On Tensions and Opportunities: Building Partnerships Between Government and Sex Worker-Led Organizations in Kenya in the Fight Against HIV/AIDS

This research reveals how the experienced “tokenism” by sex workers’ representatives who take part in HIV prevention and care partnerships hinges on a lack of expertise by “technical experts” to make use of the different types of knowledge brought to the policy table by them. The article further explores other tensions and opportunities within HIV prevention and care partnerships in Kenya, including the effect of criminalization and devolution on the partnerships. Recent strategic frameworks developed by the government explicitly mention the need for horizontal partnerships between sex workers and government as a crucial step to achieve a more unified and effective response to HIV/AIDS. In addition, during several conversations, government representatives also maintained that more horizontal partnerships can only be achieved through community participation. They defined this as taking sex workers as equal partners in policy development and program implementations. In practice, however, such partnerships have yet to become fully established. All this leads to the main question: where and why do gaps exist between policy visions and actual practices in HIV prevention and care partnerships between government and sex worker-led organizations in Kenya?


Introduction
In Kenya, and worldwide, community participation of marginalized groups such as sex workers is widely considered key to achieving horizontal partnerships in HIV/ AIDS policy development, prevention, and care. Yet, in most settings, this is still more policy than practice. Interactions between government and sex worker-led organizations in Kenya are riddled with diverging aspirations and ensuing contradictions and tensions, not least because the same government that works together with sex worker-led organizations also criminalizes sex work. With regard to the topic of community participation, the position of sex workers in Kenya draws our attention to participation by a Bcommunity^1 that is not only highly marginalized but also criminalized. Looking at community participation with regard to sex worker-led organizations thus presents us with a unique opportunity to analyze practices of community participation by and between highly ambivalent partners, i.e., between Bcriminalizer( county/government) and Bcriminalized^(sex workers). This article explores the tensions and opportunities within daily practices of community participation in HIV prevention and care partnerships in Kenya to provide new avenues to organizing inclusivity in practice. 1 The term Bcommunity^is an emic term used by both sex workers and policymakers in Kenya to refer to a group of people that allegedly shares a set of interests in a particular space and time, and which does not have to be linked to a particular Bplace^(see Bhattacharyya, 2004).
Following the international trend of government mainstreaming of HIV prevention and care (e.g., Gupta, Parkhurst, Ogden, Aggleton, & Mahal, 2008) and the key role of community participation therein, the article focuses particularly on government engagements with sex worker-led organizations in Kenya, and as such only discuss the role of NGOs and donors in passing (see Hearn, 1998 for more on NGOs). Recent strategic frameworks by the Kenyan government explicitly mention that horizontal partnerships between them are a crucial step in achieving a more unified and effective response to HIV/AIDS (see NACC, 2014a). In addition, during several interviews, government representatives also maintained that more horizontal partnerships could only be achieved through community participation. They defined this as taking sex workers as equal partners in policy development and program implementations. In practice, however, such partnerships have yet to become fully established. All this leads to our main question: why do gaps exist between policy visions and actual practices in HIV/AIDS prevention and care partnerships between government and sex worker-led organizations in Kenya?

Literature Review: Community Participation
Community participation came up as a new direction in Bplanned development^ (Li, 2007 p. 15) as part of the Bchange from below^movement, which can be traced back to Paolo Freire's Pedagogy of The Oppressed (Freire, 1970). One of the ways in which such change was to be realized hinged on including community members in previously Btop-downd ecision-making processes regarding their development, thus rendering these more Bbottom-up.^This notion of community participation gained currency within planned development discourse following the political push for more community ownership, which was backed by a growing field of community development theory (e.g., Turner, 2009;Bhattacharyya, 2004). The mainstay of these frameworks holds that community participation is not only in line with good democratic political practice; it will also translate to improved efficacy of planned development interventions. The participatory ideal within planned development resonates with the turn to Bcitizen participation^in democratic theory and practice (see Pateman 1970). The governmentality of both regimes exposes similar contradictions because participation is a political project, and the technologies of community (building) crucial to participation Bwork through, not against, the subjectivity of the poor^ (Cruikshank, 1999 p.73).
Parallel to the elaboration of community participation in development studies and practices, sex work studies and sex worker-led organizations endorse these same ideas which are articulated for example in the global sex workers movement (Network of Sex Workers Projects-NWSP) under the mottos Bnothing for us without us^and Bnothing about us without us( see http://www.nswp.org). Moreover, Wagenaar (Wagenaar, 2017) talks of Bcollaborative governance^as the most Beffective and decent way to regulate prostitution^(p. 43). However, the rationalities and technologies that underlie and give form to sex worker participation in broader policy frameworks (such as HIV/AIDS prevention and care partnerships) are to a great measure based on specific eligibility standards of program impact set by strategic actors such as government organizations (see Rose, 1999). The contractions that follow from this are at the heart of the analysis in this article.
