THE DISTRIBUTIONAL EFFECTS OF TOBACCO TAXATION: THE EVIDENCE OF WHITE AND CLOVE CIGARETTES IN INDONESIA Cover photos (clockwise from left): Man smoking in Kuta, Indonesia. Photo by Alit Pramesta / Unsplash. Bali, Indonesia. Photo by Dewa Wira / Unsplash. The Market, Ubud, Indonesia. Steve Long / Unsplash. Jakarta, Indonesia. Photo by Gede Suhendra / Unsplash. 2 // Abstract THE DISTRIBUTIONAL EFFECTS OF TOBACCO TAXATION: THE EVIDENCE OF WHITE AND CLOVE CIGARETTES IN INDONESIA Alan Fuchs: Poverty and Equity Global Practice, World Bank afuchs@worldbank.org Giselle Del Carmen: Poverty and Equity Global Practice, World Bank gdelcarmen@worldbank.org ABSTRACT Despite the well-known positive impact of tobacco taxes on health outcomes, policy mak- ers hesitate to use them because of their possible regressive effect, that is, poorer deciles are proportionally more negatively affected than richer ones. Using an extended cost-ben- efit analysis to estimate the distributional effect of white and clove cigarettes in Indonesia, this study finds that the long-run impact may be progressive. The final aggregate effect incorporates the negative price effect, but also changes in medical expenditures and in additional working years. The analysis includes estimates of the distributional impacts of price rises on cigarettes under various scenarios using 2015–16 Indonesia National Socioeconomic Surveys. One contribution is to quantify the impacts by allowing price elasticities to vary across consumption deciles. Overall, clove cigarette taxes exert an effect that depends on the assumptions of conditional price elasticity. If the population is more responsive to tobacco price changes, then people would experience even more gains from the health and work benefits. More research is needed to clarify the distributional effects of tobacco taxation in Indonesia. JEL Codes: H23, H31, I18, O15 II // Abstract ACKNOWLEDGMENTS Support in the preparation of this report has been provided by the World Bank’s Global Tobacco Control Program, co- financed by the Bill and Melinda Gates Foundation and Bloomberg Philanthropies. The authors are grateful to Patricio Marquez, Daniela Paz, Maria Eugenia Genoni, Sailesh Tiwari, Jaffar Al Rikabi, Federico Gil Sander, Pandu Harimurti, and Abdillah Ahsan for offering 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. III TABLE OF CONTENTS Abstract II Acknowldgements III List of Tables VI List of Figures VII 1. Introduction 2 2. Literature Review 6 3. Model 12 4. Data and Descriptive Statistics 14 Tobacco Prices 14 Tobacco Price Elasticity by Decile 15 Mortality and Morbidity 17 Tobacco-related Medical Costs 19 Baseline Descriptive Results 20 5. Results 24 Tobacco Price Increase 24 Medical Expenses 27 Income Gains Deriving From an Increase in Years of Working Life 29 Net Effects: Total Distributional Impact 31 6. Discussion 36 IV // Table of Contents References 38 Annexes 46 A. Model and Methodology 46 Change in Tobacco Expenditure 46 Medical Expenses 47 Increase in Years of Working Life 47 B. Tobacco Price Elasticity by Decile 47 C. Descriptive Statistics, 2010 50 D. Robustness Checks 51 E. Simulation: Tobacco Tax Revenue Redistributed to All Households 55 V LIST OF TABLES Table 1. Mean and Median Price per Stick Paid by Households, by Cigarette Type and Decile, 2016 14 Table 2. White Cigarettes: Price Elasticities, by Decile 16 Table 3. Clove Cigarettes: Price Elasticities, by Decile 17 Table 4. All Cigarette Products: Price Elasticities, by Decile 17 Table 5. Tobacco-related Deaths, by Gender and Age-Group, Indonesia, 2016 19 Table 6. Tobacco-Related Events, by Gender, 2016 20 Table 7. Years of Life Lost, by Gender, 2016 21 Table 8. Baseline Descriptive Results, 2016 22 Table 9. Direct Effect of a Price Increase through Taxes, White Cigarettes (%) 25 Table 10. Direct Effect of a Price Increase through Taxes, Clove Cigarettes (%) 25 Table 11. Reduction in Medical Costs, White Cigarettes (%) 27 Table 12. Reduction in Medical Costs, Clove Cigarettes (%) 27 Table 13. Income Gains: Production during Years Lost, by Decile, White Cigarettes (%) 29 Table 14. Income Gains: Production during Years Lost, by Decile, Clove Cigarettes (%) 29 Table 15. Net Effect on Household Expenditures, White Cigarettes (%) 31 Table 16. Net Effect on Household Expenditures, Clove Cigarettes (%) 33 Table B1. Regression Results, by Model 48 Table C1. Descriptive Statistics 50 Table D1. Net Effect on Household Expenditures, White Cigarettes, 3 Percent Discount (%) 51 Table D2. Net Effect on Household Expenditures, Clove Cigarettes, 3 Percent Discount (%) 51 Table D3. Net Effect on Household Expenditures, White Cigarettes, 10 Percent Discount (%) 54 Table D4. Net Effect on Household Expenditures, Clove Cigarettes, 10 Percent Discount (%) 54 Table E1. Net Effect on Household Expenditures (Including Direct Tax Revenue Transfer), All Tobacco Products 55 VI // Table of Contents LIST OF FIGURES Figure 1. Mean Cigarette Price, by Year and Pack Size (2010 prices) 15 Figure 2. Mean Cigarette Price, by Year and Pack Size, Real vs. Nominal Prices 16 Figure 3. White Cigarettes: Price Elasticity, Indonesia 18 Figure 4. Clove Cigarettes: Price Elasticity, Indonesia 18 Figure 5. Income Gains: Direct Effect of Tobacco Taxes (Increase in Expenditure Because of Tobacco Taxes) 26 Figure 6. Income Gains: Medical Costs of Tobacco Taxes (Reduction in Medical Expenditures) 28 Figure 7. Income Gains: Production during Years Lost, by Decile 30 Figure 8. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes (Tobacco Price Increase, Medical Expenditure, and Working Years Gained) 32 Figure 9. Total Income Effect Decomposition, Medium-Bound Elasticity (Tobacco Price Increase, Medical Expenditure, and Working Years Gained) 34 Figure D1. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes ) Tobacco Price Increase, Medical Expenditure, and Working Years Gained) (3 percent discount rate) 52 Figure D2. