TIME TO QUIT: THE TOBACCO TAX INCREASE AND HOUSEHOLD WELFARE IN BOSNIA AND HERZEGOVINA NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE TIME TO QUIT: THE TOBACCO TAX INCREASE AND HOUSEHOLD WELFARE IN BOSNIA AND HERZEGOVINA Alan Fuchs, Edvard Orlic, and César A. Cancho January 2019 Alan Fuchs (corresponding author) is senior economist with Poverty and Equity Global Practice, World Bank, Washington, DC; his email is afuchs@worldbank.org. Edvard Orlic lecturer in business economics at Bournemouth University, Bournemouth, UK. César A. Cancho is economist with Poverty and Equity Global Practice, World Bank, Washington, DC. Support in the preparation of this report has been provided under the Reducing Health Risk Factors Project in Bosnia Herzegovina Project (RHRF) supported by the Government of Switzerland and implemented by the World Bank, and the World Bank’s Global Tobacco Control Program and co-financed by the Bill and Melinda Gates Foundation and Bloomberg Philanthropies. The authors are grateful to Darko Paranos, Ian Forde, Polina Kuznetsova, Enis Baris, Emanuel Salinas Munoz, Patricio Marquez, Carlos Silva-Jau- regui and Una Sahinpasic for their comments and support. The findings, interpretations, and conclusions in this research note are entirely those of the authors. They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the countries they represent. I NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE CONTENTS Abstract # 1. Introduction 1 2. Literature Review 5 2.1. Tobacco and health 5 2.2. Tobacco control policies 6 2.3. Tobacco taxes 7 2.4. Price elasticities of tobacco consumption 8 3. Model 11 3.1. Change in tobacco expenditure 11 3.2. Medical expenses 12 3.3. Increase in years of working life 12 4. Data and descriptive statistics 15 4.1. Descriptive statistics 16 4.2. Tobacco price elasticity, by decile 18 4.3. Mortality and morbidity 20 4.4. Tobacco-related medical costs 22 5. Results 25 5.1. Tobacco price increase 25 5.2. Medical expenses 26 5.3. Income gains deriving from an increase in working life years 27 5.4. Net effects: total distributional impact 28 6. Results by entity and household type 31 6.1. Federation of Bosnia and Herzegovina vs. Republika Srpska 31 6.2. Urban vs. rural 32 7. Discussion 35 References 37 Tobacco Taxation: Evidence from Bosnia and Herzegovina Annex A. Detailed scenario outcome across entity and household type 43 A.1. Federation of Bosnia and Herzegovina 43 A.2. Republika Srpska 44 A.3. Urban population 45 A.4. Rural population 46 LIST OF FIGURES Figure 1. Trends in cigarette prices, Bosnia and Herzegovina 16 Figure 2. Average price paid by households for a 20-cigarette pack 16 Figure 3. Households with Cigarette Consumption 17 Figure 4. Share of Cigarette Consumption in Total Household Consumption 17 Figure 5. Income Gains: Direct Effect of Tobacco Taxes (Change in expenditure because of tobacco taxes) 26 Figure 6. Income Gains: Medical Costs of Tobacco Taxes (Reduction in Medical Expenditures) 27 Figure 7. Income Gains: Production during Years Lost, by Decile 28 Figure 8. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes (Tobacco price increase, medical expenditure, and working years gained) 29 LIST OF TABLES Table 1. Baseline Descriptive Results, by Decile 18 Table 2. Cross-Country Elasticity Estimates, by Decile 19 Table 3. Cigarette Price Elasticities, by Decile 19 Table 4. Deaths, by Gender, Illnesses Related to Tobacco Consumption, Bosnia and Herzegovina, 2015 20 Table 5. Events, by Age-Group, Illnesses Related to Tobacco Consumption, Bosnia and Herzegovina, 2015 21 IV // Table of Contents Table 6. Medical Cost of Treatment of Tobacco-Related Diseases (KM), 2015 23 Table 7. The Direct Effect of Price Increases through Taxes, by Decile (%) 25 Table 8. Reduction in Medical Costs, by Decile (%) 27 Table 9. Years of Working Life Lost, by Decile (%) 28 Table 10. Net Effect on Household Expenditures, by Decile (%) 29 Table 11. Net Effect on Household Expenditures, Federation of Bosnia and Herzegovina, by Decile (%) 32 Table 12. Net Effect on Household Expenditures, Republika Srpska (%) 32 Table 13. Net Effect on Household Expenditures, Urban Population, by Decile (%) 33 Table 14. Net Effect on Household Expenditures, Rural Population, by Decile (%) 33 Table A.1. Direct Effect of Price Decrease through Taxes, by Decile (%) 43 Table A.2. Reduction in Medical Costs, by Decile (%) 43 Table A.3. Years of Working Life Lost, by Decile (%) 43 Table A.4. Direct Effect of Price Decrease through Taxes, by Decile (%) 44 Table A.5. Reduction in Medical Costs, by Decile (%) 44 Table A.6. Years of Working Life Lost, by Decile (%) 44 Table A.7. Direct Effect of Price Decrease through Taxes, by Decile (%) 45 Table A.8. Reduction in Medical Costs, , by Decile (%) 45 Table A.9. Years of Working Life Lost, by Decile (%) 45 Table A.10. Direct Effect of Price Decrease through Taxes, by Decile (%) 46 Table A.11. Reduction in Medical Costs, by Decile (%) 46 Table A.12. Years of Working Life Lost, by Decile (%) 46 LIST OF BOXES Box 1. Estimating Tobacco-Related Medical Costs, Bosnia and Herzegovina 22 V NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE ABSTRACT Tobacco, a leading cause of death, is linked with high medical expenditures, lower life expectancy at birth, reductions in the quality of life, and other adverse effects. Tobacco taxes are considered an effective policy tool to reduce tobacco consumption and produce long-run benefits that may outweigh the costs associated with a tobacco price increase. However, policy makers avoid using tobacco taxes because of the possible regressive effects. In particular, poorer deciles across the income distribution are proportionally more negatively affected than richer ones by the extra tax burden. This paper uses an extended cost-benefit analysis to estimate the distributional effect of tobacco tax increases in Bosnia and Herzegovina. The analysis considers the effect on household income of an increase in tobacco prices, changes in medical expenses, and the prolongation of working years under various scenarios, based on data in three waves of the national Household Budget Survey. One critical contribution is a quantification of the impacts by allowing price elasticities to vary across consumption deciles. The results indicate that a rise in tobacco prices generates positive income variations across the lowest income groups in the population (the bottom 20 percent). At the same time, tobacco price increases have negative income effects among middle-income and upper-income groups. These effects are larger, the higher the income level. If benefits through lower medical expenses and an expansion in working years are considered, the positive effect is acerbated among the lowest income groups. The middle of the distribution sees the income effect turn from negative to positive, and the top 40 percent, although continuing to experience a negative effect, see the magnitude of this effect diminish. Altogether, these effects mean that increases in tobacco prices have a pro-poor, progressive effect in Bosnia and Herzegovina. These results also hold within entities and across urban and rural areas. JEL Codes: H23, H31, I18, O15 VII NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 1 INTRODUCTION The World Health Organization (WHO 2017a) estimates that tobacco kills more than 7 million people worldwide each year. It is the second leading cause of death and disability worldwide (Ng et al. 2014) and is among the major preventable causes of disease and premature death globally (Doll and Hill 1956; Wynder and Graham 1950). Diseases associated with tobacco use include lung cancer, stroke, ischemic heart disease, and respiratory diseases (DHHS 2004). Nearly 80 percent of the world’s smokers live in low- and middle-income countries and are less likely to be informed about the adverse health effects of tobacco use relative to individuals in high-income countries. Bosnia and Herzegovina is a major consumer of tobacco, where 47.2 percent more men and 30.0 percent more women smoke than in the high–human development index (HDI) country-average (Tobacco Atlas 2018). As a consequence of high prevalence of tobacco consumption, 29.8 percent more men and 13.7 percent more women die than on average in HDI countries (Tobacco Atlas 2018). The World Health Organization (WHO) has set the reduction of tobacco consumption as one of its primary goals. It has thus promoted the implementation of MPOWER control policies which include tobacco monitoring, smoke-free policies, smoking-cessation support programs, relevant health advice, advisory deterrents, and taxation policies (WHO 2015a). According to Levy et al. (2013), if 41 countries across the world had implemented at least one MPOWER policy advocated by WHO between 2007 and 2010, the number of smokers would have been cut by 14.8 million, and 7.4 million would have avoided death caused by smoking. Among the strategies, tobacco taxation is deemed to be one of the most efficient ways to reduce tobacco consumption. The inelastic demand of some tobacco consumers is useful for increasing tax revenues (Goodchild et al. 2016); and the higher price elasticity of younger smokers makes the tax an effective consumption deterrent (Chaloupka et al. 2002; Debrott Sánchez 2006). The reduction of tobacco consumption in Bosnia and Herzegovina has thus become a priority not only because of the health care costs, but also as a necessity in the country’s accession to the European Union (EU). Each prospective EU member is required to adjust current excise tax policies to the EU acquis communitaire. To reduce the burden on the public health system and to tackle other issues, such as productivity losses because of premature deaths and absenteeism, the government adopted an aggressive tobacco tax policy. The country now has the highest share of total and excise taxes in the average price of a pack of cigarettes globally, currently standing at 86.0 percent (WHO 2015b). The latest decision of the Office of Indirect Taxation effective from January 1, 2018, on special 1 Tobacco Taxation: Evidence from Bosnia and Herzegovina and minimum excise duties and the amount of excise duties on cigarettes represents the continuation of the harmonization of excise policy with EU standards. The specific excise on the most sold brand of cigarettes is 26.7 percent of the average price, which represents an increase of 100 percent over the corresponding tax in 2010 (WHO 2017b). The ad valorem tax of 42 percent and the fixed tax of KM 1.50 per 20-cigarette pack has increased the overall tax burden (expressed as a share of average retail price) by 17 percent relative to 2010. Overall, government revenue from tobacco excise taxes (specific and ad valorem) rose from KM 449 million in 2009 to KM 808 million in 2016 (WHO 2017b) and is expected to increase further in coming years because the country is only half way to reaching the minimum EU excise duty on tobacco, which is currently KM 3.60 per 20-cigarette pack1. Furthermore, the government introduced health warnings on cigarette packs; advertising material was banned; and smoking is prohibited in educational, health care, and social institutions. Even though increasing taxes on tobacco seem to be one of the most efficient measures for reducing tobacco consumption and increasing government revenue, its effectiveness largely depends on how the tax increase impacts the final price paid by consumers (World Bank, 1999). In 2009 when the Law on excise tax on tobacco and tobacco products was enacted the specific excise duty applied to the most sold brand was 13 percent while by the end of 2016 it increased to 27 percent (WHO 2017b). During the same period, the price of cigarettes increased from an average KM 2 to more than KM 4 in 2018 for the most popular cigarettes. The consolidation of tobacco market and constant and regular increase of excise duties has resulted in strong price increase of tobacco products which may have led to increased illicit trade as argued by major tobacco companies. In one of the rare studies on the illicit tobacco trade in Bosnia and Herzegovina, Joossens et al. (2009) estimate that the share of the illicit cigarette market in the country is between 35 percent and 45 percent, while more recent data of Euromonitor International (2018) suggest a share of the illicit trade in legal cigarette consumption at around 17 percent. However, Gallagher et al. (2018) have conducted a systematic review of 35 existing assessments of industry-funded data on ITT. They find that tobacco industry estimates are higher than independent estimates. Problems were identified with data collection (29 cases), analytical methods (22), and the presentation of results (21), which resulted in inflated ITT estimates or data on ITT that were presented in a misleading manner. According to the most recent Report on the Global Tobacco Epidemic (WHO 2017b), the age-standardized prevalence of current cigarette smoking declined from 41.5 percent in 2009 to 31.6 percent in 2015.2 A significant decrease was recorded among both genders. 1  In U.S. dollars, government revenue from tobacco excise taxes (specific and ad valorem) rose from US$329 million in 2009 to US$435 million in 2016. 2  Similarly, the age-standardized prevalence of daily cigarette smoking declined from 37.5 percent in 2009 to 28.1 percent in 2015. 2 // Introduction In addition, recent estimates of the Global Youth Tobacco Survey suggest that the prevalence of cigarette use among youth (ages 13–15) declined as well, from 11.7 percent in 2008 to 11.2 percent in 2013. The decrease in prevalence can partially be attributed to the increase in the tax burden over the years. For example, in 2008, the share of GDP per capita required to purchase 2,000 cigarettes of the most popular brand was standing at 3.1 percent, while, in 2016, the share rose to 5.9 percent, indicating that cigarettes have become less affordable (WHO 2017b). Because low-income families usually allocate a larger proportion of their budgets to purchase tobacco products, the tax increase would seem to be a regressive policy at first glance. However, the expected overall reduction in tobacco consumption associated with the tax increase would -in the long run- reduce the adverse effects of tobacco consumption, including higher medical expenditures and added years of disability among smokers, the negative effect on life expectancy at birth, reductions in the quality of life, and numerous negative externalities among first- and secondhand smokers, thus benefiting former smokers and their families. Therefore, if these indirect effects are considered, the concerns about the distributional impacts of tobacco tax policies diminish and could even no longer hold. As has been shown by Denisova and Kuznetsova (2014) and Verguet et al. (2015) for Ukraine and China, respectively, the future benefits of nonsmoking outweighs the costs attributed to tobacco taxes, especially for low income families. The increase in the tobacco tax and the subsequent reduction of tobacco consumption could therefore result in potential measurable benefits for different income groups. The aim of this paper is to quantify the effects of tobacco taxation on incomes through three factors: (1) the tobacco price increase driven by taxes increases, (2) the reduction in medical expenses associated with averted treatment costs of tobacco related diseases, and (3) the rise in revenues because of the gain in years of employment due to extension in life expectancy. To assess the impact of these effects, this paper estimates the price elasticity of tobacco for different income groups, simulates upper- and lower-bound scenarios, and calculates the welfare gains among these various income groups. The study is structured as follows. Section 2 reviews the literature on the health effects of tobacco, the economic costs associated with tobacco-related diseases, tobacco tax policies, and price elasticities. Section 3 presents methodology and model used to estimate the impact of the tobacco tax. Section 4 present the data and provides descriptive statistics. Section 5 examines the results. The final section concludes with a discussion on policy implications. 3 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 2 LITERATURE REVIEW 2.1. Tobacco and health Calculations show that close to 100 million deaths could be attributed to tobacco consumption in the 20th century (Peto and Lopez, 2004), and that if current consumption trends continue, up to 1 billion people could die from tobacco-related diseases during the present century (Jha and Peto 2014). In Bosnia and Herzegovina, more than 2,000 children (10-14 years old) and 1,053,000 adults (15+ years old) continue to use tobacco each day (Tobacco Atlas, 2018). According to the U.S. Department of Health and Human Services, tobacco consumption is responsible or contributes to many types of cancers, respiratory problems and cardiovascular diseases. The exposure to secondhand smoke has a causal relationship with many respiratory diseases in children and adults. There are more than 4,000 chemicals in tobacco smoke (of which at least 250 are harmful and more than 50 can cause cancer). Moreover, according to the WHO (2017a), secondhand smoke is responsible for over 890,000 premature deaths worldwide. Tobacco consumption has negative consequences on human capital development and imposes an increasing economic burden because smoking decreases earnings potential and labor productivity (WHO 2015a). Goodschild et al. (2018) find that tobacco-related diseases accounted for 5.7 percent of global health expenditure in 2012 with the highest proportion in Eastern Europe where smoking is responsible for around 10 percent of total health expenditure. The total economic costs of smoking, including health expenditure and productivity losses, were equivalent to 1.8 percent of the world’s gross domestic product (GDP) (US$1.85 trillion in purchasing power parity U.S. dollars). The highest share, according to these authors, was in high-income countries (US$1.12 trillion in purchasing power parity dollars), where the tobacco epidemic is the most advanced. The economic burden of smoking is highest in Eastern Europe where the costs related to smoking amount to 3.6 percent of GDP. Estimates on Bosnia and Herzegovina suggest that total economic costs attributable to smoking constitute 3.4 percent of GDP (US$1.2 billion in purchasing power parity dollars). In addition, the authors estimate that smoking-attributable health expenditure that includes the direct costs incurred in a given year (for instance, hospitalization and medications) and indirect costs, representing the value of lost productivity in current and future years because of disability and mortality, is equal to 7.5 percent of total health expenditure in Bosnia and Herzegovina. 