ArticlePDF Available


In East Africa, pastoralist systems are undergoing rapid transformation due to land enclosures, benefit distributions associated with new land uses, shifting social relations, and changing authority and governance structures. We apply a critical analysis of the institutions that mediate access and benefits across a complex mosaic of property relations within Ilkisongo Maasai pastoralist land in southern Kenya. Our analysis elucidates how global and national influences have interacted with shifting dynamics of socio-cultural norms and rules regarding access to create new benefit pathways, cascading patterns of accumulation and social differentiation, and diffuse institutional controls over land.
Full Terms & Conditions of access and use can be found at
The Journal of Peasant Studies
ISSN: (Print) (Online) Journal homepage:
Diffuse land control, shifting pastoralist
institutions, and processes of accumulation in
southern Kenya
Ryan R. Unks, Mara J. Goldman, François Mialhe, Yanni Gunnell & Charlotte
To cite this article: Ryan R. Unks, Mara J. Goldman, François Mialhe, Yanni Gunnell &
Charlotte Hemingway (2023): Diffuse land control, shifting pastoralist institutions, and
processes of accumulation in southern Kenya, The Journal of Peasant Studies, DOI:
To link to this article:
© 2023 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Published online: 16 Feb 2023.
Submit your article to this journal
Article views: 56
View related articles
View Crossmark data
Diuse land control, shifting pastoralist institutions, and
processes of accumulation in southern Kenya
Ryan R. Unks
*, Mara J. Goldman
, François Mialhe
, Yanni Gunnell
Charlotte Hemingway
Department of Geography, CNRS 5600 EVS, University Lumière Lyon 2, Bron, France;
National Socio-
Environmental Synthesis Center, University of Maryland, Annapolis, MD, USA;
PASTRES Programme,
Institute of Development Studies, University of Sussex, Brighton, United Kingdom;
Department of
Geography, University of Colorado Boulder, Boulder, CO, USA;
Department of Tropical and Mediterranean
Animal Production Systems (SELMET), LInstitut Agro/Montpellier SupAgro, Montpellier, France
In East Africa, pastoralist systems are undergoing rapid
transformation due to land enclosures, benet distributions
associated with new land uses, shifting social relations, and
changing authority and governance structures. We apply a critical
analysis of the institutions that mediate access and benets
across a complex mosaic of property relations within Ilkisongo
Maasai pastoralist land in southern Kenya. Our analysis elucidates
how global and national inuences have interacted with shifting
dynamics of socio-cultural norms and rules regarding access to
create new benet pathways, cascading patterns of accumulation
and social dierentiation, and diuse institutional controls over
Maasai; land control;
pastoralism; Kenya, access;
extended entitlements;
critical institutionalism
A growing literature has emphasized green grabbingand accumulation by disposses-
sionin pastoralist lands in East Africa (e.g. Benjaminsen and Bryceson 2012; Bersaglio
and Cleaver 2018; Bluwstein et al. 2018). This literature links changes in land control
and dierent types of associated benet streams with complex, variegated interactions
between patterns of international investment, reforms in governance processes at mul-
tiple scales, and local social responses from below(Hall et al. 2015; Lund and Boone
2013; Gardner 2012). While processes of enclosure and dispossession often pose conspic-
uous, broad-scale patterns of exclusion from land and benets, more subtle, diuse
changes in access, claiming, and exclusion at local scales can also have important impli-
cations for livelihood systems (Fairhead, Leach, and Scoones 2012; Peluso and Lund 2011).
Extensive pastoralist systems are distinguished by exible, adaptive livestock husbandry
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (
licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
CONTACT Ryan R. Unks
*Present address: Robert Schuman Centre for Advanced Studies, European University Institute, Fiesole, Tuscany, Italy.
practices and socio-cultural relations that enable a production strategy that benets from
environmental variability (Krätli and Schareika 2010). These systems have been character-
ized by their open property regimesor complex mosaics of property(Moritz 2016;
Robinson 2019), with plural, overlapping rights and authorities, and highly dynamic
social processes of inclusion and exclusion that all play key roles in access to land and
other resources (German et al. 2017; Gongbuzeren, Huntsinger, and Li 2018 ; Scoones
2021). These, as well as other dynamic socio-cultural and environmental relations,
require that a distinct set of analytical lenses be applied in order to understand the emer-
ging patterns of accumulation and social dierentiation, and ongoing, rapid transform-
ations occurring in these globally important livelihood systems (Scoones 2021).
In this paper, we focus on how informal and formal institutions shape benet path-
ways, processes of accumulation, and land control in extensive pastoralist systems that
are undergoing transformation. We do so with a case study of changes in livelihood prac-
tices, social dierentiation, and struggles among Ilkisongo Maasai over collectively titled
land in Kajiado County in southern Kenya. As seen throughout East Africa, the interaction
of complex recent changes in socio-cultural norms, and unequal abilities to access, benet
from, and control land, have profoundly impacted emerging patterns of social dieren-
tiation among pastoralists (Goldman and Riosmena 2013; Korf, Hagmann, and Emmeneg-
ger 2015; Lind et al. 2020; Unks et al. 2019). A range of factors, including structural
adjustment policies and decentralization reforms, have also bolstered recent patterns
of investment and inuence by new political actors, that have produced new sources
of capital, new pressures for land use, and cascading changes in social relations
(German, Unks, and King 2017; Lind, Okenwa, and Scoones 2020; Unks 2022). Explicit
enclosure and establishment of new boundaries within Kenyan Ilkisongo Maasai land
has resulted in both dispossession and loss of access to grazing resources. However,
changes due to new political and economic dynamics, and the expansion of wildlife con-
servation and industries such as mining and cash crop farming, have also inuenced pro-
cesses of accessing and beneting from land in more subtle ways. Focusing on how
institutions mediate dierentiated abilities to access and benet from land within a
complex mosaic of property relations, we elucidate (1) emergent pathways of accessing
and beneting from land within Ilkisongo Maasai communities; (2) how these pathways
are recursively reproducing social dierentiation; and (3) how changing livelihoods and
social dierence relate to local politics of accumulation and land control.
Previous studies among pastoralists in East Africa have shown how privatization, wild-
life conservation practices that exclude access to land, changing social relations, and new
practices of supplementing livestock mobility all shape dierential abilities to access and
benet from land (Goldman and Riosmena 2013; Unks et al. 2019). Here we expand this
approach to understand the various ways in which multiple, overlapping institutions
mediate how people benet from resources, but also shape processes of accumulation,
both of which underpin changing local politics, social relations, and views of collectively
titled land. In Kajiado County, the majority of collectively titled Maasai lands, known as
group ranches (GRs), have subdivided and dissolved since establishment in the 1970s
(Mwangi 2007a). However, several GRs surrounding Amboseli National Park (ANP) have
remained collectively titled. Amidst concerns surrounding the ecological implications of
subdivision, fencing, and farming (Groom and Western 2013), wildlife conservation
NGOs have sought to maintain open, contiguous rangelands surrounding ANP
(Western 1994), and spearheaded a range of interventions intended to sustain collective
land tenure (Unks 2022). At the time of writing, however, these GRs have begun to for-
mally subdivide. Sweeping changes in livelihood practices, market interactions, socio-cul-
tural relations, and state and non-state actor inuence over land use have also
transformed the way that Ilkisongo Maasai relate to land (BurnSilver 2009; Campbell
et al. 2005; Kituyi 1990; Lind et al. 2020; Roque de Pinho 2009; Southgate and Hulme
1996). In this context, we asked how Ilkisongo Maasai views of collectively titled land
have been impacted by changing institutional congurations of resource use and dier-
entiated abilities to benet from land. In addressing this question, we contribute to larger
debates about pastoralist property relations, land tenure, and ongoing changes in land
control in relation to global, national, and local processes.
To trace the evolution of the overlapping informal and formal rules and norms that
mediate access to variable resources and benets from land, we apply a non-normative
view of institutions (Johnson 2004), and understand them as inherently social, relational,
and historically contingent (Cleaver 2012). We draw on critical literature on institutions
(Cleaver 2012), access (Ribot and Peluso 2003), and entitlements (Leach, Mearns, and
Scoones 1999) to expand our understanding of social dierentiation across a complex
mosaic (Robinson 2019) of property relations and land use. Focusing on the relationship
between access, benet streams, and land control (Peluso and Lund 2011), we situate
socially dierentiated abilities to access, benet from, and control land and resources
in relation to a broad spectrum of social processes, local authority structures and func-
tions, and wider political and economic shifts. Through examining the webs and
bundles of powers(Ribot and Peluso 2003) that extend across complex systems of
access and benets, we highlight how dierent actors produce and reproduce asymme-
tries through their abilities to shape institutional congurations (Cleaver and de Koning
2015) and mediate processes of inclusion and exclusion (Lund 2016 ; Sikor and Lund
2009). We argue that applying this approach to agrarian questions of production,
accumulation, and politics (Bernstein 1996,2010), while considering important socio-
cultural dimensions of variegated transformations (Edelman and Wolford 2017), can
expand understandings of ongoing patterns of shifts in the control of land and associated
benet streams in extensive pastoralist systems (Scoones 2021).
In Ilkisongo Maasai land, demarcation and enforcement of new boundaries associated
with conservation and privatization has broadly restricted access to land and resources.
