Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.

com

TC Online First, published on December 11, 2013 as 10.1136/tobaccocontrol-2013-051244 Research paper

Cigarette trafcking in five northeastern US cities

Kevin C Davis,1 Victoria Grimshaw,2 David Merriman,3,4 Matthew C Farrelly,1 Howard Chernick,5 Micaela H Coady,2 Kelsey Campbell,1 Susan M Kansagra2
1

RTI International, Research Triangle Park, North Carolina, USA 2 New York City Department of Health and Mental Hygiene, New York, USA 3 Department of Public Administration, Institute of Government and Public Affairs, Chicago, USA 4 Department of Economics, University of Illinois at Chicago, Chicago, USA 5 Department of Economics, Hunter College and the Graduate Center, City University of New York, New York, USA Correspondence to Kevin C Davis, RTI International, 3040 Cornwallis Road, Research Triangle Park, NC 27709, USA; kcdavis@rti.org Received 18 July 2013 Revised 6 November 2013 Accepted 15 November 2013

ABSTRACT Background Cigarette taxation is effective in reducing tobacco use in the USA. However, these benets are reduced when taxes are unpaid. Cigarette trafcking (ie, the illegal importation of cigarettes into a high-tax jurisdiction from a lower-tax jurisdiction) is well documented in high-tax places like New York City (NYC), but the extent of trafcking in other northeastern cities is relatively unknown. Objective To estimate the extent of cigarette trafcking in Boston, NYC, Philadelphia, Providence and Washington, DC, and project the benets of reducing cigarette trafcking for recouping lost taxes and reducing smoking in these cities. Methods Littered cigarette packs were collected from a random sample of Census tracts in ve US cities. Data collection yielded 1439 total littered packs. The share of cigarette packs bearing proper local, known non-local, foreign or unknown, or no tax stamp was calculated for each city. These data were used to estimate tax revenue recovery if cigarette trafcking could be eliminated. We also estimated the extent to which eliminating cigarette trafcking would reduce cigarette consumption. Results Overall, 58.7% of packs did not have a proper local tax stamp, and 30.542.1% were attributed to trafcking. We estimate that eliminating cigarette trafcking would result in declines in youth smoking prevalence ranging from negligible in low-tax cities like Philadelphia to up to 9.3% in higher-tax NYC. We estimate that these ve cities could recoup $680729 million annually in cigarette tax revenue if cigarette trafcking was eliminated. Conclusions Reducing cigarette trafcking would increase the effectiveness of tobacco taxes in reducing smoking and generate additional tax revenue, particularly in higher-taxed cities. Federal action to reduce cigarette trafcking, such as a track-and-trace system, is needed.

USA, differing cigarette tax rates across states, lack of accountability for the cigarette supply chain and relatively free ow of interstate commerce create conditions that allow interstate cigarette trafcking. We dene cigarette trafcking as the illegal importation of cigarettes from low-tax to high-tax jurisdictions. This denition does not include importation from outside the USA nor does it include purchases of small quantities of lower-tax cigarettes for personal use, which is a legal activity. At the time our data were collected, New York City (NYC) levied the highest combined city ($1.50) and state ($4.35) cigarette excise tax in the USA. While the potential for signicant cross-border shopping by smokers is high in cities like NYC, data from 2010 show that 84% of NYC smokers report getting their cigarettes from within the city limits.12 Furthermore, data suggest that cross-border purchasing decreases over time, presumably because smokers nd cross-border shopping too inconvenient or more smokers quit.13 14 Despite the fact that a majority of NYC smokers report purchasing their cigarettes from NYC retailers, prior studies of littered cigarette packs in NYC commonly nd packs with either no stamp or stamps from distant very-low-tax states.15 16 These studies suggest that organised cigarette trafcking is signicant. The current study used the littered cigarette pack collection method17 to estimate the extent of cigarette trafcking and the potential revenue and public health benets from reductions in trafcking in ve northeastern US cities. This method, used and validated in prior studies,17 identies both the location of consumption and the location of purchase. This provides for relatively accurate estimates of cigarette trafcking that are not reliant upon potentially biased self-reports of cigarette use. This is the rst study to use this method to estimate cigarette trafcking across multiple cities.