5 Tobacco Taxation: Evidence from Bosnia and Herzegovina Tobacco price increases are also associated with expansion in productive life years. Verguet et al. (2015) analyze the health effects of a price increase in China. Their research concludes that a 50 percent rise in prices would result in 231 million years of life gained over 50 years, with a significant impact in the lowest income quintile. For Russia, Maslennikova et al. (2013) estimate that if taxes were increased to 70 percent of retail price along with other policies like banning smoking in public places, 3.7 million tobacco-related deaths would be averted, even without considering the effects of second hand smoking. Goodschild et al. (2018) estimated that in 2012 number of labour years lost due to smoking-attributable diseases came to 26.8 million years, with 18.0 million years lost due to mortality and 8.8 million years lost due to disability. 2.2. Tobacco control policies Globally, antitobacco policies include prohibiting smoking in particular locations to establish completely smoke-free environments, advertising to deter tobacco use, restrictions on tobacco sales by age, smoking cessation programs, prohibitions on tobacco sales close to schools, and taxation. These various policies have produced diverse effects in tobacco use and exposure among populations. For instance, the most common policy is related to mass media campaigns. In 2016, such campaigns addressed 56 percent of the world's population. However, people in low-income countries are less likely to be exposed to these campaigns and there is limited information about the cost effectiveness of this approach (WHO, 2015a). Durkin, Brennan, and Wakefield (2012) conclude that mass media awareness programs could promote quitting; however, their impact depends on the duration of the campaigns, especially among low-income smokers. It also depends on the message; information about the adverse health risks of smoking represents the most efficient means to reach users. Another common way to discourage tobacco consumption is through health warning labels on tobacco packages. In 2016, almost 45 percent of the world’s population was being exposed to such labelling. They are on the most cost- efficient means to discourage tobacco consumption and are widely supported by public (WHO 2015a). In Bosnia and Herzegovina, there are currently no laws prohibiting smoking in public places. Designated smoking rooms with strict technical requirements are allowed in all indoor public places under the current legislation of both the Federation of Bosnia and Herzegovina (Law on amendments to the Law on the limited use of tobacco products, 2011) and the Republika Srpska (Law on amendments to the Law on the prohibition of smoking tobacco products in public places, 2009). Cessation services are available in both entities and health insurance fully cover the costs. Nicotine replacement therapy is also available and can be purchased over the counter, but it is not on the essential drug 6 // Literature Review list and therefore is not covered by health insurance. Tobacco laws also mandate that health warning appear on every package and any outside packaging and labelling used in the retail sale and describe the harmful effects of tobacco use on health. The laws have enforced bans on five of seven forms of direct advertising, and it also bans appearances of tobacco brands on television or films. 2.3. Tobacco taxes Tobacco taxation is considered one of the most efficient measures to reduce tobacco consumption (World Bank, 1999). Therefore, by imposing an excise tax the government aims to correct a negative externality such as health risk associated with tobacco consumption and reduce exposure to second hand smoke. Apart from putting a price to tobacco consumption the aim of so called Pigouvain tax is to raise revenues which may then be used to lower taxes elsewhere or to finance the necessary prevention and control of cigarette-related diseases. Levy et al. (2014) have associated price increases with significant declines in tobacco consumption. They estimate that higher taxes are responsible for almost 50 percent of the decline in smoking. The effects of these policies mainly depend on the type of taxes (ad valorem and specific excise taxes). Ad valorem taxes are based on a percentage of the retail price. This type of tax tends to widen price differences between cigarette brands, making expensive brands relatively more expensive. However, tobacco companies can lower the tax burden by keeping prices low. For this reason, the levels of consumption and the amount of government tax revenue depend on the industry pricing strategy. Alternatively, specific excise taxes can be used by adding a fixed, monetary tax to every cigarette, regardless of its baseline price. It reduces price differences between brands, benefitting manufacturers of more expensive cigarettes. Specific tobacco excise represents a more efficient for tobacco control purpose as it increases cigarettes prices relatively more than ad valorem taxes (IARC, 2011). However, they must be adjusted periodically for inflation to accomplish their mission as otherwise specific taxes may decline over time in real terms. The taxation system in Bosnia and Herzegovina uses both type of taxes. The Law on Excise Duties in Bosnia and Herzegovina came into force in 2009. Initially, the tax base for specific and ad valorem taxes was the most popular brand of cigarettes sold in the country. However, an amendment to the law that came into effect in 2014 (Official Gazette 49/14) changed the tax base to the weighted average price of cigarettes. The latter is calculated as the total value of all cigarettes released for consumption, based on the retail selling price including all taxes, divided by the total quantity of cigarettes released for consumption. Under the Law on Excise Duties, duty on tobacco products is 7 Tobacco Taxation: Evidence from Bosnia and Herzegovina to be paid at the rate of 42 percent on the tax base, plus a special duty depending on the number of cigarettes—per 1,000 or per pack of 20 cigarettes—determined by Governing Board of the Indirect Taxation Authority. The latter is introduced to ensure the dynamics of harmonization of the excise rate with the relevant European directives. Under the amended law which came into effect in 2014, the Governing Board of the Indirect Taxation Authority will increase the specific tax until total excise tax burden reaches KM 176 per 1,000 cigarettes. In addition, minimum yearly increase cannot be lower than KM 7.50 per 1,000 cigarettes. According to latest decisions adopted in October 2018 ("Official Gazette", No. 75/18) defines that the following excise tax will be paid effective from 1st January 2019: • Proportional excise duty at the rate of 42 percent of the retail weighted average price of cigarettes (KM 4.76) • The specific excise duty of KM 82.50 per 1,000 cigarettes (KM 1.65 per pack of 20). In addition, minimal excise duty is determined in the amount of KM 143 per 1,000 cigarettes. 2.4. Price elasticities of tobacco consumption The effectiveness of tax increases on cigarette consumption is mainly determined by cigarette price elasticity. The magnitude of price elasticity is central in calibrating the effect of tobacco taxation systems because it determines the sensitivity of demand to a change in tobacco prices. There is extensive research on the price elasticity of tobacco. In low- and middle-income countries, a 10 percent increase in prices of cigarettes is associated with an average 6 percent reduction in cigarette consumption (IARC 2011). The higher price elasticity of young people makes taxes a good way to fight tobacco. Institutions such as IDB (2011), WHO (2008), The International Agency for Cancer Research (IARC 2011), the World Bank (1999) and authors such as Yeh et al. (2017) showed that a rise of 10 percent in the price of cigarettes would significantly reduce cigarette consumption as well the total death toll caused by smoking in the EU, but would be most effective in countries where GNI per capita is below US$5,500, such as Bulgaria and Romania, where the price elasticity of demand is the highest. Chaloupka et al. (2010) reported the price elasticity of cigarettes demand was in the range from -0.25 to -0.5 after revising more than 100 studies from industrialized countries. Consumption reduction was larger in low- and middle-income countries. Chaloupka and Grossman (1996) and Lewit and Coate (1982) estimate the elasticity among the under-18 population in the United States at between −1.44 and −1.31 and, among adults ages 18 years or older, at between −0.27 and −0.42. Gallus et al. (2006) estimate a price elasticity of 8 // Literature Review −0.46 for 52 countries in Europe. Denisova and Kuznetsova (2014) estimate price elasticities in Ukraine by income deciles, ranging from −0.44 for the lowest income group to −0.11 for the highest. Fuchs and Meneses (2017) also estimate decile-level price elasticities in Ukraine and find a higher average price elasticity (−0.45), ranging from −0.33 for the richest income group and −0.59 for the poorest. For India, cigarette price elasticities have been estimated for different income groups, including −0.83 and −0.26 for the lowest and highest income groups, respectively (Selveraj et al., 2015). Among several factors, there are two important ones involved in determining tobacco price elasticities, namely, income and age. People in lower-income groups tend to change consumption behavior more given a change in price, that is, they have more elastic demands, relative to higher-income groups (Jha and Chaloupka 2000). At the same time, younger groups in populations are more responsive to tobacco price increases because on average they tend to be less nicotine dependent, more affected by peer effects, and possess less disposable income. Hence the importance of the increase in tobacco prices (through taxes) to reduce tobacco consumption among the younger groups of the population. Studies have also shown that there are geographical variations in smoking behavior (Idris et al. 2007). In Eastern European countries, such as Slovenia, Romania, and Slovakia, tobacco prevalence in rural and remote areas is higher than in urban areas whereas in Western European countries, such as Germany, Sweden, Finland and Denmark, the opposite has been reported (Idris et al. 2007). This paper will take advantage of the extensive literature analyzing the health effects of tobacco, public policies, and price elasticities on the international level discussed above to analyze the potential changes in household welfare induced by an increase in tobacco taxes, as there is little empirical evidence on Bosnia and Herzegovina. 9 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE also estimate decile-level price elasticities in Ukraine and find a higher average price elasticity (−0.45), ranging from −0.33 for the richest income group and −0.59 for the poorest. For India, cigarette price elasticities have been estimated for different income groups, including −0.83 and −0.26 for the lowest and highest income groups, respectively (Selveraj et al., 2015). Among several factors, there are two important ones involved in determining tobacco price elasticities, namely, income and age. People in lower-income groups tend to change consumption behavior more given a change in price, that is, they have more elastic demands, relative to higher-income groups (Jha and Chaloupka 2000). At the same time, younger groups in populations are more responsive to tobacco price increases because on average they tend to be less nicotine dependent, more affected by peer effects, and possess less disposable 3 income. Hence the importance of the increase in tobacco prices (through taxes) to reduce tobacco consumption among the younger groups of the population. Studies have also shown that there are geographical variations in smoking behavior (Idris et al. 2007). In Eastern European countries, such as Slovenia, Romania, and Slovakia, MODEL tobacco prevalence in rural and remote areas is higher than in urban areas whereas in Western European countries, such as Germany, Sweden, Finland and Denmark, the opposite has been reported (Idris et al. 2007). This paper will take advantage of the extensive literature analyzing the health effects of tobacco, public policies, The impact and price of increasing elasticities tobacco taxes on the international in Bosnia level discussed andto above Herzegovina is estimated analyze the potential changes using an in household welfare induced extended by an increase cost-benefit in tobacco analysis taxes, as in otheras there is little recent empirical studies evidence on conducted inBosnia Easternand Herzegovina. Europe and 3. regions other Model of the world (Fuchs and Meneses 2017; Pichón-Riviere et al. 2014; Verguet et al. 2015). The paper analyzes three factors to estimate how tobacco taxes could affect The impact of increasing tobacco taxes in Bosnia and Herzegovina is estimated using an extended cost-benefit household income. First, assuming tobacco consumption does not change, tobacco analysis as in other recent studies conducted in Eastern Europe and other regions of the world (Fuchs and taxes directly Meneses affect household 2017; Pichón-Riviere income et al. 2014; as the Verguet share et al. ofThe 2015). household budgets paper analyzes three allocated to factors to estimate ∆)* how tobacco taxes could affect household income. First, assuming tobacco consumption does not change, (2) #$ ' ∗ ) , could tobacco purchases increases with the tax rise. Second, household !" #$ + ∆" medical expenses tobacco taxes directly affect household income as the share of household budgets allocated to tobacco *,, decrease as a result purchases increases withof reduced the tobacco where tax rise. Second, "consumption. #$ is the household share of medical Households product expenses could might i in total also household decrease as experience expenditure a result for a household in a of reduced tobacco a consumption. positive Households income change because of increase; price might also experience additionaland ∆"#$ is positive ayears ofthe change income labor in the change recovered consumption because through of the of additional the good that depends on t years the ofof of labor recovered through the extension product. life expectancy. 5 The aggregate effect of a tax policy is estimated as extension follows: of life expectancy. The aggregate effect of a tax policy is estimated as follows: ∆) ∆"#$ ' ∗ in *tobacco !"#$ + Change , (2) expenditure )*,, Income= Income effect effect = change in tobacco expenditure change in tobacco expenditure 566666666676666666668 + lower medical expenses 566666 (A) + lower medical 66766666668 + rise in income expenses(B) + 566676668 (1) where "#$ is the share of product i in total household expenditure To estimate for a household the variation in cigarette in a decile j; ∆. consumption is the (A) # after the price increase, the mo rise in income (C) (1) (:) change in prices (∆. (@) (/$ ) for decile j, and the share of ciga price increase; and ∆"#$ is the change in the consumption of the # ), the good tobacco that depends price elasticity on the price elasticity of the product.5 in period 0 ("#$ ). The change in expenditure for each household in each decile is prese For the first effect, a partial equilibrium approach total expenditure is used, and averaged soby that theto decile impact quantify onthe overall impact, as follows: Change in tobacco expenditure consumption because of an increase in the price of cigarettes is simulated. This approach 7 ABCD is used To estimate theto evaluate variation in the first-order cigarette effectsafter consumption of a∆ price in change Expenditure the prices. :,; =(( increase, It 1+ the relies ∆P)! model mainly 1 + ε; ∗ considers on∆P' -1)EFGHI JKLJMN:GOPJ the CD changehousehold # ), the tobacco patterns. in prices (∆.expenditure price elasticity The(/ $ ) for decile allows simulation j, and the forshare of cigarette differences in expenditure the responses in period 0 ("#$ ). The change in expenditure for each household in each decile is presented as a share of across consumption total expenditure and averaged deciles by deciletoto reflect quantify the fact the that overall pooras impact, households follows: likely have different price elasticities relative to households Medical with more resources. The different elasticities, expenses BCD A combined ∆ Expenditure =((1 + with the:,;initial ∆P)!1 + ε; ∗ ∆P' consumption patterns across deciles, explain whether -1)EFGHI JKLJMN:GOPJ (3) a price The change in medical expensesCDfrom tobacco-related diseases is estimated using equat reform will be more regressive, more neutral, cost of or more treatment progressive. of tobacco-related diseases for income decile Q is obtained from ad The cost of tobacco-related medical expenses is distributed across income decile Q acco The loss of real consumption arising from the price of households increases that consume in a product tobacco i .is in decile Q obtained Equation (4) shows the income gain Medical expenses the reduction of medical expenses because of reduced tobacco consumption in the lon as follows: The change in medical expenses from tobacco-related diseases is ∆)estimated * using equation (4), where the VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW !" #$ + ∆"#$ ' ∗ )∆),* (2) B ∆ Medical expenditure =(!1 + ε; ∗ ∆P'-1) treatment of )*(∆p_i)/p_(i,o) cost of(ω_ij+∆ω_ij tobacco-related diseases , (2) !" for #$ + ∆"decile income #$ ' ∗ ) Q is *,, :,; , obtained (2) from administrative ∆)* data. EFGHI JKLJMN:GOPJCD *,, !"