh City, in particular, people emphasized the need to be able to assert their rights under the Labor Code, the importance of establishing trade unions in private enterprises and the need for support when losing income because of accidents in the workplace. The consultation findings included recommendations to this effect and later versions of the CPRGS include an explicit commitment in the policy matrix to "review and strengthen the role of labor unions in protecting wo rker rights and working conditions" in order to "protect worker rights and working conditions in a market economy".
In this example the role of the consultations was less to promote new commitments, since the workers' rights themselves were already provided for by law. Arguably, though, the consultations played a role in moving issues of labor rights more into the mainstream discussions on poverty reduction and vulnerabilities generated by chosen growth paths.

Example Six -Providing a platform for participatory monitoring of progress
These consultations have provided strong impetus to the argument that feedback from poor communities should be part of the monitoring system for the CPRGS. First, they proved that it is possible to have meaningful communications with poor households on policy issues and that these would produce credible, useful information. Secondly, they have provided some ideas on how this might actually be done and how, methodologically, this might be tackled. One obvious suggestion is to b uild on the sequence of participatory poverty analysis, followed by consultations on the strategy to develop a participatory feedback loop in these sites.
Later drafts of the CPRGS showed a commitment to build on these consultations and encourage actors outside Government to be involved in participatory monitoring activities. There is a strong argument to be made that the experience of carrying out the PPAs and the consultations in a manner that has engaged national Government agencies has been important in making this sort of monitoring work more acceptable to MPI.

Reflections on the main lessons and experience
As part of the process of preparing this paper, the research agencies were asked to reflect on what they felt to be the main lessons and practical experience gained from the consultations. All agencies said they would be involved with the same exercise again, and that it had paid off in terms of: • Seeing change in the content of the CPRGS; • Opening up debates on policy-making to new stakeholders; • Demonstrating that participatory research could be a useful tool as plans and policies are formulated, going beyond providing background information on the causes and dimensions of poverty; • Raising their own profile with national policy-making agencies; • Developing new skills in policy analysis within their organizations; • Cementing their relationships at a local level and enabling a change of direction towards more policy-based work at the provincial level; • Capacity building of local government counterparts who had been trained to participate in the research; and • Providing an opportunity for the partner organizations to give in-depth thought to a Government strategy -" if we had not been directly engaged in managing the consultation we would not have read and internalized [the CPRGS] so thoroughly" 12 . This is not to suggest that either the process or the product was perfect. In a number of important respects the research was flawed or limited. This partly reflects the constraints of the particular situation -the timeframe and the lack of a vibrant local civil society with strong policy analysis skills. Other deficiencies provide a basis for learning and improving future practices. Reflections on the process and lessons learned include the following.

The need for stronger policy analysis skills outside Government
Whatever their shortcomings, the consultations did mark a breakthrough in terms of bringing policy-making closer to poor communities. It is the strong hope of those who have been involved that this work can be built on in the future and that innovative approaches to involving communities particularly in monitoring the implementation of the strategy should be explored. If this is to take place, then there would be real value in investing in the development of local civil society and research organizations to conduct this kind of work.
Among other factors, it was lack of capacity that restricted the thematic scope of the consultations such that some of the macroeconomic and structural reforms were excluded from the discussions. Several of the research partner agencies felt it would be beyond their skills, and the skills of those they could contract in to help, to carry out consultations with poor households on large parts of the proposed structural reform agenda. Developing this capacity in Vietnam really deserves the attention of international agencies seeking to promote broader participation in policy-making and will certainly be a long-term endeavor.

Challenges of geographical scope and coverage
It was also lack of capacity that constrained the coverage and geographical scope of the consultations. Approximately 1800 people were involved -a little more than the numbers involved in the PPAs -while Vietnam has a population of nearly 80 million people. Participative research of this nature can never hope to be fully representative. Rather, its value lies in being able to capture diversity, and this should be given precedence in the site selection and sampling design. It would have been beneficial to have extended the outreach, but this would not have been possible given that (i) only five organizations were interested in working on the consultations at such short notice and (ii) that it was felt that the consultations would have to go into some depth on the proposed policy measures if they were going to be truly useful.
Over time, as more, local organizations develop the skills to engage in this kind of work there should be potential to capture a broader range of poverty situations in this kind of research work.

Building longer-term processes for broad-based participation
The limited geographical scope and coverage highlights an important issue about the limits of what can be achieved through a consultation exercise such as this when only a fraction of the population is engaged. Done competently, it can provide important feedback that will allow policy level debates to be better informed, hopefully at a time when various alternatives are still being discussed. Clearly, though, it should not become a distraction from or a substitute for a more important agenda of ensuring greater participation of the general population in decisions about policy measures, public actions and resource allocation. The high profile that is needed to make these kinds of activities influential may divert attention away from the need for more mainstream mechanisms for state-citizen interaction and this is a risk that managers of this kind of work should be alert to. In the case of these consultations in Vietnam, a specific line of questioning was included to investigate the kind of actions that would be necessary to include in the CPRGS that would indeed foster more broad-based participation and downwards accountability in mainstream planning and budgeting activities.

