The Impact of Cigarette Excise Tax Increases on Purchasing Behaviors Among New York City Smokers

Abstract

Objectives. We examined the relationship between cigarette excise tax increases and tax-avoidant purchasing behaviors among New York City adult smokers.

Methods. We analyzed data from the city’s annual Community Health Survey to assess changes in rates of tax avoidance over time (2003–2010) and smokers’ responses to the 2008 state cigarette tax increase. Multivariable logistic regression analysis identified correlates of buying more cigarettes on the street in response to the increase.

Results. After the 2002 tax increase, the percentage of smokers engaged in tax-avoidant behavior decreased with time from 30% in 2003 to 13% in 2007. Following the 2008 tax increase, 21% of smokers reported buying more cigarettes from another person on the street. Low-income, younger, Black, and Hispanic smokers were more likely than respondents with other sociodemographic characteristics to purchase more cigarettes on the street.

Conclusions. To maximize public health impact, cigarette tax increases should be paired with efforts to limit the flow of untaxed cigarettes entering jurisdictions with high cigarette pack prices.

Increasing taxation on cigarettes is among the most effective ways to prevent and reduce smoking.¹ Higher taxes are associated with lower cigarette consumption and higher rates of quit attempts,^2–4 particularly among low-income and younger smokers⁵ and among racial/ethnic minorities.^6–8 However, high cigarette prices also lead to increased tax-avoidant behaviors among smokers,^9,10 limiting the effectiveness of taxes in reducing tobacco use.^11,12 In response to tax hikes, smokers may seek untaxed sources of cigarettes in other jurisdictions, on the Internet, or from Native American reservations.^13–18 The availability of untaxed cigarettes from illegal vendors, such as individuals selling cigarettes on the street, allows tax-avoidant smokers to continue to smoke more cigarettes at a lower price^4,10,19; thus, trafficking of black market cigarettes across state or international lines into areas with high taxes is of particular concern.^10,18,20

We sought to better understand tax-avoidant behaviors in New York City, where city, state, and federal tax increases over the past decade have resulted in the highest cigarette pack price in the nation. In 2002, combined city and state tax increases raised the price of cigarettes by $1.81; in 2008, a state tax further increased the per-pack price by $1.25. We examined trends in tax-avoidant cigarette purchasing behaviors among adult smokers in the city between 2003 and 2010. We then assessed characteristics of smokers and the prevalence of tax avoidance among smokers in 2008. We also explored the impact of the 2008 tax increase on smoking behaviors. Finally, we used multivariable logistic regression analysis to identify correlates of buying more cigarettes on the street following the 2008 tax increase.

METHODS

Data came from the New York City Community Health Survey (CHS), a population-based, random-digit-dialed telephone survey of approximately 10 000 adult residents aged 18 years or older. Although the city’s Department of Health and Mental Hygiene has conducted the CHS annually since 2002, to ensure consistency in the primary variables of interest across years, we analyzed data from 2003 to 2010 only. Each year, the survey is conducted in English as well as other languages, including Spanish, Russian, and Chinese.

Surveys conducted between 2003 and 2008 contacted eligible households through landlines only. In 2009, the survey began using landlines and cell phone numbers to contact potential respondents. The number of interviews ranged from 8665 to 9818 per year; American Association of Public Opinion Research response rates each year ranged from 29% to 50%, with cooperation rates from 63% to 97%.²¹

Instrument

The CHS adapted its instrument from the Centers for Disease Control and Prevention’s Behavioral Risk Factor Surveillance System.²² The tobacco module incorporates questions on current smoking, secondhand smoke exposure, response to increases in cigarette taxes, and smoking cessation.

