Is the CDC blowing smoke? Recommended levels of anti-tobacco spending have little effect on cigarette use.

Marlow, Michael L.

In 2007, the Centers for Disease Control and Prevention (CDC) released a report detailing "best practice" spending recommendations for state tobacco control programs. According to the report, "Research shows that the more states spend on comprehensive tobacco control programs, the greater the reductions in smoking--and the longer states invest in such programs, the greater and faster the impact."

The CDC spending guidelines use research to form "best practice" spending recommendations for each state. The report claims:

   Implementing a comprehensive
   tobacco control program structure
   at the CDC-recommended levels of
   investment would have a substantial
   impact. For example, if each
   state sustained its recommended
   level of funding for 5 years, an estimated
   5 million fewer people in
   this country would smoke. As a
   result, hundreds of thousands of
   premature tobacco-related deaths
   would be prevented. Longer-term
   investments would have even
   greater effects.

This claim should be met with skepticism. At least four reasons suggest that benefits from meeting spending targets are not as large as the CDC argues:

* The CDC recommendations draw heavily on research from just two states: California and Massachusetts. Those two states are considered models of effective programs, in part, because they have the longest funding histories. Even if highly effective, their success may not be easily exported to other states.

* The CDC ignores studies that show little to no impact from tobacco control programs.

* There is evidence, again ignored by the CDC, that little to no connection exists between state spending on tobacco control and the degree to which residents smoke.

* The CDC offers no empirical verification that implementing recommended spending targets causes significant reductions in tobacco use.

This article focuses on the last point above. Empirical investigation of the connection between spending guidelines and tobacco use is conducted to directly assess whether states that are closer to CDC guidelines actually exhibit lower tobacco sales than states that do not. Analysis focuses on two related issues. One is the fundamental question of whether tobacco sales over 2000-2007 were influenced by spending on tobacco control. The other is the policy question of whether meeting CDC spending targets matters in efforts to reduce tobacco sales.

TOBACCO CONTROL LITERATURE

Studies of tobacco control programs often focus on California and Massachusetts because of their long funding histories. Tobacco control in California began in 1988 when voters approved the California Tobacco Tax and Health Promotion Act of 1988 (Proposition 99), which increased the state surtax on cigarettes by 25 cents per pack and earmarked revenues for tobacco control programs. In 1992, a Massachusetts ballot initiative raised taxes 25 cents per pack, with the resulting revenue to be used for creation of the Massachusetts Tobacco Control Program.

Some empirical studies have indicated that the programs do indeed reduce smoking. However, the studies' reliability is unclear. A 2003 Journal of Health Economics paper by Matthew Farrelly et al. points out that most studies simply perform trend analysis on the introduction of new tobacco control programs and ignore other factors that might influence tobacco consumption. Those studies also focus heavily on California or Massachusetts. Even if their conclusions are valid--that tobacco control programs cause less smoking--it remains unclear whether extrapolation to other states is appropriate.

[ILLUSTRATION OMITTED]

Early studies that control for one or more factors outside of tobacco control programs uniformly show that the programs are highly effective in lowering tobacco use. A 1995 American Economic Review paper by T.-W. Hu et al. controlled for state excise taxes and tobacco firm media expenditures and found that state spending lowered consumption in California. A 1995 Journal of Public Health paper also by Hu et al. estimated that a 25 cent state tax hike reduced taxed sales in California when measured over about two years. Those two studies controlled for effects of time on cigarette consumption, which may control for various other factors--such as greater health concerns--that affect smoking over time. However, the Hu team's studies focus on California, the longest-lived state program, and examine effects on taxed sales over very few and very early years of a program that began in 1988.

The Farrelly paper examined tobacco control activities of all 50 states and concluded that state tobacco control expenditures lowered taxed cigarette sales over 1981-2000, after controlling for excise taxes, smuggling, time, and other state-specific factors. They also estimated that aggregate cigarette sales would have fallen by an additional 9 percent by 2000 if states had spent at minimum funding levels advocated by the CDC. A 2005 American Journal of Public Health paper by John Tauras et al. concluded that spending in the 50 states lowered youth smoking prevalence and the number of cigarettes smoked over 1991-2000, after controlling for other factors that might also influence sales.

