Abstract
Backround:
The availability of contraband cigarettes provides incentives for price-sensitive smokers to reduce their monetary costs of smoking. The objectives of this study were to examine whether Canadian smokers’ geographic proximity to First Nations reserves and attempts to quit smoking influenced the likelihood of purchasing lower-cost cigarettes from reserves.
Methods:
Data were from the International Tobacco Control (ITC) Canada Survey, a prospective survey of Canadian adult smokers conducted from 2002-2014 using telephone and online interviewing methods. Analysis was restricted to smokers from Ontario (n=2,105) and Quebec (n=1,427) participating in at least one survey wave. Smokers’ postal codes were used to calculate distance to the nearest reserve. Weighted logistic generalised estimating equations (GEE) regression examined the linear relationship between distance and the log odds of last purchasing cigarettes on reserve in each province. GEE models also examined the relationship between past-year quit attempts and the log odds of on-reserve purchasing.
Results:
Controlling for other factors, from 2002–2014, smokers from Ontario who lived 10 km closer to reserves than otherwise similar smokers had significantly higher odds of last purchasing on reserve (OR ranged from 1.16 to 1.65). Distance had little effect on smokers’ purchasing behaviours in Quebec. Moreover, in Ontario, for every 10 km increase in distance, smokers who did not try to quit had significantly greater odds of purchasing from a reserve than smokers who tried to quit (p=0.002).
Conclusion:
In order for tobacco taxation policies to achieve their maximal benefit, governments must limit potential sources of lower-cost cigarettes. Collaborative governance arrangements can ensure tobacco products sold on reserve to non-Indigenous people are appropriately taxed while allowing First Nations communities to keep the revenue generated by such taxes.
Keywords: longitudinal, tobacco, taxation, contraband, purchasing, economics, proximity
1. Introduction
Tobacco taxes reduce consumption by increasing the price of tobacco products (Chaloupka, Straif, & Leon, 2011; Chaloupka, Yurekli, & Fong, 2012; International Agency for Research on Cancer, 2011). In the face of tax increases, price-sensitive smokers may actively search for lower-priced alternatives to reduce their financial burden of continuing to smoke. Price-sensitive smokers lower their costs using a variety of tactics, including cross-border shopping, internet orders, and duty-free shops (U.S. National Cancer Institute and World Health Organization, 2016). In North America, it is also possible for smokers to purchase cigarettes from Indigenous people on tribal lands, which are usually known as First Nations reserves in Canada and Indian reservations in the United States. Cigarettes sold on tribal lands cost significantly less than those sold off tribal lands (Luk et al., 2009; Mecredy, Diemert, Callaghan, & Cohen, 2013; Soulakova, Pack, & Ha, 2018). While non-Indigenous people buying cigarettes from First Nations reserves in Canada must pay all applicable federal and provincial taxes, this is not strictly enforced (Luk et al., 2009; Mecredy et al., 2013).
The manufacture and distribution of contraband tobacco has been especially problematic in the Canadian provinces of Ontario and Quebec, despite these provinces having the lowest tobacco taxes in Canada. In these provinces, some First Nations communities legally manufacture cigarettes. For example, Grand River Enterprises (GRE), located on the Six Nations reserve near Brantford, Ontario, is the fourth largest tobacco company in Canada (Non-Smokers’ Rights Association, 2012). While GRE cigarettes can be sold legally to Indigenous people on reserve, GRE has been suspected of shipping a larger supply of cigarettes to reserves than could be smoked by the populations living on those reserves (Non-Smokers’ Rights Association, 2012, 2015; Sarson, 2019). The Royal Canadian Mounted Police (RCMP) has documented that cigarettes produced on reserve for purchase by Indigenous people are instead sold illegally to non-Indigenous people (Royal Canadian Mounted Police, 2008, 2011). In addition, the Akwesasne reserve near Cornwall, Ontario, straddles both provincial and international borders, making it difficult for provincial and federal agencies to police the illegal distribution of contraband tobacco by organised criminal networks (Leuprecht, 2016). These factors increase the availability of untaxed and under-taxed cigarettes for smokers living in Ontario and Quebec.
Data from the 2005–2006 Ontario Tobacco Survey indicate that almost 26% of adult smokers in Ontario purchased cigarettes from First Nations reserves in the previous six months while 12% of smokers reported that they usually bought cigarettes on reserve (Luk et al., 2009). Using data from nine waves of the International Tobacco Control (ITC) Canada Survey, Guindon, Burkhalter, and Brown (2017) showed that on-reserve purchases were more prevalent in Ontario than Quebec from 2004 to 2014 (Guindon et al., 2017). During that time period, the percentage of smokers from Ontario whose last purchase of cigarettes was from a reserve increased from 5% in 2005 to 23% in 2013/2014. Less than 10% of smokers from Quebec made such purchases over the entire study period (Guindon et al., 2017).
Measures have been introduced to address the Canadian contraband tobacco problem. In 2001, the Government of Quebec introduced ACCES Tabac, a partnership between provincial and federal agencies designed to reduce the supply of contraband tobacco through concerted enforcement (Leuprecht, 2016, 2018). In 2008, the RCMP launched the Contraband Tobacco Enforcement Strategy to further limit the supply of illegal tobacco (Stanbrook, 2013). Each of these programs may have had some effect in reducing the availability of contraband tobacco (Sen, 2017).
In spite of increased enforcement, the continued availability of low- and untaxed cigarettes enables price-sensitive smokers to minimise their financial costs of continuing to smoke (DeCicca, Kenkel, & Liu, 2015; Xu, Malarcher, O’Halloran, & Kruger, 2014). Previous research has found that proximity to tribal lands in the United States was associated with the purchase of low- and untaxed cigarettes from those lands (DeCicca et al., 2015; Hyland et al., 2005). In turn, price minimisation behaviours disincentive smoking cessation efforts. Mecredy et al. (2013) found that smokers who usually smoked a contraband cigarette brand were significantly less likely to attempt to quit smoking than smokers of premium or discount brands (Mecredy et al., 2013).
Building on these findings, the current study examined whether smokers’ potential access to contraband cigarettes, defined as geographic proximity to First Nations reserves, increased the likelihood of purchasing cigarettes from reserves. Because the Canadian contraband tobacco problem is concentrated in Ontario and Quebec, this study also examined whether on-reserve purchases differed by province and whether the likelihood of on-reserve purchases changed over time as a function of geographic proximity. Finally, this study examined whether past-year quit attempt attempts were associated with on-reserve purchases over and above the effect of geographic proximity.
