Contextual norms, politics and power dynamics

Structural forces resulting from colonialism have contributed to wide inequities between regions in both Indonesia and Kenya, yet the policy response to these inequities varied widely (see box 1). Prior to the introduction of devolution reforms, both countries were governed by strong centralised governments, where patronage norms and nepotism were commonplace.34 35 Our findings reveal that in both Indonesia and Kenya, these have continued since devolution, creating a threat to governance mechanisms. We found that nepotism was widely described by almost all respondents in Indonesia, indicating its continuation over time and its embedment since devolution. In Kenya, nepotism was also described, although to a much lesser extent than in Indonesia, highlighting potential opportunity for change. Politicians are often motivated to provide services that will appeal to their electorate, consolidate political support and maximise their voter base in their pursuit of re-election.39 In both countries, the provision of health (and other services) was already politically motivated prior to devolution.40 41 This continued politicisation has in some instances brought positive results towards improving access to health services, with UHC recognised as an ‘electoral asset’ in Indonesia, with popular health schemes felt to contribute towards electoral success.40 Similarly, our findings from Kenya highlight that several years into devolution, some politicians are beginning to appreciate the potential political benefits associated with expanding community health services.

It’s (primary/community health care) also good politically because you have created employment, you have encouraged someone so they show their appreciation where elections come. So politically it’s a bonus. Kenya non-health respondent, male 32 

Unfortunately, rather than proving to be the norm, these positive examples appear to have been exceptions. We found that patronage continues to influence the provision of health services since devolution in both Indonesia and Kenya, contributing to varied levels of service provision within a subnational district or county. In Indonesia, a few respondents including health officials, village heads and kaders felt that the level of development within their village varied widely according to whether the community had supported the elected leader in the previous election or not. This is in keeping with previous findings that local governments were perceived to distribute funds unevenly in favour of subdistricts with a high electoral support rate, in order to consolidate power.42 Similarly in Kenya, we found resources were commonly described by community and county technical actors, as having been channelled to the home area for the governor (elected leader in the county) or another powerful politician, with development investment at times provided according to political affiliation, rather than need. As the World Development Report (2004) identifies, poor citizens often have limited influence with politicians about public services. They may be poorly informed about services and may vote along ethnic lines, leaving the provision of public services susceptible to political patronage, with clinics built-in areas where supporters live.17

For example,the current regent, when we asked for support from him, he preferred to give support (financial assistance) to the areas where he got many votes from the previous election. Indonesia village health official, male 04

In a democracy, political parties tend to move towards the position of the median voter in order to secure (re)election, regardless of the efficiency or equity implications of those decisions.39 This is likely to lead to the neglect and further marginalisation of smaller population groups such as people living in hard-to-reach or sparsely populated nomadic areas, while resources are focused on heavily populated areas where there are a lot of votes.43 Within Indonesia, where experiences with devolution are longer, there has been a reported deterioration in public health services, with reduced access among poorer populations.30 44 The picture within Kenya so far appears mixed, with national government having introduced policy for free maternity services around the same time as devolution reforms were introduced, leading to increased utilisation of services.45 However, success of this policy has been complicated by uncertainty created by devolution and poor planning, with health facilities having been inadequately compensated for lost fees, leading health workers to raise revenue from patients, with implications for the most poor.45 Additionally, in Kenya, we found that politicians frequently gave preference to priorities deemed as popular with the electorate, such as construction of new health facilities or purchase of ambulances.33 These were frequently referred to as ‘high vote’ priorities, meanwhile health promotion or disease prevention activities were not viewed as politically favourable. Now is an opportune time for Kenya to reflect on lessons learnt from other countries with longer experiences with devolution, such as Indonesia, in order to ensure that priorities reflect those that will promote public health, as well as curative services, which are often more politically appealing. The recent identification, as UHC as one of the four central pillars for the new government, provides encouraging evidence for this.

