Does Decentralization Deliver Health Services to vulnerable Groups in Kenya? A Case Study of Busia and Kakamega Counties -Kenya


 1. BackgroundDecentralization has become an increasingly important concept in development theory and practice in the world since the 1970s. Decentralization is generally understood to be referring to policies and processes that shift the locus of decision-making and management to the local level, Hellington (2005). According to Elamon (2004), decentralization is particularly attractive because some forms of it promise direct accountability of public institutions; improved management of resources; more reliable information for planning and a more effective interagency collaboration and coordination. Furthermore, it is argued that local officials and community leaders have better knowledge and information about the local context and are therefore more aware of what can work best within the resource constraints while contributing to the achievement of the broader social, economic and development goals of central government.In the health sector, the impetus for decentralization has its roots in the Alma-Ata declaration of Primary Health Care, WHO (1978). This was enhanced by the World Health Organization’s advocacy for strengthening the health system at district level as the most appropriate level for planning and providing health services. Bankausaikaite (2005) argues that decentralization therefore appeals to the health sector largely because it is seen as a means to achieving the Primary Health Care (PHC) goals and for its potential to achieve multiple objectives. Furthermore, decentralization continues to be a subject of intense discussions in the health sector reform processes and is seen as a key strategy for achieving poverty reduction objectives and for attaining the global Millennium Development Goals (MDGs). Klouda (1995) argues that in the context of Sub-Saharan Africa, a region that has the majority of poor countries in the world, poor health indices and high burden of preventable diseases including a devastating impact of HIV/AIDS, decentralization is perceived as offering a key development strategy in addressing these complex problems. In the specific case of HIV/AIDS, decentralization is perceived as suited to stimulate far-reaching changes that can respond to the complex factors that lead to its transmission and the multiple services necessary to adequately support affected communities through multisectoral actions. Furthermore, Ribot (2002) has said that development is often argued as a motive for decentralization. It is argued that decentralization facilitates integrated local development, a necessary factor in the control of HIV/AIDS. The United Nations Development Program, UNDP (2000) has actively promoted decentralization in local government and points out that one of the earliest efforts in using decentralization to address the HIV/AIDS pandemic in sub-Saharan Africa has been through the Alliance of Mayors and Municipal Leaders on HIV/AIDS in Africa. These efforts have been supported by donor agencies such as USAID, UNDP, World Bank and Private Foundations. According to Schneider (2005) the Ford Foundation in Southern Africa (South Africa, Zimbabwe, Namibia, and Mozambique) for example, focused its grant making on strengthening local government and civic responses to HIV/AIDS at district and community levels. In Kenya, the government adopted a deconcentrated form of decentralization at independence in 1963. However, since the presidential and local elections of 2002, the government has tended to shift towards a more devolutionary form of decentralization by enacting policies and Acts of parliament that enable it to transfer funds to decentralized levels. According to NCCK ((2005) these funds include the Constituency HIV/AIDS Fund of 2003, established by legal notice 170 of the Corporations Act 1999. According to Orege (2005) this fund is primarily expected to support prevention, care and mitigation of the impact of HIV/AIDS within the communities. Managed by the National AIDS Control Council (NACC), it provides grants to civil society organization at constituency and district levels. According to NACC (2006) the Constituency AIDS Control Committees (CACCs) are responsible for the implementation, coordination and management of funded projects at the local level.The overall aim of the study is to assess how the decentralized AIDS Funds are used to respond to the social problems among children affected by HIV/AIDS in Amagoro constituency in Busia County and Mumias and Matungu constituencies in Kakamega County.2. METHODSStudy DesignA case study design was chosen for this study because of the complexity of the issues to be investigated and depth and accuracy was required.Study SettingThe study was conducted in Amagoro Constituency in Busia County; Mumias and Matungu constituencies In Kakamega County. The choice of the sites was informed by a number of factors; the highest prevalence of HIV/AIDS (14.9%) compared to the national average of (7.1); convenience and accessibility of key actors in government and beneficiary communities.Study PopulationA purposive sample of key actors directly involved in implementing and managing the decentralised HIV/AIDS programme at the national, regional county and constituency levels were included in the study. At the Community level, interviewees included key informants drawn from CBOs, management committees; members of associations of people living with AIDS. A total of ninety eight (98) individuals were interviewed across the three sites.Data collection Document analysis, key informant interviews and semi-structured interviews were the main methods used in this study. A full listing of documents collected at the national and local levels are provided in the Appendix.Data AnalysisThematic analysis was primarily used to analyse the data according to the pre-determined topics identified in the conceptual framework, namely authority, accountability, management, technical capacity and information flows.At an analytic level, a comparative analysis of findings across the study sites was done to identify county and constituency level factors that affected the implementation of the Constituency AIDS control programme.At the conceptual level, findings and lessons emerging from descriptive and analytical levels prompted reflection on the theories and assumptions underlying the promotion of decentralization as a development approach.3. ResultsThe study found that the AIDS committees delivered services to about 11% of children estimated to be affected by the pandemic. Factors affecting the delivery of services to vulnerable children identified in the study include weakness in accountability due to patronage, corruption and non-enforcement of sanctions. Others include conflict of interest among members of the committees; diversion of funds to pay sitting, lunch and facilitation allowances to the officials, complacency and inaction by national officials and outright abuse of office through technical manipulation of civil society proposals by consultants4. Conclusion This study has analysed data collected from interviews with both policy and implementing level key informants at district and community levels across the three constituencies. The information collected has been triangulated with information from secondary sources and with information collected through informal conversations with a range of stakeholders in the three constituencies. The study identifies a range of factors embedded within the project cycle that have affected the implementation and outcomes of the CACC program. Manipulation of CBOs projects at various stages of development and at various levels of the decentralized NACC structures was the greatest problem encountered by implementing agencies. The study concludes that problems affecting decentralisation are deep-seated as these are influenced by broader socio-political contextual factors.


