Effect of Increased Tobacco Taxes on Smoking Trends in Canada

Abstract

Tobacco taxation is one of the tobacco control strategies applied around the world to reduce exposure. So as to understand the full benefit available to this strategy, it is important to understand its effectiveness among different subpopulations. Are all age groups within the population affected to the same extent by standard increases in cigarette prices? Are additional strategies required to achieve optimal reduction? This paper contemplates the Canadian context to determine how cigarette price increases affect access to different age groups. The study finds that teenage and young adult smokers respond up to three times as much as other age groups for a unit increase in price. Studies found that most risk behavior associated with smoking is initiated during this age. This study therefore considers increase in cigarette prices as an effective method of reducing access to tobacco for teenagers and young adults. Even so, synergistic approaches will achieve best results for other age groups, as price increase affects them minimally with increase in age.

Introduction & Review of Literature

Research shows that there have been significant gains that have been made in recent times to reduce cigarette smoking prevalence in high-income countries, including North America. Even so, there remains significant sub-populations within these countries that still remain with high smoking rates (Bader, Boisclair, & Ferrence, 2011). This is especially because there is varied response to pathways to change in smoking and control interventions among the different subgroups. Even for subgroups which benefit from these interventions, there is still a lot of disparities in their reception. As of 2015, it was estimated that up to 13 percent of the population smoke (Bader, Boisclair, & Ferrence, 2011). Smoking has been banned in indoor public spaces and within the workplace. This blanket ban applies to the federal government, its territories and provinces. Smoking bans have been applied uniquely to each province and territory with variations applying with regard to permissions for ventilated smoking rooms and distances outside a building for which smoking is banned.

Taxation on tobacco is one such intervention. Normally passed on to the customer as a higher cigarette price, this has been seen as an effective strategy for reduction in smoking statistics and with it, the adverse health consequences (World Health Organization, 2008). It is a population-based approach that seeks to increase the price of the commodity, thereby reducing its demand among the general adult population who are users (Jha & Chaloupka, 1999). By average, whenever the price increases by ten percent, it is expected that about 4 percent of the adult population in developed countries would reduce their demand (Jha & Chaloupka, 1999). Moreover, research has shown that increasing taxation could provide benefit former smokers who quit, reduce overall tobacco consumption and encourage previously hooked smokers towards cessation. Moreover, non-smokers also benefit because of the reduced intake of secondary smoke, which is just as harmful to their health. Even then, there is little known information known about the impact of tobacco taxes on subgroups, specifically in Canada. This research paper seeks to ask whether the impact of the tobacco taxes achieves the same results for different subgroups as much as it does on the general population. Do these individuals require additional measures and initiatives to respond to smoking? Are they influenced at the same rate as the general population?

This paper uses data collected from the Canadian resources on cigarette taxation and the prevalence of smoking among different subgroups which are defined by age. The findings from this study will also be complemented by results from a knowledge synthesis done by Bader et al. (2011), who investigated the effects of tobacco taxes on smoking trends among different subgroups in North America in general. Bader et al (2011) use the better practices approach to synthesize evidence on the subject. This evidence varies in terms of its importance and subpopulation measures in Canada having used the Better Practices Model to come up with its integration (Bader, Boisclair, & Ferrence, 2011). The reason this approach has been undertaken is based on the belief that complex problems can only be solved using solutions drawn from experience and science (Moyer, Maule, Cameron, Manske, & Garcia, 2004). By using both collected data affirmed by a knowledge synthesis on the subject for the Canadian population, it becomes increasingly possible to increase knowledge on this subject of interest.

