Cashew Subsidy Voucher Pilot Impact Evaluation 2021-2022

Baseline and Endline Survey Rounds

Eligible sprayers and farmers were selected based on the ConnectCaju database that was designed and maintained by IAM and TNS. This database was developed to better track the developments within the industry and evaluate policy targeting the sector. The database collects information on farmers and sprayers including demographic data, the size and location of a farmer’s/sprayer’s orchard, and how many trees farmers/sprayers have. Moreover, IAM audits all farmers and sprayers in the database to verify the information and geo-tag orchards. Based on this database, TNS invited eligible sprayers to participate in a meeting that introduced them to the pilot, its modalities, and its objectives. Sprayers and farmers were then asked whether they would be interested in participating in the pilot and checked for their eligibility. Eligible sprayers had to: 1) Be located in the localities selected for pilot implementation, 2) Have farmer clients in the area they operate that they can service, 3) Have previous farmer clients that own between 50 and 500 trees, 4) Express an interest participating in the pilot.
Following the meetings, TNS and IAM prepared a list of 40 candidate sprayers to participate in the pilot which were randomized into Treatment and Control group. Out of these 40, 6 needed to be excluded from our sample before of the implementation of the study started due to the following reasons:

• 4 sprayers were dropped because they did not have any client farmers they work with;
• 1 sprayer could not participate in the pilot due to health reasons;
• 1 sprayer became a sprayer for one of the cooperatives and was not available for the pilot anymore.

While the randomization was done before these sprayers were identified as ineligible or decided to not participate in the pilot, this should not negatively affect the outcome of our randomization since 3 of these sprayers were assigned to the Control and the remaining 3 to the Treatment group. Moreover, we conducted balance checks on sprayer characteristics of sprayers that participated in the pilot which show that both Treatment and Control groups are comparable across a large set of observable characteristics. As a result, the pilot was implemented using a sample of 34 eligible sprayers to participate in the pilot. In the three localities, 20 sprayers are located in Namaponda, 8 in Nametil, and 6 in Nanhupo Rio.
The 34 eligible sprayers were then asked to provide a list of farmers they provided spraying services to in 2020. These lists were then updated using monitoring data from TNS on the sprayers’ new clients during the 2021 spraying season. The resulting farmer lists include a total 845 farmers across all three localities and consist of two distinct group:
• Previous farmer clients: Farmers that were listed by eligible sprayers prior to the pilot. Previous farmer clients had to fulfil the following eligibility criteria:

1. Live in the catchment area of eligible sprayers, 2) Have been clients of the eligible sprayer in 2020, 3) Have between 50 and 500 cashew trees. In total, 358 previous clients were identified to participate in the pilot.
   • New client farmers: New client farmers that were listed during the spraying season 2021 and did not have to fulfil any eligibility criteria. This group includes both farmers that have not worked with a given sprayer in 2020 (30% of new clients) and farmers that have worked with a given sprayer in 2020 but were ineligible for the pilot vouchers (70% of new clients). In total, 487 new client farmers were identified during the spraying season. Including new clients in the evaluation provides further insights into sprayers’ decisions to service clients depending on whether free inputs were available with or without vouchers.