Introduction
The main focus of the development sector debates over the years has been on whether or not there is adequate funding for development programs. However, following the adoption of the Millennium Development Goals (MDGs) by the United Nations Assembly in September 2000 by the international community to ght poverty in the developing countries, especially Africa, coupled by a substantial increased funding, for example, Dalberg (2006) has argued that from the Overseas Development Assistance (ODA) alone contributes an estimated US$100 billion a year while according to Nandini (2007) funding for HIV/AIDS Increased from US$ 300 million in 1996 to US$ 8.9 billion in 2006.
The international focus on poverty reduction and the increased funding for development assistance have shifted the focus of international development dialogue from adequate funding to issues of the effectiveness and e ciency of these large resources in delivering services to the poor and vulnerable populations. The international donor community and the recipient governments are now raising important and practical questions about the types of interventions; the best strategies that can be used to deliver services and which actors are best placed to implement these interventions.
Decentralization has been promoted by the international community as one development approach for delivering services and programs to the vulnerable populations in order to achieve the millennium development goals. Most governments in Sub-Saharan Africa have adopted decentralization as an approach to social and economic development.

Background
Dentralization has become an increasingly important concept in development theory and practice in the world since the 70s. According to Hellington (2005) decentralization is a general term referring to policies and processes that shift the locus of decision making and management to the local level. Elamon (2004) argues that It is particularly attractive because some forms of it such as local government and autonomous hospital boards promise direct accountability of public health care institutions; improved management of resources; more reliable information for planning; and a more effective inter-agency collaboration and coordination. Some proponents of decentralization argue that local o cials and community leaders have better knowledge and information about the local context and are therefore more aware of what can work best within the resource constraints while contributing to the achievement of the broader social, economic and developmental goals of central government.
In the health sector, the impetus for decentralization has its roots in the Alma-Ata declaration of Primary Health Care, WHO (1978). This was enhanced by the World Health Organization's advocacy for strengthening the health system at district level as the most appropriate level for planning and providing health services. Bankausaikaite (2005) argues that decentralization therefore appeals to the health sector largely because it is seen as a means to achieving the Primary Health Care (PHC) goals and for Its potential to achieve multiple objectives. Furthermore, decentralization continues to be a subject of intense discussions in the health sector reform processes and is seen as a key strategy for achieving poverty reduction objectives and for attaining the global Millennium Development Goals (MDGs). Klouda (1995) argues that in the context of Sub-Saharan Africa, a region that has the majority of poor countries in the world, poor health indices and high burden of preventable diseases including a devastating impact of HIV/AIDS, decentralization is perceived as offering a key development strategy in addressing these complex problems. In the speci c case of HIV/AIDS, decentralization is perceived as suited to stimulate far-reaching changes that can respond to the complex factors that lead to its transmission and the multiple services necessary to adequately support affected communities through multi-sectoral actions.
Furthermore, Ribot (2002) has said that development is often argued as a motive for decentralization. It is argued that decentralization facilitates integrated local development, a necessary factor in the control of HIV/AIDS. The United Nations Development Program, UNDP (2000) has actively promoted decentralization in local government and points out that one of the earliest efforts in using decentralization to address the HIV/AIDS pandemic in sub-Saharan Africa has been through the Alliance of Mayors and Municipal Leaders on HIV/AIDS in Africa. These efforts have been supported by donor agencies such as USAID, UNDP, World Bank and Private Foundations. According to Schneider (2005) the Ford Foundation in Southern Africa (South Africa, Zimbabwe, Namibia, and Mozambique) for example, focused its grant making on strengthening local government and civic responses to HIV/AIDS at district and community levels.
In Kenya, the government adopted a deconcentrated form of decentralization at independence in 1963. However, since the presidential and local elections of 2002, the government has tended to shift towards a more devolutionary form of decentralization by enacting policies and Acts of parliament that enable it to transfer funds to decentralized levels. According to NCCK ((2005)  According to Orege (2005) this fund is primarily expected to support prevention, care and mitigation of the impact of HIV/AIDS within the communities. Managed by the National AIDS Control Council (NACC), it provides grants to civil society organization at constituency and district levels. According to NACC (2006) the Constituency AIDS Control Committees (CACCs) are responsible for the implementation, coordination and management of funded projects at the local level.
The Constituency Development Fund, GOK (2003) aims at supporting broad priority development activities at community level including in education where up to 10% of the allocated funds can be dedicated for bursaries. However, the fund cannot be used to support religious and political activities.