Vigorous debates on what the notion of community participation exactly entails and what it excludes continue to this day (see MacKinnon, 2011;Cooke & Kothari, 2001;Chillag et al., 2002). The main bone of contention in such discussions is the location of responsibility for societal change. Community participation is generally not only seen as a means to producing horizontal partnerships, as mentioned above, but also as a vehicle to bring about large-scale societal change. Considerations of the latter bring to the fore the complexities of power relationships and draw attention to two inter-related issues that underlie most debates on community participation. The first relates to the question whether the responsibility for change lies with government or with citizens. The second is tied to organizational and political power and readiness to bring about change on a societal level. Hence, conceptualizations of community participation differ from focusing on the individual as the locus of change on the one hand to identifying structural and organizational forces as crucial sites of change on the other. The former is sometimes described as Bblame the victim^-such as in neo-liberal discourses of planned development. In its most extreme version, individuals (agents) are considered responsible for their own development. The latter is sometimes referred to as Bblame the system^-such as in planned development approaches informed by Marxism which locate responsibility for development at the level of structural forces (Tesoriero, 2010 p. 55-57). BBlaming^in this vein alludes to the location of responsibility of marginalization (i.e., exclusion) and hence the purported site of solutions (i.e., inclusion). These opposites correspond with the academic debate on structure and agency, and point at different understandings of both the causes of social injustice and the realms of (potential) transformation (see Tesoriero, 2010). Accordingly, they envision different sites and trajectories toward social change and therefore inform highly divergent practices of community participation. Most present-day interpretations of community participation in planned development discourses, however, carry a bit of both, even if the neo-liberal paradigm has decidedly taken the upper hand in these frameworks over the past few decades (Hearn, 2007).
The human rights-based approach is considered an approach to community participation which bridges said extremes. It is often described as giving back responsibility to government to engender change on a societal level while recognizing the agency of citizens to be engaged in providing directions for social justice and holding governments accountable (see Green, 2012;Rawsthorne & Howard, 2011). This approach is the most relevant to this discussion. Most sex workers-led organizations base their activities, to a greater degree, on the idea that sex workers have human rights and thus need to be involved in ensuring that their rights are protected and upheld by the duty bearers of society (Decker et al., 2015). Yet tensions also exist within the human rights framework itself. Despite its potential to include both the structural and individual levels of social change in one analytical framework on community participation, the human rights approach does not solve the inherent paradox the term participation entails, namely participating in existing power structures that oppress. The inherent tension within the human rights framework leads to various gaps between policy and practice that are explored in the sections below, but first the research methods are presented briefly.

Methods
Between 2014 and 2017, the VU University (in Amsterdam) worked together with several Kenyan partners in a research project on the economic empowerment and political positioning of sex workers in Kenya. 2 These partners were (1) Healthy Options for Young Man living with HIV/AIDS and other STDs (HOYMAS), (2) The Kenya Sex Workers Alliance (KESWA), (3) The Sex Worker Outreach Program (SWOP), (4) the International Centre for Reproductive Health in Kenya (ICRH-Kenya), and (5) AIDSfonds. In its entirety, the research made use of several different methods, including a survey under 450 male and female sex workers, participant observation in the organizations as well as meetings which were attended, and economic diaries, and weekly interviews based on these diaries, with 55 male and female sex workers.
This article focuses on only one of the research objectives geared toward analyzing the obstacles, enabling conditions and moments of dialog between strategic actors (in this case government organizations) and sex workers (in this case sex worker-led organizations) in developing horizontal partnerships in HIV/AIDS prevention and care to contribute to their political inclusion. To do this, first a 6-month policy and legislation study was conducted to explore the national context and the various forms of exclusion and stigmatization produced and implemented in national legislations, policies, and practices. This was followed by a mapping exercise of the broader stakeholder network, i.e., identifying all the actors, policies, and narratives at national and local levels that influence the social, economic, and political empowerment of sex workers. In connection to this, Van Stapele was present at meetings between government and sex worker representatives for a period of 6 months. Thirdly, 2-h semi-structured interviews were conducted with 20 government and NGO officials and 20 sex worker activists on their interactions and relationships within the context of collectively combating HIV/AIDS. In its totality, this generated a multi-dimensional understanding of the intentions, influences, and resources they performed, which influence decision-making and implementation processes.
As noted above, the sex workers' organizations that participated in the study were also research consortium partners. These two organizations participated from the moment of design until the moment of data analysis and knowledge dissemination. This explains why the authors were able to develop relationships of trust with them from the onset of the research project. Van Stapele also conducted long-term ethnographic research with members from these two organizations, which further strengthened the research relationship. 3 All partners involved carefully evaluated every step in the research process to maximize inclusivity, foster relationships of trust, and increase the relevance of the project for sex workers during implementation.

