Chapter 9: Uncertainty, Expectations and Learning
1. THE ROLE OF EXPECTATIONS In the previous chapter, we have referred to the perfect foresight hypothesis. In other words, we made the hypothesis that there is a coincidence between expectations and reality. This hypothesis is undoubtedly strong and it can be useful as a benchmark. For instance, its results can be contrasted with those derived from the adaptive expectations hypothesis, which was the most widespread formula used until the years of the first oil shocks, also named the stagflation period. As has been stressed in Part I, there is a strong interdependence between facts and theory. In the 1970s, the hypothesis of adaptive expectations was considered untenable for at least two reasons. Firstly, it seemed to contrast with the facts. Secondly, it seemed to conflict with the rationality of the agent. These new questions brought about three important changes in macroeconomics. Firstly, any macro equation has to be microfounded. Secondly, expectations must be rational. Finally, supply-side policies have become predominant. In this context, the Phillips curve (see Ferri, 2000), which seems to be compatible with demand policy, is severely questioned. Not only is its shape under rigorous scrutiny, but its mere existence is under attack (see also Sargent, 1999). 2. ADAPTIVE EXPECTATIONS In the literature, expectations have mainly referred to prices and inflation. To this purpose, the scheme by Friedman and Cagan (see Sargent, 1999, p.51) is particularly revealing: xt 2 xt21 5 (1 2 l) [ yt21 2 xt21 ] , (9.1) where l lies between 0 and 1. If one...
You are not authenticated to view the full text of this chapter or article.