#$ + ∆"#$ ' ∗ , (2) ∆) The cost of tobacco-related where "#$ is the share medical of product expenses i in A is reduction distributed total household in across tobacco income expenditure consumption decile for Q according a householdin the )*,, into long thewould a run decile sharej; ∆. beis strongly the related to a redu "#$ + ∆"#$ ' ∗ * , (2) where is the share of product i in total household expenditure for a household in a decile j; ∆. # " related diseases. The model assumes that the* a*health effects of# is the tobacco-related Where of households ω_ijconsume that the∆" is and share tobacco of product in decile i Q .in total Equation household (4) shows expenditure the income * for gains household associated with in a a decilediseases )*,, #$ ∆) ∆) ∆) price increase; #$ where is the change " #$ is the in share the of consumption product i of in !"#$ !" total the+!" good ∆" household #$ #$+ #$+'that ∆" ∗ ∆" #$depends #$,∗ ' expenditure '∗ , on (2) , (2)the (2)for price a household elasticity in j; ∆.# is the priceof reduction increase; medical and ∆" expenses #$ is the because change of diminish in reducedthe with consumption tobacco the reduction of consumption the in good tobacco in that the)*,,longconsumption. depends )*,,)*,, on the price term. Even though elasticity this assumption is al household expenditure for a household in decile aofdecile the j; j; ∆p_ ∆.# is product. the price is the price increase; short increase; and and ∆ω_i ∆" is the is the changechange in the consumption of the good that depends on the price elast changesin inthe the consumption of the good 5 term #$ because effects if tobacco-related diseases take some time of the product.5 i in total aj;decile the consumption of the good that depends on the price elasticity where where " of where the #$ is " the " product. #$ #$ share is the is the provides of 5 share product share ofof an product product i3in household total i in total household expenditure household upper-bound estimate Bof the effects of tax increases. expenditure expenditurefor a household for fora household a householdin a decile inina decile∆.#j;is j;the ∆. ∆.# is # that Changedepends ∆ Medical expenditure on =( in tobacco the price expenditure !1 +price ε; ∗ elasticity ∆P' -1)∆" of the product. VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW (4) :,; price increase; price increase;and increase;  and and #$ is ∆" the ∆" #$ is #$ change the is the in change changethe inconsumption inthe the consumption consumption of the ofgood ofthe the that goodgood depends thatthat on depends depends the onprice onthe elasticity price the price elast ela Change in tobacco expenditure EFGHI JKLJMN:GOPJCD of the Change of product. ofthe the in 5 tobacco product. product. 5 5 expenditure A reduction in tobacco consumption To estimate the variation in cigarette consumption in the longIncrease run would in workingbeafterstrongly liferelated the yearsincrease, price to a reduction the model in tobacco- considers the related 3.1.To Change estimate diseases. change inThe model prices the in (∆. tobacco variation assumes ), the To inthat tobacco estimate expenditure cigarette the pricehealth the consumption effects elasticity variation ( / of in ) after for cigarette the price increase, tobacco-related decile j , and consumption diseases the share after the immediately will of model considers the cigarette the priceexpenditure increase, the model considers Change # Change in tobacco Change inin tobacco expenditure tobacco expenditure expenditure $ diminish change with the in prices (∆. reduction in), the tobacco tobacco price consumption. Finally, elasticity Even the model /$ ) for (though decile estimates j,the this assumption and the share impact ison of cigarette implausible in the expenditure onsumption after the price increase, the model considers in period 0 ("#$ ). The# change the change inin prices expenditure (∆.#for ), the each tobacco household price in elasticity each (/$ ) decile decile jarising is income for presented , and as athe from share sharethe of of increase cigarette inexpend workin short term To in because estimate period 0 changes (the " variation ). The the effects in change in cigarette in if 6). tobacco-related expenditure To estimate consumption for each diseases the increase after household take the in some working price each time decile to years, increase, is materialize, the the presented years model asof it a life share lostof (YLL) from tobacc elasticity (/$ ) for decile j, and the share of cigarette total expenditureand expenditure #$ estimate To averaged To inTo estimate period the estimate by 0 variation decile ( the the).variation to The in variation quantify cigarette change in in the cigaretteconsumption incigarette overall expenditure consumption consumption impact, for after as each the after follows: afterprice household the the increase, priceprice in increase,the ∆) increase, model the* the considers model model considers conside the provides expenditureestimate an upper-bound of the effects" are#$ of tax distributed increases. across deciles (Q !" +each ∆" decile #$ ' ∗ ) of is, presented (2) as a shar ture for each household in each decile is presented total considers the change as a share and averaged of change change in change totalin prices prices inbyin expenditure decile ( prices prices ∆. (∆p_, # ), to the (∆. the and(∆. quantify tobacco # ),# the ), the tobacco averaged the tobacco price tobaccoby overall price elasticity price price decile impact, elasticity /$proportionally () elasticity toelasticity quantify )as for(/follows: (ε_j $ decile ) $for (the / ) for for j,decile decile overall and decile to #$ the j,the and j, and j, impact, number share and thethe as of share cigarette share follows: *,, households ofof expenditure cigarette cigarette that c expendi expen (equation 5). Subsequently, the income lost is estimated as the average income per hou to quantify the overall impact, as follows:Increase in working life in period years inin 0 period ("#$ period 0).0 ("The "). (#$ #$change The ). The in expenditure change change where " inin expenditure isexpenditure the for each share A for of for household eacheach product i inin household household each total in decile each in householdeach is decilepresented decile is is as for presented expenditure presented a share as aas of a shar housea sh the share of ∆cigarette Expenditure expenditure =( ( 1 + in Overall, ∆P period )! 1 the + 0 εmodel ∗ (ω_). ∆P' #$anticipates -1) The change that EFGHI JKLJMN:GOPJ BCD ABCD in income expenditure will increase foras each (3) the number of years lost beca total expenditure total ∆ Expenditure:,; =( 1 +total :,; expenditure expenditure( and averaged and ∆P from)!and averaged +ε 1 price by ; averaged;∗ decileby ∆P'-1) increase; by to decile quantify decile and to∆" quantify to the quantify is theoverall the the change impact, overall overall in impact, the as impact, follows: (3) as consumption A asfollows: follows: of the good that dep deaths tobacco ∆ Expenditure:,; =((1 +EFGHI JKLJMN:GOPJ consumption ∆P)!1 decline. #$+ ε ∗ ∆P' ; CD CD -1) BCD (3) P)!1 + ε; ∗ ∆P'-1) ABCD Finally, the model (3) estimates the impact on income arising of the from product.the increase 5 in working yearsEFGHI JKLJMN:GOPJ (equation CD EFGHI JKLJMN:GOPJCD ABCD ABCD ABCD estimate 6). To 3  increase of the discussion For a detailed working in the methodology,∆years, ]^_`Qab cde_f ∆ Expenditurethe see∆ years Expenditure Expenditure Coady of et:,; =( al. life (1:,; =( (2006); lost +# = ( =(∆P1 Kpodar(YLL) (chh ( + 1)! +1 ∆P + ∆P (2006). # from ∗ε )! iℎe_d ^k il^`d_f )! 1;∗ +∆P' 1 tobacco-related +ε; ε ∗-1)∆P' ∗ ∆P' -1) -1) diseases # )/n^.opeqQ^a# EFGHI JKLJMN:GOPJ EFGHI JKLJMN:GOPJ (3) (5) (3) (3) :,; ; are distributed across deciles (Q ) proportionally to the numberin Change of tobacco households expenditure that consume EFGHI JKLJMN:GOPJ tobacco CD CD CD (equation Medical expensesthe income lost is estimated as the average income per household in decile Q . 5). Subsequently, Medical expenses uFPv:Mw xJHPWB ∗EFGHI yKLJMN:GOPJB Overall, the model anticipates that Medical incomeexpenses will ∆increase Income To: =( as the !1number estimate+ ε; ∗the ∆P' -1) lostin ofvariation years cigarette because consumption after the price of premature (6) increase EFGHI JKLJMN:GOPJ B 11 The change deaths from tobacco in medical expenses consumption decline. from tobacco-related change diseases in pricesis (∆. estimated # ), the tobacco using equation price elasticity (4), where (/$ ) the for decile j, and the sha The change in medical Medical expenses Medical expenses from tobacco-related diseases is estimated using equation (4), where the expenses cost of treatment of The Medical tobacco-relatedchange expenses in medicalfor diseases in income expenses period 0fromdecile ( change infrom Q is obtained "#$ tobacco-related ). The diseases expenditureadministrative is estimated for each data. using householdequation (4), where in each decil cost of treatment of tobacco-related5 diseases for income decile Q is obtained from administrative data. bacco-related diseases is estimated using The (4), equation ]^_`Qab cde_f costwhere of= tobacco-related the (chh ∗ cost medical of iℎe_d ^k il^`d_f treatmentexpenses For of a is distributed tobacco-related total detailed )/n^.opeqQ^a expenditure discussion acrossdiseases of and the income (5) for averaged methodology, decile income by according Qsee decile decile Coady to Q is et to the obtained quantify al. (2006);share the from overall Kpodar administrative impact, (2006). as foll d The cost# of tobacco-related # medical expenses # is distributed # across income decile Q according to the share of households eases for income decile Q is obtained from administrative that data. The change consumeThe The The in medical change change tobacco cost of in inmedical in expenses medical decile tobacco-related expenses from Q .expenses Equation medical tobacco-related from from(4) tobacco-related tobacco-related shows expenses the is diseases income distributed isgains diseases diseasesestimated across is isestimated usingdecile estimated associated income equation using using with Q (4), where equation equation according (4), (4), to where whe the s households of reduction the ofthat medical consume expenses tobacco because in decileof Q . Equation reduced tobacco (4) shows the income consumption in the gainsterm. long associated with cost of cost treatment cost ofof treatment of tobacco-related treatment of oftobacco-related tobacco-related diseases diseasesfor income diseases for forincome decile income decile Q is obtained decile Q isQ is obtained from from obtained administrative from administrative administrativedata. d ∆)* !"#$ + ∆"#$ ' ∗ , (2) )*,, ∆)* ∆) !"#$ + ∆" #$ ' !" ∗ ∆" #$ + , ' (2) ∗ *, (2) duct i in total household expenditure for a household in a decile j) ; ∆. is the *,, # #$ )*,, Tobacco Taxation: Evidence from ∆)* Bosnia and Herzegovina e change in the consumption of!" the + good ∆"that #$ ' ∗ depends , household on(2)the!"price elasticity ∆) * (2) for a household , a household where is the " share of i in + ∆" #$ ' ∗ in a decile jin #$ product total expenditure for ; ∆. "#$ where #$ is the share of product ) *,, i in total #$household expenditure )*,, # is the a decile j; ∆.# is the price increase; "#$ is the share of product price and ∆" increase; isand i in total householdthe ∆"change is in the the changeconsumption in the of the consumption good ofthat the depends in a decile j; forgood thaton the depends priceon elasticity the price elasticity of expenditure household expenditure for a household ∆.#ais the #$ of the product.where 5 "#$ is the share #$ product i in total household in a decile j; ∆.# is the diture crease; and ∆"#$ is the of the change product. in theand 5 consumption price increase; ∆"#$ is theof the good change in thethat depends on of consumption thetheprice good elasticity that depends on the price elasticity roduct.5 of the product. Change in tobacco Change in expenditure 5 tobacco expenditure cigarette consumption after the price increase, the model considers the ebacco price elasticity in tobacco expenditure (/$ ) for decile Change in tobacco j, and expenditure the share of cigarette expenditure To estimate To the variation estimate the in cigarette variation inconsumption cigarette consumption after the price after increase, the pricethe model the increase, considers model the considers the e in expenditure for each change in prices household (∆. in ), the tobacco each decile price is presented elasticity ( / as ) a share for decile of j , and the share of cigarette expenditure mate change in #prices (∆. ), the tobacco price elasticity $ ( / ) for decile j , and the share of cigarette expenditure ed by the variation decile in to in quantify period 0To ( cigarette " the overall estimate ). The consumption impact, the change # variation in after as follows: expenditure the price in cigarette for each increase, the consumption household $ model considers the after in eachthedecile price is increase, presented theas model a share considers of the in period household n prices (∆.# ), the tobacco price0 (" in ). each elasticityThe change decile (/ $ ) for is in expenditure presented decile price as j, andelasticity for the sharea each share household of total in each expenditure decile is presented and averaged as by of a share (/of cigarette expenditure #$ #$ changeand total expenditure in prices averaged (∆.#by ), the deciletobacco to quantify the overall ) for $ impact, decile as j, and the share of cigarette expenditure follows: d 0 ("#$ ). The change total expenditurefor and averaged by decile to quantify is the overall as impact, asoffollows: inin decile expenditure period to quantify 0("#$ ). The A each the BCDchange household overall impact,in each in expenditure as decile follows: for each presented household ina share each decile is presented as a share of e ∆P)! :,; =((1 +and penditure 1+ε averaged ; ∗by ∆P' -1) decile to quantify the overall impact, (3) as follows: total expenditure and averaged EFGHI JKLJMN:GOPJ CD by decile to quantify the overall impact, as follows: ABCD ABCD ∆ Expenditure :,; =((1 + ∆P ∆ Expenditure )!1 + ε; ∗ )! :,; =((1 + ∆P ∆P' 1+ -1)ε ∗ ∆P'-1)EFGHI JKLJMN:GOPJ ; EFGHI JKLJMN:GOPJ (3) (3) CD CD ABCD ∆ Expenditure:,; =((1 + ∆P∆ )!Expenditure 1 + ε; ∗ ∆P'-1)  =((1 + ∆P)!1 + ε ∗ ∆P'-1) (3) ABCD (3) :,; EFGHI JKLJMN:GOPJCD ; EFGHI JKLJMN:GOPJCD 3.2. Medical expenses Medical expenses Medical expenses ses from tobacco-related The change diseases is estimated in medical using expenses equation from (4),tobacco-related where the diseases is estimated using l expenses o-related diseases for income Medical decile expenses Q is obtained from administrative data. The change in equation medical The change (4), expenses inwhere medical from the expenses tobacco-related cost of treatment diseases is diseases is using estimated of tobacco-related equation diseases (4), for where income thedecile medical expenses is distributed across income decile Q from according tobacco-related to the share estimated using equation (4), where the cost of treatmentcost of of tobacco-related treatment of diseases for tobacco-related income diseases fordecile income Q is decile obtained is from obtained administrative from data. administrative nge in medical tobacco in decile Q expenses The cost i .The ofis from Equation obtained change tobacco-related tobacco-related (4)infromshows medicalmedical the diseases income administrative expenses expenses gains from is is estimated data. associated The cost tobacco-related distributed using with across equation of tobacco-related diseases income (4), is Q where estimated decile the according medical using equation expenses(4), where isdata. the The cost of tobacco-related medical expenses is distributed across income Q∆) * decile Q to the share treatment enses because of tobacco-related of reducedcost tobacco of diseases treatment for consumption of income tobacco-relatedin decile the long isterm. Q diseasesobtained for from income !" administrative + ∆"Q#$is #$decile '∗ data. obtained , according (2) administrative share from to the data. of households that consume tobacco in decile . Equation (4) shows the income gains associated with t of tobacco-relateddistributed across income decile according iincome to the share of households that consume of households that consume tobacco Q in decile Q . Equation (4) shows the ) *,, income gains associated with medical Theof cost expenses of tobacco-related is distributed medical across expenses decile is Q according distributed across to the income share decile Q according to the share the reduction the medical reduction expenses of medical because expenses ofbecause reduced oftobacco reduced consumption tobacco in the long consumption interm. the long term. eholds that consume tobacco in decile where . " Equation is the share (4) shows of product the income i in total gains household associated expenditure with for a household in a decile j; ∆.