HIV/AIDS Prevention and Care Partnerships and Sex Workers
HIV/AIDS prevention and care partnerships in Kenya are initiated by the government and combine a great number of different actors, and since recent years also include sex workerled organizations. At present, the main partners that aim to work together in fighting HIV/AIDS among sex workers in Kenya can be divided into five levels of organization: The five levels are ranked from top (1) to bottom (5), following the traditional hierarchy of decision-making, professional status, and, especially, flow of funds between the different levels. This order is, of course, a simplification because in reality international donors and government agencies also fund certain NGOs directly, and in some cases even have direct funding links to sex worker-led organizations. Also, government organizations follow general government decrees, based on internal decision-making and funding structures. This hierarchy, however, does highlight the dominant way international donors, Kenyan government agencies, and NGOs engage with sex worker-led organizations in the fight against HIV/AIDS. It illustrates the position of the latter as occupying the bottom tier of what both government officials and sex workers described in the interviews as a Btop-downâ pproach in policy development, implementation of interventions, and allocation of funds in HIV/AIDS prevention and care in Kenya. Government officials identified the organizations at the top as well organized and those at the bottom as less organized.
The present-day strategic coalitions in fighting HIV/AIDS and the ensuing order of organizations are both rather recent phenomena. The government initiated these partnerships from 1999 onwards, whereas sex worker-led organizations were only asked by the government to join a few years back. Nowadays, the Kenyan government plays a central role in HIV/AIDS prevention, but this was not always the case. The Kenyan government under Moi (1978Moi ( -2002 responded very late to the AIDS crisis, which led to a dominance of NGOs in service provision within the HIV/AIDS sector until the late 1990s. By the close of the millennium, HIV infections had reached unprecedented levels in Kenya. This finally prompted the Moi government to declare AIDS a national disaster and establish the NACC to coordinate the national response. For a long time after this watershed moment, NGOs continued to be on the foreground in program development and implementation. Initially, international donors were hesitant to finance government programs and carried on channeling most of their funds through NGOs. Donors feared a disruption in the development and delivery of effective HIV prevention and care services. NGOs had the required expertise and local infrastructures the state lacked at the time. Hence, the two government agencies NACC and NASCOP first took up strategic coordination and facilitation and stayed away from implementation (Hershey, 2013).
Yet, over time, shifting donor attention enabled the state to gradually take center stage. This aligned with the global ambition of mainstreaming HIV prevention and care. With donor support, the two said agencies developed the necessary apparatus and gradually became the executive lead in HIV prevention and care. However, the Kenyan government continued to recognize the contribution of NGOs and still depended heavily on their capacities to roll out nationwide programs. Accordingly, donors divided their funds between both government and NGOs, which solidified partnerships between them. Government officials stated that community participation was absent in the first decade of government programming. From 2009 onwards, donor focus gradually shifted even further to the state. This shift is aptly illustrated by the Global Fund's requirement to allocate the majority of funds through state organs (interview with UNAIDS, 13 January 2015). International donors, government, and NGOs still continue their close to two decades of partnerships in HIV prevention and care to this date, but the pith of the coalition is now firmly located within the realm of the Kenyan state.
Currently, the AIDS response in Kenya is implemented through periodic strategic plans and different coalitions of international and local partners (see Hershey, 2013;Okal et al., 2009). Over the past decade, these strategic partnerships have been relatively effective. HIV prevalence and the rate of new HIV infections in Kenya are lower now than at any earlier point since the onset of the epidemic in the early 1980s. AIDS, however, remains a national priority for decades to come (NACC, 2012). Recent strategic frameworks continue the country's commitment to universal access to HIV prevention, treatment, care, and support. It emphasizes the need for an improved health sector service delivery, for the sectorial mainstreaming of the HIV response, support for communitybased efforts, and for enhanced strategic information (NACC, 2014a;b). NACC is responsible for developing strategic frameworks and for setting periodic targets. NASCOP facilitates the strategic partnerships for program implementation and manages the bulk of HIV prevention services aimed at achieving the targets. Accordingly, NASCOP is mainly involved with technical co-ordination of HIV/AIDS prevention programs in Kenya. Non-state partners are involved at every stage in the process of policy and program development and implementation through the Technical Working Group (TWG). The TWG is a government-initiated network of partner organizations currently including partners from level 3 to 5, and it is the main coalition to combat HIV/AIDS in Kenya.
As noted, sex workers only recently joined the long-forged partnerships between government and NGOs. Around the same time that government became the most significant actor in the coalitions, that is 2009, NACC and NASCOP began to emphasize the need for a more participatory process in policy development and program implementation (see NASCOP, 2010: 24). Following international research (e.g., Wagenaar, 2017) and pressures (e.g., WHO et al., 2013), government policy in Kenya became increasingly based on the premise that involving members of so-called target communities would improve the efficacy of HIV policies and programming all around.
