FORPOL-01253; No of Pages 11

Forest Policy and Economics xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Forest Policy and Economics

journal homepage: www.elsevier.com/locate/forpol

Collaborative resource management and rural livelihoods around protected areas:

A case study of Mount Elgon National Park, Uganda
Charlotte Nakakaawa ⁎, Ricarda Moll, Paul Vedeld, Espen Sjaastad, Joseph Cavanagh
Norwegian University of Life Sciences Ås,Norway

a r t i c l e i n f o a b s t r a c t

Article history: Collaborative resource management agreements (CRMAs) have been introduced to improve people–park rela-
Received 30 August 2013 tions and enhance rural livelihoods. Based on household surveys, key informant interviews, focus group discus-
Received in revised form 30 March 2015 sions and a review of literature we investigated differences in livelihood assets, park dependence and livelihood
Accepted 8 April 2015
outcomes for households with and without CRMAs. We investigated the role of park environmental incomes
Available online xxxx
(PEI) in poverty alleviation and factors influencing dependency on PEI. Results revealed significant differences
Keywords:
in household assets but no significant differences in park dependence and livelihood outcomes. People report
Collaborative resource management that PEI on average constitute 18% of total incomes. Poor households have a higher dependency on PEI and it re-
agreements duces income inequality by 13%.
Park environmental income CRMAs have a significant positive effect on total PEI but no significant effect on total household income and
Livelihoods relative environmental income. In areas with CRMAs, taungya farming and bee keeping practiced as part of the
Poverty CRMAs increase annual household incomes by 26% and 28% respectively and constitute potential pathways out
Mount Elgon National Park of poverty. However, the impact of the agreements is still low due to their limited scale and coverage and the
Uganda
targeting criteria which limits access for communities with a high dependency on park resources and high levels
of conflict.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction
Managing protected areas (PAs) in many developing countries has
been based on a conservationist approach characterized by establishing
enclosures followed by exclusion of traditional local resource users
(Brockington and Igoe, 2006; Webster and Osmaston, 2003). These
policies accrue substantial costs and reduced benefits for local people
and have resulted in many conflicts between protected area managers
and local communities (Chhetri et al., 2003; Norgrove and Hulme,
2006). Costs relate to loss of livestock, crops and land and labor costs
related to guarding crops and livestock, and loss of previous accessible
benefits such as land and natural resources (Norgrove and Hulme,
2006; Vedeld et al., 2012). As a means of reducing conflicts and net
costs imposed on local communities by PAs, Uganda like many other de-
veloping countries, embraced community-based conservation (CBC)
approaches during the late 1980s (Beck, 2000; Musali, 1998). These
approaches sought to reduce conflicts between conservationists and
local communities by involving the latter in PA management and
compensating them for restricted access to land and natural resources.
However, in practice, protected area managers have been reluctant to
⁎ Corresponding author at: Norwegian University of Life Sciences, Ås 1430, Norway.
Tel.: +47 64965334.
E-mail address: charna@nmbu.no (C. Nakakaawa).
devolve power and authority to local communities, citing among others,
short term interest of resource use and their alleged lack of capacity
(both human and financial) to collectively manage natural resources
(Nsita, 2010; Ribot et al., 2006).
The empirical evidence on the extent to which CBC initiatives have
been able to achieve their intended environmental and livelihood
goals reveals mixed results. For a recent overview of outcomes from
CBC see Brooks et al. (2013) where they in particular stress project de-
sign and capacity building in local communities in generating success. In
addition, properties of local communities such as well functioning ten-
ure regimes and supportive local cultural beliefs and institutions are
conducive for good outcomes. Some success stories of CBC enhancing
forest recovery have been reported but this mainly relates to environ-
mental goals (Ellis and Porter-Bolland, 2008; Sassen et al., 2013). In
contrast, a limited number of CBC impact studies have focused on the
livelihood outcomes (Jagger, 2012; Jumbe and Angelsen, 2004; Ribot
et al. 2010, Vedeld et al, 2012). These studies also indicate mixed results
and emphasize context dependency. In some cases CBC appears to have
actually increased land-use conflicts (Baker et al., 2011). At Mount
Elgon National Park (MENP), some empirical evidence indicates a
relationship between conservation education and positive attitudes to-
ward the park (Oonyu, 2000) but most of the literature indicates that
despite efforts to secure community support, park management is still
characterized by very high levels of conflicts between park managers

http://dx.doi.org/10.1016/j.forpol.2015.04.002
1389-9341/© 2015 Elsevier B.V. All rights reserved.

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
2 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx

and local communities (Scott, 1998; White, 2002; Norgrove, 2002;
Chhetri et al., 2003; Jankulovska et al., 2003; Norgrove and Hulme,
2006; Himmelfarb, 2012). The challenge at Mount Elgon is how to pro-
tect important ecosystem values while meeting the livelihood require-
ments of a burgeoning human population.
Still, much is not known about how local people adapt living close to
PAs, to what extent they (still) depend on environmental incomes and
on what economic scale they have benefited from collaborative re-
source use agreements that “re-allow” PA resource access. An enormous
amount of research have investigated and documented the values of
environmental resources in rural livelihoods and poverty alleviation
(Cavendish, 2000; Vedeld et al., 2007; Babulo et al., 2008; Kamanga
et al., 2009; Mamo et al., 2007; Pouliot and Treue, 2013; Vedeld et al.,
2012). These studies show that environmental incomes in many cases
contribute substantially (8–45%) to the total annual household income
and that poor households in particular have a high dependency on en-
vironmental income. However, differential dependency on environ-
mental resources has seldom been understood in combination with
contextual factors, such as location and resource access restrictions,
which influence use and dependence on environmental resources
(Jumbe and Angelsen, 2004). With the exception of a few studies
(Das, 2010; Jagger, 2012; Jumbe and Angelsen, 2004; Tesfaye et al.,
2010) most of the investigations do not consider differences in institu-
tional context governing resource access and use. Empirical evidence
does indicate substantial differences in livelihood outcomes under
different institutional contexts governing resource access and use
(Tacconi, 2007; Blomley et al., 2010; Ribot et al., 2010).
Faced with looming conservation budget deficits and the prospect of
receiving additional funding through the REDD+ mechanism, conser-
vation agencies (UWA and NFA) now have incentives to further restrict
access to the park to secure permanence of the sequestered carbon.
Deforestation, forest degradation and illegal activities in the form of
agricultural encroachment and unlicensed timber harvesting have
continued unabated (NFA, 2009) especially in areas without CRMA
(Gombya-Ssembajwe et al., 2007; Mugagga et al., 2012: Petursson,
2011). Some scholars advocate for increasing resource access restric-
tions (Mugagga et al., 2012) and concerns have been raised about a po-
tential ‘recentralization’ of decision-making, with losses of community
rights and control over protected areas (Phelps et al., 2010; Sandbrook
et al., 2010). At Mount Elgon, in areas with CRMA, UWA retains the dis-
cretion to withdraw from the agreements when deemed necessary
(Sletten et al., 2008).
This paper thus evaluates the livelihood outcomes and the effect
of CRMA at MENP and differences in dependence on environmental
income by location. The livelihood analysis is based on differences
between participant and non-participant households by examining
the relevant measures of wealth based on the sustainable livelihood
approach (SLA) (Scoones, 1998).
1. Are there significant differences in access to assets, park dependence
and livelihood outcomes between communities with and without
CRMA?
2. What are the factors influencing dependence on park environmental
income?
3. What is the role of park environmental income in rural livelihoods
and poverty alleviation in communities with and without CRMA?
2. Study context
2.1. Study area
Mount Elgon forest was initially gazetted as a forest reserve in 1938
and later upgraded to a national park in 1993 (Scott, 1998). The Mount
Elgon National Park (MENP) covers an area of about 110,971 ha, located
in Eastern Uganda in eight (8) districts — stretching between 0° 52′ and
1° 25′ N and 34° 14′ and 34° 44′ E. A large portion of the park is located in
Bukwo (26%), followed by Kween (17%) and Bulambuli (13%). Consider-
ing the proportion of the total district area covered by the park, over 50%
of Bukwo and more than a third of the Bududa (41%) and Kapchorwa
(36%) districts are under protection (Table 1). The park has a long history
of human settlement and use (Scott, 1998) and is surrounded by farm-
land which is cultivated up to and within the park boundary (see Fig. 1).
The inhabitants around the MENP belong to two major ethnic
groups — the Bagisu and the Sebei. The Bagisu's livelihoods are heavily
dependent on agriculture and access to forest resources for subsistence
and commercial purposes. The park provides a broad range of goods
(fuel wood, medicine, construction materials, pastures and forest
foods such as bamboo shoots) and services (water catchment, burial
and circumcision sites) for both the Bagisu and the predominantly pas-
toral Sebei communities (Scott, 1998; Gosalamang et al., 2008; Katto,
2004; Namugwanya, 2004; Norgrove, 2002). The main crops grown
include coffee, bananas, beans, maize, wheat and potatoes. The fertile
volcanic soils and abundant rainfall (annual mean rainfall range of
1500–2000 mm) support a high population of about 1.6 million people.
The rapid population growth (3.4% per year), high population
densities — ranging from 116 to 827 persons/km2 in the Kween and
Mbale districts respectively (Table 1) and the increasingly small agricul-
tural plots demonstrate the increasing pressure on land and park re-
sources. This is manifested in both increasing legal and clandestine
access to PA resources (Gombya-Ssembajwe et al., 2007) and also the
increasing encroachment for farmland and settlements (Mugagga
et al., 2012; Petursson, 2011; UWA, 2000). This has been a response

Table 1
Park, people and collaborative management agreements by districts, Mount Elgon National Park (MENP), as of 2013.