From the onset, different subgroups are involved in smoking. It is important to contemplate these subgroups separately to understand the full impact of increased tobacco taxes on smoking among the various subpopulations. In Bader et al (2011), six main subgroups were contemplated as high-risk populations which are prone to continuous smoking behaviors. These groups were selected because of their continuous rates of smoking and even higher risk to health-related complications arising from their smoking habits. The major consideration for this case was the basis of age, from which Bader et al (2011) obtained their model. The following statistics were collated as part of the preliminary research pointing to the prevalence of smoking in comparison with the United States:

Table 1 : Subpopulation size and corresponding smoking prevalence 

Subpopulation size and smoking prevalence (2006–2008)
Subpopulation	% of Total population		Smoking prevalence (%)
Canada	US	Canada	US
General Population	33,212,696 (total population)	303,824,640 (total population)	19%	20.8%
Youth	6.80%	7.20%	15%	22%
Young Adults	9.20%	9.90%	25%	26%
Low SES Income/Education	11.40%	17%	**	**
Dual Diagnosis	5%–10%	5%–10%	38%–57%	41%–62%
Heavy and/or Long-term Smokers	**	**	**	**
Aboriginals	3.80%	1.50%	60%	32%

As can be seen above, six major socio-economic subgroups were selected and evaluated based on their population proportion and their smoking preference. As it follows, the various subgroups are explained below:

Youth (Under 19 years)

Tobacco control has been focused on this particular age group in the recent past. Most adult smokers engaged in smoking today report that they began smoking before they were 20 years old. This ultimately means that individuals who reach adulthood without becoming exposed to smoking have a high probability of remaining non-smokers throughout their entire life. Indeed, youth are smoking but research indicates that a majority of them smoking during their teenage years have expressed a desire to stop (O'loughlin, Gervais, Dugas, & Meshefedjian, 2009). While 70 percent are interested in stopping, only a meagre 19 percent manage to remain smoke-free for a 12-month period and even fewer follow through with this new behavior for up to five years. Cessation strategies also remain ineffective for youthful smokers, meaning that there remains little motivation to follow through with cessation. The recommendation from this study stated that cessation for youthful smokers should be considered in light of taxation, legislation and appropriate public health programs.

Young Adults

Despite strong and effective public awareness programs highlighting the health hazards of smoking, this group of individuals between 18 and 24 years of age continue to smoke at high rates (Bader, Travis, & Skinner, 2007). In this age group, individuals still remain prone to becoming initiated into smoking and engaging in higher levels of smoking activity. Furthermore, research has indicated that individuals smoking during this stage in their lives are most likely going to pursue smoking through other life stages. Additionally, it is evident that smoking rates among adults and teens have reduced in the recent past; however, rates for young adults in high-income countries still remain relatively constant (Bader, Boisclair, & Ferrence, 2011).

Low Socio-economic Status

Smoking is linked to lower socio-economic status – a factor that contributes to the compounding of health inequalities between socio-economic classes. For instance, research has found that smoking preferences are higher among individuals with lower education levels and lower socio-economic groupings. For instance, a study in Scotland revealed that smoking levels among the unemployed were generally higher than among the employed (Lee, Crombie, Smith, & Tunstall-Pedoe, 1991), confirming the assertion above, that low socio-economic status is linked with higher instances of smoking. A study conducted in Korea further analyzed the employment environment of the population to determine their smoking risks (Jung, Oh, Huh, & Kawachi, 2013). The study found that precarious workers were more likely to be smokers. Their risk to heavy smoking was higher than those of standard workers. Unemployment was a risk factor in reducing smoking, encouraging quitting and discouraging relapse. Nevertheless, where there was insecure employment (a characteristic of low socio-economic status), there was more consistency in determining smoking behavior in the individual. This trend was even more consistent than that of the unemployed (Jung, Oh, Huh, & Kawachi, 2013).

Dual Diagnosis

Smokers who were diagnosed with other substance abuse/mental health disorders were found to have unique dependence on tobacco. These populations were disproportionately affected by their dependence. In Canada, for instance, between 5-10 percent of the population has a mental illness diagnosable at modern medicine. Interestingly, this population is responsible for almost half of their entire tobacco consumption for both the US and Canada, as seen above. This subgroup consists 40 percent of the total cigarettes consumed (Lasser, et al., 2000). Research has displayed the unique relationship between cigarette smoking and patients with psychiatric disorders. Individuals with such disorders have a higher chance of initiation into smoking and a slimmer chance of cessation (Smith, Mazure, & McKee, 2014). Moreover, increasing number of co-occurring mental health disorders could contribute to increased smoking severity (Smith, Mazure, & McKee, 2014). The types of co-occurring disorders could also be used as predictors of smoking preference. In similar fashion, evidence on the subject suggests that smokers also meet several criteria that are also symptoms for psychiatric disorders (Degenhardt & Hall, 2001).

