Cigarette Taxation and Demand: An Empirical Model

Article excerpt

I. INTRODUCTION

California's excise tax on cigarettes increased from 10 cents to 35 cents per pack beginning in January 1989. This paper evaluates the impact of this tax increase on cigarette consumption in California using a dynamic model that incorporates consumers' addictive behavior.

Federal, state and local governments historically have used taxation to raise revenues. Since the release of the First Surgeon General's Report on smoking and health (U.S. Department of Health, Education, and Welfare, 1964), taxation also has become an economic tool to reduce cigarette smoking. Besides taxation, governments have implemented other policies, such as smoking regulation, to control cigarette smoking. The analysis here evaluates the impact of taxation on cigarette consumption while controlling for the effects of anti-smoking regulations.

The effectiveness of a state's excise tax on cigarette consumption depends on other states' tax rates. Over the past two decades, disparities among state cigarette taxes have grown wider, and incentives for interstate cigarette bootlegging have increased (bootlegging occurs when consumers residing in high-price states buy cigarettes in low-price states). Including a bootlegging incentive variable controls the potential problem that bootlegging poses for the analysis.

Section II specifies a structural model for cigarette demand and price equations. Section III describes the data and variables. Section IV presents the equations and estimation methods. Section V contains the empirical results. Section VI offers policy implications and conclusions.

II. MODEL SPECIFICATION

The traditional approach to measuring the impact of taxation on cigarette consumption involves using a static demand model. The analysis here uses a dynamic demand model of cigarette consumption that incorporates consumers' rational addictive behavior (Becker and Murphy, 1988; Becker et al., 1994). A rational addiction demand model is given by:

(1) [D.sub.t] = [b.sub.0] + [b.sub.1][P.sub.t] + [b.sub.2][D.sub.t-1] + [b.sub.3][D.sub.t+1] + [b.sub.4][Z.sub.t] + [[Epsilon].sub.t]

where [D.sub.t] is the quantity demanded at current period t, [P.sub.t] is the price, [D.sub.t+1] is the past quantity demanded at time t-1, [D.sub.t+1] is the future quantity demanded at time t+1, [Z.sub.t] is the vector for other determinants, and [[Epsilon].sub.t], is a disturbance term. The parameter [b.sub.1] is negative.

Equation (1) states that the demand for current consumption depends on past consumption and future consumption. Becker et al. (1994) prove that [b.sub.2] [is greater than] 0, [b.sub.3] [is greater than] 0, and [b.sub.3] equals [b.sub.2] multiplied by a discount factor. That is, the more cigarettes consumed in the previous period or to be consumed in the future period, the more would be consumed in the current period.

In contrast to the rational addiction model, the habit-persistence model used by McGuinness and Cowling (1975), Fuji (1980), and Baltagi and Levin (1986) accounts for the dependence of current consumption on past addictive consumption but ignores the future effects of addiction. The habit-persistence model differs from equation (1) in that it excludes the future consumption term. Becker et al. (1994) refer to the habit-persistence model as a "myopic addiction model." Whether consumers are addicted or not, and whether they are rationally or myopically addicted, can be empirically tested by the coefficients for [D.sub.t-1] and [D.sub.t+1]. The empirical results of Becker et al. (1994), and Chaloupka (1991) support the rational addiction model.

To evaluate the responsiveness of consumption to the change in price resulting from taxation, most previous studies use a single-equation approach and focus only on the demand equation. Such an approach assumes that the cigarette industry is perfectly competitive and that supply is perfectly elastic, so that any change in excise tax would be fully reflected in the cigarette price. …