Power dynamics and new directions in the recent evolution of CBNRM in Botswana

Recently, the limited control of Botswana's community conservation organizations, or trusts, over resources has been further eroded. Community trusts now exist solely to disburse funds allocated by central government. The absence of any rights to control access to or use of their resources suggests the complete collapse of Botswana's original community‐based natural resources management (CBNRM) model. This collapse could facilitate fresh approaches that return to the original intentions currently lost behind the CBNRM acronym. Access and control can be more clearly framed through broader environmental stewardship frameworks that include moving below the homogenized community identity to specific resource user groups; broadening the conservation focus beyond monetized wildlife to include other resources and ecosystem processes; and establishing formal conservation agreements separately for specific resources, with the trust coordinating adherence to the rights and responsibilities held by individual users on the one hand, and government or public agencies on the other.


| THE CURRENT STATE OF CBNRM GOVERNANCE IN BOTSWANA
Community-based natural resources management (CBNRM) has been a fixture on the southern African conservation landscape for the past 25-30 years (Dressler et al., 2010). Underpinning the introduction of local-level conservation and development programs in the region, was a growing awareness that the alienation of rural communities through the formation of exclusionary protected areas and state control of resources-particularly wildlife-was creating resentment and resistance among local communities (Ndumeya, 2019;Nelson, 2010;Spinage, 1996). CBNRM was built on the premise that in order for rural communities to carry the costs of living with wildlife, they needed to benefit from it (Getz et al., 1999). This premise highlights the degree to which CBNRM was externally imposed and driven by conservation scientists, who were focusing on conservation as the ultimate goal, in juxtaposition to the primary expectations of the target communities, who saw CBNRM as a means to increased development (Agrawal & Gibson, 1999;Swatuk, 2005). These differing motivations are important in understanding why CBNRM in southern Africa has not been implemented in the way it was originally intended.
In Botswana, CBNRM was originally systematized through a community forming a trust to engage in joint venture partnerships (JVPs) on behalf of the community (Cassidy, 2000). The focus was initially on those communities located in wildlife management areas. Such communities are allocated a use-right head lease over a demarcated concession area, which only gives them the rights to commercial wildlife-based tourism in that area. The lease is known as a head lease because most communities sub-lease the tourism rights in the area to a JVP tour operator. The community collects lease fees in exchange for the JVP taking over the photographic or hunting (as determined by central government) rights for that area. This model precludes any true devolution of decision-making to the community level; resource use decisions are made at central government level, and then handed down to communities to implement (Blaikie, 2006;Rihoy & Maguranyanga, 2010). The only rights communities have had are: to choose whether to use their designated hunting quota or not; to have final say over which tour operator they would sub-lease to; and to decide how to spend the income received from their JVP.
In recent years, under a series of ministerial directives, there has been further, steady erosion of benefits at the community level. First, through a 2001 directive, JVP lease fees and other trust revenues were no longer being paid to communities, but kept in government coffers to disburse to communities if and when they "behaved" (Rihoy & Maguranyanga, 2010, see also Büscher & Dietz, 2005). Next, the head lease was taken away from the communities, and put towards a "Land Bank" controlled by Botswana Tourism Organisation (BTO) (Khama, 2017). This was done despite BTO's legal mandate being limited to marketing and promoting tourism in Botswana, and without involvement of the land boards-the legal authorities over communal lands. Through the Land Bank, tour operators were given the head lease over areas previously allocated to communities, and were requested by BTO to pay the community a royalty fee relating to a percentage of their operations, in recognition of that community's association with the area. According to the CEO of a leading Botswana conservation NGO, the leases needed to be taken away from communities because "these communities are undermining tourism, and we have to protect the national interest." The statement is of concern because it ignores: that these communities are themselves part of the nation and its interests; that poverty levels in the district with the most wildlife-based tourism and richest tour operators are among the highest in Botswana (Lepper & Schroenn Goebel, 2010;Mbaiwa, 2017); and the hardlearned lessons of the 1980s that exclusion and uneven development undermine conservation or encourages anti-wildlife activity (Brennan & Kalsi, 2015;Meer & Schnurr, 2013;Ndumeya, 2019;Spinage, 1996).
The "national interest" story, where the economic benefits accruing at national level are valued as more important than the developmental needs of districts or communities, is an important one for Botswana (DeMotts & Hoon, 2012;Hoon, 2014), which has a tiny population (~2.3 million), and for whom a sense of fairness has been heightened due to its emergence as a nation during neighboring South Africa's apartheid era. The ideology that all citizens are part of one "super-tribe" (Cassidy, 2000) has led to the development of a highly centralized, complex, hierarchical bureaucracy that is unwilling to let go and allow regional variation in development paths (Mooketsane, Bodilenyane, & Motshekgwa, 2017;Poteete & Ribot, 2011). Consequently, most of the country's policies have a one-size-fits-all approach, and neither ecological variations in the landscape nor ethnic variations in relation to that landscape are accommodated, with critical ramifications for CBNRM (Nelson, 2010).
Late 2019 saw the election of a new president, HE Mokweetsi Masisi, who promises to return and even strengthen community rights and their participation in wildlife-based tourism (APA Gaborone, 2019). Already the Land Bank looks to be overturned (Keakebetswe, 2019), although private investors still hold head leases over some areas formerly allocated to community trusts. A new Minister of Environment, Natural Resources and Tourism has been appointed to replace Tshekedi Khama, the former president's brother, who is widely perceived as being behind the promotion of BTO as a land-holding power beyond its constitutional mandate. Controlled hunting of wildlife has-as promised-been reinstated, with the first quota issued in late 2019. Although only a few communities lie within potential hunting areas, those that do are in the process of having their area management plans adjusted to accommodate hunting zones. It is likely that government will continue its slow move to support its 2007 CBNRM policy with an implementing strategy, and attempt to return to the original model it promoted.
Following this original path is, however, problematic. First, it is unclear whether global value shifts have left sufficient demand for hunting (cf. Dellinger, 2019) to return to its same level of operation in Botswana. This means that community expectations of a return to higher revenues may be dashed again. Second, even without the erosion of rights in recent years, the original model of CBNRM in Botswana still falls short of its theoretical benefits-from protecting the poorest to effective participation in environmental decision-making (Blaikie, 2006).