Heavy Smokers/Long-term Smokers

Heavy and long-term smokers face greater health risks associated with smoking. Research has provided intensity of smoking and its duration as strong factors contributing to the smoker’s mortality and morbidity. Although health scholars do not have an agreed definition for heavy and long-term smokers, it is generally believed that heavy smokers smoke more than fifteen cigarettes per day whereas other studies explore as high a number of 25 cigarettes daily (Godtfredsen, Prescott, & Osler, 2005; Levy, Romano, & Mumford, 2005). Long-term smokers remain undefined in literature (Bader, Boisclair, & Ferrence, 2011).

Aboriginal Populations

Aboriginal peoples in the Americas have been found to have significantly higher smoking rates compared with the general population, as evidence in Table 1. These rates have remained generally constant for the last two decades. Interestingly, one of the possible contributing factors to this situation is that inexpensive cigarettes are prevalent in these areas due to tax exemptions. In addition, aboriginal peoples have compounded factors such as low socio-economic status, inadequate access to healthcare and physical infrastructure as well as other environmental factors (Matheson, 2010).

Methods

This study intended to collect data on different age groups in Canada, which would direct its research in identifying the impact of increased tobacco taxes on smoking trends among different age subgroups.

Data Collection & Search Strategies

An extensive search was conducted to identify relevant data to the study in question. In this case, only published data was obtained on the impact and prevalence of smoking on the different age groups among Canadian populations. This study intended to collect data without regard for any socio-economic status except the age of the individual and their smoking preference. Canadian health and economic databases were consulted for the purpose of obtaining this information. For this study, the Canadian Tobacco Use Monitoring Survey was used to obtain the relevant smoker data for this study. This web resource was used because of its reliability in providing tobacco use data within the Canadian jurisdiction since 1999. The website has reported on tobacco use for individuals over the age of 15 since its inception, recording both current smokers, former smokers and never smokers. Before using this data, it was also contemplated in other contexts where it was used to corroborate the fact that the data was indeed used in other studies and was considered as viable. Moreover, a full review of the data was done before incorporating it into the study. At the end of the process, age-group data on smoking preferences was obtained alongside a twenty-year historical tobacco tax regime.

The study selected 13-year data that concluded in 2012 for the purpose of its review. The study selected a 13-year period, which would show recent trends, especially following the millennium to identify smoking trends among a group of interest, specifically millennials. Moreover, tobacco taxes were most versatile during this period, thereby creating the best space to review the market reaction to increasing tobacco taxes on smoking behaviors. Primary focus for this study surrounded cessation, consumption and smoking initiation, which could all be determined from a perspective of gains or losses in smoking numbers. This study also selected customer price index (CPI) because it is a weighted measure of prices for consumer goods and services. A general increase in the CPI shows that there was an increase in consumables, indicating an increase in cigarette pricing.

Quality Assessment

For the purpose of quality assessment, full text citations for all data resources have been provided within this document. Moreover, resources used in comparison and discussion of the results were also fully cited. Moreover, the Excel sheet used for the data analysis is also provided within the annexure of this document. The results were analyzed based on the question: “is the age subpopulation in question more responsive to cigarette pricing compared with the general population?” since this study only focused on the question of age as the limiting factor, it builds on other pieces of research which focus on different socio-economic demographics and provide the much-needed focus that analyze outcomes among the various age-groups. In the final discussion, various studies, including Bader et al (2011) were considered to compare the findings and provide recommendations.

