www.elsevier.com/locate/ypmed

Preventive Medicine 4

Increasing excise taxes on cigarettes in California: a dynamic simulation

of health and economic impacts

Sajjad Ahmad, Dr.*

Department of Planning, Policy and Design, University of California, HPRG 202 SE 1, Irvine, CA 92697-7075, USA

Available online 7 January 2005

Abstract

Background. California raised cigarette excise taxes in 1999, and may generate additional health and economic benefits by raising them

further.

Methods. A dynamic computer simulation model follows births, deaths, migration, aging, and changes in smoking status for the entire

population of California over 75 years to estimate the cumulative health and economic outcomes of these changes under several excise tax

rate conditions (up to 100% price increase).

Results. A 20% tax-induced cigarette price increase would reduce smoking prevalence from 17% to 11.6% with large gains in cumulative

life years (14 million) and QALY’s (16 million) over 75 years. Total spending on cigarettes by consumers would increase by $270 million in

that span (all going to tax revenue), and those who reduce the number of years spent as a smoker would spend $12.5 billion less on cigarettes.

Total smoking-related medical costs would drop by $188 billion. These benefits increase greatly with larger tax increases, with which tax

revenues continue to rise even as smoking prevalence falls.

Conclusions. Even considering benefits from the 1999 increase, California has not yet maximized the potential of excise taxes to lessen

the negative impacts of smoking. Additional tax increases would provide added health benefits and revenue to the state.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Tobacco policy; Price elasticity; Economic impacts; Health impacts; Cigarette excise taxes in California

Introduction

Despite the highly addictive nature of cigarettes, there is

widespread evidence that increasing their price will encour-

age smokers to decrease or eliminate their consumption and

will discourage non-smokers from taking up the habit to

begin with [1–3]. This is especially true during adolescence,

when around 90% of smokers begin smoking [4] and when

demand for cigarettes is more sensitive to price than during

adulthood [2]. For these reasons, increasing cigarette excise

taxes has tremendous potential to improve public health

outcomes by reducing smoking prevalence while, unlike

many health promotion interventions, producing a net gain

in government revenues.

0091-7435/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.ypmed.2004.10.024

* Fax: +1 949 824 8286.

E-mail address: sajjad@uci.edu.

A $0.50 per pack tax increase (raising per-pack prices

around 20%) introduced in California in 1999 with Propo-

sition 10was estimated to save over 8300 quality adjusted life

years and to generate around $700 million in government

revenues in the first year alone [5]. Because the greatest

impact of the tax hike would be felt by adolescents, the people

most likely to initiate smoking, these benefits would increase

every year until 75 years past the tax increase.

Even though the 1999 tax increase is already yielding

benefits for Californians, the state may not have tapped the

full potential of excise taxes to influence tobacco-related

outcomes. Since 2002, 35 states and the District of

Columbia have increased excise taxes (some more than

once), leaving California’s rate of 87 cents per pack as only

the 19th highest in the country-well below the national high

of $2.05 [6]. California has room to raise excise taxes again

to increase the health and economic benefits the state has

enjoyed since 1999, and policy-makers and the health

1 (2005) 276–283

S. Ahmad / Preventive Medicine 41 (2005) 276–283 277

promotion community should have an estimate of the

magnitude of these potential advantages.

This paper presents the results of a dynamic simulation

model exploring the accumulation of added benefits for

Californians over the next 75 years if excise taxes are

raised again today. First, we describe this model and the

data it employs. Then, follows a description of our

calculation of price elasticity, the calibration of the model,

and our method for estimating the outcomes of interest.

Next, we present the key findings from the model in terms

of the cumulative economic and health consequences of

various levels of cigarette price increase for California’s

population from 2003 to 2078. Finally, we discuss the

significance of the findings and some limitations to our

model.

Methods

Model overview

To predict the impacts of tobacco control interventions

on key economic and health indicators, we developed a

dynamic computer simulation model that can be easily

modified to evaluate virtually any policy that affects

smoking behaviors [7–9]. We developed the model using

Vensim [10], an object-orienting modeling environment.

The model uses a state-transition approach with a time step

of 1 year to estimate births, deaths, aging, net migration, and

smoking status distributions for the population over the next

75 years. Ultimately, the simulation calculates the antici-

pated cumulative long-term economic and public health

gains or losses assuming different levels of cigarette price

increase through taxes.