Because the CHS uses a complex sampling design, analyses require the use of a stratifying variable and a weighting variable. The stratifying variable has 34 strata that represent neighborhoods derived from the United Hospital Fund classification system.²³ The CHS began including a cell-phone-only sample with New York City exchanges in 2009. Thus, starting in 2009, the CHS added an additional stratum to represent adults living in households with only cell phones. The weighting variable adjusts for probability of selection and poststratification. Poststratification is accomplished by weighting each record up to the population of the neighborhood (as defined by the United Hospital Fund), taking into account the respondent’s age, gender, and race. Starting in 2009, the CHS weighted responses to adjust for the probability of selection and account for the distribution of the adult population into 3 telephone usage categories (landline only, landline and cell, cell only) according to data from the 2008 New York City Housing and Vacancy Survey. Analyses were also weighted to the 2000 US standard population.

Measures

Smoking behaviors.

The CHS defined current smoking as smoking on all or some days and having smoked at least 100 cigarettes in respondent’s lifetime. In our analysis of the CHS data, we defined heavy smoking as 11 or more cigarettes and light smoking as fewer than 11 cigarettes per day. The survey asked nondaily smokers how many cigarettes they smoked on the days they smoked and then how many days per month they smoked. The CHS defined a quit attempt as intentional cessation for at least a single day in the past year.²⁴

Cigarette purchasing behaviors and tax avoidance.

From 2003 to 2010, the CHS asked 1 question to assess the location of the most recent cigarette purchase and tax-avoidant behaviors: “Where did you get the last cigarette that you smoked?” Response options differentiated non-tax-avoidant purchases (“gas station, deli or other store in NYC [New York City]”) from tax-avoidant purchases (“outside NYC but in NYS [New York State],” “internet/mail,” “another person/street in NYC,” “another person/street location unknown,” “Indian reservation,” “different state,” “duty-free,” “outside the USA”). We excluded respondents who made cigarette purchases at an unknown or other location from the analysis (n = 29; 3% of respondents). The survey asked smokers whose most recent cigarette came from a carton or pack or was a single (“loosie”) about the purchase location for the most recent cigarette smoked. It did not ask this question of respondents who reported that they had bummed (n = 37; 3% of respondents) or rolled their own cigarette (n = 32; 3% of respondents).

Sociodemographic variables and smoking characteristics.

Sociodemographic variables were age, gender, race/ethnicity, borough of residence, education (for smokers aged ≥ 25 years), employment status, and household poverty, according to the federal poverty level (FPL), 2000 US Census. The logistic regression model had a dichotomous poverty measure (< 200% FPL vs ≥ 200% FPL).

Smokers’ response to the 2008 cigarette tax increase.

To assess the impact of the cigarette tax increase, the CHS asked current smokers a series of questions about how the increase affected their smoking and purchasing behaviors.

In June 2008, the price of cigarettes went up by $1.25 a pack. Since the price went up, have you: Thought seriously about quitting? Smoked fewer cigarettes? Switched to a cheaper brand? Smoked a pipe more, smoked cigars more or used chewing tobacco more? Bought more cigarettes from a person on the street? Bought more cigarettes from outside New York State, on the internet, through the mail, on an Indian reservation and/or duty free? Bought more “loosies”?

Responses were not mutually exclusive.

Statistical Analysis

To assess trends in cigarette tax avoidance among New York City smokers (2003–2010), we calculated percentages and 95% confidence intervals (CIs) for each year. We also calculated percentages for sociodemographic characteristics and smoking behaviors among adult smokers in the city in 2008 and examined the prevalence of tax-avoidant purchasing behaviors, stratified by each variable of interest. We used the χ² test to identify significant variation between groups; we followed up significant χ² test results with pairwise comparison using the t‐test. We considered differences significant at α = .05. Next, we examined smokers’ responses to the 2008 tax increase and stratified by tax-avoidant status of most recent cigarette purchased to assess any potential differences between the 2 groups.

