We study how two fiat monies, one safe and one risky, compete in a decentralized trading environment. The equilibrium value of the two currencies, their transaction velocities and agents' spending patterns are endogenously determined. We derive conditions under which agents holding diversified currency portfolios spend the safe currency first and hold the risky one for later purchases. We also examine when the reverse spending pattern is optimal. Traders generally favor dealing in the safe currency, unless trade frictions and the currency risk is low. As risk increases or trading becomes more difficult, the transaction velocity and value of the safe money increases.Key Words: money, currency substitution, search JEL Codes: E4, E5, D7
IntroductionCenturies ago the comedy writer Aristophanes lamented (in "The Frogs") that "the full-bodied coins that are the pride of Athens are never used while the mean brass coins pass hand to hand." Many observers have since debated on the organization of exchange when several monies, some 'superior' to others in some way, compete to sustain trade.A long-held notion is that an inferior currency should circulate more widely than a superior money. Those holding both monies would prefer spending the 'bad' money as soon as they can, and keep the 'good' money for future purchases. Others have favored a differing notion: it is the good currency that should circulate more widely. Hayek (1976) argued that this was the logical outcome of currency competition. People would rather spend the good money first, as it has greater purchasing power, and keep the bad money to face future trade opportunities. 1 These notions are conflicting, yet revolve around rational spending behavior. Thus, a key challenge is to determine what fundamental factors influence the use of competing monies. That is: if two currencies are accepted in trade, when will agents tend to spend the bad and hold the good one for future purchases? When will they do the opposite? A large theoretical literature has offered insight centered around arbitrary transaction costs or institutional restrictions on use of monies (see Giovannini and Turtleboom, 1994). We complement it by studying currency use as a result of decentralized and uncoordinated private decisions, absent currency-specific transaction costs and institutional restrictions.To do so we consider an economic environment in which money is essential to conduct decentralized trade. There are two intrinsically different monies: a 'bad' money characterized by purchasing power risk and a 'good' safe money. Both have explicit medium-of-exchange roles, and their equilibrium values reflect their ability to facilitate spot trades of consumption goods. This is formalized by modeling trade as a random search process among agents specialized in production/consumption. They hold currency portfolios to buy goods via pairwise trades where prices are determined via bilateral bargaining. In this context currencies compete on a 'level' trading field as currency-specific trade barriers,...