A Behavioral Micro-foundation for Cross-sectional Network Models
A continuous time stochastic model links individual preferences to network structures...
Researchers Butts and Murray-Watters have published a paper on arXiv that tackles a long-standing question: how do individual, moment-to-moment choices produce the static networks we observe? Their framework, called a behavioral micro-foundation for cross-sectional network models, uses a continuous time stochastic choice mechanism to bridge the gap between micro-level decision making and macro-level network structure.
The model is highly general: it can include agents who are not themselves part of the network (e.g., third-party influencers) and handle multilateral edge control (e.g., group decisions to form ties). Under equilibrium conditions, the network's probability distribution takes an exponential family form, meaning that analysts can estimate individual preferences using existing statistical tools like ERGMs. The graph potential separates into a preference term (reflecting agent utilities) and an entropic term (reflecting the rules of tie formation). The authors validate their approach by analyzing friendship networks in a professional organization and by modeling phase transitions in small-group dynamics.
- Uses continuous time stochastic choice to simulate how networks form from individual decisions
- Handles agents outside the network and multilateral edge control, expanding ERGM applicability
- Demonstrated on real-world professional friendship networks and small-group phase transitions
Why It Matters
Bridges behavioral microeconomics and network science, enabling better predictions of social dynamics and group formation.