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes (Tobacco price Increase, Medical Expenditure, and Working Years Gained) (10 percent discount rate) 53 Figure E1. Total Income Effect: All Tobacco Products 56 VII OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN 1 INTRODUCTION The World Health Organization (WHO) estimates that tobacco kills more than 7 million people worldwide each year (WHO 2017a). It is the second leading cause of death and disability worldwide (Ng et al. 2014) and is among the major preventable causes of dis- ease and premature death globally (Doll and Hill 1956; Wynder and Graham 1950). Dis- eases 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. Indonesia is a major producer and consumer of tobacco; 68.1 percent of men ages 15 or above and 2.5 percent of women use some form of tobacco (Zheng et al. 2018). Kretek cigarettes, often referred to as clove cigarettes, are the most popular tobacco products in the coun- try. As of 2015, an estimated 40 percent of the adult population smoked kretek cigarettes (World Bank 2017a). They are made of a combination of tobacco and ground clove mixed with special sauces unique to each brand (Arnez 2009; Malson 2003). It is estimated that more than one million farmers grow cloves in Indonesia, which accounts for more than 70 percent of the world’s cloves (World Bank 2017b). Over more than a decade, the Indonesian government has been gradually raising excise taxes on tobacco products. Yet, the taxes remain below the recommended WHO thresh- old. While the WHO recommends total taxes on tobacco products—that is, specific excise taxes, local taxes, and the value added tax—should be at least 75.0 percent of the retail price, the average rate in Indonesia is 57.4 percent (WHO 2017b). The current cigarette excise tax encompasses 12 tiers. Each tier is based on the type of product, the number of cigarettes produced, and the per unit retail price. Inevitably, such a system leads to substi- tution toward lower-priced products and thus has limited impact on smoking reduction. Moreover, the Indonesian government has a preferential tax structure for hand-rolled kretek factories because these employ a significant share of kretek workers and account for more than half of the factories in the tobacco industry. As of 2018, the excise tax on hand-rolled kreteks was 24.0 percent of the minimum retail price, while the excise tax on machine-made kreteks was 45.8 percent, depending on the production volume and range of retail prices. In addition to excise taxes on the cigarettes sold in Indonesia, the government also applies a 9.1 percent value added tax, which is collected directly from manufacturers (Ahsan et al. 2013). Even though the government has considered reforming the excise tax structure by reduc- ing and consolidating the number of tiers and raising taxes, it is concerned this might 2 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia affect the livelihoods of people in the kretek value chain. A recent World Bank (2017a) study finds that the tax increase would reduce employment in the handmade kretek industry by approximately 8 percent. Some households would lose as much as 27 per- cent–52 percent of total household income. Nonetheless, the reduction in employment in this industry is equivalent to 1.4 percent of the potential revenue gains from increasing cigarette taxes. The study estimates the annual revenue gain would be Rp 12.9 billion, while the employment loss in the handmade kretek industry would be 24,000 jobs.1 Thus, the tax and price increase would not only reduce tobacco use and related health care costs, it would generate additional revenue that would more than compensate for the income loss associated with a reduction in employment and production in the kretek industry (World Bank 2017a). Similarly, reforms in the cigarette excise tax structure have also raised concerns about the possible effects on tobacco farmers. Nonetheless, tobacco leaf is a relatively small agricultural subsector in Indonesia, representing roughly 0.30 percent of the agricultural sector and 0.03 percent of gross domestic product (GDP) (Indonesia Ministry of Agricul- ture 2010). Another World Bank (2017c) study analyzes the socioeconomic conditions of tobacco farmers in Indonesia and suggests that tobacco is not profitable for most farmers and that there is widespread poverty and food insecurity among tobacco farmers, as well as evidence of child labor during school hours. Thus, given the economic and social costs of tobacco growing, additional tax revenues from a reformed tobacco excise system could be used to help tobacco farmers switch to more profitable alternative crops. Although there are over 225,000 deaths annually from tobacco-related diseases in Indo- nesia, nearly 470,000 children ages 10–14 and 53 million adults ages 15 or older continue to use tobacco daily. In 2016, 21 percent of male deaths in Indonesia were caused by tobacco; this is higher than the average in countries with a medium human development index.2 Similarly, Indonesia is among the few countries that has not signed or ratified the WHO Framework Convention on Tobacco Control (WHO 2017c). This policy creates an attractive space for the tobacco industry to continue growing because there is limited regulation of tobacco use. Tobacco taxes are often considered regressive because the share of household budgets allocated to tobacco products is larger among low-income households than among high-income households. This paper shows that, if behavioral and indirect effects, especially health-related effects, are taken into account, this is no longer valid. The 1 The average annualized kretek income is estimated at Rp 15,500,000, and the projected decrease in wage income is 48.9 percent. The loss of 24,000 jobs in the kretek industry implies a total income loss of Rp 182 billion, equivalent to 1.4 percent of the revenue gain. 2 See “The Tobacco Atlas: Indonesia,” American Cancer Society and Vital Strategies, Atlanta, https://tobaccoatlas.org/country/indonesia/. 3 // Introduction long-run benefits of not smoking offset the costs associated with tobacco taxes among low-income groups and the overall population. Potential benefits include a reduction in medical expenditures and an increase in healthy life years, factors that translate into economic benefits that outweigh the losses created by tax increases if consumers quit or never start smoking. This paper describes and quantifies the medium- and long-run effects of tobacco tax increases on aggregate household welfare through three channels. The first implies that higher tobacco prices because of higher taxes induce a behavioral response involving a reduction and cessation in tobacco consumption that are manifest particularly among people who discontinue smoking and younger individuals who do not start smoking. The second channel is associated with a reduction in medical expenses associated with the averted treatment costs of tobacco-related diseases. The third channel is a rise in incomes because of gains in years of employment because of the prolongation of life expectancy. To assess the impact of these effects, this paper estimates the price elasticity of tobacco, simulates upper- and lower-bound scenarios, and calculates the welfare gains among var- ious 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 an overview of the data used to estimate the impact of the tobacco tax. Section 4 examines the results. The final section concludes with a discussion on policy implications. 