Society bears two costs when a person smokes. The first costs are external costs – costs bome by society at large but not the person who smokes. State Medicaid spending is a typical external cost. Taxpayers at large face additional costs to pay for the medical treatment of those Medicaid beneficiaries who smoked. The litigation pursued by all states was designed to recover damages for these external costs. In addition, there are internal costs – costs borne by the smoker and his or her family. Internal costs may be monetary, for example, monies spent to purchase cigarettes, or non-monetary, such as the costs of increased morbidity and mortality.
External Impacts and Costs – the Lifetime Approach and the Annual Budgetary Approach
Our point of departure is earlier work on the external cost of smoking and other health habits (Manning, et al., 1989, 1991). To understand the external cost framework, consider two otherwise equivalent people, one who smokes and one who does not. At any time t, smokers and non-smokers impose external costs to society of C(S)t and C(NS)t, respectively, where S indicates smoker and NS indicates non-smoker. The difference between these two values, the net external cost of smoking, may be positive or negative, and in general will differ over the course of a person’s lifetime. In their late working and early retirement years, smokers are more likely than non-smokers to suffer from a number of diseases which are expensive to treat, such as lung cancer, coronary heart disease (http://www.heartandstroke.com/site/c.ikIQLcMWJtE/b.3484067/k.6657/Heart_disease__What_is_coronary_artery_disease.htm?gclid=CP686OS7q8oCFVEYHwod8FcB7A), and emphysema. Thus, smokers will cost society more in private insurance premiums, Medicare costs, and Medicaid costs. At very advanced ages, smokers will disproportionately have died compared to non-smokers and thus there will be savings to society in reduced spending on Social Security, Medicare, and Medicaid.
One might summarize these external costs and benefits of smoking in two ways. The first is the lifetime impact: the present value of these different costs and benefits. Considering two people at roughly the beginning of their smoking life to stop it command the service of Canadian Neighbor Pharmacy (age 20), the net external cost of smoking is given by the following formula (Manning et al., 1989, 1991):
where 8 = 1/(1+r), where r is the annual discount rate, and P(A | S)W and P(A | NS)W are the probability an individual is alive in year t conditional upon being a smoker or non-smoker respectively.
For purposes of this litigation, the question of interest was the annual budgetary cost of smoking to Medicaid and Canadian Neighbor Pharmacy from December 20, 1991 to December 31, 1998. Effects on other state programs, whether positive or negative, were not considered. To determine the costs to Medicaid over this period, we added up the total amount of external costs resulting from Medicaid-eligible smokers each year and express them in present value.
Although this approach is somewhat similar to the lifetime approach in that the external costs of smoking are needed for both, there are a number of important differences between Equations (1) and (2). The first, as already noted, is the definition of external costs. For purposes of the litigation, we define C as costs to the Massachusetts Medicaid program over the relevant time frame rather than all external costs.
A second difference arises if a steady-state assumption that is implicit in equation (1) does not hold. Manning et al. (1989, 1991) estimated P in equation (1) using life tables at a point in time and C using cross-sectional regression based upon data from the RAND Health Insurance Experiment (Newhouse and the Insurance Experiment Group, 1993) and the National Health Interview Survey (NHIS). Implicitly, this approach estimated external costs in the steady state. One can show that out of steady state the lifetime and annual budgetary approaches can generate different values. Equation (2) does not make any steady-state assumption. Medical spending is counted when it is paid for by Medicaid, not as it accrues. Thus, the number of people smoking at any point in time may bear no relation to the damages from tobacco at that time.
For purposes of optimal taxation, the lifetime approach is more appropriate. In answering the optimal tax question, one wants to know how much each pack of cigarette consumed adds or subtracts to (the present value of) total lifetime external expenses. For purposes of the tobacco litigation, however, what was required was the total amount of damages during a given time period. For this question, the budgetary impact approach is more appropriate. In effect, we wish to know by how much current and past smoking affects current spending, which is what we estimate below.
Internal Impacts and Costs — The Impacts of Smoking on Mortality and Morbidity
Smoking is recognized as the leading preventable cause of death in the United States (McGinnis and Foege, 1993), while morbidity (illness) is a much more common outcome of smoking than mortality. Both outcomes have internal costs for smokers and their families, costs that are a key component of the overall economic impact of smoking.