District Mount Elgon National Park Population Type and number of active agreements

Name Park area Boundary length Total Density

Hectares % of district area km % Count of persons'000 Persons per km2 CRMAs BMAs BKAs BSCAs Total

Bukwo 28,926 55.1 108 16.3 70,5 134 0 0 0 0 0

Bududa 10,386 41.4 105 15.9 173,7 693 0 0 1 0 1
Kapchorwa 12,912 36.4 67 10.2 109,3 308 8 8 6 2 24
Kween 18,587 21.8 103 15.7 98,9 116 0 0 2 2 4
Sironko 10,222 22.9 81 12.3 233,5 524 2 2 2 5 11
Manafwa 12,983 21.6 55 8.4 355,4 590 0 0 0 0 0
Bulambuli 13,905 21.3 89 13.5 122,3 188 6 6 3 1 16
Mbale# 3050 5.9 51 7.7 428,8 827 0 0 2 0 2
Total 110,971 659 100 1,592,400 379 16 16 16 10 57

Data sources: Park area computed from the IUCN, 2011 and NFA, 2010 databases; population density calculated by dividing the 2010 district land area (NFA, 2010) by the 2011 population
estimates from UCC (2010); CRMA data was compiled from Cavanagh (2009, 2011), Moll (2011) and Hoefsloot et al. (2011); #Mbale had no active CRMA at the time when data for this
study was collected.
Key: BMA — Boundary Management Agreements, CRMAs — collaborative resource management agreements, BSCAs — Bamboo Shoot Collection Agreements; BKAs — Bee Keeping
Agreements.

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx 3

Fig. 1. Location of Mount Elgon National Park in Uganda.

to the dire need for land and other resources required to secure daily
livelihoods among natural resource dependent communities.
In addition to serving important local livelihood functions, the
Afro-mountainous flora and fauna have high national and global values
related to biodiversity conservation, carbon sequestration, tourism
development, education and research (Howard, 1991; UWA, 2000).
The park was ranked among the top ten most species rich forests
in Uganda (Davenport et al., 1996) hosting 39 endemic higher plant
species and 37 fauna species (of which nine are endemic) which are
listed as globally threatened and in need of conservation (IUCN, 2005).
MENP is also a transboundary resource (Petursson, 2011), serving as a
major water catchment area for important water systems contributing
to the Turkwell River and Lake Turkana in Kenya and to the Lake
Victoria and Nile River basins in Uganda (IUCN, 2005; Van Heist, 1994).
The park has suffered widespread encroachment dating back to the
early 1970s to date (White, 2002; Sassen et al., 2013). Park management

