Eleni Yitbarek_2023 AGRODEP Annual Conference

Eleni Yitbarek #2023 AGRODEP Conference March 22, 2023 1/24

Crop diversity and welfare dynamics: Empirical
Evidence from Nigeria
Eleni Yitbarek, Hiywot Girma, Wondimagegn Tesfaye
2023 AGRODEP Annual Conference, Kigali, Rwanda
21-23 March 2023

Outline
1 Motivation
2 Objective
3 Data
4 Estimation strategy
5 Results
6 Conclusion

Motivation
After a hiatus, promoting sustainable agricultural growth is back on
the development agenda.
Climate change is emerging as a major threats to the development of
the sector and might worsen food insecurity
Climate change might aect smallholder farmers disproportionately: a
moderate increase in temperatures will have a negative impact on the
production of stable crops
There is an public policy demand for identifying sustainable
agricultural practices that can improve welfare and help farm
households withstand the deleterious eect of climate change.

Motivation
Crop diversication is identied as the most ecologically feasible and
cost eective CSA to improve agricultural production and improve
food and nutrition security
Two possible channels:
Diversication towards nutrient dense crops has the potential to
improve nutrition for farm households
Enhancing farm household's income. Though the pathway is not
always direct and linear (food market, knowledge and preferences)
However, the direct causal link between crop diversication and nutrition is
not simple and the existing empirical evidence is mixed.

LiteratureSnapshot
The existing scant empirical evidence suggest that adopting crop
diversication improves the consumption of the poorest and reduces
poverty (Tesfaye and Tirivayi, 2020; Sibhatu and Qaim, 2018).
Nevertheless, studies has not moved down to an empirical analysis of how
crop diversication aect poverty dynamics.
Static poverty measures fail to distinguish between an individual who
has been in poverty all her life, and another who happens to have had
a small misfortune for the year the measurement was carried out.
Signicant numbers of people move into poverty throughout their
lives, one third African population move into and out of poverty
(Dang and Dabalen, 2019).

Objective
Assess the eect of crop diversication on farm household's poverty
dynamics

Data
Nigeria General Household Survey (NGHS)- 2010/11, 2012/13, and 2015/16
Living Standards Measurement Study Integrated Surveys on
Agriculture (LSMS-ISA) program of the World Bank in collaboration
with the Nigerian National Bureau of Statistics (NBS)
Panel of 5000 hh and 14,000 individuals
Geo-reference households
Temperature and precipitation data
The Climatic Research Unit (CRU-TS-4.03), University of East Anglia
Outcome variable: Poverty status of households = HH consumption
per adult equivalent + national poverty line (137,430 Naira or 382
USD per year)
Determinants of poverty status of households - household
characteristics + household head characteristics + crop diversity
indices + climate variability

Data
Table: Denition of crop diversity indices
Index Mathematical
Construction
Explanation Adaptation in this paper
Number of crops D=S Richness A Household produced S number
of crops
Shannon-Weaver D = −Σpi ln pI , D 0 Proportional abundance and Richness pi is proportion, or relative abun-
dance of a species
D 0
Composite Entropy D = −
Pp
i pi lns (pi ) (1 − 1/S), Proportional abundance and Richness p_i is proportion, or relative
abundance of a species
0 ⩽ D ⩽ 1

Data

Endogenous switching model
Cappellari and Jenkins (2002, 2004).
It models poverty transitions between two consecutive waves using a
trivariate probit model.
Four parts:

Four parts:
The determination of poverty status at t,

Four parts:
The determination of household retention between t − 1 and t,

Four parts:
The determination of poverty status at t − 1 in order to account for the
initial conditions problem,

Four parts:
The correlations between the unobservables aecting all the three processes

Four parts:
The correlations between the unobservables aecting all the three processes
The combination of all the four parts characterizes the determinants
of poverty persistence and poverty entry rates

Initial poverty status equation
p∗
it−1 = β′
xit−1 + uit−1 (1)
where:
i = 1, · · · , N indexes households and t = 1, · · · , T time span,
p∗
it−1 is the latent dependent variable,
xit−1 is a vector of controls describing i's household characteristics,
β is a vector of parameters to be estimated and the error term,
uit−1 = δi + µit−1 (the sum of an household-specic eect and an
orthogonal white noise error),
pit−1 = 1[p∗
it−10] (2)

Retention equation
r∗
it = γ′
wit−1 + εit (3)
where
i = 1, · · · , N indexes households and t = 1, · · · , T time span,
r∗
it 's is the latent retention propensity,
the error term εit = ηi + ϑit (the sum of an household-specic eect ηi plus
an orthogonal white noise error ϑit)
wit−1 is a vector of controls describing i's household characteristics and
γ is a vector of parameters to be estimated
rit = 1[r∗
it 0] (4)

Poverty transition equation
p∗
it = [(pit−1)λ′
1 + (1 − pit−1)λ′
2] zit−1 + ϵit (5)
where:
i = 1, · · · , N indexes households and t = 1, · · · , T time span
p∗
it is the latent dependent variable
λ′
1, λ′
2 are parameter vectors to be estimated
zit−1 denotes vector of controls
the error term ϵit = τi + ξit (the sum of an household specic eect τi plus
an orthogonal white noise error ξit)
(pit|pit−1, rit = 1) = 1[{(pit−1)λ′
1+(1−pit−1)λ′
2}zit−1+ϵit κt ] (6)
Note that pit is only when rit = 1
Estimation method: Maximum likelihood (multivariate approach of
Cappellari and Jenkins, 2004).

The joint distribution of the error terms uit−1, εit and ϵit is trivariate
standard normal - unrestricted and estimable.
The distribution of unobserved households' heterogeneity is parameterized
via the cross-equation correlations.
Other things being equal,
ρ1 (corr b/n unobservable aecting pit−1, and rit) and ρ3 (corr b/n
unobservable aecting rit and pit) are equal to zero, panel attrition is
random and joint estimation of retention equation can be ignored,
the model reduces to a bivariate model.

ρ2 (corr b/n unobservable aecting (pit−1 and pit) and ρ1 (corr b/n
unobservable aecting pit, and rit) are equal to zero, the initial condition
can be ignored and past poverty experience can be treated as
exogenous.

ρ2 (corr b/n unobservable aecting (pit−1 and pit) and ρ1 (corr b/n
unobservable aecting pit, and rit) are equal to zero, the initial condition
can be ignored and past poverty experience can be treated as
exogenous.
ρ1 = ρ2 = ρ3 = 0 both initial poverty and sample attrition are exogenous
and the model reduces to a univariate probit model.

Testing validity of estimation strategy

Results

Conclusion
We bring a new applied evidence from one of the poorest country in SSA to bear
on the on-going debate whether crop diversication....
We found initial conditions and sample retention are endogenous to poverty
transitions in our panel.
Adopting crop diversity is negatively associated with poverty entry but does
not aect poverty persistence.
Mitigating the eect of climate change on welfare might require adoption of
more than one CSA than one-size ts all interventions.

Thank you!
This work was supported through the Climate Research for
Development (CR4D) Postdoctoral Fellowship [CR4D-19-17].

Eleni Yitbarek_2023 AGRODEP Annual Conference

Recomendados

Recomendados

Mais conteúdo relacionado

Semelhante a Eleni Yitbarek_2023 AGRODEP Annual Conference

Semelhante a Eleni Yitbarek_2023 AGRODEP Annual Conference (20)

Mais de AKADEMIYA2063

Mais de AKADEMIYA2063 (20)

Último

Último (20)

Eleni Yitbarek_2023 AGRODEP Annual Conference