INTRODUCTION
Increasing the price of tobacco products can reduce tobacco use.1 The 2009 federal cigarette tax, for instance, was followed by immediate reductions in smoking initiation among adolescents.2 Increased cigarette taxes discourage progression from experimental to established smoking among adolescents and encourage young established smokers to quit35 and may also reduce tobacco-related morbidity and mortality,6 save millions of dollars in health expenditures79 and reduce occurrences of low-birth-weight babies.10 11 In addition to discounts and coupons that lower cigarette prices, cigarette trafcking also reduces the impact of cigarette taxes. The benets from reducing cigarette demand through cigarette taxes are not fully realised if some of those taxes are uncollected. In the

DATA AND SAMPLE Sample

We collected littered cigarette packs from a random sample of Census tracts in ve US cities: Boston, Massachusetts; New York, New York; Philadelphia, Pennsylvania; Providence, Rhode Island; and Washington, DC. These cities were selected for their proximity to one another (for ease of data collection), and their range of cigarette excise taxes (from a low of $1.60 in Philadelphia to a high of $5.85 in NYC). For sampling purposes, each city was dened by the U.S. Census Designated Place, with Providence and East Providence being combined. We randomly selected 30 Census tracts from each city using the probability proportional to size method such that the probability of a tracts selection was proportional to its weighted population
1

To cite: Davis KC, Grimshaw V, Merriman D, et al. Tob Control Published Online First: [ please include Day Month Year] doi:10.1136/tobaccocontrol2013-051244

Davis KC, et al. TobArticle Control 2013; 0:17. doi:10.1136/tobaccocontrol-2013-051244 Copyright author (or their employer) 2013. Produced

by BMJ Publishing Group Ltd under licence.

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
and employment size. Residential population size and worker population data were obtained from the Census Transportation Planning Package 2000 Part 1 and Part 2, respectively. Residential population was assigned a weight of three, whereas worker population was assigned a weight of one.15

Cigarette pack collection

Two collectors collected littered cigarette packs in each sampled Census tract. Collectors were instructed to walk the periphery of the tract, picking up all littered packs. Pack collection in each tract stopped after 45 min or after nding nine packs with cellophane, whichever came rst. The 45 min cut-off was based on our estimates of the average time it would take to walk the periphery of most Census tracts. The nine-pack limit is based on our estimates of sample size required to yield statistical power sufcient to identify effects from variation in cigarette taxes across cities. Because tax stamps are applied to cellophane wrappers, only packs with cellophane were included in the analysis. Collection was not performed if the ground was wet; for safety considerations, pack collectors limited the area in which they collected packs to 2 feet from the sidewalk or road, excluding parking lots. All packs were collected between 1 December and 20 December 2011. A signicant strength of this method is that the location of cigarette consumption is inferred with relative certainty by the location where the pack is found while the location of purchase is identied by the tax stamp. This eliminates the need to estimate the location of consumption based on self-reported cigarette use, which is often underreported in survey data. Since littered packs are not traceable to their owners, they are likely dropped without regards to the legality of the pack. Hence, estimates of cigarette consumption based on littered pack location should be relatively unbiased. Although it is possible that litterers and non-litterers may differ in their tax avoidance behaviours, some evidence18 suggests that littering is not disproportionately concentrated in any specic sociodemographic population.

Cigarette pack measures

Packs were coded by the following categories: (1) proper local stamp, (2) known non-local stamp (including packs found in NYC that bore only the New York State (NYS) tax stamp), (3) no stamp and (4) foreign or unknown stamp. Packs with a known non-local stamp have a known state origin within the USA. Packs with no stamp have an unknown origin and may originate from the three US states that do not use tax stamps (North Carolina, North Dakota and South Carolina), internet sales or Native American reservations. Packs from known nonborder states and packs with no stamp were considered to be trafcked. Although we assumed that non-border state packs resulted from organised trafcking, it is unknown what percentage of packs from bordering states were legal purchases for personal use by consumers. Therefore, we assumed that a portion of the packs from border states were trafcked, estimating a 25% lower bound and 75% upper bound as plausible. We note that while counterfeit stamps are another source of illegal cigarette sales, our coders could not recognise counterfeits without professional assistance beyond the resources of this study. Previous research suggests that the incidence of packs with counterfeit stamps is very low (less than 3.8%).17 Proportions for each tax stamp category were weighted to be representative of the city population by weighting by the inverse of the number of packs collected within each Census tract. This effectively weights each Census tract (rather than each pack) equally.17
2