Our multivariable logistic regression model identified characteristics associated with buying more cigarettes from a person on the street in response to the 2008 tax increase. The dependent variable was a dichotomous indicator of buying more cigarettes on the street (i.e., purchased more cigarettes from person on the street in response to the tax increase: yes or no). Independent variables found to be significant in bivariate analysis (P < .05) became candidates in the multivariable model. We excluded from analysis 12% of the smokers who bought more cigarettes from a person on the street (n = 24) because they answered “don’t know” to the poverty question. We ran the analysis with and without this exclusion, with similar findings. We used a priori knowledge to select potential confounding variables for the model.^8,10 We used the likelihood ratio test to assess model fit and derived adjusted odds ratios (AORs), corresponding 95% CIs, and P values from the final model.

We conducted all analyses with the survey procedures in SAS version 9.1 (SAS Institute Inc, Cary, NC) and SAS-callable SUDAAN version 10 (Research Triangle Institute, Research Triangle Park, NC) to account for the complex survey design, incorporate the survey weights, and age-standardize the estimates. In the descriptive analyses, unless otherwise noted, we age-standardized all estimates to the US 2000 standard population in 4 age strata.²⁵ We weighted all analyses to the New York City adult population.²⁶

RESULTS

Prevalence of tax-avoidant purchases of cigarettes ranged from 30% in 2003 to 13% in 2007. [Figure 1]. The highest rates of tax avoidance occurred in 2003 and 2004 (30% in each year). A 53% decline in tax avoidance occurred between 2006 and 2007 (20% to 13%). Following this decline, tax avoidance increased to 25% after the 2008 tax increase; it remained stable in 2009 and 2010. Beginning in 2005, the most common source of tax-avoidant cigarettes was “another person or on the street.” In 2005, this category accounted for 44% of all tax-avoidant cigarette purchases, increasing to 56% in 2008.

FIGURE 1—

Tax-avoidant cigarette purchases among adult smokers: New York City, 2003–2010.

Note. NYC = New York City; NYS = New York State. The following estimates should be interpreted with caution: Indian reservation (2005, 2007), Internet/mail (2006–2010), outside United States/duty-free (2007–2010). The relative SE was > 30% or the sample size was < 50, making these estimates potentially unreliable.

^aDenotes cigarette excise tax increases: NYS in 2008, federal in 2009, and NYS in 2010.

Smoker Characteristics and Tax-Avoidant Purchases

In 2008, about 50% of smokers were aged 25 to 44 years, and more than 40% were White (Table 1). The largest proportions of smokers lived in Brooklyn (31%) and Queens (28%), and about one third (36%) had a college education or more. About 42% of smokers were in the lowest income category (< 200% FPL), and 30% were in the highest income category (≥ 400% FPL). The majority of smokers were light smokers (71%), and most smoked daily (65%). About three quarters of smokers (76%) smoked their most recent cigarette from a pack, with a self-reported mean pack price of $7.40.

TABLE 1—

Characteristics of Adult Smokers and Prevalence of Tax-Avoidant Cigarette Purchases by Select Characteristics: New York City Community Health Survey, 2008