4 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN 2 LITERATURE REVIEW Blecher and van Walbeek (2008) estimate that, because the mean incomes in countries grow more quickly than average tobacco prices, cigarettes became more affordable in developing countries between 1997 and 2006. This indicates that, as smoking prevalence increases, these countries are more likely to bear the major health impacts of tobacco consumption. Furthermore, low- and middle-income countries are experiencing a rise in noncommunicable diseases. Kristina et al. (2015) estimate that, in Indonesia, cancer mortality attributable to smoking represented 74,440 deaths (30.6 percent of total cancer deaths), of which 95.0 percent were men. Cancer deaths attributable to smoking were responsible for 1.2 million years of potential life lost. This has negative consequences for human capital development and imposes an increasing economic burden because smok- ing decreases earnings potential and labor productivity (WHO 2015a). Several studies have quantified the economic costs of smoking, though most have been carried out in high-income countries. Annual direct health care costs attributable to tobacco use amounted to US$2.2 billion in Indonesia in 2015, which is equivalent to 2.5 percent of GDP (Kristina et al. 2018). In Bangladesh, the overall economic cost of tobacco use is estimated at Tk 110 billion (US$1.9 billion) or over 3 percent of GDP in 2004 (WHO 2007). In contrast, annual tobacco-related health costs are estimated at US$81 billion in the United States, nearly US$7 billion in Germany, and US$1 billion in Australia (Guindon et al. 2007). Goodchild, Nargis, and Tursan d’Espaignet (2018) find that tobacco-related diseases accounted for 5.7 percent of global health expenditure in 2012 and that the total economic costs of smoking, including health expenditure and productivity losses, were equivalent to 1.8 percent of the world’s GDP (US$1.9 trillion in purchasing power parity U.S. dollars). The highest share, according to these authors, was in high-income countries (US$1.1 trillion in purchasing power parity dollars), where the tobacco epidemic is the most advanced.3 Nearly 40 percent of these costs are concentrated in developing coun- tries, reflecting the substantial burden experienced by this group of countries. Earlier esti- mates of Lightwood et al. (2000) indicate that the gross health costs of tobacco in high-in- come countries are between 0.1 percent and 1.0 percent of GDP. Likewise, Pichón-Riviere et al. (2014) estimate the annual direct costs of tobacco-related disease in the Chilean health system at approximately 0.6 percent of GDP. 3 Goodchild, Nargis, and Tursan d’Espaignet (2018) estimate the economic cost of smoking-attributable diseases at US$15 billion in low-income countries, US$359 billion in lower-middle-income countries, and US$354 billion in upper-middle-income countries, all in purchasing power parity U.S. dollars. 6 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Tobacco price increases are also associated with expansion in productive life years. Ver- guet et al. (2015) analyze the health effects of a price increases in China and conclude that a 50 percent rise in prices would result in 231 million life years gained over 50 years and would have a significant impact among the poor. In contrast, Pichón-Riviere et al. (2014) estimate that tobacco use in Chile would reduce life expectancy by nearly 4.0 years among women and 4.3 years among men. On Indonesia, Ahsan et al. (2013) estimate that, if taxes were increased to 57 percent of the retail price, 1.96 million tobacco-related deaths would be averted. Similarly, if taxes were raised even more, to 70 percent, the number of deaths averted would increase to over 5 million. Similarly, exposure to secondhand smoke has a strong relationship with many respiratory diseases among children and adults (DHHS 2004, 2014; Mason, Wheeler, and Brown 2015; Öberg et al. 2011). In Indonesia, more than 97 million nonsmokers are constantly exposed to secondhand smoke (Barber et al. 2008). According to WHO (2017a), secondhand smoke is responsible for over 890,000 premature deaths per year. In the United States, exposure to secondhand smoke costs an estimated US$5 billion annually in direct med- ical costs and over US$5 billion more in indirect medical costs, that is, disability and lost wages (Behan, Eriksen, and Lin 2005). In the state of Indiana, the health-related costs of secondhand smoking have been estimated at more than US$1.3 billion annually (Mason, Wheeler, and Brown 2015). McGhee et al. (2006) estimate the cost of direct medical care, long-term care, and productivity losses because of secondhand smoke exposure in Hong Kong SAR, China, at approximately US$156 million annually. Tobacco taxation has been recognized as one of the most effective tobacco control strat- egies. In high-income countries, a 10 percent increase in the price of cigarettes is asso- ciated with a decrease in the demand for cigarettes of approximately 4 percent (World Bank 1999). In low- and middle-income countries, an equivalent increase is associated with an average 6 percent reduction in cigarette consumption (Chaloupka et al. 2000). Higher taxes have the additional benefits of reducing exposure to secondhand smoke and increasing government revenues. Djutaharta et al. (2005) estimate that a 10 percent increase in price would reduce total cigarette consumption in Indonesia by 6 percent. The decrease would be higher (7 percent) among low-income households and lower among wealthier ones (3 percent). Moreover, the study estimates that a 50.0 percent tax increase would raise tobacco tax revenues by 27.5 percent. A similar study by Barber et al. (2008) estimates that raising cigarette taxes to 50 percent of the retail price can decrease the expected mortality rate by 2 percent–5 percent. Adioetomo, Djutaharta, and Hendratno (2005) simulate the impact of a 10.0 percent tax increase and find that the subsequent increase in cigarette prices (4.9 percent) causes a 3.0 percent decline in consumption. Other tobacco control interventions are also relevant in decreasing demand, but have a smaller impact on tobacco consumption. Several studies have found that public health 7 // Literature Review awareness campaigns contribute to a reduction in cigarette use, but the impact has generally been small and, in some cases, temporary.4 Levy, de Almeida, and Szklo (2012) estimate that the vast reduction in tobacco use in Brazil was mostly caused by higher tobacco prices (46 percent of the impact) and, to a lesser extent, by smoke-free policies (14 percent). WHO (2015b) argues that smoke-free environments are the only way to miti- gate the harmful impacts of secondhand smoking. Policy makers can design an effective tobacco taxation system based on information on the price elasticity of tobacco products. Tax increases tend to generate more impact on tobacco consumption in low- and middle-income countries relative to high-income ones (WHO 2015b). Several studies have estimated the relationship between tobacco prices and consumption. 