We combined our littered cigarette pack ndings with data on adult smoking prevalence and average daily cigarette consumption to calculate overall consumption and consumption from trafcked sources in each city. We calculated smoking prevalence using the most recent available Behavioral Risk Factor Surveillance System (BRFSS) SMART City and County Data.19 These estimates were applied to 2010 U.S. Census city-level adult population statistics20 to estimate the total number of smokers in each city. The most recent 20102011 Tobacco Use Supplement of the Current Population Survey (TUS-CPS)21 was used to determine self-reported daily cigarette consumption for each core-based statistical area (CBSA) to calculate city-level consumption. Because smokers underreport cigarette consumption, we used established methods22 23 to adjust the TUS-CPS estimates of consumption to account for underreporting. Based on national sales data,24 smokers consumed 17.6 cigarettes per day in 2010 compared with 11.9 cigarettes per day among current daily or non-daily smokers as reported in the 2010 New York State Department of Health (NYS DOH)-sponsored National Adult Tobacco Survey (NATS). The NATS serves as the national comparison sample to the NYS DOH-sponsored New York Adult Tobacco Survey. This represents a 32.4% underreporting rate. We used this estimate to adjust each CBSA-level consumption estimate from the TUS-CPS. Adjusted local consumption estimates were then multiplied by the estimated percentage of packs consumed in each locality that were sourced via trafcking, as dened above. This percentage is based on the proportion of packs collected in each city that were from non-border states, packs with no stamp and a portion of border state packs, ranging from 25% to 75%, that were assumed to be trafcked. This yields lower and upper bound estimates of total trafcked cigarettes consumed in each locality. Finally, we calculated total potential revenue that could be recouped if trafcking was eliminated by multiplying our estimates of trafcked cigarettes by the prevailing cigarette tax rate in each city.

Statistical analysis Estimating potential revenue recovery from cigarette trafcking

Estimating the public health impact of reducing cigarette trafcking

For each city, we estimated the impact of eliminating cigarette trafcking on cigarette consumption using a cigarette price elasticity formula. This formula estimates the change in consumption as a function of percentage increases in effective cigarette prices if trafcking were eliminated in each locality. For this calculation, we used state-level average cigarette prices for nongeneric brands from the 2011 Tax Burden on Tobacco.25 Non-generic prices were used since premium brands constitute a large majority (90%) of trafcked cigarettes.16 Because NYC includes its own cigarette tax in addition to the NYS tax, we used the average NYC price per pack ($11.13) from the 2011 NYC subsample of the NYS Adult Tobacco Survey. Price increases were determined by calculating the average price paid per pack, weighted by the number of packs from each geographic source. This average was then compared with the weighted average per-pack price if all cigarettes consumed from trafcked sources were eliminated and instead consumed from non-trafcked sources. The difference between these effective prices represents the assumed price increase from eliminating trafcking. We assumed that 50% of this price differential is passed to consumers since trafcked cigarette sellers retain a portion of the price differential as prot.
Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
Table 1 Weighted proportion of tax stamp types among packs collected with cellophane by city
Type of tax stamp Proper local stamp Known non-local stamp No stamp Foreign or unknown stamp Overall % 41.3 (0.015) 34.1 (0.015) 9.0 (0.008) 15.7 (0.012) New York City (tax=$5.85) % 39.3 33.4 15.7 11.6 (0.031) (0.029) (0.024) (0.021) Boston (tax=$2.51) % 51.8 35.5 8.4 4.3 (0.037) (0.039) (0.017) (0.012) Providence (tax=$3.46) % 24.6 54.2 10.7 10.6 (0.029) (0.034) (0.019) (0.021) Philadelphia (tax=$1.60) % 73.1 (0.033) 1.2 (0.008) 1.0 (0.008) 24.7 (0.032) Washington, DC (tax=$2.50) % 18.6 45.5 9.0 26.9 (0.031) (0.037) (0.019) (0.032)

Cigarette excise taxes above are as of November 2011. Percentage of packs in each category in each city is calculated using weights designed to make packs representative of the total quantity of cigarettes consumed in that city. See Merriman34 for details. SEs are shown in parentheses.

Packs without tax stamps were treated as trafcked and were assumed to originate from one of the three states that do not stamp (North Carolina, North Dakota and South Carolina) for purposes of assigning a price to non-stamped packs. These states have a combined average non-generic retail pack price of $5.05. Packs that had foreign or unknown tax stamps were excluded from this calculation because their prices could not be determined. Our estimates of price increases in the absence of cigarette trafcking were combined with estimates of price elasticities for cigarette demand to determine the impact of eliminating cigarette trafcking on consumption. An extensive literature has used econometric models to estimate cigarette price elasticities. For example, a study by Nonnemaker and Farrelly26 suggests that a 10% increase in cigarette prices results in a 3.1% decline in youth smoking initiation (price elasticity=0.31). Other studies estimate price elasticities of 0.56 for youth prevalence,27 0.40 for adult cigarette consumption28 and 0.19 for adult prevalence.29 While there is an extensive literature on price elasticities of cigarette demand, we used the estimates noted above because they are relatively recent and are focused on US populations. Percentage declines in each outcome were calculated given the estimated percentage increase in effective prices in each city if cigarette trafcking was eliminated. This was done via algebraic manipulation of the elasticity formula (elasticity=% change in

Y/% change in X) that yields percentage declines in outcomes for given percentage increases in cigarette taxes.

RESULTS
Data collection yielded 1439 littered packs, of which 1183 (82.2%) had cellophane. A total of 344 packs (74.0% with cellophane) were collected in NYC, 253 packs (90.9% with cellophane) in Boston, 305 packs (83.3% with cellophane) in Providence, 280 packs (77.5% with cellophane) in Philadelphia and 257 packs (87.9% with cellophane) in Washington, DC.