The model schematic is given in Fig. 1. The population

is divided into three smoking status categories based on

their smoking behavior participation—never smokers, who

have smoked less than 100 cigarettes in their lifetime,

current smokers, who have smoked 100 or more cigarettes

in lifetime and have smoked in the past 30 days, and former

smokers, who have smoked 100 or more cigarettes in

lifetime, but none in the past 30 days. The initial number of

Californians by age and gender come from US Census [11].

The initial smoking status breakdown by age and gender is

drawn from the Behavioral Risk Factor Surveillance

Fig. 1. Model s

System [12] and the Teenage Attitudes and Practices

Survey [13], and current levels of cigarette consumption

for California come from California Department of Health

Services [14].

The simulation model begins with data on the starting

state of the population (population count and smoking

status) by age and gender. Individuals may enter the

simulated population, leave it, or change smoking status

within the model according to various change probabilities

(fertility, mortality, and net migration, rates of smoking

initiation, cessation and relapse, and price elasticity of

smoking participation) by age group, smoking status, and

gender. As individuals move through the model, their ages

increase and their subsequent change probabilities are

adjusted accordingly.

Change probabilities vary over time as a function of age

and gender, and sometimes, smoking status. Fertility [15] is

the probability of a live birth as a function of a woman’s

age. The number of live births for each year is computed by

multiplying the number of women in each age cohort by the

corresponding age-specific fertility rate. For each live birth

in the simulation, a new never-smoker of random gender

and age zero enters the population, and proceeds through the

subsequent years of the simulation. Mortality hazard is the

probability of death at a given time, and was estimated as a

function of age, gender [16], and smoking status [17]

assuming a Weibull distribution. The number of deaths each

year is computed by multiplying the number of people of

each combination of age, gender, and smoking status by that

group’s respective mortality hazard, and taking the sum

across all groups for each year. Net migration (immigration

minus emigration) rates reflect changes due to people

moving in an out of the population for each gender and

age group and is extracted by age and gender from

California Department of Finance data [18].

Age- and gender-specific initiation, cessation, and

relapse probabilities are derived from multiple sources.

In each case, we estimated change probabilities from

survey data by comparing smoking status for 2 consec-

utive years, using regression methods to fit separate hazard

functions by age, gender and interaction terms. Initiation

probability was estimated in this way from the Tobacco

Use Supplement of the Current Population Survey data.

Cessation and relapse probabilities were derived in a

similar fashion from National Health Interview Survey

chematic.

S. Ahmad / Preventive Medicine 41 (2005) 276–283278

data for adults, and Teenage Attitudes and Practices

Survey II [13] for youth.

For every year that follows, the number of never

smokers within each age by gender group is multiplied by

the corresponding age- and gender-specific initiation

probability to determine how many people leave the never

smoker category to join the current smoker category. The

number of current smokers (by age and gender group) is

multiplied by group-specific cessation probabilities to

obtain the number of individuals subtracted from the

current smoker group and added to the former smoker

group. Similarly, the number of former smokers (by age

and gender group) is multiplied by group-specific relapse

probability to indicate how many former smokers will

become current smokers again.

Public health outcomes are measured in quality-adjusted

life years (QALYs), a measure recommended by the US

Task Force on Cost-Effectiveness in Health and Medicine

[19] to quantify both the number and the health-related

quality of years of life lived by a person. QALYs are

calculated by assigning to each year of life a quality weight

from zero to one, with zero indicating complete disability

and one indicating optimal quality, and then summing the

weighted life years.

Our model employs quality of life (QOL) estimates that

vary with age, gender, and smoking status. We derived these

estimates from a polynomial regression on data from two

major sources. Kaplan’s Quality of Well Being Scale (RM

Kaplan, personal communication, 1999) provided QOL

estimates for ages 17 and up. Analysis by Erickson et al.

[20] of National Health Interview Survey data provides

comparable QOL estimates for children and adolescents.

Current smokers have lower health-related quality of life

than former smokers, who have lower health-related quality

of life than never smokers. For example, the QOL values for

males in the 40- to 44-year age group are 0.82, 0.88, and 0.90

for current, former, and never smokers, respectively. The

QOL values for females in the same age group and smoking

status are 0.83, 0.87 and 0.88, respectively. Health-related

quality of life values decrease as age increases.

