The rise of money : an evolutionary analysis of the origins of money

Abstract

This paper shows that if there are more goods than money in all trading periods then nonconvertible fiat money is evolutionarily successful in complex economies. This result is developed in Kiyotaki and Wright (1993) search model of money, using a learning algorithm developed by Marimon et. al (1990). When we include an evolutionary model similar to Kandori, Mailath, and Rob (1993) we find that fiat money is frequently evolutionarily successful. To be precise let x be the probability someone finds someone who has a good they want –small x represents a complex economy– and let µ be the fraction of people holding money in any trading period. As long as µ < (1-2x)/(2-2x) fiat money will be evolutionarily successful.