Impact Evaluation of the Solar Irrigation Program in Yemen

The study will include 1 intervention where randomly selected wells (their respective well owners and smallholder farmers connected to that well) will be provided with financial grants and technical support. Financial grants with a ceiling of $15,000 will be provided to each well owner to procure solar panels and solar-powered water pumps. A financial grant of $1,400 will be provided to a smallholder farmer to procure drip irrigation system for 0.5 hectares of farming land, and other agricultural inputs such as seeds and seedlings, mulch, and fertilizers to promote modern, sustainable practices and maximize yields and product quality. The financial grants will follow a matching grant scheme of 80/20 – 80% project subsidy, 20% beneficiary contribution to ensure sustainability and ownership of the assets. Technical support component of the intervention will be in the form of technical training through agronomists and linking the beneficiaries with other market players such as input suppliers and microfinance institutes (MFIs). Training will cover the following topics: i) technical training on the solar-powered water pump components, operation, different types of solar-powered pumps and their respective advantages and disadvantages, ii) technical, economic, and environmental feasibility of solar-powered pumps, iii) importance and methodology of linking solar-powered pumps to a modern, water-conserving irrigation network, iv) daily and periodic maintenance of solar-powered pumps, and v) safety and security measures during installation and operation. The wells (their respective well owners and smallholder farmers connected to those wells) randomly assigned to control and spillover groups will not receive financial grants nor technical support.

2023-10-28

2024-06-30

1. Production
2. Income
3. Consumption

1. Production variable will measure the land area planted with crops and the quantities harvested. As such, yields will be calculated for each crop planted by smallholder farmers in September 2022 and September 2023 planting seasons, using self-reported land area and self-reported quantities harvested.

2. Income outcome variables such as total income and agricultural income will be generated from household questionnaire that will ask about income sources and transfers in the last 12 months including agricultural income, income from family businesses, properties, and wages, as well as cash transfers, loans, and social assistance.

3. Consumption will be measured using a 24-hour recall of consuming food items from different food groups, and through a food consumption coping strategy index based on a 7-days recall.

1. Water consumption
2. Water extraction
3. Farming practices
4. Household time use
5. Social cohesion
6. Access to basic services
7. Food availability

1. Water consumption outcomes for smallholder farmers will be collected using a household questionnaire:
i. Availability of water for agricultural purposes will be measured as frequency at which water was available for crop irrigation during September 2022 and 2023 planting seasons. Availability of water for non-agricultural purposes will be measured through the main source of drinking water for the household such as well, stream, water trucking, etc.
ii. Efficiency of water use for agricultural purposes will be measured as an indicator variable where the use of drip irrigation systems will represent efficient use of water (water preserving), whereas flooding areas will represent inefficient use of water due to evaporation, seepage, or pours into the drainage.
iii. Cost of water for agricultural purposes will be measured as: i) the amount paid in Yemeni Rial (YR) per week for irrigation water during September 2022 and 2023 planting seasons, and ii) the monetary value (YR) of crop given to water provider during September 2022 and 2023 seasons. The cost of water for non-agricultural purposes will be measured as YR paid for water consumption in the past 30 days.

2. Water extraction outcomes will be collected using a questionnaire for water suppliers, i.e., well owners:
i. Number of solar-powered well pumps.
ii. Number of farmers connected to the well.
iii. Well water level in meters.
iv. The volume of water pumped in meters cubed per hour.
v. The use of pumped water for different purposes in the past week will be measured as an indicator variable for the following categories of use: i) on crops owned by the well owner, ii) on sharecropped fields, iii) through irrigation contracts to provide water for nearby farmers at a cost, iv) through non-irrigation contracts to provide water for nearby households at a cost, v)providing water for nearby households for non-irrigation purposes at no charge.
vi. Well water supply duration throughout September 2022 and 2023 agricultural seasons.
vii. Well water supply duration outside of September 2022 and 2023 agricultural seasons.
viii. The cost of operating the well in the last week will be measured through the price, quantity, and delivery charge of diesel.
ix. Price of pumped water will be measured in terms of YR/hour, YR/day, YR/ meter cubed of pumped water, diesel compensation per hour of operating the pumps (l/h), and diesel compensation per meter cubed of pumped water (l/meter cubed).

3. Improved farming practices such as crop rotation, terracing, etc. will be measured through farmers’ awareness and use of these practices in the past two agricultural seasons. (Improved) farming practices will also include the use of complementary agricultural inputs that will be measured through the use of improved seed, chemical and organic fertilizer, pesticide, hired and household labor in September 2022 and 2023 seasons.