For example, the third strategic plan was the first time that it explicitly states government's intention to work together with sex workers and men who have sex with men (NACC, 2012). Now, different policy frameworks and guidelines state the premise that effective HIV program management, planning, and implementation must be a participatory process (see NACC, 2014a, b). Recent strategic aims of government are to consolidate this by moving away entirely from intervention-driven (i.e., top-down) to key population-driven (i.e., bottom-up) interventions (NACC, 2014a, b). A policymaker at NACC, however, shared during an interview in January 2015: BPolicy-wise, we are moving in the right direction, but it still has to become a reality on the ground.Ô n 12 February 2016, a meeting was organized by the authors with partners in the TWG to discuss the preliminary findings of this research. The meeting took place in Nairobi and was attended by representatives from the government and different local and international NGOs, and sex worker-led organizations and activists. 4 At some point, the discussion turned to the challenges of working together in the TWG. A representative of an LGBTQ+ coalition, which also includes several groups of gay male sex workers, shared that many sex workers who participated in such meetings often felt as if they are not taken seriously as partners. She asked the group if their participation could thus be understood as a Btoken gesture.Ŝ he alluded to the fact that Bcommunity participation^is often a donor requirement (see WHO et al., 2013;UNAIDS, 2012;Cornish, Campbell, Shukla, & Banerji, 2012). A government representative responded that he often did not know how to make use of the information sex workers provided during policy meetings, and then asked the research team to explore this further. Other participants nodded in agreement.
During this meeting with TWG members, a common tendency regarding community participation surfaced and was made visible to all who were present. Representatives of socalled targeted communities are invited to the policy table where their contributions are somehow not included. This is experienced by the community as a form of Btokenism^(e.g., Kanter, 1987). Commonly, community participation in HIV prevention and care partnerships in Kenya goes as follows. The leaders of recognizable sex worker-led organizations are invited by the Kenyan government to take part in existing arrangements (in this case the TWG), and as such the asymmetrical power relationships that bring forth these structures remain intact. The fact that representatives of sex workers are included-as opposed to including themselves-delineates their power in influencing decision-making processes. This shows that questions such as Bwho includes,^Bwho is included,^and Bon whose terms^are crucial in determining practices of community participation, trajectories of envisioned change, and potential outcomes. The potential for social change therefore relies not only on the wish to change by the government but also on the ability of these existing structures to become more inclusive and effective. However, another interesting point came out of this meeting, namely it became clear that government officials involved in the TWG, upon receiving information from Btargeted groups^did not know how to use this information effectively. This tension as well as others will be discussed in the BResults^section below.

Results
The above reveals the incongruity of encounters between government and sex worker representatives in the context of HIV/ AIDS prevention and care partnerships. The study reveals three overarching tensions that contribute to the gaps between policy and practice in the participation of sex worker-led organizations in government initiated partnerships on HIV prevention and care partnerships; these are (1) criminalization, i.e., sex work is still criminalized in Kenya; (2) devolution, a system of government within which the local governments takes center stage in the administration of localities within a nation; and (3) different appraisal of expertise, the way in which the different types of knowledge and organizational experience that partners bring to the policy table are valued.
Criminalization The Penal Code (Laws of Kenya, 2014) is the national legal framework that criminalizes sex work through Bthird party involvement^(i.e., it targets parties which organize or facilitate sex work but not sex workers themselves). This legal framework delineates the legal space within which the country's 47 counties can develop by-laws regarding sex work. In practice, this means sex work is further criminalized through county by-laws targeting individual sex workers and not only brothel owners for instance (see more below).
The first factor is tied to the criminal status of sex work in Kenya. Different laws criminalize sex work in Kenya. As mentioned above, the Penal Code is the national legal framework which does not directly criminalize sex work, but it prohibits the role of third parties in organizing and profiting from sex work. The regulation of sex work on county level differs per county, but in general criminalizes sex work by also targeting individual sex workers and not only third parties. For example, the current Nairobi BCity By-Laws^state under General Nuisance that Bloitering, importuning or attempting to procure a female/male for prostitution purposes^is an offense. Hence, most sex workers are arrested by county police (locally dubbed BKanjo^) and taken to a county court where they are charged on the basis of county by-laws (FIDA, 2008). An interviewed police officer who sometimes participates as a consultant to the TWG stated: You don't break the law [i.e. the Penal Code] when you have sex, the act itself, but when you get money, yes. No one has been charged yet with living of proceeds, it is very hard to prove. Most prostitutes are arrested for loitering for immoral purposes [by county police]. When they are taken to [the county] court, they just say yes and pay 500 Kenyan Shillings. If they say no it means a long court case and a lot of money, which they don't have. It is very illogical because it costs a lot of money to arrest a sex worker. The police officer, the vehicle, the court people, the judge, everything needs money and the court only gets 500 Kenyan Shillings.
For economic purposes alone we should legalize sex work.