The Constituency Bursary Fund (2003) aims at cushioning poor households from impacts of poverty, unstable economy and the devastating impacts of the HIV/AIDS pandemic. The fund targets children from poor households, from arid and semi-arid areas and children affected by HIV/AIDS including orphans. 5% of the allocation is set aside for the girl-child and other children under special circumstances such as the disabled.
The Local Authorities Transfer Fund (LATF) is expected to improve service delivery and governance in local authorities. According to DFID (2002) and Schuler (2004) Participatory Local Authorities Service Delivery Plans (LASDAP) involving bene ciary communities are a pre-requisite for full funding of the LATF.
The Poverty Eradication Fund was established in 1999 as a follow-up to an o cial commitment to the UN summit on Social Development held in Copenhagen in 1995. The Fund aims at providing basic social services, enhancing economic growth and placing emphasis on urban livelihoods. It targets individual members of self-help groups and provides access to affordable credits to enable them to implement income generating projects that improve their livelihoods.
It is signi cant that for the rst time in the history of independent Kenya that legislation at least in theory transfers nancial and decision making powers to the local levels. This type of decentralization would perhaps give the people a sense of ownership of their development programs, allow them to allocate resources according to their preferences, and mobilizes them to use their skills and knowledge to improve their conditions. However, as Nduva (2005), Nyamu (2005) and Kimenyi (2005) argue, there are already indications of ambiguity, lack of implementation and accountability frameworks, which threaten the success of this decentralization effort.
This study is primarily concerned with the constituency AIDS Control Fund which was established to support priority HIV/AIDS interventions including addressing social problems among affected children. This decentralized fund has been chosen for study because not only is it addressing the greatest social, economic and political challenge affecting Kenya and the rest of the continent, but also because according to Klouda (1995) HIV/AIDS is a marker of critically weak areas in approaches to development. The study will examine the operational and developmental relationships between this fund and other decentralized community development oriented funds. These include the Constituency Development Fund (CDF) the Constituency Bursary Fund (CBF) and Local Authorities Transfer Fund (LATF), which also aims at responding to the communities priority needs at the local level. These funds are of relative importance to the prevention, treatment, mitigation, care and support of vulnerable children and families affected by HIV/AIDS because each of them contributes to addressing one or several of the multiple impacts of the pandemic.
Despite the political and technical rhetoric about decentralization as a key development strategy, implementation of decentralization policies has consistently been problematic. Most researchers Khalegian (2003); Smoke (2003); Elamon (2004); Saltman and Bankauskaite (2005) have pointed out the major disconnect between policy pronouncements, health outcomes and the limited empirical data on implementation of decentralization. Gilson (1994); Smoke (2003); Bankauskaite (2005) and Victoria (2008) have argued that this lack of empirical data has contributed to the inadequate understanding of the decentralization process. There are therefore many unanswered questions about how decentralization works, for whom, in what circumstances and with what outcomes. Campbell (2003) has argued that although worldwide, there has been emphasis and strong focus on local, coordinated and comprehensive responses to the HIV/AIDS pandemic as a development approach to addressing the root causes of the pandemic, there has been very little research done in understanding the challenges at the local level. The Kaiser Network (2003) observes that despite large amounts of funds owing to Africa for ghting the escalating deaths, not adequate resources are reaching the community.
A review of the African literature on decentralization reveals two important sets of factors affecting development and implementation of decentralization policy. These factors relate to power, the actors who hold it and governance.
The way power is distributed and how it is exercised at various levels of the decentralized process; how roles, responsibilities and authority are assigned, to whom and with what resources; have been shown to have a great in uence on the policy implementation process. These power related factors are exercised in different contexts and by a wider range of actors that include government o cials; international organizations and donors; policy communities and intended bene ciaries of the decentralization process.
The second set of factors includes governance and the management capacity of the actors to design, implement, monitor and evaluate decentralized policy interventions. The literature shows that understanding the behaviors and attitudes of actors; communication systems; program design and accountability mechanisms and reporting procedures; allocation and ow of nancial resources; inter-sectoral collaboration and coordination; community participation; strengthening the capacity of all management and implementation structures; and clarifying responsibility for management of human resources are critical factors for its successful implementation.
Underlying both these factors is the process and context within which the policy is developed and implemented. The nature of the policy process; the key actors; the level of commitment and interest of the actors; the contextual factors especially situational, economic, political and cultural have been shown to substantially in uence the policy process.