# is the !1 + ε; ∗ ∆P'-1) tobacco in decile that i. Q Equation B (4) shows the income gains associated with the reduction of VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW of households consume #$ tobacco in decile (4) Q . Equation (4) shows the∆) income gains associated with ction of medical expenses the reductionbecause EFGHI JKLJMN:GOPJ of reduced price of medical increase; tobacco expenses and consumption ∆" because is the of in change reduced the in long !" the tobacco term. + consumption ∆" #$ consumption ' ∗ of * , the good(2) that depends on the price elasticity )*,, in the long term. CD #$ #$ longmedical expenses because toof reduced tobacco consumption VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW B in the longB(4) term. umption in ∆the Medical expenditure run would ∆ Medical expenditure be strongly =(!1 :,;of the+ ε:,; related product. ∗ ∆P' ; =( !1 5a +-1) reduction ε ; ∗ ∆P'-1) in tobacco-  EFGHI JKLJMN:GOPJ (4) EFGHI JKLJMN:GOPJ CD assumes that the health effects of tobacco-related where "#$ diseases is the share willof product immediately B i in total household expenditure for a household in a decile j; ∆.# is the VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW CD cal expenditureA reduction:,; =( !1 in + ε; ∗ ∆P' tobacco -1) consumption in the long run would (4)  VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW be strongly related to a reduction B in tobacco- A ∆ n tobacco consumption. Even though reduction Medical expenditure in tobacco Changethis price consumption assumption increase; :,; =( !1 tobacco + inEFGHI JKLJMN:GOPJ andε is in ∗ the ∆P' implausible ; expenditure ∆"#$ long -1)  is the in CD run change would be the EFGHI JKLJMN:GOPJ strongly in the consumption related of to thea reduction good that (4) in tobacco- depends on the price elasticity related diseases.related The model diseases. The assumes model thatassumes the health that effects the health of tobacco-related effects of diseases tobacco-related CD will immediately diseases will immediately in the effects if tobacco-related tion in tobacco consumption A reduction in the diseases in longof tobacco run take the would some product. consumption time be strongly 5 in to thematerialize, related long runto a itreduction would be in tobacco- strongly related to a reduction in tobacco- diminish with diminish the reduction withthe the in tobaccoin reduction consumption. tobacco Even though consumption. Even this assumption though this is implausible assumption in the implausible isincrease, in the considers the imate of The diseases. the effects modelA of tax increases. assumes that To health estimate effects the tobacco-related ofvariation in the cigarette diseases will consumption immediately after the price short term reduction related because short diseases. term changes in because tobacco The in model the effects changes consumption assumes in the that the if tobacco-related effects in if health long tobacco-relatedeffects diseases run ofwould take tobacco-related some diseases be time take strongly todiseases some related materialize, time will to it the model to immediately materialize, it h with the reduction in tobacco diminish with consumption. the change Change reduction in Even pricesin though (∆.# ),of tobacco in tobacco this the assumption tobacco expenditure consumption. price Even implausible iselasticity though (/ $ ) in this for the decile j, and assumption the share ofin is implausible cigarette the expenditure provides a an upper-bound provides reduction an estimate upper-bound in tobacco-related of the estimate effects of the diseases. tax effects increases. The of tax model increases. assumes that the health effects rm ars because changes in the effects short term because if tobacco-related in period changes 0 (" in #$ ). the diseases The change effects take some in expenditure for if tobacco-related time to materialize, each household diseases it take somein eachto time decile is presented materialize, it as a share of s an upper-bound estimate of an effects theupper-bound To of taxestimate estimateincreases. the variation in cigarette toconsumption after the price follows:the model considers the increase, Increaseof tobacco-related provides in working Increase inlifeworking total years diseases expenditure life years will and ofimmediately the averaged effects by of diminish tax increases. decile with the quantify reduction the overall impact, in tobacco as the impact on income arising from the changeincrease inin prices working (∆.#years tobacco price elasticity (/$ ) for decile j, and the share of cigarette expenditure ), the(equation e in in working working lifethe years, consumption. years years of life lostEven (YLL) thoughfrom this assumption tobacco-related diseasesis implausible in the short term because Finally, the Increase model the in working estimates in period the life impact years 0on ( #$ ). The arising "income change in expenditure from the increase forin each working household years in each decile is presented as a share of (equation ABCD s (Q ) proportionally Finally, to changes the number in model the of estimates effects households if the ∆ impact Expenditure that tobacco-related consume on income =( ( tobacco diseases 1 arising + ∆P )! take from 1+ some the ε ; ∗ increase ∆P' time -1) in working EFGHI JKLJMN:GOPJ to materialize, years (equation (3) 6). To estimate the increase total workingin inincrease expenditure years, and the years averaged :,; of life by decile lost of (YLL) to quantify from the overall tobacco-related impact,diseases asit provides follows: theincome he model lost estimates is 6). the estimated To estimate impact as the on the income average arising income per working from the household years, increase in the in decile years working Q . life yearslost (YLL) (equation from tobacco-related CD diseases are distributedanFinally, acrossthe model deciles estimates (Q ) deciles the proportionally impact to on the incomenumber arising of from the households increase that consume in working that tobacco years (equation stimate the increase s that income in upper-bound are will increase 6). distributed working Toas theyears, estimate number across the estimate the of years increase years ofin of(Q )the life lostworking proportionally lost because effects (YLL) years, of from of prematurethe tax years increases. the totobacco-related number of life households ofdiseases lost (YLL) from tobacco-related consume tobacco (equation 5). Subsequently, (equation 5). Subsequently,the income the lost is income estimated ∆ Expenditure lost as the is estimated :,; =(( average 1+ asconsume the ∆P income )! average 1 + tobaccoper ε; ∗ household ∆P'-1)per income in decile household EFGHI JKLJMN:GOPJ ABCD Q . in diseases decile Q . (3) across deciles ributeddecline. mption are ( Q ) proportionally distributed across to the deciles number(Qwill of ) proportionally households tonumberthat theas number households of lost that consume tobacco Overall, the model Overall, anticipates the model that anticipates income that increase income will asincrease the the years ofnumber ofbecause years lost of premature because of premature CD n 5). Subsequently, the income lost is estimated as the average income per household in decile Q . ∗ the deaths from model anticipates iℎe_d ^k il^`d_f 3.3. (equation deathstobacco that Overall, Increase from income # )/n^.opeqQ^a 5). consumption the model Subsequently, tobaccoMedical will # inconsumption increase years anticipates the income expenses decline. as the (5) that of decline. incomeof number working lost is estimated years will life lost because increase as the average income per household in decile Q . as the number of premature of years lost because of premature rom tobacco consumption deaths decline. from tobacco consumption decline. ]^_`Qab cde_f The change ]^_`Qab cde_f# = (chh# ∗ iℎe_d ^k il^`d_f# )/n^.opeqQ^a# diseases Finally, the # = model (chh # ∗ estimates iℎe_d ^k il^`d_f in medical the impactexpenses # on )/n^.opeqQ^a from income tobacco-related # arising (5) from the is estimated (5)increase using equation (4), where the in working 1)  uFPv:Mw xJHPWB ∗EFGHI yKLJMN:GOPJB cost Medical of treatment(6) expenses of tobacco-related diseases for income decile Q is obtained from administrative data. = (chhyears ab cde_f# EFGHI JKLJMN:GOPJ (equation ∗ iℎe_d ^k il^`d_f # ]^_`Qab cde_f 6). To estimate # )/n^.opeqQ^a the #increase (5)in )/n^.opeqQ^a working years, the years (5) of life lost (YLL) from The # = cost(chh ∗ iℎe_d ^k il^`d_f of# tobacco-related medical # expenses is distributed across income decile Q according to the share B uFPv:Mw xJHPW B ∗EFGHI yKLJMN:GOPJ uFPv:Mw xJHPW # ∆ Income =( !1 + ε ∗ ∆P' -1) B ∗EFGHI yKLJMN:GOPJ B B (6) Finally, ∆ : Income tobacco-related the ; model =( !1 + estimates ε  The ofdiseases ∗ ; change households ∆P' the -1) are  in thatimpact medical distributed consume on income expenses across arising from deciles from tobacco-relatedthe increase ( Q.) Equation proportionally (6) diseasesin working isto theyears estimated number(equation using 6). equation To withthe (4), where : EFGHI JKLJMN:GOPJ Btobacco EFGHI JKLJMN:GOPJ in B decile (4) shows the income gains associated me =( !1 + ε ∗ ∆P' estimate -1)  the uFPv:Mw xJHPW increase B ∗EFGHI yKLJMN:GOPJ the cost in reduction working of treatment of years,B of medical uFPv:Mw xJHPW the years tobacco-related expenses B (6) of life because ∗EFGHI yKLJMN:GOPJ lost (YLL) diseases of reduced B from for tobacco-related income tobacco decile Q consumption is diseases obtained in are the distributed fromlong administrative term. data. e methodology, see Coady : ; of∆ households Income et al. (2006); =(!1Kpodar + ε; consume that :EFGHI JKLJMN:GOPJ ∗ ∆P' (2006). B-1) tobacco (equation 5). Subsequently, (6) income lost is the across deciles The cost (f ) proportionally of tobacco-related to the number EFGHI JKLJMN:GOPJofmedical households expenses B that is distributed consume tobacco across income 5). (equation decile Q according to the share Subsequently, 5 For a detailed discussion of the methodology, see Coady et (2006); al.Coady Kpodar (2006). estimated 5 For the income a detailed as lost the discussion is average of ∆ estimated households of the asincome the that methodology, average per consume household see income tobacco peret in al. in household decile decile (2006); inQ .. Equation Overall, the VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW Kpodar decile (2006). f . Overall, model (4) shows the model theanticipates income anticipates B gains associated with that Medical expenditure :,; =(!1 + ε; ∗ ∆P'-1) (4) etailed discussion of the income methodology, will increase the Coady of see increase asreduction the et the al. number (2006); of medical of Kpodarsee years expenses (2006).lost because because of of reduced premature tobacco EFGHI JKLJMN:GOPJ deaths consumption from CD tobacco in the consumptionlong term. that 5 For a detailedwill income discussion A reduction as the number inmethodology, tobacco consumption of Coadyyears et al. in lost the because (2006); long run Kpodar would of premature (2006). deaths from be strongly related to a reduction in tobacco- decline. tobacco consumption related diseases. decline. ∆ Medical expenditure The model assumes =(!1 +that ε ∗ the ∆P' -1) effects health of tobacco-related diseases VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW B will immediately (4) diminish with the reduction :,; in tobacco ;consumption. Even though this assumption EFGHI JKLJMN:GOPJ CD is implausible in the xyyJ ∗z{|}~ Ä zÅÇ~}ÉJ short term A reduction because nopqfrs tuvpw in tobacco changes Dconsumption = in the effects in the if tobacco-related long run would diseases (5) take be strongly some to related time to materialize, a reduction it in tobacco- ÑIÖÜ|áDàJ providesrelated an upper-bound diseases. The model estimate assumes of thethat effectsthe of healthtax increases. effects of tobacco-related diseases will immediately diminish with the reduction in tobacco consumption. âZeäcaã å^\eh ∗YZ[\] ç_`^abc[de^ Even though this assumption is implausible in the ∆ Income Increaseshort term inc =( B1 working + εQ ∗ because ∆PG changes life -1) in the effectsV if tobacco-related years V diseases take (6) some time to materialize, it YZ[\] ^_`^abc[de^V provides an upper-bound estimate of the effects of tax increases. Lastly,the Lastly, the totaltotal income Finally, income the gains gains model in each in each estimates income incomethe impact group aregroup are estimated on income estimated arising the by adding by from adding the increase results of the the results in increase working inof years (equation tobacco expenditures, the increase in the 6). tobaccoIncrease To reduction estimate in working the increase in medical treatments, expenditures, life in years working the reduction and the gain years, thein working in medical years of life treatments, lost years (equation (YLL) and1).from tobacco-related the gain in diseases are distributed across deciles (Q ) proportionally to the number of households that consume tobacco working years (equation Finally,5). (equation 1). the model estimates Subsequently, the incomethe impact lost on income arising is estimated as the from average theincome increase per in household working years (equation in decile Q. 4. Data and descriptive statistics Overall,6). To theestimate model the increase anticipates in working that income will years, the years increase as the of number life lost of (YLL) years from tobacco-related lost because diseases of premature Data on household deaths are distributed from consumption tobacco andacrossconsumptiondeciles (on expenditure ) proportionally Qdecline. tobacco products to in the number Bosnia andof householdscome Herzegovina consume tobacco that from (equation 5). Subsequently, the income lost is estimated as the average income per household in decile Q . three waves of the Household Budget Survey (2007, 2011 and 2015). The survey covers consumption Overall, the model ]^_`Qab cde_f anticipates that income will increase as the number of(5) # = (chh# ∗ iℎe_d ^k il^`d_f# )/n^.opeqQ^a# years lost because of premature expenditure information deaths on from a range tobacco commodities decline. of consumption including tobacco products, using a 14-day reference period. It also collects household characteristics such as age, gender and educational qualifications. The uFPv:Mw xJHPWB ∗EFGHI yKLJMN:GOPJB advantage of surveys ∆ Income is that ]^_`Qab cde_f : =(one !1may + ε;detect ∗ ∆P'the -1) price paid by consumers and account for (6) (5) and sales. promotions 12 // Model # = (chh# ∗ iℎe_d ^k il^`d_f EFGHI JKLJMN:GOPJ # )/n^.opeqQ^a B # However, the price paid is not independent of the characteristics of the buyers, for instance, as heavy smokers may consume cheaper brands, buy greater quantities, shop at lower-priced uFPv:Mw xJHPW B ∗EFGHI yKLJMN:GOPJ retailers, B or engage in tax-avoiding 5 For ∆a Income detailed : =( !1 + ε; of discussion ∗ ∆P' -1) the methodology, see Coadydepends et al. (2006); Kpodar (2006). (6) behaviors (WHO 2011). In addition, the accuracy of the responses EFGHI JKLJMN:GOPJ B on the ability of household head to accurately respond on questions about expenditure by other household members. Despite these 13 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 4 DATA AND DESCRIPTIVE STATISTICS Data on household consumption and expenditure on tobacco products in Bosnia and Herzegovina come from three waves of the Household Budget Survey (2007, 2011 and 2015). The survey covers consumption expenditure information on a range of commodities including tobacco products, using a 14-day reference period. It also collects household characteristics such as age, gender and educational qualifications. The advantage of surveys is that one may detect the price paid by consumers and account for promotions and sales. However, the price paid is not independent of the characteristics of the buyers, Lastly, the total income gains in each income group are estimated by adding the results of the increase in Lastly, Lastly, the total thegains income income income total in gains ingroup each income areadding group by estimated by adding the increase results ofinthe increase in for instance, tobacco as heavy expenditures, smokers Lastly, the the each reductionmay total in consume income gains medical are estimated cheaper in each treatments, brands, income and thegroup gainbuyare in the results greater estimated working of quantities, by(equation years the shop adding the results of the increase tobacco expenditures, tobacco expenditures, the reduction the reduction in medical in medical treatments, Lastly, and the total the gain income in working gains 1). (equation years in each income 1). group are1). estimated by adding the tobacco expenditures, the treatments, and reduction in the gain medical in working treatments, andyears (equation the gain in working years (equation 1). me gains in each income group are estimated at bylower-priced retailers, adding the results of the or engage increase tobacco expenditures, in in tax-avoiding behaviors the reduction (WHO in medical 2011). In treatments, and the gain in wor addition, s, the reduction in medical treatments, and the gain in working years (equation 1). the accuracy 4. Data of Data and the and responses statistics descriptive depends on the ability of household head to accurately 4. Data and statistics 4.descriptive descriptive statistics 4. Data and descriptive respond on questions about expenditure 4.other by statistics Data and descriptive household statistics members. Despite these descriptive statistics Tobacco price elasticity by