Noteworthy, recent strategic plans of NACC 2011 and 2014 the possibility of de-criminalizing sex work to improve access to healthcare for sex workers. However, the plans also state that attempts to de-criminalize sex work have faced significant resistance among different religious and cultural groups. The recognition by NACC of the positive effect decriminalization can have is in line with international tenets in the fight against HIV/AIDS which perceive the illegal status and negative attitudes as obstacles to effective HIV prevention interventions for sex workers. In addition, around the same time, guidelines for peer education among key populations were published which emphasized the importance to Bcreate an enabling environment in which sex workers have access to appropriate, affordable, acceptable and assessable health services without being penalized^ (UNAIDS, 2012;WHO et al., 2013). Similar statements are made in the National Guidelines for HIV/STI programs for sex workers developed by NASCOP (2010). These guidelines express concern about key populations' experience of barriers Bto accessing services because their behaviors are criminalized and stigmatized making them marginalized and hard to reach members of society^, and posits that HIV/STI and other reproductive health interventions should not only be Baccessible but also acceptablet o sex workers (2010 p. 23; Nyblade et al. 2015). The said guidelines even explicitly state that such interventions need to Brespect sex workers' human rights and accord them basic dignity (e.g. services are voluntary)^; moreover, interventions should be based on Bsex workers' views, knowledge and life experiences^(2010 p. 29). Finally, the guidelines recognize the essential role sex workers play as Bpart of the solution( 2010 p. 29).
All government representatives interviewed posited that the human rights concerns and directions for participation of sex workers articulated by NACC and NASCOP stand on tense footing with the existing legal frameworks of national and local governments. Moreover, they also shared that powerful parts within national and local levels of government in Kenya do not acknowledge the human rights perspective as a viable way to approach sex work because it is still criminalized by law. Ensuing tensions fundamentally shape the ambivalence with which these agencies engage with sex workers. Accordingly, one of the main barriers to more equal partnerships between sex worker-led organizations and the government is the division within government on sex work, and the lack of political leverage by NACC and NASCOP to solve ensuing tensions.
The government representatives unanimously declared that the government is stretched between two opposites. One part of the Kenyan government (represented by NACC and NASCOP) is guided by a form of Bhealth pragmatism^and by a push from international donors to become more inclusive and acknowledge sex workers' human rights in the fight against HIV/AIDS (see UNAIDS, 2014). The other part is directed by an alleged moral panic regarding sex work and continuously fight any attempt to decriminalize it. This leads to a split within government whereby it is possible that one arm of the government tries to engage sex worker-led organizations as partners and work together to, for instance, distribute condoms to sex workers, while another arm uses the same condoms as evidence to arrest them (Abdalla, 2015). Several government representatives avowed that the lack of power to solve this internal divide is illustrated by the recurrent emphasis by NACC and NASCOP of the need for more evidence to convince colleagues in other parts of the government of the urgency for decriminalization of sex work. This demand for more data paradoxically does not take into account the already existing body of evidence which illustrates the links between criminalization, stigma, and human rights violations of sex workers (e.g., Bruckert & Hannem, 2013;Harcourt et al., 2010).
However, in the context of this article, it is important to focus on the impact this divide within government has on the partnerships between government organizations such NACC and NASCOP and sex worker-led organizations. The lack of trust between these, frequently expressed by all the sex workers we interviewed, is a direct result from the inability of the two government agencies to solve said tension. When one partner (the government) is responsible for the marginalization, harassment and discrimination of the other, how can one imagine these two partners to build enough trust to achieve a more equal basis for interaction? Indeed, the two government agencies proclaim a human-rights-based approach and favor decriminalization, yet the government as a whole does not, and sex worker-led organizations are acutely aware of this. Consequently, sex worker-led organizations have stated that they harbor low expectations of what they can accomplish in their partnerships with NACC and NASCOP and prefer to strategize with international partners rather than with the national agencies.
Representatives from different sex worker-led organizations have time and again shared that they are aware that their invitation to partner in the TWG is fueled by the growing donor focus on the participation of key population in the fight against HIV/ AIDS. Additionally, they realize that their invitation to participate is conditional and their presence will cease to be welcomed when international pressure no longer demands it. Even if this is an unfair judgment on individual efforts within said agencies, it clearly demonstrates how the leaders perceive the limitations of government efforts to engage with them as partners. Thus, their interactions can be described as a rather Bweak^partnership, meaning that these partnerships are experienced, at least on the side of the sex workers, as highly contingent on external factors such as donor demands rather than on internal motivations. In contrast, focusing on international partnerships enables sex worker-led organizations to validate and expand their efforts and build their capacities in the long run. As a result, leaders often experience these as Bstrong^partnerships. Yet, this also contributes to more tension between the government and sex worker-led organizations for the latter do not primarily depend on the government to provide services to their constituencies. It can be argued that, at present, the government needs the participation of sex worker-led organizations more than the other way around since Bcommunity participation^is a key requirement of most funding NACC and NASCOP receive from international donors. In contrast, participation with government agencies is not a requirement for most funds sex worker-led organizations receive from international NGOs. This, of course, affects the partnership and adds to the complexities of the relationships. It also demonstrates that sex worker-led organizations have some leverage and are not just located at the bottom tier.
An interview with a program manager at UNAIDS pointed out the important role of international donors in this respect.
