Evolution of trust in structured populations

TL;DR

This paper analyzes how trust evolves in structured populations using a simplified trust game, revealing how different updating rules and population structures influence trust and betrayal dynamics.

Contribution

It provides an analytical framework comparing structured and well-mixed populations, highlighting the impact of various strategy updating rules on trust evolution.

Findings

Trust collaboration is highest in well-mixed populations without untrustworthy trustees.

Structured populations show reduced trust with certain updating rules, especially IM and DB.

Untrustworthy trustees destabilize trust, leading to coexistence of trustworthy and untrustworthy trustees.

Abstract

The trust game, derived from an economics experiment, has recently attracted interest in the field of evolutionary dynamics. In a recent version of the evolutionary trust game, players adopt one of three strategies: investor, trustworthy trustee, or untrustworthy trustee. Trustworthy trustees enhance and share the investment with the investor, whereas untrustworthy trustees retain the full amount, betraying the investor. Following this setup, we investigate a two-player trust game, which is analytically feasible under weak selection. We explore the evolution of trust in structured populations, factoring in four strategy updating rules: pairwise comparison (PC), birth-death (BD), imitation (IM), and death-birth (DB). Comparing structured populations with well-mixed populations, we arrive at two main conclusions. First, in the absence of untrustworthy trustees, there is a saddle point between investors and trustworthy trustees, with collaboration thriving best in well-mixed populations. The collaboration diminishes sequentially from DB to IM to PC/BD updating rules in structured populations. Second, an invasion of untrustworthy trustees makes this saddle point unstable and leads to the extinction of investors. The 3-strategy system stabilizes at an equilibrium line where the trustworthy and untrustworthy trustees coexist. The stability span of trustworthy trustees is maximally extended under the PC and BD updating rules in structured populations, while it decreases in a sequence from IM to DB updating rules, with the well-mixed population being the least favorable. This research thus adds an analytical lens to the evolution of trust in structured populations.