Results

An initial sorting function was applied to the data providing Consumer Price Indices for tobacco in Canada between 1999 and 2012 and the following result was obtained:

Table 2 : Sorted Canada CPI and Age Group Data (Smallest to Largest Mean Preference) 

Year/CPI		Canada
Age Groups		Mean Preference	15-19	20-24	25-44	45+
2012	157.93	16.1	10.9	20.3	19.9	13.8
2010	150.24	16.7	12.2	22.1	20.4	14.1
2011	155.86	17.3	11.8	21.5	19.7	15.9
2009	145.43	17.5	13.0	23.0	20.4	15.4
2008	141.18	17.9	14.8	27.3	19.9	15.4
2006	132.52	18.6	14.8	27.3	21.7	15.4
2007	138.36	19.2	15.2	25.5	20.9	17.6
2005	128.92	18.7	18.0	26.0	23.4	13.9
2004	124.88	19.6	18.4	27.8	24.8	14.2
2003	116.20	20.9	18.3	30.5	25.4	15.8
2002	99.99	21.4	22.0	30.6	25.4	16.3
2001	75.78	21.7	22.5	32.1	25.0	16.8
2000	67.14	24.4	25.3	32.3	29.6	18.2
1999	64.07	25.2	27.7	35.4	29.9	18.5

From the above data, it is clear that there is a reducing preference for cigarettes with an increasing consumer price index. From an initial analysis, it is clear to see that the higher price is a continuous deterrent for the general population regarding their smoking habits. The mean preference moves from its initial place at 27.3% nationally to 16.2%, which represents an 11.1 percentage point change within the fifteen year period. On the other hand, the CPI for the product had gone full circle, displaying a 147% increase. This is perhaps an interesting point of view that will be visited in the discussion section. For teenagers under 15 years, the response was also noted having shifted 36%. Individuals between 15 and 19 years old experienced a significant 61% deterrence, whereas young adults (20-24 years) experienced 43% deterrence. The other age groups experienced 33% and 25% deterrence respectively. The study went further to contemplate each age group individually for the study period. The general population was evaluated first to find the general trend:

Table 3 : CPI and National Mean Preference 

Year	CPI	Mean Preference
1999	64.07	16.1
2000	67.14	16.7
2001	75.78	17.3
2002	99.99	17.5
2003	116.20	17.9
2004	124.88	18.6
2005	128.92	19.2
2006	132.52	18.7
2007	138.36	19.6
2008	141.18	20.9
2009	145.43	21.4
2010	150.24	21.7
2011	155.86	24.4
2012	157.93	25.2

From the mean preference for the entire country, one can see the progression more clearly, where increasing tobacco prices result in lowered interest in smoking trends. A graphical presentation of the same shows the same trend: an increasing CPI leading to lower overall smoking preference:

Figure 1 : Comparing CPI and national smoking trends 

To fully establish this relationship, the statistical analysis also involved a correlation model that would review the trend over the fifteen years to determine the strength of the tax incentive for the general population as a whole. The following model was obtained thereafter:

Table 4 : Correlation Model for CPI against General Population Preference 

	CPI	Mean Preference
CPI	1
Mean Preference	-0.97634	1

With a correlation coefficient of -0.97, this indicates a strong negative relationship between CPI and the mean preference, giving the third confirmation that increased cigarette pricing arising from increased taxation indeed affected negatively the consumption of cigarettes among Canadians. Noting that the study was interested in the different age groups to determine where cigarette taxation had the most impact, the rest of the age groups were individually considered for trends in smoking tendency reduction. A regression model was also created to contemplate the mean preference against CPI for the 13-year period analyzed. The following represents the results from the regression analysis:

Table 5 : Regression Analysis for CPI against mean preference 

SUMMARY OUTPUT
Regression Statistics
Multiple R	0.976343637
R Square	0.953246897
Adjusted R Square	0.949350805
Standard Error	7.279167652
Observations	14

From above, the R-squared value stands at 0.9532, indicating that there is a very strong fit to the regression line if it was plotted. This indication will be discussed in detail within the subsequent discussion section to determine the implications. The following results were obtained for the various age groups. The study equally investigated the age group between 15 and 19 years. Increasing CPU generally led to reducing interest in smoking with few outliers in the years 2004 and 2007. A graphical presentation of this data will also yield the same trend:

Table 6 : CPI and Preference for 15-19 year olds 

Year	CPI	15-19
1999	64.06667	27.7
2000	67.14167	25.3
2001	75.78333	22.5
2002	99.99167	22
2003	116.2	18.3
2004	124.875	18.4
2005	128.9167	18
2006	132.5167	14.8
2007	138.3583	15.2
2008	141.1833	14.8
2009	145.4333	13
2010	150.2417	12.2
2011	155.8583	11.8
2012	157.9333	10.9