We make space for dialogue, and lobby to influence policy. In the most recent strategic frameworks, we pushed for the more controversial agenda points. We do what NACC and NASCOP can't do, but with their mandate. When we push it, it adds considerable weight to an agenda, and government can't ignore us. We use our power to help NACC and NASCOP, and local NGOs, to move a more progressive agenda. We push for change, and provide tools for advocacy, that is our mandate. We have very little money, we are more policy oriented. But, we also initiate joint fund raising, like the Global Fund proposal, we made sure key populations are involved as partners, we make sure they are recognized and included in the targets as key populations.
In the above excerpt, this representative posits that UNAIDS helps NACC and NASCOP to resolve the tensions within government regarding sex work. In this vein, UNAIDS does not only push for a more progressive agenda on behalf of key populations but also on behalf of these government agencies that do not have the same leverage but do share such policy visions. This also, again, reveals that said agencies experience considerable headwind within government, which hinders them to follow through on their policy intentions, a situation that would grow more acute were it not for the powerful back-up of international partners.
However, these already rather temporal solutions forged through the mediation of international partners are severely backtracked by another development. All the complexities involving the realization of more horizontal partnerships between government and sex worker-led organizations were decidedly augmented when devolution made its entree.

Devolution
Government and sex worker representatives all pointed at devolution as the second factor contributing to tensions between policies. Devolution denotes the decentralization of government, which commenced in Kenya after the promulgation of the new constitution in 2010 (Greste, 2010). Most prominent in the new constitution's dispensations was the objective of devolution, that is, the transfer of some of the central government functions to the 47 county governments with the aim of ensuring equality and equity in service deliveries. By-laws inherited from the former local government's municipal councils were still enforced during the period research was conducted, which partly coincided with the transition period (2013)(2014)(2015). However, The Transition to Devolved Government Act () gives the county governments powers to also develop their own by-laws as long as these are in line with the national legal frameworks. Interestingly, the ratification of the new constitution coincided with NACC and NASCOP's ambition to include representatives of key populations as participants in HIV prevention and care partnerships. This is not a coincidence because citizens' participation in the governance of their affairs is central to devolution-as per articles 10(2), 69(1) (d), 174(c), and 184(1) (c) of the constitution of Kenya. The County Government Act sections 99 to 101 also stress the importance of civic education aimed at creating an informed citizenry that would actively participate in governing society (Republic of Kenya 2013). Similar to the aforementioned government agencies, local governments also struggle with the execution of Bpublic participation,^especially where it concerns marginalized communities.
What is more, devolution made the still relatively new and rather ambivalent relationships between the government and sex worker-led organizations decidedly more difficult. On the one hand, public participation was considered central to a devolved system of government, yet on the other, counties were also granted power to decide on the nature of such partnerships. In the past few years, county governments have gradually taken over national HIV services. To enable a smooth transition, NASCOP has seconded its staff to the relevant county government programs (referred to as County AIDS/STI Coordinating Programs-CASCOPs) to oversee the continuity of services during transition. A policy-maker at NACC related that, despite time passed, the current challenge remains encouraging county authorities to take on the responsibility of developing and implementing HIV prevention strategies together with key populations (Interview 12 January 2015). This is trickier than it appears at first because counties have the mandate to develop their own budgetary priorities and can use financial arguments to cease partnerships and programs. To illustrate this, the NACC representative stated: Male sex workers are even more vulnerable than female sex workers. They are even more mobile, they live nomadic lifestyles. That mobility and also the clandestine lifestyle adds to their weak social positions. It will be very difficult for them to operate a business. That stigma in society is very high, for example I was in Kilifi (a county in Kenya) to discuss devolution of our programs and the Muslim clerics said we can discuss [female sex workers], but [male sex workers]? No, there are no [male sex workers]. Kilifi! You know how many there are? With women, they are against divorce so they feel pity for women who have to do sex work after divorce. In Islam, men have to provide for their wives for life. So, they see it as their duty even to help these women, who lack life skills because they have been married young, but men? No.
To counter this, NACC has teamed up with national and local sex worker-led organizations, among others, to (again) compile (more) evidence and convince reluctant counties of the urgency of and lobby for the inclusion of all sex workers in strategic partnerships. The policy-maker at NACC explained that there are still many roadblocks that have to be tackled before partnerships between sex workers and government are realized on a county level. Alongside complicating decision-making structures and money flows, the above quote reveals that devolution is also problematic because there is more space for county governments to exclude sex workers from participation as many are guided by an alleged moral panic of (particular forms of) sex work. The tenet of devolution, namely public participation, should provide ample opportunities to local sex worker-led organizations to forge partnerships with county governments within the context of CASCOP, yet the lack of political will in various counties complicates this. In this context, NACC and NASCOP play a similar role like the international partner and mediate between county governments and sex worker-led organization. However, the uncertainty of money flows between government and counties cuts deep into their leverage.

