Ambiguity Aversion, Entrepreneurship and Innovation

We consider the take-up of a new technology whose bene\u85ts are only vaguely known (am-biguous). A \u85rm consists of two agents operating either the old or the new technology, and sharing output according to an ex ante Pareto e ¢ cient contract. Individuals employ the Arrow-Hurwicz criterion when making decisions in ambiguous environments, and are hetero-geneous with respect to their degree of ambiguity aversion. We show that Pareto e ¢ cient contracts resemble ordinary wage contracts. Standard equilibrium conditions therefore deter-mine the masses of \u85rms employing each technology, as well as the equilibrium wage rates. Using a dynamic version of this economy, in which agents learn about the new technology over time, we deduce properties of the di¤usion pro\u85le for the innovative technology. This involves a non-standard inference problem, since agents have a set of prior beliefs. When the new technology is superior to the old, the familiar S-shaped di¤usion pro\u85le emerges under a wide range of conditions. More surprisingly, when the new technology is inferior, a boom-bust pro le is possible. Even though bad news predominates, the ambiguity-reduction e¤ect of new information initially overcomes the bad news e¤ect, giving rise to the boom phase of the pro le