2. Methods
2.1. Data Sources
This study used data from the first nine waves of the International Tobacco Control Canada Survey. The ITC Canada Survey was part of a larger prospective cohort survey of nationally representative samples of smokers from Canada, the United States, the United Kingdom, and Australia, conducted from 2002 (Wave 1) to 2014 (Wave 9). In Wave 1 of the ITC Canada Survey, 2,214 adult smokers aged 18 and older were randomly sampled from 14 geographic strata defined by each of the 10 Canadian provinces. The largest urban centres in British Columbia (Vancouver), Ontario (Toronto and the Greater Toronto Area), and Quebec (Montreal) comprised additional sampling strata in those provinces (Thompson et al., 2006).
In the first six waves, respondents were interviewed by computer assisted telephone interviewing (CATI). In the remaining waves, both CATI and computer assisted web interviewing were used. In all follow-up waves, respondents lost to attrition were replaced with new respondents using the original sampling design. Respondents who quit smoking were also followed over time. Approximately 2,000 respondents were interviewed in each of the first six waves; the sample size was reduced to approximately 1,750 respondents in Wave 7 and 1,500 respondents in each of Waves 8 and 9 (ITC Project, 2011; Thompson et al., 2006). In the first three waves, data were collected during a single calendar year, while in all remaining waves, data collection spanned two calendar years. All respondents provided informed consent to participate in the ITC Canada Survey; ethical approval was provided by the Office of Research Ethics, University of Waterloo (ORE #10556, #12978, #17469).
Overall, 3,611 current smokers from Ontario and Quebec participated in at least one wave of the ITC Canada Survey. Of these, five respondents moved from Ontario to Quebec over the course of the study and were excluded from the analysis. Of the remaining respondents, 3,562 (98.8%) provided sufficient postal code information to identify approximate residential locations. The statistical analysis was based on 3,532 smokers who provided complete data for all covariates used in the generalised estimating equation (GEE) regression models specified below. These smokers contributed 9,063 observations to the analysis. Location information for almost all of these observations (9,042 or 99.8%) was based on postal code. Geographic coordinates for these postal codes were obtained from geospatial data files supplied by DMTI Spatial through Scholars GeoPortal (DMTI Spatial Inc., 2014; Ontario Council of University Libraries, 2013). Approximate locations for the remaining 21 observations were based on the geographic centroid of the forward sortation area (FSA) in which respondents lived using data from Statistics Canada (Statistics Canada, 2011). These data were also obtained through Scholars GeoPortal.
2.2. Measures
2.2.1. Purchasing cigarettes from a First Nations reserve
The primary outcome measure for this study was whether respondents made their last purchase of cigarettes from a First Nations reserve. Purchase locations were assessed using responses to the question “Where did you last buy cigarettes/roll-your-own tobacco for yourself?” One of the possible responses was “from a First Nations reserve”. Additional open-ended responses to a second question identified other respondents who reported they last purchased cigarettes on reserve (e.g., “Native smoke shop” “the reserve” “Six Nations reserve”).
2.2.2. Distance to First Nations reserves
The main explanatory variable was geographic distance between respondents’ residential locations and the closest First Nations reserve. The geographic coordinates of residential locations were identified using the postal code of each respondent’s home address, as outlined above. Separate geospatial data identifying the locations of First Nations reserves in Ontario and Quebec were obtained from Natural Resources Canada (Natural Resources Canada, 2017). These data files identified the administrative boundaries of all First Nations reserves in those provinces (one file per province). These separate files were merged into a single geospatial layer prior to extracting geographic centroids from the polygons identifying administrative bounders using QGIS Version 2.18. In order to compute the Euclidean distance from respondent locations to the nearest First Nations reserve, it was necessary to re-project each geospatial data layer to the North American Equidistant Conic projection (EPSG:102010) to minimise distance distortions. The distance from residential locations to the nearest First Nations reserve was then computed using QGIS.
2.2.3. Self-reported Cigarette Prices
Smokers also reported the price they paid for their last purchase of cigarettes. For smokers who last purchased from a reserve, the reported purchase price corresponds to an on-reserve purchase. Smokers reported the amount of cigarettes purchased for carton, pack, or loose (single cigarette) purchases. Smokers then reported the total price paid for their entire purchase. In Waves 3 through 6 and in Wave 9, smokers purchasing on reserve could also report the number of cigarettes purchased on reserve, often by a “bag” of 200 cigarettes. Price per cigarette (total purchase price divided by total cigarettes purchased) was computed for each smoker and then converted to a standardized price per “pack” of 25 cigarettes. Nominal prices were adjusted for inflation to 2014 Canadian dollars (last quarter of 2014, where $1 CAD = $0.862 USD (Organization for Economic Co-operation and Development, 2015, 2016)). Inflation-adjusted pack prices below $1/pack or greater than $20/pack were considered outliers and treated as missing for the analysis. Overall, 73 prices were treated as outliers across all waves (0 to 8 observations per wave, with the exception of Wave 2 where 39 observations were considered outliers).
2.2.4. Covariates
Both sociodemographic measures and smoking behaviours were included as covariates in regression models. Sociodemographic measures were sex, age group (18–24, 25–39, 40–54, 55+), ethnicity (First Nations, other, white), annual household income (low: <$30,000; moderate: $30,000–$59,999; high: ≥ $60,000; income not reported), and highest level of education (low: completed high school or less, moderate: trade school/community college/some university, high: completed university/post-graduate degree).
Measures of smoking behaviors were smoking status (daily, non-daily smoker), amount smoked per day (≤10, 11-20, 21–30, ≥ cigarettes/day), exclusive use of roll-your-own (RYO) cigarettes (exclusive use, smokes factory-made cigarettes alone or in combination with RYO cigarettes), intentions to quit smoking in the next 6 months (no plans, any plans), and attempts to quit in the previous year. Attempts to quit was computed from responses to questions identifying whether a respondent had ever tried to quit, when their last quit attempt started, or when their last quit attempt ended. Respondents who ever tried to quit and whose last attempt either started or ended in the previous 365 days were classified as having made an attempt to quit smoking in the previous year. Finally, since some respondents were surveyed multiple times over the course of the study, and because previous participation in the survey is known to influence respondent’s answers to important outcomes, an additional covariate measuring each respondent’s time-in-sample was also included in all regression models (Driezen & Thompson, 2011; Thompson, Boudreau, & Driezen, 2005).
2.3. Statistical Analysis
Unweighted descriptive statistics were used to characterise the sample of smokers participating in the ITC Canada Survey. The average distance to First Nations reserves was then estimated for two groups of smokers: those who last purchased cigarettes from a reserve and those who did not. Mean distances were estimated within each province for each survey wave using the sampling weights; variances were estimated using Taylor series linearization methods. Mean self-reported cigarette prices per standardized “pack” of 25 cigarettes were estimated in a similar fashion.