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. Cigarette price elasticities across income groups in India range from −0.83 for the lowest income group and −0.26 for the highest (Selvaraj, Srivastava, and Karan 2015). Gallus et al. (2006) estimate a price elasticity of −0.46 for 52 countries in Europe. In the United Kingdom, price elasticity is estimated at −0.5 and, in Hungary, between −0.44 and −0.37 (Szilágyi 2007; Townsend, Roderick, and Cooper 1994). 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 (2017a) 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. In Moldova, Fuchs and Meneses (2018) find a price elasticity of -0.51 for the poorest decile and -0.26 for the richest. The price elasticity of cigarettes in Indonesia ranges from −0.26 to −0.76. Bird (1999) uses annual aggregate data on Indonesia from 1970 to 1994 to estimate a price elasticity of the demand for cigarettes at −0.43 in the long run and −0.60 in the short run. Similarly, Djutaharta et al. (2002) use aggregate time series data on Indonesia from 1970 to 1996 to estimate an elasticity of −0.57, while, in 1970–2001, they the elasticity at −0.35. Based on the 1999 Indonesia National Socioeconomic Survey, Adioetomo, Djutaharta, and Hen- dratno (2005) estimate a total price elasticity of cigarette demand of −0.61 and find that price elasticity decreased in absolute terms at higher income levels: the price elasticity is −0.67 among the low-income group, −0.33 among the middle-income group, and −0.31 among the high-income group. For clove cigarettes, de Beyer and Yurekli (2000) use a log-linear model with 1980–95 data to estimate a price elasticity of demand of −0.51. Nas- rudin et al. (2013) use the Indonesian Family Life Survey to estimate cigarette price elas- 4 Atkinson and Skegg (1973); Bardsley and Olekalns (1999); Stavrinos (1987); Sumner (1971); Townsend (1987); Townsend, Roderick, and Cooper (1994); Witt and Pass (1981). 8 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia ticity: −0.15 and −0.16 among the first two income deciles and −0.20 to −0.28 among the third to ninth deciles. Witeolar, Rukumnuaykit, and Strauss use the household panel of the Indonesian Family Life Survey in 1997 and 2000 to estimate a price elasticity of −0.80. Two key factors that determine tobacco price elasticities are age and income. Young adults and individuals in low-income groups are more responsive to price changes rel- ative to their peers. This makes them particularly susceptible to tobacco tax increases because they tend to be less dependent on nicotine, more affected by peer effects, and possess less disposable income (Jha and Peto 2014). Several studies in the United States have consistently shown that younger groups show higher elasticities relative to older ones (CDC 1998; Chaloupka and Grossman 1996; Lewit and Coate 1982). Barkat et al. (2012) estimate that a uniform excise tax accounting for 60 percent of the retail price of cigarettes would reduce youth smoking prevalence by 27 percent, preventing over 2 mil- lion youth from taking up smoking. 9 // Literature Review 10 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN 3 MODEL The impact of an increase in tobacco taxes in Indonesia is estimated using an extended cost-benefit analysis as in other studies (Pichón-Riviere et al. 2014; Verguet et al. 2015). The paper analyzes three factors to estimate how tobacco taxes could affect household income. First, assuming tobacco consumption does not change, tobacco taxes directly affect household income because the share of household budgets allocated to tobacco purchases rises as the tax increases. Second, household medical expenses could decrease as a result of reduced tobacco consumption. Third, households could also experience a positive income change because of additional years of labor recovered through the extension of life expectancy. The aggregate effect of a tax policy is estimated as follows: Income effect = change in tobacco expenditure (A) + lower medical expenses (B) + higher income (C) (1) For further details on the methodology, see annex A or refer to Fuchs and Meneses (2017a). 12 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN 4 DATA AND DESCRIPTIVE STATISTICS Tobacco Prices Data on the household consumption of and expenditure on tobacco products in Indo- nesia come from the 2015–16 Indonesia National Socioeconomic Surveys. The surveys ask households about daily expenditure and consumption for different tobacco products. These questions allow an estimation of the average price paid by households for tobacco products and of average price elasticities among expenditure groups. However, the sur- vey is not based on a good representative sample of higher-income groups because non- response rates among these groups are high. Table 1 shows significant price variation across deciles and minimal variation by cigarette type. For instance, the poorest decile in 2016 paid an average Rp 495 for clove cigarettes, whereas the richest decile spent Rp 909. For white cigarettes, the variation across deciles is within a similar range. Using Euromonitor data, one can compare the estimated prices from the expenditure survey.5 The most expensive 20-pack of cigarettes in 2017 cost Rp 24,300, while the cheapest cost Rp 17,200 (figure 1). Moreover, while nominal cigarette prices increased from 2011 to 2017, the real prices remained relatively unchanged (figure 2). Table 1. Mean and Median Price per Stick Paid by Households, by Cigarette Type and Decile, 2016 PRICE TYPE DECILE 1 2 3 4 5 6 7 8 9 10 Clove Cigaretes 495 530 557 587 630 691 741 781 831 909 Mean White cigarettes 523 559 571 615 633 714 746 792 812 884 Median Clove cigarettes 469 511 539 574 618 671 730 733 795 863 White cigarettes 479 535 556 597 633 689 730 730 771 821 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Prices are in 2010 real rupiah. Deciles are constructed using per capita household expenditure. 5 See Euromonitor International (database), Euromonitor International, London, http://www.euromonitor.com/. 14 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Figure 1. Mean Cigarette Price, by Year and Pack Size (2010 prices) 20,000 PRICE RUPIAH 15,000 10,000 5,000 0 2011 2012 2013 2014 2015 2016 2017 YEAR 12-pack 16-pack 20-pack Source: Authors’ calculations from kretek roller survey. Note: The percentage shares are weighted by the size of household. Tobacco Price Elasticity by Decile Although several models have been tested to estimate decile-level elasticities, the fixed effects with controls model is preferred. (See annex B for more on the methodology and the various iterations of the estimation of tobacco price elasticities by decile.) The estimated average white cigarette price elasticity in Indonesia is −0.63, which is within the estimated elasticities found in the literature on developing countries (a range of −0.4 to −0.8), as well as previous estimates on Indonesia. As expected, lower-income deciles exhibit higher white cigarette price elasticities relative to richer deciles. For instance, the poorest decile has a medium-bound elasticity of −0.77, whereas the richest has an elasticity of −0.55 (table 2). The standard error of these estimates is approximately 0.10, producing a 95 per- cent confidence interval of −0.20, +0.20.6 The effect of a tax increase under different scenar- ios is shown by simulating a lower- and upper-bound elasticity for each decile. The former tends to involve income groups that would not change consumption patterns, such as rural 6 To maintain comparability with similar studies, the same average upper and lower bounds have been applied to all deciles. Applying individual bounds by decile produces minimal changes in the results. 