Countries in sub-Saharan Africa are particularly vulnerable to climate change, given their limited capacity to adapt. As such, climate smart agriculture through adoption of resilient agricultural practices technologies has become a topical issue in the recent development policy agenda. This paper examines the influence of household dynamics on adoption of resilient agricultural technologies and subsequently the effect on household resilience using primary data collected from 300 households in Machakos and Makueni counties of semi-arid Kenya. The study uses the household type as the gender identifier, however, we recognize that female-headed households are not a homogenous group and hence we disaggregate the female-headed households into households managed by women when the husband is not physically present i.e. working away from home (de facto female-headed) and households managed by single, widowed, divorced, or separated women (de jure female-headed). Multivariate probit model and ordered logit models were used to model simultaneous adoption decisions by farm households facing multiple resilient technologies, which can be adopted singly or in different combinations. The analysis reveal that there exist a significant correlation between resilient agricultural technologies, implying that the adoptions of resilient technologies are interrelated. The results further establish that household type is a critical factor determining adoption behaviour and consequently the household level of resilience. Disaggregating the female-headed households, we find that de facto female-headed households are more likely to adopt multiple resilient technologies and become relatively resilient compared to other household types. From the findings, policies and interventions grounded on a sound understanding of the capacities and complexities of different household types, interrelationship between technologies and organising farmers into groups have the potential to influence adoption of resilient agricultural technologies and improve their resilience to climatic shocks.
Household Dynamics in Adoption of Climate Resilient Agricultural Technologies in Semi-arid Kenya
1. Household Dynamics in Adoption of Climate Resilient
Agricultural Technologies in Semi-arid Kenya
Daniel Kangogo and Pascal Sanginga
Presented by;
Daniel K. Kangogo
Our Common Future under Climate Change Conference, 7-10 July, 2015
Paris, France
3. Research Background
Climate Change is Real!
Sub-Saharan Africa is largely vulnerable to adverse impacts of climate change
given their inadequate capacity to adapt.
Hence, strengthening agricultural production systems is fundamental to
improving household resilience.
At national levels, this requires substantial investment in drought and heat
tolerant seed varieties among other interventions.
Most of the interventions have focused on adoption of single technologies to
improve productivity, yet to build resilience, diversified adoption of resilient
technologies is critical.
Emerging concerns are moving from increasing yields to building resilience
4. Research Background
Resilience, ability to anticipate, adapt to, and recover from the effects of shocks in
a manner that protects livelihoods, accelerates and sustains recovery
There is an urgent need to understand the critical resilience dimensions in
the face of changing climate
This study was carried out within the Canadian International Food Security
Research Fund (CIFSRF) project of IDRC
Objective- to enhance food security in developing countries by funding applied
agricultural research.
How- through a participatory approach to evaluate agricultural practices in semi-
arid Kenya
Objective- to catalyze adoption of appropriate agricultural practices
5. Project Intervention
Some of the practices implemented through the project
Indigenous chicken
Improved maize varieties Improved green gram varieties
Improved pigeon pea varieties
Water management
Improved sorghum varieties
6. Research Problem
Over time researchers have focused on the adoption of single technologies
with the aim of increasing productivity
No study on adoption of multiple technologies to improve household
resilience to climate change.
In this study we;
o Analyse simultaneous adoption of a portfolio of climate resilient
technologies to demonstrate how adoption decisions can be used to
explain household resilience.
(Households that have diversified livelihood options are relatively resilient)
Past studies have compared Male Headed Households (MHHs) Vs. Female
Headed Households (FHHs)
o This way of analysis masks the dynamics that come along with different household
structures
7. Household Dynamics
We distinguish 3 different types of households
o MHHs – households where male and female are present
o de facto FHHs – female headed households with absentee
husband
o de jure FHHs – female headed households with no male
(widowed, divorced, separated or never married)
This allows for the analysis of different household structures.
8. Research questions
Do the different household structures exhibit different adoption behaviours?
How does household structure influence the adoption of climate resilient
farming technologies?
Do household structure affect household resilience?