Weighted logistic regression was then used to estimate the adjusted prevalence of last purchasing cigarettes from a First Nations reserve. Regression models accounted for the stratified sampling design and were estimated using GEE to account for repeated measures using an exchangeable working correlation matrix. Regression models controlled for all covariates listed above. A province X wave interaction term was included to test whether the overall temporal trend differed by province. In addition, the adjusted percentages of on-reserve purchases were compared between provinces and within provinces, relative to 2002 (Wave 1). Statistical tests controlled for multiple comparisons using the false discovery rate (FDR) (Benjamini & Hochberg, 1995).
Separate GEE models were estimated for each province to estimate the linear effect of distance to nearest reserve on the log odds of last purchasing cigarettes on reserve. In these models, a wave X distance interaction effect was included to test whether the effect of distance changed over time. A final set of logistic regression models were estimated for each province to examine the association between attempts to quit in the previous year and last purchasing cigarettes on reserve. In these models, a quit attempt X distance interaction effect was specified in addition to a main effects model. SAS (version 9.4) was used for estimating descriptive statistics while SAS-callable SUDAAN (version 11.0.3) was used to estimate weighted mean distances and all GEE models.
3. Results
3.1. Sample Characteristics
Over the course of the study, 2,105 smokers from Ontario and 1,427 smokers from Quebec participated in at least one wave of the ITC Canada Survey. On average, respondents from Ontario remained in the study slightly longer than respondents from Quebec (mean time-in-sample = 2.79 waves vs. 2.50 waves, respectively). Table 1 presents the respondent characteristics based on their initial wave of recruitment. About 53% of respondents from either province were female. The age distribution of respondents was similar in each province. A slightly greater percentage of respondents from Ontario were ethnic minorities, with 3.7% of smokers reporting First Nations ancestry compared to only 1.4% of smokers in Quebec. Slightly more than one third of respondents from Ontario had high household incomes (≥ $60,000/year) compared to only one fifth of smokers from Quebec. Respondents from Ontario tended to be better educated, with 57% reporting moderate to high education levels, compared to only 48% of respondents from Quebec. While a smaller percentage of smokers from Ontario exclusively smoked RYO cigarettes compared to Quebec smokers (1.6% vs. 8.4%, respectively), a similar percentage of smokers from either province smoked on a daily basis. Daily cigarette consumption patterns were similar in each province and just over 40% of smokers reported they planned to quit smoking within the next six months.
Table 1.
Characteristics of smokers from Ontario and Quebec participating in the ITC Canada Survey (2002–2014) at the time of recruitment.
| Ontario (n = 2,105) |
Quebec (n = 1,427) |
|||
|---|---|---|---|---|
| % | (Freq.) | % | (Freq.) | |
| Sex | ||||
| Female | 52.7 | (1,110) | 54.1 | (772) |
| Male | 47.3 | (995) | 45.9 | (655) |
| Age group | ||||
| 18-24 | 11.4 | (240) | 11.8 | (169) |
| 25-39 | 28.9 | (609) | 29.8 | (425) |
| 40-54 | 37.1 | (782) | 36.4 | (520) |
| 55+ | 22.5 | (474) | 21.9 | (313) |
| Ethnicity | ||||
| First Nations | 3.7 | (77) | 1.4 | (20) |
| Other | 9.5 | (200) | 6.2 | (88) |
| White | 86.8 | (1,828) | 92.4 | (1,319) |
| Household income | ||||
| Low | 23.7 | (498) | 32.9 | (469) |
| Moderate | 32.6 | (687) | 37.9 | (541) |
| High | 35.8 | (754) | 20.4 | (291) |
| Not reported | 7.9 | (166) | 8.8 | (126) |
| Education | ||||
| Low | 43.1 | (908) | 51.6 | (737) |
| Moderate | 38.9 | (819) | 33.9 | (484) |
| High | 18.0 | (378) | 14.4 | (206) |
| Wave of recruitment | ||||
| Wave 1 | 38.7 | (815) | 35.3 | (504) |
| Waves 2-5 | 35.9 | (756) | 39.3 | (561) |
| Waves 6-9 | 25.4 | (534) | 25.4 | (362) |
| Time-in-sample* | ||||
| Mean (SD) | 2.79 | (2.09) | 2.50 | (1.98) |
| Smoking status | ||||
| Non-daily | 7.6 | (160) | 7.1 | (101) |
| Daily | 92.4 | (1,945) | 92.9 | (1,326) |
| Cigarettes/day | ||||
| 31+ | 4.4 | (92) | 4.7 | (67) |
| 21-30 | 21.3 | (449) | 25.4 | (363) |
| 11-20 | 42.5 | (895) | 40.3 | (575) |
| ≤ 10 | 31.8 | (669) | 29.6 | (422) |
| Product type | ||||
| RYO | 1.6 | (33) | 8.5 | (121) |
| FM or both | 98.4 | (2,072) | 91.5 | (1,305) |
| Plans to quit in next 6 months | ||||
| No plans | 59.6 | (1,253) | 57.5 | (820) |
| Plans to quit | 40.4 | (849) | 42.5 | (605) |
Based on last wave of participation.
3.2. On-reserve purchasing
Figure 1 presents the model adjusted percentage of smokers who last purchased cigarettes from a First Nations reserve from 2002 to 2014. The temporal trend differed significantly between provinces (p = 0.009). While on-reserve purchases were relatively rare in 2002 and 2003 in both provinces, by 2006 (Wave 4), a significantly greater percentage of smokers last purchased cigarettes on reserve in Ontario compared to Quebec (FDR p < 0.001). For the remainder of the study, smokers from Ontario were significantly more likely to make their last purchase of cigarettes on reserve than smokers from Quebec (all FDR p < 0.001).
Figure 1.
Adjusted percentage of smokers last purchasing cigarettes from a First Nations reserve in Ontario and Quebec from 2002 to 2014 (n = 3,532; *p < 0.001; †p < 0.01; and ‡p < 0.05 test the percentage in the given year against the baseline year within province, controlling for multiple comparisons using the false discovery rate). Shaded areas represent the 95% confidence interval. See Supplementary Table A1 for all GEE parameter estimates.
In Ontario, the prevalence of on-reserve purchasing was significantly higher in 2006 than in 2002 (FDR p < 0.001). On-reserve purchasing remained significantly higher in subsequent years compared to 2002 (all FDR p < 0.001). In Quebec, however, the prevalence of last purchasing on reserve peaked in 2006 and 2008 at 5.6%. By 2010, only 3.1% of smokers reported last purchasing on reserve, similar to the baseline percentage of 1.4% (FDR p = 0.188).