The studies discussed so far examined years in which many states did not actively fund programs. The CDC only began publishing funding data in 2000 because many states did not actively fund programs until after the Master Settlement Agreement in 1998. It remains unclear whether the experience of tobacco control programs prior to when most states actively began funding the programs easily translates into recommendations for many states that only began funding around 2000. It is also unclear if experiences in the few states with relatively long funding histories easily convey to the many states without such histories.

BEST PRACTICES SPENDING TARGETS

The CDC's 2007 "best practices" report claims that a range of $15 to $20 per capita is a reasonable annual target for each state to fund tobacco control activities. Appropriate activities include anti-smoking ads, cessation interventions (intensive counseling services and cessation medications), and enforcement of age restrictions on the purchase of tobacco.

Table 1 displays how state spending on tobacco control over 2000-2007 compared to CDC-recommended minimum levels. Percentages of CDC-minimum levels are shown. Two states--Maine and Mississippi--averaged over 100 percent over this period, and three states--Michigan, Missouri, and Tennessee--spent nothing. Table 2 displays total spending in years 2000-2007 as well as average percentages of CDC minimums in each of those years. Total spending (in 2005 dollars) has been declining since 2003, and percent of CDC minimums has ranged between 33.6 percent and 46.9 percent. Aggregate spending is $5.3 billion, which is roughly $18 per capita. The CDC is thus arguing that states should have spent at least $8 billion more during those years, for a total of $13.3 billion, to meet minimum recommended levels of funding. That works out to roughly $44.30 per capita over those years. The summary measures appear to indicate ample variation with which to assess CDC claims that consistent meeting of recommended spending targets over time leads to significant reduction in tobacco use.

As mentioned previously, the CDC stresses the importance of maintaining state funding over time at levels that meet or exceed minimum targets. This suggests that it takes many years of tobacco control activity before full effects of programs can be detected. Spending is hypothesized to act as advertising: successful campaigns (in this case) reduce demand both today and tomorrow. The Hu et al. American Economic Review paper, the Farrelly et al. Journal of Health Economics paper, and my 2006 Cato Journal paper each found some evidence that cumulative measures of spending that discounted past spending on tobacco control were associated with less tobacco use.

This article employs discount rates of 5, 10, 15, and 20 percent in its analysis of the effectiveness of the CDC recommendations. This article also considers a cumulative measure of funding adequacy as defined by how closely states have met spending recommendations over time. Table 3 displays cumulative measures of tobacco control spending (in 2005 dollars) per capita. Values are shown for 2007, and so values over 2000-2006 are included at discounted rates. Average values are $19.53 and $11.68 for discount rates of 5 and 20 percent, respectively. Table 4 displays cumulative measures of funding adequacy, as defined as percentage of CDC minimum spending targets, at discount rates of 5 and 20 percent. Average values are 274 percent and 167 percent for discount rates ors and 20 percent, respectively. Cumulative measures of funding levels and adequacy thus range widely across states throughout the period.

MODELING EFFECTS OF TOBACCO CONTROL

Taxed sales of cigarettes are estimated using a pooled and balanced regression model over 2000-2007. A fixed effects model is estimated to control for state-specific factors outside of the model. Total sample size is 400 observations and represents all states across the eight-year time period. The following relationship for taxed cigarette sales is estimated:

[CIG.sub.it] = f([PRICE.sub.it], [PCY.sub.it], [UE.sub.it], [BAN.sub.it], [SMUG.sub.it], [CONTROL.sub.it])

The subscript i refers to the 50 states and t refers to years 2000 to 2007. The dependent variable [CIG.sub.it] is the number of tax-paid per-capita cigarette sales (in packs) and is obtained from William Orzechowski and Robert Walker's 2008 monograph The Tax Burden of Tobacco. The log of [CIG.sub.it] is examined, as is commonly done, and allows direct estimation of the price elasticity of demand because the log of the price variable is included on the right-hand side of the equation. [PRICE.sub.it] is the price (in 2005 dollars) of a pack of cigarettes, as reported by Orzechowski and Walker, and is hypothesized to be inversely related to cigarette consumption.

Real per-capita personal income [PCY.sub.it] is obtained from U.S. Department of Commerce data. The sign on [PCY.sub.it] is ambiguous because, while cigarette demand may be income-elastic and therefore exhibit a positive sign, higher income may also lower sales given common perceptions that higher-income individuals smoke less. Unemployment level [UE.sub.it] controls for the state of the economy and its sign is ambiguous as well because it is unclear if high unemployment is associated with more smoking from greater emotional distress or less smoking from greater financial distress.

