Based on 2008 financial crisis, the least credit worthy received the lowest cost of capital (banks). This demonstrates that "credit worthiness" is political, and a tool used to charge consumers, but is not applied consistently to politically connected organizations and in fact, it demonstrates a flaw in the current financial system that is designed to reward bankers regardless of risk management, as long as politics is the true currency, not monetary currency

One approach is a longitudinal study of subjects, determining stated commitment and then measuring actual actions at periodic intervals thereafter. That can be used to suggest commitment in other matters and hence a measure of credibility.

Let me expand the question: intertemporal commitment and sticking to it.

Mainstream behavioral economics is still using Fisher-type (1930s, can you believe it?) utility discounting model, for instance hyperbolic discounting. I have simple counterexamples showing such model is weak, and does not suit very important use cases.

What should be the approach for accumulating measurements of past behavior, and forecast of deviation risks in the future from commitment?

Agent based models with Belief Desire Intent (BDI in literature) refers to utility, ultimately desire, desirability, commitment. It has developed in the field of software agents, maybe less for physical agents, and even less for economic agents.

Should we revisit the "desire/preference" framework, with commitment/capability/ability with added human (or machine/organisation?) deviation in behavior (e.g. Not meeting a commitment, failing to deliver or adapting the firm of delivery), evolution etc...

Comments welcome