Abstract:Our study seeks to identify the social contagion of in-group gifts: if gifts trigger their recipients to send more gifts subsequently, the actual impact of a gift on group dynamics would be greatly amplified. Causal identification of contagion is always challenging in observational data; To identify gift contagion, we leverage a natural experiment using a large sample of 36 million online red packets sent within 174,131 chat groups on WeChat, one of the largest social network services worldwide. Our natural experiment is enabled by WeChat's random gift amount algorithm, with which the amount that a user receives does not depend on her own attributes. We find that, on average, receiving one more dollar causes a recipient to send 18 cents back to the group within the subsequent 24 hours. Moreover, this effect is much stronger for recipients or those who receive the largest share from a red packet, suggesting a group norm according to which the luckiest draw recipients should send the first subsequent red packets. Additionally, we find that gift contagion is affected by in-group friendship network properties, such as the number of in-group friends and the local clustering coefficient.

Identify and Understand Pay-It-forward Reciprocity Using Millions of Online Red Packets.