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
4 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx
has been riddled with conflicts related to contested unclear bound-
aries and resource access restrictions (Norgrove and Hulme, 2006;
Himmelfarb, 2005, 2012), although various initiatives have been im-
plemented to restore, sustainably manage and conserve the forest in
order to meet both local livelihood objectives and global conservation
goals (UWA, 2000, 2009). As a way of explicitly acknowledging humans
and human interests in the conservation landscape, MENP was declared
a Man and Biosphere Reserve in 2005. This was part of the global at-
tempts to reconcile environmental protection with sustainable develop-
ment goals. Since the early 1990s, CRMAs were initiated to reduce
conflicts (Chhetri et al., 2003). However, the different initiatives to date
have failed to alleviate the tension and conflicts between the UWA and
the local communities (Norgrove and Hulme, 2006; Cavanagh, 2012).
The Mount Elgon region is also susceptible to natural calamities such
as floods, landslides, drought and famine and recent studies reveal that
climate change is expected to result in an increase in the incidence of
such disasters (Mugagga et al., 2012). Similarly, the frequency of crop
failure and famines is expected to increase. Access to park resources in
such periods constitutes a vital de facto ‘safety net’ and gap filling coping
strategies for poor households (Angelsen and Wunder, 2003). Access
to park resources through CRMA can help prevent poor households af-
fected by landslides or floods from sinking deeper into poverty during
such difficult times.
2.2. Collaborative resource access, management and use agreements
at MENP
CRMAs in Uganda are implemented within a decentralized gover-
nance framework (Government of Uganda, 1997) in which local gov-
ernments are organized into a five-tier system of Local Councils (LCs);
from level one (LC1 — the villages) to level five (LC5 — the Districts).
The District (LC5) is the highest governance unit that links up with
the central government. Below the District is the County or Municipality
Council (LC4) in the rural and urban settings respectively, followed by
the sub-county (LC3) which is the basic unit of local government —
both political and administrative (Turyahabwe et al., 2007). Below the
LC3 are the Parishes (LC2). The CRMAs are established at the parish
level.
At Mount Elgon, the CRMAs were initiated by IUCN in 1992 to reduce
conflicts and improve people–park relations in communities directly
bordering MENP. The first CRMAs were signed in 1996 with the
Mutushet and Kapkwai parishes in Kapchorwa district (Beck, 2000;
Musali, 1998; Scott, 1998; White, 2002). As of March 2013, Kapchorwa
had the highest number of agreements followed by Bulambuli and
Sironko while conflict ridden communities in Bukwo and Manafwa
had none. To date there are no clear criteria for establishing CRMA. So
far, CRMAs are established in parishes with low levels of conflict and
where communities are able to maintain good relations with the UWA
staff. Implicitly, these are communities which are less dependent
on park resources (Cavanagh, 2009; Moll, 2011). This has resulted in
a highly uneven distribution of the CRMA (Table 1). The agreements
give communities rights to access specific park resources such as fire-
wood, medicinal plants and bamboo shoots (Bamboo Shoot Extraction
Agreements) or to undertake specific activities such as bee keeping
(Bee Keeping Agreements) and Taungya farming (Boundary Manage-
ment Agreements) within the park boundaries.
3. Methods
3.1. Data sources and sampling procedure
Primary data were collected in two phases — January to March 2011
(Moll, 2011) and June to December 2011 (Cavanagh, 2011). Data col-
lection involved mixed qualitative and quantitative methods which in-
cluded a household survey, key informant interviews and focus group
discussions.
Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
In order to reduce sampling error due to sample heterogeneity
(Angelsen et al., 2011), the target population was restricted to the
Bagishu communities covering four districts (Bulambuli, Sironko,
Mbale and Manafwa) around Mount Elgon National Park in Eastern
Uganda. Within each district, one sub-county was purposively selected
based on prior knowledge of presence of parishes directly bordering
MENP with or without collaborative resource access agreements
(CRMAs). Two parishes with and two parishes without CRMA were
chosen which were all bordering the park, easy to access and had a sim-
ilar duration of the agreements. Within the parishes, relevant villages
were identified with the help of local chairpersons contacted prior
to conducting the household survey. At the village level, in the absence
of current, reliable census records and the resources to construct an
exhaustive sampling frame, household sampling was based on a sys-
tematic random sampling procedure — every third household on a
predetermined route through the village was selected. Twenty (20)
households were selected from each of the four study sites. In total 80
household interviews were conducted to collect largely quantitative
household socio-economic data from 40 participant and 40 non-
participant households (Moll, 2011).
The household survey was augmented by key informant interviews
held with the UWA staff responsible for negotiating the agreements
with communities, focus group discussions, key informant interviews
and semi-structured interviews with various stakeholders in the com-
munities and Community Based Organizations (CBOs) involved in
implementing the CRMA (Cavanagh, 2011).
To compliment and triangulate the primary data, we used secondary
data and performed a literature review to identify other relevant studies
conducted on MENP governance, livelihoods and dependence on
park environmental incomes. The main sources of information include
Master's theses (8 from UMB including Katto, 2004; Kiggundu, 2007;
Luzinda, 2008; Namugwanya, 2004; Sletten, 2004), PhD studies (Beck,
2000; Norgrove, 2002; Muggaga, 2011; Petursson, 2011) and other as-
signments conducted by researchers from other institutions (Hoefsloot
et al., 2011). Similarly, the methods used in these studies included
household surveys, focus group discussions, key informant interviews
and semi-structured interviews. The paper also draws on other field
studies on livelihoods conducted between 2009 and December 2011
(Vangen, 2009; Cavanagh, 2009, 2012) as part of a broader research pro-
ject (2007–2011) in East Africa on Protected Areas and Poverty (PAPIA).
These studies investigated the impacts of protected areas on forest
dependent communities. We also draw on our extensive experience
from other assignments in the area over the last 10 years (Baatvik
et al., 2002; Larsen et al., 2008; Vedeld et al., 2005).
3.2. Data analysis
Data management and analysis were conducted using Excel, Stata
and SPSS version 19.
3.2.1. Estimating household income
The income accounting methods used in this study draw on Vedeld
et al. (2007) and Angelsen et al. (2011). According to Angelsen et al.
“rural household income includes three broad components:
• Value added from self-employment, for example, in agriculture, for-
estry or other business;
• Wage earnings and rents from renting out land or other forms of
capital;
• Transfers, for example, remittances and pensions” (Angelsen et al.,
2011).
The total household income was calculated as a summation of the
total household subsistence and cash income from farm (crop and live-
stock), park dependence (park environmental income), off-farm and
non-farm activities, and income derived from remittances.
 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx
Farm income included cash and subsistence income from crop and
livestock products. Cash income included income from the sale of crop
and livestock products while subsistence income was estimated as the
value of products directly consumed by the household or given away
to friends and relatives as gifts, multiplied by the local unit price.
Input costs for fertilizers, pesticides, seeds, hired equipment, hired
labor, ropes and transportation were deducted to obtain the net value.
Family labor was not considered as a cost due to the thin labor market.
Changes in the value of crop land or livestock were excluded from in-
come calculations (Angelsen et al., 2011).
Off-farm, non-farm income and remittances were self-reported during
the household interviews. Off-farm income included cash income
earned from engaging in farming activities on someone else's land or
from herding other people's cattle while non-farm income included
cash income derived from employment or self-employment that is not
related to farm activities, such as teaching, petty trade, formal employ-
ment, riding a ‘boda-boda’ (local term for motorcycle) or trading with
agricultural inputs or outputs.
3.2.2. Forest (park) dependence: absolute and relative (park) environmental
income
“Environmental income (EI) is the capture of value added in ex-
change or consumption of natural capital within the first link in a mar-
ket chain, starting from the point at which the natural capital is
extracted or appropriated” (Sjaastad et al., 2005). Forest dependence re-
fers to the extent to which local communities use and depend on forest
(park) resources. The actual magnitude of such dependence is referred
to as ‘absolute environmental income’ and when expressed as a share
or proportion of total household income, this share is referred to as
the ‘relative forest environmental income’ or more specifically as the
‘relative park environmental income’ if the accounting is restricted to
products obtained from park resources (Cavendish, 2000; Vedeld
et al., 2007; Tumusiime et al., 2011). In this paper, we focus on park en-
vironmental income (PEI) due to high dependency on Mount Elgon Park
resources as a result of limited tree cover beyond the park boundary
(Sassen et al., 2013). To estimate PEI, we used household self-reported
quantities used or sold and average local market prices. During the
household survey, respondents reported the weekly, monthly or annual
amounts of the different products harvested/gathered, consumed/used,
given away to other households (gift) and sold. Weekly and monthly
estimates were aggregated to make annual income estimates. The
main source of PEI in this study was firewood due to the existing re-
source access restrictions. Illegal dependency has been reported but
due to difficulty in estimating illegally extracted resources, these were
not included. Other services such as water – with missing markets –
were also not included.
Total household income was then determined as a summation of
farm income, PEI, off-farm income and remittances. In order to make in-
come comparisons across households, household size was converted to
adult equivalent units (AEU) as described in Cavendish (2000). The total
household income was divided by AEU to control for differences in
household age and sex composition. We then generated arbitrary pov-
erty groups, based on both total household income and total household
asset values (Nielsen et al., 2012). We ranked the households from the
lowest to highest total household income and asset values, classified
them into three different categories and investigated whether there
were any significant differences in household classification. Results re-
vealed no significant differences in the classification of the sample
households — that is, asset poor households were also found to be in-
come poor. We thus chose to use annual household income to split
the sample into three income groups — characterized as poor (27), me-
dium (27) and rich (26).
3.2.2. Estimating the income poverty line
Uganda's poverty line is estimated using the basic needs approach
which is based on the cost of obtaining the minimum daily per capita
Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
5
food (calorie) requirements for adult males (Appleton, 2001). The
poverty line has two components — a food poverty line, which is the
cost of a food basket that can meet the calorie requirements and an
allowance for non-food requirements. Drawing on Appleton (2001),
Yamano et al. (2004) estimated the food poverty line based on data col-
lected during the 2003 REPEAT survey. The food poverty line was esti-
mated based on the cost of a ‘food basket’ containing 39 food items
consumed by the poorest 50% of the survey population. The non-food
requirements were estimated as the predicted non-food spending
of a household with total consumption just sufficient to meet the cost
of recommended 3000 kcal per capita per day for adult males. The
3000 kcal per day was deflated with adult-equivalents to control for dif-
ferences in household composition (age–gender differences) and corre-
sponding differences in food requirements within households (Yamano
et al., 2004). The resulting poverty line is held constant in real terms
and compared with incomes from different years, using the consumer
price index (CPI) to deflate nominal incomes. Similarly, in this study,
after adjusting household size for sex and number of people of differ-
ent ages and correcting for inflation using a 2010 consumer price
index of 145.17 as the deflator to cater for changes in the cost of living
(UBOS, 2011), we estimated a poverty line of Uganda Shillings 155,224
per person per year (primary data used in this paper was collected in
January–March 2011).
Using this poverty line as the cut-off point, we grouped the sample
into two income groups — those below the poverty line as ‘poor’ and
those above as ‘non-poor’. According to this grouping, about 33.8%
(27 households) of the sample households were poor. We considered
this estimate to be reasonable as it is comparable to the national
(31.1%) and regional (headcount poverty index for eastern region —
35.9%) estimates of Uganda's population living in poverty (UBOS,
2011). The World Bank currently uses a cut-off point of $1.25/day
(Ravallion et al., 2009), but using this value in this study is not very
meaningful because of the variability in costs of acquiring basic liveli-
hood items and the predominantly subsistence livelihood patterns of
most of the households in the study area.
3.2.3. Investigating factors influencing dependency on park environmental
income (PEI)
We used multiple regression models to investigate factors influenc-
ing dependency on park environmental income (PEI) in absolute and
relative terms. We expected differences in dependency on PEI caused
by differences in livelihood assets and other contextual variables
(Kamanga et al., 2009; Tumusiime et al., 2011).
Human capital assets were estimated based on household size and
composition measured in adult equivalent units to cater to differences
in age and sex of the household members; and consumer–worker ratios
as proxies for household demands, potential consumption and availabil-
ity of labor required to engage in park extraction activities. Other human
capital variables assumed to influence household dependency on PEI in-
clude sex of the household head and spouse and their education levels
(as proxies for access to information, knowledge, and skills).
Physical capital assets include size of land and number and type of
livestock (expressed as Tropical Livestock Units — TLU). These variables
have been found to influence dependence on environmental resources
through their impact on labor and time allocation decision making pro-
cesses (Mamo et al, 2007; Kamanga et al., 2009; Kar and Jacobson,
2012). From this literature, we expected a positive or negative relation-
ship between these variables and park dependency because both agri-
culture and livestock production systems constitute an important
source of alternative income relieving pressure on park resources or
the park may be an important source of grazing resources increasing
park dependency especially during the dry season (Katto, 2004). The
monetary value of household assets was estimated by summing the
value of their land holdings (based on average prices) and the respon-
dents' reported monetary value of personal belongings, machines and
livestock.
6 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx

