Fragility Of Asymptotic Agreement Under Bayesian Learning

September 21, 2021 in Uncategorized by

Bayes`sche LearningSignal distribution small amount asymptotic asymptotic concordance Asymptotic dissent of regular variability of small individual uncertainty of many signals conditional distribution exponential distribution fast variable tail significant amount of asymptotic belief Payment parameters Assuming that individuals know the conditional distributions of signals under the indications of payment parameters, the existing results come to the conclusion that, since individuals observe an infinite number of signals, the r beliefs about the parameters will eventually merge. We first show that these results are fragile when individuals are unsure of signal distributions: in such a model, a tiny individual uncertainty about signal distributions can lead to a considerable amount (without disappearance) of differences between asymptotic beliefs. We then characterize the conditions under which a low degree of uncertainty leads to only a small amount of asymptotic disagreements. After our characterization, this is the case when uncertainty about signal distributions is generated by a family with “rapidly varying tails” (such as the normal or exponential distribution). However, if this family has “regularly varying tails” (such as the Pareto, Log-Normal, and t-distributions), a small amount of uncertainty results in a considerable amount of asymptotic disagreements. @MISC{Acemoglu09fragilityof, author = {Daron Acemoglu and Victor Chernozhukov and Muhamet Yildiz}, title = {Fragility of Asymptotic Agreement under Bayesian Learning}, year = {2009}} If you are aware of the missing articles citing these articles, you can help us create these links by adding the relevant references in the same way as for each article above. If you are a registered author of this article, you can also check the “Citations” tab in your RePEc Author service profile, as some citations may be waiting for confirmation. Please note that fixes may take a few weeks to go through the different RePEc services. If CitEc has detected a reference but no RePEc element has been linked to it, you can use this form to help you.

If you have written this article and are not yet registered with RePEc, we advise you to do so here….