Different appraisals of expertise
In the above, we have looked at the way criminalization and devolution affect the goal of the government to achieve more horizontal partnerships with sex worker-led organizations. The criminal status of and alleged moral panic concerning sex work are the more obvious tensions that impact policy ambitions on national and county levels, thus preventing the acceptance of sex workers as equal partners, the latter with great risks on cooperation for change. One aspect involved in all this merits further elaboration. Most sex worker activists that were interviewed for this research expressed concern that their expertise was undervalued and as such underused by other TWG members. The observations done by van Stapele at meetings corroborated their sentiments, as she observed and wrote in her field notes the many different moments their contributions were ignored.
As relatively recent members of the TWG, sex worker-led organizations are included to assist government bodies and NGOs in outreach work concerning more effective distribution, prevention, monitoring, and forecasting of medical supplies and safe sex items. Also, sex worker-led organizations are invited to take part in the execution of program activities.
As of yet, this has, however, not led to more equal partnerships in terms of decision-making regarding policy and programming. Sex worker representatives shared that while they are present at the table, they nonetheless often feel as if they do not contribute to the process in any significant way. One even described it as: BIt is like they [government representatives] don't even hear us^(interview with a sex worker representative, 6 August 2015).
Sex workers bring to the table an enormous amount of experiential knowledge. Moreover, these organizations have a wide and immediate reach among their peers and the ability to improvise as changing situations emerge and unfold. For example, if a particular place of work (for example a bar) is abandoned due to a risk of police violence, the peer educators of sex worker-led organizations will immediately know where it has moved to and change their outreach plan accordingly. Also, sex worker-led organizations can mobilize members quickly to address urgent matters. On 12 October 2015, a fifth sex worker in Nakuru was murdered within a short time span. It took the sex worker-led organizations in Nairobi, which is situated about 3 h by public transport from Nakuru, less than 4 h to mobilize a large crowd from Nairobi to demonstrate together with sex workers in Nakuru and demand proper investigations from the local police (personal observation by Van Stapele, 12 October 2015). This type of knowledge and reach, coupled with flexibility and commitment, are key features of sex worker-led organizations. In other words, sex worker-led organizations have a vast network of peers and relatively up-to-date knowledge of everyday experiences, whereabouts, and needs of fellow sex workers. This also implies that they have their own practice-based and Badhocratic ( Dunn, 2012) ways of organizing that sometimes clash with the bureaucratic or professional forms of organizing by government and NGOs, respectively.
In bureaucratic organizations such as government-based organizations, there are high levels of formalization as well of standardization of tasks and skills, including expectations of how others should Borganize^and what kind of knowledge counts. The same goes for professional organizations such as NGOs. Most government and NGO policy officials in the health sector are accustomed to ground policy and program development on quantitative data-such as derived from surveys. The taken-for-grantedness of a Bformal view^on knowledge and policy development prevents partners from making use of other forms of knowledge and knowledge development, especially experience-based and ad hoc knowledge. Policy professionals in this sense are accustomed to and thus have an inclination to use Brationally ordered^informationas deducted from surveys and graphs-and lack skills to make use of life histories and other forms of qualitative data. They might not even recognize the latter as knowledge. This greatly complicates collaboration between government and sex worker-led organizations, as the next example illustrates.
On 20 January 2015, van Stapele observed a meeting where TWG members participated in an exercise facilitated by the University of California San Francisco to evaluate quantitative data and identify possible policy priority areas. Two factors stood out during this seminar that help understand what stands in the way of attaining more horizontal partnerships. First, the representatives of the different TWG actors participating in the discussion did not seem to share a common language. Language here means the same jargon that was used in discussing data and ensuing implications for policy development. Most of the sex workers participating in the discussions did not understand some of the technical terminology used, which excluded them from certain parts of the debates. Second, the contributions by the sex workers were not taken up in the final conclusions. During the debates, several sex workers voiced their rejection of particular conclusions that were drawn on the basis of the quantitative data. They supported their position by narrating personal experiences that illustrated their arguments, which were dismissed by a few other participants as Bsubjective.^It was thus clear that the manner in which the sex workers supported their objections clashed with the presentation of the conclusions based on the quantitative data, whether this was the reason their comments were ignored can only be speculated. Some sex workers left the meeting earlier, and one later explained that she did not feel Bappreciated^and was angry that the other TWG members continued with conclusions she deemed highly problematic based on her own experience as a sex worker and that of her friends.
The meeting described in the above reveals that there does not seem to be any room for negotiation regarding how to include different types of knowledge, each valuable in their own ways. Instead, formalized knowledge takes a hegemonic position. As brought out by the government representative during the February 2016 meeting, discussed earlier, he and other of the government policy-makers participating in the TWG lack the expertise sex workers have and do not know how to use it. Concurrently, many sex workers lack the technical competences needed to evaluate quantitative data and translate these to policies or patterns of lived experiences. Yet, the organizational power dynamics underlying these partnerships favor the latter skill-sets and are thus more exclusive to the sex workers than they are to the other TWG members. This analysis of different types of knowledge is not usually an issue that is consciously reflected upon in studies concerning community-based participation. However, this example revealed how these tensions reinforce unequal power relations between the actors and infer ways to making such spaces more inclusive and thus more effective.
