Bubbles in the Field

• Agricultural Economics Workshop

Date and Location

Thursday, October 7, 2021, 4:10 PM - 5:30 PM
Online Meeting, Zoom

Abstract

Dispositive evidence of bubbles in the field has proven elusive. Even if bubbles were not rare, they would be hard to find because of an identification problem. A bubble occurs when an asset’s price exceeds its fundamental value, but fundamental values are latent (to both economists and the market). Thus, bubble tests have required pivotal assumptions, such as correct specification of a structural valuation model, which are fragile and opaque. I propose another approach to detect bubbles in the field. I use the price of a close substitute as a quasi-control asset, which models the asset of interest’s path of prices for a no-bubble counterfactual. I show this strategy directly tests the no-bubble null hypothesis. I also show the strategy is a generalization of other bubble tests that have similarly palatable assumptions, but have failed to detect a bubble. As a necessary aside, I derive Cantelli’s inequality for finite samples, which I use to show the identifying assumptions need not be fragile. I apply the strategy to two independent cases in the field. In both, I find strong, robust, and economically-significant evidence of bubbles, which is backed by data supporting the identifying assumptions. The evidence I find of bubbles in the field supports the growing consensus amongst theorists and lab experimentalists that bubbles do indeed exist in asset prices.