3.3. Geographic proximity and on-reserve purchasing
In both provinces, smokers who did not make their last purchase of cigarettes from a First Nations reserve lived further away from a reserve, on average, compared to smokers who did (Table 2). This was true across all survey waves. Among smokers who made their last purchase on reserve, the average distance between smokers’ residences and the nearest reserve increased from Wave 1 (2002) to Wave 6 (2007/2008). This suggests that smokers living further away from reserves at the time of Wave 6 were more inclined to purchase cigarettes on reserve. This increase in average distance corresponds to the peak prevalence of on-reserve purchasing in both provinces. After 2008, the average distance between smokers’ residences and the nearest reserve decreased among smokers who made their last purchase on reserve, suggesting that smokers living further away from a reserve were less inclined to purchase cigarettes from a First Nations reserve after 2008.
Table 2.
Average distance (in km) between smokers’ residential locations and the nearest First Nations reserve and average self-reported price paid per standardized “pack” of 25 cigarettes* by province and last purchase type (on/off reserve).
| Ontario |
Quebec |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| On reserve | Off reserve | On reserve | Off reserve | |||||||||
| Wave (Year) | n | Mean | (SE) | n | Mean | (SE) | n | Mean | (SE) | n | Mean | (SE) |
| Distance (km) to Nearest Reserve | ||||||||||||
| 1 (2002) | 21 | 27.6 | (4.09) | 789 | 48.4 | (0.87) | 5 | 14.4 | (2.96) | 496 | 37.3 | (1.47) |
| 2 (2003) | 29 | 27.0 | (3.55) | 739 | 47.5 | (0.91) | 5 | 10.8 | (3.16) | 426 | 38.9 | (1.73) |
| 3 (2004) | 42 | 25.2 | (3.17) | 669 | 46.3 | (0.97) | 15 | 20.4 | (4.65) | 421 | 36.4 | (1.69) |
| 4 (2005/2006) | 77 | 29.9 | (2.53) | 587 | 47.7 | (1.07) | 23 | 38.8 | (5.13) | 362 | 33.6 | (1.45) |
| 5 (2006/2007) | 112 | 35.3 | (2.53) | 519 | 47.5 | (1.14) | 21 | 31.5 | (4.61) | 409 | 34.3 | (1.50) |
| 6 (2007/2008) | 137 | 37.0 | (2.29) | 481 | 47.7 | (1.20) | 24 | 37.0 | (6.59) | 364 | 34.5 | (1.57) |
| 7 (2008/2009) | 126 | 34.6 | (2.54) | 416 | 47.4 | (1.24) | 15 | 35.1 | (5.72) | 318 | 35.4 | (1.83) |
| 8 (2010/2011) | 109 | 29.0 | (2.18) | 339 | 44.9 | (1.52) | 8 | 24.9 | (3.43) | 249 | 35.0 | (1.94) |
| 9 (2013/2014) | 90 | 27.6 | (2.31) | 363 | 46.7 | (1.35) | 5 | 27.0 | (5.04) | 252 | 38.8 | (2.73) |
| Average Price Paid/Pack* | ||||||||||||
| 1 (2002) | 17 | 6.22 | (0.99) | 763 | 8.03 | (0.05) | 4 | 4.50 | (0.59) | 443 | 7.83 | (0.06) |
| 2 (2003) | 22 | 5.17 | (0.45) | 697 | 9.04 | (0.05) | 5 | 7.43 | (1.61) | 340 | 8.74 | (0.11) |
| 3 (2004) | 39 | 5.54 | (0.27) | 639 | 9.31 | (0.07) | 13 | 4.95 | (0.38) | 378 | 8.69 | (0.09) |
| 4 (2005/2006) | 64 | 4.56 | (0.24) | 572 | 9.28 | (0.08) | 14 | 3.18 | (0.26) | 324 | 8.63 | (0.10) |
| 5 (2006/2007) | 97 | 3.72 | (0.22) | 491 | 9.40 | (0.11) | 16 | 2.99 | (0.48) | 364 | 8.60 | (0.10) |
| 6 (2007/2008) | 112 | 3.38 | (0.18) | 443 | 9.25 | (0.12) | 12 | 2.06 | (0.25) | 317 | 8.55 | (0.12) |
| 7(2008/2009) | 97 | 3.15 | (0.21) | 387 | 9.40 | (0.15) | 8 | 2.05 | (0.43) | 277 | 8.73 | (0.12) |
| 8 (2010/2011) | 90 | 4.05 | (0.16) | 315 | 9.67 | (0.13) | 5 | 4.16 | (0.87) | 213 | 8.19 | (0.13) |
| 9 (2013/2014) | 82 | 4.10 | (0.19) | 319 | 9.56 | (0.16) | 5 | 2.28 | (0.49) | 206 | 8.68 | (0.15) |
All prices are adjusted for inflation to 2014 Canadian dollars, where 1 CAD = 0.862 USD in the fourth quarter of 2014.
Table 2 also presents the average self-reported price paid for a standardized “pack” of 25 cigarettes among smokers who last purchased cigarettes on or off reserve. Prices were based on smokers’ last purchase and reflect real prices adjusted to 2014 CAD. In Ontario, smokers who last purchased on reserve paid anywhere from $1.81 to $6.25 less per pack over the course of the study compared to smokers purchasing off reserve. In Quebec, smokers purchasing on reserve faced similar cost savings, ranging from $1.31 to $6.68 per pack.
Separate logistic regression models were then estimated within each province to examine the effect of distance to the nearest reserve on the log odds of having last purchased cigarettes on reserve (Figure 2). Province-specific models controlled for sociodemographic covariates and smoking characteristics. A wave X distance interaction effect was also included to test whether the effect of distance varied over time. There were significant interaction effects in both provinces, even after controlling for other factors (p = 0.005 in Ontario; p = 0.016 in Quebec).
Figure 2.
Adjusted odds of last purchasing cigarettes from a First Nations reserve in Ontario (n = 2,105) and Quebec (n = 1,427) from 2002 to 2014. Odds ratios are adjusted for all factors shown, as well as the main effects of time-in-sample, distance, and survey wave (not shown). See Supplementary Table A2 for all GEE parameter estimates.
In Ontario, a 10 km reduction in proximity to a First Nations reserve was associated with significantly increased odds of having last purchased cigarettes on reserve. While there was an association between proximity and on-reserve purchasing in every survey wave, the weakest effect was observed in Wave 6, when a 10 km reduction in proximity was associated with only 16% greater odds (aOR = 1.16; 95% CI: 1.05–1.28) of last purchasing on reserve (Figure 2). This compares to adjusted odds ratios of 1.55 in Wave 1, 1.40 in Wave 8 and 1.46 in Wave 9. These findings suggest that in 2007/2008, smokers were willing to travel greater distances to purchase cigarettes on reserve than at other times.