The smoking ban variable [BAN.sub.it] controls for the effect of state-wide smoking bans in restaurants. Values of [BAN.sub.it] are set to 0 prior to bans and 1 thereafter to capture effects on taxed sales. Data on bans are obtained from a 2008 publication by the group Americans for Nonsmokers Rights. Some researchers have found that smoking restrictions lead to less smoking, which suggests that its sign will be negative.

Tax-paid cigarette sales do not fully reflect in-state consumption when smokers purchase some portion of cigarettes across state borders. [SMUG.sub.it] is defined as the ratio of the own-state price to the average for bordering states and is hypothesized to be inversely related to taxed sales because higher values indicate higher incentives for cross-border smuggling. Values for Hawaii and Alaska are set to 1 because they do not border other states and so their smokers are assumed to not purchase from other states. Data are obtained from Orzechowski and Walker (2008).

[CONTROL.sub.it] measures tobacco control spending and is measured in two alternative ways. The first measure is spending (again, in 2005 dollars) per capita ([SPENDING.sub.it]) and the second is as a percentage of CDC minimums ([ADEQUACY.sub.it]). As discussed above, [SPENDING.sub.it] is commonly used in the literature and [ADEQUACY.sub.it] measures CDC-defined funding adequacy. Data are obtained from a 2008 publication by the group Campaign for Tobacco-Free Kids. Cumulative measures of these variables are also examined to test whether differences in funding levels exert lagged effects on tobacco use over time and across states.

If states with relatively few smokers display less tolerance for smoking and therefore spend more on tobacco control, my econometric estimates of the effect of tobacco control spending will be inaccurate. Instrumental variable estimation is one method of dealing with this endogeneity problem and involves selection of a new variable that is both highly correlated with the independent variable in question (tobacco control spending) and also uncorrelated with all the other causes of smoking. Real state per-capita tobacco settlement funds [MSA.sub.it] are used as an instrument for [SPENDING.sub.it] and [ADEQUACY.sub.it] because those funds should influence funding availability, but it is unlikely that those funds independently influence cigarette sales. (As a check for whether [MSA.sub.it] is a weak instrument, we can determine whether the F-statistic exceeds 10 when testing the hypothesis that coefficients on all instruments are zero. In this case, the F-statistic for [SPENDING.sub.it] was 18.1, and for [ADEQUACY.sub.it] was 19.50, indicating that [MSA.sub.it] is a good instrument.) Settlement revenues were based on a formula that included smoking-attributable state Medicaid expenses and, while the agreement did not dictate how funds were to be allocated, tobacco control advocates argued that states should use those funds to significantly expand tobacco control spending. Evidence, however, indicates that those funds often go toward closing state government deficits and costs associated with general health care programs. Tobacco settlement data are obtained from Orzechowski and Walker's book.

CONTEMPORANEOUS SPENDING AND ADEQUACY

Table 5 displays estimates of tobacco control spending on cigarette sales. Columns (1) and (2) display estimations with and without instruments for tobacco control spending, and columns (3) and (4) display estimations with and without instruments for adequacy. Price coefficients, which measure price elasticities, lie between -0.82 and -0.84 and are in line with the expectation that demand for cigarettes is price inelastic. Per-capita income is found to exert negative and significant influences on sales. Unemployment and smoking ban variables do not exert significant effects. Smuggling is found to exert the hypothesized negative effect on sales in all estimations except in column (1). Coefficients on both measures of tobacco control spending are all positive and significant. Positive coefficients run counter to CDC arguments that sales fall with higher spending. The implication is that higher contemporaneous spending raises cigarette sales, and this result is found for estimations with and without instrumental variables.

At least two possibilities might explain positive coefficients on spending and adequacy measures:

* A "James Dean" effect might exist whereby spending on tobacco control raises social taboos against smoking and then (perversely) causes more smoking. There is some evidence that this happens with younger smokers following the introduction of smoking bans.

* More likely, specification error may exist when effects of spending on sales are longer-lived than the contemporaneous relation in Table 5. Tables 6 and 7 show that cumulative effects of spending and adequacy on sales are estimated to be either negative or zero, and so positive coefficients in Table 5 are likely the result of specification error.