Figure 2 : CPI Changes for Teenage Smokers 

From the change statistics, this group experienced the highest change in terms of impact of CPI and smoking preferences. Different reasons could be behind this phenomenon. Even so, due to the significant change in smoking behaviors, this group was also subjected to regression analysis to determine the relationship between the two phenomena. The following results were obtained:

Table 7 : Regression analysis for teenage smokers 

SUMMARY OUTPUT
Regression Statistics
Multiple R	0.977309953
R Square	0.955134743
Adjusted R Square	0.951395972
Standard Error	7.130690358
Observations	14

An interesting outcome is that the R-squared value for the teenage smoking group is higher than that of the mean group (15+). This could be the result of many reasons including the fact that the mean group captures a larger sample population. On the other hand, it could also be an indicator that increased tariffs on tobacco in fact affect teenage smokers’ ability to buy and therefore reduce exposure. Moreover, teenage smokers have minimal disposable income, explaining why this group is most affected by increasing cigarette prices. This could have good implications on the initiation and exposure to smoking for individuals as it seems to cut down smoking at the most risk-prone age group. The fact that the R-sqaured value is also at 0.955 indicates that there is a very good fit to the regression line, as displayed below:

Figure 3 : Regression line plot for teenage smokers 

To bring this age group into focus, a further correlation analysis was done to determine the nature of the relationship between in commodity prices and smoking behavior. The relationship was almost a negative linear relationship; indicating that increasing tobacco prices significantly affected the teenage smoker’s ability to keep up with their habit.

Table 8 : Correlation table for teenage smokers 

	CPI	15-19
CPI	1
15-19	-0.97731	1

The effect of other age groups was equally contemplated in the course of this study. The full review of the data is attached within the Excel sheet annexure. Young adults were also affected by increasing cigarette prices. However, over the 13-year period, increasing CPI did not achieve a 50% change in cigarette smoking populations among this age-group, indicating that there could be other factors in play. For instance, it is well-known that individuals within this age bracket already have sources of income, such as employment or business, which sustain their smoking habits, unlike for teenage smokers. Although this group had a negative linear relationship for their increasing cigarettes prices, it was much lower than that of the overall mean preference:

Table 9 : Correlation table for young adult smokers 

	CPI	20-24
CPI	1
20-24	-0.93712	1

With increasing age groups, this trend seems to persist, with increasing cigarette tariffs affecting the 45+ age group at a meagre 25%. Regression models for these groups indicate the same thing, despite the relationships and regression lines remaining fairly strong at 0.9. These differences among the age groups therefore display trends and factors that are contributors to continuous smoking within these age groups.

Table 10 : Regression for 25-45 smokers 

Regression Statistics
Multiple R	0.936730681
R Square	0.877464368
Adjusted R Square	0.867253065
Standard Error	11.78440601
Observations	14

R-squared value for this age group is 0.8775, indicating a weaker plot line for this age group compared to young adults and teenage smokers. Perhaps the least R-squared value was obtained for the 45+ age group, where there was almost half of the values outside of the regression line as displayed below:

Table 11 : Regression for 45+ smokers 

SUMMARY OUTPUT
Regression Statistics
Multiple R	0.72477
R Square	0.52529
Adjusted R Square	0.48574
Standard Error	23.1947
Observations	14