Prevalence of cigarette pack origins

Across all ve cities combined, 41.3% of packs had the proper local tax stamp, 34.1% had known non-local stamps from other states or areas, 9.0% had no tax stamp and 15.7% had foreign or unknown tax stamps (table 1). Overall, 58.7% of packs did not have a proper local tax stamp, and between 30.5% and 42.1% of packs were attributed to trafcking.

Geographic sources of known non-local and non-stamped cigarette packs

Among all NYC packs that were missing proper local tax stamps, 44.7% had stamps from the state of Virginia, where the excise tax is $0.30 (table 2). Other sources included 30.9% with

Table 2 Geographic distribution of packs collected with known non-local stamps or no stamps by city and state issuer
New York City Issuing state California Connecticut Delaware Florida Georgia Kentucky Maine Maryland Massachusetts New Hampshire New Jersey New York State only Pennsylvania Virginia No stamp* Tax rate (per pack) ($) 0.87 3.40 1.60 1.34 0.37 0.60 2.00 2.00 2.51 1.68 2.70 4.35 1.60 0.30 0.49 (average) Total packs 1 0 2 3 5 2 0 0 0 0 7 8 2 55 38 Per cent 0.8 0.0 1.6 2.4 4.1 1.6 0.0 0.0 0.0 0.0 5.7 6.5 1.6 44.7 30.9 Boston Total packs 0 1 0 0 0 0 1 0 0 4 0 0 54 2 23 Per cent 0.0 1.2 0.0 0.0 0.0 0.0 1.2 0.0 0.0 4.7 0.0 0.0 63.5 2.4 27.1 Providence Total packs 0 0 0 0 0 0 0 0 117 2 0 0 7 2 27 Per cent 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 75.5 1.3 0.0 0.0 4.5 1.3 17.4 Philadelphia Total packs 0 0 0 0 0 0 0 0 0 0 1 0 0 1 2 Per cent 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 25.0 0.0 0.0 25.0 50.0 Washington, DC Total packs 0 0 0 0 0 0 0 39 1 0 0 0 0 61 21 Per cent 0.0 0.0 0.0 0.0 0.0 0.0 0.0 32.0 0.8 0.0 0.0 0.0 0.0 50.0 17.2

Tax rates on this table are as of November 2011. This table contains raw counts and is not relatively weighted to total city consumption. *Average tax rate among the three states (North Carolina, South Carolina and North Dakota) that do not use cigarette tax stamps.

Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
Table 3 Estimated cigarette consumption and potential tax revenue recovery by city
Cigarettes per day among adult current smokers 14.2 16.3 12.4 17.9 18.5 Percentage of packs trafficked (%) 47.949.9 36.838.4 29.655.4 1.92.2 29.059.9 Total packs consumed from trafficked sources (in millions) 108.7113.3 8.18.4 1.52.9 1.21.3 6.413.2 Total 2011 revenue that could be regained (in millions) ($) 636.1662.7 20.321.2 5.39.9 1.92.2 16.033.0

City/market (tax) New York City Boston Providence Philadelphia Washington, DC

Adult smoking prevalence* (%) 13.7 14.4 16.7 15.9 13.0

Adult smoker population total 878 075 73 993 22 775 188 042 65 082

*Smoking prevalence derived from the 2010 Behavioral Risk Factor Surveillance System (BRFSS) SMART City and County Data. Adult smoker population totals derived from the 2010 U.S. Census city-level adult population statistics and aforementioned city-level estimates of smoking prevalence. Cigarettes smoked per day estimates are derived from the 2010 core-based statistical area-level Tobacco Use SupplementCurrent Population Survey, adjusted for underreporting. Lower range is based on the assumption that 25% of neighbouring border sales are from trafficking, and upper range is based on the assumption that 75% of neighbouring border sales are from trafficking.

no stamps, 6.5% with a NYS-only stamp, 5.7% from New Jersey and 4.1% from Georgia. In Boston, improper stamps were most frequently from Pennsylvania (63.5%) and sources that did not use stamps (27.1%). This may be due to Pennsylvanias $1.60 tax rate compared with Massachusetts $2.51 per pack tax. Massachusetts, however, is the primary source of improperly stamped packs in higher-taxed Providence, accounting for 75.5% of packs without proper stamps. In Philadelphia, only 1.2% of packs had non-local stamps. The predominant sources of non-local packs in Washington, DC, were Virginia (50.0%), Maryland (32.0%), and non-stamp sources (17.2%).