It is also apparent that t-values associated with spending and adequacy coefficients are lower in instrumental value estimations. This commonly occurs because standard errors are biased downward in ordinary least squares estimations when endogeneity is present. This suggests that endogeneity is a problem and so only instrumental variable estimation of cumulative effects from spending and adequacy are displayed in the following tables.

SPENDING AND ADEQUACY

Table 6 displays estimations of cumulative control spending on cigarette sales with four alternative discount rates. Prices and income exert significant and negative effects in all four estimations, and effects of unemployment and smoking bans (except in column 4 where it is weakly significant and negative) cannot be distinguished statistically from zero. Smuggling exerts weakly significant and negative effects in all estimations. Cumulative spending measures exert no significant effects on sales in all estimations. Again, evidence does not support CDC claims that states that spend more will also have less tobacco use.

Table 7 displays estimations based on measures of cumulative spending adequacy. Effects from price, income, and unemployment mirror those of the previous table. Although smoking bans exerted negative and weakly significant effects in one instance in the previous table, it never exerts a significant effect in Table 7. Smuggling exerts significant and negative effects only when adequacy measures are discounted at 15 and 20 percent; it exerted negative effects in all estimations of the previous table. Spending adequacy exerts negative and significant effects when discounted at S and 10 percent. No effects are found when adequacy measures are discounted at 15 and 20 percent. Therefore, there is some support for CDC claims concerning funding adequacy when this measure is discounted at rates of 5 and 10 percent, but not at rates of 15 and 20 percent.

It is also worth noting that the effects of funding adequacy discounted at a 5 percent rate are statistically significant at the 95 percent confidence level, and at the 90 percent confidence level when funding adequacy is discounted at a 10 percent rate. In this log-linear specification, estimated coefficients of -0.0001 on funding adequacy at discount rates of 5 and 10 percent can be interpreted as unit changes in the cumulative value of adequacy associated with a 100 * -0.0001 percent change in taxed cigarette sales. Taxed sales fall by 0.011 percent for each additional unit. Based on average taxed sales in 2007 of 68 packs per capita, an additional unit rise in ADEQUACY would lower annual sales by 0.68 packs per capita. This estimate suggests economic effects are trivial.

CONCLUSIONS

Empirical evidence does not generally support the CDC claim that states that spend more on tobacco control deter more tobacco use than states that spend less. Contemporaneous spending on tobacco control is never found to exert an inverse effect on sales, and at times is found to exert a significant and positive effect on sales, contrary to the claims of the CDC. The true effect, however, appears to be zero based on current and past spending discounted at various rates. There is limited support for CDC claims regarding its recommendations on funding adequacy when this spending measure is discounted at rates of 5 and 10 percent, but not at rates of 15 and 20 percent. When significant, however, these effects arise at fairly low levels of confidence and with trivial effects on cigarette sales, and therefore suggest very cautious support for the CDC recommendations concerning adequacy. These conclusions are based on a battery of tests that consider various measures of contemporaneous and past spending and adequacy and are conducted over an eight-year period in which over $5 billion (in 2005 dollars), or roughly $18 per capita, was spent on tobacco control.

This study raises questions about the process by which the CDC determines its spending recommendations and whether the process is designed to reach a particular conclusion about tobacco control policy rather than to uncover policies that may best allocate resources toward controlling tobacco use. There may be a similarity to what I noted in a 2008 Econ Journal Watch paper on why the CDC and various researchers conclude that indoor smoking bans exert either positive or no adverse economic effects on restaurants and bars when, in fact, published studies demonstrate that numbers of businesses harmed are not zero. Factors include biases by governments and researchers that favor government solutions to perceived smoking problems, ample funding for researchers that conclude that bans exert no economic harm, simply ignoring industry-funded research that indicates some degree of harm, and tacit agreement between many researchers to not openly scrutinize the quality of colleagues' published research on this topic. It would be interesting to explore whether any of those factors might be influencing the policy process whereby the CDC makes spending recommendations regarding tobacco control. Those factors might also explain why the CDC is not compelled to demonstrate the effectiveness of its recommendations.