Problem statement
This study attempts to address some of these concerns by trying to understand how these complex factors affected the provision of care and support to children affected by HIV/AIDS through the implementation of the policy (Constituency AIDS Control Fund) in three constituencies. No such study has been conducted in the research sites.

Justi cation and signi cance
This study is justi ed because: 1. HIV/AIDS is a leading cause of morbidity and mortality and is a marker of critically weak areas in development 2. Orphans and vulnerable children are the visible faces of the HIV/AIDS pandemic as they mirror the suffering and impacts of the AIDS pandemic in society.
3. There has been a large amount of money from donors going to address the AIDS pandemic

Most governments agree that those affected by AIDS need social services and protection
This study is signi cant because the study ndings are expected to contribute to a better understanding of critical factors affecting implementation of decentralization policy at local level. From an academic perspective, the study will contribute to narrowing gaps in knowledge which has been identi ed in the literature by providing empirical evidence on the implementation of decentralization policy at grass-roots level.

Aims and Objectives
The overall aim of the study is to assess how the decentralized AIDS Funds are used to respond to the social problems among children affected by HIV/AIDS in Amagoro, Mumias and Matungu constituencies.
Objectives 1. To describe and identify factors affecting the implementation of the Constituency AIDS Control program in responding to the social problems among children affected by the HIV/AIDS pandemic.
2. To assess and compare how the constituency AIDS Committees in the three study constituencies had responded to the speci c needs of vulnerable children affected by HIV/AIDS and to identify the factors affecting their performance.

Research Questions:
1. How were decentralized AIDS funds used to provide services to children affected by AIDS 2. What factors affected the performance of key actors in the delivery of these services?
3. What lessons can be drawn from this case study about the challenges of decentralization of health services in Kenya and their implication for Africa?

Conceptual Framework
A framework initially developed by Hutchison and Lamont (2004) to evaluate the entire decentralization system was adopted and modi es to analyze the factors in uencing the implementation of the Constituency AIDS Control Programme and its outcomes on vulnerable children.

Study Design
A case study design was chosen for this study because of the complexity of the issues to be investigated and depth and accuracy was required.

Study Setting
The study was conducted in Amagoro Constituency in Busia County; Mumias and Matungu constituencies In Kakamega County. The choice of the sites was informed by a number of factors; the highest prevalence of HIV/AIDS (14.9%) compared to the national average of (7.1); convenience and accessibility of key actors in government and bene ciary communities.

Study Population
A purposive sample of key actors directly involved in implementing and managing the decentralised HIV/AIDS programme at the national, regional county and constituency levels were included in the study. At the Community level, interviewees included key informants drawn from CBOs, management committees; members of associations of people living with AIDS. A total of ninety eight (98) individuals were interviewed across the three sites.

Data collection
Document analysis, key informant interviews and semi-structured interviews were the main methods used in this study. A full listing of documents collected at the national and local levels are provided in the Appendix.

5 Data Analysis
Thematic analysis was primarily used to analyze the data according to the pre-determined topics identi ed in the conceptual framework, namely authority, accountability, management, technical capacity and information ows.
At an analytic level, a comparative analysis of ndings across the study sites was done to identify county and constituency level factors that affected the implementation of the Constituency AIDS control programme.
At the conceptual level, ndings and lessons emerging from descriptive and analytical levels prompted re ection on the theories and assumptions underlying the promotion of decentralization as a development approach.