decile Tobacco Tobacco price price elasticity elasticity by decile by decile shortcomings, the use of price surveys household Tobacco to elasticity calculate decile price elasticities is a common by Tobacco price elasticity by decile elasticity by decile on in Data on practice Data household Data both household consumption on household developed consumption and consumption ofdeveloping of and expenditure and expenditure on andon tobacco expenditure of countries products on (Fuchs and in Bosnia Bosnia tobacco Meneses and products inHerzegovina 2018).Bosnia and Herzegovina come fromfrom the the Household Data on household Budget consumption survey oftobacco products and expenditure that has in on tobacco been and Herzegovina products in Bosnia and Herzegovin collected in three three waves (2007, 2011 and and 2015). come come from the Household Household Budget survey Budget that Datahas on survey been household that collected has been consumption in collected waves of in and expenditure three (2007, waves 2011 (2007, on2011 2015). tobacco products in and 2015). The advantage of come surveys is from that the one Household may detect Budget the price survey paid by that has been consumers and collected account in three for waves promotions (2007, 2011 and 2015 consumption of and expenditure on tobacco In products The advantage Bosnia in and The Bosnia advantage of surveys Herzegovina, and of Herzegovina is that surveys theone may Household is that detect come onethe Budgetfrom may price the detect paid Survey Household the by price consumers does paidBudget not by and contain survey consumers account datathatand for onhas been account promotions prices collected in three wave for promotions and sales. However, Theprice the advantage paid oftied is surveystopaid one may detect is thatendogeneity certain the price concerns paid by because consumers heavy smokers and mayaccount for promotion sehold Budget survey that has been collected and in three sales. waves and However, (2007, sales. the2011However, priceand paid the 2015). is price tied to The advantage is tied to certainof surveys is that endogeneity one may concerns detect becausethe price heavy paid may and by consumers smokers of consume tobacco cheaper products. and brands, sales. buy Therefore, However, greateran thecertain quantities, indirect price shop method endogeneity paid at is lower-priced usingtied to concerns unit certain values because endogeneity retailers,is used. heavy engage concerns The in smokers because tax-avoiding Household may heavy smokers ma veys is that one may detect the price paid by consumers consume cheaper consume and account brands, cheaper for promotions buy brands, greater buy quantities,and greater sales. shop However, quantities, at lower-pricedshopthe price at paid lower-priced retailers, is tied engage toin certain retailers, engage tax-avoiding endogeneity concerns bec in tax-avoiding behaviours (WHO, consume 2011). cheaperthe addition, In addition, brands, accuracy buyof greater the quantities, responses dependsshop at onlower-priced the ability of retailers, household of household engage in tax-avoidin concerns the price paid is tied to certain endogeneityBudget becausebehaviours (WHO, 2011). consume addition, Inaccuracy the cheaper accuracy brands, of the buy greater responses quantities, depends on shop the at lower-priced ability of household retailer behaviours head to Survey (WHO, accurately inheavy2011). Bosnia respond smokers behavioursIn and on may the Herzegovina (WHO, questions 2011). about asks behaviours ofhouseholds In addition, expenditure the responses (WHO, accuracy theother by “how 2011). depends In much of household theon addition, the [was] responses members. ability the spentdepends Despite accuracy in of on the ability of househo these the responses depends o ands, buy greater quantities, shop at lower-pricedhead to retailers, accurately headengage to accurately respond in tax-avoiding on respond about questions on questions expenditure about byexpenditure other household by other household members. Despite members. these Despite these shortcomings, the use headof to accurately household respond surveys calculate to and,on questions price about expenditure elasticities isquestions a commonby other household practice in both members. Despite thes both the 2011). In addition, the accuracy of the responses last 14 days depends shortcomings, onshortcomings, the the on cigarettes ability use of of the and household household use of household tobacco” surveys head to to calculate accurately surveys “how priceto calculate many respond packages elasticities on priceis a elasticities of common about cigarettes is expenditure a common practice [were]in by other practice household in both developed and developing shortcomings, countries the (Fuchs use of and household Meneses, surveys 2018). to calculate price elasticities is a common practice in bot spond on questions about expenditure by other household developed and developed members. developing and Despite developing countries these (Fuchs shortcomings, countries and (Fuchs and Meneses, the 2018). use of household Meneses, 2018). surveys to calculate price elasticities is a c purchased during the 14-day and developed period.”developing These countries questions (Fuchs allow and an Meneses, estimate 2018). se of household surveys to calculate price elasticities is a common practice in both developed and developing countries of (Fuchsthe average and Meneses, 2018). oping countries (Fuchs and Meneses, 2018).price Household The Household The paid by The Budget household Household Budget survey at survey for BiH Budget three for BiH asks households survey different asks households for BiH asks points in “how time “how much households and much to [was] “how estimate [was] spent much spent in in the last [was] individual the last spent 14 14 in pricedays days the on on last 14 days on cigarettes and and tobacco” Theand, Household “how many Budget many survey for packages BiH cigarettes of cigarettes asks households [were] “how much purchased during [was] spent 14-day the 14-day in the last 14 days o cigarettes cigarettes tobacco” and and, tobacco” “how The Household “how and, packages many of packages Budget of [were]survey cigarettes for BiH purchased [were] asks purchased during households the “how during much the [was] spe 14-day elasticities. period.” 4 These Outliers questions cigarettes are eliminated allow and us to tobacco” for estimate each theand, year “how average that many priceare three paid packages by standard household of cigarettes deviations at three [were] from different purchased the points during the 14-da get survey for BiH asks households “how much [was] spent period.” period.” These questions in the These last 14 questions allowThese days on us to estimate allow cigarettes us the to estimate average and price thetobacco” average paid the and, price by household “how paid by manyhousehold at three packages different at of three points cigarettes different [were] points purc in time and to to estimateperiod.” individual questions price allow elasticities. us 6 toWe estimate eliminate average outliers pricethat paid areby household three standard at three different poin co” and, “how many packages of cigarettes mean in[were] time purchased underand in estimate time the andindividual during assumption theestimate to 14-day that these price individualperiod.” purchases elasticities. price These 6 tend We questions elasticities.to 6 allow reflect eliminate We data outliersus to eliminate estimate problems. that outliers are the Once three average that a standardareprice three paid by househol standard deviations from the in time mean, under and thetoassumption estimate individual that these price purchases elasticities. tend 6 We eliminate reflect to reflect datato problems.outliers Oncethat are three standar ions allow us to estimate the average price paid by household deviations from deviations atthethreemean, from different under the mean, points the under assumption inthe time that and purchases assumption these to estimate that these tend individual purchases to price tend data elasticities. reflect data problems. 6 Once We eliminate problems. Once outliers measure a measure of oftheprice price of deviations of cigarettes cigarettes from in BiH the in mean, Bosnia is obtained, under and assumption Herzegovina thethe tobacco price is that these purchases obtained, elasticities the for tobacco total tend to population reflect price is data problems. Onc mate individual price elasticities. We eliminate 6 a measureoutliers athat of price measureare of price three ofacigarettes standard inofBiHcigarettes is deviations in BiH obtained, theis from obtained, tobacco the mean,the tobacco price under the elasticities priceassumption for elasticities total that populationfor these totalis purchases population tend is to ref estimated using the measure following of price of cigarettes in BiH is obtained, the tobacco price elasticities for total population equation: mean, under the assumption that these purchases tend tousing estimated elasticities forestimated reflect the thedata total using problems. following populationthe following Once equation: is estimated a measure equation: using of price of cigarettes the following in BiH is obtained, the tobacco price elasticiti equation: estimated using the following equation: f cigarettes in BiH is obtained, the tobacco price elasticities for total population is estimated using the following equation: ollowing equation: lnz z#{ = =| |} + +ln |z pan =∗ ∗ | }##++| | Å#{∗ pan + Ç + |Ä Å#{ + Ç#{ (7) (7) ln #{ } | ~ pan ~#{ ln z  + #{ = Ä |~Ä |Å ++ } #{ |~Ç # #{ pan #{ ∗ # + |Ä Å#{ + Ç#{ (7) (7) ln z#{ = |} + |~ pan ∗ # + |Ä Å#{ + Ç#{ z#{ = |} + |~ pan ∗ # + |Ä Å#{ + Ç#{ where z where z#{ is Q_ is the where isthe the quantity (7) z#{ is quantity ofthe cigarette of quantity cigarette packs smoked of cigarette packs per month packs smoked month smoked per by household per month month by byhead in income QQ in household income head decile iQ in d; income decile d; where #{ quantitywhere of z is thepacks cigarette #{ cigarette quantity smoked per of cigarette packs by household smoked perhousehold head month head by household decile d; head Q in income decile n the average n the averagen monthly the average monthly price price of monthly of cigarette price pack pack (bothz of where cigarette (both imported imported #{ is the pack andimported quantity (both and domestic); of cigarette domestic); # and  the domestic); the consumption packs smoked consumption  # the decile per month by household consumption decile decile h inof income decile d; n the P the average average monthly monthly price of cigarette price of cigarette pack (both pack imported (both # and domestic); imported and # the consumption deci ntity of cigarette packs smoked per month by individual household of individual Q ; Å#{head Q of; Å Q the in individual individual income #{ the individual Q ; Å decile characteristics the d; individual n the average (age, education, characteristics #{ characteristics (age, education, location, gender). monthly (age, price location, education, of cigarette gender).location, pack (both gender). imported and domestic);  of individual Q ; Å#{ the individual characteristics (age, education, location, gender). hly price of cigarette pack (both imported and domestic);  domestic); theconsumption D# the consumption deciledecile of of individual Q individual i;; X_i Å#{ thethe individual individual characteristics characteristics (age, (age,education, location, gend Table he individual characteristics (age, education, location, Table 1 1 shows gender). shows Table the the tobacco 1 showsprice tobacco price elasticity the tobacco elasticity across price across income elasticity income acrossdeciles. deciles. income The average The averagedeciles. price price The average elasticity elasticity -0.32 is -0.32 price is elasticity is -0.32 education, which is in location, line with Table gender). estimates 1 shows obtained theby tobacco Fuchs price and elasticity Menenes foracross Moldova income (2018) deciles. and The are not average far from price elasticity is -0.3 which which is in line with estimates is in line with obtainedestimates Table obtained byestimates Fuchs and 1 shows by Fuchs Menenes the and fortobacco Menenes Moldova price for elasticity Moldova (2018) and across are (2018) income not far(2018) andfromaredeciles. The averag not far from those estimated by which Chaloupka, is inet.line al. with (2010). As obtained expected, lower by Fuchs income and Menenes deciles exhibit for Moldova higher elasticities and are not far from obacco price elasticity across income deciles. average Theestimated those those price estimated elasticity by Chaloupka, by iset.Chaloupka, -0.32 al. (2010). which et. expected, As is in line al. (2010). Aswith lower estimates expected, income lower obtained deciles income exhibit by Fuchs deciles higher and Menenes exhibit elasticities for Moldova (2 higher elasticities relative richer to richer deciles.those For estimated instance, by the Chaloupka, poorest et. al. decile has (2010). As expected, amedium-bound medium-bound lower income elasticity of− − 0.73, deciles whereas exhibit higher elasticitie estimates obtained by Fuchs and Menenes for Moldova relative to relative (2018) andto deciles. are richer not For deciles. far instance,fromthe instance, Forpoorest those estimated the poorest decile has a by Chaloupka, decile et. has a medium-bound al. (2010). elasticity of As expected, elasticity 0.73, whereas of − lower 0.73,income whereasdeciles the richest has an relative elasticity to of richer − 0.10. deciles. To showFor instance, the effect theof poorest different decile has scenarios, a medium-bound we simulate a elasticity lower of −0.73, wherea haloupka, et. al. (2010). As expected, lower income deciles the exhibit the richest has an elasticity richest higher has an elasticity elasticities of −has 0.10. To of relative − 0.10. show the toTo richer effect show of deciles. the different For effect instance, of different scenarios, the we poorest scenarios, simulate decilewe has a simulate a lower we simulatemedium-bound a lower el bound elasticity elasticity andof theupper an richest bound an elasticity elasticity. of − These 0.10. estimates To show have the effect of differences different ofdifferences -0.2 and +0.2 scenarios, +0.2 with a low les. For instance, the poorest decile has a medium-bound bound bound elasticity and elasticity an − 0.73, upper and whereas bound an upperelasticity. theThese bound richest has an elasticity. estimates elasticity Thesehave estimates of −0.10. differences have of To -0.2 show and the of -0.2effect with different of +0.2 and with scenar the elasticity estimated bound in eq.(7). elasticity and an upper bound elasticity. These estimates have differences of -0.2 and +0.2 wit eq.(7). asticity of −0.10. To show the effect of different scenarios, the elasticity we simulate estimated a lower in eq.(7). bound elasticity and an upper bound elasticity. These estimates have differenc 4  the elasticity estimated in The purchased quantity of the elasticity cigarettes estimated in eq.(7). and total an upper bound elasticity. These estimates have differences of -0.2 and +0.2 withexpenditure the on cigarettes elasticity during 14-day estimated period are converted to monthly terms. in eq.(7). Cigarette prices are obtained by converting nominal monthly expenditure on cigarettes to real monthly expenditure using ed in eq.(7). national monthly CPI index. Each pack of cigarettes is expressed in 2010 prices. 15 purchased The purchased The 6 quantity The of cigarettes purchased 6 quantity of cigarettes quantityand and total of total expenditure cigarettes and total expenditure onexpenditure during cigarettes during 14-day period on cigarettes period are during are converted 14-day to period to are converted to of cigaretteson cigarettes 14-day converted 6 monthly terms. 6 Cigarette The purchased prices quantity obtained are obtained by and total expenditure on cigarettes during 6 converting nominal monthly expenditure on cigarettes to real 14-day period are converted monthly terms.monthly Cigarette terms. prices monthly Cigarette are terms. prices Cigarettebyare obtained The converting prices purchasedby converting quantity nominal obtained are 04. of monthlynominal cigarettes expenditure by converting monthly nominal expenditure and total on expenditure cigarettes monthly on to oncigarettes real expenditure to during cigarettes real 14-d monthly tity of cigarettes and total expenditure on cigarettes during monthly expenditure monthly 14-day expenditure using period using aremonthly expenditure converted monthly CPI using CPI toindex index for monthly for COICOP CPI index monthly COICOP for terms. 04. Each COICOP Each pack Cigarette of 04.cigarettes Each prices pack are expressed is expressed obtained by in 2010 in2010 of cigarettes is prices. expressed converting in on nominal cigarettes 2010 to re prices. expen monthly monthly expenditure using monthly CPI indexpack of cigarettes for COICOP 04. is Each pack of prices. cigarettes is expressed in 2010 price Tobacco Taxation: Evidence from Bosnia and Herzegovina 4.1. Descriptive statistics Figure 1 presents trends in cigarette prices in Bosnia and Herzegovina in 2009, when the Law on Excise Duties on cigarettes was introduced. Over the years, one can see a clear upward trend in the prices which is mainly caused by constant increase in the excise burden that on average accounts around 76 percent of the retail price. Figure 1. Trends in cigarette prices, Bosnia and Herzegovina 7 66.7 KM 6 5 4 33.2 Price annual growth rate 3 25.0 Retail price of 20 cigarettes per pack 2 13.6 11.7 1 8.3 7.4 9.9 0 Source: Calculations based on official data of the -1 Indirect Taxation Authority. -12.0 -2 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Figure 2. Average price paid by households for a 20-cigarette pack 4.0 Average price (KM) 3.5 2007 3.0 2011 2.5 2015 2.0 1.5 Source: Calculation based on the Household Budget Survey. 