(Editor’s note: This response is from the authors of the Center’s recent study on the effect of taxes on cigarette smuggling.)
February 1, 2013
Mr. Danny McGoldrick
Vice President for Research
Tobacco Free Kids
1400 I Street NW, Suite 1200
Washington, DC 20005
Dear Mr. McGoldrick:
By now you have probably seen our rejoinder to your letter online in The Kalamazoo Gazette. Unfortunately, the space limitation provided to us for such letters limits the degree to which we can explain our model and its workings, among other thoughts, so we thought we would write you and offer more detail. A small portion of the text below is taken directly from our letter.
It is not our methodology that is flawed, but your interpretation. Our model begins by estimating state level per-adult cigarette consumption by statistically comparing state level legal tax-paid cigarette sales per adult to the CDC's Behavioral Risk Factor Surveillance System estimate of state level smoking prevalence and a linear time trend.
The smoking prevalence variable provides an estimate of the percentage of each state’s population that smokes cigarettes; this value varies substantially across states. It seems you have mistakenly interpreted the linear trend as forcing per-adult consumption to be constant across all the states. The correct interpretation is that the linear time trend models a constant rate of decline in smoking intensity across the state. Specifically, consumption is estimated to decline by roughly 2.1 percent each year in each state.
We also estimated a model that allows for the rate of decline to vary across states, but it statistically added no additional explanatory power to the model (and did not change the other estimates in any significant way). Such a finding suggests that the rate of decline in consumption did not vary much from the 2.1 percent average. As such, we opted for the simpler model, a decision supported in almost all econometric textbooks.
The consumption estimates from this process are then compared to the observed state-level sales to derive an estimate of illicit cigarette consumption. This estimate is then modeled against the state cigarette tax and tax differentials to compute our estimate of tax-induced smuggling activity.
We are disappointed that you would blithely dismiss our own cigarette tax and smuggling research when so much other research exists from government and university economists that basically come to the same conclusion. For instance:
Michael Lovenheim’s 2008 National Tax Journal article — using data that ended in 2002 — only focuses on casual smuggling and concludes conservatively that “between 13 and 25 percent of consumers purchase cigarettes” in border locations or Indian Reservations. That is, they smuggle.
Perhaps more importantly, Lovenheim writes that “the central implication of this study is cross-border smuggling confounds many of the potential health and revenue gains from cigarette taxation.” This confounding should be of particular interest to your organization.
As an aside, there is a little irony in the fact that that the taxing jurisdiction Lovenheim identifies as having the highest percentage of “consumers who smuggle” (more than 60 percent) is Washington, D.C., home to the Tobacco Free Kids' headquarters.
It is worth noting that even when smuggling is not confounding public health goals the “substitution effect” is. Last August the Centers for Disease Control published its study “Consumption of Cigarettes and Combustible Tobacco — United States, 2000-2011" and reported that it appears “certain smokers” were switching types of tobacco to avoid high cigarette taxes. The report suggests this is perhaps due to the hike in the federal excise tax in 2009.
The 1995 study by Morris Coats titled “A Note on Estimating Cross-Border Effects of State Cigarette Taxes” — and also published in the National Tax Journal — concluded that “[A]bout four-fifths of the sales response to state cigarette taxes is due to cross-border sales.” In other words, published reports about declines in legal cigarettes bought in each state do not reflect the true declines in smoking.
In 2004 Mark Stehr published “Cigarette tax avoidance and evasion” in the Journal of Health Economics and came to a similar conclusion. He notes that “up to 85 percent of the tax paid sales response” to cigarette consumption is a result of trying to avoid taxes.
This obviously does not sum up the state of the literature, but it is worth asking if you believe investigators at state of New Jersey or the CDC and various university researchers are just “parrots” of others who pull their strings, or are they dispassionate scholars who have come to roughly similar conclusions about taxes and smuggling?
We don’t disagree with your assertion that higher tax rates help reduce smoking, but there is much evidence to suggest that it is not to the degree many people — perhaps including the good folks at Tobacco Free Kids — believe.
The timing of your letter, by the way, is quite interesting. Last Jan. 10, less than two weeks before it was published, 30,000 pounds of loose, smuggled tobacco was confiscated by Canadian officials at the Ambassador Bridge (which connects the U.S. and Canada between Detroit and Windsor). Smuggling tobacco is a serious problem. This is a direct function of high excise taxes and evidence seems to suggest that it undermines very laudable health goals.
The Center takes its research very seriously. Its scholars are confident that their research can stand up to rigorous peer reviews, but you did not provide one. Instead you called on readers to be skeptical based on a misunderstanding of our research. This in itself was unscholarly, if not uncharitable, particularly in light of significant other research by a wide array of researchers who basically come to similar conclusions.
It is our hope that in the future you might at least contact us about our methodology or consult the wide body of other evidence on the subject of higher excise taxes and smuggling.
Director of Fiscal Policy
The Ohio State University
Cc: Matthew Meyers