This study considers model uncertainty about tail risk, the possibility of learning and
how this affects choice. I designed an experiment involving repeated risk taking where assets can yield steady streams of good outcomes but eventually inflict a major loss that annihilates all previous gains. Learning about those assets' risk/reward profiles is crucial yet challenging. The main findings are: 1) When asked to perform a stylized version of the task, participants managed to learn in a Bayesian way; 2) However, many still chose to invest in these assets, apparently owing to an overwhelming desire to "pick pennies." These findings suggest that the key issue with tail risk is not the most commonly expected one, namely, that people cannot assess it, but rather that people cannot deal with it properly due to purely behavioral issues related to limited self-control.
%0 Unpublished Work
%1 citeulike:14014151
%A Payzan-LeNestour, Elise
%D 2016
%K 91g50-corporate-finance 91g40-credit-risk 91-01-game-theory-economics-social-behaviour-science-instructional-exposition 91e40-memory-and-learning 62f15-bayesian-inference 62c12-empirical-decision-procedures-empirical-bayesian-procedures
%T Fooled by Randomness? Financial Decision-Making under Model Uncertainty
%X This study considers model uncertainty about tail risk, the possibility of learning and
how this affects choice. I designed an experiment involving repeated risk taking where assets can yield steady streams of good outcomes but eventually inflict a major loss that annihilates all previous gains. Learning about those assets' risk/reward profiles is crucial yet challenging. The main findings are: 1) When asked to perform a stylized version of the task, participants managed to learn in a Bayesian way; 2) However, many still chose to invest in these assets, apparently owing to an overwhelming desire to "pick pennies." These findings suggest that the key issue with tail risk is not the most commonly expected one, namely, that people cannot assess it, but rather that people cannot deal with it properly due to purely behavioral issues related to limited self-control.
@unpublished{citeulike:14014151,
abstract = {{This study considers model uncertainty about tail risk, the possibility of learning and
how this affects choice. I designed an experiment involving repeated risk taking where assets can yield steady streams of good outcomes but eventually inflict a major loss that annihilates all previous gains. Learning about those assets' risk/reward profiles is crucial yet challenging. The main findings are: 1) When asked to perform a stylized version of the task, participants managed to learn in a Bayesian way; 2) However, many still chose to invest in these assets, apparently owing to an overwhelming desire to "pick pennies." These findings suggest that the key issue with tail risk is not the most commonly expected one, namely, that people cannot assess it, but rather that people cannot deal with it properly due to purely behavioral issues related to limited self-control.}},
added-at = {2017-06-29T07:13:07.000+0200},
author = {Payzan-LeNestour, Elise},
biburl = {https://www.bibsonomy.org/bibtex/22b7703c89ad5ff23c4db7f6f91686123/gdmcbain},
citeulike-article-id = {14014151},
citeulike-attachment-1 = {payzan-lenestour_16_fooled.pdf; /pdf/user/gdmcbain/article/14014151/1064267/payzan-lenestour_16_fooled.pdf; 66c6f668178112c5e41b735d35b439d258041636},
comment = {circulated by jackiedent 2016-04-17 before TEDx Sydney 2016},
day = 22,
file = {payzan-lenestour_16_fooled.pdf},
interhash = {7ac3f40a8b462d65a6eeabfe29cd5b34},
intrahash = {2b7703c89ad5ff23c4db7f6f91686123},
keywords = {91g50-corporate-finance 91g40-credit-risk 91-01-game-theory-economics-social-behaviour-science-instructional-exposition 91e40-memory-and-learning 62f15-bayesian-inference 62c12-empirical-decision-procedures-empirical-bayesian-procedures},
month = mar,
posted-at = {2016-04-17 10:42:12},
priority = {2},
timestamp = {2022-06-06T08:43:59.000+0200},
title = {{Fooled by Randomness? Financial Decision-Making under Model Uncertainty}},
year = 2016
}