15 // Data and Descriptive Statistics Figure 2. Mean Cigarette Price, by Year and Pack Size, Real vs. Nominal Prices PRICE 25,000 RUPIAH 20,000 15,000 10,000 5,000 0 2011 2012 2013 2014 2015 2016 2017 YEAR Real Prices 12-pack Real Prices 20-pack Nominal Prices 12-pack Nominal Prices 20-pack Source: Authors’ calculations from kretek roller survey. Note: The percentage shares are weighted by the size of household. residents or older individuals, while the latter tends to show a longer-term scenario, reflect- ing the effect the tobacco tax would have on younger individuals. After a few decades, only these would still be alive; the total average effect of the price increase would therefore be approximated more accurately by the upper-bound price elasticity. Table 2. White Cigarettes: Price Elasticities, by Declie PRICE DECILE ELASTICITY 1 2 3 4 5 6 7 8 9 10 Lower bound −0.567 −0.498 −0.468 −0.444 −0.422 −0.403 −0.393 −0.382 −0.366 −0.345 Medium bound −0.767 −0.698 −0.668 −0.644 −0.622 −0.603 −0.593 −0.582 −0.566 −0.545 Upper bound −0.967 −0.898 −0.868 −0.844 −0.822 −0.803 −0.793 −0.782 −0.766 −0.745 Source: Estimates based on Indonesia National Socioeconomic Survey 2015–16. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average dif- ferences of −0.2 and +0.2, respectively, with the medium-bound elasticity. Demographic controls include: the age, education, and gender of the household head, the share of individuals by age-group in each household, and urban status. The model also controls for year fixed effects. Only 2015–16 survey years were used to estimate price elasticities because previous surveys asked about the number of cigarette packs consumed, whereas the most recent years ask about quantities of individual cigarettes. 16 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table 3. Clove Cigarettes: Price Elasticities, by Decile PRICE DECILE ELASTICITY 1 2 3 4 5 6 7 8 9 10 Lower bound −0.440 −0.386 −0.354 −0.334 −0.318 −0.302 −0.293 −0.283 −0.272 −0.256 Medium bound −0.640 −0.586 −0.554 −0.534 −0.518 −0.502 −0.493 −0.483 −0.472 −0.456 Upper bound −0.840 −0.786 −0.754 −0.734 −0.718 −0.702 −0.693 −0.683 −0.672 −0.656 Source: Estimates based on Indonesia National Socioeconomic Surveys 2015–16. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average dif- ferences of −0.2 and +0.2, respectively, with the medium-bound elasticity. Demographic controls include: the age, education, and gender of the household head, the share of individuals by age-group in each household, and urban status. The model also controls for year fixed effects. Only 2015–16 survey years were used to estimate price elasticities because previous surveys asked about the number of cigarette packs consumed, whereas the most recent years ask about quantities of individual cigarettes. Table 4. All Cigarette Products: Price Elasticities, by Decile PRICE DECILE ELASTICITY 1 2 3 4 5 6 7 8 9 10 Lower bound −0.447 −0.345 −0.313 −0.293 −0.276 −0.261 −0.251 −0.241 −0.230 −0.213 Medium bound −0.647 −0.545 −0.513 −0.493 −0.476 −0.461 −0.451 −0.441 −0.430 −0.413 Upper bound −0.847 −0.745 −0.713 −0.693 −0.676 −0.661 −0.651 −0.641 −0.630 −0.613 Source: Estimates based on Indonesia National Socioeconomic Surveys 2015–16. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average dif- ferences of −0.2 and +0.2, respectively, with the medium-bound elasticity. Demographic controls include: the age, education, and gender of the household head, the share of individuals by age-group in each household, and urban status. The model also controls for year fixed effects. Only 2015–16 survey years were used to estimate price elasticities because previous surveys asked about the number of cigarette packs consumed, whereas the most recent years ask about quantities of individual cigarettes. In contrast, average price elasticities are lower for clove cigarettes than for white cigarettes (−0.52). The poorest decile shows a medium-bound price elasticity of −0.64 for clove ciga- rettes, which is lower than the −0.77 for white cigarettes (tables 2–4, figures 3–4). Mortality and Morbidity Data on mortality, years of life lost, and morbidity are obtained from the Global Burden of Disease Study.7 Ischemic heart disease, chronic respiratory diseases, and tuberculosis are among the most prevalent diseases among men and women in Indonesia. In 2016, 7 See GBD Results Tool (database), Global Burden of Disease Study 2016, Global Health Data Exchange, Institute for Health Metrics and Evaluation, Seattle, http://ghdx.healthdata.org/gbd-results-tool. 17 // Data and Descriptive Statistics Figure 3. White Cigarettes: Price Elasticity, Indonesia -0.2 PRICE ELASTICITY -0.4 -0.6 -0.8 -1.0 1 2 3 4 5 6 7 8 9 10 DECILE Lower Bound Medium Bound Upper Bound Figure 4. Clove Cigarettes: Price Elasticity, Indonesia -0.2 PRICE ELASTICITY -0.4 -0.6 -0.8 -1.0 1 2 3 4 5 6 7 8 9 10 DECILE Lower Bound Medium Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Surveys 2015-2016. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average differenc- es of −0.2 and +0.2, respectively, with the medium-bound elasticity. 