A Short Description of the Model

Several years and many cigarette tax hikes have passed since we first obtained the data necessary to publish our original estimates of smuggling rates. To update our research with three additional years of data, we recently reran the model we constructed in 2008. That model was built only after an extensive review of the academic literature yielded strong evidence that substantial cigarette smuggling exists. These studies included, but were not limited to:

  • “How Far to the Border?: The Extent and Impact of Cross-Border Cigarette Smuggling,” by Michael Lovenheim and published in the National Tax Journal in March 2008. Lovenheim found that “between 13 and 25 percent of [U.S.] consumers purchase cigarettes in border localities.”[*], [3]

  • “Excise Tax Avoidance: The Case of State Cigarette Taxes,” a May 2008 working paper by Philip DeCicca, Donald Kenkel and Feng Liu. This study used 2003 survey data to estimate the percentage of smokers in each of the 50 states and the District of Columbia who were casual cigarette smugglers. In a 2010 update of the study, the authors estimated that in 2003, 6.2 percent of Michigan’s smokers engaged in casual cigarette smuggling.[†]

  • “Cigarette Tax Avoidance and Evasion,” by Mark Stehr and published in the Journal of Health Economics in March 2005. Stehr found that “up to 85 percent of the tax paid sales response”[4] was due to tax avoidance, rather than actual reductions in tobacco use.[‡]

The simplest way to describe our model is that it compares legal, per-capita sales of cigarettes to survey data on the percentage of smokers in each state. The difference between legal sales and reported rates of smoking provides a basis for estimating a state’s smuggling rate.[§] If the difference is positive, it indicates that the state is exporting cigarettes to other locations, making the state a “source state” for smuggled cigarettes. If the difference is negative, it indicates the state is importing cigarettes from elsewhere, making the state a “destination state” for smuggled cigarettes. A fuller treatment of the model’s construction — including a description of important variables — is provided in the Appendix.

The model contains a variable to measure the degree of international smuggling between the United States and Mexico or Canada. In these estimates, shipments are assumed to go only one way: from Mexico to the United States, or from the United States to high-tax Canada — not the other way around.

While the model includes two other countries in its calculations, it excludes two American states — Alaska and Hawaii — from smuggling measurements. These two states present unique challenges to modeling smuggling because they are not contiguous to the continental United States. North Carolina is also excluded from the model, since it is the model’s base source stage for commercially smuggled cigarettes, and other states’ taxes are measured against its own.[¶]

[*] In 2010, economist David Merriman published his paper “The Micro-Geography of Tax Avoidance: Evidence from Littered Cigarette Packs in Chicago.” Merriman found that about 75 percent of the discarded cigarette packs collected in Chicago were not acquired there. Slightly more were smuggled from Indiana than were found to have the Chicago tax stamp. David Merriman, “The Micro-Geography of Tax Avoidance: Evidence from Littered Cigarette Packs in Chicago,” American Economic Journal: Economic Policy 2, no. 2, 61 (2010).

[†] See Philip DeCicca, Donald S. Kenkel and Feng Lieu, “Excise Tax Avoidance: The Case of State Cigarette Taxes,” NBER Working Paper Series (2010): 56 (Table 2). In the course of our research, we produced annual smuggling rates for each of the 47 states in our model. Our unpublished estimate is that in 2003, 7.9 percent of Michigan’s total cigarette consumption involved casual smuggling imports — reasonably consistent with the figure produced by DeCicca, Kenkel and Lieu. Their estimate of the percentage of Michigan smokers who casually smuggled in 2007 was 7.23 percent, while our casual smuggling estimate for that year is 8.39 percent of state consumption — reasonably close again. DeCicca, Kenkel and Lieu, “Excise Tax Avoidance: The Case of State Cigarette Taxes,” NBER Working Paper Series (2010): 54 (Table 2).

[‡] Other useful U.S.-specific studies included R. Morris Coats, “A Note on Estimating Cross-Border Effects of State Cigarette Taxes,” National Tax Journal 48, no. 4 (1995);Jerry G. Thursby and Marie C. Thursby, “Interstate Cigarette Bootlegging: Extent, Revenue Losses, and Effects of Federal Intervention,” National Tax Journal 53, no. 1 (2000).

[§] The model is not designed specifically to capture smoking “intensity”— the amount smoked by people per day, for instance — but it does include a national trend variable reflecting the almost linear decline in U.S. smoking intensity over time.

[¶] North Carolina frequently plays this role in statistical studies of cigarette smuggling. Because North Carolina’s tax differential with itself would be zero, and because cigarettes would not be smuggled from North Carolina to itself, including the state in the study could bias the estimates.

[3] Michael F. Lovenheim, “How Far to the Border?: The Extent and Impact of Cross-Border Casual Cigarette Smuggling,” National Tax Journal 61, no. 1 (2008): 7.

[4] Mark Stehr, “Cigarette Tax Avoidance and Evasion,” Journal of Health Economics 24 (2005).