However, attention to ner-scale changes in rules and norms of access to multiple
resources, highlights complex, socially dierentiated benet pathways where explicit
control over land is otherwise dicult to identify. A systematic understanding of the
ways in which dierent households secure benets from land which we refer to as
benet pathways provides a lens for understanding complex changes in land control
and accumulation at multiple spatial scales, and highlights the interaction between
direct benets from land (e.g. farming, livestock) and new, indirect ways of beneting
from land (e.g. leasing for conservancies and mining).
By mapping these benet path-
ways, we trace explicit connections between socio-cultural change, livelihood dieren-
tiation, and uneven abilities to regulate the use of resources at multiple scales within
We distinguish these benet streams as corresponding to the distinction of direct use rightsand indirect use rights
over land made by Sikor, He, and Lestrelin (2017).
collectively titled lands. By showing the relationship between multiple benets from land,
multiple dimensions of inclusion and exclusion from these benets, and the social and
political relationships that mediate this inclusion and exclusion, we elucidate diuse pro-
cesses of land control and accumulation.
In what follows, we rst review intersecting bodies of literature on institutions, access,
entitlements, and land control. We then present empirical ndings alongside supporting
literature to show how changes in rules and norms and other socio-ecological relations
have shaped dynamics of accessing and beneting from resources in Ilkisongo Maasai
land. We then relate these changes to trends of social dierentiation and new patterns
of accumulation within land that was de jure recognized as collectively titled, but with
complex de facto patterns of access and claiming. Finally, we discuss linkages between
benet pathways and livelihood outcomes, perceptions of collectively titled land, and
processes of land control. We show how under shifting external inuences, benet path-
ways have become recongured along old and new lines of inclusion and exclusion
according to authority, wealth, political aliation, socio-cultural identity, and geographic
Pastoralism, institutions, entitlements, and access
Mobile pastoralism is a livelihood activity that is well-adapted to drylands (Behnke,
Scoones, and Kerven 1993), in large part due to exible, informal rules and norms of
resource access that enable constant adjustment in response to spatio-temporal environ-
mental variability (Niamir-Fuller 1999; Oba 2001). However, pastoralistspractices of mobi-
lity and resource access have frequently been marginalized relative to other land uses,
and pastoralists themselves often have limited political power within state governance
structures (Turner 2017; Niamir-Fuller 1999; Waller 2012). Nation states have tended to
refer to pastoralist areas as underutilizedor degradedas a justication for economic,
political, and wildlife conservation interventions (Bluwstein et al. 2018; Scoones et al.
2019), and for exerting control over populations, asserting legibility, and extracting
value (Turner 2017; Waller 2012). This has led to a frequent lack of appropriate land
tenure and wider political and economic support for pastoralist livelihood practices
(Turner and Schlecht 2019), and to a perpetuation of narratives of the death of pastoral-
ism(Weldemichel and Lein 2019).
Common pool resource (CPR) studies have emphasized how pastoralist institutions
mitigate risk associated with high variability of rainfall (Niamir-Fuller 1999), and examined
the main factors driving the seeming paradox of pastoralists choosing to privatize collec-
tively titled land (Mwangi 2007a; Lesorogol 2008). In Kenya, CPR analysts have contributed
powerful arguments about how British colonial-era policies, along with those
implemented by the Independent Kenyan state after 1963, undermined Maasai land man-
agement systems (Mwangi 2006,2007a,2007b; Mwangi and Ostrom 2009a,2009b).
However, persistent conceptual problems have limited CPR theorys explanatory power
in pastoralist contexts. In particular, CPR theorys emphasis on clearly dened groups,
with powers to exclude outsidersenabling sustainable management of resources
(Ostrom 1990), has not reected the exible, overlapping, and uid social processes
used by pastoral peoples to access spatio-temporally variable resources (Fernández-
Giménez 2002 ; Turner 2011). Alternate conceptualizations have attempted to overcome
these limitations through characterizing pastoralist resource governance as varying from
open property regimesthat are regulated primarily through informal rules (Moritz 2016)
to complex mosaicsinvolving a mixture of open property,common property, private,
open access, and/or state regulated land (Robinson 2019). Additionally, given the
inherent complexity of pastoralist systems of access and the importance of bottom-up
social processes, they also elude normative, managerial framings that focus on how the
tof institutions determines the potential to regulate resource use (e.g. Cumming,
Cumming, and Redman 2006; Epstein et al. 2015; Young and Gasser 2002). While a
wide range of scholars have emphasized how natural resource institutions are fundamen-
tally linked to social conict, inequality, and asymmetries in decision making and auth-
ority (Agrawal 2003; Cleaver 2012; Kabeer 2000), CPR studies have generally tended to
not view institutional congurations as constituted through socially dierentiated pro-
cesses (see Cote and Nightingale 2012; Cleaver and de Koning 2015). In contrast, critical
institutionalists have emphasized how institutions are produced through relational social
processes that involve diverse economic, emotional, moral, and social motivations; asym-
metrical power relations; and struggles over meaning (Bennett et al. 2018 ; Cleaver 2012;
Cleaver and de Koning 2015; Cote and Nightingale 2012; Hall et al. 2014; Mehta et al. 1999;
Whaley 2018 ). These scholars center how institutions tend to have complex origins (e.g. a
mix of customary, colonial state, post-colonial state, and non-governmental); are
inuenced by a wide range of political, economic, historical, and discursive processes
over time; and often represent powerful interests rather than those of communitiesas
a whole (Cleaver 2012;Hall et al. 2014; Leach, Mearns, and Scoones 1999; Mehta et al.
1999; Mosse 1997; Whaley 2018). This highlights the need for a relational understanding
of property in legally plural contexts (Berry 1989), where rules and norms such as land
access rights are dynamic, historically contingent, and contested, and reect and (re)pro-
duce social dierence and relations of inclusion and exclusion (Cleaver 2012; Kabeer
2000). These insights are particularly relevant in contexts of state-imposed collective
tenure regimes, which may in ways resemble commons(Ostrom 1990), but that have
been prescriptively imposed and are inherently at odds with the bottom-upregulatory
concerns of CPR scholars (Berry 2018; Borras and Franco 2010;Lietal.2010; Mwangi and
Ostrom 2009a,2009b; Peters 2004 ; Sikor, He, and Lestrelin 2017).
The conceptual frameworks of access(Ribot and Peluso 2003) and extended entitle-
ments(Leach, Mearns, and Scoones 1999) have both greatly advanced understandings of
rural livelihoods by expanding beyond frameworks that primarily focus on land tenure
and rights. Analysis of access elucidates how property itself is a product of social relations,
among others, and focuses on the webs and bundles of powersthat mediate the ability
of dierent actors to access, and maintain access to resources, but also the ability to regu-
late, reorganize, and control the access of others in diverse contexts (Ribot and Peluso
2003). The closely related extended entitlements framework is more acutely focused on
livelihoods, and explicitly addresses how householdsdierential control over various
social, economic, and material resources shapes their ability to gain benets from multiple
other resources (Leach, Mearns, and Scoones 1999). It expands Sens original (1984) focus
on legal rights, broadening this framing to emphasize how the ability of households or
individuals to utilize and benet from resources is mediated by interactions between
endowments (i.e. assets) and diverse institutions, rather than by rights alone (Leach,
Mearns, and Scoones 1999). An analysis of the interaction between the spectrum of a
households endowments, and the institutions that shape their access to resources, can
be used to mapthe relative entitlements of households, and to assess the role that insti-
tutions play in mediating benet streams from multiple resources (Leach, Mearns, and
Scoones 1999). Studies of pastoralism drawing jointly from the frameworks of access
(Moyo, Funk, and Pretzsch 2017) and extended entitlements (Goldman and Riosmena
2013; Unks et al. 2019) thus enable a ne-grained analysis of political, social, and econ-
omic relations that sustain and recongure institutions which mediate dierentiated
access and benets to livelihoods from land across complex mosaicsof dierent prop-
erty relations (Robinson 2019).
Land control in extensive pastoralist systems
New land uses, global political and economic forces, and local contestations have increas-
ingly been characterized as producing new frontiers of land control,dened as practices
that x or consolidate forms of access, claiming, and exclusion(Peluso and Lund 2011
pp. 668). Changing processes of access to and control of resources in pastoralist contexts
is challenging to understand because of the heterogeneous and variable distribution of
resources over time and space, overlapping webs of social relations, plural notions of
property and rights, and diverse labor practices (Scoones 2021). Increasingly commodied
labor and livestock, privatization of land and resources, and increasing social dieren-
tiation can indicate apparent shifts in these production systems (Bassett 2009; Korf,
Hagmann, and Emmenegger 2015; Scoones 2021). However, land control can also
involve processes of local accumulation, for example through the commodication of
resources (such as pasture and sand), which exacerbates socio-economic dierentiation,
modies access, and can lead to new territorial assemblages (Korf, Hagmann, and Emme-
negger 2015). Furthermore, attention to shifts in the logics of livestock production in
response to changing political economic conditions, but embedded in distinct pastoralist
socio-cultural systems, are important for understanding increasingly capitalist relations
and new patterns of controlling land (Schareika, Brown, and Moritz 2021). Critical agrarian
studies scholars have increasingly shown how land acquisitions and environmental gov-
ernance reforms advocated by international actors are often facilitated by support from
dierent state representatives, including local authorities, with important implications
for local communities (Lanz, Gerber, and Haller 2018; Wolford et al. 2013). Those in pos-
itions of authority, such as local state representatives, are given new powers to recon-
gure access and make claims within geographical boundaries (Peluso and Lund 2011),
and serve as intermediaries to balance the interests of the state with local social relations
(Lund 2016). However, they also frequently disproportionately promote the interests of
new outside actors, and leverage local norms and practices of inclusion and exclusion
along lines of identity to their advantage (Lund and Boone 2013; Lund 2016). These auth-
orities acting as intermediaries can work to recongure land in ways that doesnt consti-
tute clear privatization, for example, within pastoralist communities in Kenya, where
authorities have gained an uneven ability to modify institutions that mediate the
access of others (Bergsalio and Cleaver 2018 ; German, Unks, and King 2017).