NYC, 5.35.4% in Boston, 3.13.8% in Providence, 0.1% in Philadelphia and 2.16.1% in Washington, DC. Because NYC had the highest cigarette tax rate in the USA, the potential gains from eliminating trafcking are greatest there (table 5). We estimate that eliminating cigarette trafcking in NYC would result in relative declines of 5.25.5% in youth smoking initiation, 9.510.0% in youth smoking prevalence, 6.87.1% in adult cigarette consumption and 3.23.4% in overall adult smoking prevalence.

DISCUSSION
This study suggests there is substantial cigarette trafcking across the ve cities that were examined as 58.7% of cigarette packs collected had non-local stamps, foreign stamps or no stamps. Virginia was the most signicant source of NYC packs with known non-local stamps, with non-stamp sources being a secondary, signicant source of non-local cigarettes. The large quantity of cigarette packs from these sources suggests these packs are the result of organised trafcking into NYC rather than consumer-driven purchasing. The proportion of packs without stamps in NYC (15.7%) is considerably lower than reported in the recent Chernick and Merriman15 study of NYC (24.0%) and the recent Kurti et al16 study of the South Bronx (42%). Because cigarettes sold on many reservations do not have tax stamps, this suggests that the 2011 state regulations restricting wholesalers from selling untaxed cigarettes30 to reservations have deterred trafcking from reservations. This may also explain the increase in Virginia-sourced cigarettes compared with the relatively lower estimates of Virginia-sourced cigarettes found by Chernick and Merriman15

Lost tax revenue

Total consumption in 2010 was estimated to range from 227.1 packs per smoker in Providence to 338.3 packs per smoker in Washington, DC (data not shown). Based on our estimates of the total proportion of these packs that are trafcked into each city, we estimate that if trafcking was eliminated, the potential revenue recouped would be $636.1662.7 million in NYC, $20.321.2 million in Boston, $5.39.9 million in Providence, $1.92.2 million in Philadelphia and $16.033.0 million in Washington, DC (table 3). This represents a total of $680729 million in revenue that could be recouped.

Public health impact

The effective price for all packs consumed was $8.16 in NYC, $7.08 in Boston, $7.55 in Providence, $6.13 in Philadelphia and $6.05 in Washington, DC (table 4). When cigarette trafcking is eliminated, the effective price increases by 16.917.8% in

Table 4 Estimated effective prices paid by smokers on legal, imported and total cigarettes by city
City/market (average cigarette price) New York City ($11.13) Boston ($7.87) Providence ($8.22) Philadelphia ($6.14) Washington, DC ($7.50) Effective price paid per pack consumed (all packs) ($)* 8.16 7.08 7.55 6.13 6.05 Effective price paid per pack consumed (if trafficking eliminated via track and trace) ($) 10.9211.06 7.837.85 8.028.12 6.156.15 6.316.80 % Change in effective price if trafficking eliminated via track and trace 16.917.8 5.35.4 3.13.7 0.10.1 2.16.1

*Calculated as the average price per pack across all packs collected, weighted by the number of packs found from each source. Calculated as the weighted average price per pack consumed if illegal packs are eliminated (ie, illegal packs from long-distance non-border states are given a weight of zero). Lower range is based on the assumption that 25% of neighbouring border sales are from illicit trafficking, and upper range is based on the assumption that 75% of neighbouring border sales are from trafficking. Change in prices discounted to assume that traffickers retain 50% of effective tax differentials as profit (ie, only 50% of potential profit is passed to consumer).

Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
Table 5 Estimated impact of eliminating trafficking on key smoking-related outcomes among youth and adults by city
Reduction in youth initiation (price elasticity=0.31)* (%) 5.25.5 1.61.7 1.01.1 0.00.0 0.71.9 Reduction in youth smoking prevalence (price elasticity= 0.56)* (%) 9.510.0 3.03.0 1.72.1 0.10.1 1.23.4 Reduction in adult cigarette consumption (price elasticity= 0.40)* (%) 6.87.1 2.12.2 1.21.5 0.00.0 0.82.4 Reduction in adult smoking prevalence (price elasticity= 0.19)* (%) 3.23.4 1.01.0 0.60.7 0.00.0 0.41.2

City/market New York City Boston Providence Philadelphia Washington, DC

*Lower range is based on the assumption that 25% of neighbouring border sales are from illicit trafficking, and upper range is based on the assumption that 75% of neighbouring border sales are from illicit trafficking.