Readings

* "Adult Smoking Intervention Programmes in Massachusetts: A Comprehensive Approach with Promising Results," by H. Robbins, M. Krakow, and D. Warner. Tobacco Control, Vol. 11, Sup. 2 (2002).

* "Decreased Youth Tobacco Use in Massachusetts 1996-1999: Evidence of Tobacco Control Effectiveness," by S. Soldz, T. W. Clark, E. Stewart, C. Celebucki, and D. Walker. Tobacco Control, Vol. 11, Sup. 2 (2002).

* "Determinants of State Tobacco-Control Expenditures," by Michael L. Marlow. Applied Economics, Vol. 40, No. 7 (2008).

* "Do Expenditures on Tobacco Control Decrease Smoking Prevalence?" by Michael L. Marlow. Applied Economics, Vol. 41 (2009).

* "Do Tobacco-Control Programs Lower Tobacco Consumption? Evidence from California," by Michael L. Marlow. Public Finance Review, Vol. 35 (November 2007).

* "Has the California Tobacco Control Program Reduced Smoking?" by J. P. Pierce, E. A. Gilpin, S. L. Emery, M. M. White, M. Rosbrook, and C. Berry. Journal of the American Medical Association, Vol. 280 (1998).

* "Honestly, Who Else Would Fund Such Research? Reflections of a Non-Smoking Scholar," by Michael L. Marlow. Econ Journal Watch, Vol. 5, No. 2 (2008).

* "Impact of the American Stop Smoking Intervention Study on Cigarette Consumption," by M. W. Manley, J. P. Pierce, E. A. Gilpin, B. Rosbrook, C. Berry, and L. Wun. Tobacco Control, Vol. 6, Sup. 2 (1997).

* "Impact of the Massachusetts Tobacco Control Programme: Population Based Trend Analysis," by Lois Biener, Jeffrey E. Harris, and William Hamilton. BMJ, Vol. 321 (August 5, 2000).

* "Reducing Cigarette Consumption in California: Tobacco Taxes vs. an Anti-Smoking Media Campaign," by T.-W. Hu, H.-Y. Sung, and T. E. Keeler. American Journal of Public Health, Vol. 85 (1995).

* "State Expenditures for Tobacco-Control Programs and the Tobacco Settlement," by Cary P. Gross et al. New England Journal of Medicine, Vol. 347 (October 3, 2002).

* "State Tobacco Control Spending and Youth Smoking," by John A. Tauras et al. American Journal of Public Health, Vol. 95 (February 2005).

* "The Effects of Smoking Ban Regulations on Individual Smoking Rates," by Hielke Buddelmeyer and Roger Wilkins. University of Melbourne discussion paper 1737, 2005.

* "The First Decade of the Massachusetts Tobacco Control Program," by Howard K. Ko et al. Public Health Reports, Vol. 120 (Sept.-Oct. 2005).

* "The Impact of Clean Indoor Air Laws and Cigarette Smuggling on Demand for Cigarettes: An Empirical Model," by A. A. Yurekli and P. Zhang. Health Economics, Vol. 2 (March 2000).

* "The Impact of Tobacco Control Program Expenditures on Aggregate Cigarette Sales: 1981-2000," by Matthew C. Farrelly, Terry F. Pechacek, and Frank J. Chaloupka. Journal of Health Economics, Vol. 22 (2003).

* "The State Antismoking Campaign and the Industry Response: The Effects of Advertising on Cigarette Consumption in California," by T.-W. Hu, H.-Y. Sung, and T. E. Keeler. American Economic Review, Vol. 85 (May 1995).

* The Tax Burden on Tobacco: Historical Compilation 2007, by William Orzechowski and Robert C. Walker. 2008.

* "Tobacco Control Programs and Tobacco Consumption," by Michael L. Marlow. Cato Journal, Vol. 26 (Fall 2006).

* "Trends in Prevalence of Current Smoking: Massachusetts and States without Tobacco Control Programmes, 1990-1999," by J. M. Weintraub and W. L. Hamilton. Tobacco Control, Vol. 11, Sup. 2 (2002).

BY MICHAEL L. MARLOW

California Polytechnic State University

Michael k. Marlow is professor of economics in the Orfalea College of Business at California Polytechnic State University.