Throughout East Africa, complex patterns of sedentarization, increased reliance on
grain-based diets, crop cultivation, markets, and other livelihood changes have exacer-
bated inequality and led to changing norms of mutual assistance among pastoralists
(Fratkin and Roth 1990; Fratkin 2001; Little et al. 2001; McPeak and Little 2005; Potkanski
1999). In the context of land surrounding ANP, Ilkisongo Maasai elites have shaped access
to resources within local communities as they negotiate new land uses with powerful
outside interests such as cement manufacturers and wildlife conservation NGO represen-
tatives (Unks et al. 2021 ). Processes of inclusion and exclusion in access to resources and
benets from land along the lines of clan, age-set, and gender are also known to play a
growing role in political divisions within communities (Southgate and Hulme 2000). These
processes of inclusion and exclusion, along with wider changes in social relations and
dierentiated abilities to benet from land, have important implications for processes
of accumulation, but also for land control. In what follows, we illustrate a case study
where we take a critical focus on shifting institutional congurations that mediate
systems of accessing and beneting from land, and show how new benet streams,
uneven processes of accumulation, and shifting authority, worked together to produce
a new system of diuse land control. In this case study, Ilkisongo Maasai authorities, in
particular, were able to treat collectively titled land as private property, and this,
coupled with their inuence on social processes that mediated access and benets, recur-
sively reproduced local patterns of accumulation, socio-economic dierentiation, and
control over land.
Study area description and history
Our study focused on three GRs of the Ilkisongo Maasai section (Eselenkei, Olgulului
Ololorashi, Mbirikani), which are situated between the Chyulu Hills to the east,
Figure 1. Map of the study area and the three focal group ranches.
subdivided former GRs to the north and west, and Tanzania to the south (Figure 1). The
climate is semi-arid; the annual rainfall regime is bimodal, with rainy seasons typically
occurring in MarchMay and NovemberDecember; and frequent droughts (Altmann
et al. 2002).
A series of colonial and post-independence land appropriations, formalization of
institutional congurations, and decreased political autonomy all restricted the exi-
bility of Maasai pastoralism in Kenya (Rutten 1992; Mwangi and Ostrom 2009a). The
British colonial authority coerced the Maasai into signing a series of treaties that
reduced their land by 5070% and limited them to current day Kajiado and Narok coun-
ties (Hughes 2006; Reid 2012). Additional land was appropriated for private European
ranches (Spear 1993), for game reserves under the national parks ordinance of 1945,
and for privatized ranches claimed by Maasai authorities (Campbell 1993; Galaty
1992; Kituyi 1990). Within the land of the Ilkisongo section, Amboseli, along with the
Chyulu Hills and Tsavo areas, were designated as national reserves in 19471948; and
all were later gazetted as national parks following independence (Hughes 2006).
Based upon pejorative assumptions about Maasai land use, a series of interventions
were introduced by colonial authorities (and subsequently maintained by the postcolo-
nial Kenyan state). These measures aimed to transform livestock husbandry practices
into commercial production, undermine communal land management, encourage indi-
vidualistic thinking, and promote cultivation of cash crops on settled land (Fratkin 2001;
Rutten 2008; Waller 2012; Hemingway et al. 2022).
The colonial-era East African Royal Commission (Dow Commission of 1952) along
with the post-independence Lawrence Report formed the policy basis for the establish-
ment of GRs, beginning in 1968 with the Land Adjudication Act. GRs implemented collec-
tive land tenure, brought grazing management plans designed to reduce stocking rates
and limit mobility, developed permanent water sources, and introduced measures to
prevent livestock disease and promote commercialized livestock keeping (Campbell
1981; Grandin 1991; Mwangi 2007b; Rutten 2008; Waller 2012). Many GRs were not
designed to reect the need for seasonal access to ecologically variable resources, and
those that did include seasonal forage and water access considerations did not account
for required mobility during droughts (Campbell 1981; Coldham 1982; Rutten 1992;
Rutten 2008; Waller 2012). Collective GR land tenure recongured access rights and
undermined the socio-cultural basis of pastoralistsmobile production strategies
(Okoth-Ogendo 1986; Okoth-Ogendo 1989). As such, GRs constrained the exibility of his-
torical resource management institutions and social relations that previously mediated
responses to ecological variability, leading to failure of collective management
(Mwangi 2006;2007a;2007b; Mwangi and Ostrom 2009a,2009b). In Kajiado, most GRs
were soon subdivided into private parcels (Mwangi 2007a). This was motivated by a
complex suite of factors including fears of appropriation of land by Maasai elites, non-
Maasai, and the state, as well as deepening inequality within the GRs, national prioritiza-
tion of crop cultivation and land privatization, and Maasai individuals use of land as col-
lateral to obtain loans (Galaty 1992; Galaty 1994; Rutten 1992; Mwangi 2007a). Insecurity
in land tenure, however, continued after subdivision due to GR representatives securing
land for themselves and their relations, while many, especially women and youth, lost
access to land and resources in the process (Galaty 1994; Mwangi 2007b; Rutten
1992). Highly unequal distributions of land parcels during subdivision, and land-grabs
by non-Maasai politicians, have been common (Galaty 1994; Galaty 2013; Mwangi 2007b;
Ntiati 2002; Rutten 1992; Waller 1976).
In the three GRs considered in this study, full subdivision had not yet occurred, but all
had begun formal subdivision processes at the time of writing. Until the 1970s, GR land
was primarily used for pastoralism, but competing land uses such as wildlife conserva-
tion, crop cultivation, and extractive industries had steadily increased (Campbell et al.
2000; Campbell et al. 2005; Roque de Pinho 2009; Southgate and Hulme 1996). Since
1977, Ilkisongo Maasai dry-season settlement within ANP, a key dry-season source of
forage and water, had been restricted (Campbell 1981). Widespread farming began in
wetlands in response to drought and livestock loss following this prohibition (Campbell
1981; Campbell et al. 2000; Campbell et al. 2005). There was increased demand to
produce cash crops for urban markets, to extract sand from seasonal river channels
for urban construction, and to dedicate land to wildlife conservation (BurnSilver 2009;
Campbell et al. 2005). Partial subdivision of land, while not ocially recognized, had pre-
viously occurred in all three GRs within high-productivity swamps that were designated
for drainage and cultivation in OlgululuiOlolorashi in the 1970s, and in Mbirikani in the
early 2000s (BurnSilver and Mwangi 2007; BurnSilver, Worden, and Boone 2008; South-
gate and Hulme 2000). Other areas were also subdivided for trade centers and towns, or
designated for the construction of ecotourism lodges. Drier upland areas remained pri-
marily used for livestock herding. Livestock populations had uctuated over time with
variability in rainfall, but as the Maasai population had increased, per capita livestock
numbers had declined, relative numbers of small stock had increased, and household
diets had become increasing dependent on grains (BurnSilver, Worden, and
Boone 2008; Grandin 1991; Nkedianye et al. 2020). Maasai livelihoods continued to
depend heavily on livestock but increasingly relied on cultivating crops, waged labor,
and trade (BurnSilver 2009; Hemingway et al. 2022), with lower numbers of livestock
among completely sedentarized households (Kimiti et al. 2018).
So-called community-based conservation(CBC) projects had gained prominence in
GRs surrounding ANP. Wildlife exploited water and forage within swamps inside the
park during the dry season, but typically moved to areas with more nutritious forage
outside of the park during the rainy season. A number of wildlife conservation NGOs
were active within these GRs, and monetary income to GRs from these activities occurred
primarily through: (1) ecotourism lodges leasing parcels of land as conservancies, with
direct payments to GR representatives, intended to provide water infrastructure, edu-
cation, and medical facilities (Campbell 1999; Rutten 2002); (2) leasing of subdivided
lands from communities, bundled together as group conservancies, where individual
payments were made to each individual title holder; (3) compensation schemes that
oer payments when livestock are lost to predators (Okello, Bonham, and Hill 2014); (4)
jobs as wildlife rangers and scouts on all three GRs; and (5) educational bursaries provided
by NGOs and Kenya Wildlife Service (KWS). The overall monetary incomes produced from
conservation had been reported to be small on all three GRs (BurnSilver 2009; Rutten
2002; Western 1994) and also unevenly distributed among households and across GRs
(Groom and Harris 2008; Roque de Pinho 2009). Very few tourist operators were from
the communities, and the returns from leased lands were largely under the command
of Maasai authorities and elites (BurnSilver 2009; Campbell et al. 2000; Roque de Pinho
2009; Rutten 2002; Western 1994).
