William Zame, Peter Bossaerts, Charles Plott
Almost all asset pricing theories make predictions about portfolio choices as well as about asset prices. However, most tests of asset pricing theories focus exclusively on the price predictions -- perhaps because the portfolio predictions are "obviously" wrong. This seems a paradox: how can the price predictions be right if the portfolio predictions on which they rest are wrong?
This work offers a model that resolves this paradox, experimental evidence consistent with the theoretical model, and econometric analysis that ties the two together.