69th International Atlantic Economic Conference

March 24 - 27, 2010 | Prague, Czech Republic

Time Discounting and Smoking Behavior Under Tax Hikes

Saturday, 27 March 2010: 17:05
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
By analyzing our broad survey of Japanese adults from 2005 to 2008, the purpose of this article is to examine empirically the effects of time discounting and its behavioral biases on smoking behavior, and their responses to a hike in cigarette taxes. We incorporate two time discounting biases that has been reported in dynamic choice theory: (i) hyperbolic discounting, under which a person discounts the immediate future more intensively than the distant future, and thereby makes time-inconsistent decisions; and (ii) the sign effect, by which a person discounts positive future payoffs more intensively than negative future payoffs. We hypothesize that hyperbolic discounters smoke more than exponential discounters; that reductions in cigarette consumption after a tax hike amongst naïve hyperbolic smokers is larger than those amongst the others; and that decreases in cigarette consumption after a tax hike amongst hyperbolic smokers are larger than those amongst exponential smokers.
     The novelty of the present research is in the following two points. First, we identify respondents’ types by whether discounting is exponential or hyperbolic and by whether decision making is naïve or sophisticated, and estimate differences in smoking behavior among the distinct types of respondents. To do so, we construct the data for time discounting, hyperbolic discounting or procrastination, and the sign effect from response data to questions on various hypothetical intertemporal decisions including choices between a small amount of present money and a larger amount of future money. In addition, naïve and sophisticated respondents were identified by using response data to two questions regarding procrastination in doing homework assignments: one asked respondents when they used to plan to do homework in summer vacations in their school days; and the other asked them when they in effect used to do homework. Respondents were classified as naïve if they used to procrastinate doing homework more than they planned whereas they were classified as sophisticated otherwise. Second, by using the actual time series data of the cigarette tax, which is exogenously determined by the Japan’s government, we quantify how the natural experiment of cigarette tax hikes in 2006 affects smoking behavior of each type of respondents. To do so, we use our four-year panel survey data, in which roughly three to five thousands of Japanese adults reply to questionnaires on preferences and various attributes each year.
     The main results of this article are as follows: firstly, Cigarette consumption displays significantly positive correlation with discount rates and procrastination, and negative correlation with the sign effect; secondly, hyperbolic, procrastinating, and naïve respondents decrease their after-tax-hike cigarette consumption more than the others, implying that, irrespective of the preannouncement of a future tax hike, they postpone smoking moderation until the tax hike actually takes place; finally, the government’s revenue from cigarette tax peaks at a JPY 29.92 (around USD 0.33 using the conversion rate [90.29] in October 2009) higher tax per cigarette than the present actual level.