Clustering Drives Cooperation on Reputation Networks, All Else Fixed

TL;DR

This paper shows that in fixed-size, regular social networks, the clustering coefficient significantly influences cooperation levels, providing a clear mechanism linking network structure to social trust and cooperation.

Contribution

It demonstrates that clustering coefficient, rather than other network properties, drives cooperation in fixed-size, degree-regular social networks, making the mechanism accessible to non-network specialists.

Findings

Clustering coefficient correlates with cooperation levels.

Network size and degree are held constant in the analysis.

Clustering influences cooperation independently of other network properties.

Abstract

Reputation-based cooperation on social networks offers a causal mechanism between graph properties and social trust. Recent papers on the `structural microfoundations` of the society used this insight to show how demographic processes, such as falling fertility, urbanisation, and migration, can alter the logic of human societies. This paper demonstrates the underlying mechanism in a way that is accessible to scientists not specialising in networks. Additionally, the paper shows that, when the size and degree of the network is fixed (i.e., all graphs have the same number of agents, who all have the same number of connections), it is the clustering coefficient that drives differences in how cooperative social networks are.