Harnessing markets for improved nutrition: A Case Study of Zomba
Presentation PSRI
1. Co-existence & Livelihood
Opportunities of Pastoralists &
Migrant Agricultural Communities
in Kajiado County, Kenya
Omondi Ea,
Prof. Oucho Jb,
Dr. Odipo Gb
a Student PSRI – UoN, Kenya
b My Supervisors PSRI – UoN, Kenya
2. Problem Statement
• The coexistence of the two major livelihood systems in
Kimana creates pressure to the rangeland ecology
• causing vulnerability and reduced productivity
• This affects households differently depending on the
level of resilience or vulnerability
• Resilience is the function of livelihood capital asset
endowment and the position of the household in the socio-
political fabrics of a society
• There is potential threat to survival in terms of food
insecurity
• Therefore, one livelihood group is potentially dependent on
the other with endowment of each livelihood group
complementing each other
3. Problem Statement
• Such elements of complementarities between the two
livelihood systems have not been exploited
• This study sought to undertake livelihood analyses of
migrant and non-migrant communities in Kimana area
of Kajiado County,
• comparing livelihood strategies of each group and their
related food-security outcomes
• The analysis revealed how the livelihood of one group is
impacted upon by the presence of livelihood activities
of another group operating in the same livelihood
ecology
4. Research Objectives
Resilience to food insecurity x Livelihood Strategy Groups
Objectives:
a) theoretical: The classic and the neoliberal
- why resilience analysis?
- why livelihood analysis?
b) empirical: based on AWF Livelihoods Data 2015
- To identify livelihood strategies pursued by the agro-pastoralist
communities and indigenous peasants in Kimana
- To compare the households' asset endowment of the two
communities (agrarian and pastoralists), and the institutional context
surrounding their access and use
- To compare the extent of food availability and dietary diversity, among
households of the agrarian and pastoralist communities
Advantages in combining RA and LA
5. Justification
• Resource competing livelihood systems creates pressure
to Kimana’s ecology
• Resulting in depletion of its resource base and reduced
productivity
• The outcome of this situation is that the wellbeing of in-
migrants would be affected first before that of non-
migrants resulting into conflicts
• This study therefore would flag potentials for socio-
cultural and social-economic synergies between non-
migrants and the in-migrants
6. Data Set and Methodology
AWF Livelihoods data in Kilimanjaro landscape (2009-2015)
The survey instrument was organized in four questionnaires:
(1) a household questionnaire;
(2) a 14-day household expenditure diary to record consumption and
purchases;
(3) a market price questionnaire; and,
(4) a community questionnaire.
Sample Size: Population of 1,215 HHs, Sampled 104 HH members
Share of male population was 49.3%.
Average age: 22.17 years (males: 21.9 years. females: 22.3 years).
Rural households: 73.9% of the sample.
Religion: Protestants (45.64%), Traditionalists (14.8%), Catholic
(11.5%) and other Christians (10.0%).
8. Results Interpretation/Discussion
Table 5: Demographic and other characteristics of the sampled households
Pastoralists;
N=31
Farmers;
N=55
Other;
N=18 Total
Mean HH size 8.4 5.16 5.33 6.29
No Education (%) 14 (45.16) 16 (29) 7 (38.9) 37 (37.57)
Primary Education (%) 17 (54.83) 37 (67.27) 8 (44.4) 62 (59.61)
Secondary Education (%) 0 (0.0) 2 (3.6) 3 (16.7) 5 (4.8)
Workforce 2.42 1.96 1.89 2.09
Dependency Ratio (%) 67.7 58.6 59.6 61.9
Mean Area Cultivated 13 4.58 5.94 7.84
Female Headed Households (%) 0 20 16.67 13.46
Male Headed Households (%) 100 80 83.33 86.53
Youth headed HH (<40 yrs old) (%) 45.16 30.9 44.44 37.5
HH heads aged 41-65 yrs old (%) 51.6 52.7 50 51.92
HH heads aged >65 yrs old (%) 3.2 16.36 5.56 10.57
HH average number of children 6 3.25 3.33 4.19
9. Major Source of Income
Livelihood group
Sell Food
Crops n
(%)
Sell
Livestock
n (%)
Sell Agric.