| WHAT BOTSWANA'S CBNRM IS AND IS NOT AFTER 30 YEARS
Part of the challenge in Botswana is that the acronym CBNRM has lost its original meaning. CBNRM as a word has come to mean "communities getting money from wildlife-based tour operations in their area." In the current Botswana context, CBNRM is not "communitybased," does not involve "natural resources," and does not include any "management," to a degree that goes beyond Blaikie's earlier exposé where he highlighted, among other challenges, how CBNRM had a "problematic lack of official recognition"; lack of government support; and a persistence-in spite of theoretical advances in common property theory-in government blaming disempowered local communities for environmental degradation (2006).
The current structure of CBNRM cannot be said to be "community-based" when decisions are imposed on communities by higher-level authorities. The stroke-of-a-pen removal of revenue control and head leases through ministerial directive exposes how little control there ever was. Neither can CBNRM be said to speak to "natural resources." For the most part, trust activities have focused on wildlife-based tourism, a resource which they may not use directly, and at the expense of other resources and other dimensions of resource use (i.e., subsistence vs tourism use). Finally, no actual "management" activities take place under the current CBNRM model (Jones & Murphree, 2004). Establishing and implementing of zoning, harvesting seasons, off-take limits, and monitoring are all activities that should take place for a range of resources, including fish, thatching grass, tree poles, reeds, and medicinal plants (such as grapple plant-Harpagophytum procumbens), among others, but typically do not.