1.0 Note: Deciles were created using per capita 0.5 household expenditure. 0.0 1 2 3 4 5 6 7 8 9 10 Income deciles Figure 2 shows variation in average cigarette (domestic and imported) prices over time and income deciles. While there is no significant variation of cigarette prices across deciles, the prices obtained from the Household Budget Survey follow the same trend as those reported by official statistics in Figure 1, albeit lower on average. These differences are likely caused by household recall error and the way how prices are calculated in the Household Budget Survey. Because prices in the Household Budget Survey are calculated by using household expenditure on cigarettes it may also account for illicit trade, therefore underestimating the proportion of income aimed at cigarette consumption and resulting in lower prices. In addition, the Household Budget Survey does not report cigarette prices by brand, and therefore the differences across income deciles may be underestimated as well. 16 // Data and Descriptive Statistics Table 1 summarizes the most important indicators, including total monthly household expenditures and the share of expenditures on smoked tobacco products for year 2015. The share of households using tobacco products is more prevalent among higher income families (47 percent in the highest decile vs 33 percent in the lowest income decile). On average, households spend 6 percent of their income on tobacco. For instance, the lower income households spend slightly more on tobacco (7 percent) relative to 5 percent in higher income households. When looking at monthly mean per person consumption, it is clear that higher income households spend three times more on tobacco and cigarettes consumption. This indicates that higher incomes households choose more expensive products although their consumption of cigarettes is very similar to lower income deciles. In addition, higher income deciles prefer cigarettes consumption over other types of tobacco products such as cigars or leaf tobacco. Figure 3 reports the proportion of households smoking cigarettes over the analyzed period. The reported numbers suggest that tobacco control policies, especially taxation policies have had an impact on consumption of cigarettes across all income deciles, especially for lower income households. Similarly, Figure 4 suggests that the overall share of cigarette consumption in total household consumption has decreased over time as well. Figure 3. Households with Cigarette Consumption 0.70 Proportion of HH 0.60 0.50 0.40 Source: Calculations based on the Household 0.30 Budget Survey. 2007 0.20 2011 Note: Deciles were created using per capita house- 0.10 2015 hold expenditure. 0.00 1 2 3 4 5 6 7 8 9 10 Income deciles Figure 4. Share of Cigarette Expenditures in Total Household Consumption 3.5 Proportion of expenditures 3.0 Source: Calculations based on the Household 2.5 Budget Survey. 2.0 Note: Deciles were created using per capita 1.5 household expenditure. The proportions of cig- 2007 arette consumption are calculated for all house- 1.0 2011 holds, regardless of cigarette consumption. 0.5 2015 0.0 1 2 3 4 5 6 7 8 9 10 Income deciles 17 Tobacco Taxation: Evidence from Bosnia and Herzegovina Table 1. Baseline Descriptive Results, by Decile DECILE 10 AVERAGE DECILE 4 DECILE 8 DECILE 6 DECILE 9 DECILE 3 DECILE 5 DECILE 2 DECILE 7 DECILE 1 INDICATOR Household per capita monthly 81 122 150 175 202 235 274 325 405 687 298 expenditure (US$) Household total monthly expenditure (US$) 309 456 526 561 613 694 740 834 977 1343 760 Proportion of households smoking cigarettes 0.21 0.26 0.29 0.30 0.35 0.35 0.34 0.37 0.40 0.43 0.34 Proportion of households smoking all 0.33 0.38 0.38 0.38 0.44 0.43 0.40 0.44 0.45 0.47 0.42 tobacco products Proportion of expenditure on cigarettes 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.05 0.06 0.05 0.06 Proportion of expenditure on all 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.05 0.06 0.05 0.06 tobacco products Amount spent on cigarettes (US$/month) 22 34 34 36 40 43 48 50 56 62 46.64 Amount spent on all tobacco products (US$ 23 34 33 36 39 43 47 48 56 61 45.19 /month) Cigarettes smoked in the last seven days 76 90 91 84 92 91 92 94 97 106 93 (only smokers) Age when started to smoke 19 19 19 19 20 20 20 19 20 20 20 Household size 3.93 3.80 3.51 3.23 3.01 2.97 2.72 2.57 2.41 2.05 3.09 Proportion of households with woman head 0.22 0.18 0.18 0.20 0.22 0.25 0.27 0.32 0.31 0.36 0.26 Proportion of households with higher 0.04 0.04 0.06 0.06 0.08 0.10 0.12 0.15 0.18 0.29 0.13 educational attainment Prop. Urban area 0.29 0.34 0.33 0.37 0.38 0.42 0.46 0.47 0.51 0.59 0.43 Source: Estimates based on 2015 version of the Household Budget Survey. Note: Deciles created using per capita household expenditure. The proportions of cigarette/tobacco consumption and expendi- ture levels/shares are calculated only for households reporting positive cigarette/tobacco consumption. 4.2. Tobacco price elasticity, by decile Table 2 shows the tobacco price elasticity across income deciles across countries. It is evident that in all countries reported, the elasticity estimates are larger for lower income deciles suggesting higher price elasticity for those income groups. 18 // Data and Descriptive Statistics Table 2. Cross-Country Elasticity Estimates, by Decile 1 2 3 4 5 6 7 8 9 10 Bangladesh −0.81 −0.66 −0.63 −0.57 −0.47 −0.42 −0.42 −0.32 −0.30 −0.25 Chile −0.64 −0.58 −0.52 −0.47 −0.41 −0.35 −0.29 −0.24 −0.18 −0.12 Indonesia −0.64 −0.59 −0.55 −0.53 −0.52 −0.50 −0.49 −0.48 −0.47 −0.46 Moldova −0.51 −0.39 −0.40 −0.34 −0.32 −0.32 −0.32 −0.25 −0.24 −0.26 Russian Federation −0.48 −0.41 −0.38 −0.34 −0.32 −0.31 −0.29 −0.26 −0.25 −0.21 South Africa −0.36 −0.26 −0.24 −0.31 −0.34 −0.17 −0.24 −0.21 −0.13 −0.22 Ukraine −0.59 −0.51 −0.52 −0.46 −0.44 −0.43 −0.42 −0.41 −0.36 −0.33 Source: World Bank. In the case of Bosnia and Herzegovina, the average price elasticity reported in Table 3 is -0.56 which is higher than other estimates obtained for Ukraine and Moldova (Fuchs and Meneses, 2017; 2018) and those reported in Table 2. However, there are not far from those estimated by Chaloupka et. al. (2010) for low and middle-income countries. As expected, lower income deciles exhibit higher elasticities relative to richer deciles. For instance, the poorest decile has a medium-bound elasticity of −1.08, whereas the richest has an elasticity of −0.34. Table 3. Cigarette Price Elasticities, by Decile PRICE 1 2 3 4 5 6 7 8 9 10 NATIONAL ELASTICITY Lower bound −1.182 −0.970 −0.773 −0.784 −0.681 −0.656 −0.606 −0.530 −0.530 −0.402 −0.603 Medium bound −1.081 −0.871 −0.687 −0.700 −0.604 −0.580 −0.532 −0.461 −0.461 −0.336 −0.557 Upper bound −0.980 −0.773 −0.602 −0.615 −0.527 −0.504 −0.458 −0.392 −0.392 −0.271 −0.511 Source: Estimates based on the Household Budget Survey 2007, 2011, 2015. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average differences of −0.2 and +0.2, respectively, with the medium-bound elasticity. To show the effect of different scenarios, a lower-bound elasticity and an upper-bound elasticity are simulated. Lower-bound elasticity tends to reflect income groups that would not change consumption patterns, such as rural residents or older individuals, while the upper-bound elasticity tends to show a longer-term scenario, reflecting the effect the tobacco tax would have on younger individuals. After a few decades, only these would still 19 Tobacco Taxation: Evidence from Bosnia and Herzegovina be alive; the total average effect of the price increase would therefore be approximated more accurately by the upper-bound price elasticity. Table 4. Deaths, by Gender, Illnesses Related to Tobacco Consumption, Bosnia and Herzegovina, 2015 WOMEN MEN TOTAL Cerebrovascular diseases 2,670 1,999 4,668 Ischaemic heart disease(s) 2,092 2,533 4,626 Lung cancer 406 1,212 1,618 Chronic lower respiratory diseases 312 466 778 Stomach cancer 200 329 530 Pulmonary heart disease and diseases of pulmonary circulation 215 216 431 Influenza and Pneumonia 175 197 372 Pancreatic cancer 143 213 356 Leukemia 99 105 204 Bladder cancer 48 146 194 Kidney cancer 48 89 137 Esophagus cancer 22 43 65 Uterus cancer 27 0 27 Tongue cancer 3 19 22 Other forms of heart disease 3,410 2,611 6,021 Other respiratory diseases 94 116 210 Other acute lower respiratory infections 3 3 6 Total 9,967 10,300 20,267 Source: Institute of Public Health, Federation of Bosnia and Herzegovina. Note: Data for Bosnia and Herzegovina are obtained by dividing estimates for the Federation of Bosnia and Herzegovina by the share of the country’s population in the Federation of Bosnia and Herzegovina. 4.3. Mortality and morbidity Data on mortality used in the estimations are obtained from Tobacco Atlas (2018) based on the methodology developed by the Global Burden of Disease (GBD).5 Estimates for 2015 indicate that 8,725 deaths were caused by tobacco-caused diseases. For illustrative and comparative purposes, detailed data obtained from the Institute of Public Health of the 5  GBD estimates are based on over 80,000 different data sources and therefore may differ from national statistics due to differences in data sources and methodology. 20 // Data and Descriptive Statistics Federation of Bosnia and Herzegovina are presented on a number of illnesses associated with tobacco consumption, but which also include deaths resulting from other causes (for example, cerebrovascular diseases could also be caused too by high blood pressure, diabetes, or high blood cholesterol). For this reason, the estimates reported in the Tables 4 and 5 are higher than the ones coming from the Tobacco Atlas. The illness-level mortality and morbidity rates estimates for Bosnia and Herzegovina are obtained assuming that the Republika Srpska incidence is proportional to the population. The data are disaggregated according to the gender of the deceased. According to these calculations, in 2015, approximately 20,267 deaths were attributed to the illnesses associated to tobacco consumption (Table 4) of which 10,300 were men. Ischemic heart disease, other forms of heart disease and cerebrovascular diseases are among the most prevalent forms. Similarly, more than 287 thousand cases for the illnesses associated to tobacco consumption were reported in Bosnia and Herzegovina the same year (Table 5). Table 5. Events, by Age-Group, Illnesses Related to Tobacco Consumption, Bosnia and Herzegovina, 2015 15–18 19–64 65+ TOTAL Chronic lower respiratory diseases 3,278 28,431 22,822 54,531 Other forms of heart disease 216 23,882 28,383 52,482 Ischaemic heart disease(s) 57 24,098 23,377 47,533 Influenza and Pneumonia 4,280 27,755 10,814 42,849 Other acute lower respiratory infections 5,779 22,539 8,855 37,173 Cerebrovascular diseases 24 7,020 9,765 16,809 Other diseases of upper respiratory tract 2,449 9,122 4,037 15,608 Other respiratory diseases 1,157 5,817 3,523 10,497 Pancreatic cancer 14 1,997 1,940 3,951 Lung cancer 2 1,050 850 1,901 Bladder cancer 2 741 839 1,582 Leukemia 24 776 579 1,380 Stomach cancer 0 406 364 770 Uterus cancer 0 434 277 711 Total 17,282 154,070 116,424 287,776 Source: Institute of Public Health, Federation of Bosnia and Herzegovina. Note: Incidence is defined as the number of new cases of a given disease during a given period in a specified population. It is also used for the rate at which new events occur in a defined population. It is differentiated from prevalence, which refers to all cases, new or old, in the population at a given time. Data for Bosnia and Herzegovina are obtained by dividing estimates for the Federa- tion of Bosnia and Herzegovina by the share of the country’s population in the Federation of Bosnia and Herzegovina. 21 Tobacco Taxation: Evidence from Bosnia and Herzegovina 4.4. Tobacco-related medical costs After producing estimates of the deaths and other events related to tobacco, the study examines the medical costs of treatments of tobacco-related diseases. Data on these costs are obtained from official Republika Srpska sources. To estimate medical costs for the entire country, the following steps outlined in box 1 have been undertaken to approximate the costs reported in Table 6. Box 1. Estimating Tobacco-Related Medical Costs, Bosnia and Herzegovina 1. The cost of treating the diagnoses associated with the use of tobacco in Republika Srpska was divided by the total costs of curative treatment in Republika Srpska as reported in national health accounts. 2. The ratio obtained in the first step was then used to obtain the equivalent costs for the Federation of Bosnia and Herzegovina by multiplying the former by the total costs of curative treatment for the Federation of Bosnia and Herzegovina obtained from national health accounts. 3. Once the costs were obtained for the Federation of Bosnia and Herzegovina and Republika Srpska, these were added together to obtain the total cost of the treatment of diseases related with tobacco consumption. Because data for Republika Srpska show the distribution of costs according to diseases, these costs could also be approximated for the Federation of Bosnia and Herzegovina by carrying out additional steps, as follows: 4. The associated costs for each disease in Republika Srpska are divided by the total costs related to tobacco treatment for men and women separately. 5. After obtaining corresponding ratios for each disease, these were added up to obtain total shares for each gender separately. 6. The total ratios by gender (74 percent for men and 26 percent for women) are then multiplied by the costs related to tobacco treatment for the Federation of Bosnia and Herzegovina obtained in step 2 (KM 57,296,290) to obtain the total costs for each gender. 7. In the final step, the corresponding ratios for each disease in Republika Srpska are multi- plied by the total costs for men and women in the Federation of Bosnia and Herzegovina. Annual direct health care costs attributable to the treatment of health problems related to tobacco use in Bosnia and Herzegovina in 2015 amount to US$51 million, of which more than two-thirds are related to medical treatments among men. Most medical costs can 22 // Data and Descriptive Statistics be attributed to the treatment of Ischaemic heart disease(s), other forms of heart diseases and cerebrovascular diseases which are also the major cause of death reported in Table 3. It is worth noting that the medical costs reported in Table 3 are related to treatment of diagnosis related to tobacco use and do not take into account additional cost estimates of cytostatic, specific drugs, and biological therapies associated with the use of tobacco and tobacco products. If these extra costs are included, the total medical costs of tobacco use amount to US$60.5 million.6 Table 6. Medical Cost of Treatment of Tobacco-Related Diseases (KM), 2015 MEN WOMEN TOTAL Ischaemic heart disease(s) 14,155,628 4,354,722 18,510,350 Cerebrovascular diseases 9,395,247 5,393,228 14,788,475 Influenza and Pneumonia 6,008,061 2,129,306 8,137,367 Lung cancer 5,360,088 951,297 6,311,385 Chronic lower respiratory diseases 3,279,764 1,134,855 4,414,618 Leukemia 2,824,143 1,085,343 3,909,485 Stomach cancer 1,708,732 388,360 2,097,093 Bladder cancer 1,405,250 271,243 1,676,493 Pancreatic cancer 1,051,805 498,696 1,550,501 Pulmonary heart disease and diseases of pulmonary circulation 973,576 550,168 1,523,745 Kidney cancer 663,524 159,503 823,027 Esophagus cancer 229,175 20,213 249,388 Tongue cancer 170,153 33,697 203,849 Uterus cancer - 11,312 11,312 Other forms of heart disease 12,948,611 5,169,431 18,118,042 Other respiratory diseases 4,234,399 1,374,626 5,609,025 Other diseases of upper respiratory tract 2,208,568 487,384 2,695,952 Other acute lower respiratory infections 1,225,030 184,802 1,409,833 Total (KM) 67,841,754 24,198,186 92,039,940 Total (US$) 37,664,754 13,434,480 51,099,234 Note: Calculations based on official statistics of Republika Srpska and the Federation of Bosnia and Herzegovina. 6  Because these extra costs are only available for Republika Srpska, the procedure outlined in box 1 is followed to obtain costs for the Federation of Bosnia and Herzegovina. The extra medical costs associated with the use of drugs to treat tobacco-related diseases amount to US$9.4 million. 23 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 5 RESULTS To investigate the distributional effects of an increase on tobacco taxes, the effects on prices, medical expenditures, and years of working life were estimated, aggregating these three into a single measure. The price elasticities estimated in Table 3, including the lower- and upper-bound elasticities, facilitate an understanding of how the results could change under different assumptions. 