Subscribe to E-Bulletin Donate to CEGA

Social Networks and the Decision to Insure

Development Challenge

Smallholder farmers in developing countries are often forced to cope with unpredictable weather shocks such as severe floods and droughts. These shocks can severely reduce farmers’ yields and profits for the season.  In recent years, weather index insurance products have been designed to protect farmers from such shocks. Despite potential benefits, adoption of insurance products has remained low. Little evidence exists on how to effectively increase use of weather insurance among farmers.

Context

In China, nearly half of farmers produce rice, making it the most important staple crop in the country. In 2009, the People’s Insurance Company of China (PICC) developed an insurance product at the request of the government to better protect rice farmers from unpredictable shocks. Farmers received a payout if a natural disaster – heavy rain, flood, windstorms, or extreme temperatures – led to at least a thirty percent loss in yields, and payouts increased linearly with the rate of loss.  The maximum payout was 200 RMB per mu.[1] The study included 5,332 households in 185 villages in Jiangxi province. Nearly two thirds of households in the study had experienced a weather shock in the past year.

Evaluation Strategy

Researchers sought to identify whether and how social networks in China affected the adoption of a weather index insurance product. They asked participating households to rank five close friends with whom they discuss rice production or finances to construct each village’s social network.  The farmers were then randomly assigned to one of two educational sessions held three days apart. They were further randomized to either a simple or an intensive session within the round creating four treatment arms. Those in T3 and T4 were then randomly subdivided into three groups: one third received no additional information, one third were provided overall attendance and take-up rates from the first round, and one third were given detailed lists of who purchased insurance. The randomization scheme allowed researchers to identify the mechanisms driving any social network effects. 

The researchers also conducted two village level randomizations. First, they randomly varied the price of insurance in one set of villages. They also varied the default option for buying the insurance (opt-in vs opt-out to buy). See the figure below for details of the randomization. Village level randomization allowed researchers to quantify the value of the social network effect.

Results and Policy Implications

The results showed that households in the first round intensive information sessions had higher take-up rates.  The adoption of insurance among this group was 50 percent – nearly 15 percentage points--higher than those that received the simple session in the first round. The results also indicate significant, positive social network spillover effects. Having a friend in a first round intensive session increased a farmer’s take-up rate by more than six percentage points. Having two close friends attend the first round increased take-up rates among second session farmers by 20.6 percentage points while having more than two had only a slightly higher effect. Farmers are also more strongly influenced by those they consider closest in their social networks. Farmers who received the intensive session in the second round were less likely to be influenced by their peers suggesting that the impact of social networks depends on whether farmers received direct financial education. The research identified that the mechanism through which social network effects work was primarily the diffusion of knowledge about insurance, not knowledge of others’ purchase decisions. Finally, they found that social networks effects decreased farmers’ sensitivity to price increases.

The results suggest that providing intensive financial education to a few farmers may increase broader adoption of new insurance product as knowledge is disseminated through their social networks. It is important to target the most influential and central farmers within the village to effectively use these networks. 

Timeline

2010

[1] 1 RMB = .15 USD; 1 mu = .067 hectare

Photo Credit: McKay Savage via flickr Creative Commons. China - Yangshuo 29 - Rice Paddy Terraces.