Discussion and Recommendations

From the above data, multiple conclusions could be drawn regarding smoking habits among the different age groups. One of the critical applications that this data has revealed is the impact of increased cigarettes taxation on young smokers. Literature has highlighted the problem of smoking in Canada, and indeed around the world, as one that begins with teenagers and youth being exposed to smoking behaviors (O'loughlin, Gervais, Dugas, & Meshefedjian, 2009). Moreover, a majority of these youth have expressed the desire to stop smoking and pursue healthy lifestyles. Even then, these smokers are not able to achieve the 12-month smoke free period. Research therefore advocates for increased awareness programing and effective tax regimes to reduce access to young smokers, as one of the mechanisms of achieving lasting behavioral change within this group (O'loughlin, Gervais, Dugas, & Meshefedjian, 2009). Moreover, one of the motivating factors among smokers in Canada is the availability of cheap tobacco and cigarettes – a factor responsible for unchanged smoking rates among aboriginal Canadians (Matheson, 2010). Therefore, the effective result obtained in raising tobacco tariffs to reduce teenage and young adult smoking rates is encouraging and supports research in the area that advocates for increased taxation as one of the mechanisms to reduce tobacco exposure in society. Research in the area found that the youth were up to three times more responsive to increased tobacco taxes and subsequent increase in cigarette prices (Bader, Boisclair, & Ferrence, 2011). Cigarette taxes were linked to increased levels of cessation (DeCicca, Kenkel, & Mathios, 2008), reduced progression to advanced levels of smoking (Cawley, Markowitz, & Tauras, 2004), and reduced initiation, especially within the Canadian context (Sen & Wirjanto, 2010). The findings above support this body of evidence by displaying that teenage and young adult smokers are the most responsive to cigarette taxation increases, with up to twice the response rate for individuals between 25-45 years and three times the response for individuals over 45 years of age.

Even then, age is not a singular factor to consider when contemplating reduction mechanisms for smoking preference among the youth. Different factors could be at play, which come into play in addition to age, which could affect youth smoking patterns, including gender, peer influences, school status and public sentiment. For instance, experimental smokers are less likely to stop due to increased prices whereas more regular smokers are more responsive to price changes (Harris & Chan, 1999). Again, younger teens could be affected by price because their smoking habit largely depends on borrowing, whereas older teens have the purchasing power that sustains their habit.

These socio-economic factors also provide part explanation for the different responses for individuals over the age of 25 and 45+ age group. For instance, cigarette taxation affecting low income individuals from these age groups has been found to increase cessation, quitting and behavioral change. The reason why this reduction is much lower than that of young smokers is because of the access to additional funds from employment and disposable income, which encourages smoking despite higher pricing. Perhaps the small changes in smoking preferences could be attributed to changes in individuals from lower socio-economic statuses who are more responsive to cigarette price changes. Research has indicated that increased tax affects individuals from lower socio-economic status more than it does individuals with higher income, since such individuals pay higher tax per capita (Kamin, 2002). Contributions from economists argue that cigarette tax is progressive because it influences behavioral change, even at the population level (Colman & Remler, 2008). One of the reasons provided for this reasoning is that increased cigarette pricing results in reduced uptake – an outcome that increases accrued benefits of non-smoking.

In reducing smoking in Canada, studies have examined the effect of independent tobacco control prices alongside price adjustment. Studies have found that tobacco control policies in play within Canada all play a role in reducing cigarette smoking tendencies in the population. Some of the most effective policies included clean indoor air and youth access laws. Clean indoor air placed restrictions on smoking especially in public places, workplaces and school premises whereas access laws limited availability to youth and provided warnings at points of sale. As a result, this study recommends the following:

Applying increased cigarette taxation as a means of reducing youth access to cigarette smoking.

Applying strong synergistic approaches to reducing youth access to smoking, including stringent clean air policies and access laws.

Conducting research on applicable youth awareness campaigns to approach youth smokers and promote cessation.

Conclusion

This study set out to evaluate the effect of tobacco taxes on smoking habits among different age groups in Canada. Having obtained CPI data for Canadian consumables and smoker preference data from Canadian Tobacco Use Monitoring Survey, this study considered different factors that affected smoking trends in Canada, as explored in literature. Using regression analysis, this study found disproportionate effects of price increase among the various age groups, with young adults and teenagers being affected by up to three times as much as older adults. Among other things, this has been attributed to individual buying power, access to disposable income and other socio-economic factors. As a singular mechanism to reduce smoking preference among youth, price increases are more effective among teenagers and young adults compared to other age groups within the population. Even so, price increases for cigarettes are felt at a population level, as displayed by the results above. As such, this paper concludes that there is a population-wide response to cigarette price increases.