Total medical care costs for youth came from the Medical

Expenditure Panel Survey [21] and medical costs for adults,

by age, gender, and smoking status were derived from

Hodgson’s work [22]. The increase in medical costs over the

simulated period was determined with an autoregressive

moving average model introduced by [23] first forecasting

medical CPI and medical cost data and then adjusting the

medical cost data by the medical CPI to obtain costs in 2003

U.S. dollars. For every year of the simulation, the model

assigns each individual the estimated average value for each

health and economic outcome based on the person’s age,

gender, and smoking status at that time point. These annual

outcome values are then summed across the 75-year run for

the whole population to determine the final outcomes.

We start the present simulation in the year 2003 and

simulated annual transitions over 75 years. Because we are

not evaluating the future benefits of an action against

present costs, we did not discount our projected results.

Estimating price elasticity

Price elasticity is the change in the demand for a product

that results from a change in the price of that product. It is

almost always negative, indicating that increasing prices

reduces demand. The amount of smoking behavior change

that can be expected to result from a given change in

cigarette taxes can be captured using price elasticity. If taxes

are raised to increase cigarette prices by p percent and the

price elasticity of smoking participation is e, the resulting

expected proportionate change in smoking prevalence will

be the product of p and e. For example, if the price of

cigarettes were $2.00 per pack, and taxes were increased to

raise the price to $3.00 per pack (a 50% increase), and if the

price elasticity of smoking participation were �0.20 for a

given group, the proportionate change in smoking preva-

lence would be �10% (50% times �0.20). If in this scenario

the smoking prevalence for that group before the tax

increase were 20%, one should expect prevalence to drop

by 2% (10% of 20%) to 18%.

Consistent with the work of other researchers [1,2], we

estimated age-specific price elasticity of smoking participa-

tion with a weighted ordinary least squares model using a

standard model of consumption. The estimation function

can be represented as follows:

Ykijt ¼ a þ b1P

Cjt þ b2X it þ b3Rj þ b3Tt þ eijt

The dependent variable is dichotomous, indicating whether

the ith individual in the jth state is a smoker in year t. The

independent variables represent socioeconomic and other

factors typically associated with cigarette consumption. PjtC

is the price of cigarettes in state j and year t. Xit is a vector

of individual characteristics including gender, health status,

age, race, education, marital status, and income. Rj is a set

of region or state dummies, Tt is a set of year dummies, and

eijt is a random disturbance term.

The elasticity estimation used data from the Centers for

Disease Control and Prevention’s (CDC) Behavioral Risk

Factor Surveillance System (BRFSS)—a representative

survey of each of the fifty states and the District of Columbia

stratified by age, sex and race. Because the BRFSS over-

samples residents of small states and slightly under-samples

whites, men, and younger adults, all estimates are deter-

mined using appropriate sample weights. Annual data from

each state from 1993 to 2000 (except for 1993, when

Wyoming did not participate) are pooled into a single data

set, which is then merged with state-specific cigarette tax and

price data [24]. Around 6% of available observations were

omitted due to missing data, resulting in a final sample size

of 1,000,013 observations for the selected time period.

Table 1 presents price coefficients with standard errors

and price elasticities used in the dynamic simulation model

Table 1

Price effects on smoking

Dependent variable 15 to 17 18 to 23 24 to 29 30 to 39 40 to 64 65 and over

Status as a current smoker – �0.0320 �0.0265 �0.0175 �0.0171 �0.0124

– (0.010) (0.008) (0.005) (0.004) (0.004)

Elasticity for current smoker status �0.831 �0.3565 �0.2957 �0.1809 �0.1979 �0.3286

This table presents weighted ordinary least squares estimates from the BRFSS data, 1993–2000. There are 1,000,013 observations. All regressions estimate

controls for gender, age, race, education, income, martial status, health status, and region effects although not reported. Standard Error’s are in parenthesis.

Coefficient estimates from Harris and Chan [2] for the age group of 15- to 17-year-olds are not comparable since they used a slightly different method.

