The Impact of Targeting Mechanisms on Efficiency and Equity of Irrigation in Mozambique

Last registered on November 13, 2015

Pre-Trial

Trial Information

General Information

Title
The Impact of Targeting Mechanisms on Efficiency and Equity of Irrigation in Mozambique
RCT ID
AEARCTR-0000945
Initial registration date
November 13, 2015

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
November 13, 2015, 8:28 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Region

Primary Investigator

Affiliation
World Bank

Other Primary Investigator(s)

PI Affiliation
World Bank
PI Affiliation
UC San Diego
PI Affiliation
Cornell University

Additional Trial Information

Status
In development
Start date
2015-11-17
End date
2017-09-30
Secondary IDs
Abstract
Irrigation is more important than ever in ensuring sustainable livelihoods for farmers In the face of increasing climate uncertainty. Yet, in the poorest parts of the rural world, the rates of irrigation are low. In Mozambique, only 8% of all farmers have access to irrigation. One source of Mozambique’s inability to develop widespread irrigation infrastructure has been an inability to sustain and protect previous investments in the sector. A lack of attention to creating local institutions capable of maintaining the infrastructure has historically led to infrastructure decaying before reaching its productive potential (FAO, 2005). Forming strong institutions begins with selecting the right set of beneficiaries who can function and cooperate together to maximize productive potential of schemes.

The evaluation aims to shed light on if different approaches to select beneficiaries can reduce elite capture and result in a more equitable distribution of benefits. However, the potential trade-offs between inclusion of smaller farmers and successful management remain an empirical question. We will exploit exogenous variations in the composition of water users groups induced by the random assignment to the two targeting regimes to shed light on the causal relation between group composition and collective action over operation and maintenance of the schemes, and final impact on production.
External Link(s)

Registration Citation

Citation
Christian, Paul et al. 2015. "The Impact of Targeting Mechanisms on Efficiency and Equity of Irrigation in Mozambique." AEA RCT Registry. November 13. https://doi.org/10.1257/rct.945-1.0
Former Citation
Christian, Paul et al. 2015. "The Impact of Targeting Mechanisms on Efficiency and Equity of Irrigation in Mozambique." AEA RCT Registry. November 13. https://www.socialscienceregistry.org/trials/945/history/6017
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The project will provide 56 small scale irrigation kits in the southern province of Gaza. The project will be implemented in the four drought affected districts of Guija Mabalane, Chicualacuala and Massengena. Since the number of farmers that are interested in receiving irrigation equipment will likely exceed the number that can be covered, a within village level selection procedure is necessary. The evaluation will test two different approaches to carry out the within village selection of beneficiaries to understand how different protocols can result in a more equitable distribution of benefits and if there are potential trade-offs between inclusion of smaller farmers and productivity gains. The main features of the models are:

Model 1 - Score-based targeting. Selection happens according to a fixed set of criteria for placing the schemes and is designed to prioritize the smaller farmers in the community.

Model 2 - Decentralized Community-driven targeting. The community will have the freedom to decide who they want to benefit from the kit.
Intervention Start Date
2015-11-17
Intervention End Date
2017-03-30

Primary Outcomes

Primary Outcomes (end points)
Data will be collected both at the community, farmer and plot level. The key outcome variables are:

Name // Definition // Measurement Level

- Yield // Total Revenue per hectare (mean and variance) // Individual/plot
- Profit // Total income received from crop sales net input costs (mean and variance) // Individual/plot
- Irrigation usage // Number of hours the system is used per day // Scheme
- Plot size // Plot size (hectare) within the irrigation kit // Individual
- Equity of plot size // Z score of plot size (hectare) within the irrigation kit // Individual
- Number of beneficiaries // Number of farmers covered by the kit // Scheme
- Total cultivated land // Sum of area of all cultivated plots of beneficiaries (hectare) // Individual
- Within scheme land distribution // Z score of total cultivated land of beneficiaries // Scheme
- Functioning of the irrigation equipment // Number of problems reported; cost of repairs; time required for repairs // Scheme
- Input usage // Plot level outcome
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The evaluation will aim to respond the research questions by testing two different approaches to select beneficiaries within the selected community. The identification strategy will be based on random allocation of protocols of models 1 and 2 to the project communities. The unit of randomization will be the community. The list of communities will be provided by the project team and will include 56 locations in total, of which 28 will be allocated to each of the models. The randomization will be stratified by district and allocated kit size. Identification of the program effects will be based on comparing the outcomes of interest between communities assigned to Model 1 and communities assigned to model 1. Additional specifications of treatment effects will use treatment assignment to one of the two models as an instrument for receiving irrigation equipment or constructing a regression discontinuity from the eligibility scores. The survey sampling frame will be based on 1) all farmers included in the irrigation kit and 2) a random sample of farmers outside the kit yet inside the eligible area.
Experimental Design Details
Randomization Method
Randomization done in office by a computer using Stata.
Randomization Unit
The evaluation will aim to respond the research questions by testing two different approaches to select beneficiaries within the selected community. The identification strategy will be based on random allocation of two protocols of beneficiary selection to the project communities. The unit of randomization will be the community. The list of communities will be provided by the project team. The randomization will be stratified by district and allocated kit size.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
56 communities
Sample size: planned number of observations
1400 farmers. In each community on average 25 farmers will be selected to participate in the survey. The number of beneficiaries will vary from 1 - 20 farmers.
Sample size (or number of clusters) by treatment arms
In each of the treatment arms we will have 28 communities.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Our sample size is limited by the number of communities that will be treated and the number of farmers that will be surveyed in each of the communities. The project will build around 56 irrigation schemes. In each of the communities, we will survey a random sample of in total of 25 farmers, resulting in a total sample of 1,400. Given these parameters on cluster size and the number of observations within each cluster, we calculate the minimum detectable effects (MDEs) assuming a power of 0.8 and an alpha of 0.1. Where there is no information available, we calculate them under potential scenarios of intra-cluster correlation (ICC). Below we present the power calculations for the main indicator directly targeted by the intervention: proportion of small farmers in the eligible area that are selected. The final outcome indicator of interest is yield. Proportion of small farmers The main outcome indicator of interest is proportion of small farmers in the eligible area that are included in the scheme. We will select a random sample of 25 farmers within the eligible area to determine the distribution of cultivated land in the area. From the TIA data we see that in the region around 75% of farmers cultivate 2 or less hectares. We will compare the probabilities of these 19 small farmers to be selected under the two different selection protocols. Without reliable information on ICC we report the MDEs under a wide range of parameters. Departing from a scenario where under model B half of the small farmers are included, which yields the most conservative results, the MDE of an increase in small farmers being included ranges from 0.11 -0.31. Yield Our sampling frame for yield will be based on the on average 15 households included in the scheme. To calculate the MDE on yield we use data from a Smallholders Survey collected in an adjacent region of Mozambique (Kondylis, Mueller, and Zhu, 2015). Mean maize revenue per hectare (MZN/ha) is 1051, standard deviation 1186 and ICC 0.10. With a baseline and one follow up for yield outcomes we should be able to detect a 20.1% increase in revenue.
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials