The Cushioning Benefits of Biased Beliefs

Abstract

We present a model of dynamic investment and production in which producers may have biased beliefs in which they overextrapolate recent demand conditions into the future. This bias leads producers' beliefs to exhibit insufficient mean reversion, as these producers underestimate the degree of mean reversion in the demand process. In a volatile industry, while biased beliefs lead firms to make sub-optimal investment decisions in the short-run, they can be beneficial in the long-run by counteracting the general trend in the industry, "cushioning" the industry against prolonged downturns and aiding faster recovery. As an empirical case study, we consider oil exploration in Alaska. We present evidence that firms in this industry were subject to extrapolation bias, leading to drilling of lower-profit wells after recent price increases. Calibration of our model to Alaska oil exploration shows that the cushioning effect can be large: in a typical episode of oil price decline arising from a sequence of adverse demand shocks, the cushioning effect reduces the decline of the oil price by 8.2% and accelerates the price recovery by 27%. This showcases the potential positive implications that biased beliefs can have on industry dynamics.

Keywords: cushioning, extrapolation, oil exploration

JEL Classification: G02, G12, G30