These models’ microfoundations rely on expected utility theory (EUT) and the rational expectations hypothesis (REH). EUT is based on an axiomatic approach that represents how an individual chooses among risky portfolios given her forecasts of their returns. In traditional portfolio-balance models, EUT implies that the one-period-ahead expected excess return - the risk premium - depends on the ex ante variance of returns. The rational expectations hypothesis (REH) is used to portray individualsíforecasts of the return and variance. REH assumes that these forecasts differ from ex post outcomes by white noise errors. This assumption is based on a premise that underpins much of modern macroeconomics: the process underpinning outcomes at every point in time can be represented adequately by a single, time-invariant, conditional probability distribution.
