Decomposition of social networks and household purchase of insurance as knowledge products

This paper proposes a nested model of household social networks and re-examines how social networks affect household insurance consumption by decomposing inner and outer networks in tandem. Using the 2013 and 2015 waves of China Household Finance Survey (CHFS), we find that the inner networks substitute and the outer networks supplement insurance consumption. Our results help reconcile the inconsistency of the evidence on how social networks affect household insurance demand in the literature. Both the inner and outer networks contribute to the urban–rural gap of household insurance ownership. Mechanism analysis indicates that the inner networks affect household insurance purchase through risk-sharing and the outer networks through information dissemination and learning. Outer networks are of value to screen insurance companies by word-of-mouth. Surprisingly, insurance renewal analysis suggests that inner kinship networks have no persistent impeding effects, while the mutable outer networks persistently complement household insurance consumption. Households with male heads acquire stronger risk-sharing of inner networks than female heads.