# The assumptions of Rational Expectations Models

What are the assumptions between rational expectations models and how restricted are there for the following results of economic theory? Where can I find them all gathered in some textbook or in the internet and what is the objection against REE models?

The assumptions are that the agents understand the situation, model or game they are in, that they are rational - in economic sense of the word and try to maximize their utility.

I think that one reason why you might have trouble finding a list of assumptions behind rational expectations is that rational expectations don’t really rely on their own assumptions rather they are application of rationality (with its set of assumptions) and utility maximization to uncertainty. Rationality and utility maximization under uncertainty simply give you rational expectations as rational expectations are just the mathematical/statistical expectations of any uncertain situation.

To illustrate this if you have a gamble $$G$$ where you toss a fair coin with payoffs $$T=100$$ and $$H=-200$$ the statistical expectations are given by:

$$E[G]=0.5(100)+0.5(-200)=-50$$

Now rational expectations are also -50 as if you understand the game, and want to maximize your utility the best possible expectation of the outcome of gamble is just to calculate it statistically.

You can find assumptions behind utility maximization and rationality in almost any economic textbook, but the most detailed explanations is in game theory textbooks. For example, Tadelis game theory: an introduction has clear treatment of the assumptions behind rationality.

It’s not really possible to say how restrictive rational expectations are without comparing them to some other model of expectations and considering the specifics model you want to apply them.

The critique of rational expectations is that the predictions of rational expectations don’t always hold in empirical testing. However, they are still being widely used even though we know they are not completely correct as there is no other consistent general theory of expectation formation that would provide better predictions than rational expectations even though in some isolated individual instances different expectation theories may actually provide better predictions.

• Thank you very much! Nov 13 '19 at 5:41
• One other thing to point out is that, not only does RE assume that the agent knows the model. It's also assumes that all the agents know that the other agents know the model also. In his text, ( which I highly recommend ) , Pesaran shows that, without this assumption, it's very hard-impossible to solve RE models. Nov 14 '19 at 2:15
• @markleeds you are completely right - that’s why I was talking in the plural agents but maybe I should have been more explicit about it
– 1muflon1
Nov 14 '19 at 12:04
• @1muflon1: no problem. your answer was fine. sometimes the literature is not so clear about the agent assumptions but pesaran definitely is. I really like the explanations in his text the most out of the many RE explanations. some are more confusing than helpful. all the best. Nov 15 '19 at 1:32
• Answer is grossly incorrect. Nov 16 '19 at 6:22