The value of flexible, readily-available resources
The development of strategies and policy positions is a messy and contested process. It is often not clear at the outset what opportunities will emerge as the process moves along. For agencies seeking to influence policy, it is not always possible to predict what resources will be needed or when. In the Vietnam case, the agencies involved had to respond very quickly to MPI's initiative with an exercise that (using the roughest of estimates) could not have cost l ess than $150,000 when staff time is included. Several of the partner agencies had funds available from internal sources that could be used to fund the consultations. Funding and human resource gaps were filled by the World Bank using a Trust Fund established with funds from the UK DFID. Without this supplementary support the whole exercise would have been quite diminished and less influential (many of the additional activities -such as the regional consultations -were also funded from the same source). One can project from this experience that if there are going to be further efforts to support the Government in these kinds of activities, then agencies engaged in this work will need to position themselves to be responsive when unexpected opportunities arise.

Involvement of sectoral agencies and line ministries
At the time when MPI requested assistance in getting the consultations underway, the engagement of line ministries and sub-national levels of Government in strategy development had been inconsistent. MPI was the main Government counterpart in this work and there was less awareness of this work and input from other ministries as the research framework was designed. Had it been foreseen at the outset that one channel for internalizing the findings in the re-drafting process would be through the line ministries and their interactions with MPI, then it would have been sensible to seek their involvement more vigorously at early stages of the research design. If the Ministry of Education and Training and t he Ministry of Health had been involved in designing the questions on accessing basic social services, for example, then one can speculate that they might have felt more closely associated with the findings. This would be an important dimension to explore in developing similar exercises that serve a monitoring role.

Highly experienced and adaptable facilitation teams are essential
Given the challenges involved in carrying out this type of study, team selection becomes critically important in order to draw in the required range of facilitation skills and experience. The fieldwork presented a considerable challenge to all the facilitation teams in terms of learning how to actually handle the consultation process. Many of the facilitators had previous experience in carrying out the PPAs, but little participatory research with such an explicit policy focus had been carried out in Vietnam, and there is still only limited experience to draw on from other countries -this was new territory for everyone involved. It was found that new types and combinations of facilitation and analytical skills were required. These could only be anticipated in advance to a certain extent. It was more a case of learning-by-doing and adjusting the approach on a day-to-day basis during the fieldwork period.
In addition to having excellent basic facilitation skills, the following pointers can be given with respect to the requirements for team composition: • The teams should ideally represent the range of institutional partners and an appropriate balance between researchers and practitioners. In this case, the facilitation teams were variously drawn from staff of the research agencies, local government staff and officials, staff of local NGOs and socio-economic research institutes and contracted participatory research specialists; • The teams should include people that have a sound knowledge of the overall orientation and content of government policies and in-depth understanding of the I-PRSP / CPRGS process, as well as people who have an understanding of local government policy applications and implementation on the ground; and • The teams should consist of people with the required range of disciplinary backgrounds and in-depth knowledge of the social welfare, productive, financial, construction and other sectors as covered by the strategy.
Team building is also important to consider. The facilitation teams from each site were involved at an early stage in the formulation of the research framework and methodology, and group training for the teams and local counterparts was provided in each province at the beginning of the fieldwork period. In addition, there was a considerable amount of interaction between the teams during the fieldwork, with cross visits being made in some instances that helped improve coordination and sharing of experience as the consultations took place.

Linking policy actions to local realities
The practical challenge for the facilitation teams was in how to link macro-level policies to local realities in such a way that would enable the participants to make constructive inputs to these policies. On the one hand, the consultations were rooted in real issues and concerns for the local participants (for instance, how to create better employment opportunities for the poor and how to improve the provision and performance of social services). On the other hand, the information being sought from them was on their viewpoints, ideas and opinions on an arena and level of decision-making that was in practice (or had been until now) far removed from them. The risk is that without highly experienced facilitation it is easy to loose the thread of the consultation process and the discussions can become too abstract and general. Alternatively, the discussions can become too involved in local detail, which is then very difficult to aggregate into policy messages.
One main difficulty arose out of the sheer volume and range of the policy actions and issues under discussion. In Ho Chi Minh City, for instance, the broad scope of the consultations initially made some of the facilitators anxious. This was because a majority of the team members had a practical social work background and they were more familiar with facilitating community development work rather than facilitating policy discussions. The heavy content also meant it was not always possible to use the participatory techniques effectively. This was because the facilitators were trying to get through a large number of topics and having to deal with huge amounts of information. In the first 3 days of the fieldwork in Ho Chi Minh City, all groups were attempting to cover all five research areas. However, this proved too much for both the facilitators and participants and was quickly adjusted so the topics were subsequently divided between the groups.