S. Ahmad / Preventive Medicine 41 (2005) 276–283 279

for six age groups: 15 to 17, 18 to 23, 24 to 29, 30 to 39, 40

to 64, and 65 and over. Because the BRFSS does not include

individuals under the age of 18, price elasticity for the 15 to

17 age group was taken from Harris and Chan [2], which

uses a model similar to ours and reports similar price

elasticity estimates to ours for other age groups. Estimates

for the remaining five groups were calculated from BRFSS

data as described above. Note that younger age groups

contain fewer years than older age groups because price

sensitivities tend to be less stable in younger people. Price

elasticities are lowest for 30- to 64-year-olds, higher for 18-

to 29-year-olds and people over 65, and highest for 15- to

17-year-olds. The jump in elasticity observed in young (18–

29) and elderly (65+) adults results may be due in part to the

lower incomes typical of these ages compared to 30- to 64-

year-olds. In general, people with lower incomes are

typically more sensitive to price changes [25]. The age

group-specific elasticities we computed were used in the

dynamic simulation model to estimate changes in smoking

behavior resulting from increased cigarette taxes.

Model calibration

To ensure the accuracy of the model, we calibrated it

against reliable external estimates of (1) smoking prevalence

(never, current and former), (2) population size (by age and

gender), and (3) life expectancy (by age, gender, and

smoking status). We used external estimates from California

Department of Health Services [14], California Department

of Finance [26], and American Academy of Actuaries [27]

to compare model output for smoking prevalence, popula-

tion count, and life expectancy, respectively. We compared

each model output with reliable existing external estimates,

and then made adjustments to select model parameters to

improve correspondence. We repeated this exercise until all

model outputs were within 1% of corresponding external

estimates.

To calibrate smoking prevalence, we started the model in

1995, loading it with historical data. We then ran the model

forward and observed the predicted prevalence of current

smokers, former smoker, and never smokers in every year

from 1995 to 2003. We corrected discrepancies in never

smoker prevalence by changing the initiation rate and

discrepancies in current and former smoker prevalence by

changing cessation and relapse rates. As an example, in

2000, our model estimated that 17.1% of the adult

population would be current smokers, this estimates

compared favorably to estimates from California Depart-

ment of Health Services [14], which reported prevalence as

17.1%, respectively.

To calibrate population size, we ran the model forward

through the year 2075. We compared population counts for

years 2025 and 2050 with California Department of Finance

projections for those years [26]. To improve the corre-

spondence between model estimates and Census projec-

tions, we made slight increases to fertility and decreases to

mortality. In the end, model estimates of population size for

both genders and all age groups differed from Census

projections by less than 1% for the future years 2025, 2050,

and 2075 (for all years, R2 = 0.99). For example, population

estimated by model in 2050 is 55.53 million compared to

census projections of 54.78 million.

Finally, we compared the simulated life expectancy of

current, former, and never smokers to external estimates

from the American Academy of Actuaries [27] and Hatton

Financial [28] revising mortality rates to improve corre-

spondence. Life expectancy estimates vary by age and

gender, but as one example, the model estimated that a 45-

year-old female never smoker would live 39.37 additional

years and a female current smoker of the same age would

live 33.94 additional years. These life expectancy estimates

compare favorably with insurance industry estimates of

39.33 and 33.89 years, respectively.

Calculating public health and economic impacts

The model assesses the health and economic consequen-

ces of various levels of taxation by first running a status quo

scenario with no tax increase using an average price per

pack of cigarettes of $3.95 [29]. Then, subsequent simu-

lation runs estimate outcomes for tax increases that raise

prices by 20, 40, 60, 80 and 100% (where 100% means the

tax increase doubles the price from the status quo

condition). Impacts of raising taxes are calculated as the

difference between outcomes for each intervention run (with

a price increase) and outcomes for the status quo run (with

no price increase). The model estimates the accumulation of

these outcomes for a 75-year period to ensure that the full

impact of the excise tax increase is experienced by the entire

population. After 75 years, virtually every individual in the

population will have experienced the influence of assumed

tax scenario for his or her entire lifetime.

Table 2

75-year health impacts of cigarette price increases

Price

increase

(%)

Adult smoking

prevalence

(%)

Cumulative life

years gained

(million)

Cumulative

QALYs gained

(million)

0 17.1 – –

20 11.6 14.08 15.95

40 8.8 21.78 24.79

60 7.1 26.70 30.45

80 6.1 30.18 34.44

100 5.3 32.82 37.45

S. Ahmad / Preventive Medicine 41 (2005) 276–283280

Smoking prevalence is the proportion of the adult

population in the current smoker category at the end of

the 75-year run. Cumulative life years gained for each level

of tax increase is difference of the total of life years lived by

the population in each intervention condition and the status

quo run. Quality-adjusted life years gained (QALYs) is

calculated similarly, taking the difference of the sums of the

products of each year of life lived and its age, gender, and

smoking status-specific QOL weight for each intervention

condition and the status quo.