This result is corroborated in Figure 3, which presents the adjusted predicted marginal probability of last purchasing cigarettes on reserve as a function of proximity. From Wave 1 to Wave 3, there was only a small probability (< 5%) of purchasing on reserve among smokers living more than 25 km from a reserve in either province. By Wave 6, however, 25% of Ontario smokers living 25 km from a reserve were predicted to have last purchased cigarettes on reserve while as many as 14% of Ontario smokers living 75 km from a reserve were predicted to have done so. By Waves 8 and 9, a smaller percentage of Ontario smokers living further from a First Nations reserve were predicted to have last purchased on reserve (25% and 21% of smokers living 25 km from a reserve, respectively, and only 6% and 4% of smokers living 75 km from a reserve, respectively). Thus, in later waves, Ontario smokers living further away from a First Nations reserve were less likely to have last purchased cigarettes on reserve.
Figure 3.
Adjusted predicted marginal probability of smokers purchasing cigarettes from a First Nations reserve as a function of distance to the nearest reserve in Ontario and Quebec from 2002 to 2014 (n = 3,532).
Different effects were observed in Quebec. In that province, proximity to reserves was associated with purchasing on reserve only in Waves 1, 2, and 8, where a 10 km reduction in distance was associated with 1.38 to 4.19 times the odds of last purchasing on reserve (Figure 2). In the remaining waves, distance was not associated with purchasing on reserve. Figure 3 demonstrates how the predicted probability of purchasing on reserve remained relatively constant irrespective of proximity to the nearest reserve.
Across all waves, other factors besides proximity were associated with last purchasing on reserve. In Ontario, income, cigarettes smoked/day, and intentions to quit smoking were associated with last purchasing cigarettes on reserves (all p < 0.01, Figure 2). In that province, younger smokers had significantly lower odds of last purchasing on reserve compared to smokers aged 55 and older. Low- (aOR = 2.20; 95% CI: 1.57–3.08) and moderate- (aOR = 1.79; 95% CI: 1.33–2.40) income smokers had significantly higher odds of last purchasing on reserve compared to high-income smokers as did smokers who smoked more than 11 cigarettes per day (vs. 10/less). Smokers not planning to quit smoking in the next 6 months also had higher odds of last purchasing cigarettes on reserve relative to smokers who planned to quit in the next 6 months (aOR = 1.36; 95% CI: 1.13–1.65). In Quebec, other factors associated with last purchasing cigarettes on reserve were sex (p = 0.002), ethnicity (p = 0.045) and education (p = 0.035). Across all waves, male smokers had significantly greater odds of purchasing on reserve compared to female smokers (aOR = 2.34; 95% CI: 1.35–4.06). In addition, smokers of First Nations ancestry had significantly greater odds of purchasing on reserve compared to white smokers (aOR = 6.11; 95% CI: 1.46–25.51) as had the least educated smokers relative to the most educated smokers (aOR = 3.08; 95% CI: 1.10–8.56).
3.4. Attempts to quit smoking and on-reserve purchasing
The final set of logistic GEE models examined whether past-year attempts to quit smoking were associated with purchasing cigarettes from First Nations reserves. Controlling for other factors, smokers from Quebec who did not try to quit in the previous year were no more likely to purchase cigarettes from a reserve than smokers who did try to quit (p = 0.550, Table 3). In Ontario, however, there was a significant interaction between proximity and past-year quit attempts (p = 0.002). Across all waves, the odds of purchasing from reserves decreased significantly the further smokers lived from reserves. For every 10 km increase in distance to the nearest reserve, smokers who did not try to quit had 0.79 times the odds of last purchasing on reserve (95% CI: 0.74, 0.85) while smokers who did try to quit had lower odds (aOR = 0.67 for a 10 km increase; 95% CI: 0.60, 0.74). Put another way, distance was less of a barrier for smokers who did not try to quit relative to those who did.
Table 3.
Odds of last purchasing cigarettes on reserve in Ontario* and Quebec† across all survey waves as a function of covariates and attempts to quit smoking in the previous year.
| Ontario (n=2,105) |
Quebec (n=l,427) |
|||||
|---|---|---|---|---|---|---|
| Covariate (reference) | OR | (95% CI) | p‡ | OR | (95% CI) | p‡ |
| Sex (female) | ||||||
| Male | 1.10 | (0.83, 1.46) | 0.517 | 2.32 | (1.33, 4.04) | 0.003 |
| Age group (55+) | ||||||
| 18-24 | 0.40 | (0.21, 0.73) | 0.014 | 0.25 | (0.07, 0.92) | 0.177 |
| 25-39 | 0.65 | (0.43, 0.97) | 0.65 | (0.28, 1.50) | ||
| 40-54 | 0.83 | (0.59, 1.17) | 0.93 | (0.46, 1.87) | ||
| Ethnicity (White) | ||||||
| First Nations | 1.00 | (0.47, 2.14) | 0.672 | 7.37 | (1.77, 30.69) | 0.022 |
| Other | 0.74 | (0.38, 1.44) | 1.27 | (0.48, 3.34) | ||
| Income (high) | ||||||
| Low | 2.29 | (1.64, 3.19) | < 0.001 | 1.29 | (0.56, 2.98) | 0.78 |
| Moderate | 1.75 | (1.31, 2.36) | 1.42 | (0.66, 3.06) | ||
| Not reported | 1.53 | (0.95, 2.46) | 1.63 | (0.56, 4.79) | ||
| Education (high) | ||||||
| Low | 0.99 | (0.65, 1.52) | 0.442 | 2.92 | (1.03, 8.33) | 0.058 |
| Moderate | 1.20 | (0.78, 1.85) | 1.56 | (0.48, 5.11) | ||
| Smoking status (daily) | ||||||
| Non-daily smoker | 0.63 | (0.37, 1.07) | 0.086 | 1.16 | (0.36, 3.78) | 0.805 |
| Smokes RYO/FM (FM/both) | ||||||
| Smokes RYO | 1.19 | (0.62, 2.27) | 0.608 | 1.71 | (0.90, 3.23) | 0.101 |
| Cigarettes/day (≤ 10) | ||||||
| 31+ | 2.76 | (1.71, 4.45) | < 0.001 | 3.24 | (0.99, 10.62) | 0.151 |
| 21-30 | 1.80 | (1.28, 2.52) | 1.84 | (0.83, 4.06) | ||
| 11-20 | 2.05 | (1.54, 2.74) | 2.18 | (1.07, 4.41) | ||
| Quit intentions (within 6 months) | ||||||
| No plans to quit | 1.23 | (1.02, 1.50) | 0.032 | 1.35 | (0.85, 2.17) | 0.207 |
| Past-year attempts to quit (any attempt) | ||||||
| No attempt to quit | N/A | 1.16 | (0.71, 1.88) | 0.550 | ||
| Distance to reserve | ||||||
| 10 km increase | N/A | 0.90 | (0.81, 1.00) | 0.045 | ||
| Past-year quit attempts X distance | ||||||
| Did not try vs. tried X 10 km increase | 1.19 | (1.15, 1.22) | 0.002 | N/A | ||
| 10 km increase, if tried to quit | 0.67 | (0.60, 0.74) | < 0.001 | N/A | ||
| 10 km increase, if did not try | 0.79 | (0.74, 0.85) | < 0.001 | N/A | ||
Notes:
Separate GEE logistic regression models were estimated for Ontario and Quebec. Odds ratios for time-in-sample and survey wave were excluded for brevity. For the Ontario sample, the past year quit attempts X distance interaction effect was statistically significant (p = 0.002). See Supplementary Table A3 for all GEE parameter estimates.
