Prototyping the macroeconomic impact of cryptomoney with a small agent-based model

Cryptomoney (Bitcoin, Ethereum ...) are uncontrolled by central authorities, have no monetary or interest rate policies attached to them. While their mass is not significant, they provide enormous potential as speculative underlier. Agents dealing with them could be a vast, unchecked public, and their expectations and optimization on future values may be very far from rationality. The objective of this paper is to offer a framework to simulate the possible outcome of interacting independent, heterogenous classes of consumers, banks, governments, goods manufacturers, with an autonomous generation of cryptomoney. Starting from a classic IS-monetary policy and bank credit framework, a flow model is derived, and independent agents are created. They have independent optimization processes and visibility scopes. They act in uncertainty and with partially guessed variables. The model is driven through a 40-50 time frames with a DSGE like logic to explore transitory regimes. Excel VBA is used for the core prototype, then the model is finally rewritten in Python - speculative bubbles may form around the cryptocurrency used. The model zooms on what other parameters in the behaviour equations (e.g psychological discount rate, propensity of consumption, preference for liquidity ...) may dampen or heighten speculations. - speculators are endowed with a specific reinforced learning mechanism, TD-Learning. The model studies the optimum time intervals for enhancing or forgetting the learning effect in speculators.