Experimental games to teach farmers about weather index insurance in Kenya

3ie evidence programme:  Agricultural Insurance Evidence Programme
Author(s): Sarah Janzen¹, Nicholas Magnan², Conner Mullally³, Alessandra Garbero⁴, Karl Hughes⁵, Judith Oduol⁵, Bailey Palmer², Soye Shin²
Institutional affiliation(s): Montana State University¹, University of Georgia², University of Florida³, International Fund for Agricultural Development⁴, World Agroforestry Center⁵ 
Grant-holding institution: Montana State University
Main implementing agency: Montana State University
Sex disaggregation: Yes
Gender analysis: No
Equity focus: Yes
Study type: Formative evaluation

Context

This study was conducted in Tharaka South sub-county of Tharaka Nithi. Tharaka Nithi is located in the greater Meru region of the former Eastern Province, near Mount Kenya. The entire study sample resides in Tharaka South, one of the sub-counties of Tharaka Nithi, and mostly people live in Chiakariga ward. According to local government officials, most people in Tharaka South are subsistence farmers growing sorghum or green gram, with fewer growing millet, maize, cowpeas or pigeon peas. The land is semi-arid and not naturally very fertile, but farmers often do not have the liquidity to purchase fertilizers. Households generally supplement their agricultural income with handicrafts or small trade, but farming is the main source of income. Few if any households use irrigation, which is fairly representative of the national average. Most households keep livestock, including chickens and cows. Across Kenya, 75 per cent of the population participates in agriculture, and 75 per cent of agricultural production is by small-scale farmers (CIA World Factbook 2017, USAID 2017).

Unlike many agricultural innovations, learning about insurance products and other risk reducing technologies can take a long time. If shocks that result in payouts are infrequent, individuals who purchase insurance may not see that it pays out in bad years until a bad year occurs. Furthermore, for risks that are highly covariate, such as drought, one might not be able to readily learn from their peers, as experiences will not vary much within a given season. Thus, providing the opportunity for farmers to rapidly experience different outcomes with and without insurance could be an effective way to educate farmers and increase demand (Cai and Song, 2017).

Intervention design

This study evaluated two innovative interventions designed to increase demand for basis-risk focused weather index insurance (WII) products among smallholder producers in Tharaka South, sub-county of Tharaka Nithi county of Kenya. One intervention focused on an improved WII contract using unique high-resolution satellite data. The second was an experiential game that taught producers about basis risk and WII as a risk management tool.

The study theorised that the innovative designs of the interventions would increase farmer engagement and knowledge of basis risk, thereby increasing demand for WII.

In order to incentivize participation and simulate real risks in the game, small payouts relative to the actual price of insurance were used. The study measured demand for insurance by conducting a semi-binding auction immediately after the insurance game or PT game and examining actual uptake of insurance. This occurred prior to the normal insurance sales period.

Evaluation design and methodology

The study employed a 2x2 randomised control trial in which farmers were randomly assigned a contract type – high-resolution or low-resolution (HR or LR)  –  and then receive a basic information treatment or the same basic treatment plus the experiential game.

1. LR insurance, basic insurance information, PT game
2. LR insurance, basic insurance information, insurance game
3. HR insurance, basic insurance information, PT game
4. HR insurance, basic insurance information, insurance game

457 farmers who had previously worked with an insurance aggregator partner were surveyed, plus 30 farmers who had not but would become program farmers through the study. Farmers targeted for the study were semi-subsistence producers with no more than 10 acres of land.

Primary evaluation questions

1. Are farmers sensitive to basis risk?
   a)    Is demand for HR insurance higher than for LR insurance?
2. Does experiential learning (the game) affect demand for insurance?
   a)    Does playing the game increase or decrease demand for WII?
   b)    Does playing the HR game have a different effect on demand for HR insurance than playing the LR game has on demand for LR insurance? c)    Do negative basis risk events (to the farmer and to others) in the game decrease demand for WII?
   d)    Do positive basis risk events (to the farmer and to others) in the game increase demand for WII?
3. Does experiential learning (the game) affect attitudes toward and knowledge of insurance?
   a)    Does playing the game alter attitudes/understanding?
   b)    Do negative basis risk events (to the farmer and to others) in the game alter attitudes/understanding?
   c)    Do positive basis risk events (to the farmer and to others) in the game alter attitudes/understanding?

Primary findings

Farmers offered the LR insurance participating in the experimental game demanded KSH 4,123 more insurance relative to the status quo (a 31 percent increase). Farmers offered HR insurance not participating in the game demanded KSH 5,093 more insurance (a 38 per cent increase).  There were no significant effects of the game on attitudes for farmers offered either insurance product. Interventions increased stated demand for insurance, but did not increase actual uptake even when insurance prices were substantially below market prices.

Implications

Farmers are sensitive to the resolution index insurance products. Games can be effective at promoting stated insurance demand, although there is uncertainty as to under what conditions. Researchers recommend continuing this line of research but with changes to the index insurance product, by bundling with credit. Additionally, yield insurance rather than weather index insurance should be offered because farmers may prefer it due to coverage of multiple perils and is based on yields and not the relationship between weather and yields.