and is consistent with littered pack patterns found in the South Bronx before and immediately after these regulations.16 We also nd that Pennsylvania is a signicant exporter of cigarettes in the northeast. This is not surprising given its lower cigarette tax ($1.60 per pack) compared with nearby states. Pennsylvania was also the most signicant source of known nonlocal packs found in Boston. Because Massachusetts and Pennsylvania are not bordering states, this suggests organised trafcking rather than consumer cross-border shopping. Packs found in Philadelphia were also much more likely than other cities to have the proper local tax stamp. Although only 18.6% of Washington packs had a DC stamp, this should be interpreted cautiously. The majority (82.0%) of non-local packs collected in Washington, DC, originated from either Virginia (50.0%) or Maryland (32.0%). Washington, DC, has one of the highest proportions of commuters in the USA, so these packs may represent legal cross-border shopping. However, cigarette trafcking cannot be ruled out. While NYS regulated the sale of cigarettes to Native American reservations to reduce the supply of untaxed cigarettes, our results indicate that this may have shifted the supply of illegal cigarettes to lower-taxed states such as Virginia. This illustrates the fact that the actions of a single state to reduce trafcking may have minimal impact, requiring a coordinated national policy to reduce cigarette trafcking. One potential strategy would require manufacturers to apply codes or similar devices on tobacco products to enable federal and local ofcials to track the product from the point of manufacture to the nal point of retail sale and to trace packs to determine legal custody of the product. While this system would probably not eliminate trafcking, it would allow law enforcement to identify patterns of illegal trade when untaxed tobacco is found and facilitate investigation of potential trafcking when there are discrepancies between predicted and reported tobacco sales. In 2013, NYC and other public health organisations petitioned the Food and Drug Administration (FDA) to use their authority under the Family Smoking Prevention and Tobacco Control Act to create such regulations.31 Given the potential health benets and revenue recovery, FDA should consider such a track-and-trace system to improve oversight of the US tobacco supply chain and to reduce cigarette trafcking. Cigarette taxes are an effective intervention for preventing and reducing smoking,32 33 and maintaining the integrity of these taxes by requiring a track-and-trace system would extend these benets. We estimate that in NYC the effective price per pack would increase by as much as $2.90 if trafcking was eliminated and would result in decreases in youth and adult smoking.
Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

This study has several limitations. First, we did not have data on total cigarette exportation from each city. Thus, our estimates of potential tax revenue may be understated and estimates of potential revenue recovery overstated. In addition, our revenue estimates assume that all littered packs from long-distance states are trafcked. However, it is possible that some of these packs represent incidental state-to-state transfers, particularly in NYC, which is frequented by visitors from all areas of the USA.i Even if only unstamped packs represented trafcking, the estimated revenue recovery in NYC if trafcking was eliminated would be approximately $259 million. However, this is a conservative estimate given the high percentage of littered packs in NYC that come from Virginia, a known source for organised trafcking. If we further assume that only unstamped packs and Virginia-sourced packs represent illegal trafcking into NYC, the estimated revenue recovery increases to approximately $547 million. This variability illustrates the substantial impact of Virginia-sourced cigarettes on revenue estimates and more generally the impact of assumptions about how trafcking is dened. We also note that while our analysis suggests that elimination of trafcking may generate substantial recovery of revenue, we do not account for the additional costs associated with increased enforcement. Hence, our estimates should be viewed as upper bounds of potential revenue that could be recouped from such policies. Another limitation is that packs from foreign/unknown sources were excluded since we were unable to estimate the price of cigarettes from these sources. Further research is warranted given the sizeable sample of foreign/unknown cigarettes and the potential for international cigarette trafcking. Additionally, although prior research34 shows that the origins of littered packs do not signicantly differ from the origins of packs disposed of in trash receptacles, we did not replicate that analysis for this study. In calculating the health impact, we assumed that non-local or unstamped packs were purchased at a lower price that was equivalent to half the difference in effective prices if trafcking was eliminated. Trafckers may, however, sell these packs at lower or higher prices to consumers. A nal limitation is that the actual proportion of border-state packs that were trafcked is unknown. We assumed that this proportion is likely between 25% and 75% of border-state packs, and we used these estimates to generate modest lower bound and

We note that among all packs with known state origins and packs with no stamps, we found a small number of packs, three in total, with higher tax stamps in lower tax cities (one Connecticut pack in Boston, one New Jersey pack in Philadelphia and one Massachusetts pack in Washington, DC) 5