Context of the study sites
This section presents a brief description of the three case study sites in order to provide an understanding of the local context within which the CACC program was implemented. It presents an overview of the administration and history of the areas; the socio-economic characteristics of the in terms of population, adult literacy, main economic activities, access to health services, the situation of vulnerable children including the nutritional status of the <5s; and a listing by constituency and focus area of the international organizations working on HIV/AIDS in the three constituencies.  (2007) is a more robust government document that guides its national planning and development activities; is based on empirical research as opposed to estimates; and is the most up to date available source of statistics in the country, the author adopts the data and information provided by this source for further analysis and discussions in this thesis.
According to the KIHB survey, KBS (2007) the total population of the three constituencies is 838,654 of whom 585,967 reside in Mumias and Matungu constituencies while 252,687 are resident in Amagoro constituency. Understanding the distribution of the population by sex of the household head was important for this study to determine the number of female headed households in each constituency and to compare across the sites, this would serve to identify vulnerable individuals in the study areas that require to be targeted by decentralized interventions such as the CACC program. According to the KIHBS survey, KBS (2007) the number of male and female headed households were almost equal in the two study districts with Mumias and Matungu constituencies (Mumias district) having 67.6% male and 32.4% female compared to 66.5% male and 33.5% female headed households in Amagoro constituency (Teso district).
The mean household size was slightly higher in Amagoro at 5.9 as opposed to 5.2 for Mumias and Matungu constituencies. Adult literacy was relatively higher in Mumias and Matungu constituencies at 80.2% compared to 76% in Amagoro constituency. Action for child protection (2003)), and NACC (2007) in its guidelines to CACCs to undertake an inventory of orphaned and vulnerable children and the Kenya National Children's Act (2001) de ne vulnerable children in terms of the child's ability for self-protection.
One of the main objectives of this study is to assess how the constituency AIDS control committees in the three constituencies responded to some of the social problems of Orphans and vulnerable children (OVCs). UNICEF and UNAIDS (2004) (2000), Csete (2001) and Johnson, Ferguson et al (2002) have found that most vulnerable children in this region are exposed to these problems.
One of the main objectives of this study is to assess how the constituency AIDS control committees in the three constituencies responded to some of the social problems of Orphans and vulnerable children (OVCs). UNICEF and UNAIDS (2004) de nes social problems among OVCs as including inability to attend school; inadequate food; shelter; materials needs; access to health services; exposure to discrimination; sexual abuse; exploitation of child labour; and life on the street among vulnerable children. In separate studies in western Kenya, several authors including Othather (2000), Csete (2001) and Johnson, Ferguson et al (2002) have found that most vulnerable children in this region are exposed to these problems.

Organizations Working on Vulnerable children affected by HIV/AIDS in the study constituencies
This study looked at service provision to the vulnerable children within the study communities. It was therefore important to identify the organizations working on HIV/AIDS and development related matters in the study areas.
In Table 4 below, I present a listing by constituency of the international and national organizations that were working in the HIV/AIDS eld by program area of focus and show the speci c activities implemented.  It would seem that the broad developmental approach that was emphasized in the 2000 guidelines is progressively being eroded as more emphasis seems to be placed on the diseases speci c medical interventions as can further be seen in the list of main activities to be implemented to achieve program goals ((NACC 2006).

Con ict of interest among CACC Members
Although NACC's guidance to the CACCs requires that members should be conscious of the con ict of interest and should follow NACC guidelines to resolve con icts, this study found that: In Matungu constituency 7 out of 11 interviewed policy level key informants within the CACC said that they run, have control over or had assisted to establish an implementing CBO in the constituency.
In Amagoro constituency all the 8 policy level key informants within the CACC stated that one policy level actor had established a consultancy rm outside the district and that a senior civil servant from the district had established an NGO (CACC 2003); while a third policy level actor run a CBO and was linked to several other CBOs in the district where he had substantial control and in uence over decision making.
In Mumias constituency, all the 13 policy level key informants said that there was only policy level actor who had established a consulting rm in town. The number of CBOs under the control of the CACC members was relatively larger in Matungu constituency compared to the other two constituencies for reasons that were not clearly explained by both policy and implementing level key informants in the constituency.
The con ict of interest by those in the governance positions of the CACC program raised many ethical and accountability questions which did not seem obvious to most policy and implementing level key informants. Slightly more than half of these respondents saw no con ict of interest while the rest were certain of a con ict of interest. When asked if their positions as CACC members and as chairpersons or patrons of the CBOs were tantamount to a con ict of interest, most of the policy level informants especially in Matungu constituency argued that there was no con ict of interest, the quote below captures the perception of most of the members who saw no con ict of interest: " Not at all, some people may think because I belong to CACC that is why my group got the money but its just perception. We are told when we approve proposals at CACC level in Nairobi there is random sampling of these proposals. Councillors in CACC have wondered why proposals in their wards are not funded yet other wards are funded".
They argued that it was perfectly legitimate that as community members, they provide leadership to their communities and these included formation of development entities such as CBOs. Furthermore, one key informant, who was also the chairman of the association of PLWHAS in one of constituencies, argued that it would be immoral to deny his group funding because he sits in the CACC.
For those policy level key informants who saw a con ict of interest, the arguments touched on two key issues, rstly, promotion of lack of transparency and corruption. To this extent, one well connected and positioned key informant narrated the experiences of the CACC in one of the constituencies, where the chairperson was closely associated with one of the CBOs. Throughout his tenure as chair and the life of the project.
The second point made by the policy level key informants was that running or control of CBOs perpetuated inequities in the allocation of resources within the CACC and dependency on o cials by the CBOs. They pointed out several examples of how members associated with implementing CBOs or consultancy rms had canvassed their counterparts to approve proposals submitted by the CBOs these members were associated with. The quote below from one policy level key informants sums up the thinking of those who spoke on this matter. "It was the elite members of the community who came up with groups, went for funding which they could easily get away with it since CACC o ce had no funding to monitor their activities" In separate follow-up in-depth interviews with three well placed informants who had a deep insight into the operations of the CACCs in the three constituencies, policy level key informants con rmed the corruption allegations being levelled against some of the o cials by implementing agencies and community leaders. Asked why the o cials had resorted to such corrupt behaviors, the responses although expressed in different ways pointed to three main reasons.
Firstly, the un-funded mandate of monitoring and supervision of CBOs; the policy level key informants argued that the o cials were forced to solicit funds from CBOs to enable them to travel and meet subsistence expenses since NACC expected its local o cials to certify that they had witnessed the execution of the activities reported by implementing organizations. Said one policy level key informant: CACC members go to attend functions of CBOs in their area. The CBOs motivate the members by giving them transport of about Ksh.200, we do not leave these meetings of CBOs to the program o cials only Secondly, the key informants argued that corruption was the prevailing culture within government ministries and that even if the o cials in question desisted from it, others were doing it anyway.
Finally, key informants stated that among the community based members of the CACC technical committees, the main reasons for their involvement in corrupt deals within their own communities were attributed to poverty. The informants argued that with a NACC approved allowance of only Ksh.200, it was di cult for these members to meet their travel and subsistence expenses during eld work as well as meet their family needs.
Another approach commonly used by o cials to divert resources as stated by some policy and a majority of implementing level key informants in the study sites with apparent collusion of their superiors was issuing of directives which were enforced by threats to the local implementing organizations. For example, some key actors in two of the three study constituencies directed all the implementing agencies that only MOH appointed consultants were to provide technical and management services to the funded projects. This directive which was implemented with little resistance served to transfer power to appoint consultants and facilitators from the implementing CBOs and the bene ciary communities to the local o cials.
Those who resisted the directive were threatened that their expenditure returns and further funding which had to be approved by the concerned actors would not be sanctioned. Indeed the few implementing organizations in Mumias and in Amagoro constituencies that challenged this directive reported consistently experiencing negative relations with these key actors during project implementation as expressed in the following key informant from one of these CBOs: Our relationships with CACC were not good; characterized by suspicion and negative follow-up simply because we declined to give 10% kick-back as demanded.