For the Quebec sample, estimated odds ratios were from a main effects model only. The past year quit attempts X distance interaction effect was not statistically significant (p = 0.077). See Supplementary Table A3 for all GEE parameter estimates.
Wald χ2 omnibus test
N/A = not applicable
4. Discussion
The results of this longitudinal study of contraband purchasing among provincially representative samples of smokers from Ontario and Quebec demonstrate that potential access to low- and untaxed sources of cigarettes influences smokers’ purchasing behaviours. While the purchase of contraband cigarettes from First Nations reserves peaked in both provinces around 2008, such purchases were more prevalent among Ontario smokers from 2006 to 2014 compared to Quebec smokers. Moreover, potential access, as defined by the geographic distance between smokers’ residences and the nearest reserve, was consistently associated with a significantly increased likelihood of Ontario smokers purchasing contraband cigarettes from a reserve. While geographic proximity played almost no role in the purchasing behaviours of smokers from Quebec, the consistent effects over time among Ontario smokers corroborate existing cross-sectional research from the United States demonstrating a link between geographic proximity to low- and untaxed sources of cigarettes and purchasing behaviours (DeCicca et al., 2015; Hyland et al., 2005; Xu et al., 2014). In Ontario, potential savings of as much as $6.25 CAD per pack ($5.39 2014 USD) may more than offset any additional costs incurred by travelling to nearby reserves.
Over the course of the study, there were important differences between Ontario and Quebec in attempts to limit the availability of contraband tobacco. As early as 2001, Quebec intraduced ACCES Tabac, a program designed to limit the supply of contraband tobacco through concerted enforcement. Additional enforcement strategies implemented by the RCMP in 2008 may have helped maintain the relatively low prevalence of contraband purchasing in Quebec. As Sen (2017) notes, increased enforcement may have had some effect in reducing the availability of contraband tobacco in Quebec, which may partly explain the lack of a consistent association between geographic proximity and on-reserve purchasing in that province. These findings underscore the importance of minimizing the potential availability of contraband tobacco, thereby eliminating sources of low- and untaxed cigarettes for price-sensitive smokers who actively seek out lower-cost cigarettes.
On-reserve purchases peaked in 2008 in both provinces. At this time, smokers from Ontario were willing to travel greater distances to purchase cigarettes on reserve. Interestingly, the timing of these factors coincided with the beginning of the global financial crisis. While the effects of the economic downturn may not have yet been felt by smokers from Ontario at the beginning of the recession, it is possible that price-sensitive smokers attuned to prevailing economic conditions might be more inclined to seek out lower-cost cigarettes. If so, smokers might be willing to travel greater distances to obtain cheaper cigarettes, especially if the price of contraband products offset any additional travel costs.
Attempts to quit smoking were also associated with last purchasing cigarettes on reserve, at least in Ontario. In that province, there was a significant interaction between proximity to the nearest reserve and whether smokers tried to quit smoking in the previous year. Specifically, living further from a reserve was less of a deterrent to purchasing from a reserve for smokers who did not actively try to quit in the previous year compared to smokers who did. It is plausible that smokers who did not actively try to quit in the previous year were more willing to travel further distances to purchase lower-cost cigarettes.
In light of these findings, additional research is needed to better understand how potential access to lower-cost cigarettes influences smokers’ purchasing behaviours. In this study, geographic distance to First Nations reserves was used as a measure of potential accessibility to lower-cost cigarettes. However, the way in which smokers navigate their environments on a daily basis may further influence their purchasing behaviours. For example, while some smokers may live near sources of lower-cost cigarettes, if their daily travel patterns do not bring them into areas where they can readily purchase lower-cost cigarettes, they may not make such purchases. On the other hand, smokers who are less interested in quitting may be more inclined to travel longer distances to purchase lower-cost cigarettes. Novel activity-space research developed in the field of time geography may identify how smokers travel through their environments on a daily basis and whether specific travel behaviours influence cigarette purchasing behaviours (Rainham, McDowell, Krewski, & Sawada, 2010; Wang & Kwan, 2018). Such approaches may also consider differences between smokers such as levels of addiction and willingness to quit and whether the effects of typical travel patterns are moderated by such differences. This approach has implications for the purchase of other substances including alcohol and cannabis.
4.1. Strengths and limitations
This study relied on a unique source of longitudinal data to assess temporal trends in the purchase of contraband tobacco from First Nations reserves among smokers from two different Canadian provinces where the use of contraband tobacco is most prevalent. As a result, this study could assess differences in trends between provinces. It was also possible to examine how potential access to contraband influences smokers’ purchasing decisions. Geographic proximity was measured consistently over the entire study period for all survey respondents and used as a means to assess potential access to contraband cigarettes under the assumptions that smaller distances between smokers’ residences and the nearest First Nations reserve represented greater accessibility to low- and untaxed sources of cigarettes. In Ontario, smokers’ proximity to reserves consistently increased the odds of purchasing contraband cigarettes from reserves over the course of this 12-year study.
That said, this study has some limitations that must be considered. First, distance was treated as a continuous covariate in GEE models which assumes a monotonic linear relationship between distance and the likelihood of last purchasing cigarettes on reserve. Since this is a strong assumption, a sensitivity analysis was conducted to examine the effect of distance as a categorical measure using different classification schemes (supplementary Table A4). In Ontario, the adjusted predicted marginal probability of last purchasing on reserve decreased as distance to the nearest reserve increased. This was true irrespective of the classification scheme used. In Quebec, distance had little effect on the probability of last purchasing on reserve. In both provinces, the overall main effect of distance under each classification scheme was broadly consistent with the wave-specific effects displayed in Figure 3. Thus, while a linear relationship is a strong assumption, it seems reasonable for these data.
Second, not all respondents provided valid postal code information; as a result, some smokers were excluded from the analysis. However, valid postal code information was provided by almost 99% of respondents, therefore any bias resulting from the exclusion of these smokers should be minimal.