18 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table 5. Tobacco-related Deaths, by Gender and Age-Group, Indonesia 2016 FEMALE MALE 25–34 35–54 55–74 75+ TOTAL 25–34 35–54 55–74 75+ TOTAL Total Tobacco-related deaths 148 3,797 8,649 5,147 17,740 3,901 40,160 69,136 22,172 135,368 Ischemic heart disease 65 1,808 5,092 2,720 9,684 1,705 18,017 26,742 7,890 54,354 Cerebrovascular disease 38 1,264 2,714 2,107 6,123 1,165 14,888 27,204 7,634 50,892 Tuberculosis 11 153 357 283 804 810 4,279 5,626 2,162 12,876 Chronic obstructive pulmonary 9 184 282 31 507 76 916 5,057 3,281 9,329 disease Tracheal, bronchus, and lung 12 199 114 3 328 68 1,196 3,051 839 5,153 cancer Cervical cancer 4 62 17 0 83 - - - - - Lip and oral cavity cancer 3 32 17 0 52 18 150 179 52 399 Larynx cancer 0 20 25 1 46 5 111 348 90 554 Stomach cancer 2 18 7 0 27 8 103 194 55 359 Nasopharynx cancer 1 16 7 0 24 11 167 244 38 460 Colon and rectum cancer 1 16 6 0 24 10 48 50 13 121 Esophageal cancer 1 15 7 0 22 3 96 152 31 282 Bladder cancer 0 4 3 0 7 2 37 118 52 209 Other chronic respiratory diseases 0 3 1 0 4 4 14 4 1 23 Leukemia 0 2 1 0 3 8 48 51 10 117 Liver cancer due to other causes 0 1 1 0 2 7 90 118 24 240 Source: GBD Results Tool (database), Global Burden of Disease Study 2016, Global Health Data Exchange, Institute for Health Metrics and Evaluation, Seattle, http://ghdx.healthdata.org/gbd-results-tool. approximately 153,108 deaths were attributed to tobacco-related diseases, of which nearly 88 percent were men (135,368) (table 5). Nearly 2.2 million cases of tobacco-related disease were reported in Indonesia the same year (table 6). A total of 4,155,652 years of life lost were estimated for all tobacco-related deaths in 2016 (table 7).8 7 See GBD Results Tool (database), Global Burden of Disease Study 2016, Global Health Data Exchange, Institute for Health Metrics and Evaluation, Seattle, http://ghdx.healthdata.org/gbd-results-tool. 8 Years of life lost is the estimated average number of years a person would have lived if he or she had not died prematurely. 19 // Data and Descriptive Statistics Table 6. Tobacco-Related Events, by Gender, 2016 EVENT ALL WOMEN MEN Ischemic heart disease 673,152 269,411 403,741 Tuberculosis 622,909 244,059 378,849 Chronic obstructive pulmonary disease 500,616 166,757 333,859 Cerebrovascular disease 334,295 158,159 176,136 Cervical cancer 25,192 25,192 Colon and rectum cancer 16,695 8,260 8,435 Tracheal, bronchus, and lung cancer 13,227 3,372 9,854 Stomach cancer 12,642 5,616 7,025 Lip and oral cavity cancer 6,429 3,395 3,034 Bladder cancer 6,177 1,750 4,426 Leukemia 4,675 2,034 2,641 Nasopharynx cancer 3,771 919 2,852 Liver cancer due to other causes 3,386 370 3,016 Larynx cancer 2,694 478 2,216 Esophageal cancer 1,729 539 1,190 Total tobacco-related events 2,227,589 890,313 1,337,275 Source: GBD Results Tool (database), Global Burden of Disease Study 2016, Global Health Data Exchange, Institute for Health Metrics and Evaluation, Seattle, http://ghdx.healthdata.org/gbd-results-tool. 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. Tobacco-related Medical Costs The most recent study estimating the economic cost of smoking in Indonesia was con- ducted in 2015. The study estimates the treatment costs of 19 chronic diseases attribut- able to smoking. Annual direct health care costs attributable to tobacco use amount to US$2.2 billion (equivalent to 2.5 percent of GDP in 2015). The treatment costs of hyper- tension, chronic obstructive pulmonary diseases, and ischemic heart disease showed the highest cost burden (Kristina et al. 2018). Similarly, a 2008 study estimated the economic loss attributable to tobacco use at US$18.5 billion (NIHRD 2009). Kristina et al. (2015) estimate the indirect mortality costs of premature cancer deaths and years of potential life lost attributable to smoking in Indonesia. They estimate smoking-attributable cancer 20 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table 7. Years of Life Lost, by Gender, 2016 YEARS ALL WOMEN MEN 30–34 220,583 8,060 212,523 35–39 302,657 14,839 287,819 40–44 400,801 29,242 371,559 45–49 495,617 40,812 454,805 50–54 576,637 63,748 512,889 55–59 609,999 49,992 560,007 60–64 546,819 55,091 491,729 65–69 404,476 52,486 351,991 70–74 308,666 42,070 266,596 75–79 178,786 29,140 149,645 80 plus 110,610 22,656 87,954 Total 4,155,652 408,136 3,747,516 Source: GBD Results Tool (database), Global Burden of Disease Study 2016, Global Health Data Exchange, Institute for Health Metrics and Evaluation, Seattle, http://ghdx.healthdata.org/gbd-results-tool. mortality at nearly a third of total cancer deaths (30.6 percent) and the years of life lost at 1.2 million. Furthermore, cancer mortality costs related to smoking amounted to US$1.3 million in 2013. Baseline Descriptive Results Table 8 summarizes the most important indicators, including total monthly household expenditures and the share of expenditures on smoked tobacco products, for 2016. As of 2016, there was at least one smoker in 64 percent of all households. The share was high- est among households in deciles 4–6 (an average 69.1 percent) and lowest in high-in- come households (52.0 percent). Most households consume clove cigarettes (59.7 per- cent) rather than white cigarettes (5.6 percent). On average, households spent 6.8 percent of their total incomes on white and clove cigarettes. Lower-income households spent the highest share. For instance, the poorest decile spent 7.5 percent relative to less than 4.0 percent among the richest. The results do not significantly differ if one analyzes the expenditures for tobacco products using 2010 as the base year (see annex C). 21 // Data and Descriptive Statistics INDICATOR DECILE 1 DECILE 2 DECILE 3 DECILE 4 DECILE 5 DECILE 6 DECILE 7 DECILE 8 DECILE 9 DECILE 10 ALL Average Household Monthly per capita 233,370 310,272 374,363 449,352 532,701 631,259 764,097 947,436 1,286,670 3,952,778 1,031,795 Expenditure (Rupia) spending on tobacco products. Proportion of exp. on all tobacco 7.5 8.8 9.2 9.5 9.5 9.3 8.5 7.6 6.2 3.3 6.8 products a Proportion of exp. white cigarettes 0.3 0.4 0.5 0.6 0.6 0.7 0.7 0.8 0.8 0.6 0.6 Proportion of exp. clove cigarettes 7.2 8.3 8.7 8.9 8.9 8.6 7.8 6.8 5.3 2.7 6.2 Households that smoke any tobacco 57.1 64.0 66.6 69.3 69.5 68.5 66.9 65.0 60.5 51.9 63.9 product (%) Households that smoke white 2.3 3.3 3.7 4.6 4.9 5.5 5.9 7.1 8.6 9.5 5.6 Table 8. Baseline Descriptive Results, 2016 cigarettes (%) Note: Deciles were created using per capita household expenditure. Households that smoke clove 55.3 61.8 64.0 66.1 65.9 64.5 62.4 59.6 53.5 43.4 59.7 cigarettes (%) Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Woman-headed households (%) 16.5 15.0 14.6 13.8 14.2 14.2 15.0 14.6 15.2 17.1 15.0 Age, household head 50.3 48.9 47.9 47.8 47.7 47.4 47.6 47.1 46.4 45.4 47.7 Share of kids ages 0-5 in household (%) 12.0 10.7 10.7 10.1 9.6 8.8 8.2 7.8 7.0 5.3 9.0 Share of individuals ages 6-17 in 22.4 22.2 22.1 20.9 20.2 19.7 18.1 17.1 15.3 12.2 19.0 household (%) Share of adults ages 18-49 in 39.2 42.4 44.5 45.9 47.1 48.5 49.4 51.1 54.4 58.0 48.1 household (%) Share of individuals ages 50+ in 26.4 24.7 22.7 23.1 23.1 23.0 24.3 24.0 23.3 24.5 23.9 household (%) a. Proportional to total budget per household in each decile; households that smoke = 1 if a household member reports 22 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN AN 5 RESULTS To analyze the distributional effects of an increase in tobacco taxes, the effect is estimated on prices, medical expenditures, and years of working life, aggregating these three indi- cators into a single measure. The price elasticities estimated in tables 2–4, including the HAS lower- and upper-bound elasticities, facilitate an understanding of how the results could change under different assumptions. Tobacco Price Increase Income changes that arise