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
aggressive upper bound estimates of potential revenue recovery and public health impact. However, the estimated potential revenue recovery will be overstated if the actual proportion of border-state packs that are trafcked is less than 25% and understated if this proportion is more than 75%. In summary, cigarette trafcking can undermine the public health benets of cigarette taxes. Prior research in NYS has shown that cigarette tax evasion is associated with fewer smokers attempting to quit and reduced intentions to quit.35 Our study suggests that current tax policies would benet from better systems, such as track and trace, to reduce cigarette trafcking and ensure more robust public health benets from cigarette tax policy.
2 Huang J, Chaloupka FJ. The impact of the 2009 federal tobacco excise tax increase on youth tobacco use. NBER Working Paper No. w18026. 2012 April. Available at SSRN: http://ssrn.com/abstract=2056178 Katzman B, Markowitz S, McGeary KA. The impact of lending, borrowing, and anti-smoking policies on cigarette consumption by teens. Working Paper No. 9944. Cambridge, MA: National Bureau of Economic Research, 2002. Liang L, Chaloupka FJ. Differential effects of cigarette price on youth smoking intensity. Nicotine Tob Res 2002;4:10914. Taurus JA. Public policy and smoking cessation among young adults in the United States. Health Policy 2004;68:32132. Moore MJ. Death and tobacco taxes. Rand J Econ 1996;27:41528. Abt Associates Inc. Independent evaluation of the Massachusetts Tobacco Control Program, Sixth Annual ReportJanuary 1995 to June 1999. Report prepared for the State of Massachusetts, 2000. California Department of Health Services, Tobacco Control Section. California tobacco control update. Sacramento, CA: California Department of Health Services, 2000. Lightwood J, Glantz S. Short-term economic and health benets of smoking cessation. Circulation 1997;96:108996. Lightwood J, Phibbs CS, Glantz S. Short-term health and economic benets of smoking cessation: low birth weight. Pediatrics 1999;104:131220. Newhouse JP. Assessing the effects of tobacco policy changes on smoking-related health expenditures. In: Bearman P, Neckerman K, Wright L. eds What would happen if Americans stopped smoking? New York: Columbia University Press, 2011:20632. New York City Department of Health and Mental Hygiene. Epiquery: NYC Interactive Health Data System-Community Health Survey 2010. http://nyc.gov/ health/epiquery (accessed 19 Feb 2013). New York City Department of Health and Mental Hygiene. Epiquery: NYC Interactive Health Data SystemCommunity Health Survey 20032010. http://nyc.gov/health/epiquery (accessed 19 Feb 2013). Coady MH, Chan CA, Sacks R, et al. The impact of cigarette excise tax increases on purchasing behaviors among New York City smokers. Am J Public Health 2013;103: e5460. Chernick H, Merriman D. Using littered pack data to estimate cigarette tax avoidance in NYC. National Tax J 2013;66:63568. Kurti MK, von Lampe K, Thompkins DE. The illegal cigarette market in a socioeconomically deprived inner-city area: the case of the South Bronx. Tob Control 2013; Epub ahead of print. http://tobaccocontrol.bmj.com/content/early/2012/08/04/ tobaccocontrol-2011-050412.abstract Merriman D. Cigarette tax avoidance in the Chicago area: A contribution to the literature. Working Paper, Institute of Government and Public Affairs of the University of Illinois. 2008. http://igpa.uillinois.edu/system/les/documents/ WP-MerrimanCigLitter.pdf Williams E, Curnow R, Streker P. Understanding Littering Behavior in Australia. Community Change Consultants Report for The Beverage Industry Environment Council. Victoria, June 1997. Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System Survey Data. Atlanta, GA: Centers for Disease Control and Prevention, 2010. U.S. Census Bureau. 20062010 American Community Survey. 2010. http://quickfacts.census.gov/qfd/states/ U.S. Department of Commerce, Census Bureau. National Cancer Institute-sponsored Tobacco Use Supplement to the Current Population Survey (201011). 2012. http://riskfactor.cancer.gov/studies/tus-cps/. Data les and technical documentation (technical documentation Web site: http://www.census.gov/cps/methodology/ techdocs.html). Warner K. Possible increases in the underreporting of cigarette consumption. J Am Stat Assoc 1978;73:31418. Farrelly MC, Nonnemaker JM, Watson KA. The consequences of high cigarette excise taxes for low-income smokers. PLOS One 2012;7. http://www.plosone.org/ article/info%3Adoi%2F10.1371%2Fjournal.pone.0043838 AC Nielsen Co. ScanTrack Cigarette Item Rank Reports 20102011. Schaumburg, IL: AC Nielsen Co., 2011. Orzechowski & Walker. The Tax Burden on Tobacco: Historical Compilation. Vol 46. Arlington, VA, 2011. Nonnemaker JM, Farrelly MC. Smoking initiation among youth: the role of cigarette excise taxes and prices by race/ethnicity and gender. J Health Econ 2011;30:5607. Carpenter C, Cook PJ. Cigarette taxes and youth smoking: new evidence from national, state, and local Youth Risk Behavior Surveys. J Health Econ 2008;27:28799. Gallet CA, List JA. Cigarette demand: a meta-analysis of elasticities. Health Econ 2003;12:82135. Franz GA. Price effects on the smoking behavior of adult age groups. Public Health 2008;22:13438. New York State Department of Taxation and Finance. State Court Ruling Allows Implementation of Tax Law Provisions Related to Cigarettes Sold on Indian