The number of total packs consumed is the sum of the

products of the number of current smokers in the population

and the average cigarette consumption for a current smoker in

California (53 packs per year [14]) across all 75 years of the

simulation. Reduction in packs consumed is the difference

between the total packs consumed for each intervention run

and the status quo. Money saved by would-be smokers is an

estimate of how much less money individuals who spent

fewer years as a current smoker in each intervention condition

would have been spent compared to their projected cigarette

spending in the status quo condition. This is calculated by

multiplying the reduction in packs consumed in each

intervention condition by the base average per-pack price

of $3.95 in the status quo condition. Increased cost to

smokers, on the other hand, represents the total increase in

spending on cigarettes for those who continue to smoke in

each intervention condition and is calculated by multiplying

the total number of packs consumed by the increase in price

per pack from the base price. Net private spending on

cigarettes, the difference in the increased cost to smokers and

money saved by would-be smokers, represents the net effect

of an excise tax increase on total spending on cigarettes for all

consumers in California over the 75-year period.

Additional tax revenue is the difference between total

taxes collected in each intervention condition (the increased

per pack tax rate times the number of packs sold at the

higher price) and the status quo (the base case tax rate of

$0.87 per pack [6], times the projected number of packs sold

without a price increase). Cumulative medical cost savings

is the difference in total public and private medical

expenditures for each intervention condition minus those

for the status quo.

Results

75-year estimates of health and economic benefits of tax

increases

Over the course of the 75-year simulation for California,

raising excise taxes would have considerable impact on

smoking prevalence, health outcomes, and economic out-

comes at all levels of price increase.

Table 2 presents the effects of each level of tax increase

on smoking prevalence, life years and QALYs. Compared to

the status quo, a price increase as small as 20% would drop

adult smoking prevalence from 17.1% to 11.6%, reducing

the number of cumulative deaths at the 75-year mark by

557,800 for a total of 14 million life years and almost 16

million QALYs saved across those 75 years. Doubling the

price (a 100% price increase) would drop adult smoking

prevalence to 5.3% by 2075 and would boost the health

benefit to a cumulative gain of 32.8 million life years and

almost 37.5 million QALYs over the same time span.

The economic effects of excise tax increases, presented

in Table 3, would free up a large amount of money for those

motivated to avoid smoking, while increasing tax revenues

from increased tax contributions from those who continue to

purchase cigarettes at the higher price.

The reduction of smoking prevalence and the concom-

itant improvement of health outcomes would free up for

other uses money that would have been spent on tobacco

products and medical care. With a 20% tax-induced price

increase, cigarette consumption would drop by close to 3.2

billion packs over the 75 years, saving those who reduce or

eliminate time spent as a current smoker nearly $12.5 billion

over the 75 years. A 100% price increase from the status

quo would reduce consumption by about 7.9 billion packs

for savings of over $31 billion. Between $188 billion and

$412 billion would be saved on medical costs over that time

given a 20% to 100% price increase.

Even with the savings created by reduced smoking

participation, however, Californians, on the whole, will

spend anywhere between $270 million (in the 20% tax-

induced price increase scenario) and $13.9 billion (in the

100% tax-induced price increase scenario) more on cigarette

purchases over the 75-years compared to the status quo if

taxes are increased. All of the additional costs incurred by

smokers, however, will go toward increased tax revenues,

which can offset some of the social costs of smoking. These

increases in tax revenues would persist at all levels of tax

increase even as smoking prevalence declines. Increasing

taxes to raise prices by 20% would generate a total of $10

billion in additional tax revenues. If taxes increased until

cigarette prices doubled, tax revenues would increase by

over $38 billion compared to the status quo, even though

smoking prevalence would have dropped by almost two

thirds. If prices were increased by as much as 100%, the

total tax per pack would have to increase from the current

level of $0.87 per pack to $4.82-still well below the CDC’s

Table 3

75-year economic impacts of cigarette price increase

Price

increase

(%)

Total packs

consumed

(billion)

Reduction

in packs

consumed

(billion)

Money saved

by would-be

smokers

($ billion)

Increased

costs to

smokers

($ billion)

Net private

spending on

cigarettes

($ billion)

Additional tax

revenue

($ billion)

Cumulative

medical cost

savings

($ billion)

0 19.28 – – – – – –

20 16.12 3.16 12.47 12.74 0.27 10.0 187.8

40 14.25 5.03 19.85 22.52 2.67 18.14 286.1

60 13.00 6.28 24.79 30.81 6.02 25.26 345.1

80 12.09 7.18 28.38 38.21 9.84 31.97 384.0

100 11.39 7.88 31.14 45.01 13.87 38.15 411.6

S. Ahmad / Preventive Medicine 41 (2005) 276–283 281

estimate of $11.38 per pack of smoking-related medical care

costs and productivity loss borne by the state [6].