Directives to pay allowances
The implementing CBOs in Mumias and Matungu constituencies were further directed to pay government o cials lunch and sitting allowances whenever they participated in any functions related to project implementation. Table 2 below provides a snapshot of the situation in the two constituencies.  The nding that program staff and o cials were being by CBOs for technical assistance was contrary to NACC guidelines to CACCs that such technical assistance should be provided free (NACC, 2006).
Payments of lunch and sitting allowances to the local o cials by implementing organizations were not reported in Amagoro constituency. According to the policy level informants and project records, the only allowance that was paid to CACC members was ksh.200 later increased nationally to Ksh. 500 by NACC being members travel and subsistence allowance during regular program meetings.
The misuse of power and diversion of resources was not con ned to the CACCs alone, this study found collusions between some NACC o cials at the national level, CACCs o cials, some senior civil servants; and some staff of the implementing CBOs. Several major concerns were raised by policy and implementing level key informants across the three constituencies. Three examples are provided to illuminate the main issues:

Complacency and Inaction by the National AIDS Control Council -NACC
Inaction by NACC towards in uential outside actors who used their strategic positions in collusion with o cials in the governance structures of the program to get large grants in the name of local communities but turned the funds to their personal use; An example of this mechanism of diverting funds by local o cials was reported in Amagoro constituency where a senior o cial in the provincial administration headquarters, in collusion with a senior district level actor, with strong connections with NACC o cials at national level, received a grant of Ksh. 2.1 million to implement an orphan support project in the district but which was never implemented. When the proposed bene ciary community under the leadership of their chief got information about the project, they reported the matter to the CACC o ce since they had not been consulted or seen any activities being implemented on the ground.
The CACC technical committee investigated the matter and were satis ed that it was indeed fraud. In their letter (CACC 2003), addressed to the Director of NACC and copied to provincial and District AIDS Control Committees, the CACC Technical Committee explained their ndings in great detail and called on NACC to take appropriate action against the concerned NGO. They also registered a complaint about the collusion of some NACC o cials from National level with the o cials of the CBO under investigation. They pointed out that they had been reliably informed by the manager of the CBO that 5 senior NACC o cials had visited the CBO under investigation in November 2003 without informing the CACC members. To date however, no action has been taken against the CBO and the concerned individual, now retired and actively involved in local politics.

