73rd International Atlantic Economic Conference

March 28 - 31, 2012 | Istanbul, Turkey

Time discounting biases and smoking decisions

Saturday, 31 March 2012: 5:05 PM
Myong-Il Kang, Ph.D , Department of Business Administration, Korea University, Kodaira, Tokyo, Japan
Shinsuke Ikeda, Ph.D. , The Institute of Social and Economic Research, Osaka University, Ibaraki, Osaka, Japan
The purpose of this article is to empirically examine whether smoking decisions are related to time discounting and its two biases: (i) hyperbolic discounting, in which the discount rate to payoff in the immediate future is higher than the discount rate in the distant future; and (ii) the sign effect, by which a person discounts positive future payoffs more intensively than negative future payoffs.

We hypothesize, that (i) hyperbolic discounters smoke more than exponential discounters; that (ii) people with the sign effect smoke less than others; and that (iii) naïve hyperbolic discounters, who are unaware of their hyperbolic discounting, do not take into account their time-inconsistency problem in smoking behavior and therefore smoke more than the sophisticated.

For testing our hypotheses, we use the broad survey of Japanese adults and distinguish respondents’ types of discounting, such as the degree of impatience, hyperbolic discounting and the incidence of the sign effect, from questions on various hypothetical intertemporal choices between a small amount of present money and larger amount of future money. We also distinguish between naïve hyperbolic discounters and sophisticated ones by using two questions regarding procrastination in doing homework assignments: one asks respondents when they used to plan to do homework in summer vacations in their school days; and the other asks them when they in effect used to do homework. To control for other factors affecting cigarette consumption, a variety of independent variables are constructed, such as the degree of risk aversion, the gender, the age of respondents, the level of income and the education.

We estimate a two-part model of smoking behavior based on the model developed by Cragg(1971), which uses probit methods to estimate a smoking participation equation in the first step and least squares methods to estimate cigarette consumption by smokers in the second step.

The main results of this article are as follows: firstly, the degree of impatience affects smoking participation for males and cigarette consumption for females; secondly, the procrastination tendency of homework affects smoking participation and cigarette consumption for males and smoking participation for females; thirdly, naïve hyperbolic discounters smoke more than the sophisticated in female samples; lastly, the incidence of the sign effect decreases smoking amounts of male respondents.