Table 1
Keeping Up With the CDC?
Annual state spending on tobacco
control for 2000-2007 as percentages
of CDC minimums

                  Average      Minimum      Maximum
                 (percent)    (percent)    (percent)

Alabama              7            1            22
Alaska               48           17           77
Arizona              96           66          132
Arkansas             82           0           103
California           57           45           82
Colorado             54           16          110
Connecticut          5            0            19
Delaware             56           0           119
Florida              26           1            56
Georgia              30           5            49
Hawaii               77           39           95
Idaho                11           5            17
Illinois             30           13           71
Indiana              64           31          101
Iowa                 36           26           49
Kansas               4            3            6
Kentucky             15           9            23
Louisiana            25           2            42
Maine               139          123          168
Maryland             67           30           99
Massachusetts        56           7           136
Michigan             0            0            0
Minnesota            94           65          122
Mississippi         108           0           165
Missouri             0            0            0
Montana              35           4            74
Nebraska             35           3            53
Nevada               30           22           33
New Hampshire        14           0            28
New Jersey           42           23           67
New Mexico           35           16           56
New York             45           31           89
North Carolina       19           0            40
North Dakota         27           0            38
Ohio                 73           35           97
Oklahoma             25           8            46
Oregon               31           14           53
Pennsylvania         49           0            80
Rhode Island         25           10           33
South Carolina       5            0            8
South Dakota         15           8            31
Tennessee            0            0            0
Texas                9            5            12
Utah                 44           39           47
Vermont              68           57           82
Virginia             40           32           57
Washington           68           45           82
West Virginia        41           38           42
Wisconsin            45           32           68
Wyoming              41           12           80

SOURCE: Campaign for Tobacco-Free Kids, "History of State
Spending for Tobacco Prevention," 2008

Table 2
National Totals
Annual state tobacco control

         Total    Percent
       Spending    of CDC
       ($2,005)   Minimum

2000   $771.3M        42.5
2001    813.1M        46.1
2002    814.0M        46.9
2003    715.9M        42.1
2004    561.3M        33.9
2005    538.2M        33.6
2006    533.9M        34.4
2007    562.6M        37.2

SOURCE: Campaign for Tobacco-Free Kids,
"History of State Spending for Tobacco
Prevention," 2008

Table 3
Per-Capita Spending
Discounted cumulative per-capita
tobacco control spending, in 2005 dollars

                       5%          20%

Alabama           $    2.69    $    1.08
Alaska            $   41.36    $   28.41
Arizona           $   32.57    $   18.37
Arkansas          $   37.66    $   24.02
California        $   18.43    $   10.57
Colorado          $   20.20    $   13.58
Connecticut       $    2.14    $    1.13
Delaware          $   42.04    $   33.18
Florida           $    8.21    $    3.43
Georgia           $    9.89    $    4.85
Hawaii            $   46.11    $   27.41
Idaho             $    6.12    $    3.51
Illinois          $   10.60    $    5.27
Indiana           $   24.47    $   12.16
Iowa              $   16.23    $    9.13
Kansas            $    1.70    $    1.14
Kentucky          $    6.27    $    3.28
Louisiana         $   11.00    $    7.40
Maine             $   82.67    $   48.32
Maryland          $   25.40    $   13.58
Massachusetts     $   20.44    $    8.86
Michigan          $    0.00    $    0.00
Minnesota         $   36.44    $   20.29
Mississippi       $   48.04    $   24.66
Missouri          $    0.00    $    0.00
Montana           $   25.27    $   17.94
Nebraska          $   18.30    $    9.45
Nevada            $   12.26    $    7.27
New Hampshire     $  785.00    $    3.12
New Jersey        $   15.22    $    8.00
New Mexico        $   17.85    $   11.96
New York          $   15.91    $   10.76
North Carolina    $    6.68    $    5.20
North Dakota      $   24.38    $   17.10
Ohio              $   27.27    $   16.38
Oklahoma          $   10.71    $    7.24
Oregon            $   12.69    $    6.46
Pennsylvania      $   18.37    $   12.04
Rhode Island      $   15.77    $    8.80
South Carolina    $    1.89    $    1.01
South Dakota      $   11.68    $    6.15
Tennessee         $    0.00    $    0.00
Texas             $    2.73    $    1.50
Utah              $   19.32    $   11.74
Vermont           $   60.19    $   35.05
Virginia          $   14.52    $    8.49
Washington        $   25.74    $   16.51
West Virginia     $   22.50    $   13.49
Wisconsin         $   17.72    $    9.51
Wyoming           $   42.99    $   31.79