labour n
(%)
Do petty
jobs n (%)
Do Pond
Fishing n
(%)
Other
activities
n (%)
In-migrants (n=55) 33 (60.0) 0 (0.0) 4 (7.2) 5 (9.09) 4 (7.2) 9 (16.36)
Non-Migrants (n=31) 27 (81.7) 3 (9.67) 1 (3.2) 0 (0.0) 0 (0.0) 0 (0.0)
Other (n=18) 5 (27.78) 1 (5.56) 1 (5.56) 8 (44.44) 0 (0.0) 3 (16.67)
Total (n=104) 65 (62.5) 4 (3.8) 6 (5.77) 13 (12.5) 4 (3.85) 12 (11.54)
10. Food availability and HDDS means
comparison, independent t- test
Number of months of
food availability
(MONTFAVL)
Household dietary diversity
score (HDDS)
n Mean n Mean
In-Migrants
(Agro-
Pastoralists)
55 8.209 55 5.2
Non-Migrants
(Pastoralists)
31 9.919 31 6.839
p-Value 0.01 0.001
11. Children's nutritional status in Kimana and
those at County and National levels
Study Results
Kajiado
County
(%)*
National (%)*
In-Migrants
(Agro-
Pastoralists)
Non-
Migrants
(Pastoralists)
Overall Overall Rural Areas
Stunted (HAZ
≤ -2SD)
6.06 5.12 5.55 18.2 26.0 5
Wasted
(WHZ≤
-2SD)
12.12 17.95 15.28 3.0 4.0 17
Under-weight
(WAZ ≤ -2SD)
8.1 11.0
MUAC
GAM
(MUAC<12.5)
3.03 7.69 5.55 - - -
SAM
(MUAC<11.5)
0 2.56 1.38 - - -
*Source: KDHS (2014)
12. Anthropometric measurements (independent
t-test comparison) results
N=72
MEAN
HAZ
MEAN
WAZ
MEAN
WHZ
MEAN
MUAC
MEAN NO. OF
MEALS
In-Migrants
(Agro-
Pastoralists)
33 -2.4609 -1.3164 0.3133 13.721 2.52
Non-
Migrants
(Pastoralists)
39 -1.3441 -1.0251 -0.2903 14.359 2.64
p-value 0 0.226 0.079 0.082 0.32
13. Comparison of livelihood asset endowment
(independent t-test)
N
Mean
HH Size
Mean year
of
schooling
Mean
workforce
Mean
dependency
Ratio (%)
Mean
cultivated
Area (Acre)
In-Migrants
(Agro-
Pastoralists)
31 8.42 3.55 2.42 69.93 12.9
Non-
Migrants
(Pastoralists)
55 5.16 4.95 1.96 58.57 4.22
p- Value 0.038 0.001 0.003 0.00
14. Household Dietary Diversity Score (HDDS)
HDD = 6.77 + 1.23 (AGRPST) + 0.18 (EDUCSPOU) + 0.07
(AREACULT) – 0.03 (DEPRATE) – 1.75 (FISHING)
A significant empirical Ordinary Least Square Regression Model
(OLSRM) emerged at 5% (F7,89=10.3, p<0.05, R=63.8% and
R2=40.8%)
15. Estimated coeff. of factors affecting
HDD
Variable Co-efficient Std Error p-Value
Collinearity
Tolerance
Mean
CONSTANT 6.771 0.682 0.000
AGRPST 1.229 0.430 0.005 0.531 0.320
EDUCSPOU 0.182 0.050 0.000 0.669 3.290
AREACULT 0.070 0.035 0.047 0.544 7.691
DEPRATE -0.033 0.009 0.001 0.916 61.920
FISHING -1.754 0.763 0.024 0.929 N/A
EDUCHH -0.074 0.056 0.191 0.701 4.520
R 0.638
R2 0.408
ADJUSTED R2 0.368
F-VALUE
(F7,89) =
10.322
1.440 0.000
AGRPST Agro-pastoral livelihood as a dummy variable of livelihood ( coded: 1 = agro-
pastoral, 0 = not agro-pastoral)
EDUCSPOU Level of education of household spouse (quantified by number of years of
schooling)
AREACULT Household's total area cultivated in acres
DEPRATE Household dependence rate(percentage ratio of number of dependants to total
household members)
FISHING Fishing activity as a dummy variable of source of income (coded: 1 = fishing, 0
= not fishing)
EDUCHH Level of education of household head (quantified by number of years of
schooling)
16. Factors Influencing Household Food
Availability and Stability
MONTFAVL = 12.3 + 0.1(AREACULT) + 0.15(EDUCSPOU) –
0.04(AGEHH) – 1.36 (NONMIGRANT)
The OLSRM was far very significant at 5% (F7,89=7.08,
p=0.000, R=59.8% and R2=35.8%)
17. Variable Co-efficient Std Error p-Value
Collinearity
Tolerance
Mean
CONSTANT 12.306 1.301 0.000
EDUCSPOU 0.147 0.060 0.016 0.87 3.290
AREACULT 0.100 0.046 0.032 0.582 7.691
AGRLABOU -0.573 0.609 0.350 0.868 N/A
WFORCE -0.468 0.380 0.222 0.821 2.120
AGEHH -0.038 0.017 0.025 0.751 46.680
DEPRATE -0.014 0.014 0.316 0.78 61.920
PEASNT -1.359 0.512 0.009 0.608 N/A
R 0.598
R^2 0.358
ADJUSTED R^2 0.307
F-VALUE
(F7,89) =
7.075
1.965 0.000
AREACULT Household's total area cultivated in acres
EDUCSPOU Level of education of household spouse (quantified by number of
years of schooling)
AGRLABOU Agricultural labour as a dummy variable of source of income (coded:
1 = agricultural labour, 0 = not agricultural labour)
WFORCE Work force ( number of working persons in the household)
AGEHH Chronological age of household head in years
DEPRATE Household dependence rate (percentage ratio of number of
dependants to total household members)
PEASNT
Peasantry as a dummy variable of livelihood (coded: 1 = peasant, 2 =