The decisions might be subjected to a lot of political interference… They (politicians) make decisions based on votes, will this get me more votes? So they do not tend to see that when we prevent malaria or prevent diarrhoea that you are going to get votes. Kenya county health respondent, male 02

The influence of power, leadership capacity and values on priority-setting and decision making

In both Indonesia and Kenya, in the absence of a clear process and adequate capacity for determining guiding values and priorities, changing power dynamics following devolution became more important, with priorities being selected in line with the degree of ‘power over’ the decision-making process and values held by key decision-making actors. Capacity for setting priorities was widely discussed within Kenya, although it was less commonly described in Indonesia. This may have been as a consequence of the differing timeframes since the introduction of devolution reforms, since earlier study in Indonesia highlighted inequities in capacity between regions.46 A recent study in Kenya highlighted the vital importance of strong leadership capacity following reforms, providing examples of subcounty level managers having innovated and demonstrated flexibility to ensure essential services continue, despite uncertainty during the period immediately following devolution.47 Capacity of local officials has widely been acknowledged as critical to the success of devolution reforms.19 48–50 In Kenya, we observed differences in the levels of capacity for key decision makers between counties.33 Here the key leaders for health are the county executive committee member for health and the chief officer for health who may or may not have public health (or even health) or budgeting experience. Capacity also varied among other actors involved with decision making, including community members and local politicians to understand health comprehensively (including preventive, promotive, rehabilitative and curative services) and their role for setting health priorities. This limited capacity was felt by national-level respondents to impede wise priority-setting.

… the capacity is not there, their priority is wrong sometimes and so decision-making sometimes gets messed up; we are ending up with decisions that will not necessarily address the health issues that we have. Kenya national respondent, female 12

Across both countries, there was wide variation at subnational level in the values held by actors involved with planning and budgeting for health services. The process of priority-setting is inherently complex and value laden.51 As a result of differing values between actors, gaining a common consensus can be problematic,51 and the process for identifying priorities should therefore involve decisions about which values or principles should dominate.23 In literature from Indonesia, differing values contribute to varied priorities. At times, this was felt to cause leaders to prioritise development programmes deemed most profitable to them, which often led to neglect of health and social development programmes.52 From our study findings in Indonesia, respondents described that the priority-setting values of the subnational leaders in the study location did not lead to prioritisation of maternal and child health, which was felt to lead to limited allocation of funding, compared with other priorities. This contrasted with another area that had similarly high maternal mortality rates but had prioritised maternal health in local policy and budget allocation. These findings are in keeping with previous study in Indonesia.53

I know the government would do budget reducing, but they did too much especially in the budget for health sector… The government doesn’t care about the number of maternal and infant deaths. It looks like they just think that the maternal and infant death is a destiny. Indonesia village midwife, female 15

In Kenya, there was no clear guidance surrounding the role of values in priority-setting. Values varied considerably between counties and between actors within a single county, leading to differing health priorities during health planning and budgeting. Politicians tended to prioritise the need for an accountable rights-based approach, typically emphasising the importance of the public participation meetings to validate priorities directly with community members, while technical actors tended to emphasise the need for effective, efficient and equitable priorities.33

In both Indonesia and Kenya, there were positive examples of priority-setting that have prioritised provision of quality, responsive health services. In Indonesia, published examples of positive deviant cases highlight innovation by district leaders to deliver services responsive to local needs, or which seek to improve both the quality and quantity of services provided.40 54 Meanwhile, in Kenya, selected counties are placing high value on community health services, leading to allocation of county government funding, with one positive deviant county having sought to educate and empower actors across all levels (including community, political and technical actors) to understand health holistically and to actively participate in priority-setting.33 This was felt to have contributed to community identification of community health services as a priority for inclusion within the county health plan and budget. The varied duration of devolution reforms allows us to identify lessons, such as the need for rigorous evaluation and learning between well-performing and poorly performing districts, which was felt to have been a missed opportunity in Indonesia.40 In Kenya, forums such as the county executive committee forum (a platform for county executive committee members from all counties to meet together) provide scope for informal sharing of best practices and has provided an opportunity for county executive committee members with an interest in community health approaches to promote awareness with their counterparts around community health and other critical health issues.