Contextual factors such as location (captured as distance to the park)
and institutional arrangements governing resource access and use such
as participation in CRMA are also expected to influence forest depen-
dence (Jumbe and Angelsen, 2004; Katto, 2004; Tesfaye et al., 2010).
Households from locations with CRMA have access to park resources,
meaning higher absolute and relative forest incomes. Distance to the
park boundary was included to capture location specific variations
which may influence dependence on park resources and to control for
variations in park dependence as a result of proximity to the resource
(Abebaw et al., 2012; Mamo et al., 2007).
We tested for correlations between variables using Pearson's corre-
lations and ran regressions to investigate the factors affecting household
incomes and dependency on park environmental income. Since our in-
come data does not have a normal distribution, the Kruskal–Wallis
test — a non-parametric version of ANOVA, was used to compare ranked
sample means for the different locations and income groups. For signif-
icant variables, Bonferroni tests were used for pair-wise comparisons
between means of different locations and income groups. Pair-wise
comparisons between households with and without CRMA were made
using standard Student's t-tests.
3.2.4. Investigating the role of environmental income and CRMA in rural
livelihoods
We analyzed variations in dependency on environmental resources
by the three income groups, location and participation in CRMA and in-
vestigated the role of environmental income (EI) based on variations in
the Gini coefficients and the Foster–Greer–Thorbecke (FGT) class of mea-
sures on poverty prevalence, depth and severity “with” and “without”
PEI (Vedeld et al., 2007). Poverty prevalence refers to the proportion
of households whose income is below the poverty line. It provides
an indication as to whether a particular intervention significantly im-
pacts the income or consumption potential of the poor. Poverty depth
refers to the income shortfall as a proportion of the poverty line while
severity is a measure of the variation in income distribution among
households below the poverty line. Accounting for purchasing power
parity, the poverty line estimated in 3.2 was used to calculate the
poverty indices.
Gini coefficients were used to estimate the impacts of environmental
income on inequality. The coefficient is a commonly used measure of in-
come inequality and indicates the distributional effects of environmen-
tal income. A Gini coefficient of zero denotes perfect income equality,
while a coefficient of one implies that all income is held by a single indi-
vidual (Ray, 1998). Gini coefficients were calculated for participants and
non-participants with and without environmental income.
4. Results and discussion
We first analyzed peoples' access to assets by income group, location
and if they participated in CRMA or not. We then looked at the same fac-
tors in relation to household income, before we did an in depth analysis
of poverty and on dependence on environmental incomes.
4.1. Access to livelihood assets
There were significant differences in human capital assets. House-
hold heads and spouses in rich households have more education than
poor households (p = 0.0099 and p = 0.05 respectively). Medium
and rich households have similar human and physical capital assets.
Rich households have access to more physical assets; such as the num-
ber of plots and size of land (Table 2). In general, households have
limited access to land (0.13 ha/AEU). Previous studies in the Mount
Elgon area (Nayenga, 2003) and other areas with fertile land amid exten-
sive protected areas (Naughton-Treves et al., 2011; Tumusiime et al.,
2011) indicate similar findings.

Table 2
Household assets by income group, Mount Elgon National Park, Uganda, 2011.

Variables (units) Sample Income group

Mean Poor Medium Rich

Observations (n) 80 27 27 26

Human capital
Household head's age (years)⁎⁎ 46.98 50.63a 39.93b 50.5ab
Household head's education (years in school)⁎⁎⁎ 4.84 2.78a 6.31bc 5.5bc
Spouse's education (years in school)⁎⁎ 4.14 2.81a 5.00b 4.90b
Household size (count of persons) 6.25 6.37 6.26 6.12
AEU 5.03 5.33 4.90 4.86
Consumer–worker ratio 2.05 1.89 2.21 2.06
Household head male (percentage) 75.00 26.00 26.00 23.00

Physical capital
Number of plots (count)⁎⁎ 2.80 1.81a 2.52a 4.04b
Household land (total hectares)⁎⁎⁎ 1.45 0.82a 1.18a 2.30b
Land (hectares/AEU)⁎⁎⁎ 0.33 0.19a 0.25a 0.55b
Livestock (TLU) 1.34 1.28 1.25 1.49
Value of household assets (UGX)⁎⁎⁎ 4,557,589 2,527,277a 4,038,689b 7,204,847c

Household external and contextual variables

Distance from the park
Kilometers 1.43 1.45 1.25 1.61
Minutes⁎⁎ 23.19 21a 18.52ab 30.31ac

Institutional variables
Participation in CRMA (count (%))⁎⁎ 40 9 (23)a 13 (33)ab 18 (45)bc
Participation in BMA (count (%)) 6 (30) 6 (30) 8 (40)

Outcomes
Household income
Total⁎⁎⁎ 1,765,609 417,077 1,556,526 3,383,133
Per AEU 314,464 45,596 219,335 692,460

Significance tests: ANOVA for continuous variables followed by Bonferroni tests for pair wise comparisons between means for different income groups. Observations followed by the same
letter in rows were not significantly different. Categorical variables — Chi square. We present the counts and percentages of respondents in each group.
⁎⁎ Significant at 5%.
⁎⁎⁎ Significant at 1%.

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx 7

Table 3
Asset distribution and household external factors by household location and participation in CRMA, Mount Elgon National Park, Uganda, 2011.

With CRMA n = 40 Without CRMA, n = 40

Variable District Sironko Bulambuli All with CRMA Mbale Manafwa All without CRMA

Sub county Bugitimwa Buginyanya Bubyangu Tsekululu

Human capital
Male headed HHs Count (%) 18 (90) 18 (90) 36 (90) 19 (95) 20 (100) 39 (97.5)
HH head's age Years 41.5 (3.28) 46.6 (3.6) 44.05 (2.44) 47.85 (3.67) 51.95 (4.17) 49.9 (2.76)
HH head's education⁎⁎ Years 4.6 (0.57)a 6.8 (0.89)b 5.7 (0.55) 3.21 (0.98)a 4.65 (0.74)a 3.95 (0.61)
Spouse's education Years 4.63 (1) 3.53 (0.75) 4.06 (0.62) 4.88 (0.75) 3.63 (0.6) 4.22 (0.48)
HH size AEU 4.51 (0.43) 4.76 (0.5) 4.64 (0.33) 5.38 (0.57) 5.48 (0.67) 5.43 (0.44)
Consumer–worker ratio Ratio 2.28 (0.25) 1.86 (0.12) 2.07 (0.14) 2.18 (0.19) 1.89 (0.17) 2.04 (0.13)

Physical capital
Own land⁎⁎⁎ Total hectares 0.91 (0.15)b 0.64 (0.11)b 0.77 (0.1) 0.52 (0.12)a 0.29 (0.08)a 0.40 (0.07)
Own land/AEU⁎⁎⁎ Hectare/AEU 0.21 (0.04)b 0.16 (0.03)b 0.18 (0.02) 0.11 (0.04)a 0.06 (0.01)a 0.08 (0.02)
Plots⁎⁎⁎ Count 3.6 (0.5)b 3.6 (0.6)b 3.61 (0.39) 2 (0.28)a 1.95 (0.32)a 1.97 (0.21)
Livestock TLU 1.21 (0.2) 1.63 (0.26) 1.42 (0.16) 1.1 (0.24) 1.42 (0.29) 1.26 (0.19)
HH asset value⁎⁎⁎ UGX (mill) 4.63 (0.75)b 8.30 (1.3)c 6.47 (0.80) 2.56 (0.66)a 2.74 (0.6)a 2.65 (0.45)

Household external factors

Count (%)
Poverty status (⁎⁎) Poor 4 (20)b 6 (30)b 10 (25) 8 (40)a 9 (45)a 17 (42.5)
Medium 6 (30) 6 (30) 12 (30) 7 (35) 8 (40) 15 (37.5)
Rich 10 (50) 8 (40) 18 (45) 5 (25) 3 (15) 8 (20)
Distance to park⁎⁎⁎ Meter 2330.5 (492.12) 1265 (163.6) 1797.75 (269.8) 1067 (195.8) 1070.5 (149.74) 1068.75 (121.7)
Minutes 36.4 (5.71) 19.35 (2.78) 27.87 (3.42) 21.7 (2.78) 15.3 (2.8) 18.5 (2.02)

HH = Household.
Significance tests: ANOVA for continuous variables followed by Bonferroni tests for pair wise comparisons between means for different income groups. Means are given followed by
the standard errors in parentheses. Observations followed by the same letter in rows were not significantly different. For categorical variables — Chi square. We present the counts and
percentages of respondents in each group.
⁎⁎ Significant at 5%.
⁎⁎⁎ Significant at 1%.

A majority (86%) reports to own less than 1 ha of land and they use
all the land for crop and livestock production. The average size of land
reported in this study is much lower than estimates from earlier studies
for instance in Mbale — 0.11 vs. 0.26 ha per capita (Namugwanya, 2004)
and Manafwa — 0.06 vs 0.25 (Mugagga, 2011). These results could be
indicative of the increasing land scarcity due to the exponential popula-
tion growth. It helps explain the increasing pressure on the park evident
from recent (1995 — 2006) encroachment and clearing of forest cover to
make way for agriculture (Mugagga, 2011; Peturrson et al., 2013). The
situation is exacerbated by the prevalent land politics through which
communities are continuously enticed to encroach on the park by
local politicians (Nsita, 2010).
Categorizing the data by location (captured by district and distance
to the park) and participation in CRMA (Table 3) reveals interesting
differences. There is a significant difference in land access and Sironko,
the location farthest way from the PA, has three times higher land access
than that in Manafwa, the location closest to the park. The number of
poor people is also much higher in Manafwa.
We also analyzed asset access by CRMA participation. People with
agreements have higher education levels, more land, higher asset
values, are rich and for some reason, they also live farther way from
the park. This can be explained by the fact that UWA avoids making
agreements with communities which have high levels of conflict and
that conflicts are more frequent with communities very close to the
park and that have high dependence on park resources (Katto, 2004;
Moll, 2011; Cavanagh, 2011).
4.2. Livelihood strategies and income outcomes
Does total income and various income sources vary by income
group, location and participation in CRMA?
4.2.1. Income sources and diversification by income group
The most important livelihood strategy and the primary source of in-
come is agriculture (crop and livestock) contributing about 48% of the
total annual household income (Table 4). Other sources of income
include off-farm activities (23%) and environmental income which
constitutes (18%) of the total household income. This estimate is com-
parable to other studies conducted in the MENP area (Katto, 2004;
Namugwanya, 2004; Mugagga, 2011). The main crops grown include
maize, cassava, coffee and beans. Other crops include Irish potatoes,
onions, passion fruits, tomatoes, peas, millet, wheat, soybean and le-
gumes. The key constraints to agricultural production include land
shortage and declining soil fertility. There is limited use of fertilizer.