4 5 6 7

What this paper adds This is the rst study to compare cigarette trafcking across ve major cities in the northeastern USA. More than half of all cigarette packs in ve northeastern cities bore non-local tax stamps or no stamps at all. Cigarette trafcking appears to be signicant in the northeast, with Virginia being the most signicant source of non-local cigarettes in NYC and Pennsylvania the most signicant source of non-local cigarettes in Boston. Our evidence, when compared with the results of Chernick and Merriman,15 suggests that as a result of recent state regulation in New York restricting the sale of untaxed cigarettes to Native American reservations, trafckers may have shifted their supply source from reservations to lower-tax states, such as Virginia. Revenue that could be recouped if cigarette trafcking could be reduced is between $680 and 729 million in total across these ve cities. This study suggests that current tax policies should be further complemented by better systems, such as track and trace, to ensure more robust public health benets from increased taxes and to reduce cigarette trafcking. Federal taxes also eliminate tax differentials that eliminate incentives to trafc cigarettes.

10 11

12

13

14

15 16

17

18

19

20

Acknowledgements We would like to thank Shannon Farely and Michael Johns (New York City Department of Health and Mental Hygiene) for reviewing the manuscript and Lesley Stalvey (New York City Department of Health and Mental Hygiene) for administrative assistance. We would also like to thank Joseph McMichael (RTI International) for sampling and statistical support and Stephanie Stolzenberg (RTI International) for data collection support and management. Contributors KCD led the study design, writing and directed data analysis. VG participated in manuscript writing and study design. DM contributed to study design, analytic methods and manuscript review. MCF participated in study design, data collection methods and review of nal manuscript. HC contributed to cigarette pack data collection protocols, study design and manuscript review. MHC participated in manuscript writing and study design. KC conducted all data analysis, statistical programming and assisted in manuscript writing. SMK participated in manuscript writing, manuscript editing and nal manuscript review. Funding This work was supported by the New York City Department of Health and Mental Hygiene. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed.

21

22 23

24 25 26

27

28 29 30

REFERENCES
1 Wilson LM, Avila Tang E, Chander G, et al. Impact of tobacco control interventions on smoking initiation, cessation, and prevalence: a systematic review. J Environ Public Health 2012;961724.

Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Research paper
Reservations to Go Forward. June 21, 2011. Technical Memorandum TSB-M-11(4.2) M. Retrieved from http://www.tax.ny.gov/pdf/memos/multitax/m11_4_2m_7_2s.pdf New York City Department of Health & Mental Hygiene, et al.Citizen Petition. March 6, 2013. Retrieved from http://www.regulations.gov/#!documentDetail; D=FDA-2013-P-0285-0001 Warner K. Tobacco policy research: Insights and contributions to public health policy. In: Warner K. ed Tobacco Control Policy. San Francisco, CA: Jossey-Bass, 2006:386. 33 Centers for Disease Control and Prevention (CDC). Best practices for comprehensive tobacco control programs2007. Atlanta, GA: Centers for Disease Control and Prevention, 2007. Merriman D. The micro-geography of tax avoidance: evidence from littered cigarette packs in Chicago. Am Econ J 2010;2:6184. Davis KC, Farrelly MC, Li Q, et al. Cigarette purchasing patterns among New York smokers: Implications for health, price, and revenue. Report prepared for New York State Department of Health, 2006.

31

34 35

32

Davis KC, et al. Tob Control 2013;0:17. doi:10.1136/tobaccocontrol-2013-051244

Downloaded from tobaccocontrol.bmj.com on January 9, 2014 - Published by group.bmj.com

Cigarette trafficking in five northeastern US cities

Kevin C Davis, Victoria Grimshaw, David Merriman, et al. Tob Control published online December 11, 2013

doi: 10.1136/tobaccocontrol-2013-051244

Updated information and services can be found at:

http://tobaccocontrol.bmj.com/content/early/2013/12/11/tobaccocontrol-2013-051244.full.html

These include:

References P<P Email alerting service

This article cites 13 articles, 3 of which can be accessed free at:

http://tobaccocontrol.bmj.com/content/early/2013/12/11/tobaccocontrol-2013-051244.full.html#re f-list-1

Published online December 11, 2013 in advance of the print journal. Receive free email alerts when new articles cite this article. Sign up in the box at the top right corner of the online article.

Notes

Advance online articles have been peer reviewed, accepted for publication, edited and typeset, but have not not yet appeared in the paper journal. Advance online articles are citable and establish publication priority; they are indexed by PubMed from initial publication. Citations to Advance online articles must include the digital object identifier (DOIs) and date of initial publication.

To request permissions go to:

http://group.bmj.com/group/rights-licensing/permissions

To order reprints go to:

http://journals.bmj.com/cgi/reprintform

To subscribe to BMJ go to:

http://group.bmj.com/subscribe/