Cumulative advantage is a double-edge sword for cooperation

The Matthew effect emphasizes the influence of early advantage on shaping long-term development by amplifying it over time, and its implications for public cooperation are yet to be fully understood. In this letter, we propose and study a spatial public goods game driven by cumulative advantage, where players who achieve high payoffs in a given round receive greater allocations in the next. The simulation results show that the Matthew effect leads to an irreversible polarization of individual wealth on the network. Such polarization makes moderate cooperation levels more prevalent, which helps to explain the widespread coexistence of prosocial and antisocial behavior. Meanwhile, heterogeneous networks may restrict the polarization of wealth, but also inhibit the evolution of cooperation, requiring a reconsideration of the commonly held view that heterogeneous networks enhance cooperation.