9. Methodology
Using multistage sampling procedure, 300 households were surveyed: 240
project participants and 60 non-project members from Machakos and
Makueni Counties, Kenya
To analyse adoption decisions and household resilience, three technologies
were considered;
o Maize, Green grams and Indigenous chicken
and their combinations (level of diversification) in the form of:
o Maize + green grams
o Maize + IC
o Green grams + IC
o Maize + green grams + IC
Crop ent.
Crop-poultry ent.
11. Econometric models
Multivariate Probit (MVP) model – since farmers adopt
technologies as compliments or substitutes
MVP takes into account the potential correlation
between adoption decisions (-/+)
However, the MVP model does not draw distinction
between households that adopted one technology and
those that adopted multiple technologies
Ordered logit model to determine the influence
household structures on the resulting household
resilience category
12. The ordered logit model allows for the analysis of the
factors that influence the adoption of single technologies
and the various combination.
An ordinal dependent outcome is generated from the
nature of household adoption behaviour
Ordinal Outcome Score
if a household adopted any single technology
Maize/Green grams/IC
0
if adopted Maize + green grams 1
if adopted Maize + IC or Green grams + IC 2
if adopted maize + green grams + IC 3
Econometric models
13. Descriptive results
Q1. Do the different household types exhibit different adoption behaviours?
technologies?
Table 1. Descriptive statistics comparing different household types
Variables MHHs De facto FHHs De jure FHHs
Proportions Proportions Proportions
Number of observations 195 49 56
Indigenous chicken (IC) 0.93 0.92 0.91
Improved maize varieties 0.74 0.84 0.57***
Improved green gram varieties 0.76 0.93 0.79*
Technology combination
Improved maize varieties and IC 0.69 0.78** 0.52*
Improved green gram varieties and IC 0.53 0.51 0.54
Improved maize and green gram varieties 0.45 0.53 0.41
Improved maize, green gram and IC 0.42 0.49 0.39
*-MHHs vs. de facto *-De facto vs. de jure
14. Q2. How does household structures influence the adoption of climate resilient technologies?
Regression results
Table 2. Multivariate Probit model results
Improved maize
technology
Improved green
gram technology
Indigenous chicken
Explanatory variables Coeff. Std. Err. Coeff. Std. Err. Coeff. Std. Err.
Male-headed household -0.888** 0.445 -0.857** 0.410 0.275 0.437
De jure FHHs -0.980** 0.502 -0.714 0.481 0.897*** 0.574
Belong to mkt group 0.047 0.238 0.538** 0.236 -0.040 0.363
Ln(Off-farm income) -0.071* 0.043 0.038 0.034 0.074 0.048
Project member 0.761** 0.302 -0.323 0.298 0.007** 0.396
Note: De facto female-headed household is the reference category where other household types are compared.
15. Yi= 0 (collapse)
Yi= 1 (recover, but
worse than before)
Yi= 2 (bounce back)
Yi= 3 (bounce back
better)
Explanatory variables ME SE ME SE ME SE ME SE
MHHs 0.040 0.042 0.013 0.014 0.022** 0.027 -0.075 0.082
De jure FHHs 0.050** 0.066 0.014 0.017 0.017 0.015 -0.081** 0.096
Belong to mkt group -0.077** 0.034 -0.024** 0.011 -0.041* 0.022 0.141** 0.063
Ln(Distance to mkt) 0.033** 0.017 0.010* 0.006 0.016* 0.009 -0.060** 0.030
Ln(Farm income) -0.021** 0.009 -0.007** 0.003 -0.010** 0.005 0.038** 0.016
Project member -0.077 0.051 -0.021 0.013 -0.021** 0.011 0.119* 0.068
County -0.084** 0.034 -0.025** 0.011 -0.040** 0.019 0.150*** 0.058
Regression results
Table 2. Multivariate Probit model results
Q3. How does household structures influence the adoption of climate resilient technologies?
Note: De facto female-headed household is the reference category where other household types are compared.
16. Conclusions
If we consider households structures to consist of only MHHs and FHHs we
miss important development outcomes.