The question of how a pure fiat currency is enforced and comes to have a non-zero value has been much debated (Selgin, 1994). What is less often addressed is, in the case where the enforcement is taken for granted and we ask what value (in terms of goods and services) the currency will end up taking. Establishing a decentralised mechanism for price formation has proven a challenge for economists: "Since no decentralized out-of-equilibrium adjustment mechanism has been discovered, we currently have no acceptable dynamical model of the Walrasian system." (Gintis, 2006) In his paper, Gintis put forward a model for price discovery based on the evolution of the model's agents, i.e. \poorly performing agents dying and being replaced by copies of the well performing agents." It seems improbable that this mechanism is the driving force behind price discovery in the real world. This paper proposes a more realistic mechanism and presents results from a corresponding agent based model.