What needs to be explored then is the receptivity and capability of dominant organizations (such as the government and NGOs) to include community representatives to not only take part in discussions but also, and more importantly, to initiate agenda setting and genuinely influence decision-making. If not, any potential for achieving more horizontal partnerships is at risk and sex worker representatives continue to participate on the terms of the more powerful (see Oloka-Onyango & Tamale, 1995).

Discussion
The term Bopportunities^mentioned in the title points to the fact that a highly marginalized and criminalized group of citizens, namely sex workers, are sitting at the government policy table together with NGOs. These representative organizations not only help in developing policies, they also partner with the government and NGOs in the implementation of various health interventions, and they receive government funds to reach out to their peers and fight stigma and violence. These are huge achievements in any situation, but especially when considered in the Kenyan context where most marginalized and criminalized groups are excluded from such opportunities. However, this article also reveals that these partnerships are riddled with tensions and contradictions, which hinder the widely held ambition to make these partnerships more horizontal and thus more bottom-up and effective.
This article has given specific insight into the dynamics involved in community participation of sex workers in governmental policy concerning HIV/AIDS prevention. It questions homogenous notions of the concept community participation and in doing so reveals that while community participation is a political ideal in the sex worker rights movement, it is nonetheless contextually defined. In Kenya, this means that who is involved and what it entails is still predominantly determined by governmental agencies.
Consequently, three tensions arise in the everyday practices of community participation. In the first place, a schism exists within the government, which can be acknowledged as a cause of mutual mistrust at the policy-making table. On the one hand, legislation criminalizes sex work. On the other hand, the government is obliged by international donors to include sex workers in policy-making regarding HIV/AIDs. Although international organizations such as UNAIDS prove to be crucial in bolstering NACC and NASCOP to stand their ground in relationship to the more hostile attitude toward sex work harbored by dominant sections in the national government, this schism remains intact. Secondly, the processes of devolution taking place on a national level have diluted some of the hard-fought efforts made by NACC and NASCOP in their interaction with sex worker-led organizations. These agencies have less control now since counties decide on their own policy priorities and budget choices, and they lack the leverage enjoyed by international organization to push for partnerships with sex workers on county levels, although to a certain extent they do play a similar role in this situation. While this development is relatively new and the outcomes are yet to be seen, all TWG members are committed to persuading counties to engage with sex workers. However, these tensions involved between the two governmental levels divert attention away from implementing the proposed policies because strategic actions taken to persuade organizations at county level consume a lot of time and energy. Hopefully, the national TWG will be able to help these localized partnerships to at least forego some of its earlier problems and as such win back a bit of time lost. Finally, this article demonstrates that it is not only the power configurations and the concomitant decision-making processes that deter achieving horizontal community participation but also the exclusion of alternative types of knowledge, more specifically experiential knowledge, as a source of evidence to create policy. Moreover, the lack of professional skills to understand and make use of each other's expertise puts said partnerships at great risk. As such, this should best be addressed upfront, that is before or in the early stages of developing HIV prevention and care partnerships.
All this leads to the conclusion that there is an imminent need for a discussion among the strategic actors in the TWGand similar networks-on horizontal partnerships and the role of community participation therein. This article showed that partners have quite diverging assumptions of what horizontal partnerships entail, even in situations where they seem to depart from similar policy frameworks. Making explicit each actor's role, responsibility and expertise-and how these all work together in achieving greater efficacy in health service delivery for key populations-would alleviate some of the tensions explored in the above. Yet, horizontal partnerships as a key goal to achieving equality will continue to stagnate if the Kenyan government, be it national or local, solely and exclusively engages with sex workers from a health perspective. Governmental bodies which uphold and implement the laws which criminalize sex work activities make it impossible to approach sex work from a perspective which recognizes the human and labor rights of sex workers. Horizontal partnerships between government and sex work communities will remain perpetually postponed if the schism that exists between the health perspective and the criminalization of sex work remains unresolved.
In addition to matters of responsibility for societal change, the tensions within community participation that are highlighted in the above draw attention to issues of potentiality and conditionality. Important questions to consider are: what potential do different types of knowledge contribute to the process? And, what conditions need to be in place for these knowledges to be fully taken up as part of the process? These build additional dimensions to responsibility and power that are crucial when discussing community participation, especially when it concerns groups that are exceptionally vulnerable and which potentially suffer even more from Btokenism.^Including potentiality and conditionality in our analysis of community participation allows for more nuances in pinpointing the kind of responsibilities each actor potentially can have. This also offers directions on how collaborations between politically diverging actors are to be encouraged, consolidated, and innovated through dialog and common language. Such encounters also engender a better understanding of how community engagement affects and upholds both Bsystem^and Bvictim^and may thus inform new modes of community participation that not only (potentially) unsettle such binaries, but which also open up new avenues for inclusive partnership based on local (as opposed to governmental) technologies and logics.
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.