Another bias to consider is that postal codes were used to identify smokers’ residential locations in order to compute the geographic distance between smokers and the nearest reserve. In Canada, postal codes are only approximate indicators of residential location. Moreover, postal codes are more accurate in urban areas than rural areas, because rural postal codes span much larger land areas than urban postal codes. While 80% of smokers included in this study were from urban areas, including an urban/rural covariate in GEE models estimating the effect of distance on purchasing from a First Nations reserve had no discernible influence on estimated odds ratios.
Another factor possibly affecting the accuracy of computed distances is that they were computed to the centroid of the nearest First Nations reserve, not to the exact location where cigarettes were purchased. Furthermore, distances were approximate because they were calculated as Euclidean distances, rather than actual distance travelled. What is more, distance effects were included in GEE models as linear effects while the exact functional relationship between distance and the probability of last purchasing cigarettes on reserve may be nonlinear. However, the results reported here clearly demonstrate that when contraband tobacco is readily available, as in Ontario during the study period, potential access, as measured by approximate physical distance to contraband sources, clearly influences smokers’ purchasing behaviours.
Computed distances also assume that smokers travel from their homes to reserves to purchase cigarettes. It does not account for their daily travel patterns, such as trips to work which may bring them in closer proximity to a reserve. Computed distances also assumed that contraband cigarettes were available from all reserves which likely does not reflect the true availability of contraband.
Finally, some on-reserve cigarette purchases may have been legitimate purchases. Specifically, Indigenous people, having status as per the Canadian Indian Act, may legally purchase tobacco on reserves. In order to examine whether potentially legitimate purchases of on-reserve cigarettes influenced the results, a sensitivity analysis was conducted that excluded all respondents of First Nations ancestry (n = 77 in Ontario and n = 20 in Quebec). In both provinces, these exclusions had almost no effect on the estimated odds of last purchasing contraband cigarettes on reserve as a function of distance to the nearest reserve (compare Figure 2 to Supplementary Figure A1).
5. Conclusion
In conclusion, this study demonstrates that the purchase of contraband tobacco from First Nations reserves was more prevalent among smokers from Ontario than among smokers from Quebec from 2006 to 2014. As early as 2001, the province of Quebec introduced measures to curb the availability of contraband tobacco whereas Ontario did not. The potential availability of contraband tobacco in Ontario seems to influence the purchasing behaviours of some price-sensitive smokers in that province. In particular, geographic proximity to potential sources of contraband on First Nations reserves was strongly associated with smokers’ purchasing behaviours in Ontario over a 12-year period while unrelated to smokers’ purchasing behaviours in Quebec. The effect of geographic proximity to potential sources of contraband was further moderated by attempts to quit smoking in Ontario, such that increasing distance was less of a deterrent of last purchasing on reserve among smokers who did not try to quit in the previous year compared to smokers who did. In the absence of policies limiting the availability of lower-cost cigarettes, these findings point to the need to better understand how price-sensitive smokers may travel through their environments and how these travel behaviours might influence purchasing behaviours.
In order for tobacco taxation policies to achieve maximal benefit for all smokers, it is important for governments to limit potential sources of lower-cost cigarettes that price-sensitive smokers might rely on to continue smoking. In some Canadian provinces, provincial governments actively partner with First Nations communities to ensure the prices of tobacco products sold on-reserve are consistent with those sold off-reserve. In Manitoba, for example, tobacco products sold on-reserve are subject to a tobacco tax equal to the province tobacco tax, and First Nations communities in that province are allowed to keep the revenue generated by those taxes (Canadian Cancer Society, 2017). By partnering with First Nations communities, provincial governments may be able to limit potential sources of lower-cost cigarettes. From an international perspective, such collaborative governance can be an effective way to minimize cigarette price differentials between countries, especially in the European Union, where smokers may travel between nations with relative ease in order to purchase lower-cost cigarettes across national borders.
Supplementary Material
Acknowledgments
Funding
The Canada arm of the ITC Four Country Survey was supported by grants from the US National Cancer Institute (R01 CA100362, R01 CA090955, P50 CA111236 (Roswell Park Transdisciplinary Tobacco Use Research Center), and P01 CA138389), the Canadian Institutes of Health Research (MOP-57897, MOP-79551, MOP-115016), and the Canadian Tobacco Control Research Initiative (014578). Geoffrey T. Fong was supported by a Senior Investigator Grant from the Ontario Institute for Cancer Research and a Senior Prevention Scientist Award from the Canadian Cancer Society Research Institute. G. Emmanuel Guindon holds the Centre for Health Economics and Policy Analysis (CHEPA)/Ontario Ministry of Health and Long-Term Care (MOHLTC) Chair in Health Equity, an endowed Chair funded in part by the MOHLTC and an Ontario Ministry of Research, Innovation and Science, Early Researcher Award. None of the funding sources had any involvement in study design, data collection, interpretation of results, in the writing of this manuscript, nor in the decision to publish this manuscript.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest
Geoffrey T. Fong and David Hammond have served as expert witnesses on behalf of governments in litigation involving the tobacco industry. All other authors have no competing interests to declare.