Table 4
Household annual income (UGShs/AEU) by income group, Mount Elgon National Park, Uganda, 2011.

Household income source Sample mean Income group

n = 80
Poor Medium poor Rich
n = 27 n = 27 n = 26

Agriculture* 180,198.0 (48.12) −25,212.8 (−41.69)a 106,615.2 (39.24)a 469,922.2 (58.2)b

Off-farm activities* 109,802.3 (29.32) 31,363.5 (51.86)a 102,429.4 (37.69)ab 198,914.5 (24.65)b
Environmental resources 67,886.5 (18.13) 53,623.5 (88.67) 61,196.6 (22.52) 89,645.3 (11.10)
Remittances* 16,626.9 (4.44) 701.6 (1.16)a 1491.6 (0.55)a 48.882.5 (6.05)b
Total income* 374,513.7 (100) 60,475.7 (100) 271,732.80 (100) 807,364.5 (100)
Number of income sources 2.59 2.59 2.37 2.81

1 USD = 2350; * indicates significant differences in mean income across income groups at p b 0.05. In parentheses are relative incomes in percentages. Bonferroni tests for pair wise
comparisons between mean incomes for different income groups — observations followed by the same letter in rows were not significantly different.

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
8 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx

Increase in agricultural output is mainly achieved by expanding the area Table 6

under cultivation. This greatly increases demand for new land for cultiva- Socio-economic factors influencing absolute environmental income, MENP, Uganda, 2011.

tion manifested in the increasing land use conflicts emanating from the Absolute environmental Coefficient Robust std t value Pr(Nt)
high levels of park encroachment (Mugagga, 2011; Sassen et al., 2013). income error
Livestock income is derived from local chicken, cattle and goat Intercept −149,559.4 63,920.5 2.34 0.022
breeds. Other animals include pigs and turkeys. Off-farm income (29%)
Human capital
contributes much to total household income. This was partly attributed Household size (AEU)⁎⁎⁎ −041,149.4 10,331.6 3.98 0.000
to the great losses from the 2010 landslide which buried crop fields Education of HH −012,613.2 07,682.3 −1.64 0.106
and resulted in the death of many animals. Typical off-farm income gen-
Physical capital
erating activities include shop-keeping, providing local transport – ‘boda Size of land (hectares)⁎⁎ −205,972.4 90,756.8 −2.27 0.027
boda’ riding – and casual laboring. These activities are carried out all year
round and options were not affected by the landslide. Financial capital
Total household income⁎⁎ −0.063514 0.03087 2.01 0.044
Concerning distribution, rich households have higher incomes from
all kinds of activities compared to the other groups. In addition, they Household external factors
have higher relative incomes from agriculture in particular and lower Distance (m)⁎⁎ −30.26008 15.0917 −2.01 0.049
Participation in CRMA⁎⁎ 192,658.7 73,188.7 2.63 0.011
dependence on environmental incomes (lower relative environmental
income). Poor people in the sample have lost their agricultural incomes R-sq = 0.28; adjusted R = 0.21.
⁎⁎ Significant at 5%.
through landslides and depend on environmental incomes in particular ⁎⁎⁎ Significant at 1%.
indicating its important safety function. They also have higher relative
off-farm employment incomes — working for others. Rich households
have access to significantly higher remittances than the medium and without CRMA have somewhat higher PEI (but non-significant)
poor households. (Table 5). This may be the case because they are situated closer to the
Although we found no significant differences in absolute or total en- park.
vironmental income (Table 4) between income groups, as revealed in Recalling from Table 3, households in Bugitima have access to almost
other studies (Babulo et al., 2008; Tumusime et al., 2011), poor house- three times and twice as much land as those in Bubyangu and Tsekululu
holds have a higher dependence on or relative park environmental in- respectively. In addition, they participate in CRMA which provide them
come (PEI). PEI thus contributes to over 88% of the poor people's total the opportunity to utilize park boundary land for annual crop cultiva-
annual household income compared to 23% and 11% for the medium tion through the Taungya system. The crops grown are mainly for in-
and rich households respectively. As revealed in Table 2, rich house- come generation. Households in Bugitima earned significantly high
holds have access to more livelihood assets (more labor and higher cash (p = 0.0002) and subsistence (p = 0.0021) incomes. However,
educational levels and more physical assets such as land). This enables since both Bugitima and Buginyanya were participating in CRMA, and
these households to diversify their income activity portfolio. As a result, we found significant differences between the two locations but no sig-
income diversification is significantly higher among the rich whereas nificant differences between Buginyanya and locations (Bubyangu and
no significant differences were found between the medium and poor Tsekululu) without CRMAs, we could not attribute the differences in
households (Table 4). income to participation in CRMAs. However, we do not rule out this
possibility as studies conducted by Cavanagh (2009) among Taungya
4.2.2. Income sources by location permit holders in Kapchorwa district revealed that such households
There is a significant difference in total income across locations in earned an additional annual income of USD121.68 equivalent to 26.5%
that Sironko has higher total incomes than the others and in particular of the total average annual household income.
higher than Manafwa (more than twice the size). It is particularly agri-
cultural incomes and off-farm incomes that are higher. However, there 4.3. Explaining factors influencing PEI
are no statistically significant differences among the sources of income
between locations. Agriculture accounts for around 40–50% of the in- The regression analysis revealed that the common significant human
comes, environmental incomes 13–26% and off-farm incomes 25–32% and physical capital variables were the household size and the size of
of total incomes (Table 5). land owned (Table 6). We observed a significant positive correlation be-
tween human (household size) and financial capital (absolute park en-
4.2.3. Income and diversification by participation in CRMA vironmental income) variables. Large households have access to more
Analyzing income sources by participation reveals some important labor to engage in other income generating activities or labor intensive
differences. Households that participate in CRMA have much higher forest extractive activities, such as illicit timber extraction. Though not
total incomes, but agricultural, off-farm and environmental incomes are significant in our results, education offers knowledge, skills and employ-
also higher. The relative incomes however, reflect that households ment opportunities for engaging in alternative livelihood strategies

Table 5
Absolute and relative household annual income (UGX/AEU) by location and participation in CRMA, Mount Elgon National Park, Uganda, 2011.

Participation status

With CRMA Without CRMA (n = 40)

Location Sironko Bulambuli All with CRMA Mbale Manafwa All without CRMA

Bugitima Buginyanya Bubyangu Tsekululu

Agriculture
Off farm activities
Environmental resources
Remittances
Total annual household
289,470.8 (50.4) 131,731.7 (39.2) 210,601.3 (46.3) 167,623.6 (48.7) 131,966.0 (54.2) 149,794.8 (51.0)
179,195.8 (31.2) 106,658.0 (31.7) 142,926.9 (31.4) 93,093.5 (27.0) 60,261.9 (24.8) 76,677.7 (26.1)
74,305.1 (12.9) 88,285.7 (26.3) 81,295.4 (17.9) 58,856.0 (17.1) 50,099.1 (20.6) 54,477.6 (18.5)
31,049.2 (5.4) 9618.6 (2.9) 20,333.9 (4.5) 24,792.8 (7.2) 1047.1 (0.4) 12,920.0 (4.4)
574,020.9 (100)a 336,294.0 (100)ab 455,157.4 (100) 344,366.0 (100)ab 243,374.1 (100)ab 293,870.1 (100)

1 USD = 2350; n = 80, 20 households from each parish; all p-values were calculated using non parametric Mann–Whitney test or Kruskal–Wallis test on ranks. * is significant at 10%.
In parentheses are percentages.

