K. Rebecca Scott
Rational Habits in Gasoline Demand (Energy Economics; http://dx.doi.org/10.1016/j.eneco.2012.02.007)
Abstract: Understanding the dynamics of demand for gasoline---the combustion of which is responsible for 20% of U.S. CO2 emissions---is essential to deciding how to rein in consumption. Rational habits have not been broached in the gasoline demand literature, and yet they may be a key shaper of demand: if consumers are forward-looking, then habits will render their demand sensitive to price dynamics. This sensitivity implies that the usual measures of price elasticity will underproject consumers' response to policy interventions.
This paper examines the implications of rational habits on gasoline demand. Using a simple model encompassing myopic and rational habits, I demonstrate that an agent with rational habits will respond to anticipated future price changes and react more strongly to permanent than to temporary price changes, with this distinction increasing in the strength of the habit. I then estimate several habits models using panel data on U.S. states for the years 1989-2003. These estimates provide compelling evidence of habits for gasoline consumption and some evidence for rational habits.
Working paper (revised 23 Nov. 2011)
Abstract: The combination of habits and a forward outlook suggests that consumers will be sensitive not just to prices but to price dynamics. In particular, rational habits models suggest 1. that price volatility and uncertainty will reduce demand for a habit-forming good and 2. that such volatility will dampen demand's responsiveness to price. These two implications can be tested by augmenting a traditional partial-adjustment or error-correction model of demand. I apply this augmented model to data on gasoline consumption, as rational habits provide a succinct representation for the investment and behavioral decisions that determine gasoline usage. The trade-offs among 2SLS, system GMM, and pooled mean group (PMG) estimators are considered, and my preferred PMG estimator provides evidence for the two implications of rational habits in a panel of 29 countries for the years 1990-2009. The sensitivity of certain results to the choice of estimator offers a cautionary illustration of the cost of assumptions such as coefficient heterogeneity. Given the evidence uncovered in favor of rational gasoline habits, such habits may help to explain some of the cross-country variation in "total" price elasticity. These habits also imply that the effect of price volatility must be taken into account when projecting the impacts of potential policies on gasoline consumption.
Previous version (2011)
Abstract: When consumers are forward-looking with respect to their demand for a habit-forming good, traditional measures of price elasticity are misleading. In particular, such measures will underestimate sensitivity to long-run price shifts---and therefore underestimate the potential effect of policy instruments that act through price. Correcting elasticities for the behaviour of the price process requires a model with forward-looking consumers, a habit-forming good, and uncertain relative prices. With appropriate restrictions on the type of price uncertainty, this paper shows that it is possible to solve for the optimal consumption path under any price process. Simulations then sketch out how habits and the price process shape demand. Petrol demand motivates the model and illustrates its implications.