References
- Benjamini Y, & Hochberg Y (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B (Methodological), 57, 289–399. [Google Scholar]
- Canadian Cancer Society. (2017). Overview Summary of Federal/Provincial/Territorial Tobacco Control Legislation in Canada. Ottawa, Canada: Retrieved September 4, 2019, from http://convio.cancer.ca/documents/Legislative_Overview-Tobacco_Control-F-P-T-2017-final.pdf [Google Scholar]
- Chaloupka FJ, Straif K, & Leon ME. (2011). Effectiveness of tax and price policies in tobacco control. Tobacco Control, 20, 235–238. doi: 10.1136/tc.2010.039982 [DOI] [PubMed] [Google Scholar]
- Chaloupka FJ, Yurekli A, & Fong GT. (2012). Tobacco taxes as a tobacco control strategy. Tobacco Control, 21(2), 172–180. doi: 10.1136/tobaccocontrol-2011-050417 [DOI] [PubMed] [Google Scholar]
- DeCicca P, Kenkel D, & Liu F (2015). Reservation prices: An economic analysis of cigarette purchases on Indian reservations. National Tax Journal, 68, 93–118. doi: 10.17310/ntj.2015.1.04 [DOI] [Google Scholar]
- DMTI Spatial Inc. (2014). CanMap Postal Code Suite (User Manual No. v2014.2). Markham, Canada. [Google Scholar]
- Driezen P, & Thompson ME. (2011). Comparing policy measures across multiple ITC countries: Adjusting for time-in-sample. University of Waterloo; Waterloo, Canada: Retrieved from https://www.itcproject.org/files/ITC_Technical_Report_time-in-sample-adjustment_Dec2011.pdf [Google Scholar]
- Guindon GE, Burkhalter R, & Brown KS. (2017). Levels and trends in cigarette contraband in Canada. Tobacco Control, 26, 518–525. doi: 10.1136/tobaccocontrol-2016-052962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hyland A, Bauer JE, Li Q, Abrams SM, Higbee C, Peppone L, & Cummings KM. (2005). Higher cigarette prices influence cigarette purchase patterns. Tobacco Control, 14(2), 86–92. doi: 10.1136/tc.2004.008730 [DOI] [PMC free article] [PubMed] [Google Scholar]
- International Agency for Research on Cancer. (2011). IARC Handbooks of Cancer Prevention: Tobacco Control, Volume 14 Effectiveness of Tax and Price Policies for Control of Tobacco. Lyon, France. [Google Scholar]
- ITC Project. (2011). International Tobacco Control Policy Evaluation Survey (ITC): Four Country Project Wave 2–8 Technical Report. University of Waterloo; Waterloo, Ontario, Canada: Retrieved January 7, 2019, from http://itcproject.org/files/4c-w28-tech-report-sept2011.pdf [Google Scholar]
- Leuprecht C (2016). Smoking Gun: Strategic Containment of Contraband Tobacco and Cigarette Trafficking in Canada. MacDonald-Laurier Institute; Ottawa, Canada. [Google Scholar]
- Leuprecht C (2018). Cause and Effect: Why Higher Taxes Require Better Contraband Enforcement to Reduce Smoking. MacDonald-Laurier Institute; Ottawa, Canada. [Google Scholar]
- Luk R, Cohen JE, Ferrence R, McDonald PW, Schwartz R, & Bondy SJ. (2009). Prevalence and correlates of purchasing contraband cigarettes on First Nations reserves in Ontario, Canada. Addiction, 104(3), 488–495. doi: 10.1111/j.1360-0443.2008.02453.x [DOI] [PubMed] [Google Scholar]
- Mecredy GC, Diemert LM, Callaghan RC, & Cohen JE. (2013). Association between use of contraband tobacco and smoking cessation outcomes: a population-based cohort study. Canadian Medical Association Journal, 185(7), E287–E294. doi: 10.1503/cmaj.111861 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Natural Resources Canada. (2017). Aboriginal Lands of Canada. Retrieved April 5, 2017, from https://www.nrcan.gc.ca/earth-sciences/geography/topographic-information/download-directory-documentation/17215
- Non-Smokers’ Rights Association. (2012). Backgrounder on the Canadian Tobacco Industry and its Market (2012 Edition). Ottawa, Canada: Retrieved August 14, 2019, from https://nsra-adnf.ca/wp-content/uploads/2016/08/NSRA_Industry_Backgrounder_2012.pdf [Google Scholar]
- Non-Smokers’ Rights Association. (2015). Backgrounder on the Canadian Tobacco Market (Spring 2015). Ottawa, Canada: Retrieved August 14, 2019, from https://nsra-adnf.ca/wp-content/uploads/2016/08/2015_Canadian_Tobacco_Market.pdf [Google Scholar]
- Ontario Council of University Libraries. (2013). Scholars GeoPortal Wiki. Retrieved April 5, 2017, from https://spotdocs.scholarsportal.info/display/scholarsgeoportal/wiki
- Organization for Economic Co-operation and Development. (2015). Main Economic Indicators – complete database. Paris, France: Retrieved January 20, 2017, from 10.1787/data-00052-en [DOI] [Google Scholar]
- Organization for Economic Co-operation and Development. (2016). US Dollar to National Currency Spot Exchange Rate for Canada [CCUSSP01CAQ650N]. Federal Reserve Bank of St. Louis; St. Louis, United States: Retrieved October 24, 2019, from https://fred.stlouisfed.org/series/CCUSSP01CAQ650N [Google Scholar]
- Rainham D, McDowell I, Krewski D, & Sawada M (2010). Conceptualizing the healthscape: Contributions of time geography, location technologies and spatial ecology to place and health research. Social Science & Medicine, 70, 668–676. doi: 10.1016/j.socscimed.2009.10.035 [DOI] [PubMed] [Google Scholar]
- Royal Canadian Mounted Police. (2008). Contraband Tobacco Enforcement Strategy. Ottawa, Canada. [Google Scholar]
- Royal Canadian Mounted Police. (2011). Contraband Tobacco Enforcement Strategy: Third Progress Report (2010–2011). Ottawa, Canada. [Google Scholar]
- Sarson L (2019). “You Cannot Trade What Is Not Yours”: Indigenous governance and the NAFTA negotiations. American Review of Canadian Studies, 49, 332–347. doi: 10.1080/02722011.2019.1618616 [DOI] [Google Scholar]
- Sen A (2017). Smokes, smugglers and lost tax revenues: How governments should respond (Commentary No. 471). C. D. Howe Institute; Toronto, Canada. [Google Scholar]
- Soulakova JN, Pack R, & Ha T (2018). Patterns and correlates of purchasing cigarettes on Indian reservations among daily smokers in the United States. Drug and Alcohol Dependence, 192, 88–93. doi: 10.1016/j.drugalcdep.2018.07.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stanbrook MB. (2013). Fighting contraband tobacco: An opportunity for first nations leadership and partnership. Canadian Medical Association Journal, 185, 547. doi: 10.1503/cmaj.130377 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Statistics Canada. (2011). Forward Sortation Area Boundary File, Reference Guide (Catalog no. No. 92-179-G). Ottawa, Canada. [Google Scholar]
- Thompson ME, Boudreau C, & Driezen P (2005). Incorporating time-in-sample in longitudinal survey models Statistics Canada International Symposium Series 2005: Methodological Challenges for Future Needs. Ottawa, ON. [Google Scholar]
- Thompson ME, Fong GT, Hammond D, Boudreau C, Driezen P, Hyland A, … Laux FL. (2006). Methods of the International Tobacco Control (ITC) Four Country Survey. Tobacco Conrol, 15(Suppl III), iii12–iii18. doi: 10.1136/tc.2005.013870 [DOI] [PMC free article] [PubMed] [Google Scholar]
- U.S. National Cancer Institute and World Health Organization. (2016). The Economics of Tobacco and Tobacco Control National Cancer Institute Tobacco Control Monograph 21. NIH Publication No. 16-CA-8029A. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute; and Geneva, CH: World Health Organization; Retrieved January 7, 2019, from https://cancercontrol.cancer.gov/brp/tcrb/monographs/21/index.html [Google Scholar]
- Wang J, & Kwan MP. (2018). An analytical framework for integrating the spatiotemporal dynamics of environmental context and individual mobility in exposure assessment: A study on the relationship between food environment exposures and body weight. International Journal of Environmental Research and Public Health, 15, 2022. doi: 10.3390/ijerphl5092022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xu X, Malarcher A, O’Halloran A, & Kruger J (2014). Does every US smoker bear the same cigarette tax? Addiction, 109, 1741–1749. doi: 10.1111/add.12630 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.