Back to History

Fields Changed

Registration

Field Before After
Abstract This study examines how attachment to social networks, home villages, and ancestral farmland constrains climate adaptation among smallholder farmers in Pakistan. We recruit 1,440 households across 120 villages stratified by flood exposure. Two crossed household-level randomized treatments are administered during face-to-face surveys: an information treatment providing historical flood frequency and projected climate change impacts, and an insurance framing treatment presenting the same index crop insurance product as income protection, adaptation-enabling, or attachment-preserving. We measure insurance demand through a discrete choice experiment and multiple price list. We measure attachment constraints on adaptation through a willingness-to-accept experiment in which respondents are presented with scenarios that independently vary where they would live, what their social network would be, and which land they would farm (if any). This design allows decomposition of the total welfare cost of displacement into place, network, and plot components. This study examines how attachment to social networks, home villages, and ancestral farmland constrains climate adaptation among farmers in Pakistan. We recruit 1,440 households across 120 villages stratified by flood exposure. Two crossed household-level randomized treatments are administered during face-to-face surveys: an information treatment that provides historical flood frequency and projected climate change impacts, and an insurance framing treatment that presents an identical index crop insurance product as income protection, adaptation-enabling, or attachment-preserving. We record their past and planned adaptation decisions, demand for insurance, and relocation preferences using displacement vignettes that vary their people, place, and plot.
Trial Start Date June 27, 2026 August 24, 2026
Last Published June 29, 2026 08:37 AM August 19, 2026 12:25 PM
Intervention (Public) Households are randomly assigned to two independent treatments. The first is a flood information treatment: treated households receive the number of flood events in their union council over the past 10 years and a climate projection for the next 10 years. Control households receive no information. The second is an insurance framing treatment: households are randomly assigned to one of three descriptions of an otherwise identical index insurance product. F1 (income protection) frames the insurance as protecting income in a bad harvest year. F2 (adaptation) frames it as enabling continued farming and long-term investment despite growing flood risk. F3 (attachment) frames it as helping farmers stay on their land and in their community even after a bad year. Under their assigned framing, all households complete a discrete choice experiment across 7 insurance contract pairs and a multiple price list for three spatial trigger levels (village, sub-district, district). Households are randomly assigned to two independent treatments. The first is a flood-information treatment in which treated households receive information on the number of flood events in their union council over the past 10 years and a climate projection for the next 10 years. Control households receive no information. The second is an insurance framing treatment in which households are randomly assigned to one of three descriptions of an otherwise identical index insurance product. F1 (income protection) frames the insurance as protecting income in a bad harvest year. F2 (adaptation) frames it as enabling continued farming and long-term investment despite growing flood risk. F3 (attachment) frames it as helping farmers stay on their land and in their community even after a bad year.
Intervention Start Date June 27, 2026 August 27, 2026
Intervention End Date July 31, 2026 September 30, 2026
Experimental Design (Public) This is a household-level randomized controlled trial conducted in 120 villages in Pakistan. The study sample totals 1,440 households (12 per village). Villages are stratified into three flood-exposure strata (low, medium, high), with 40 villages per stratum. Within each village, 12 households are randomly assigned to one of the six treatment cells formed by crossing the information treatment (2 levels) with the insurance framing treatment (3 arms). The main analysis estimates the effects of the information treatment and framing arms on insurance demand and WTA outcomes, including their interaction, with village fixed effects included in all specifications. Additionally, we test the role attachment plays in mediating the decision to insure, adapt, and relocate. This is a household-level randomized controlled trial conducted in 120 villages in Pakistan. The study sample totals 1,440 households (12 per village). Villages are stratified into three flood-exposure strata (low, medium, high), with 40 villages per stratum. Within each village, 12 households are randomly assigned to one of the six treatment cells formed by crossing the information treatment (2 levels) with the insurance framing treatment (3 arms). The primary analysis focuses on the role attachments play in impacting climate adaptation, insurance, and migration decisions. Leveraging our experimental variation, we investigate whether the relationship is driven by beliefs, experiences, or other factors.
Secondary Outcomes (End Points) 1. Flood risk beliefs and perceived basis risk. 2. Insurance product comprehension. 3. Adaptation intentions, past adaptation behaviors, and policy support 4. Place attachment 5. Plot attachment 6. Network attachment 7. Risk and time preferences. 1. Flood risk beliefs and perceived basis risk. 2. Insurance product comprehension. 3. Adaptation intentions, past adaptation behaviors, and policy support 4. Place attachment 5. Plot attachment 6. Network attachment
Back to top