References 

Bader, P., Boisclair, D., & Ferrence, R. (2011). Effects of tobacco taxation and pricing on smoking behavior in high risk populations: a knowledge synthesis. International journal of environmental research and public health, 8(11) , 4118-4139.

Bader, P., Travis, H. E., & Skinner, H. A. (2007). Knowledge synthesis of smoking cessation among employed and unemployed young adults. American Journal of Public Health, 97(8) , 1434-1443.

Cawley, J., Markowitz, S., & Tauras, J. (2004). Lighting up and slimming down: the effects of body weight and cigarette prices on adolescent smoking initiation. Journal of health economics, 23(2) , 293-311.

Colman, G. J., & Remler, D. K. (2008). Vertical equity consequences of very high cigarette tax increases: if the poor are the ones smoking, how could cigarette tax increases be progressive? Journal of Policy Analysis and Management: The Journal of the Association for Public Policy Analysis and Management, 27(2) , 376-400.

DeCicca, P., Kenkel, D., & Mathios, A. (2008). Cigarette taxes and the transition from youth to adult smoking: smoking initiation, cessation, and participation. Journal of health economics, 27(4) , 904-917.

Degenhardt, L., & Hall, A. W. (2001). The relationship between tobacco use, substance-use disorders and mental health: results from the National Survey of Mental Health and Well-being. Nicotine & Tobacco Research, 3(3) , 225-234.

Godtfredsen, N. S., Prescott, E., & Osler, M. (2005). Effect of smoking reduction on lung cancer risk. Jama, 294(12) , 1505-1510.

Harris, J. E., & Chan, S. W. (1999). The continuum ‐ of ‐ addiction: cigarette smoking in relation to price among Americans aged 15–29. Health Economics, 8(1) , 81-86.

Jha, P., & Chaloupka, F. (1999). Curbing the Epidemic: Governments and the Economics of Tobacco Control. Washington, DC, USA: World Bank Publications.

Jung, Y., Oh, J., Huh, S., & Kawachi, I. (2013). The Effects of Employment Conditions on Smoking Status and Smoking Intensity: The Analysis of Korean Labor & Income Panel 8th–10th Wave. PloS one, 8(2) , e57109.

Kamin, D. (2002). Justifying the Tobacco Tax. Swarthmore, PA, USA: Swarthmore College. Retrieved from http://www.swarthmore.edu/x10949.xml

Lasser, K., Boyd, J. W., Woolhandler, S., Himmelstein, D. U., McCormick, D., & Bor, D. H. (2000). Smoking and mental illness: a population-based prevalence study. Jama, 284(20) , 2606-2610.

Lee, A. J., Crombie, I. K., Smith, W. C., & Tunstall-Pedoe, H. D. (1991). Cigarette smoking and employment status. Social science & medicine, 33(11) , 1309-1312.

Levy, D. T., Romano, E., & Mumford, E. (2005). The relationship of smoking cessation to sociodemographic characteristics, smoking intensity, and tobacco control policies. Nicotine & Tobacco Research, 7(3) , 387-396.

Matheson, J. (2010). Estimating Price Elasticity for Tobacco in Canada’s Aboriginal Communities . Retrieved from Job Market Paper: http://works.bepress.com/cgi/viewcontent.cgi?article=1003&context=matheson

Moyer, C., Maule, C., Cameron, R., Manske, S., & Garcia, J. (2004). Better Solutions for Complex Problems: Description of a Model to Support Better Practices for Health, Version 04.07.27. Toronto, ON, Canada: Canadian Tobacco Control Research Initiative.

O'loughlin, J., Gervais, A., Dugas, E., & Meshefedjian, G. (2009). Milestones in the process of cessation among novice adolescent smokers. American journal of public health, 99(3) , 499-504.

Sen, A., & Wirjanto, T. (2010). Estimating the impacts of cigarette taxes on youth smoking participation, initiation, and persistence: empirical evidence from Canada. Health economics, 19(11) , 1264-1280.

Smith, P. H., Mazure, C. M., & McKee, S. A. (2014). Smoking and mental illness in the US population. Tobacco control, 23(e2) , e147-e153.

World Health Organization. (2008). WHO Report on the Global Tobacco Epidemic. Geneva, Switzerland: World Health Organization.