from an increase in tobacco prices are estimated for each LY decile based on lower-, medium-, and upper-bound elasticities. Using the price elas- ticities and the shares of household expenditure on tobacco by decile, the effects of an increase in tobacco prices can be simulated (see annex A, equation A2). To show the effect of the elasticities on prices, tables 9 and 10 also include estimates of a com- plete 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 rise by 25 percent, given the medium-bound elasticity for clove cigarettes in the bottom decile (−0.64) in table 3 and the proportion of clove cigarette expenditures among the bottom decile (4.3 percent) in table 8, the expected decrease in household expenditures would be 0.36 percent (table 10).9 This represents a loss in welfare because consumers would devote a higher share of their incomes CCO to the purchase of the same amount of tobacco, thereby reducing the consumption of other goods. In contrast, the analysis for white cigarettes implies an overall smaller decrease, whereby, for example, the same medium-elasticity analysis for the first decile yields a 0 percent expected decrease in household expenditures. These results hold for all analyzed scenarios. The effect for white cigarettes appears to be relatively progres- THE sive, that is, it affects upper-income groups in a larger proportion, whereas, for clove cigarettes, the effect is U shaped, with an increasing effect from decile 1 through 6 that then reverses so that the top decile is less affected than the poorest decile (lower- and medium-bound elasticity cases; see figure 5 and table 10). BELOW 9  The analysis assumes the prices of both clove and white cigarette increase by 25 percent. However, given the complex tobacco-tax structure in Indonesia, there may be a scenario whereby prices increase less for clove cigarette than for white cigarettes. NDED 24 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table 9. Direct Effect of a Price Increase through Taxes, White Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Complete pass-through −0.07 −0.11 −0.12 −0.15 −0.16 −0.17 −0.18 −0.19 −0.21 −0.14 Lower-bound elasticity −0.02 −0.04 −0.05 −0.07 −0.07 −0.08 −0.09 −0.10 −0.11 −0.08 Medium elasticity 0.00 −0.01 −0.02 −0.03 −0.03 −0.04 −0.05 −0.05 −0.06 −0.05 Upper-bound elasticity 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 −0.01 −0.01 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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 10. Direct Effect of a Price Increase through Taxes, Clove Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Complete pass-through −1.81 −2.09 −2.18 −2.22 −2.23 −2.14 −1.94 −1.70 −1.33 −0.67 Lower-bound elasticity −0.81 −1.08 −1.22 −1.29 −1.34 −1.33 −1.23 −1.10 −0.88 −0.46 Medium elasticity −0.36 −0.56 −0.67 −0.74 −0.79 −0.80 −0.75 −0.67 −0.55 −0.29 Upper-bound elasticity 0.09 −0.04 −0.13 −0.18 −0.23 −0.26 −0.26 −0.25 −0.21 −0.12 Source: Estimates based on Indonesia National Socioeconomic Surveys 2015–16. Note: Deciles were created using per capita household expenditure. Lower- and upper-bound elasticities show average dif- ferences of −0.2 and +0.2, respectively, with the medium-bound elasticity. Demographic controls include: the age, education, and gender of the household head, the share of individuals by age-group in each household, and urban status. The model also controls for year fixed effects. Only 2015–16 survey years were used to estimate price elasticities because previous surveys asked about the number of cigarette packs consumed, whereas the most recent years ask about quantities of individual cigarettes. 25 // Results Figure 5. Income Gains: Direct Effect of Tobacco Taxes (Increase in expenditure because of tobacco taxes) A. WHITE CIGARETTES 0 INCOME GAINS % -.05 -.1 -.15 -.2 0 2 4 6 8 10 DECILE CI 95% Direct pass through B. CLOVE CIGARETTES Lower Bound Medium Bound Upper Bound 0 INCOME GAINS % -.05 -.1 -.15 -.2 0 2 4 6 8 10 DECILE CI 95% Direct pass through Lower Bound Medium Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. 26 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Medical Expenses Tables 11 and 12 and figure 6 report the income effect of a reduction in medical expenses by cigarette type. As noted above, the model assumes that the health effects of tobac- co-related diseases will immediately diminish with the reduction in tobacco consump- tion. Although this assumption is implausible in the short term because changes in the effects of tobacco-related diseases take time to materialize, it provides an upper-bound estimate of the effects of tax increases. Nonetheless, discounted rates of medical cost savings have been analyzed and the results hold (see annex C). The overall results indicate that, for both types of cigarettes, the reduction in medical expenditures is highly progres- sive, disproportionally benefiting lower-income groups, especially for clove cigarettes. This derives from two factors: (1) the higher price elasticity and (2) a lower income base that massively benefits from the reduction in medical costs. Table 11. Reduction in Medical Costs, White Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.17 0.16 0.14 0.14 0.12 0.11 0.10 0.10 0.09 0.05 Medium elasticity 0.23 0.22 0.20 0.20 0.18 0.17 0.15 0.15 0.14 0.08 Upper-bound elasticity 0.29 0.28 0.26 0.27 0.24 0.22 0.20 0.20 0.19 0.11 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: The table reports the share of total consumption for each decile Table 12. Reduction in Medical Costs, Clove Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.29 0.21 0.17 0.14 0.11 0.09 0.07 0.06 0.04 0.02 Medium elasticity 0.42 0.32 0.27 0.22 0.19 0.15 0.12 0.10 0.07 0.03 Upper-bound elasticity 0.54 0.44 0.36 0.31 0.26 0.22 0.17 0.14 0.10 0.04 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: The table reports the share of total consumption for each decile 27 // Results Figure 6. Income Gains: Direct Effect of Tobacco Taxes (Increase in expenditure because of tobacco taxes) A. WHITE CIGARETTES .3 INCOME GAINS % .25 .2 .15 .1 .05 0 2 4 6 8 10 DECILE CI 95% Medium Bound B. CLOVE CIGARETTES Lower Bound Upper Bound .5 INCOME GAINS % .4 .3 .2 .1 0 0 2 4 6 8 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. 