Identify and understand pay-it-forward reciprocity using millions of online red packets

Yuan Yuan, Tracy Xiao Liu, Chenhao Tan, Qian Chen, Alex Pentland, Jie Tang

Introduction

Social contagion—the spread of behaviors through a network—is a fundamental concept in social science, yet proving its existence is notoriously difficult. When friends exhibit similar behaviors, such as digital gift-giving, researchers must distinguish between true contagion (one person's action causing another's), homophily (similar people befriending each other), and external factors (like holidays). This paper addresses this challenge by analyzing "red packets"—digital monetary gifts—on a major East Asian social network.

Methodology: A Natural Experiment in Code

The researchers utilized a massive, anonymized dataset comprising 3.4 million users and 36.6 million red packets. To isolate causal effects, they exploited the platform's randomized payout algorithm. When a user sends a "random-amount" red packet to a group, the total sum is divided among recipients based on a stochastic process.

The Red Packet user interface flow, from group selection (a) to sending (c) and receiving (f).

Because the exact amount a recipient receives is determined by random variance (controlled for the total amount and opening order), the researchers could treat the payouts as a stratified randomized experiment. This allowed them to measure how receiving an extra unit of currency causally impacts the recipient's subsequent likelihood of sending a gift back to the group within 24 hours.

Key Findings: The Contagion Effect

The study found clear evidence of gift contagion. On average, receiving an additional dollar causes a recipient to send approximately 18 cents back to the group. Interestingly, this effect is driven primarily by the "extensive margin"—meaning that receiving money encourages people who wouldn't have otherwise given to participate, rather than simply increasing the gift sizes of existing senders.

The "Luckiest Draw" Norm

A pivotal discovery was the influence of the "Luckiest Draw" social norm. The platform's interface highlights the individual who receives the largest share of a packet with a gold crown icon.

The results screen (f) highlights the 'Luckiest Draw' recipient, who faces a strong social obligation to send the next packet.

Recipients designated as the "Luckiest Draw" are nearly ten times more likely to send the next red packet compared to other recipients. This suggests that the platform's UI design actively reinforces a group norm of reciprocity, where the biggest winner feels a "fairness-based" obligation to keep the chain going.

Social and Network Dynamics

Beyond immediate reciprocity, the study explored how these digital gifts affect the broader social fabric:

  • Network Growth: Receiving larger random payouts correlates with an increased likelihood of adding other group members as direct friends. For every 100 CNY received, users added an average of 0.055 new friends within the group over the following week.
  • Clustering Effects: Interestingly, high network clustering (where a user's friends are also friends with each other) appears to dampen contagion. Users in highly interconnected clusters were less susceptible to the pressure of reciprocating gifts.
  • Contextual Variation: The contagion effect was significantly more pronounced during festival periods, such as the Lunar New Year, and within groups composed of family members.

Conclusion and Future Directions

This research provides robust causal evidence that digital gifts trigger a "pay-it-forward" cycle. While the study identifies the "Luckiest Draw" as a powerful driver, the exact psychological mechanisms—whether pure reciprocity, guilt, or social status—remain a subject for future field experiments. Additionally, while these dynamics are deeply rooted in East Asian cultural concepts of guanxi (relationships), they offer valuable insights into how platform design can foster social solidarity and engagement in digital communities worldwide.

Summary updated

BibTeX
@article{yuan2019identify,
  author = {Yuan Yuan and Tracy Xiao Liu and Chenhao Tan and Qian Chen and Alex Pentland and Jie Tang},
  title = {{Identify and understand pay-it-forward reciprocity using millions of online red packets}},
  journal = {Manag. Sci.},
  year = {2019},
}
Figure