Discussion

The dynamic simulation model shows the great promise

of excise tax increases to reduce cigarette consumption

and improve health outcomes while generating tax

revenues, reducing medical expenditures, and freeing up

billions of dollars to be employed in non-tobacco related

sectors of the economy. The economic benefits to the state

continue to grow with price increases of at least up to

100%, even though cigarette consumption drops by over

two-thirds.

A simulation model is only as good as the parameters

upon which it operates. Our estimates of input variables

were based on the best available data, and where there was

uncertainty, we tried to err toward figures that would

produce in the model more conservative estimates of the

health and economic benefits. For example, price elasticity

for adolescents’ demand for cigarettes has been estimated

anywhere from 0.8 to 1.4 [5]. The figure we computed,

0.831, is on the low end of this range. If the actual elasticity

is, in fact, greater than this, tax revenues would be smaller,

smoking prevalence would be lower, and health benefits,

health care savings, and money not spent on tobacco would

be greater than the values reported in this paper.

We only consider smoking participation rates–not the

quantity of cigarettes smoked–in our model. Although the

quantity of consumption is related to health and economic

outcomes, we did not have the data to model accurately

the impact of price change on the amount smoked. It is

reasonable to assume, however, that increasing the price of

cigarettes would not increase the amount of cigarettes

consumed by those continue to smoke. At worst, the

results presented underestimate the benefits of a tax

increase.

Because adolescents have the highest risk for smoking

initiation of any age group and exhibit the highest price

elasticity for smoking participation, increasing taxes is

particularly effective at improving long-term population

health outcomes by reducing the number of new smokers

who enter the system. Given that adults rarely initiate

smoking, curbing teen initiation will drastically reduce the

population’s exposure to cigarettes [5], the incremental

health benefit that accrues to the population annually

increases with each passing year in every taxation scenario.

To give one example, in the 20% tax-induced price increase

scenario, the population accrues an additional 24,200

QALYs compared to the status quo in the 10th year of the

simulation. By the 75th year of the simulation, the

incremental annual health benefit would have grown to

543,500 QALYs.

The level of net private spending on cigarettes increases

with a rapidly accelerating slope as a function of the level of

tax-induced price increase modeled (producing a large

increase in spending from $270 million in the 20% scenario

compared to $13.9 billion in the 100% scenario). This is due

in large part to the fact that the number of packs consumed

decreases with a rapidly decelerating slope as a function of

price increase. The increase in private spending on

cigarettes price due to higher per-pack prices is offset by a

much larger drop in consumption from the 20% to 40% tax

increase scenarios than it is from the 80% to 100%

scenarios. Therefore, small increases in the per-pack price

of cigarettes are offset by large decreases in the number of

packs bought, but large increases in price are offset by

relatively small decreases in cigarette purchases. This

differential accumulates over the 75-year simulation period

to produce the substantial increase in private spending on

cigarettes reported. As mentioned before, all additional

private spending on cigarettes will become additional tax

revenue, which can be applied toward offsetting the societal

costs of smoking.

We did not find it appropriate to report discounted values

for the outcomes of interest. Discounting is employed when

a cost incurred in the present is compared to a benefit

anticipated in the future, and in this simulation, no positive

net costs are incurred at all. 75-year economic outcomes are

reported in 2003 dollars for ease of interpreting the true

value of the benefits incurred.

Unaddressed in the simulation model is the important

issue of smuggling, which may drop price elasticities for

cigarettes below our estimates and may introduce additional

law enforcement costs. Among the 50 states, there is a

strong correlation between cigarette excise tax rates and the

prevalence of interstate and international smuggling [30].