SOURCES: Campaign for Tobacco-Free Kids,
"History of State Spending for Tobacco Prevention,"
2008; author's calculations

Table 4
CDC Recommendations
and Total Spending

Discounted cumulative measures of
funding adequacy as a percentage of
CDC minimum targets

                      Percent of CDC
                      recommendation

                    5%          20%
                 discount    discount
                   rate        rate

Alabama             41          17
Alaska              335         236
Arizona             637         378
Arkansas            566         369
California          379         224
Colorado            376         263
Connecticut         33          18
Delaware            418         333
Florida             159         68
Georgia             190         96
Hawaii              514         315
Idaho               74          44
Illinois            193         98
Indiana             408         208
Iowa                236         137
Kansas              25          17
Kentucky            98          53
Louisiana           176         120
Maine               924         555
Maryland            435         240
Massachusetts       342         151
Michigan             0           0
Minnesota           615         353
Mississippi         694         364
Missouri             0           0
Montana             251         183
Nebraska            225         120
Nevada              202         126
New Hampshire       83          33
New Jersey          274         148
New Mexico          244         168
New York            314         217
North Carolina      139         110
North Dakota        189         134
Ohio                487         300
Oklahoma            171         119
Oregon              200         105
Pennsylvania        340         226
Rhode Island        162         92
South Carolina      31          18
South Dakota        98          53
Tennessee            0           0
Texas               56          32
Utah                300         190
Vermont             451         269
Virginia            267         160
Washington          469         310
West Virginia       276         169
Wisconsin           295         163
Wyoming             297         224

Table 5
CDC Recommendations and Cigarette Sales

Effects of spending and funding adequacy on the log of cigarette
sales, with and without instrumental variables

                                                          (2)
                                                    Instrument for
                                      (1)              Spending

Real price per pack (PRICE)       -0.8430 ***          0.8430 ***
                                    (10.35)             (10.35)

Per-capita income (PCY)          -3.2E-05 ***        -2.9E-05 ***
                                     (764)              (6.37)

Unemployment (UE)                   0.0037              0.0011
                                    (0.48)              (0.14)

State-wide smoking                  -0.0198             -0.0249
ban in restaurants (BAN)            (1.12)              (1.38)

Ratio of in-state price to          -0.1863            -0.2148 *
border states' prices (SMUG)        (1.51)              (1.70)

Real per-capita spending on      0.011495 ***         0.02778 **
tobacco control (SPENDING)          (4.72)              (2.17)

Tobacco control spending
as percentage of CDC
minimums (ADEQUACY)

Constant                         10.39062 ***        10.14091 ***
                                    (26.71)             (24.82)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0805              0.0825

F-statistic                         146.07              138.75

Observations                          400                 400

                                                          (4)
                                                    Instrument for
                                      (3)              Adequacy

Real price per pack (PRICE)       -0.8205 ***         -0.8205 ***
                                    (9.84)              (9.84)

Per-capita income (PCY)          -3.22E-05 ***       -2.93E-05 ***
                                    (7.56)              (6.37)

Unemployment (UE)                   0.0029              -0.0011
                                    (0.39)              (0.14)

State-wide smoking                  -0.0238             -0.0249
ban in restaurants (BAN)            (1.34)              (1.38)

Ratio of in-state price to         -0.2286 *           -0.2148 *
border states' prices (SMUG)        (1.84)              (1.70)

Real per-capita spending on
tobacco control (SPENDING)

Tobacco control spending          0.0008 ***           0.0022 **
as percentage of CDC                 3.98                2.17
minimums (ADEQUACY)

Constant                          10.3865 ***         10.1327 ***
                                    (26.78)             (24.74)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0812              0.0825

F-statistic                         143.40              138.75

Observations                          400                 400

NOTES: *** significant 1% level (two-tailed test); ** significant
at 5 level; * significant as 10 percent level; t-statistics in
parentheses

Table 6
Cumulative Tobacco Control Spending and Cigarette Sales

Effects of cumulative control spending on cigarette sales with
instrumental variable for spending

                                      (1)                 (2)

Real price per pack (PRICE)       -0.8241 ***         -0.8286 ***
                                    (9.76)              (9.79)