Abuse of o ce
This study found abuse of position and technical power by o cials. These manifested in a number of ways according to both policy and implementing level key informants across the three constituencies. For example in Mumias constituency, a key informant narrated how a CACC member established a business in town where he developed a typical proposals for HIV/AIDS prevention, duplicated it and sold copies to several CBOs at a fee of between Ksh.3,000 and 6,000 each. He also promised the CBOs that he will ensure their proposals are positively reviewed and funded. Several key informants said that the Chair of the CACC at the time, a highly respected individual in the community was informed about the malpractice and tabled it in the CACC meeting where a resolution to dismiss the affected individual. Accordingly to several policy level key informants within CACC, the decision to dismiss the individual was based on breach of the con ict of interest guidance from NACC (2006). The quote below from a key informant in the same constituency is illustrative: We had a CACC man in town who had proposals duplicated for all the groups, he insisted on all the proposals that he wrote must be passed, the person was in CACC and we sent him away, the church where he used to be active also kicked him out of its projects. The policy level key informant quoted above explained that the activities of the ex-CACC member were detrimental to the funded CBOs because even though the CBOs were funded, these organizations were unable to implement the activities stated in the proposals. Such organizations also found it di cult to write and submit narrative progress and nancial reports to NACC as required and were forced to hire other persons to do this for the organization.

Manipulation of CBOs & NGOs project proposals
The misuse of power and diversion of resources was not con ned to the CACCs alone, this study found collusions between some NACC o cials at the national level, CACCs o cials, some senior civil servants; and some staff of the implementing CBOs. Several major concerns were raised by policy and implementing level key informants across the three constituencies. Two examples are provided to illuminate the main issues: 1. Inaction by NACC towards in uential outside actors who used their strategic positions in collusion with o cials in the governance structures of the program to get large grants in the name of local communities but turned the funds to their personal use; An example of this mechanism of diverting funds by local o cials was reported in Amagoro constituency where a senior o cial in the provincial administration headquarters, in collusion with a senior district level actor, with strong connections with NACC o cials at national level, received a grant of Ksh 2.1 million to implement an orphan support project in the district but which was never implemented. When the proposed bene ciary community under the leadership of their chief got information about the project, they reported the matter to the CACC o ce since they had not been consulted or seen any activities being implemented on the ground.
The CACC technical committee investigated the matter and were satis ed that it was indeed fraud. In their letter (CACC 2003), addressed to the Director of NACC and copied to provincial and District AIDS Control Committees, the CACC Technical Committee explained their ndings in great detail and called on NACC to take appropriate action against the concerned CBO. They also registered a complaint about the collusion of some NACC o cials from National level with the o cials of the CBO under investigation. They pointed out that they had been reliably informed by the manager of the CBO that 5 senior NACC o cials had visited the CBO under investigation in November 2003 without informing the CACC members. To date however, no action has been taken against the NGO and the concerned individual, now retired and actively involved in local politics.

Unethical practices by Consultants and O cials-
According to the implementing key informants, the proposal writing and development phase was the critical stage at which all their projects were de ected. They argued that it was at this stage that priority activities identi ed by the CBOs and budgeted were Most policy and implementing level key informants however said that the funds were largely mismanaged and misappropriated at all three levels, NACC, CACC and CBOs. One very well positioned key informant who had the insights of the CACC program in one of the constituencies, argued that: "In the district, the problems of nancial accountability emanated from the department of Social Services and Community development, where, conditions for registration of CBOs were not complied with due to improper incentives given to the concerned o cials by prospective CBOs. Secondly, the agreement between NACC and the CBOs required certi cation from commercial banks that the CBOs should have held accounts with the respective bank for at least 6 months preceding the certi cation, but this too was outed by the banks. Lastly, when it came to nancial matters, NACC dealt directly with the CBOs and only involved CACCs when there were problems".