5.1. Tobacco price increase Income changes that arise from an increase in tobacco prices are estimated for each decile based on low, medium-, and upper-bound elasticities. Using the price elasticities and the share of household expenditure on tobacco by decile, one may simulate the effects of an increase in tobacco prices. To show the effect of the elasticities on prices, Table 7 also includes estimates of a complete pass-through scenario, whereby the increase in prices is completely transferred to consumers without a reduction in consumption. For instance, if one assumes that the prices for tobacco products rose by 25 percent (in line with price growth between 2015 and 2018), given the medium-bound elasticity for cigarettes in the bottom decile (−1.08) in Table 3 and the proportion of cigarettes expenditures in the bottom decile (7 percent) in Table 1, the expected increase in household expenditures would be 0.13 percent (Table 7). This represents a rise in welfare because consumers would react strongly to price increases and therefore devote a smaller share of their incomes to purchasing the same amount of tobacco, thereby augmenting the consumption of other goods. The results for all income deciles and elasticity scenarios are shown in Table 7. Across all scenarios, the direct effects after the second income deciles is a welfare loss, but, in none of the cases, does the shock seem to be regressive. In contrast, the effect of Table 7. The Direct Effect of Price Increases through Taxes, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Complete pass-through −0.37 −0.47 −0.45 −0.47 −0.56 −0.53 −0.54 −0.53 −0.55 −0.48 Low-bound elasticity 0.17 0.10 −0.02 −0.01 −0.08 −0.09 −0.13 −0.18 −0.19 −0.24 Medium elasticity 0.13 0.04 −0.06 −0.06 −0.14 −0.14 −0.18 −0.22 −0.23 −0.28 Upper-bound elasticity 0.08 −0.02 −0.11 −0.11 −0.19 −0.19 −0.23 −0.27 −0.28 −0.32 Source: Estimates based on the Household Budget Survey. Note: The table shows the share of total consumption for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices, and this does not affect the quantity purchased. 25 Tobacco Taxation: Evidence from Bosnia and Herzegovina a price rise is relatively progressive, that is, it affects upper-income groups (the last two deciles) in a larger proportion. In the complete pass-through scenario, the price shock is also progressive, that is, the negative effects are still smaller among lower-income groups than among high-income groups. Figure 5. Income Gains: Direct Effect of Tobacco Taxes (Change in expenditure because of tobacco taxes) .2 95% CI Income Gains Complete pass through 0 Upper Medium -.2 Lower -.4 Source: Author estimation using a price shock of 25%. Estimates based on House- -.6 hold Budget Survey. Note: Estimates assume a price shock of 0 2 4 6 8 10 25 percent. Income decile 5.2. Medical expenses Tables 8 and Figure 6 report the income effect of a reduction in medical expenses. The total medical costs are obtained from Table 6 and augmented with additional costs’ estimates of cytostatic, specific drugs and biological therapies involved in the diagnosis associated with the use of tobacco and tobacco products. Because these costs are only available for Republika Srpska, the same procedure outlined in box 1 is followed to obtain the costs in the Federation of Bosnia and Herzegovina. The model assumes that the health effects of tobacco-related diseases will immediately diminish with the reduction in tobacco consumption. Although this assumption is implausible in the short term because changes in the effects of tobacco-related diseases take some time to materialize, it provides an upper-bound estimate of the effects of tax increases. The overall results indicate that the reduction in medical expenditures is progressive, benefiting lower-income groups. This derives from two factors: (1) the higher price elasticity and (2) a lower income base that benefits from the reduction in medical costs. The income gains would vary between 0.51 and 0.05 percentage points in the case of the lower-bound elasticity assumption, between 0.47 and 0.04 percentage points in the case of the medium-bound elasticity, and between 0.42 and 0.03 percentage points in the case of the upper-bound elasticity (Figure 3). These results show the importance of the elasticity assumptions; they also stress the relevance to the possible elasticity variations across income groups. 26 // Results Table 8. Reduction in Medical Costs, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Lower-bound elasticity 0.51 0.36 0.25 0.22 0.19 0.17 0.13 0.10 0.09 0.05 Medium elasticity 0.47 0.32 0.22 0.20 0.17 0.15 0.11 0.09 0.08 0.04 Upper-bound elasticity 0.42 0.28 0.20 0.18 0.15 0.13 0.10 0.08 0.07 0.03 Source: Estimates based on the Household Budget Survey. Note: The table reports the share of total consumption for each decile. Figure 6. Income Gains: Medical Costs of Tobacco Taxes (Reduction in Medical Expenditures) .5 Income Gains 95% CI .4 Upper Medium .3 Lower Source: Author estimation using a price .2 shock of 25%. Estimates based on the Household Budget Survey. .1 Note: Estimates assume a price shock of 25 percent. 0 0 2 4 6 8 10 Income decile 5.3. Income gains deriving from an increase in years of working life The cost of working life lost because of tobacco consumption is estimated based on the assumption that the impact of lower tobacco use on health and work-generated income is direct. The 8,725 deaths obtained from the Tobacco Atlas attributed to tobacco consumption in 2015 are distributed using the mortality profile. For each death, working years lost are divided across deciles proportionately to the number of households that consume tobacco in each income group. The results show that the reduction in tobacco consumption and the expected reduction in work years lost have a positive, albeit negligible impact on welfare in all income groups (Figure 7; Table 9). However, given that elasticities differ across deciles, the gains are more pronounced among lower-income groups across all three scenarios. 27 Tobacco Taxation: Evidence from Bosnia and Herzegovina Figure 7. Income Gains: Production during Years Lost, by Decile .0005 Income Gains % 95% CI Upper .0004 Medium Lower .0003 Source: Author estimation using a price shock of 25%. Estimates based on the Household Budget Survey. .0002 Note: Estimates assume a price shock of 0 2 4 6 8 10 25 percent. Income decile Table 9. Years of Working Life Lost, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.0005 0.0005 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 0.0003 Medium elasticity 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 0.0003 0.0003 0.0003 0.0003 Upper-bound elasticity 0.0004 0.0004 0.0003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0002 Source: Estimates based on the Household Budget Survey. Note: The table reports the share of total consumption for each decile. Years of life lost have been estimated using all deaths from tobacco-related diseases. 5.4. Net effects: total distributional impact Once the effects of tobacco tax policy on prices, medical expenditures, and increased years of employment are calculated separately, one may examine the total distributional impact. Based on all three scenarios, the aggregate effect of an increase on tobacco taxes is positive and progressive; in the long run, poorer deciles benefit more than richer ones in the sense that their income is positively affected. The positive effect of reduced medical expenses and years of life gained offset the negative price effect which was evident in third and higher income deciles in Table 7. It seems that population that smokes will reduce tobacco consumption sufficiently to allow health and work benefits to offset the price increases. The positive income effects are evident among low and middle income population under all scenarios. It seems that young persons (upper-bound elasticity estimates) from low income households benefit less from reduction in consumption and the positive effects tend to decline and become negative again in the middle of income distribution. On average, the results suggest that positive effects are more pronounced in medium and especially lower-bound elasticity scenarios. 28 // Results The assumptions in this model do not include other possible policies, such as smoking cessation programs, antismoking advertising, youth outreach, or policies financed through the new tax revenue. Therefore, these results are in line with the literature, showing important the important role that taxation plays in lowering tobacco usage. Table 10. Net Effect on Household Expenditures, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.685 0.456 0.236 0.216 0.111 0.074 −0.002 −0.074 −0.094 −0.192 Medium elasticity 0.595 0.361 0.159 0.143 0.036 0.004 −0.067 −0.133 −0.154 −0.239 Upper-bound elasticity 0.506 0.268 0.083 0.069 −0.040 −0.065 −0.133 −0.192 −0.213 −0.286 Source: Estimates based on the Household Budget Survey. Note: The table reports the share of total consumption for each decile. Figure 8. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes (Tobacco price increase, medical expenditure, and working years gained) .6 Income Gains % 95% CI .4 Upper Medium .2 Lower 0 Source: Author estimation using a price -.2 shock of 25%. Estimates based on the Household Budget Survey. -.4 Note: Estimates assume a price shock of 25 percent. 0 2 4 6 8 10 Income decile 29 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 6 RESULTS BY ENTITY AND HOUSEHOLD TYPE 6.1. Federation of Bosnia and Herzegovina vs. Republika Srpska To shed more light on the results in the previous section, a separate analysis is run by entity and household type to explore possible differences. Across all scenarios, the elasticity estimates shown in Table 3 are maintained. However, the parameters related to medical costs and years of life lost have changed. Because the Tobacco Atlas only reports total years of life lost, latter is approximated by calculating the proportion of population living in the corresponding entity and multiplying this by the total years of life lost. Tables 11 and 12 report the net income effects in the Federation of Bosnia and Herzegovina and Republika Srpska, respectively. In both entities the effects of tobacco taxation are positive among lower- and middle-income groups across all elasticity scenarios. However, the positive effects are more pronounced in Republika Srpska, with the exception of poorest income decile. Results presented in annex A (Tables A.1 and A.4) suggest that positive net income effects are driven by the positive direct effects of the tobacco price increase in the two lowest income deciles. If one considers indirect effects that take into account the medical expenses associated with the averted treatment costs of tobacco-related diseases, the welfare gains are even greater. The overall results indicate that the reduction in medical expenditures is progressive, benefiting lower-income deciles to a greater extent (Tables A.2 and A.5). For instance, because of the price increase, households in the first income decile would experience an increase in incomes of 0.51 percentage points in the Federation of Bosnia and Herzegovina and 0.49 percentage points in Republika Srpska, while households in the highest income bracket would increase their incomes by only 0.04 and 0.05 percentage points in the Federation of Bosnia and Herzegovina and Republika Srpska, respectively, under the medium elasticity scenario. Similar results hold also across the other two scenarios. As was the case in the baseline model, the expected reduction in work years lost have a positive impact on welfare among all income groups across all three scenarios. The gains are more pronounced in the lower-income groups, especially in the Federation of Bosnia and Herzegovina (Table A.3). 31 Tobacco Taxation: Evidence from Bosnia and Herzegovina Table 11. Net Effect on Household Expenditures, Federation of Bosnia and Herzegovina, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.738 0.449 0.241 0.226 0.101 0.075 0.003 −0.081 −0.078 −0.180 Medium elasticity 0.643 0.362 0.169 0.153 0.035 0.005 −0.057 −0.137 −0.133 −0.225 Upper-bound elasticity 0.550 0.275 0.097 0.079 −0.032 −0.065 −0.116 −0.194 −0.186 −0.269 Table 12. Net Effect on Household Expenditures, Republika Srpska (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.739 0.515 0.275 0.279 0.164 0.097 0.022 −0.061 −0.090 −0.227 Medium elasticity 0.639 0.406 0.191 0.189 0.078 0.017 −0.048 −0.136 −0.157 −0.284 Upper-bound elasticity 0.541 0.299 0.107 0.098 −0.009 −0.063 −0.118 −0.210 −0.224 −0.340 6.2. Urban vs. rural The estimated net effects on household expenditure for household type show that urban population benefits disproportionally more. This is especially pronounced for lower income households across all three elasticity scenarios. The aggregate effect of an increase on tobacco taxes is positive and progressive for low and middle income households and is mostly driven by positive contribution in the reduction of medical costs (Tables A.8 and A.11). By looking at medium price elasticity scenario, one can see that the reduction of medical costs for urban population as a result of an increase in tobacco prices by 25 percent will lead to almost two times larger effect for lowest income decile in comparison to rural population. As is the case in other models, the effects of reduction in the years of life lost has positive effects across all income deciles, but it is very low in magnitude to be able to offset any negative price effects. When it comes to direct effects, the results suggest that an increase in prices by 25 percent benefits lower income deciles in both groups of households. This increases household welfare because consumers in two lowest income deciles are more price sensitive and react strongly to price increase by reducing the purchase of cigarettes. However, this positive direct effects turn negative for middle and higher income deciles, 32 // Results by Entity and Household Type but in none of the cases does the shock seem to be regressive. On the contrast, the effect of price increase is relatively progressive, that is, it affects upper-income groups (the last two deciles) in a larger proportion. Table 13. Net Effect on Household Expenditures, Urban Population, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.87 0.49 0.27 0.28 0.13 0.07 0.03 −0.05 −0.08 −0.17 Medium elasticity 0.76 0.40 0.18 0.19 0.06 0.00 −0.04 −0.11 −0.14 −0.22 Upper-bound elasticity 0.66 0.31 0.08 0.09 0.00 −0.07 −0.11 −0.16 −0.20 −0.26 Table 14. Net Effect on Household Expenditures, Rural Population, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.546 0.455 0.199 0.195 0.112 0.052 −0.009 −0.079 −0.118 −0.214 Medium elasticity 0.473 0.357 0.131 0.126 0.042 −0.012 −0.077 −0.136 −0.179 −0.264 Upper-bound elasticity 0.401 0.259 0.063 0.056 −0.028 −0.075 −0.146 −0.194 −0.240 −0.313 33 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE 7 DISCUSSION There has been extensive research on the negative effects of tobacco consumption and on the benefits of various public policy mechanisms aimed at reducing tobacco use. The implementation of tobacco taxes, both specific and ad valorem is considered one of the most effective ways to discourage tobacco use. Nonetheless, this policy has a direct impact on household incomes, especially among low-income households that are more likely to smoke, have limited access to health insurance and adequate health care. The question whether tobacco taxes are regressive is particularly important because the welfare effects derived from increased taxes depend on price elasticity of this product across different sectors of the population. The price elasticity determines the magnitude of the income shock and the benefits gained from decline in tobacco consumption. Using three waves of household budget surveys, the study calculated the price elasticity of tobacco for the population in Bosnia and Herzegovina, obtaining an average price elasticity of −0.56 and estimates for the 10 income deciles. The elasticity for the lowest income group is −1.08 and −0.34 for the highest income group. This appears to be the first tobacco price elasticity estimate across income groups ever produced on Bosnia and Herzegovina. To extend the analysis, elasticities have also been simulated for the younger population (upper-bound) and the older (lower-bound) population and the long-term scenario to assess the net welfare gains from this policy. Therefore, apart from the direct impact on household income, other benefits of lower tobacco consumption are considered, including a reduction in medical costs and an increase in the potential working years associated with good health. Thus, it is critical to justify the maintenance or intensification of the use of tobacco taxes by means of a demonstration of the aggregate monetary gains or losses generated. Moreover, the policy should focus on low-income households that are more likely to smoke and, hence, tend to be the most highly affected by consumption taxes. One of the main motivations of this study is to weigh the main costs and benefits of tobacco taxation to determine if, in the end, the policy is regressive. Results show that, considered by itself, a price increase for tobacco through higher taxes generate negative income variation for middle and high-income groups in the population, but positive ones for lower income groups because of the overall prices increase. These negative effects are particularly accentuated under the upper-bound elasticity scenario. Based on assumptions of a more comprehensive approach, including benefits through lower medical expenses and an increase in working years the positive monetary effect for lower income groups are further accentuated and bring about positive income effects to 35 Tobacco Taxation: Evidence from Bosnia and Herzegovina middle income households as well. The reduction in medical expenses is the main driver of the increase in net incomes because of the reduction in tobacco-related problems that require expensive treatments. In all three elasticity scenarios, the benefits of medical expenses are greater than the benefit of the increase in working years. Net income effects indicate that although increased prices may lead to reduction in tobacco consumption for some groups they do not necessarily lead to tobacco cessation which would bring substantial benefits evident in substantial reduction in medical costs and increased working life. Overall monetary effects are positive, and results suggest that older population and those with lowest income experience more positive and greater income effects. The limited data on medical costs and mortality, which are available only for Republika Srpska and the Federation of Bosnia and Herzegovina, respectively, obliged a certain adjustment to the data and may therefore represent a lower bound on the potential benefits of reduced medical expenditure and aggregate wealth effects. More investigation is therefore needed to produce more accurate estimates of the distributional effects of tobacco policy. Under an alternative scenario, if one simulates the effects of a price increase in different entities and for different household types, the main effects are corroborated. The results suggest that overall monetary effects are positive among lower- and middle-income households, regardless of the location or household type. All households benefit disproportionately more from decreased medical expenses and an increase in working life in the future, but lower income households also benefit from a direct increase in tobacco prices because they substitute tobacco consumption for other types of goods. The net effects are positive and progressive. Thus, overall, the benefits of increased prices go to lower-income households. The three price elasticity scenarios do not exactly mimic the short- versus the long-term effects of a tobacco tax. There is evidence that adult smokers will introduce changes in their behavior if price increases, the lower-bound elasticity would tend to represent this behavior more closely. In contrast, younger groups of the population show less elastic demand, similar to the upper-bound elasticity. After a few decades, one expects the impact of the tax policy to resemble the upper-bound elasticity scenario. In summary, the results provide evidence that supports possible reduction in tobacco taxes. In addition, given the limited use of tobacco cessation strategies, taxation policy should be accompanied by prohibiting smoking in public places, mass media campaigns on the negative effects of tobacco use and increase the bans on all forms of direct and indirect advertising. An integrative approach to anti-tobacco policies that favors coordination between taxation and behavioral change policies is thus needed to generate positive social returns. 