Community accountability and empowerment

Within Indonesia and Kenya, we found that limited actions have been introduced since devolution to promote citizens’ understanding of health or their decision-making role. Across both Indonesia and Kenya, we found a host of barriers to effective citizen participation. While avenues for citizen participation are described in policy for both countries, in practice, a lack of funding and tokenistic implementation hinder opportunity for genuine involvement of the community in identifying priorities. These findings are in keeping with previous studies in Kenya and in other devolved countries, including Tanzania and Philippines, where genuine community participation in health planning at the health facility level was limited.55–57

In Indonesia, community participation is described in policy through Musyawarah Perencanaan dan Pembangunan, also known as musrengbang. These are multisectoral annual community consultative meetings for planning and development, involving community members, local community and religious leaders and representatives of local organisations. However, these meetings are not implemented as intended in many regions.42 Musrengbang are often only partially effective, as power is still retained by the village elite,58 with limited community knowledge or interest to participate.44 Within our study, the community role in priority-setting was not commonly discussed by respondents in Indonesia. Instead respondents viewed community participation as the responsibility of the kader (who attended monthly village women’s meetings) or the puskesmas, rather than the community themselves. At times, this resulted in the village midwife (who represents the puskesmas) being blamed by the village head for health issues occurring within the catchment community.

If the puskesmas staff gave them all information about the jamkesda (District health Insurance) or Jamkesmas (Community health insurance), the village people would not have the wrong perception about that. Indonesia head of village, male 06

In Kenya, a similar finding emerged in terms of the role for the community in priority-setting described in policy, compared with the meaningfulness of their role in practice. Pre-existing avenues for citizen participation in accountability and decision making include community and health facility management committees that continue with relatively unchanged roles and responsibilities postdevolution. New avenues for engagement of citizens with state actors have been introduced since devolution through public participation fora, whereby county authorities are responsible to facilitate public participation and involvement.59 Barriers to effective citizen participation highlighted through our study include: failure to address barriers to attendance and active participation of hard-to-reach and marginalised groups, such as lack of funding for transport for those in remote areas, dominance of discussion by local ‘elites’, advertisement of meetings at short notice in English language newspapers and a failure to address local patriarchal norms that limit women’s and youth’s active participation in the presence of men/elders. Respondents described that the setting of priorities was often strongly influenced by power plays and local politics, in keeping with previous findings that acknowledged the influence of social hierarchies, social, cultural and religious norms, economic and political divisions and power asymmetries on participation in health facility management committees.60 61 Many of these barriers to citizen participation in selecting priorities have previously been described in Indonesia.42

Illiterate ones are never involved. They will not even know some of these processes [public participation]. Kenya county non-health respondent, male 19

Community health workers, positioned at the interface between the community and the health system, are potentially key actors to promote community empowerment and engagement with the decision-making process as culturally and socially embedded members of their community.62 63 First, however, they themselves must be empowered, since they live, work and experience the same social norms and discriminations as the citizens within their community.64 They will need the support of both their community and the health system in order to perform this unique role.65

Limitations

We used qualitative methods with a selected population making our findings difficult to generalise, quantify and rank in importance; instead our findings provide in-depth insights to the perspectives and experiences of stakeholders at different levels of the health system. There were a number of methodological challenges within our study that we carefully considered during the joint comparative analytical process. First, within Kenya, priority-setting was viewed from a range of respondent perceptions, with a strong emphasis on the subnational county level. Meanwhile in Indonesia respondent perceptions were primarily sought from community and CTC provider levels. Second, the study was carried out within one district in Indonesia, which was purposively selected for its poor maternal health indicators. By contrast, in Kenya, 10 counties were selected purposively to reflect a range of health indicators. Third, the topic guides used in each country were different, although they did explore similar issues. Fourthly, the devolution processes were carried out over different time periods: with Kenya in the immediate period postdevolution, while Indonesia is 15 years postdevolution. This brought challenges to the direct comparative analysis and also enabled us to review the impact of devolution from longer and shorter term analytical lenses. Despite these limitations and differences, what is interesting is the strong commonality between contexts and respondents across both countries.