S. Ahmad / Preventive Medicine 41 (2005) 276–283282

Because interstate price differentials create a powerful

incentive to transport cigarettes across great distances, some

smuggling activity is inevitable [25,31]. Aside from cutting

into the tax revenues projected in the model, high rates of

smuggling would reduce the effect of the excise tax on

cigarette demand (and consequently, on smoking behavior

change) by making cheaper cigarettes available to consum-

ers on the black market. Although the exact magnitude of

these effects are not known, research suggests that the

impact of smuggling is not large, and smuggling may reduce

but will not eliminate the economic benefits of excise tax

increases [31].

Another important issue is the potential recessivity of

cigarette excise tax [32]. In general, as a proportion of

disposable income, excise taxes may represent a larger

burden to poor people than to more affluent individuals. In

the context of smoking, however, because price elasticities

are higher for people with lower incomes, some argue that

an excise tax increase would actually reduce the tax burden

and improve health outcomes to a similar or greater degree

for poor individuals compared to rich people as their total

cigarette consumption drops by a greater amount [25,33].

Even if this is true in the aggregate, however, policy-makers

should be sensitive to the possibility that high excise taxes

may put an unreasonable burden on individuals who do not

quit—particularly if a lack of resources is an impediment to

cessation.

Although these caveats should be considered when

discussing further tax increases, it is clear that the overall

benefits of raising cigarette taxes are considerable. Cal-

ifornia has not exhausted the potential of excise tax

increases to add to the health and economic benefits it

already enjoys. Raising cigarette taxes to levels equal to or

greater than those of highest-taxing states should be

considered a viable policy option to reduce further the

social burden of widespread smoking.

Acknowledgments

The National Institute of Drug Abuse through a grant to

the University of California Irvine Transdisciplinary Tobacco

Use Research Center (TTURC) (PHS Grant DA 13332)

supported this work. Robert Kaplan generously provided

unpublished quality of life data collected using the Quality of

Well Being scale. Gregor Franz, and John Billimek provided

research assistance. Finally, I thank Tammy Tengs and our

TTURC colleagues for their useful suggestions.

References

[1] Lewit EM, Coate D. The potential for using excise taxes to reduce

smoking. J Health Econ 1982;1:121–45.

[2] Harris JE, Chan SW. The continuum-of-addiction: cigarette smoking

in relation to price among Americans aged 15–29. Health Econ 1999;

8:81–6.

[3] Chaloupka F, Wechsler H. Price, tobacco control policies and smoking

among young adults. J Health Econ 1997;16(3):359–73.

[4] U.S. Department of Health and Human Services. Preventing tobacco

use among young people: a report of the Surgeon General. Atlanta,

Georgia: U.S. Department of Health and Human Services, Public

Health Service, Centers for Disease Control and Prevention, National

Center for Chronic Disease Prevention and Health Promotion, Office

on Smoking and Health; 1994.

[5] Kaplan RM, Ake CF, Emery SL, Navarro AM. Simulated effect of

tobacco tax variation on population health in California. Am J Public

Health 2001;91(2):239–44.

[6] McMahon K. State cigarette tax rates and rank, date of last increase,

annual pack sales and revenues, and related data. Available at: http://

tobaccofreekids.org/research/factsheets/pdf/0099.pdf. Accessed on

July 20, 2004.

[7] Tengs T, Ahmad S, Moore R, Gage E. Federal policy mandating

safer cigarettes: a hypothetical simulation of the anticipated

population health gains or losses. J Policy Anal Manage 2004;23(4):

857–72.

[8] Tengs T, Ahmad S, Savage J, Moore R, Gage E. The AMA proposal

to mandate nicotine reduction in cigarettes: a simulation of the

population health impacts. Prev Med 2005;40:170–80.

[9] Ahmad S. The cost-effectiveness of raising the legal smoking age in

California, Medical Decision Making [in press].

[10] Vensim [object-oriented modeling environment]. Release 4.2. Har-

vard, MA: Ventana Systems Inc; 2000.

[11] U.S. Bureau of the Census, Population Division. Detailed state

projections by single year of age, sex, race, and Hispanic origin:

1995 to 2025. Available at: http://www.census.gov/population/www/

projections/stproj.html. Accessed on May 20, 2003.

[12] National Center for Health Statistics (NCHS). Data warehouse on

trends in health and aging. Cigarette smoking by year, state, sex, race-

Hispanic origin, and age 1993–2001. Behavioral Risk Factor

Surveillance System, 2001 (BRS01b). Available at: http://www.cdc.

gov/nchs/agingact.htm. Accessed May 20, 2003.

