Tools
Notify Me of Updates
Related Links
Tests of Rational Expectations in a Stark Setting
We test the hypothesis that forecasts by participants in a stark experimental setting are the same as the rational expectation. At least for a process as simple as a random walk, relatively sophisticated as well as relatively unsophisticated participants' forecasts can be characterized as consistent with rational expectations. We find no support for the proposition that one-step-ahead forecasts are predictably wrong and substantial support for the converse proposition. The variance of forecast errors, though, is greater than for the rational expectation. Some participants have differences between two-step-ahead and one-step-ahead forecasts that are statistically significant, but the direction of bias is not systematic across individuals.