Content area
Full Text
Abstract
Using a simple model and state-level cross-section U.S. data from 1993 to 1999, quantile-regression estimates of price elasticity and income elasticity for cigarette demand are obtained. It is noted that price elasticity shows a sizable variation across the high and low quantity-quartiles. There is a similar variability in the income elasticity, but most of these estimates lack statistical significance. Besides providing an indication of the variation in the price (and income) elasticity for different consumption levels, the exercise suggests some interpretative caution in regard to estimates from constant-elasticity models. (JEL L13, L66, I18)
Introduction
Demand elasticities for cigarettes have been of considerable interest to scholars for a long time. These have some analytical significance and can also yield policy implications relative to taxation on cigarettes and regulation of the industry. The Centers for Disease Control and Prevention (CDC, 2000) and Chaloupka and Warner (2000) have provided extensive surveys of the relevant literature.
In their useful meta-analysis, Gallet and List (2003) considered 523 price-elasticity estimates from 86 studies. While describing the wide variation in model specification, data coverage, and estimation procedure, they noted (p. 825) that most estimates were from a double-log specification, used quantity in per-capita terms, were based on time-series observations aggregated at the country level, and used ordinary least-squares (OLS) estimation procedure.
Two broad points may be noted from the fine description by Gallet and List (2003) of the many studies they considered. First, there does not seem to be any work that systematically explored variations in the demand elasticities across different consumption levels, although there are some that considered variations across the U.S. states or periods.1 Second, very few studies seem to have obtained estimates from state-level cross-sections, despite the fairly sizable sample size (around 50) for such cross-sections.
The main purpose of this paper is to explore variations in price (and income) elasticities across high- and low-consumption contexts. This is done by using a simple double-log demand model and estimating quantile regressions from state-level cross-sections for several recent years. The main finding is that there is a sizable variation in the price elasticities across high- and low-consumption contexts, and the variability seems to accord with the basics of demand theory in terms of a general rise in (absolute)...