Working Paper: NBER ID: w18351
Authors: Vivi Alatas; Abhijit Banerjee; Arun G. Chandrasekhar; Rema Hanna; Benjamin A. Olken
Abstract: We use unique data from 600 Indonesian communities on what individuals know about the poverty status of others to study how network structure influences information aggregation. We develop a model of semi-Bayesian learning on networks, which we structurally estimate using within-village data. The model generates qualitative predictions about how cross-village patterns of learning relate to different network structures, which we show are borne out in the data. We apply our findings to a community-based targeting program, where villagers chose which households should receive aid, and show that networks the model predicts to be more diffusive differentially benefit from community targeting.
Keywords: network structure; information aggregation; community targeting; Indonesia; social learning
JEL Codes: D83; D85; H23; O12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| network characteristics (D85) | predict how well information is aggregated (D80) |
| better-connected households (R20) | more accurate ranking of economic status (D31) |
| increase in average degree (C29) | decrease in error rates in ranking (C52) |
| shorter path lengths (R12) | more likely to be accurately ranked (C52) |
| non-responses in ranking (C83) | correlated with network structure (D85) |
| social distance (Z13) | more likely to say do not know relative wealth (D89) |