The following analysis is based upon eleven aggregate months of eld work by the rst
author. Participant observation and ninety-eight interviews (sampling stratied by age,
gender, clan, wealth, and livelihood practices) were conducted across the three
dierent GRs (OlgululuiOlolorashi, Eselenkei, Mbirikani). Interviews were limited to GR
residents of Maasai ethnicity to focus on understanding their views of collectively titled
Interviews were conducted primarily in Maa and translated into English by a trans-
lator, with several also in Kiswahili or English. Interviews were conducted under informed
consent and recorded only if additional consent was granted. Interviews followed a semi-
structured format, with a subset of questions replicated in each interview that allowed for
responses to be coded and quantied across interviews. Themes addressed in the inter-
view questions included:
(1) Changes in informal and formal rules or norms that mediated livelihood practices (e.g.
changes in land tenure, land use, access relations, authority structures, and mutual
(2) How the above changes in institutions related to the ways individuals have adjusted
their livelihood practices.
(3) Views of collective land management.
In order to challenge preliminary interpretations, the rst author reported study
ndings at eight public meetings in locations where interviews were completed. All resi-
dents were invited and encouraged to critically discuss study ndings (noted in the text
when referenced). Below we rst report ndings on changes in institutions alongside key
literature on past institutional congurations. We follow this with an analysis of dieren-
tial access and entitlements as shaped by institutions in this dynamic context. We then
relate overlapping changes in benets from land to changing views of collective land
Changes in institutions mediating herding access
In addition to restrictions in access to vital dry season resources, a number of other shifts
in rules and norms had limited Maasai herding practices in the study area over time.
Though Maasai historically practiced transhumant pastoralism, multiple practices were
used by colonial and post-colonial governments to coerce sedentarization. Colonial-gov-
ernment-appointed Maasai chiefs(see Waller 1976; Hughes 2006) played central histori-
cal roles in the formation of permanent settlements (emparnat) intended to sedentarize
families near boreholes, pipeline junctures, and other water access points, and to institute
grazing management plans focused near these settlements (BurnSilver, Worden, and
Boone 2008; BurnSilver 2009; Fratkin 2001)(Figure 1). Sections (olosho, pl.: iloshon) are
the largest geographic divisions within Maasai territory and represent distinct socio-lin-
guistic groups. Rules implemented by chiefs further divided these into imurrua,
Members of Kamba, Kikuyu, and Luo ethnic groups typically resided in farming areas or urban centers and did not
usually keep livestock that utilized collective land.
10 R. R. UNKS ET AL.
areas that contain permanent settlements as well as dry- and wet-season grazing areas.
Historically, access to these areas was open to other groups, and numerous male and
female interlocutors over 50 years of age (hereafter: male and female elders), told us
that before GR formation, movements between areas that came to be designated as
imurrua and GRs were common. For example, people who lived in areas situated
within OlgululuiOlolorashi GR formerly regularly moved to areas currently within Mbir-
ikani GR during the beginning of the short rainy season (oloorkisirat), and those living
within Mbirikani GR moved to areas within Olgulului-Ololorashi during the beginning
of the long rainy season (inkokua). These movements with livestock across GR boundaries
over time became much less common, and were actively prevented by chiefs aside from
in a few localities with permanent settlements near to GR boundaries. Subdivision had
also occurred to the south within the Ilkisongo section (i.e. Kimana south to Oloitokitok),
and to the north and west in the Ilkaputei and Matapato sections, constraining movement
to these areas in the recent past (Figure 1). Many interlocutors told us that they still some-
times moved to these lands within subdivided areas, but that it required a relationship
(e.g. agnate, ane, or stock partner), and increasingly, a rental agreement; as one male
elder remarked: before they used to move freely, now even your friend asks you for some-
thing(see also BurnSilver and Mwangi 2007).
At the time of the study, within GRs, movements with livestock between imurrua had
also become increasingly limited, but access was sometimes granted to outsiders under
drought conditions. Movement across these numerous boundaries required permission
from chiefs, GR committees, and members of an imurrua, with denial being common.
As one male elder commented, they were sometimes told to return to your place, you
are here because you didnt save your grass. Movement was also sometimes discouraged
by access to water being limited or charged for. Importantly, both imurrua and water
point access were sometimes, but not always, limited by clan aliation. Although each
GR considered here contains a mix of clans (engilat) and sub-clans, GRs were allocated
along lines of clan, and were sometimes dominated by one group, to the exclusion of
others. Rivalries were particularly strong between three clans: the ilaitayok and ilaiser,
who are both clans within the orokitengmoiety; and ilmolelian, a clan of the odomongi
moiety. Each imurrua tended to be dominated by one of these groups. Farming had also
limited livestock access within GRs to perennial, spring-fed, swamp areas that formerly
were heavily relied upon during drought, but which had been privatized, drained, and
converted to farms. Finally, leased areas within GRs that have become conservancies
(Figure 1) had restricted grazing access, despite sometimes allowing grazing during
droughts, albeit only after other reserve areas had been exhausted, and no overnight
cattle enclosures were permitted within them (see also Rutten 2002).
Changes in norms of mutual assistance
Changes in mutual assistance had created additional limitations on mobility. Two
common types of mutual assistance: individual and clan-based, are core features of
Maasai social organization (Galaty 1981; Grandin 1991; Potkanski 1999). Individual
Sometimes referred to in the literature as enkutoto, but imurrua is the Kenyan Ilkisongo term. Imurrua also refers to the
sites of former homesteads.
assistance through food and livestock occurs between patrilineal and anal bonds as well
as between close friends that share age-set bonds formed among the ilmurran (typically
unmarried males initiated into an age-set through circumcision and charged with security
and cattle herding) (Galaty 1981; Grandin 1991; Potkanski 1999). Clan-based assistance is
distinct from individual assistance, and can take the form of multiple clan members con-
tributing livestock for various purposes, e.g. bride-price, nes, or following loss of livestock
(see Galaty 1981; and Potkanski 1999 for detailed discussions). These agnate, ane,
clan, and friendship bonds are the basis of numerous important relationships that also
facilitate securing seasonal grazing access across dierent spatial scales (Homewood
and Rodgers 1991).
Among the Ilkisongo, mutual assistance through kinship and other extended social
networks beyond immediate family had decreased with the increasing inuence of live-
stock markets, cultivation practices, and employment (Campbell 1999), as also noted
among Maasai in other sections (Galaty 1981; Potkanski 1999). Ilkisongo female elders
in particular tended to emphasize that mutual assistance was becoming increasingly loca-
lized within imurrua, and isolated by family, age-set, and clan, especially among men.
Many male and female elders described these changes in behavior as symptomatic of
wider changes in norms, including overall decreases in respect (enkanyit). However,
some, also usually female elders, disagreed that mutual assistance had decreased, some-
times even claiming that, overall, group-level unity (naibosho) had increased and that
new types of assistance had become more important than individual mutual assistance
(osotua) especially with a recent proliferation of new norms around church and
womens groups that encouraged pooling of resources, income, and labor support.
These socio-cultural changes were frequently discussed as being closely related to
decreases in shared herding labor. Whereas in the past numerous neighboring house-
holds would combine their herds and household labor, especially during the dry
season where mobility and herd splitting were most important (see also Sperling and
Galaty 1990), it was increasingly common for individual households to herd their own live-
stock. However, some emphasized that those without paid herders would sometimes still
combine their herds, and herd in shifts, especially when accessing dry-season grazing
areas and the national parks during droughts. Some related these decreases in shared
herding labor to changes in joint herding and food-sharing inkangitie (singular.:
enkang; residential compounds of several nuclear families, often of patrilineal descent)
(Grandin 1991), as combining herds and sharing labor primarily occurred within these
inkangitie, or between family and close friends. While inkangitie historically usually con-
tained dozens of households that herded collectively, their physical sizes and the number
of people sharing inkangitie, have both diminished on average (see also BurnSilver,
Worden, and Boone 2008; Archambault 2016). In explaining these changing relations,
male and female elders regularly emphasized the role of rising inequality, with decreasing
mutual assistance between people from dierent levels of wealth, divergent individual
preferences for specic livestock husbandry or housing practices, and also underlying
changes in religion and formal education (see also Borona 2020).
Increases in older women and men herding, and young girls and boys herding only
when not in school were also prominent, and male and female elders usually explained
the increasing prevalence of children attending Kenyan schools as another key factor
underlying changes in shared labor. Since the colonial era there have been government
12 R. R. UNKS ET AL.
attempts to weaken the social bonds of ilmurran through forced school attendance
(Knowles and Collett 1989; Kituyi 1990; Zaal 1999). Historically, Maasai often fought
such pressures, and resisted sending their children to school. This had changed,
however, with many people, especially elder women, associating school attendance
with higher social status. However, male and female elders typically told us that
changes in livestock husbandry were also related to education in three ways: (1) the
quality of practices of care of cattle, especially regarding mobility and protection from
predators, had deteriorated with ilmurran regularly attending school; (2) the need to
pay for hired herders or to resort to alternative arrangements had increased; and (3)
sales of livestock to pay school fees had increased. Although some ilmurran refused to
attend school, and some families could not aord school, young males from wealthier
families nonetheless nearly always went to school, as one wealthy male elder commented,
I will not allow my children to herd. In turn, however, the families of ilmurran that were
not enrolled in school and that were available to herd cattle had increasingly expected
compensation from families they assisted.
A number of other inter-related changes in social relations also had important impli-
cations for mutual assistance, particularly changes in authority structures, decision-
making processes, and clan and age-set politics. The creation of GRs imposed a new struc-
ture of authority, where elected or appointed representatives gained the power to make
decisions on behalf of GR members about land use (Mwangi and Ostrom 2009b ). Govern-
ment-appointed Maasai chiefs, GR representatives, and politicians have in general,
gained exclusive access to, and disproportionate benets from, land, with sweeping impli-
cations for mutual assistance and concepts of ownership (Campbell 1993; Campbell et al.