| THE UNDERMINING OF CBNRM RIGHTS AS A POWER DYNAMIC
The recentralization of CBNRM can be interpreted as a reinforcement of state dependency and the engineering of "equality" at the expense of equity-taking away information, and taking away resources, keeping people economically and politically poor at the local level in order to benefit the national level (see for example, DeMotts & Hoon, 2012;Rihoy & Maguranyanga, 2010). Locally, there are some who believe that the removal of first, revenue control and second, head lease control was less to do with central government dissatisfaction of management quality and their stated need to protect the burgeoning tourism industry (see for example Keakebetswe, 2012), and more about concerns that rural communities were forgetting their place, and that the state preferred to maintain control over wildlife as a "patronage resource" (Gibson, 1999). To the political elite, rural communities were becoming too politically strong, particularly where CBNRM trust representatives support opposition political parties (Hoon, 2014). Under this interpretation, revenue and head lease rights were removed not to protect the larger community, but to ensure that small communities do not become too powerful in the context of a tiny nation that has only ever elected one political party (Hillbom, 2011;Poteete, 2009). If a small village in a remote area becomes more autonomous, it would set an unacceptable (to central government and the ruling party) precedent for all villages to want to take over control of the resources in their surroundings.
Despite government attempts to control decisionmaking, CBNRM has given residents of rural Botswana communities an opportunity to learn that they do have a voice, and can have power over their resources (Gujadhur, 2000). For example, they have learned different strategies for bargaining and have accommodated a longer planning horizon, and both groups and individuals within different communities use such strategies to protect their individualized interests (Blaikie, 2006;Boggs, 2000;Swatuk, 2005). In some instances, communities have by-passed government to negotiate directly with tour operators, or have tried to develop their own tourism enterprises-either through the trust or as individual citizens. Some communities outside of wildlife management areas have formed alternative structures such as area management committees, and established their own plant resource management and monitoring schemes (Cassidy, Magole, & Botumile, 2017;Magole, Magole, & Moemedi, 2017). Community members now understand their capacity for navigating the quite alien and alienating foreign-dominated tourism industry (Mbaiwa, 2017), even when their view of it might differ considerably from a view external to the community. Whereas outsiders may ascribe an intrinsic or aesthetic value to wildlife, community members have had to switch from seeing wildlife as either a threat to livelihoods or a consumptive resources, to a view of tolerance that allows their interests to overlap with those of the tourism industry, where wildlife is a source of monetary revenue (cf. also Magome & Fabricius, 2013).
Further, because many rural communities are small, positions of leadership tend to circulate among a small number of people. This means that other community institutions, such as village development committees have also acquired a set or strategies for negotiation, communication and planning. Even as government appears locked into its rhetoric of static and monolithic communities, power dynamics and interest groups have emerged through the introduction of CBNRM (DeMotts, Haller, Hoon, & Saum, 2009;Swatuk, 2005).
The combination of stronger village-level engagement skills with the conceptual disconnect of the current CBNRM model suggests that community conservation in Botswana is ripe for redefinition. Stepping outside of the over-monetized, wildlife-based tourism CBNRM model cuts away the highly politicized yet stagnant rhetoric that exemplifies a brittle, locked-in system on the brink of release (Gunderson & Holling, 2002). Community conservation in Botswana is at a point of transformability that would allow for reorganization based on novelty (Allen & Holling, 2010;Walker, Holling, Carpenter, & Kinzig, 2004). This is an opportunity for a fresh start that should be based on a modular, "small wins" (Weick, 1984) approach that responds to local social and environmental variations, where each community works at within-community level with resources its members decide are of conservation importance.