Per-capita income (PCY)          -3.41E-05 ***       -3.48E-05 ***
                                    (6.45)              (6.67)

Unemployment (UE)                   -0.0017             -0.0024
                                    (0.21)              (0.29)

State-wide smoking                  -0.0289             -0.0296
ban in restaurants (BAN)            (1.57)              (1.61)

Ratio of in-state price to         -0.2244*            -0.2269*
border states' prices (SMUG)        (1.76)              (1.78)

                                 Discount Rate

                                      5%                  10%

Real per-capita spending on         0.0004              0.0008
tobacco control (SPENDING)          (0.50)              (0.79)

Constant                          10.4208 ***         10.4727 ***
                                    (23.99)             (24.00)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0830              0.0830

F-statistic                         136.90               13705

Observations                          400                 400

                                      (3)                 (4)

Real price per pack (PRICE)       -0.8347 ***         -0.8420 ***
                                    (9.83)              (9.89)

Per-capita income (PCY)          -3.56E-05 ***       -3.62E-05 ***
                                    (6.93)              (7.19)

Unemployment (UE)                   -0.0032             -0.0040
                                    (0.40)              (0.50)

State-wide smoking                  -0.0302            -0.0306*
ban in restaurants (BAN)            (1.65)              (1.68)

Ratio of in-state price to         -0.2268*            -0.2248*
border states' prices (SMUG)        (1.78)              (1.77)

                                 Discount Rate

                                      15%                 20%

Real per-capita spending on         0.0014              0.0022
tobacco control (SPENDING)          (1.11)              (1.44)

Constant                          10.5321 ***         10.5940 ***
                                    (24.06)             (24.17)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0829              0.0815

F-statistic                         137.30               13765

Observations                          400                 400

NOTES: *** significant 1% level (two-tailed test); ** significant
at 5 level; * significant as 10 percent level; t-statistics in
parentheses

Table 7
Cumulative Spending, CDC Recommendations, and Cigarette Sales

EFfects of cumulative adequacy on cigarette sales with instrumental
variable for adequacy of spending

                                      (1)                 (2)

Real price per pack (PRICE)       -0.7956 ***         -0.7939 ***
                                    (9.47)              (9.40)

Per-capita income (PCY)          -2.61E-05 ***       -2.72E-05 ***
                                    (5.03)              (5.28)

Unemployment (UE)                   -0.0044             -0.0040
                                    (0.55)              (0.50)

State-wide smoking                  0.0199               ~
-0.0213
ban in restaurants (BAN)            (1.09)              (1.16)

Ratio of in-state price to          -0.1933             -0.2027
border states' prices (SMUG)        (1.52)              (1.60)

                                 Discount Rate

                                      5%                  10%

Tobacco control spending          -0.0001 **           -0.0001 *
as percentage of CDC                (2.22)              (1.94)
minimums (ADEQUACY)

Constant                          9.9678 ***          9.9985 ***
                                    (23.08)             (22.98)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0824              0.0826

F-statistic                         138.84              138.37

Observations                          400                 400

                                      (3)                 (4)

Real price per pack (PRICE)       -0.7941 ***         -0.7965 ***
                                    (9.34)              (9.31)

Per-capita income (PCY)          -2.82E-05 ***       -2.93E-05 ***
                                    (5.56)              (5.86)

Unemployment (UE)                   -0.0034             -0.0027
                                    (0.43)              (0.34)

State-wide smoking                  -0.0227             -0.0240
ban in restaurants (BAN)            (1.24)              (1.31)

Ratio of in-state price to         -0.2106 *           -0.2163 *
border states' prices (SMUG)        (1.66)              (1.71)

                                 Discount Rate

                                      15%                 20%

Tobacco control spending            -0.0001             0.0001
as percentage of CDC                (1.64)              (1.30)
minimums (ADEQUACY)

Constant                          10.0418 ***         10.0962 ***
                                    (22.91)             (22.91)

R-squared                            0.95                0.95

Mean dependent variable              4.26                4.26

S.E. of regression                  0.0827              0.0829

F-statistic                         137.91              137.50

Observations                          400                 400

NOTES: *** significant 1% level (two-tailed test); ** significant
at 5 level; * significant as 10 percent level; t-statistics in
parentheses