2005; Galaty 1981). There had also been an increased salience of interactions between
Maasai and conservation NGO representatives, where employees and GR representatives
were perceived as disproportionately beneting nancially from wildlife conservation
activities, and also forming new patron/client relationships with NGO representatives
(Unks et al. 2021). While there have been long-standing tensions between adjacent
age-sets as they have grappled over power and roles, men from younger age-sets have
gained historically unprecedented power by occupying formal GR authority positions
that enabled enhanced relationships of patronage with national political actors and
that undermined the authority of male elders (Campbell 1999; Southgate and Hulme
2000). These tensions have been raising across Maasai areas (for discussion of this in Tan-
zania, see Goldman 2020, and Hodgson 2011). Finally, competition between clans, and
clan-based alliances with national politicians, had also reconcentrated ows of patronage,
which had been detrimental to other forms of mutual assistance relations (Southgate and
Hulme 2000).
Dierentiated access to grazing resources during drought
At the time of study, there were few locations that could be accessed with livestock
outside of GRs, and access to these places was highly socially stratied. During the
drought of 20162017, once forage within GRs was exhausted, reliable grazing access
was secured by some on subdivided, privately titled lands through social relations (e.g.
family, friends, or people in positions of inuence that did so on promise of future
favors), but also through payments. Several herders said they had traveled to Tanzania
or other distant places in search of grazing. The most common destinations at this time
were Chyulu Hills National Park (CHNP), Tsavo West National Park (TWNP), ANP, and
neighboring subdivided GRs. Because pipelines no longer provided water to some perma-
nent settlements, herders were legally able to access water in the swamps inside ANP as
per agreements made following exclusion from having settlements within ANP. Access to
the swamps was permitted for four hours a day, and only in areas where livestock could
be out of sight of tourists. During droughts, however, some herders took the risk of over-
staying this time because grass within ANP was the only available aside from that in
private land, TWNP, or CHNP. Taking this risk to sustain cattle was only possible for
those with sucient herding labor because this enabled them to avoid KWS guards
(who they said were likely to drive livestock away or beat herders, if encountered).
While formally not permitted to access, some also entered TWNP and CHNP, where
access was sometimes negotiated by bribe. If caught by guards, however, and negotiation
was impossible, they said that livestock were driven into remote areas and lost to preda-
tors, and herders ned, beaten, or jailed.
According to male and female elders, in the past, dry season watering points were key
determinants of where households temporarily settled during dierent seasons. However,
patterns of restriction, locations of designated watering points, and settlement areas, and
the ability to transport water, had transformed these patterns. In particular, the ability to
use a vehicle to transport animals, water, and grain to supplement livestock had become
extremely important during drought. As a result, some were able to disproportionately
exploit distant grazing areas that were formerly limited to times when water was available
in seasonal ponds. Although some interlocutors claimed that water carried by truck was
equally available to all, others indicated that individuals with trucks who provided water
would prioritize assistance to those with close relations (e.g. kinship, clan- and sub-clanship,
neighbors). Further, use of supplemental feed such as grain, bailed grass, and crop residues
during drought was a recent practice that became widespread in the area in 20162017. As
one male elder remarked we are feeding them like people now(see also Goldman and
Riosmena 2013). Those who had greater access to these assets and a means of transporting
them, along with water, were able to reach areas far from homesteads and watering points,
where the forage had been grazed less. Those with these assets were thus able to remain
within dry season grazing zones, to graze farther into distant areas, and to reduce walking
strain on animals. Male and female elders both frequently said these changes were leading
to deepening inequality compared to the recent past when everyone grazed in the same
direction and traveled the same distance from settlements.
Market interactions, and ability to gain cash assets that shape mobility
and farming
Overall, Ilkisongo Maasai had been increasingly engaging with livestock and crop
markets in nearby town centers to meet growing needs for cash (e.g. livestock-related
costs, livestock purchases, school fees, transportation, medical expenses, and family
food). This had been facilitated by market demands in Nairobi, and recently constructed
paved roads (see also Campbell 1999). However, because of price uctuations in these
markets, being able to sell livestock strategically when market prices are high had
14 R. R. UNKS ET AL.
gained increasing priority. A sizeable herd, secondary income, or alternate source of food
were regularly emphasized by interlocutors as being crucial for reducing the impact of
herd otake, especially to avoid sales when market prices were low such as during
times when school fees were due or during drought. Indeed, many indicated that their
herd sizes had decreased primarily through necessary sales at extremely low prices,
rather than due to animal mortality during the 20162017 drought.
Many noted that they had been motivated to increasingly adopt improvedbreeds of
livestock (e.g. Sahiwal cattle, Dorper sheep, see also BurnSilver, Worden, and Boone 2008)
by the higher market prices of these animals, despite the majority stating that they
consumed more forage, could not walk long distances, required more intensive
herding practices, and were more likely to die during drought. However, those with
larger herds, and livestock brokers, who in both cases frequently bought and sold
cattle locally and transported them to markets, explained how they were able to utilize
the market to their advantage (see also Evangelou 1984 ; Zaal 1999). These individuals
shared their extensive knowledge of prices at dierent markets throughout the Amboseli
basin according to livestock breeds, animal health conditions, and dierent purposes (e.g.
milking cows, breeding bulls, castrated steers for meat, etc.). Traders, in particular, also
usually had other sources of income and were able to prot from buying and reselling
(often strategically, under favorable market conditions) animals that others could not
because they lacked the labor force or time to transport them to markets.
Farming strategies were similarly dierentiated, with some focusing on cash crops,
some on subsistence farming, and others leasing their informally titled land, usually for
very low returns. Those cultivating for subsistence tended to grow maize or beans for
family consumption, storage, or local sale in small quantities, with maize in particular
having relatively low input costs and being easier to maintain. Growing tomatoes as a
cash crop involved prohibitive costs for many (e.g. digging wells, plowing land, purchas-
ing generators, pipes, chemicals, and seeds) and high amounts of nancial risk due to
uctuating market costs and loss from pests and wildlife (see also Hemingway et al.
2022). The majority of cash-crop farms in many areas were leased to, or cultivated in part-
nership with non-Maasai, who often took responsibility for labor and for deterring wildlife
from foraging in plots. These plots were primarily located within former (drained)
swamps, but also in locations near pipelines in Mbirikani GR, and in close proximity to
the Eselengei River (Figure 1). While most said that there was not a bias in how plots
were allocated amongst clans, they were more likely to say that plot distribution was
skewed in terms of quality toward the wealthy and/or GR representatives. Soil type,
plot size, elevation relative to the water distribution furrow, distance to the river or pipe-
line, and distance to main transit corridors all impacted the farms viability. Depending on
location, bribes allowed some to inuence the amount and timing of water in irrigation
furrows, or to access pipeline water. In locations without furrows or pipelines, farming
was largely limited by ability to hand dig wells or pay for a private borehole. Cash was
also required to pay for purchasing or renting, maintaining, and fueling water pumps.
High cost chemical fertilizer and pesticide reliance had also increased conspicuously,
especially for growing tomatoes, and experienced tomato farmers tended to stress the
high susceptibility of these crops to pests.
Changing ecological stressors and dierentiated abilities to benet from
Male and female elders explained how the above changes in mobility, social relations,
and cash incomes were also interacting in new ways with three key ecological factors:
livestock disease, changing wildlife behavior, and changing rainfall patterns. They indi-
cated that several livestock diseases had increased in recent memory, requiring more
frequent use of antibiotics and acaricides, and that improvedlivestock breeds were
more sensitive to some of these diseases. Many explained that livestock that visited
CHNP and TWNP were exposed to East Coast Fever (oltikana) and trypanosomiasis
(engoroto), and so migrating there required a source of cash for preventative
disease treatment. Numerous interlocutors also explained that predators such as
lions and hyenas feared people less today (see also Goldman, de Pinho, and Perry
2013; Unks et al. 2021) and so were a greater threat to livestock, especially when
using dry-season grazing areas, an issue that was compounded when labor to
defend the livestock from predators was lacking. Male and female elders told us,
nearly unanimously, that seasonal rains had become less predictable (e.g. more fre-
quently failing, or arriving later than usually expected) since the late 1990s. They
also commonly added that the impact of changes in rainfall had exacerbated the
lack of open areas to move to, with many male and female elders repeating the
phrase everywhere is now occupied.
Many told us that the 20162017 drought was the rst time in their experience that
grass had not been available anywhere (aside from the parks and private lands), and
that 20162017 was the rst time they had not migrated during a drought. Most
explained that their decision to not migrate was related to recent experiences, such as
during the 2009 drought, and they recounted that they had lost large amounts of live-
stock, had diculties nding water, and had negative interactions with park rangers.
Those who chose not to migrate also sometimes elaborated and said that those who
had remained within their GRs had lost less livestock than them in 2009, or that personal
risks, increased risks of predation and disease, labor and asset limitations, the new strat-
egy of feeding grains to livestock, and/or prevalence of less mobile livestock breeds, had
inuenced their decisions to remain within GRs. Some also concluded that they had lost
fewer animals by remaining within their GRs in 20162017 compared to those who
moved. However, others attributed their changing assessment of mobility primarily to
decreases in cattle numbers, and some from these households with few livestock said
they were able to sustain them through the drought by grazing within high-productivity
private farm plots in former swamps.
