What is it about?
Suppose you want to test rational expectations across some group. If the outcome is a continuous variable and the forecasts are also continuous, then you can test whether on average the difference between outcomes and forecasts is 0. Also, if you compute the square of the difference between outcomes and forecasts (or the absolute value), you can test whether this squared difference is small and whether using a linear forecast with observable control variables is a better predictor than the agents' expectations. However, you cannot do this with qualitative data (say, Excellent, Good, Satisfactory), because you cannot say the difference between Excellent and Good is a certain number such as 1 or 3, all you can say is that outcomes are ordered. This article provides a valid asymptotic rationality test for any kind of qualitative data (whether it is ordered or not).
Featured Image
Why is it important?
Most available data on the forecasts or expectations of agents about the future is qualitative. Even if the outcome is a quantitative variable, very often the forecast is qualitative in the sense that respondents may answer qualitative forecasts (ex: Higher, similar, worse). Therefore this econometrics paper has the potential to be applied to all sorts of surveys with economic and business forecasts or expectations.
Perspectives
I think this article should be interesting for applied economists that use economic forecasts. It is also interesting for economists interested in human capital and education. In the study I show how to apply the econometric tests to a panel data of parents and students. I then show that their expectations are rejected to be rational and that this irrationality or lack of forecasting knowledge implies that students study less time than they should.
Carlos Madeira
Banco Central de Chile
Read the Original
This page is a summary of: Testing the rationality of expectations of qualitative outcomes, Journal of Applied Econometrics, April 2018, Wiley,
DOI: 10.1002/jae.2625.
You can read the full text:
Resources
Article (free download)
This article provides an adequate statistic for testing the rationality of point predictions of categorical outcomes under a subjective median or mode assumption.
Slides for the Parental Expectations Workshop (University of Essex, May 2017)
Slides for the ISER conference on Parental Expectations and Achievement https://www.iser.essex.ac.uk/2017/04/03/how-do-parental-choices-affect-children-s-achievements
Codes and Data
Codes to implement the Rationality Tests and instructions to replicate the Monte Carlo exercises and the empirical analysis using a real Panel Dataset of Parents and Students' expectations.
Contributors
The following have contributed to this page
