Viability of an Irrigation Development Intervention in Tigray: an Application of Stochastic Impact Evaluation

Abstract

Irrigation dams and other forms of flood and rainwater harvesting infrastructure may make important contributions to help farmers reduce poverty, improve food and nutrition security, and adapt to the impacts of climate change. Water harvesting and retention systems are of particular importance in arid and semi-arid areas, such as Northern Ethio-pia, where rain fed agricultural production faces severe drought risks. Dam construction incurs high costs, however, and governments or other investors are often unsure of whether such investments are justified. Assessing the feasibility of water harvesting infrastructure investments and optimising expected returns from them requires detailed ex ante appraisal. Due to the inherently complex and uncertain consequences of agricultural investments and often severe data scarcity, traditional cost-benefit assessment methods face limitations. Stochastic Impact Evaluation (SIE) attempts to overcome the particular challenges of evaluating investments in such contexts. Here we assess the viability of an irrigation dam intervention in Raya valley, Ethiopia. To achieve this, we worked with stakeholders to generate a causal model of the planned intervention's impact pathway. We then applied an SIE approach based on integration of Monte Carlo simulation, Partial Least Squares regression, and Value of Information analysis. The model was developed and estimates for the input variables were collected from ten subject matter experts via expert data elicitati-on methods. Preliminary results indicate that the effect of the proposed dam project varies for the different stakeholders involved. Further analysis is underway to identify the variables with high information value and whose measurement could best inform the investment decision. If further information is needed in order to decide on a preferable course of action, decision makers should target these variables for measurement. Our results demonstrate that SIE is an effective approach for providing guidance to decision-makers in agricultural development, in the face of system complexity and uncertainty.

Full text

Tropentag, September 20-22, 2017, Bonn
“Future Agriculture:
Socio-ecological transitions and bio-cultural shifts”
Viability of an Irrigation Development Intervention in Tigray: an
Application of Stochastic Impact Evaluation
Negusse Yigzaw1,2, Cory Whitney2,1, Chris-Ackello Ogutu3, John Mburu3, Eike
Luedeling1,2
1World Agroforestry Centre (ICRAF), Kenya
2University of Bonn, Center for Development Research (ZEF), Germany
3University of Nairobi, Dept. of Agricultural Economics, Kenya
Abstract
Irrigation dams and other forms of flood and rainwater harvesting infrastructure may
make important contributions to help farmers reduce poverty, improve food and nutriti-
on security, and adapt to the impacts of climate change. Water harvesting and retention
systems are of particular importance in arid and semi-arid areas, such as Northern Ethio-
pia, where rain fed agricultural production faces severe drought risks. Dam construction
incurs high costs, however, and governments or other investors are often unsure of whether
such investments are justified. Assessing the feasibility of water harvesting infrastructure
investments and optimising expected returns from them requires detailed ex ante apprai-
sal. Due to the inherently complex and uncertain consequences of agricultural investments
and often severe data scarcity, traditional cost-benefit assessment methods face limitati-
ons. Stochastic Impact Evaluation (SIE) attempts to overcome the particular challenges
of evaluating investments in such contexts. Here we assess the viability of an irrigation
dam intervention in Raya valley, Ethiopia. To achieve this, we worked with stakeholders
to generate a causal model of the planned intervention’s impact pathway. We then applied
an SIE approach based on integration of Monte Carlo simulation, Partial Least Squares
regression, and Value of Information analysis. The model was developed and estimates for
the input variables were collected from ten subject matter experts via expert data elicitati-
on methods. Preliminary results indicate that the effect of the proposed dam project varies
for the different stakeholders involved. Further analysis is underway to identify the varia-
bles with high information value and whose measurement could best inform the investment
decision. If further information is needed in order to decide on a preferable course of action,
decision makers should target these variables for measurement. Our results demonstrate
that SIE is an effective approach for providing guidance to decision-makers in agricultural
development, in the face of system complexity and uncertainty.
Keywords: Dam construction, investment viability, Irrigation, probabilistic simulation, Stochastic
Impact Evaluation
Contact Address: Negusse Yigzaw, World Agroforestry Centre (ICRAF), Nairobi, Kenya, e-mail: N.Yigzaw@cgiar.
org