A range of strategies were taken across households to sustain livestock in the 2016
2017 drought. The characteristics of householdsdierent strategies were usually
described to us as associated with dierentiated abilities to access and benet from
dierent resources. Most emphasized that herding labor was a key determinant of their
relative mobility. Herding labor varied greatly among families, and most considered ilmur-
ran labor in particular to be essential for maintaining and protecting cattle in remote areas
with high numbers of predators. Another commonly mentioned asset was multiple
enkang locations, which many told us both enhanced herdersaccess rights across GR
or imurrua boundaries, and also reduced labor constraints because it enabled splitting
16 R. R. UNKS ET AL.
of labor and herds between locations (e.g. it facilitated tending to milking cows and
nursing or weak livestock within permanent settlements while the remainder of the
herd was mobile). Furthermore, some said that those that had one enkang in farming
areas, in addition to others in locations more conducive to the needs of livestock, had
an optimal arrangement.
Having a source of income in addition to farming and livestock also enabled some
households to enhance their ability to benet from livestock and farming (as also reported
by Campbell 1999). These sources of cash usually came from family members employed in
Nairobi, working in eco-lodges or as tour guides, as scouts or guards in conservation
organizations, or in cement manufacturing. Temporary employment incomes were also
common, with some women working to harvest cash crops, and some men working to
mine sand or driving motorcycle taxis. Others managed businesses such as general
stores, restaurants, bars, hair salons, butcheries, or rental houses. Furthermore, many
stated that their livestock- and farming-related sales and costs were closely interrelated,
and often complementary, e.g. selling livestock to invest in farming, and then reinvesting
proceeds from farming into livestock. Some with higher cash assets were able to invest in
feed during the drought, with some even able to buy stocks of feed to sell to others for a
prot. Some pointed to important dierences in quality of feeds, and stressed that lower
price feeds sometimes contained additives (such as sawdust) that were harmful to live-
stock and had sometimes caused their death during the drought. Those with greater
cash assets were also more likely to buy and transport water, more able to pay for
labor and for grazing fees in subdivided areas, to pay nes and/or bribes when
moving, and to pay for transport of weak animals. They were also better able to buer
herd losses from predation, disease, and starvation during the drought. One striking
example of the interrelation of these multiple assets was a prominent county-level poli-
tician and absentee pastoralist, who reportedly did not lose any of his herd of 500
cattle. In contrast, others who lacked assets such as additional income, adequate labor,
supplemental feeds, transport capabilities, and paid grazing access, said they were
heavily impacted by the drought and by other interacting stressors such as markets, live-
stock diseases, and wildlife predation. Similar outcomes were described surrounding
farming, where those with cash assets to pay for bribes to furrow and pipeline managers,
to cover the costs of pumping water from rivers or wells, and to pay for chemical inputs
and seeds tended to report more benecial outcomes from farming. Many, however, also
viewed farming cash crops as enabling them to be less sensitive to drought and also
enabling them to support livestock, and to reduce sensitivity to otake by waiting to
sell livestock until market prices had risen. This showed similarity to other contexts in
the region, where the ability to invest in and benet from cash crops (Little et al. 2001)
has been shown to be closely related to dierentiation in mobility, market interactions,
non-livestock sources of income, and the structure of risk (Lesorogol 2008; McPeak and
Barrett 2001).
The politics of access and benet pathways
Though numerous factors had spurred the process of subdivision underway at the time of
writing (see Unks 2022), to understand how individuals viewed the overall structure of
access and benets, we explicitly asked about their views on subdivision of GRs, which
we then related to wider views of collective land tenure and management. Though in a
minority, some expressed fears about subdivision. Of those fully against subdivision or
with mixed views of it (Figure 2(a)), the fears that they expressed most commonly con-
cerned loss of access to livestock forage, and to a lesser extent, a mixture of concerns
over widespread sales, and/or loss otherwise, of lands, unequal distribution of land in
the subdivision process, fears of people from other ethnic groups moving onto Maasai
land, and increased hatredamong Maasai
(Figure 2(c)).
The overwhelming majority indicated that subdivision would be a favorable change
(Figure 2(a)). The most common motivator they listed was a desire to manage their house-
hold activities and land (eramatare) without consulting others, which was also usually
described as closely related to desires for individual tenure security, which was the
third most commonly referenced motivator
(Figure 2(b)). Both of these motivators
were often explained as enabling conditions to build permanent homes, drill boreholes,
and dig private water reservoirs. The second most common reason for supporting subdi-
vision was mistrust of GR representatives (Figure 2(b)). Though many maintained public
respect for representatives through public narratives of respect (enkanyit), unity (nai-
bosho), and mutual assistance (osotua) all of which are paramount for Maasai they
often expressed deep seated reservations about representatives in these private
Indeed, it was a common sentiment in all GRs that the three primary representatives
(chairman, secretary, treasurer) were involved in nancial mismanagement, were the
main beneciaries of new land use projects, and did not make decisions on GR
membersbehalves. In particular, GR representativesdecisions about leasing land to out-
siders, fears about their ability to make decisions about land in the future, and unequal
distribution of nancial benets from projects they had arranged were frequently
stated to be the source of this distrust. As one male elder remarked they realized even
if they didnt subdivide that it would be sold ; someone might come with a new
idea for a conservancy and so we decided to subdivide). There were widespread views
that representatives were eatingland everywhere and that they were selling without
consulting. This applied to leasing of land to the Simba Cement Company on Mbirikani
Figure 2. Views of subdivision in three group ranches. (a) Support of subdivision; (b) reasons motiv-
ating supporters of subdivision; (c) reasons to not subdivide.
All of which were commonly referenced as having occurred in nearby Kimana, a fully subdivided group ranch.
Commonly expressed as wanting to know ones place.
18 R. R. UNKS ET AL.
GR, sand mining agreements on Eselenkei GR, and wildlife conservation agreements on all
three GRs, where the conditions of payments, and decisions about the use of these funds,
were often conducted in private among GR committee members. Members of Eselenkei
GR frequently described the negotiations and agreements with respect to Selenkay Con-
servancy (Figure 1) as lacking transparency and primarily beneting GR representatives
through their ability to control funds from entry fees, bed night fees, and other
incomes intended for use by the GRs as a whole. Many members of OlgululuiOlolorashi
GR, where subdivided plots were bundled together and leased as Kitenden Community
Wildlife Conservancy to a conservation organization (Figure 1), said that they had not
been informed of the contents of the agreement and suspected that a public meeting
had been manipulated by representatives to make it appear to NGO representatives
that GR members had reached a consensus about leasing the land. Many also stated
that processes of plot allocation in local town centers had foregone public discussion
about surveying processes and fees, and that no documents from surveyors had been
made available to them.
Others expressing support for subdivision saw it as a way of making the benets
gained from livestock more equitable, sometimes explaining that it would enable them
to lease land to those with greater livestock, or negotiate more protable conservancy
leases (equality,Figure 2(b)). These were also often related to the way that chiefs,GR
representatives, politicians, Ilkisongo NGO employees, and their extended families had
accumulated assets derived from leasing land, from wildlife that broadly used collective
land, and that enabled them to enhance the benets derived from farming and livestock.
Some also emphasized a widening social divide between lower and higher wealth
families, politicians, and representatives. This was related through examples of elites
not adhering to grazing rules (e.g. GR representatives and inuential county government
ocials whose herds utilized dry season grazing areas at times others were restricted). GR
representatives were also frequently described as being heavily inuenced in their
decisions about land by patron/client relationships, especially those with higher level
national politicians.
The four most prevalent reasons for supporting subdivision (distrust, managing
without consultation, tenure security, and concern for equality) were often explained
as closely related to one another, and all were tied to direct and indirect benets from
land such as farm plot allocation, water access, bursaries, cash payments from GR
ocials for medical or ceremonial purposes, grazing access, employment, and predation
compensation. In order to understand the complexity of these interrelations and how
they impacted households along lines of social dierence, we next briey summarize
how clan, age-set, wealth, political party, and gender were explained to us as interacting
and forming axes of a constantly evolving relational system of inclusion and exclusion.
While some stated that clan politics had decreased in terms of limited water rationing
and grazing movements (see also Southgate and Hulme 2000), a majority indicated that
the inuence of clan politics had increased surrounding elections, job selection for wildlife
conservation, and distribution of indirect benets from land. Importantly, bribes and dis-
tribution of benets among clan members, and members of allied clans, were vital for GR
representatives to build and sustain support among clan leaders. Many also stated that
national political alliances increasingly inuenced election outcomes, with national poli-
ticians being able to determine who was nominated, to sway elections, and create day-to-
day barriers for elected committee members who choose not to support them (see also
Kibugi 2008). Many said age-set politics had gained an increasing inuence, for example
in Eselenkei where four representatives are from the younger (Irkiponi) age-set and were
widely seen as challenging representatives from older age-sets (i.e. Ilkeleyani and Ilk-
ishumu). Through participant observation, we also documented that dierent subgroups
(e.g. clan, age-set, and political alliances) strategized to push specic agendas in public
meetings, especially in terms of employment in wildlife conservation and sand mining,
plot allocation, or other income benets that representatives had inuence over. We
also observed that while GR representatives commonly engaged in public deliberation,
as has been the norm historically, they also engaged in frequent private meetings with
outside actors (e.g. wildlife conservation NGO representatives, businesspeople, politicians,
surveyors). These all likely contributed to perceptions that GR representatives have made
decisions that primarily suited the interests of their extended families, and their clan and
political allies.