Variable	% (95% CI) or Mean (SE)	Prevalence, % (95% CI) or Mean (SE)	Pc
Total sample	100	24.7 (21.0, 28.9)
Demographics
Age, y			.27
18–24	12.9 (9.5, 17.2)	29.7a (16.0, 48.4)
25–44	47.5 (43.2, 51.8)	20.2 (15.3, 26.1)
45–64	30.7 (27.4, 34.3)	24.4 (19.6, 29.9)
≥ 65	8.9 (7.1, 11.0)	32.5 (22.2, 44.8)
Gender			.5
Male	55.5 (51.3, 59.6)	25.5 (20.0, 32.0)
Female	44.5 (40.4, 48.7)	25.2 (20.3, 30.8)
Race/ethnicity			.73
Non-Hispanic White	40.4 (36.5, 44.5)	24.2 (18.7, 30.7)
Non-Hispanic Black	24.2 (20.8, 27.9)	29.2 (21.5, 38.4)
Hispanic	24.2 (20.4, 28.3)	21.3 (14.7, 29.9)
Asian	9.2 (6.8, 12.3)	23.9 (14.0, 37.7)
Other	2.1 (1.3, 3.4)	20.7a (10.5, 36.7)
Borough			.77
Bronx	16.4 (13.8, 19.5)	21.8 (15.2, 30.3)
Brooklyn	30.6 (27.0, 34.3)	23.1 (16.6, 31.2)
Manhattan	18.1 (15.1, 21.4)	27.7 (20.0, 37.0)
Queens	27.6 (23.8, 31.8)	25.9 (18.8, 34.5)
Staten Island	7.3 (5.7, 9.2)	23.1 (15.0, 33.9)
Education			.28
≤ high school	20.0 (19.9, 27.4)	29.1 (19.7, 40.7)
High school graduate	23.9 (20.8, 27.7)	24.3 (18.0, 32.0)
Some college	20.2 (17.7, 24.6)	25.2 (17.7, 34.6)
≥ college	35.9 (28, 35.3)	23.7 (18.2, 30.3)
Employment status			.05
Employed	64.6 (60.6, 68.5)	23.3 (17.9, 29.8)
Unemployed	9.9 (7.5, 12.9)	25.2 (15.2, 38.7)
Not in labor force	25.5 (22.2, 29.1)	32.5 (24.6, 41.6)
Household poverty, % FPL			.52
< 200	42.4 (38.1, 46.8)	26.6 (20.9, 33.3)
200–399	17.9 (15.0, 21.2)	21.7 (14.4, 31.3)
≥ 400	30.4 (26.7, 34.4)	23.4 (17.4, 30.8)
Don’t know	9.3 (6.9, 12.5)	29.9 (17.4, 46.4)
Current smoking characteristics
Consumption, cigarettes/d			.16
Light (< 11)	70.5 (66.9, 73.9)	26.3 (21.8, 31.1)
Heavy (≥ 11)	29.5 (26.1, 33.1)	20.6 (14.6, 28.2)
Total	9.5 (0.3)	9.80 (0.9)	.74
Smoking frequency			.07
Nondaily	35.4 (31.4, 39.7)	30.9 (24.1, 38.7)
Daily	64.6 (60.3, 68.6)	21.7 (17.9, 26.2)
Quit attempt in the past y			.02
Yes	61.5 (57.3, 65.5)	21.1 (16.5, 26.5)
No	38.5 (34.5, 42.7)	30.2 (24.2, 36.9)
Cigarette pack price, $	7.40 (0.1)	5.48 (0.3)	<.001
Source of most recent cigarette			<.001
Pack	76.0 (72.2, 79.3)	17.5b (13.6, 22.2)
Single	6.6 (4.7, 9.2)	44.7b (31.6, 58.5)
Carton	10.3 (8.3, 12.8)	67.5 (56.3, 77.0)

Note. CI = confidence interval; FPL = federal poverty level (according to 2000 US Census). Characteristics of current smokers and mean pack price are not age-standardized to the US 2000 standard population.

^aEstimate should be interpreted with caution: relative SE (a measure of estimate precision) is > 30% or the sample size is < 50, making the estimate potentially unreliable.

^bSignificant difference within source of most recent cigarette (reference level = pack), among tax-avoidant smokers, derived from pairwise t test (P < .05).

Overall, about one quarter (25%) of smokers made tax-avoidant purchases in 2008. Tax avoidance was highest among those who bought their most recent cigarette from a carton versus a pack (68% vs 18%; P < .001) or loosie (68% vs 45%; P = .013) and among smokers who did not make a quit attempt in the past year (30% vs 21% who attempted to quit; P = .02). The mean per-pack price among tax-avoidant smokers was $5.48.

In response to the 2008 tax increase, more than half (61%) of smokers seriously thought about quitting, 46% smoked fewer cigarettes, and 13% switched to a cheaper brand (Table 2). Smokers also engaged in tax-avoidant purchasing behaviors. About one fifth (21%) reported buying more cigarettes from a person on the street, one fifth (21%) bought more cigarettes outside New York State, on the Internet, through the mail, at a duty-free shop, or on an Indian reservation, and 15% bought more loosies.