not peasant)
Notes de l'éditeur
In-migrants occupy barren lands unsuitable for agriculture, On the other hand, indigenous people have the land, but lack producer goods, Draught-oxen, and other physical capitals
Consequently, there is an increase in the number of landless pastoralists who own some livestock but have no access to large tracts of grazing land. The classic and the neoliberal views are such that a large and increasing population that is predominantly agrarian and is sustained by a fragile ecosystem such as pertains in Kimana, easily over-consumes the limited resources available. They blame resource degradation on the management practices and communal tenure system prevalent in rural societies like Kimana
The broad objective of this study was to undertake livelihood analyses of migrant and non-migrant communities in Kimana, Kajiado County by comparing their livelihood strategies and the resulting food security outcomes.
Two regression analyses were done in which two dependent variables measuring food security, were each regressed with a number of independent variables.
Two regression analyses were done in which two dependent variables measuring food security, were each regressed with a number of independent variables.
On average the in-migrants’ livelihood provides 68.42% annual security to food, compared with non-migrant livelihood which provides only 82.5%.
on average about 15.3% of surveyed households (children) were underweight (low weight-for-age).
The severe acute malnutrition (SAM) (MUAC<11.5cm) is 1.38% whereas the global acute malnutrition (GAM) (i.e. MUAC<12.5cm) is 16.3%.
While non-migrants’ households have more number of months of food availability and higher dietary diversity score than non-migrants’ households, the nutritional status of their children is worse than that of the in-migrants’. This can be explained by the entitlement theory as documented by Sen (1981).
The t-test comparison of mean values of these nutritional outcome indices in the two livelihood groups show that the mean Weight-for-Age z-score, mean Weight-for-Height z-score and mean MUAC between the two groups are not statistically significant at p≥ 0.05 while the mean Height-for-Age z-score for in-migrants’ children are significantly lower than that of non-migrants’ children (p<0.05)
Non-migrants are relatively more educated with a mean year of schooling of 4.95 against 3.55 of in-migrants. The in-migrants also have a high dependency ratio of 69.93% against 58.57% of the non-migrants who also had a smaller household workforce of 1.96 persons per household against 2.42 of the in-migrants. The mean cultivated area is higher for the in-migrants, 12.9 acres against 4.22 acres of the non-migrants. All the variables are statistically different at 5% significance level (p<0.05).
Of the six variables deemed to have influence on the dependent variable, five were found to have significant influence. These are fishing as income generating activity, dependency ratio, total cultivated area of household in acres, education level of household spouse and agro-pastoral livelihood (p<0.05)
Reflecting from this table, the independent variables explain 40.8% of the variations in the dependent variable (R2=0.408). The F-value is highly significant (p=0.000) at 5% probability level, indicating that the regression model is statistically significant in explaining variation in the dependent variable.
Of the seven variables deemed to have influence on the dependent variable (food availability and stability), four were found to have significant influence. These are total cultivated area of the household in acres, education of household spouse, age of household head and peasant livelihood (p<0.05).
The included variables explain 35.8% of the variations in the dependent variable (R2=0.358). The F-value is highly significant at 5% probability level (p=0.000), indicating that the independent variables all together are statistically significant in explaining variation in the dependent variable.
Another interesting observation in this finding is that it is the education of spouse (women) and not household head (men) that has shown to positively influence the food availability and stability in the household.