As also noted by others, Maasai women are often excluded from formal decision-
making processes regarding land and nances, yet share an increasing burden of
herding and farm labor (Wangui 2008; see also Yurco 2018). In our study area, women fre-
quently emphasized that they were not typically employed by sand mining or conserva-
tion projects. Neither did they regularly participate in the leasing of GR land, or attend
meetings about land uses, GR nances, or meetings with external actors. As a result, dis-
trust of GR representatives was especially common among women. However, at the same
time these patterns of exclusion and distrust highlight the central role that Ilkisongo
Maasai women were playing in envisioning new ways of sharing land and exploring pos-
sibilities for more even distributions of benets, and how these might be bolstered by
subdividing land. Despite on-going processes of exclusion, women had an increasing
inuence at household, as well as imurrua-level decisions, because of their increasing
ability to control cash through farming labor and other jobs, to advocate for the education
of children, to create new practices of mutual support through womens groups and
church groups, to speak during some meetings, and increasingly, to generate political
pressures through social mobilization. While women continued to be excluded from
most formal positions of authority, they were organizing to advocate for new relations
and decision-making processes surrounding land use (see also Archambault 2016), con-
testing decisions about land use made by GR committees composed exclusively of
men, contesting the alliance-building practices of authorities across political levels, and
fostering new types of alliances intended to counter-act political and clan divisions
among elder men. Women were also creating new alliances with ilmurran from the
youngest (ilnyankulo) age-set, who were also excluded from decision-making processes
(see also Goldman, Davis, and Little 2016). One conict on Mbirikani GR in particular
was frequently referenced, where the ilmurran and women jointly contested an arrange-
ment made by GR representatives for sand mining, leading to a conict claiming the lives
of one olmurrani and a police ocer. References to this, and other similar public contesta-
tions, were often described as closely linked to wider mobilizations by women to nego-
tiate the norms of sharing and beneting from land.
20 R. R. UNKS ET AL.
Here we explore the patterns presented above and discuss the ways in which socially
dierentiated pathways of beneting from collectively titled land are intertwined with
the changing political, social, economic, and ecological context of Ilkisongo Maasai
land, and how new diuse patterns of land control closely relate to both surplus
accumulation and social patterns of inclusion and exclusion that have emerged. Key
drivers of institutional changes during the colonial, post-colonial, and more recent
NGO-led wildlife conservation eras,
which have all shaped access to resources, can
be grouped as follows:
(1) State interventions, land appropriations, and collective tenure arrangements (GRs);
(2) Changes in land use designations and access rules due to GR governance, subdivision
of neighboring GRs, and internal wildlife conservation policies;
(3) Benet streams from cash cropping and wildlife conservation;
(4) Social changes that emphasize increasingly individualized practices and changing
(5) Market and technological changes that impact livelihood-based benets from land;
(6) The ability of local authorities to modify the access of others to land, and to treat col-
lectively titled land as private property;
(7) Reconguration of patterns of mutual assistance to increasingly focus on specic
relations (e.g. external political actors, NGO representatives, clans), leading to new
patterns of inclusion and exclusion from indirect benets of land.
The relationship between access and benets from land
The drivers listed above produced a number of institutional changes (Figure 3(b)) that
shaped patterns of dierentiated access (Figure 3(d)), which in part mediated dierential
household abilities to benet from land (Figure 3(f)). Due to a suite of changes in mobility,
markets, and social relations, accumulation of livestock required an array of new assets
that distinguished it from historical pastoralist practice (Figure 3(c)). Below, we emphasize
four key ways that the institutional basis of access interacted with these household assets
to shape direct benets from land (Figure 3(h)).
Firstly, access to forage was dependent on individual household endowments, and was
mediated by institutional changes at multiple scales. Access to forage outside of GRs during
times of scarcity was dierentially limited by cash assets and ability to pay for grazing lease
fees, labor, bribes, medicine, and water. As forage outside of GRs was primarily within pro-
tected or privatized areas, this access was also mediated by social relations that formed the
basis of arrangements to access (illicit and otherwise). While all had de jure and de facto
access to dry-season grazing within GRs, access to these areas was also heavily inuenced
by individual householdsabilities to provide water and supplemental feeds to livestock and
to transport livestock, and thus was highly socially stratied, according to household cash
assets, technology (e.g. veterinary supplies, vehicles), and social relations. Further, the ability
See Unks (2022) for a detailed discussion of the inuence of wildlife conservation on governance processes.
to sustain cattle through mobility within and outside of GRs to avoid exposure, and other
measures to decrease sensitivity to ecological stressors such as disease and wildlife(Figure 3
(i)) rested on household assets that were supplemented by assets gained from farming or
other household income (Figure 3(g)). In other words, the ability to benetfromlandwas
not just about access, but was also tied to a wider set of household endowments and inter-
related abilities to benet from multiple resources (i.e. ones extended entitlements set).
Changes in required entitlement sets, in turn shaped the highly dierentiated decision
making of households in response to social and ecological uncertainty and impacted
how livelihood stressors were experienced.
Secondly, institutional changes in social norms of mutual assistance led to increasingly
individualized responses to uncertainty. This placed heightened importance on individual
household assets (i.e. endowments), which shaped abilities to benet (i.e. entitlements),
and deepened social dierentiation in abilities to adjust to ecological variability and
new constraints on mobility. Changes in shared labor and resources among households
were linked to long-term social changes due to multiple factors. These included historical
interventions to end ilmurran practices (see Kituyi 1990) and national reforms in Kenya
that shifted the burden of education and social services onto households (Galaty 2013).
These had salient, cascading impacts on families who lacked household labor and/or
the nancial assets to hire labor needed to sustain livestock during droughts, with impor-
tant gendered implications. Changing social alliances along lines of wealth, political
aliation, clans, age-sets, and gender, and between a subset of Maasai and wildlife con-
servation NGO representatives also recongured patterns of mutual assistance. Further,
the ability to gain cash assets through employment, and through other indirect
Figure 3. Key factors shaping benet streams and control of collectively titled group ranch land. Boxes
e and h specify benets associated with the distinction of direct use rightsand indirect use rights
over land made by Sikor, He, and Lestrelin (2017).
22 R. R. UNKS ET AL.
benets from land (Figure 3(e)), was closely related to social alliances (e.g. especially
through clan and patron/client relations).
Thirdly, the ability to generate cash assets, and the ability to selectively interact with
markets, allowed some households to strategically use livestock markets to gain
benets while others were forced to sell livestock at inopportune times, such as when
market prices had crashed. Success in livestock keeping was closely linked to farming
success as well as indirect benets from land (Figure 3(e)), and assets gained which, in
turn, supported livestock production, highlighting the importance of the overlap in
dierent benet pathways represented in Figure 3. Farming outcomes were in great
part determined by the ability to raise cash assets for farm inputs, the ability to strategi-
cally interact with markets, to prevent or buer impacts of wildlife and pests, and to form
social relations that shaped plot allocation.
Fourthly, changes in mobility, labor, markets, and intensication of farming also inter-
acted with technology and knowledge to shape access to land, and had dierential liveli-
hood outcomes. Improvedlivestock breeds and cash crops brought new abilities to
benet from land for some, but also introduced new livelihood stressors through sensi-
tivity to scarcity of grass, livestock disease, crop pests, and wildlife herbivory in farms
(Figure 3(i)). These also required other technologies (e.g. chemical, transportation), and
new forms of knowledge about veterinary drugs, farm chemicals, and markets (e.g. that
of livestock brokers) for their successful use. They also required additional assets to
buer their increased sensitiviity, including increased need for labor, water, and secure
access to grazing in locations like private lands.
Accumulation, the politics of access control, and cascading socio-
economic dierentiation
The disparities in abilities to directly benet from land through livestock that we described
above were impacted by householdsvariable control of diverse assets that historically have
been central to livestock production, but also the ability to control a variety of new herding
assets (e.g. hired labor, grazing fees, supplementary livestock feed, transported water) .
These disparities in livestock production were compounded by abilities to gain assets
from farms, to benet from market interactions, and to gain indirect benets from land
(e.g. lease payments, employment income, bursaries that reduced education costs). Receiv-
ing indirect benets from land, in particular enhanced some householdsoverall abilities to
navigate institutions that mediated benets from both livestock and farming, enabling
them to generate surpluses in livestock, cash assets from sales of livestock and farm
produce, and to recursively reinvest these in the above pathways to maximize their pro-
duction. A narrow subset of households (notably, wildlife conservation NGO employees
and GR representatives) were able to reap new indirect benets from land, and were sub-
sequently able to exploit direct benets from land to a greater extent than other GR resi-
dents, deepening these dierences in abilities to access and benet from land. However,
the ability to navigate these overlapping benet pathways, and the overall accumulation
Importantly, it should not be concluded that previous social dierentiation did not exist (Scoones 2021). Historical
inequalities likely played a strong role in the assets households controlled at the time of study.
produced through these imbricated pathways, in turn also translated into enhanced abil-
ities to shape institutions and to enable diuse control over land (Figure 3(a)).
