Testing Expectations Hypotheses Flashcards
what is rational expectations theory?
the optimal future forecast that uses all available information
- no uncertainty
- you know the distribution of the outcome
rational expectations hypothesis implies EMH - forecast errors are not predictable - prices behave randomly - the only thing that changes prices is some new news
what is the Coibon and Gorodnichenko 2015 test against the rational expectations hypothesis
regressing forecast error on your forecast revision
rational expectations : B1 and B0 = 0
- there is no relationship between FR and FE –> you can not improve your forecast
what is the Coibon and Gorodnichenko 2015 test against the rational expectations hypothesis
RESULTS
what does B1 < 0 imply
if you hear good news and think prices will go up you revise your FR upwards
B1 is not equal to 0 = can use FR to improve future forecasts –> violates REH
B1 < 0 = forecasters over react to news
- when the increase their FR - the forecast is negative
B1 > 0 = forecasters under react to news
- has not made sufficient adjustment - when they increase FR - the forecast error increases
what does revising belief mean
how people change their minds after they hear new news/information
what is bayes rule used for = rational expectations
It provides a way to update or revise the probability of an event occurring based on new evidence or information
- conditional probabilities = likelihood of event when some info has already been revised = the more info you have the more info you eliminate
what are the 2 alternative theories to rational expectations hypothesis (since they have been widely rejected)
- rational inattention = predicts under reaction to news
- hypothesis of diagnostic expectations = predicts over reaction to news
what is rational expectations
individuals make predictions about the future conditions based on all available information
- perfectly use info when they make decisions
- bayes rule = how beliefs change when new information is added
what is rational inattention
- too costly (money and time) to collect and process al information - so inattentive to new information
- still use Bayes but ignore pieces of info
- use system 1 = quick thinking
- will UNDER PREDICT + Under react
what is diagnostic expectations
- if you get info that increases the probability - you think this is representative
- under diagnostic expectations a forecaster overweighs the more representative state = more weight that under rational expectations
- so over predicts - over reacts to new information