Program Outcomes For Vulnerable Children Affected By Aids
The general ndings of the CACC program shows a shared perception among all key informants interviewed across the three constituencies that the 16,262 orphaned and vulnerable children targeted by NACC (2005) bene ted materially and through access to education and health care.
These ndings raised more questions that required to be answered.
First how many of these children were actually reached by the services or received material bene ts?
Secondly, what nancial resources were allocated to CBOs to provide services to these children? How much of the funds were actually received by the CBOs and reasons for the shortfall if any and Thirdly, on what activities was the money received by CBOs spent to meet the social needs of the Vulnerable children To answer the above questions across the three constituencies was extremely di cult for a number of reasons. First, insu cient detailed, disaggregated and itemized expenditure records, secondly, some of the CBOs were not willing to divulge the details of their nancial records and thirdly, the CACC o ces in the three constituencies did not have such detailed expenditure records of CBOs in their areas.
In order to answer the three questions raised above and to illuminate how the resources made available through the CACC program addressed the social problems among these children, an in-depth analysis of the Matungu CACC was conducted.
Matungu constituency with 3,219 vulnerable children affected by AIDS was chosen for the in-depth analysis for three main reasons.
Firstly, compared to the other two constituencies, it has the largest number of CBOs working on mitigating the impact of AIDS on vulnerable children (18 out of 23 implementing agencies covered in this study), secondly, the CBOs in this constituency had detailed and disaggregated project income and expenditure records which were not available in the other constituencies and in the CACC o ces.
Finally, all the CBOs interviewed were open and willing to share the nancial records with the researcher.
Financial data for the approved grants and the amounts received was collected from the implementing CBOs and veri ed with the records held in the CACC o ce; this is presented in table 13 below. The table also presents the numbers of orphaned children under the care of each of the CBOs in the constituency; investments made by the CBOs as long-term measures to sustaining orphan support and direct material bene ts that the orphans actually received. Having established the number of children that were reached by the CACC program and also the amount of money that reached the CBOs, one question remained un-answered, what activities were funded by the Ksh.4 million received by the CBOs?
In an effort to answer this question, the author was guided by a set of principles and assumptions derived from NACC (2006) guidance to the CACCs; community practices related to community participation in organized events; and by rates for lunch allowance set by CBOs while working in their respective areas which were expected to guide the implementation of the CACC program in Matungu constituency.
The rst of these principles was that the CACCs are mandated by NACC (2006) to provide free technical and management assistance to the CBOs, FBOs and NGOs working in their areas and to be guided by the ethics in the performance of their duties.
Secondly, that strategies adopted in the implementation of program actives should be guided by context speci c practices within the study areas. According to key informant interviews with formal and informal community based leaders in Matungu, Koyonzo and Indangalasia locations in the constituency, the practice within Matungu constituency related to community participation is as follows: During community meetings usually organized through the provincial administration, and typically lasting on average between 2-3 hours, community members attend freely.
During government supported events such as Independence Day celebrations, soft drinks are offered to o cials, community leaders and traditional dancers and other artists in attendance.
During other community based cultural functions such as circumcision, weddings and funeral ceremonies, food and drinks are usually prepared and served to the invited guests.
Thirdly, expenditure records of some of the CBOs complimented by key informant interviews with implementing agency staff indicated that while working in their own communities CBO staff claimed a lunch allowance of between Ksh.50 -Ksh.80 The fourth assumption made is that the funds invested in income generating ventures by the CBOs are assumed to be bene ts accruing to the orphans and vulnerable children because accordingly to all implementing level key informants and informal conversations with community leaders across the constituency, the pro ts from these investments are used to sustain support for social needs of the orphans.

Analysis of funds for CBOs projects
Guided by the above principles, assumptions and with the expenditure information provided by CBOs in Matungu constituency, it was possible to estimate the amount of money that each of the 18 CBOs used in the implementation of three common activities across the constituency. These activities included: launching of the CBOs Community sensitization seminars lasting on average 4 days, range 3-5 days

Presentation of material bene ts to orphans
The costs to the program would be: Based on the above calculations, about Ksh. 1.2 million or 30% of funds received by CBOs in the constituency were used in activities that did not directly bene t the vulnerable children affected by AIDS.
The study also found that the program staff and local administration o cials had instructed CBOs to ensure that the o cials and community members were invited to each community function to witness the distribution of blankets, clothing, beans and maize to caregivers or guardians of the orphaned children.
During this occasion, the orphans were paraded before the o cials and the community and photo-graphs taken of them as evidence that NACC funds were reaching the vulnerable children. Policy level key informants explained and justi ed this activity as promoting accountability of the CBOs and empowering the bene ciary community.

Conclusion
This study has analyzed data collected from interviews with both policy and implementing level key informants at district and community levels across the three constituencies. The information collected has been triangulated with information from secondary sources and with information collected through informal conversations with a range of stakeholders in the three constituencies.
The chapter identi es a range of factors embedded within the project cycle that have affected the implementation and outcomes of the Consistent with the above ndings, the study identi es people in key positions within and outside the CACC program as having captured program resources. This is a new phenomenon that requires being further explored. It was neither identi ed in the literature that was reviewed nor in the conceptual framework for evaluating decentralization.
The study nds weak or lack of accountability by key actors in terms of resource utilization, ful lment of mandated responsibilities and functions at all levels of the program. The lack of accountability seems to be perpetuated by a context characterized by patrimoliasm, nepotism and institutionalized misappropriation of resources. Permission to contact this study was obtained the Kenyan Ministry of Education and the Ministry of Health in the study areas.

CONSENT FOR PUBLICATION
This manuscript does not contain any individual person's data and therefore consent for publication is not required -N/A

AVAILABILITY OF DATA AND MATERIALS
All data generated or analyzed during this study are included in this published article [and its supplementary information les]

COMPETING INTERESTS
"The author declares that he has no competing interests"

FUNDING
The funding for this research was provided by the Women's Health and Action Research Centre (WHARC); Benin City, Edo State, Nigeria. http://wharc.freehosting .net/ email: wharc@hyperia.com The funding body had no role in the design of the study and collection, analysis, interpretation of data and in writing the manuscript.