| A WAY FORWARD TO MORE MEANINGFUL CBNRM IN BOTSWANA
The modular, small wins approach is particularly suited to remote, rural communities where access and connection to a broader network are limited. It allows each project to be appropriate to its location, and for communities to work with the specific ecological and social attributes at hand. The starting point for any project would a natural resource that has current use-value, and the usergroup with interest in that resource. Working with user groups sets the focus for conservation intervention below the homogenized community identity (Agrawal & Gibson, 1999;Gibbes & Keys, 2010), and better matches the scale of project benefits with that of environmental costs, which are most typically felt at household, not community, level (Cassidy et al., 2017;Silva & Mosimane, 2013;Wyborn & Bixler, 2013).
Coupled with the hierarchical shift of focus for conservation interventions, would be a shift in narrative. The paradigm of communities being compensated for bearing the costs of living with wildlife casts them as passive victims and is disempowering (Dressler et al., 2010). Instead, using concepts such as "custodianship" or "stewardship" which focus on environmental responsibilities (Bennett et al., 2018), rather than solely on rights and benefits, reintroduces a sense of agency at the local level. Stewardship acknowledges that those living closest to, and most dependent on, a resource are best placed and motivated to protect it and manage its sustainable use. Community conservation would refocus on actual resource use practices, and an explicit link between management of a specific resource, its sustainability and its contribution to livelihoods. For example, in a Kenyan community area, carnivore conservation is centered on livestock husbandry (Woodroffe, Frank, & Lindsey, 2007), while in Uganda's Rwenzori National Park, forest management agreements are tied-less successfully-to permissions for subsistence use of resources (Jagger, Sellers, Kittner, Das, & Bush, 2018).
In a small wins, modular system of community conservation, the options for incentivizing partnerships are much wider. Government departments responsible for the full range of the country's natural resources can spatialize their own management responsibilities and respond to localized variation in environmental and social conditions. By working with an existing community organization to register users and establish resource user rights, rules, offtake quotas, seasons and licenses for resources in a bounded area, the relevant government agency would partner with the community to re-establish resource access control (Ostrom, Burger, Field, Norgaard, & Policansky, 1999;Vollan & Ostrom, 2010). Not only does such control strengthen local leadership, but it also out-sources basic management tasks away from distance government offices to those institutions on the ground.
Mechanisms for participation in stewardship-based resource use projects should be predicated on active participation by resource group members. Partnerships would be on the basis of conservation agreements separately around each specific resource (Niesten, Zurita, & Banks, 2010). The agreement is a negotiated contract that outlines a set of rights and responsibilities, and typically links social or financial benefits directly to conservation management or environmental behaviors (Niesten et al., 2010). An example well-suited to the Botswana context includes assisted access to beef markets or development of water infrastructure in exchange for range management practices such as collective herding, rotational grazing, or reduction of wildlife predation through kraaling. Another would relate to harvesting of plant resources such as thatching grass or grapple plant, with transport and preferred markets providing the incentive for adherence to harvesting seasons, techniques and quotas. While each individual resource user would sign on to the agreement, there is still an important role for a centralized community organization, which would engage in permit allocation, collection of permit revenue, establish season and locale for resource use, and implement monitoring of both the resource and its use. Where a conservation trust already exists, it could take on this role. However, other equally valid organizations include the village development committee or the tribal leadership, because the role of the conservation organization would shift from one of leaseholder, to one of coordinator.
In the Botswana context, stewardship agreements would involve resource users, a community organization, an external incentivizing and networking agency, as well as the relevant government agency (as legal authority for a given resource, and assistance with networking). By linking stewardship and an associated conservation agreement with an identified, bounded area, the issue of exclusion of nonlocal citizens can be addressed. Access to the resource in that area is solely on the basis of participation in the agreement; and anyone wanting to use the area would need to sign up. The government agency would need to assist in developing by-laws for a given resource in a given location. This suggests a three-stage process, and two distinct arenas for engagement: (a) working with both government and the broader community to identify an area for managing a given resource, and establish resource rules within that area; (b) establishing commitments through a conservation agreement whereby any person wishing to use that area agrees to abide to the rules; and (c) separately working with people from the nearest community to patrol and report to the government agency any infringements of the rules.
The agreement would be hierarchical, enhancing functionality and resilience. Resource group members sign an agreement with the community-level organization, which in turn would sign an agreement relating to group oversight with both the incentivizing and government agencies. Each contract would contain noncompliance punitive measures; at the level of the user, this could be withdrawal of a person's inclusion in the permitting system and hence their access to participation in revenue from the resource, while at the level of the community failure to enforce compliance could lead to withdrawal of incentives for the community. In this way, benefits are directly associated with the use of each resource, and all three parties would have the ability to influence the sustainability of the resource.
To be meaningful at the landscape-and nationallevel, such modular efforts need to be scaled up through collaborations that bring together local communities with shared resource interests, supporting NGOs and government departments, as well as the private sector (cf. Nelson & Parrish, 2018). Such collaborations would create a nested system of environmental stewardship that is vital for cross-fertilization of ideas and lessons learned, as well as for the development of a broader movement that coalesces core principles. Scaled collaboration also provides opportunities for linking to funding, access to technical capacity, and the development of enabling policy (Wyborn & Bixler, 2013).