This ability to shape institutions was most apparent when considering the collective
land tenure of GRs and its associated governance structure. By law, GR representatives
had become gures who could modify access institutions, control institutions that
further shaped both direct and indirect benet pathways (Figure 3), and also serve as
direct arbiters in the distribution of indirect benet streams. The case of Ilkisongo
Maasai is somewhat distinct from observations by Mwangi (2007a) elsewhere in
Kajiado, due to interventions to prevent subdivision driven by concerns of sustaining
wildlife mobility that have sustained and extended collective land tenure and GR auth-
ority structure (Unks 2022). At the time of this study, the GRs considered remained collec-
tively titled, and GR representatives retained a highly uneven inuence over decisions
about rules of land use and access, including having the authority to make arrangements
concerning land use as if it were their private land. This role was enhanced by nego-
tiations that passed through their domain of authority, both as a result of decentralization
reforms intended to expand conservation, and due to increasing pressures for industrial
cement production and cash crop farming. At the same time, leasing of land by
these ocials had broadly limited the access of GR residents to community lands, and
leases for industrial use and wildlife conservation alike were viewed as land grabs
facilitated by, and primarily beneting, GR representatives (see also Jacquet 2017; Unks
et al. 2021).
Owing to their likely ability to reap nancial benets, GR representatives also stood as
clear examples of individuals who have been able to expand their direct and indirect
benets from land (i.e. their extended entitlements). These dierential abilities to control
benet pathways should not, however, be interpreted as themselves generating clear
social divisions. This dierentiation also does not negate the well-known social and/or redis-
tributional values of livestock within Ilkisongo society. Rather, lines of social dierentiation
depended on complex interactions between a range of endowments that specichouse-
holds had due to family, kin, clan, age-set, and stock-partner relations, but also new
patron/client relations with political, industrial, and non-governmental actors. GR represen-
tatives were mediators between various social, political, and economic interests linking local
communities with outside actors (see also Lund and Boone 2013;Lund2016). They worked
to maintain their relationships with outsiders while simultaneously rearming their auth-
ority and political support within GRs through their control over the distribution of
benets, especially indirect benets from land associated with wildlife conservation
NGOs, which they in turn used to arm support among social alliances. GR representatives
were also not uniformly wealthy, and some of the wealthiest members of GRs were neither
representatives nor political elites. Further, GR representativesdecision making was also
heavily inuenced by higher-level politicians, and representatives of wildlife conservation
NGOs and extractive industries. Through both the land tenure and governance system of
the GR, representatives were responsible for mediating negotiations between powerful
actors and the constituency they represented. While these various relationships likely
beneted them, it is vital to understand these ocials, as well as others such as chiefs, as
embedded within a complex socio-cultural context of alliances, where they served as
gate-keepers over numerous informal and formal rules of access that reinforced complex
24 R. R. UNKS ET AL.
benet pathways that were selectively benecial to some, and that enabled dierentiated
accumulation within the overall livelihood system.
The complex political and social alliances which indirect benets from land were dis-
tributed through generated cascading benet streams and solidied political and social
alliances, and recursively entrenched dierent abilities to shape benet pathways and
patterns of accumulation. The patterns of indirect benet distribution also were closely
related to important changes in norms of mutual assistance that shaped livelihood prac-
tices and labor relations. Distribution of indirect benets reinforced patterns of mutual
assistance by clan in particular, reinforcing processes of inclusion and exclusion. These
complex socially dierentiated processes of beneting from land, and accumulation of
assets, were also creating new diuse patterns of land control that do not neatly fall
under strict notions of enclosure or territorialization. Land control rather, was increasingly
consolidated through the ability to shape institutions that bridged social alliances, and
that enhanced dierent groupsabilities to include and exclude along lines of social dier-
ence. Further, these dynamics were mediated by institutions that cut across the varied
systems of access within the mosaic of dierent property relations (Robinson 2019) and
enabled some groups to preferentially pool their benets across a mosaic of tenure
types, including collective, where most land and resources did not t clear denitions
of private or common property. These patterns of land control, therefore, were best
understood through attention to the complex mixture of overlapping institutions that
dierentially mediated direct and indirect benets along lines of identity, social relations,
and wealth.
Similar to dynamics noted by Mwangi (2007b), Ilkisongo Maasai GR members had sup-
ported shifting property relations toward privatization (i.e. subdivision) in hopes of pursu-
ing individual advancement, and of addressing growing inequality, lack of transparency in
decision-making, and elite capture of benets from land. Subdivision was seen by many as
enabling new arrangements of sharing land between individual households. The benets
of creating new practices of land sharing and mutual assistance were expressed most
commonly among women, who were often excluded from the new indirect benets of
land (see also Unks et al. 2021) as well as from historical production and accumulation
(Hodgson 1999). However, subdivision seems extremely unlikely to deliver the promises
that it oered in terms of addressing inequity in benets from land, tenure insecurity, or
decision-making. Continued access to land has been observed by others in nearby former
Maasai GRs that have subdivided (BurnSilver and Mwangi 2007), but is limited by a range
of inequalities and structural relations within communities (Jeppesen and Hassan 2022).
As we discuss in greater depth elsewhere (Unks 2022) the emerging land tenure
system at the time of writing was being heavily inuenced by conservation NGOs, GR
representatives, and state authorities, and prioritized securing wildlife conservation corri-
dors and the land investment interests of a range of local, county, and national political
actors. GR representatives were playing a key role in the subdivision process and plot allo-
cation. The Community Land Act (CLA) of 2016 was intended to address well-known
conicts over the distribution of land rights along lines of gender and age-sets, and the
inherent problems of GR governance, and registration under the CLA would have
required subdivision to follow an alternative, more closely regulated process under
newly elected authorities (Alden Wily 2018). However, GR representatives have proceeded
with subdivision following the process that existed prior to the CLA, preempting the CLA
from taking eect, preventing dissolution of their authority, and sustaining their ability to
shape the subdivision process. In interviews, GR representatives had repeatedly stated
that non-Maasai groups would be able to claim Ilkisongo Maasai land if the GR was regis-
tered under the CLA, and this narrative likely helped them to gain political support within
GRs for subdivision prior to the CLA registration deadline (Unks 2022). Given that the ben-
eciaries of land distribution are registered GR members who are primarily male elders,
subdivision threatens to leave high numbers of women and men from younger age-
sets without titles, potentially deepening patterns of inequity in access and benets
from land. In addition to the uneven distributions of land that have been seen in subdivi-
sion process elsewhere nearby (Mwangi 2006; Rutten 2002), it was widely rumored that
GR representatives and county level authorities were also purchasing large expanses of
land from those unable to aord surveyors fees. If true, this will lead to them controlling
large swaths of grazing land, further deepening their ability to pursue enhanced accumu-
lation strategies, to dominate livestock production and markets, and also likely to deepen
their indirect benets from land through wildlife conservation and ecotourism activities.
Furthermore, several large (greater than 10,000 ha) parcels of land have been designated
as group conservancies across GR membersdesignated individual plots, that have been
secured through long-term leases with NGOs and that include lease conditions, payment
structures, and land use restrictions that appear to mirror existing conservancies. The land
tenure system taking shape at the time of writing, thus appears, to sustain many of the
characteristics of collective tenure that motivated subdivision of GRs.
Analyses of pastoralist livelihoods in East Africa have often focused on contrasts made
between private and common property regimes. While these analyses have drawn impor-
tant attention to how pastoralist institutions mediate and regulate resource use, they
have been less attentive to social processes of inclusion and exclusion, and the multiple,
socially dierentiated ways of beneting from and controlling land. In this paper, we con-
sidered the institutional basis of access to and benets from land among Kenyan Ilkisongo
Maasai, and how this system has undergone transformation through international and
national political and economic inuences. Drawing from a critical approach to the
study of institutions, we examined dynamics of access to and benets from land across
a complex mosaic of property relations. This allowed us to distinguish the socially dier-
entiated, multiple ways of beneting from land, and to explicitly connect these to overall
livelihood outcomes and patterns of accumulation in a context of high variability and
uncertainty. This showed how struggles surrounding distribution of benets from live-
stock, crop cultivation, and other new land uses such as wildlife conservation, are inter-
twined with changing socio-cultural norms such as mutual assistance and patronage,
social processes of inclusion and exclusion, asymmetrical inuence over land access
rules, and a complex interplay of national and international political and economic
forces. This elucidated how institutional changes that cut across a complex mosaic of
property relations in this rapidly transforming system have reproduced socially dieren-
tiated benet streams and cascades of accumulation that have generated diuse patterns
of land control.
26 R. R. UNKS ET AL.
We express our deepest gratitude to the people of Mbirikani, Eselenkei, and Olgulului-Ololorashi for
generously sharing their time and knowledge, teaching us about their lives, and for welcoming us
into their homes. This work would not have been possible without the companionship, guidance,
and assistance in translation of Charles Ntimama Mutunkei, Richard Solonka ole Supeet, and
Sakimba Kennedy Kimiti. The article beneted greatly from comments by Ian Scoones and two
anonymous reviewers.
Disclosure statement
No potential conict of interest was reported by the author(s).
The French Agence Nationale de la Recherche (ANR) supported the research program [ANR-16-
CE03-0004] MaGnuM. Writing was supported by the National Socio-Environmental Synthesis
Center (SESYNC) under funding received from the National Science Foundation [DBI-1639145].
Notes on contributors
Ryan Unks is a visiting fellow at the Robert Schuman Centre for Advanced Studies at the European
University Institute, and an aliate of the PASTRES (Pastoralism, Uncertainty and Resilience: Global
Lessons from the Margins) programme funded by the European Research Council.
Mara J. Goldman is an associate professor in the Department of Geography, and the director for the
Environment and Society Program a