4. Household time use outcome will represent the total number of hours spent on each task by main male and female household members in the past 7 days. The tasks included will be fetching water, raising livestock, household chores, etc.

5. Social cohesion outcomes will be measured using respondents’ attitudes toward common goods and social relations.

6. Access to basic services will be measured as the ownership, type, and characteristics of household dwelling and its distance from services and markets.

7. Availability of different foods to be purchased locally will be measured at the community level.

Recruitment of participants will be carried out by The Small and Micro Enterprise Promotion Service (SMEPS). SMEPS will first work with local authorities to identify areas that would be suitable for the project. Then, SMEPS will identify eligible wells and eligible smallholder farmers connected to each eligible well through their selection criteria. The criteria for eligibility of wells will be as follows: i) the well owner has not received any kind of similar support in agriculture, ii) the well owner is providing water for agriculture to 5-9 smallholder farmers, iii) the well is permanent and its water productivity is greater than smallholder farmers’ needs, iv) the depth of the well does not exceed 150 meters, and v) the well owner is willing to pay the 20% financial contribution to install solar panels and solar-powered water pump. Upon identifying eligible wells, SMEPS will identify eligible farmers connected to each eligible well. The eligibility criteria for smallholder farmers will be: i) the smallholder farmer should be a practicing farmer, i.e., planning to grow horticultural products in September 2023 planting season, ii) the smallholder farmer should own 0.25-2 hectares of agricultural land, iii) the smallholder farmer should be supported by the eligible well, with a water transmission network from the irrigation source (from the well to the farm), and iv) the smallholder farmer should be willing to work, contribute, and commit during the season and duration of the project.

Using the list of all eligible wells, we will randomly select and assign wells in the following way: i) 16 wells (8 shallow wells below 40 meters, and 8 deep wells above 40 meters) will be randomly assigned to treatment group, and ii) 25 wells will be randomly assigned to control group, out of which we will then assign 5 wells closest to the treatment wells to spillover group. The smallholder farmers that are connected to randomly selected wells will be enrolled in the study. The assignment of these farmers to treatment, control, and spillover group will match the random assignment of wells, i.e., smallholder farmers connected to a treatment well will be assigned to treatment group, smallholder farmers connected to a control well will be assigned to control group, and smallholder farmers connected to a spillover well will be assigned to spillover group.

The intervention will be implemented by SMEPS through connecting well owners and smallholder farmers to input suppliers. More specifically, SMEPS consultants will determine well owners’ and smallholder farmers’ needs, collect prices from local markets, analyze price offers, prepare and deliver acceptance letters to suppliers, receive and inspect the inputs, install and test the inputs, complete authorization process and disburse funds to the suppliers. Additionally, SMEPS will conduct a set of trainings to the well owners and smallholder farmers on sustainable irrigation and farming practices. Interaction Consultancy will carry out two rounds of data collection for the study consisting of baseline and endline surveys. All well owners and smallholder farmers assigned to control and spillover groups will be surveyed at the same interval as treatment well owners and smallholder farmers but will not receive financial grants nor technical training.

Not available

In office by a computer using Stata.

Well (assignment to treatment, control, or spillover).

Yes

41 wells (level of randomization and clustering).

5-9 households per well, 205-369 households for baseline and follow-up surveys.

1. Treatment: 16 wells (80-144 households).
2. Control: 20 wells (100-180 households).
3. Spillover: 5 wells (25-45 households).

Horticulture yields, as the measurement of smallholder farmers’ productivity, will be our primary outcome and they will be measured at the smallholder-farmer level. Our power calculations were based on tomato yields and were done for a two-way comparison between control and treatment group. Based on FAO data for tomato yields in Yemen of 12 kg/ha, a 0.05 significance level, 0.8 statistical power, and 0.1 intra-cluster correlation, a sample of 16 treatment and 20 control clusters (wells) and a cluster size of 6 smallholder farmers (per well) provides a minimum detectable effect size of 2.8 kg/ha difference in yields. This represents around 24% increase in yields, and results in Cohen’s d of 0.47 which is considered a small to medium effect size (Cohen, 1988). References: Cohen, J. (1988) Statistical Power Analysis for the Behavioral Sciences (2nd ed.). Hillsdale, NJ Lawrence Erlbaum Associates.