TABLE 2—

Response to the 2008 State Cigarette Tax Increase Among Adult Smokers: New York City Community Health Survey, 2008

Response to Tax Increasea	Weighted Population (n = 959 000)	% (95% CI)
Seriously thought about quitting	583 000	61.2 (57.0, 65.2)
Smoked fewer cigarettes	432 000	45.7 (41.5, 50.0)
Switched to cheaper brand	118 000	13.3 (10.6, 16.6)
Smoked more pipes or cigars or used chewing tobacco	24 000	3.0 (1.7, 5.0)
Bought more cigarettes from a person on the street	205 000	21.3 (17.8, 25.3)
Bought more cigarettes from outside the state, on the Internet, through the mail, on an Indian reservation, or duty-free	193 000	20.8 (17.7, 24.4)
Bought more single cigarettes	152 000	14.9 (12.1, 18.3)

Note. CI = confidence interval.

^aSurvey asked, “In June 2008, the price of cigarettes went up by $1.25 a pack. I'd like to ask you a few questions about how this increase affected you.” Responses were not mutually exclusive.

Correlates of Street Buying

In the final adjusted model, after control for all other variables, smokers aged 18 to 44 years were more likely than other smokers to buy more cigarettes from a person on the street in response to the 2008 cigarette tax increase (AOR = 2.20; 95% CI = 1.32, 3.67; Table 3).

TABLE 3—

Multivariable Logistic Regression Analyses Predicting Buying More Cigarettes From a Person on the Street in Response to the 2008 State Cigarette Tax Increase: New York City Community Health Survey, 2008

Variable	AOR (95% CI)
Demographics
Age, y
18–44	2.20* (1.32, 3.67)
≥ 45 (Ref)	1.00
Gender
Male (Ref)	1.00
Female	0.66 (0.37, 1.19)
Race/ethnicity
Non-Hispanic White/Asian/other (Ref)	1.00
Non-Hispanic Black	9.19* (4.44, 19.01)
Hispanic	6.66* (2.75, 16.13)
Borough
Bronx	1.79 (0.92, 3.48)
Other boroughs (Ref)	1.00
Household poverty, % FPL
< 200	2.41* (1.33, 4.35)
≥ 200 (Ref)	1.00
Current smoking characteristics
Tax-avoidant behavior at most recent cigarette purchase
Yes	3.50* (1.84, 6.67)
No (Ref)	1.00
Consumption, cigarettes/d
Light (< 11; Ref)	1.00
Heavy (≥ 11)	1.50 (0.74, 3.06)
Smoking frequency
Nondaily (Ref)	1.00
Daily	1.69 (0.83, 3.44)

Note. AOR = adjusted odds ratio; CI = confidence interval; FPL = federal poverty level (according to 2000 US Census).

*P < .05.

In comparison with Whites, Asians, and members of other races, Black smokers had more than 9 times the odds (AOR = 9.19; 95% CI = 4.44, 19.01) and Hispanic smokers had almost 7 times the odds of buying more cigarettes on the street after the 2008 tax increase (AOR = 6.66; 95% CI = 2.75, 16.13). Smokers in the lowest income group had more than twice the odds (< 200% FPL; AOR = 2.41; 95% CI = 1.33, 4.35) and smokers who purchased their most recent cigarette from a tax-avoidant location had more than 3 times the odds (AOR = 3.50; 95% CI = 1.84, 6.67) as other smokers of purchasing more cigarettes on the street.

DISCUSSION

Although higher cigarette taxes have contributed to the large decrease in the adult smoking prevalence over time in New York City, from 22% in 2002 to 14% in 2010,²⁷ our findings confirm earlier research suggesting that raising taxes on cigarettes produces increases in tax-avoidant behaviors.^10,14,28 However, we also found that tax-avoidant behaviors declined with time. Following the 2002 city and state cigarette excise tax increase, which raised the per-pack price of cigarettes in New York City by $1.81, the prevalence of tax-avoidant purchasing behaviors among adult smokers reached 30% in 2003. No new taxes were introduced until 2008, and correspondingly, tax avoidance declined over that period, to 13% in 2007. Following the 2008 state excise tax increase, which raised the per-pack price by $1.25, the prevalence of tax avoidance nearly doubled, to almost 25% of smokers.

In 2008, the tax avoidance rate was higher among smokers who bought their most recent cigarette from a carton, perhaps because tax-avoidant smokers were seeking to stock up on cigarettes purchased in other jurisdictions or to minimize the number of illegal transactions by buying in bulk.¹⁹ Tax avoidance was higher among those who did not report a quit attempt in the past year. Because higher cigarette prices are known to be associated with increased quit attempts,⁵ the availability of cheap cigarettes through untaxed sources may have provided a disincentive for tax-avoidant smokers to quit.

Among all smokers, responses to the 2008 tax increase included several tax-avoidant behaviors. More than one fifth bought more cigarettes from someone on the street, slightly more than those who bought more cigarettes outside the state, on the Internet, through the mail, in a duty-free shop, or on an Indian reservation. Indeed, since 2005, the most common source of untaxed cigarettes in New York City has been people on the street, underscoring the importance of street vendors as distributors of untaxed cigarettes.

These findings led us to examine street purchases of cigarettes in response to the 2008 tax increase and to identify sociodemographic and smoking characteristics associated with this behavior. Black and Hispanic smokers were more likely than Whites, Asians, and members of other racial groups to purchase more cigarettes on the street. Smokers younger than 45 years and smokers classified as tax avoidant by their most recent cigarette purchase were also more likely than other respondents to purchase more cigarettes on the street in response to the 2008 tax hike.

Our population-based data set prevented us from examining the association of these characteristics with street purchases of untaxed cigarettes in greater detail. However, these findings update Frieden et al.’s study of New York City’s 2002 cigarette tax increase, which found that 12% of all tax-avoidant purchases following the tax hike were from another person and that Whites were less likely than other racial/ethnic groups to purchase cigarettes from another person.⁸ Our findings also update Shelley et al.’s study of the “$5 man,” describing the proliferation of black market cigarette street sales in New York City’s low-income neighborhoods following the 2002 tax increase.¹⁸ Street purchases of untaxed cigarettes in the city may be increasing among price-sensitive smokers.¹⁰

Tax-avoidant smokers in 2008 paid an average of only $5.48 per pack, in contrast with the $7.30 per pack paid by the general smoking population. When this difference is considered within the context of studies that have documented greater price sensitivity among smokers in lower socioeconomic classes,^4,29 and in light of our finding that low-income smokers were more likely to buy more cigarettes on the street following the 2008 tax increase, the nearly $2 difference in price may present a further potential incentive drawing price-sensitive smokers toward tax-avoidant behaviors. Youths are also more price sensitive than adults, and the availability of cheaper cigarettes may contribute to youth smoking as well.

In the past decade, national, state, and local efforts have attempted to limit trafficking from the locations examined in the CHS, including Native American reservations, online, and through the mail³⁰; yet black market cigarette sales continue. In 2005, the US Bureau of Alcohol, Tobacco, Firearms, and Explosives, along with several state attorneys general, reached an agreement with all major credit card companies and PayPal to ban credit card payment processing for Internet cigarette vendors; a similar agreement was reached with all major private delivery companies, such as UPS and FedEx.¹⁵ In 2010, Congress enacted the Prevent All Cigarette Trafficking Act,³¹ imposing significant requirements on Internet and mail-order sellers of tobacco products and banning cigarette shipping through the US Postal Service. Although New York State had passed legislation to prohibit direct deliveries of tobacco products to consumers,³² a loophole prevented the state from regulating the Postal Service until the trafficking law took effect.