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx 9

Table 7 household pursues, and the farm net income (Table 7). Households
Socio-economic factors influencing relative environmental income, MENP, Uganda, 2011. are less dependent on forest environmental income the higher their
Estimate Robust std error t value Pr(Nt) farm income and the more income activities they pursue, the further
Intercept⁎⁎⁎ −0.807292 0.21533 −3.75 0.000
away they reside from the park and the better educated its members
are. However, magnitudes of all but one coefficient — for number of
Human capital income activities, are rather small and may be difficult to interpret. Par-
Household size (AEU)⁎⁎ −0.063340 0.02826 −2.24 0.028
Education of HH⁎⁎ −0.040050 0.01628 −2.46 0.017
ticipation in CRMA and the number of total household members have a
positive influence on forest dependence.
Financial capital
Number of income activities⁎⁎ −0.172650 0.07370 −2.34 0.022
Farm net income⁎⁎ −1.21e−07 3.91e−08 −3.10 0.003
4.4. Role of environmental incomes and CRMA in poverty alleviation

Household external factors
Distance (m) −0.004960 0.00356 −1.39
Participation in CRMA 0.214804 0.12364 1.74
R-sq = 0.28; adjusted R = 0.21.
⁎⁎ Significant at 5%.
⁎⁎⁎ Significant at 1%.
thereby reducing dependence on environmental income (Vedeld et al.,
2007).
The average size of landholding had a significant negative effect on
absolute environmental income. Larger landholdings exercise a signifi-
cant negative impact on environmental income. Similarly, an increase
in total household income, number of income sources and farm income
had significant negative effects on environmental resources in absolute
terms (Table 6). Similar results have been reported from Ethiopia
(Mamo et al., 2007; Abebaw et al., 2012), India (Narain et al., 2008),
Malawi (Fisher, 2004) and Zimbabwe (Cavendish, 2000). The results
are also in line with the general observation that dependence on envi-
ronmental resources is dominated by poor households that lack alterna-
tive sources of livelihood (Vedeld et al., 2007).
As observed in other studies by Naughton-Treves et al. (2011), dis-
tance from the park boundary had no significant effect on total house-
hold income and only a slightly significant negative effect on absolute
environmental income (Table 6). This could be because it is easier and
cheaper for park adjacent communities to access and trade in park envi-
ronmental resources; those far away find it more economical to concen-
trate on farm, off-farm and non-farm activities. This in return balances
the economic benefits that each of these communities accrue from the
pursuit of its income portfolio.
Participation in CRMA had a significant positive effect on absolute en-
vironmental income (Table 6), but no significant effect on total household
income. Other studies revealed that most people do not have knowledge
about the existence of such programs, where to find them and how to
benefit from them (Mugagga, 2011; Katto, 2004). Those who expressed
knowledge about them complained that the people concerned were inac-
cessible. Dependence on park resources could be improved if access to
such programs and extension services could be enhanced and farm
households are mobilized to form needs driven cooperative groups.
Relative forest environmental income is negatively correlated with
park distance, years of education, number of income activities that a
4.4.1. Effect of PEI and CRMA on poverty incidence, depth and severity
0.169
Poverty incidence (head count poverty), depth and severity values
0.087
are higher for locations without CRMA than those for locations with
access to CRMA. Access to PEI reduces poverty incidence, depth and se-
verity by 16.3%, 11.6% and 11.4% respectively.
The decline in poverty incidence is higher in locations without CRMA
(20%) compared to locations with CRMA (12.5%) while the decline in
poverty depth and severity is higher in locations with CRMA (15.4%
and 12.3%) than those without CRMA (7.8% and 10.2%) (Table 8).
4.4.2. Effect of PEI and CRMA on income inequality
PEI reduces income inequality by 13%. The overall Gini coefficient is
0.52, and rises up to 0.59 when park income is excluded. Similar results
have been reported in other studies supporting the income equalizing
effect of environmental income especially among poor households
which have limited alternative income sources (Fonta and Ayuk,
2013; Kamanga et al., 2009). By location, we observe a 10%, 8%, 4.4%
and 4% in Buginyanya, Bubyangu, Bugitima and Tsekululu respec-
tively. Similarly, the Gini coefficients for households with CRMA (from
0.48 to 0.55) are higher than those for households without CRMA
(0.51 to 0.59).
As shown in Table 4 most (42%) of the households without CRMA are
poor. CRMA may increase existing inequalities as returns from Taungya
farming (USD121.68) and bee keeping (USD133) constitute additional
sources of annual household income for households with access to
CRMA (Cavanagh, 2009). At the same time, Taungya farming and bee
keeping are potential pathways out of poverty for those households as
they increase annual household incomes by 26.5% and 28% respectively.
4.4.3. MENP as a safety net
Although this study did not quantitatively explore ways in which
households rely upon MENP as a “natural insurance” or as a safety
net/coping strategy in the wake of natural disasters and shocks, local
narratives revealed that through illegal or clandestine access, MENP
provided a safety net response strategy employed by asset strapped
households affected by flooding and landslides in Bulambuli and
Bududa respectively (Mugagga, 2011). Such claims are further sup-
ported by both qualitative and quantitative studies conducted around
Mount Elgon (Katto, 2004; Norgrove and Hulme, 2006). Similarly, in
communities around the Budongo forest reserve, Debela et al. (2012)

Table 8
Poverty measures by location and participation in CRMA around Mount Elgon National Park Uganda, 2011.

Poverty measures All Location CRMA

District Mbale Sironko Bulambuli Manafwa Yes No

Units Sub-county Bubyangu Bugitima Buginyanya Tsekululu

With forest environmental income

Prevalence Head count 0.275 0.300 0.150 0.300 0.350 0.225 0.325
Depth Poverty gap 0.157 0.149 0.065 0.212 0.201 0.138 0.175
Severity Poverty gap2 0.116 0.101 0.054 0.150 0.190 0.107 0.125

Without forest environmental income

Prevalence Head count 0.438 0.500 0.250 0.450 0.550 0.350 0.525
Depth Poverty gap 0.273 0.292 0.140 0.365 0.293 0.292 0.253
Severity Poverty gap2 0.230 0.245 0.114 0.341 0.215 0.230 0.227

Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
10 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx
found that flooding, in particular, resulted in coping strategies among
households that involved intensified use of forest resources (more so
than shocks related to human health issues), and also found that house-
holds that were asset poor were more likely to turn to intensified forest
use. Restricting access to park resources in periods of natural disaster is
likely to push such households deeper into poverty (Naughton-Treves
et al., 2011; Tumusiime et al., 2011). Establishing CRMAs in such
locations may cushion affected households from the effects of such
shocks.
5. Conclusions and policy recommendations
This study contributes to the broader literature on the role of envi-
ronmental incomes in rural livelihoods and more specifically the debate
concerning the effects of participatory forest governance on forest users'
livelihoods and poverty alleviation. Our aim was to investigate differ-
ences in dependence on park environmental income (PEI), the role
of PEI in poverty alleviation and the factors influencing dependence
on PEI in communities with and without CRMA. Our results revealed
no significant differences in park dependence. PEI contributes 18–26%
of the total annual household income. Previous studies have highlighted
the important role of collaborative resource governance in enabling or
constraining rural livelihoods (Jumbe and Angelsen, 2006; Tsefaye
et al. 2010). Overall, PEI reduces headcount poverty by 17%; and,
although surprisingly more (20%) for households without CRMA
compared to 12.5% for households with CRMA. These results point
to the need for more detailed studies on the dynamics involved in
implementing CRMA. Ideally the CRMA are intended to offer access
and reduce potential costs and conflicts in park adjacent communities.
So far, conflict ridden communities have continuously challenged the
legitimacy of the existing agreements and have in some cases opted
not to negotiate any agreements, resorting to clandestine use of park re-
sources (Cavanagh, 2012) and contributing to increasing deforestation
and forest degradation in areas without CRMA (Sassen et al., 2013). In
order to achieve both environmental and livelihood objectives, any ef-
forts to conserve the park should ensure that these communities are
not continuously marginalized since drawing on past experience; they
will continuously resist such marginalization and will jeopardize any ef-
forts to sustainably manage MENP (Norgrove and Hulme, 2006). Empir-
ical findings from Tanzania show that governments may try to alienate
resource users on the grounds that their livelihood strategies are re-
sponsible for the current forest carbon loss (Beymer-Farris and Bassett,
2012). Thus, there is a need to target communities whose livelihoods
are intricately dependent on the park.
The major factors influencing dependence on PEI were household
size, the education level of the household head, ability of the household
to diversify its income sources, distance to the park and participation in
collaborative resource access agreements. Our findings also revealed
that PEI is more important to the poor, contributing over 80% of their an-
nual household. However, the criteria and mechanism for establishing
CRMA is not clear and the current ad hoc approach intended to reduce
transaction costs systematically favors those communities that have ac-
cess to other resources and are able to comply with UWA regulations
and keep away from the park. This will in the long run marginalize
the intended beneficiary communities which have in return resorted
to illicit harvesting of park resources and have no incentive to protect
the park. Formal access to park resources through upscaling CRMA
coverage would improve their livelihoods and probably contribute to
reducing income poverty in the region.
Acknowledgments
The field studies were supported by the Norwegian Research Council
under the Protected Areas and Poverty in Africa (PAPIA) project -
NORGLOBAL 178645 and a Post Doctoral Fellowship to the first author
from the Norwegian University of Life Sciences and the Department
Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
of International Environment and Development Studies. We thank
George Matanda for the help in all stages of fieldwork. We are especially
grateful to the communities that participated in the study and the anon-
ymous reviewers' comments which improved the manuscript.
References
Abebaw, D., Kassa, H., Kassie, G.T., Lemenih, M., Campbell, B., Teka, W., 2012. Dry forest
based livelihoods in resettlement areas of Northwestern Ethiopia. Forest Policy
Econ. 20, 72–77.
Angelsen, A., Wunder, S., 2003. Exploring the Forest–Poverty Link: Key Concepts. Paper,
Issues and Research Implications CIFOR Occasional.
Angelsen, A., Larsen, H.O., Lund, J.F., Smith-Hall, C., Wunder, S. (Eds.), 2011. Measuring
Livelihoods and Environmental Dependence — Methods for Research and Fieldwork.
Earthscan, Edinburgh.
Appleton, S., 2001. Changes in poverty and inequality. In: Collier, P., Reinnikka, R. (Eds.),
Uganda's Recovery: The Role of Farms, Firms and Government. World Bank,
Washington DC, pp. 83–121.
Baatvik, S.T., Kaboggoza, J.R.S., Kabutha, C., Vedeld, P., 2002. Mt. Elgon Regional Ecosystem
Conservation Programme (MERECP) Appraisal Report. Noragric Report, NLH, As,
Norway.
Babulo, B., Muys, B., Nega, F., Tollens, E., Nyssen, J., Deckers, J., Mathijs, E., 2008. Household
livelihood strategies and forest dependence in the highlands of Tigray, Northern
Ethiopia. Agric. Syst. 98, 147–155.
Baker, J., Milner-Gulland, E., Leader-Williams, N., 2011. Park gazettement and integrated
conservation and development as factors in community conflict at Bwindi Impene-
trable Forest, Uganda. Conserv. Biol. 26 (1), 160–170.
Beck, P.A., 2000. Conservation, development and collaboration: analyzing institutional
incentives for participatory conservation in Uganda. PhD. Dissertation. Indiana
University, Indiana.
Beymer-Farris, B., Bassett, T., 2012. The REDD menace: resurgent protectionism in
Tanzania's mangrove forests. Glob. Environ. Chang. 22, 331–341.
Blomley, T., Ramadhani, H., Mkwizu, Y., 2010. Bohringer. A. Hidden harvests: unlocking
the economic potential of community-based forest management in Tanzania. In:
German, L., Karsenty, A., Tiani, A. (Eds.), Governing Africa's Forests in a Globalized
World. Earthscan Press, UK, pp. 136–140.
Brockington, D., Igoe, J., 2006. Eviction for conservation: a global overview. Conserv. Soc.
4 (3), 424–470.
Brooks, J., Waylen, K.A., Mulder, M.B., 2013. Assessing community-based conservation
projects: a systematic review and multilevel analysis of attitudinal, behavioral,
ecological, and economic outcomes. Environ. Evidence 2 (1), 2.
Cavanagh, C., 2009. Jobs and power: the role of employment multipliers and leakages in
the political economy of conservation at Mount Elgon National Park, Uganda. PAPIA
Project Field Report. Norwegian University of Life Sciences, Ås.
Cavanagh, C. 2011. Impacts on poverty and poverty reduction from protected areas
in Africa: four cases”. Prepared for the Norwegian Institute for Nature Research
(NINA), unpublished.
Cavanagh, C., 2012. Anthropocene Conservation: Governing Environmental Change, Bio-
diversity, and Local Resistance at Mount Elgon, Uganda. (M.Sc. Thesis). Norwegian
University of Life Sciences, Ås.
Cavendish, W., 2000. Empirical regularities in the poverty–environment relationship of
rural households: evidence from Zimbabwe. World Dev. 28 (11), 1979–2003.
Chhetri, P., Mugisha, A., White, S., 2003. Community Resources Use in Kibale and
Mt. Elgon National Park, Uganda. Parks, pp. 28–49 13, pp.
Das, N., 2010. Incidence of forest income on reduction of inequality: evidence from forest
dependent households in milieu of joint forest management. Ecol. Econ. 69 (8),
1617–1625.
Davenport, T., Howard, P., Dickinson, C., 1996. Mount Elgon National Park Biodiversity
Report. Forest Department, Kampala, Uganda.
Debela, B.L., Shively, G., Angelsen, A., Wik, M., 2012. Economic shocks, diversification and
forest use in Uganda. Land Econ. 88 (1), 139–154.
Ellis, E.A., Porter-Bolland, L., 2008. Is community-based forest management more effective
than protected areas? A comparison of land use/land cover change in two neighboring
study areas of the Central Yucatan Peninsula, Mexico. For. Ecol. Manag. 256, 1971–1983.
Fisher, M., 2004. Household welfare and forest dependence in southern Malawi. Environ.
Dev. Econ. 9 (2), 135–154.
Fonta, W.M., Ayuk, E.T., 2013. Measuring the role of forest income in mitigating poverty
and inequality: evidence from south-eastern Nigeria. Forests, Trees Livelihoods 22
(2), 86–105.
Gombya-Ssembajwe, W., Buyinza, M., Kakuru, O.V., 2007. Quantification of monetary
losses due to illegal pitsawing in Budongo Forest, Uganda. Int. For. Rev. 9, 842–849.
Gosalamang, D., Vedeld, P., Gombya-Ssembajjwe, W., 2008. From forest reserve to national
park. Change in Legal Status and Impacts on Livelihoods and Biodiversity Resources,
Mt. Elgon, Uganda. Working Paper No. 44. Noragric/UMB, Ås.
Government of Uganda (GoU), 1997. The Local Government Act, 1997. Ministry of Local
Government. Government Printer, Entebbe.
Himmelfarb, D., 2005. Moving people, moving boundaries: the socio-economic effects of
protectionist conservation, involuntary resettlement and tenure insecurity on the
edge of Mt. Elgon National Park, Uganda. Agroforestry in Landscape Mosaics Working
Paper Series: World Agroforestry Centre. Tropical Resources Institute of Yale Univer-
sity and the University of Georgia.
Himmelfarb, D., 2012. In the Aftermath of Displacement: A Political Ecology of Disposses-
sion, Transformation, and Conflict on Mt. Elgon, Uganda. (PhD Dissertation). Univer-
sity of Georgia, Athens.
 C. Nakakaawa et al. / Forest Policy and Economics xxx (2015) xxx–xxx
Hoefsloot, H., Kahata, J., Nsita, S.A., 2011. Mount Elgon Regional Ecosystem Conservation
Programme: End Review Report. LTS International Ltd., Penicuik, UK.