| KEY DEPARTURE POINTS
While the political landscape across which Botswana's tourism has played out has constantly evolved, one aspect has remained constant-no real power has ever been intentionally devolved to the local level. In this small wins, modular approach, devolvement is outside of the discussion. Projects start on the ground, with pre-existing interests and usage. Small modular projects that are stand-alone are not only successes in and of themselves; they also contribute robustness to community conservation across the broader landscape. A diversity of projects allows for experimentation and adaptation for greater social learning (Armitage, 2005;Berkes, 2009). "Organic" partnering at local and intermediate levels can lead to structural complexity, increasing the overall resilience of community conservation and rural livelihoods (Olsson, Folke, & Berkes, 2004). The Namibian conservancy system, although not previously explored through a "small wins" lens, is a useful example of how the grassroots base of each conservancy can be seen as a module which is then linked through support networks provided by NGOs (such as IRDNC and WWF-Namibia), to provide an expanded landscape of collective conservation (Bollig, 2016). Namibian conservancies think much "smaller" than Botswana's trusts-they aim to get revenue, however little, to participating households (Bandyopadhyay, Humavindu, Shyamsundar, & Wang, 2009;Silva & Mosimane, 2013) rather than focus on generating huge sums at community level. Namibian conservancies also bring together a range of different land uses and income-generating activities based more on the resources and ecosystem types available than on a centralized policy principle (Bollig, 2016;Mannetti, Goettert, Zeller, & Esler, 2017;Naidoo et al., 2016).
Importantly, a small wins approach in Botswana would be a major shift away from the high stakes revenue generated by JVPs with wildlife tourism ventures. This shift is particularly relevant at a time when global interest in Africa's wildlife puts increasing pressure on governments generally, and Botswana's particularly due to its high elephant population (Salerno, Cassidy, Drake, & Hartter, 2018), and global growth in wildlife tourism increases its value to national economies (cf. Cornelissen, 2017). As such, government is likely to become even less willing to devolve any control below its central governance structures.
However, there are several positive points associated with these lower stakes. The benefits will reach the households with the greatest interest in the resource. It will not involve people who do not want to be involved, or who are only engaging because they see a chance to access financial rewards. On a modular basis that focuses on a wider variety of natural resources, stewardship projects can reach an exponentially greater number of communities in Botswana, since the great majority of communities lie outside wildlife areas. When combined with a networking platform coordinated by an impartial institution, such as an NGO, or independently structured platform for exchange, the benefits of interest-based, community-level conservation can be rolled out across a range of different ecosystem types, and for a range of different natural resources.
In this way, without directly engaging with wildlife, habitat can be managed across a much broader landscape. Instead of "fast cash" being captured (and toooften lost) at community level, a stewardship model would enhance livelihood stability, and ensure that the resources of interest at any given location, are conserved.

ACKNOWLEDGMENTS
I would like to thank the editors and reviewers for their helpful guidance in strengthening this manuscript.

CONFLICT OF INTEREST
This article was written without any funding, and with no institutional interests at stake. As such I declare no conflict of interest.
AUTHOR CONTRIBUTION I confirm that the work in this manuscript is wholly my own.

ETHICS STATEMENT
This manuscript is solely the work of the author. No funding was received for this analysis, and there is no known conflict of interest. The author has not published the ideas in this manuscript in any other forum.

DATA AVAILABILITY STATEMENT
No original data were generated for this manuscript, and all works cited are in the public domain.