New York State enacted regulations in 2011 requiring cigarette wholesalers to impose the state tax on their products before selling them to Native American reservations; at the municipal level, New York City has sued retailers on Native American reservations for selling millions of packs of cigarettes to nontribal customers, resulting in a flood of cheap cigarettes into the city.³³ Additional measures to limit black market cigarette sales are urgently needed to maximize the public health benefits of increased taxes.

Limitations

New York City’s CHS is a population-based survey that relies on self-reported data from noninstitutionalized adults with functioning residential or mobile telephones. Its cross-sectional design limited our ability to draw causal inferences across years. However, the surveys were large, were conducted in multiple languages, and provided a representative sample of New York City residents; smokers’ characteristics and behaviors correlated well with predictions from the literature.³⁴ In addition, the tax-avoidant purchasing variable was derived from the purchase location of the most recent cigarette smoked, so our measure may have underestimated the true rate. Retailers engaged in illegal sales of untaxed cigarettes were not captured as a tax-avoidant source in our data.

Because one of the response options for the question assessing smokers’ responses to the 2008 tax increase listed multiple purchase locations (i.e., outside the state, on the Internet, through the mail, at a duty-free shop, or from an Indian reservation), we were unable to examine the proportion of cigarettes purchased from the individual tax-avoidant sources in this data set. Finally, New York City is a unique urban environment with high population density. Our findings regarding street sales of cigarettes may not be applicable in suburban and rural areas of the country, where tax-avoidant purchases are more likely to come from other sources explored in our study.

Because of the prominence of street sales of cigarettes as a source of untaxed cigarettes, we focused on this variable and did not explore in detail sales of cigarettes via the Internet, duty-free outlets, other states, and Indian reservations. However, in New York State, Hyland et al. found that smokers buying lower-priced cigarettes from Indian reservations were less likely to make a quit attempt,¹³ and in New Jersey, Kim et al. found that smokers who purchased cigarettes online were primarily motivated by lower prices and might smoke more cigarettes when they were purchased this way.³⁵ The overlap between Internet cigarette vendors and Indian reservation sales was identified by the New York State Department of Health in 2010¹⁴ and was explored in a 2012 article.³⁶ We intend to consider how best to explore these additional sources of lower-priced cigarettes.

Conclusions

Most studies of tax avoidance in New York State and elsewhere have examined the effect of higher cigarette excise taxes on the proliferation of untaxed cigarettes from a variety of sources.^10,12,14 Ours was among only a few to focus on illegal street purchases of untaxed cigarettes. Determining the overall proportion of cigarettes purchased from tax-avoidant sources in New York City is an avenue for future study. In addition, more quantitative research is needed to understand the association between the sociodemographic characteristics and smoking behaviors of tax-avoidant smokers identified here and the growth in illegal street purchases of untaxed cigarettes in the city. Qualitative research would also be helpful in exploring tax-avoidant behaviors and related attitudes in more depth. Together, such studies would provide a more comprehensive understanding of tax-avoidant smokers in New York City, including behaviors surrounding illegal street purchases of untaxed cigarettes.

In 2013, New York City along with other organizations, filed a Citizen Petition with the Food and Drug Administration urging the establishment of a national track-and-trace system. A national track-and-trace system would prevent untaxed cigarettes from compromising the public health benefit of cigarette excise taxes.³⁷ By allowing cigarette packs to be tracked from manufacturer to retailer, the system would prevent diversion of cigarettes into black markets, complementing the efforts of the city and state to limit illegal trafficking. New York City has the highest per-pack cigarette price in the nation, which has likely contributed to the large reductions in smoking prevalence in the city.³⁸ However, among smokers, tax-avoidant behaviors have risen, particularly among price-sensitive smokers. Tax increases should be paired with enforcement strategies that limit the flow of untaxed, cheap cigarettes into jurisdictions with high cigarette pack prices.