Howard, P.C., 1991. Nature Conservation in Uganda's Tropical Reserves Forest Depart-
ment. Ministry of Environmental Protection Uganda.
IUCN, 2005. Mount Elgon Regional Ecosystem Conservation Programme (MERECP) —
Programme Document (Version May 2005) IUCN. Nairobi, Kenya.
Jagger, P., 2012. Environmental income, rural livelihoods, and income inequality in
western Uganda. Forests Trees and Livelihoods 21 (2), 70–84.
Jankulovska, A., Vedeld, P., Kaboggoza, J.R.S., 2003. Coaching — poaching? Governance,
local people and wildlife around Mount Elgon National Park, Uganda. Noragric
Working Paper 31. Norwegian Agricultural University, Ås
Jumbe, C.B.L., Angelsen, A., 2004. Forest Dependence and Participation in CPR Manage-
ment. Paper Presented at the European Association for Ecological Economics.
Lisbon, Portugal.
Kamanga, P., Vedeld, P., Sjaastad, E., 2009. Forest incomes and rural livelihoods in
Chiradzulu District, Malawi. Ecol. Econ. 68 (3), 613–624.
Kar, S.P., Jacobson, M.G., 2012. NTFP income contribution to household economy and
related socio-economic factors: lessons from Bangladesh. For. Policy Econ. 14 (1),
136–142.
Katto, F., 2004. Sustainable Livelihoods and Environmental Income Dependence around
Mt. Elgon. MSc Thesis. Noragric, Norwegian University of Life Science, Ås.
Kiggundu, K.J., 2007. Institutional Sustainability of Collaborative Resource Use Agree-
ments in Mount Elgon, Uganda. (MSc Thesis). Noragric, Norwegian University of
Life Sciences, Ås.
Larsen, T.S., Kamugasha, B.N., Karani, I., 2008. Mid-term Review of Mount Elgon Regional
Ecosystem Conservation Programme (MERECP). Noragric Report No. 44. Norwegian
University of Life Sciences, Ås.
Luzinda, H., 2008. Mobile boundary and mobile people: involuntary resettlement of the
Benet people in Mt. Elgon National Park, Uganda. M.Sc. Thesis. Noragric, Norwegian
University of Life Sciences, Ås.
Mamo, G., Sjaastad, E., Vedeld, P., 2007. Economic dependence on forest resources: a case
from Dendi District, Ethiopia. For. Policy Econ. 9 (8), 916–927.
Moll, R., 2011. The privilege to cooperate: Assessing the impact of collaborative resource
management agreements on local communities' livelihoods around Mount Elgon
National Park, Uganda. Master Thesis. Department of Economics, University of Oslo,
Oslo, Norway.
Mugagga, F. 2011. Land use change, landslide occurrence and livelihood strategies on
Mount Elgon slopes, eastern Uganda. Unpublished PhD Dissertation, Nelson Mandela
Metropolitan University, Port Elizabeth.
Mugagga, F., Kakembo, V., Buyinza, M., 2012. Land use changes on the slopes of Mount
Elgon and the implications for the occurrence of landslides. Catena 90, 39–46.
Musali, Paul, 1998. Issues, challenges and prospects of collaborative management of
protected areas: a case of introducing people's participation in the management of
Mt. Elgon National Park. East Afr. Geogr. Rev. 20 (2).
Namugwanya, M., 2004. People's dependence on environmental income for survival and
livelihoods: a case study of Mt. Elgon National Park Uganda. (M.Sc. Thesis). Noragric,
Norwegian University of Life sciences, As.
Narain, U., Gupta, S., van't Veld, K., 2008. Poverty and resource dependence in rural India.
Ecol. Econ. 66, 161–176.
Naughton-Treves, Lisa, Alix-Garcia, Jennifer, Chapman, Colin A., 2011. Lessons about parks
and poverty from a decade of forest loss and economic growth around Kibale National
Park, Uganda. PNAS 108 (34), 13919–13924.
Nayenga, R.N., 2003. Landlessness Amidst Abundance. Poverty Monitoring and Analysis
Unit, Ministry of Finance, Planning and Economic development, Uganda.
Nielsen, M.R., Pouliot, M., Riyong, K. Bakkegaard, 2012. Combining income and assets
measures to include the transitory nature of poverty in assessments of forest
dependence: evidence from the Democratic Republic of Congo. Ecol. Econ. 78,
37–46.
NFA, 2009. National Biomass Survey 2005. National Forestry Authority, Springs Road,
Kampala.
NFA, 2010. National Forestry Authority (NFA) Land Use Land Cover database. Accessed
from the National Forestry Authority. Springs Road, Kampala, Uganda.
Norgrove, L., 2002. Parking resistance and resisting the park. PhD Dissertation. The Theory
and Practice of National Park Management, A Case Study of Mount Elgon, Uganda.
University of Manchester, Manchester, UK.
Norgrove, L., Hulme, D., 2006. Confronting Conservation at Mount Elgon, Uganda. Dev.
Chang. 37 (5), 1093–1116.
Nsita, S., 2010. In search of forest governance reform in Uganda. Background Paper for the
Workshop On Forest Governance In Uganda, 15–16 June.
Oonyu, J., 2000. Impact of Non-formal Environmental Education Programmes on
Environmental Awareness, Attitudes and Conservation Practices of Local Com-
munities Living Around Mt. Elgon National Park. PhD Dissertation. Makerere
University, Kampala.
Please cite this article as: Nakakaawa, C., et al., Collaborative resource management and rural livelihoods around protected areas: A case study of
Mount Elgon National Park, Uganda, Forest Policy and Economics (2015), http://dx.doi.org/10.1016/j.forpol.2015.04.002
11
Peturrson, J.G., Vedeld, P., Sassen, M., 2013. An institutional analysis of deforestation pro-
cesses in protected areas: the case of transboundary Mt. Elgon, Uganda and Kenya.
For. Policy Econ. 26, 22–33.
Petursson, J.G., 2011. Institutions and Transboundary Protected Area Management: The
Case of Mount Elgon, Uganda and Kenya. PhD Dissertation. Norwegian University of
Life Sciences, Ås.
Phelps, J., Webb, E.L., Agrawal, A., 2010. Does REDD + threaten to recentralise forest
governance? Science 328, 312–313.
Pouliot, M., Treue, T., 2013. Rural people's reliance on forests and the non-forest environ-
ment in West Africa: evidence from Ghana and Burkina Faso. World Dev. 43, 180–193.
Ravallion, M., Chen, S., Sangraula, P., 2009. Dollar a day revisited. World Bank Econ. Rev.
23 (2), 163–184.
Ray, D., 1998. Development Economics. Princeton University Press, Princeton, NJ.
Ribot, J., Agrawal, A., Larson, A., 2006. Recentralizing while decentralizing: how national
governments reappropriate forest resources. World Dev. 34, 1864–1886.
Ribot, J.C., Lund, J.F., Treue, T., 2010. Democratic Decentralization in Sub-Saharan Africa:
Its contribution to forest management, livelihoods, and enfranchisement. Environ.
Cons. 37, 35–44.
Sandbrook, C., Nelson, F., Adams, W.M., Agrawal, A., 2010. Carbon, forests and the REDD
paradox. Oryx 44 (3), 330–334.
Sassen, M.D., Sheil, K.E., Giller, C.J., Ter Braak, F., 2013. Complex contexts and dynamic
drivers: understanding four decades of forest loss and recovery in an East African
protected area. Biol. Conserv. 159, 257–268.
Scoones, I., 1998. Sustainable Rural Livelihoods. A Framework for Analysis. IDS Working
Paper 72.
Scott, P., 1998. From Conflict to Collaboration: People and Forest at Mount Elgon, Uganda.
IUCN East Africa Regional Office, Nairobi, Kenya.
Sjaastad, E.P., Vedeld, P., Angelsen, A., Bojø, J., 2005. What is environmental income? Ecol.
Econ. 55 (1), 37–46.
Sletten, M., 2004. Collaborative Management in Mount Elgon National Park, Uganda.
(MSc Thesis). Norwegian University of Life Sciences, Ås.
Sletten, M., Vedeld, P., Kabbogoza, J., 2008. To co-operate or not to cooperate: a study of
collaborative management planning in Mount Elgon National Park, Uganda. Noragric
Working Paper No. 46. Norwegian University of Life Sciences, As.
Tacconi, T., 2007. Decentralization, forests and livelihoods: theory and narrative. Glob.
Environ. Chang. 17, 338–348.
Tesfaye, Y., Roos, A., Campbell, B., Bohlin, F., 2010. Forest incomes and poverty alleviation
under participatory forest management in the Bale Highlands, Southern Ethiopia. Int.
For. Rev. 12, 66–77.
Tumusiime, D.M., Vedeld, P., Gombya-Ssembajjwe, W., 2011. Breaking the law? Illegal
livelihoods from a protected area in Uganda. For. Policy Econ. 13, 273–283.
Turyahabwe, N., Geldenhuys, C.J., Watts, S., Obua, J., 2007. Local organizations and
decentralized forest management in Uganda: roles, challenges and policy implica-
tions. Int. For. Rev. 9 (2), 581–596.
UBOS, 2011. Statistical Abstract. Uganda Bureau of Statistics (UBOS), Kampala, Uganda.
Uganda Communications Commission (UCC), 2010. Rural Communications Development
Fund (RCDF): UCC Support Through the RCDF Programme. Uganda Communications
Commission, Kampala, Uganda.
Uganda Wildlife Authority (UWA), 2000. Mt. Elgon National Park General Management
Plan (MENP GMP), 2000–2010. Uganda Wildlife Authority, Kampala.
Uganda Wildlife Authority (UWA), 2009. Mount Elgon National Park General Manage-
ment Plan, 2009–2019. Uganda Wildlife Authority, Kampala.
van Heist, M., 1994. Accompanying Report With the Land Unit Map of Mount Elgon
National Park. Mount Elgon Environmental Conservation Project, Mbale.
Vangen, C., 2009. Evicted in the Name of Nature: The Process of Eviction and its Impact on
Local Rural Livelihoods in Mount Elgon, Uganda. (MSc Thesis). Norwegian University
of Life Sciences, Ås.
Vedeld, P., Rooij, A.V., Sundnes, F., Jørgensen, I.T., 2005. Final Appraisal of the Mt. Elgon
Regional Ecosystem Conservation Programme (MERECP). Noragric Report No. 25.
Norwegian University of Life Sciences, As.
Vedeld, P., Angelsen, A., Bojo, J., Sjaastad, E., Berg, G.K., 2007. Forest environmental
incomes and the rural poor. For. Policy Econ. 9, 869–879.
Vedeld, P., Jumane, A., Wapalila, G., Songorwa, A., 2012. Protected areas, poverty and
conflicts: a livelihood case study of Mikumi National Park, Tanzania. For. Policy
Econ. 21, 20–31.
Webster, G., Osmaston, H.A., 2003. A History of the Uganda Forest Department 1951–
1965. Commonwealth Secretariat, London.
White, S.P., 2002. People–park conflicts in Mt. Elgon. The role of collaborative management
in conflict resolution. Paper presented on the National Conference on Mountains and
Highlands in Uganda 3rd–4th October 2002. Makerere University, Kampala, Uganda.
Yamano, T., Sserunkuuma, D., Otsuka, K., et al., 2004. The 2003 REPEAT Survey in Uganda: Re-
sults. Foundation for Advanced Studies on International Development Tokyo, Japan
(downloadable from http://www3.grips.ac.jp/~yamanota/REPEATinUgandaReport.pdf).