Pedro L. Ferreira

What humans do when exposed to uncertainty, incomplete information, and a dynamic environment influenced by other agents remains an open scientific challenge with important implications in both science and engineering applications. In these contexts, humans handle social situations by employing elaborate cognitive mechanisms such as theory of mind and risk sensitivity. Here we resort to a novel theoretical model, showing that both mechanisms leverage coordinated behaviors among self-regarding individuals. Particularly, we resort to cumulative prospect theory and level-k recursions to show how biases towards optimism and the capacity of planning ahead significantly increase coordinated, cooperative action. These results suggest that the reason why humans are good at coordination may stem from the fact that we are cognitively biased to do so.

The cognitive mechanics of human decision-making is affected by a large set of high-level processes. Some of these processes, called cognitive biases, are often regarded as failures, since they prescribe behavior which is not deemed as rational. Furthermore, in social settings, humans employ a process known as theory of mind which enables them to create and manage a dynamic model of the mental states of others, allowing for the prediction of future actions to better inform current behavior.

Can cognitive biases promote coordination? Can increasingly sophisticated levels of theory of mind promote coordination? In this thesis, we answer these questions by showing how coordination among agents measuring value using the prescriptive Expected Utility Theory (EUT) differs from the coordination among agents measuring value using the descriptive Cumulative Prospect Theory (CPT), in two experimental settings: a normal-form stag hunt game allows us to study how coordination differs when agents use EUT and CPT as theories of value, while a Markov game of stag hunt focuses on studying the effects of increasingly sophisticated policies among both EUT- and CPT-agents, using the recursive theory of mind level-k model that captures bounded rationality.

We show that CPT-agents are better able to coordinate in both experiments, compared to EUT-agents. Furthermore, in the Markov stag hunt, while coordination with both EUT and CPT stand to gain from increasingly sophisticated policies, CPT-agents do not require as much sophistication as EUT-agents do to coordinate to the same extent. We can thus conclude that, while some of these cognitive biases are viewed as failures in individual decision-making, they actually make social interaction easier.