36 // Discussion REFERENCES Coady, D. P., M. El Said, R. Gillingham, K. Kpodar, P. A. Medas, and D. L. Newhouse. 2006. “The Magnitude and Distribution of Fuel Subsidies: Evidence from Bolivia, Ghana, Jordan, Mali, and Sri Lanka.” IMF Working Paper WP/06/247, International Monetary Fund, Washington, DC. Chaloupka, F. J., and M. Grossman. 1996. “Price, Tobacco Control Policies, and Youth Smoking.” NBER Working Paper 5740 (September), National Bureau of Economic Research, Cambridge, MA. Chaloupka, Frank J., K. Michael Cummings, Chris P. Morley, and Janet K. Horan. 2002. “Tax, Price, and Cigarette Smoking: Evidence from the Tobacco Documents and Implications for Tobacco Company Marketing Strategies.” Tobacco Control 11 (1): i62–i72. Chaloupka, F. J., R. Peck, J. A. Tauras, X. Xu, and A. Yurekli. (2010). “Cigarette excise taxation: the impact of tax structure on prices, revenues, and cigarette smoking”. Working paper number 16287, National Bureau of Economic Research (NBER), Cambridge, MA. Debrott Sánchez, David. 2006. “Economía del Control del Tabaco en los países del Mercosur y Estados Asociados: Chile.” Pan American Health Organization, Washington, DC. Denisova, Irina, and Polina Kuznetsova. 2014. “The Effects of Tobacco Taxes on Health: An Analysis of the Effects by Income Quintile and Gender in Kazakhstan, the Russian Federation, and Ukraine.” Discussion Paper 92765 (October), World Bank, Washington, DC. DHHS (U.S. Department of Health and Human Services). 2004. The Health Consequences of Smoking: A Report of the Surgeon General. Atlanta: Office on Smoking and Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, DHHS. Doll, Richard, and A. Bradford Hill. 1956. “Lung Cancer and Other Causes of Death in Relation to Smoking: A Second Report on the Mortality of British Doctors.” British Medical Journal 2 (5001): 1071–81. Drope J, Schluger N, Cahn Z, Drope J, Hamill S, Islami F, Liber A, Nargis N, Stoklosa M. 2018. The Tobacco Atlas. Atlanta: American Cancer Society and Vital Strategies. Durkin, Sarah, Emily Brennan, and Melanie Wakefield. 2012. “Mass Media Campaigns to Promote Smoking Cessation among Adults: An Integrative Review.” Tobacco Control 21 (2): 127–38. 37 Tobacco Taxation: Evidence from Bosnia and Herzegovina Euromonitor International. 2018. Passport - tobacco. London, GB: Euromonitor International. Federation Ministry of Health and Institute for Public Health 2014. “Survey of the Health Status of the Population of the Federation of BIH in 2012, with a comparison to the survey conducted in 2002.” Fuchs, A., and F. J. Meneses. 2017. “Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine.” World Bank, Washington, DC. Fuchs, A., and F. J. Meneses. 2018. “Tobacco Price Elasticity and Tax Progressivity in Moldova.” World Bank, Washington, DC. Gallagher A.W.A, Evans-Reeves K.A., Hatchard J.L., Gilmore, A. B. 2018. Tobacco industry data on illicit tobacco trade: a systematic review of existing assessments. Tobacco Control, Published online. Gallus, S., A. Schiaffino, C. La Vecchia, J. Townsend, and E. Fernandez. 2006. “Price and Cigarette Consumption in Europe.” Tobacco Control 15 (2): 114–19. Goodchild, M., A.M. Perucic, and N. Nargis. 2016. “Modelling the impact of increasing tobacco taxes on public health and finance”. Bulletin of the World Health Organization 94: 250–57. Goodchild, M., N. Nargis, and E. Tursan d’Espaignet. 2018. “Global Economic Cost of Smoking Attributable Diseases.” Tobacco Control 27 (1): 58–64. House of Representatives of the Federation Parliament. 2011. Zakon o izmjenama Zakona o ograničenoj uporabi duhanskih prepađevina [Law on amendments to the Law on the limited use of tobacco products]. Official Gazette of the Federation of Bosnia and Herzegovina. 2011;50:203 (http://www.fmoh.gov.ba/images/federalno_ministarstvo_ zdravstva/zakoni_i_strategije/zakoni/zakoni_PDF/Duhanske_preradjevine_Izmjene_50_11.pdf ) (in Bosnian). IARC (International Agency for Research on Cancer, World Health Organization). 2009. Evaluating the effectiveness of smoke-free policies. IARC Press, International Agency for Research on Cancer. IDB (Inter-American Development Bank) Arpón, P., Bitrán, R., & Escobar, L. (2011). Consecuencias del aumento del impuesto al tabaco en Chile en el año 2010. 38 // References Idris BI, Giskes K, Borrell C, Benach J, Costa G, Federico B, Helakorpi S, Helmert U, Lahelma E, Moussa KM, Ostergren PO, Prättälä R, Rasmussen NK, Mackenbach JP, Kunst AE. (2007). “Higher smoking prevalence in urban compared to non-urban areas: time trends in six European countries”. Health Place 13 (3): 702–12. Jha, P., and F. J. Chaloupka. 2000. The economics of global tobacco control. British Medical Journal 321 (7257): 358. Jha, Prabhat, and Richard Peto. 2014. “Global Effects of Smoking, of Quitting, and of Taxing Tobacco.” New England Journal of Medicine 370 (1): 60–68. Joossens L, Merriman D, Ross H, Raw M. 2009. How eliminating the global illicit cigarette trade would increase tax revenue and save lives. Paris: International Union Against Tuberculosis and Lung Disease. Kpodar, K. 2006. “Distributional Effects of Oil Price Changes on Household Expenditures: Evidence from Mali.” IMF Working Paper WP/06/91, International Monetary Fund, Washington, DC. Levy, D., Rodríguez-Buño, R. L., Hu, T. W., & Moran, A. E. (2014). The potential effects of tobacco control in China: projections from the China SimSmoke simulation model. Levy DL, Ellis JA, Mays D, Huang AT. (2013). “Smoking-related deaths averted due to three years of policy progress”. Bull World Health Organ 91(7):509–18. Lewit EM, Coate D.1982. “The potential for using excise taxes to reduce smoking”. Journal of health economics 1 (2):121–145. Maslennikova, G. Y., Oganov, R. G., Boytsov, S. A., Ross, H., Huang, A.-T., Near, A., … Levy, D. T. (2014). Russia SimSmoke: the long-term effects of tobacco control policies on smoking prevalence and smoking-attributable deaths in Russia. Tobacco Control, 23(6), 484–490. http://doi.org/10.1136/tobaccocontrol-2013-051011 National Assembly of the Republika Srpska. Закон о измјенама и допунама Закона о забрани пушења дуванских производа на јавним мјестима [Law on amendments to the Law on the prohibition of smoking tobacco products in public places]. Official Gazette of the Republika Srpska. 2009;92:16 (http://www.vladars.net/sr-SP-Cyrl/Vlada/Ministarstva/MZSZ/ Documents/ Zakon%20o%20izmjenama%20i%20dopunama%20Zakona%20o%20zabrani%20 pusenja%20duvanskih%20proizvoda%20na%20javnim%20mjestima%202.pdf ) (in Serbian) 39 Tobacco Taxation: Evidence from Bosnia and Herzegovina Ng, Marie, Michael K. Freeman, Thomas D. Fleming, Margaret Robinson, Laura Dwyer- Lindgren, Blake Thomson, Alexandra Wollum, et al. 2014. “Smoking Prevalence and Cigarette Consumption in 187 Countries, 1980–2012.” Journal of the American Medical Association 311 (2): 183–92. Peto, Richard, and Alan D. Lopez. 2004. “The Future Worldwide Health Effects of Current Smoking Patterns.” In Tobacco and Public Health: Science and Policy, edited by Peter Boyle, Nigel Gray, Jack Henningfield, John Seffrin, and Witold Zatonski, 281–86. New York: Oxford University Press. Pichón-Riviere, A., A. Bardach, J. Caporale, A. Alcaraz, F. Augustovski, F. Caccavo, C. Vallejos et al. 2014. “Carga de Enfermedad Atribuible al Tabaquismo en Chile.” Documento Técnico 8 (April), Instituto de Efectividad Clínica y Sanitaria, Buenos Aires. Selvaraj, S., S. Srivastava, and A. Karan. 2015. “Price Elasticity of Tobacco Products among Economic Classes in India, 2011–2012.” British Medical Journal. Verguet, Stéphane, Cindy L. Gauvreau, Sujata Mishra, Mary MacLennan, Shane M. Murphy, Elizabeth D.Brouwer, Rachel A. Nugent, Kun Zhao, Prabhat Jha, and Dean T Jamison. 2015. “The Consequences of Tobacco Tax on Household Health and Finances in Rich and Poor Smokers in China: An Extended Cost- Effectiveness Analysis.” Lancet Global Health 3 (4): e206–16. WHO (World Health Organization). 2008. WHO report on the global tobacco epidemic 2008: the MPOWER package. ———. 2015a. Global Report on Trends in Prevalence of Tobacco Smoking. Geneva: WHO. ———. 2015b. WHO Report on the Global Tobacco Epidemic 2015: Raising Taxes on Tobacco. Geneva: WHO. ———. 2016. Framework Convention on Tobacco Control. 2016. Convention Secretariat and World Health Organization. http://www.who.int/fctc/en/. ———. 2017a. WHO Tobacco Fact Sheet. Geneva: WHO. http://www.who.int/mediacentre/ factsheets/fs339/en/. ———. 2017b. WHO Report on the Global Tobacco Epidemic 2017: Monitoring tobacco use and prevention policies. Geneva: WHO. World Bank. 1999. Curbing the Epidemic: Governments and the Economics of Tobacco Control. Report 19638 (May). Development in Practice Series. Washington, DC: World Bank. http:// documents.worldbank.org/curated/en/1999/05/437174/curbing-epidemic-governments- economicstobacco-control. 40 // References Wynder, Earnest L., and Evarts A. Graham. 1950. “Tobacco Smoking as a Possible Etiologic Factor in Bronchiogenic Carcinoma: A Study of Six Hundred and Eighty-Four Proved Cases.” Journal of the American Medical Association 143 (4): 329–36. Yeh, C-Y., Schafferer, C., Lee, J-M., Ho, J-M. and Hsieh, C-J. (2017). “The effects of a rise in cigarette price on cigarette consumption, tobacco taxation revenues, and of smoking- related deaths in 28 EU countries - applying threshold regression modelling”. BMC Public Health 17:676 41 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE ANNEX A. DETAILED SCENARIO OUTCOME ACROSS ENTITY AND HOUSEHOLD TYPE A.1. Federation of Bosnia and Herzegovina Table A.1. Direct Effect of Price Decrease through Taxes, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Complete pass-through −0.36 −0.41 −0.41 −0.46 −0.49 −0.53 −0.48 −0.51 −0.49 −0.45 Low-bound elasticity 0.17 0.09 −0.01 −0.01 −0.07 −0.10 −0.12 −0.17 −0.17 −0.23 Medium elasticity 0.13 0.04 −0.06 −0.06 −0.12 −0.15 −0.16 −0.22 −0.21 −0.26 Upper-bound elasticity 0.08 −0.01 −0.10 −0.11 −0.17 −0.20 −0.21 −0.26 −0.25 −0.30 Source: Estimates based on the Household Budget Survey. Note: The table shows the share of total consumption for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices, and this does not affect the quantity purchased. Table A.2. Reduction in Medical Costs, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Low-bound elasticity 0.565 0.362 0.254 0.235 0.173 0.170 0.119 0.093 0.088 0.046 Medium elasticity 0.516 0.325 0.226 0.210 0.154 0.150 0.105 0.081 0.077 0.039 Upper-bound elasticity 0.468 0.288 0.198 0.184 0.134 0.131 0.090 0.069 0.065 0.031 Source: Estimates based on the Household Budget Survey. Table A.3. Years of Working Life Lost, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.0003 0.0003 0.0003 0.0003 0.0002 0.0003 0.0002 0.0002 0.0002 0.0002 Medium elasticity 0.0003 0.0003 0.0002 0.0003 0.0002 0.0003 0.0002 0.0002 0.0002 0.0002 Upper-bound elasticity 0.0003 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0001 Source: Estimates based on Household Budget survey 43 Tobacco Taxation: Evidence from Bosnia and Herzegovina A.2. Republika Srpska Table A.4. Direct Effect of Price Decrease through Taxes, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Complete pass-through −0.42 −0.55 −0.48 −0.56 −0.60 −0.59 −0.55 −0.63 −0.60 −0.57 Low-bound elasticity 0.20 0.12 −0.02 −0.01 −0.09 −0.11 −0.13 −0.21 −0.20 −0.29 Medium elasticity 0.15 0.05 −0.07 −0.07 −0.15 −0.16 −0.19 −0.27 −0.26 −0.33 Upper-bound elasticity 0.10 −0.02 −0.12 −0.13 −0.21 −0.22 −0.24 −0.32 −0.31 −0.38 Source: Estimates based on the Household Budget Survey. Note: The table shows the share of total consumption for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices, and this does not affect the quantity purchased. Table A.5. Reduction in Medical Costs, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Low-bound elasticity 0.54 0.40 0.29 0.29 0.25 0.20 0.16 0.15 0.11 0.06 Medium elasticity 0.49 0.36 0.26 0.26 0.23 0.18 0.14 0.13 0.10 0.05 Upper-bound elasticity 0.45 0.32 0.23 0.23 0.20 0.16 0.12 0.11 0.08 0.04 Source: Estimates based on the Household Budget Survey. Table A.6. Years of Working Life Lost, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.0002 0.0002 0.0001 0.0002 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 Medium elasticity 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Upper-bound elasticity 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Source: Estimates based on the Household Budget Survey. 44 // Annexes A.3. Urban population Table A.7. Direct Effect of Price Decrease through Taxes, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Complete pass-through −0.37 −0.39 −0.59 −0.60 −0.48 −0.55 −0.53 −0.48 −0.54 −0.45 Low-bound elasticity 0.18 0.08 −0.02 −0.01 −0.07 −0.10 −0.13 −0.16 −0.18 −0.22 Medium elasticity 0.13 0.03 −0.08 −0.08 −0.12 −0.15 −0.18 −0.20 −0.23 −0.26 Upper-bound elasticity 0.08 −0.01 −0.15 −0.14 −0.16 −0.20 −0.23 −0.24 −0.28 −0.30 Source: Estimates based on the Household Budget Survey. Note: The table shows the share of total consumption for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices, and this does not affect the quantity purchased. Table A.8. Reduction in Medical Costs, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Low-bound elasticity 0.691 0.411 0.292 0.295 0.205 0.172 0.157 0.108 0.105 0.051 Medium elasticity 0.631 0.369 0.260 0.264 0.181 0.152 0.138 0.094 0.091 0.043 Upper-bound elasticity 0.573 0.328 0.228 0.232 0.158 0.132 0.119 0.080 0.077 0.034 Source: Estimates based on the Household Budget Survey. Table A.9. Years of Working Life Lost, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.0003 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0001 Medium elasticity 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0001 0.0002 0.0001 Upper-bound elasticity 0.0002 0.0002 0.0002 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Source: Estimates based on the Household Budget Survey. 45 Tobacco Taxation: Evidence from Bosnia and Herzegovina A.4. Rural population Table A.10. Direct Effect of Price Decrease through Taxes, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Complete pass-through −0.31 −0.51 −0.41 −0.45 −0.50 −0.50 −0.57 −0.52 −0.59 −0.52 Low-bound elasticity 0.15 0.11 −0.01 −0.01 −0.07 −0.09 −0.14 −0.18 −0.20 −0.26 Medium elasticity 0.11 0.05 −0.06 −0.06 −0.12 −0.14 −0.19 −0.22 −0.25 −0.30 Upper-bound elasticity 0.07 −0.02 −0.10 −0.10 −0.17 −0.18 −0.24 −0.27 −0.30 −0.34 Source: Estimates based on the Household Budget Survey. Note: The table shows the share of total consumption for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices, and this does not affect the quantity purchased. Table A.11. Reduction in Medical Costs, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO +25% Low-bound elasticity 0.400 0.346 0.212 0.204 0.186 0.141 0.129 0.096 0.079 0.043 Medium elasticity 0.366 0.311 0.189 0.182 0.165 0.124 0.113 0.084 0.069 0.036 Upper-bound elasticity 0.332 0.276 0.165 0.160 0.144 0.108 0.097 0.071 0.059 0.029 Source: Estimates based on the Household Budget Survey. Table A.12. Years of Working Life Lost, by Decile (%) DECILES PRICE SHOCK 1 2 3 4 5 6 7 8 9 10 SCENARIO Low-bound elasticity 0.0002 0.0003 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 Medium elasticity 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0001 Upper-bound elasticity 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0002 0.0001 0.0001 0.0001 Source: Estimates based on the Household Budget Survey. 46 // Annexes 47 NEARLY 80 PERC OF THE WORLD’S SMOKERS LIVE IN LOW- AND MIDDL INCOME COUNTR AND ARE LESS LIK TO BE INFORMED ABOUT THE ADVE HEALTH EFFECTS OF TOBACCO USE