[13] U.S. Department of Health and Human Services. National Center for

Health Statistics. Teenage attitudes and practices survey II, 1993. Ann

Arbor, MI: Inter-university Consortium for Political and Social

Research; 1994 [ICPSR 6375].

[14] California Department of Health Services (CDHS). California and

US adult per capita cigarette consumption. Packs per fiscal year,

1980/1981–1999/2000. Available at: http://www.dhs.ca.gov/tobacco/

documents/PR4-01graphs3.pdf. Accessed October 10, 2003.

[15] California Department of Health Services (CDHS). Table 2–30:

general fertility Rate and birth rates by age of mother, California

counties. Available at: http://www.dhs.ca.gov/hisp/chs/OHIR/vssdata/

tables.htm. Accessed October 10, 2003.

[16] U.S. Bureau of the Census, Population Division. 1996. Population

projections of the United States by age, sex, race, and Hispanic origin:

1995 to 2050 – middle series vital rate inputs. Available at: http://

www.census.gov/population/www/projections/natvital.html. Accessed

August 31, 1999.

[17] U.S. Department of Health and Human Services, National Center

for Health Statistics. National health interview survey. Health

promotion and disease prevention supplement. Washington, D.C.

Available at: http://www.icpsr.umich.edu. Accessed March 28,

2003.

[18] California Department of Finance. County population estimates and

components of change, July 1, 2001–2002, with historical 2000 and

2001 estimates. Available at: http://www.dof.ca.gov/html/demograp/

e-2.xls. Accessed October 13, 2003.

[19] Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in

health and medicine. New York7 Oxford Univ. Press; 1996.

[20] Erickson P, Wilson R, Shannon I. Years of healthy life. National

Center for Health Statistics, Healthy People 2000 Statistical Notes 7;

1995.

[21] Agency for Healthcare Research and Quality (AHRQ). Medical

S. Ahmad / Preventive Medicine 41 (2005) 276–283 283

expenditure panel survey. Available at: http://www.meps.ahrq.gov.

Accessed July 20, 2004.

[22] Hodgson TA. Cigarette smoking and lifetime medical expenditures.

Milbank Q 2001;70(1):81–125.

[23] Box GEP, Jenkins GM. Time series analysis: forecasting and control.

San Francisco7 Holden-Day; 1976.

[24] Orzechowski W, Walker R. The Tax burden on tobacco volume 35,

2000. Arlington, VA7 Orzechowski and Walker; 2001.

[25] Gruber J, Sen A, Stabile M. Estimating price elasticities when there is

smuggling: the sensitivity of smoking to price in Canada. J Health

Econ 2003;22(5):821–42.

[26] California Department of Finance. Population projections by race/

ethnicity for California and its counties 2000–2050. Available at:

http://www.dof.ca.gov/HTML/DEMOGRAP/DRU_Publications/

Projections/P1.htm. Accessed July 20, 2004.

[27] American Academy of Actuaries. Final report of the American

Academy of Actuaries’ commissioner’s standard ordinary task force.

Washington, DC: American Academy of Actuaries. Available at:

http://www.actuary.org/life/CSO_0702.htm. Accessed March 29,

2003.

[28] Hatton Financial Inc. Life expectancy table 2002. Available at: http://

www.hatton.ca/lifeexpectancies.htm. Accessed March 29, 2003.

[29] Orzechowski W, Walker R. The Tax burden on tobacco volume 38,

2003. Arlington, VA7 Orzechowski and Walker; 2003.

[30] Fleenor P. How excise tax differentials affect interstate smuggling and

cross-border sales of cigarettes in the United States. Washington, DC7

Tax Foundation; 1998 [Background Paper No. 26].

[31] Farrelly MC, Nimsch CT, James J. State cigarette excise taxes:

Implications for revenue and tax evasion. Research Triangle Inter-

national Report. Available online at: http://www.rti.org/pubs/

8742_Excise_Taxes_FR_5-03.pdf. Accessed July 20, 2004.

[32] Remler DK. Poor smokers, poor quitters, and cigarette tax regressiv-

ity. Am J Public Health 2004;94(2):225–9.

[33] Lindblom E. State cigarette taxes benefit lower-income smokers and

families. Available at: http://tobaccofreekids.org/research/factsheets/

pdf/0147.pdf. Accessed July 20, 2004.