Getting to the underlying governance and institutional issues
Another challenge in facilitating this type of consultation refers to the analysis of institutional issues and local governance. Institutional analysis is commonly a weak aspect of much participatory research. For instance, in the PPA studies carried out in 1999, this was consistently the least well dealt with set of issues. While the PPAs generally gave a coherent and wellsubstantiated picture of the profile of poverty situations in different regions of the country, and of the factors that contribute to poverty and household livelihood strategies, the institutional analysis was generally fragmented and remained inconclusive. In these consultations a more concerted and structured attempt was made to understand these issues from the local perspective that proved to be quite successful.
An important lesson from this is in how to actually approach and discuss such issues in the focus group meetings. It can be difficult to address institutional topics through general lines of questioning and sensitive to address them directly. For instance, when discussing implementation of the Government legislation on grassroots democracy, putting general questions to local people such as "have you heard of the Democracy Decree" or "are you more empowered now than in the past" will commonly meet with a baffled response. Rather, more concrete lines of questioning relating to the actual mechanisms of implementation of the legislation are likely to yield a more forthcoming response. Similarly, such issues can be approached from the practical perspective of the way in which, and extent to which, the administrative reform measures and greater public participation are being achieved through the management and provision of services on the ground. For instance, such questions may include: • Are village meetings held more regularly now than in the past?
• Who does and does not participate in these meetings?
• What things are discussed in the village meetings?
• In the village meetings do you discuss commune plans and budgets?
• Do you know if 'commune supervision boards' have been established in the commune to supervise commune infrastructure works? • If so, do you know who is on the board and what its function is?

Identifying equity concerns in policy implementation
An important facet of the consultation methodology should be to identify areas in which equity is a major concern amongst local participants or is an issue that needs to be addressed in rolling out parts of the poverty reduction strategy. This can be done through: (i) ranking the proposed policy actions and analysis of which social groups are most and least likely to benefit from them, and (ii) assessing differential access to services and how equity can be ensured in the implementation of the proposed policy actions and services in the future.
Such equity concerns were covered more fully in some consultation sites than in others. In general, however, it was found that insufficient time was available during the fieldwork to fully explore and disaggregate the potential benefits and impacts of the proposed policies in this way. In particular, more time would be required to identify the range of social groups in each locality and to ensure their adequate representation in the consultation meetings. Deeper investigation would also be required with respect to the overall provision and volume of services vis-à-vis factors that may prohibit equity in access to these services and the enforcement of rights. It was only possible therefore to make a general assessment of equity issues in these consultations and this is certainly an aspect that should be strengthened in the future through monitoring the CPRGS implementation.

Building on the consultations -future monitoring of the CPRGS
The CPRGS has a strong accountability framework in which the Government sets out what it is trying to achieve and how progress will be measured, at least at the broad outcome level. The General Statistics Office (GSO) has put in place a system of biannual household surveys. These will measure household expenditures to track material changes in household livelihoods and will gather data on other dimensions of poverty. There will be great value in supplementing this information with qualitative surveys, which can provide feedback on many dimensions of the CPRGS less well captured by quantitative data. This could also identify why and in which areas progress is being achieved, or not.
One option would be to return to these local consultation sites at regular intervals as the CPRGS is being implemented. There is now in place a network of local officials and community members who are familiar with the research techniques and who have articulated their appreciation of this kind of participatory research directed at improving policy. This is a valuable resource and it would be a shame not to use it in future work. The future research could ask communities and local officials to respond to questioning along the following broad lines (obviously refined much further): "The Government was hoping to achieve X over the past 2 years. Did this happen, here, in your area? And if not, what factors might have prevented this from occurring? And what could be done in the future to address these constraints?" Both the fieldwork and the analysis would be challenging, but -done well -this could provide extremely useful, focused information that could feed directly into the process of improving national plans and policy-making.
These findings could also feed into future household survey design, informing the scope and detail of questionnaire used. The findings from the PPAs have informed the discussions on the design of the household survey that has been carried out in 2002. As an example, the issues that arose in discussions with unregistered urban migrants in the Ho Chi Minh City PPA led to a stronger effort in the 2002 GSO household survey to ensure that households without formal registration in urban areas would be included in the sampling frame. The would be a good argument for using future participatory research to ensure that the household survey is addressing the relevant dimensions of poverty.
In such future monitoring of the CPRGS, it would be beneficial to expand the number and range of sites in which this type of exercise would be carried out. For instance, to augment the work in Ho Chi Minh City, it would be useful to include a wider range of urban situations, including selection from the rapidly growing secondary cities and peri-urban centers. Coastal areas and communities could also be better represented, as these are currently areas of rapid economic development and highly complex interactions between wealth, prosperity and poverty situations. The only truly mountainous location included in the PPAs and these consultations is the northern province of Lao Cai, which could be complemented by an additional upland area. As part of this effort, it would be essential to broaden the base of competent researchers, and to continue to build the capacity and skills of local NGOs, socio-economic research institutions and other agencies to carry out this type of complex policy-oriented research.
There would also be value in engaging line ministries more actively in the design of any future participatory monitoring of the CPRGS. This would broaden the sense of ownership across national government and would hopefully strengthen the possibility that the findings are acted on. One way of seeking this engagement would be through the various Government-donor-NGO working groups that are active in a number of sectors. It could also have other benefits in supplementing the data that is collected through administrative reporting channels and informing line ministry responses to the data. Broadening the range of agencies involved in the design and implementation will -almost certainly -lengthen the time it takes to develop a workable research framework. One response to this problem would be to start work very soon to build consensus on the need for this work and to start developing a research framework.