28 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Income Gains Deriving From an Increase in Years of Working Life The study estimates the cost of working life lost because of tobacco consumption based on the assumption that the impact of lower tobacco use on health and work-generated income is direct. For each tobacco-related death, the working years lost shown in table 7 (that is, forgone income) are divided across deciles proportionately to the number of households that consume tobacco in each income group. The model anticipates that income will increase as the number of years lost because of premature deaths from tobacco consumption decline (see annex A for methodology details). The results show that the reduction in tobacco consumption and the expected reduction in years of work lost have a positive impact on welfare (see figure 7; tables 13–14). For white cigarettes, the positive impact increases with the income decile, that is, it has a highly regressive effect. In contrast, for clove cigarettes, the impact is more evenly distributed across deciles, gen- erating an important impact on lower-income groups. Table 13. Income Gains: Production During Years Lost, by Decile, White Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.08 0.11 0.12 0.14 0.15 0.16 0.18 0.21 0.26 0.30 Medium elasticity 0.11 0.15 0.17 0.21 0.22 0.24 0.27 0.32 0.40 0.47 Upper-bound elasticity 0.14 0.19 0.22 0.27 0.29 0.32 0.35 0.43 0.54 0.65 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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. Table 14. Income Gains: Production During Years Lost, by Decile, Clove Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.14 0.15 0.14 0.14 0.14 0.13 0.13 0.12 0.11 0.09 Medium elasticity 0.21 0.22 0.22 0.23 0.23 0.22 0.22 0.21 0.19 0.17 Upper-bound elasticity 0.27 0.30 0.30 0.31 0.31 0.31 0.30 0.30 0.27 0.24 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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. 29 // Results Figure 7. Income Gains: Production During Years Lost, by Decile A. WHITE CIGARETTES .6 INCOME GAINS % .5 .4 .3 .2 .1 0 2 4 6 8 10 DECILE CI 95% B. CLOVE CIGARETTES Medium Bound Lower Bound Upper Bound .3 INCOME GAINS % .25 .2 .15 .1 0 2 4 6 8 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. Years of life lost have been estimated using all deaths from tobacco-related diseases. 30 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Net Effects: Total Distributional Impact In this subsection, the results above have been added to estimate the aggregate effect of an increase in tobacco taxes (see equation 1). For white cigarettes, the aggregate effect of a 25 percent increase in tobacco taxes is positive, but regressive; in the long run, poorer deciles benefit less than richer ones. The positive effect of reduced medical expenses and years of life gained offset the negative price effect (figure 8; table 15). Table 15. Net Effect on Household Expenditures, White Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.23 0.22 0.21 0.22 0.20 0.19 0.19 0.21 0.23 0.27 Medium elasticity 0.33 0.35 0.35 0.38 0.36 0.37 0.37 0.42 0.48 0.51 Upper-bound elasticity 0.44 0.49 0.49 0.54 0.53 0.55 0.56 0.63 0.72 0.75 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: The table reports the share of total consumption for each decile. In contrast, for clove cigarettes, the effect is highly dependent on the elasticity scenario (figures 8 and 9, table 16). Using a lower-bound elasticity, the results show that the net effect of clove tax policy is negative and also regressive. A population that is not sensitive to tobacco price changes will not reduce consumption sufficiently to allow for health and work benefits to offset cost increases. In the case of the medium-bound elasticity, the tax has a progressive impact because of a positive effect on the two first deciles and a negative effect on higher-income groups. For upper-bound elasticity, the tax would have a positive effect on all income groups and show a progressive pattern, that is, bene- fiting lower-income groups in larger proportion. Although this effect is driven mostly by the elasticity variance among income deciles, the elasticity level is also relevant. The tax would have a positive effect on most income groups, but also show a progressive pattern, that is, greater benefits for lower-income groups. 31 // Results Figure 8. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes (Tobacco Price Increase, Medical Expenditure, and Working Years Gained) A. WHITE CIGARETTES .7 INCOME GAINS % .6 .5 .4 .3 .2 0 2 4 6 8 10 DECILE CI 95% B. CLOVE CIGARETTES Medium Bound Lower Bound Upper Bound 1 INCOME GAINS % .5 0 -.5 -1 0 2 4 6 8 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. Years of life lost have been estimated using all deaths from tobacco-related diseases. 32 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table 16. Net Effect on Household Expenditures, Clove Cigarettes (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity −0.39 −0.72 −0.91 −1.01 −1.09 −1.11 −1.03 −0.92 −0.73 −0.35 Medium elasticity 0.26 −0.01 −0.18 −0.29 −0.37 −0.42 −0.41 −0.37 −0.29 −0.09 Upper-bound elasticity 0.91 0.70 0.54 0.44 0.34 0.26 0.22 0.19 0.16 0.16 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: The table reports the share of total consumption for each decile. In conclusion, under any tobacco price elasticity scenario, the overall effect in the case of white cigarettes is always positive and regressive. Under a low tobacco price elasticity scenario, the results for clove cigarettes show a negative overall effect. In the medi- um-bound elasticity scenario, the results for clove cigarettes are mixed, that is, positive among lower-income groups and negative among higher-income groups. In the case of an upper-bound elasticity scenario, the results would be income gains across all deciles and a progressive pattern. Moreover, the impact would be particularly important among lower-income deciles. The assumptions in this model do not include other possible pol- icies, such as smoking cessation programs, antismoking advertising, youth outreach, or policies financed through the new tax revenue. These results are in line with the literature, showing the key role taxation plays in lowering tobacco usage. 33 // Results Figure 9. Total Income Effect Decomposition, Medium-Bound Elasticity (Tobacco Price Increase, Medical Expenditure, and Working Years Gained) 0.8 0.7 PERCENTAGE POINTS % 0.6 0.5 0.4 0.2 0.3 0.0 0.1 -0.2 -0.4 -0.1 -0.6 -0.3 -0.8 -1.0 -0.5 1 2 3 4 5 6 7 8 9 10 DECILE 0.8 Price YLL 0.7 PERCENTAGE POINTS % Medical Exp. Net e ect 0.6 0.5 0.4 0.2 0.3 0.0 0.1 -0.2 -0.4 -0.1 -0.6 -0.3 -0.8 -1.0 -0.5 1 2 3 4 5 6 7 8 9 10 DECILE Price YLL Medical Exp. Net e ect Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. Years of life lost have been estimated using all deaths from tobacco-related diseases. Y-axis on the right refers to the line graph, y-axis on the left refers to the bar graphs. 34 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, TAXES REMAIN B THE RECOMMEN AN 6 DISCUSSION Despite the wealth of research on the negative effects of tobacco consumption and on the benefits of various public policy mechanisms aimed at reducing tobacco use, ques- tions remain about the progressivity or regressivity that these entail. The implementation HAS of tobacco taxes 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 and have more limited access to health insurance and adequate health care. Moreover, the net effect of an increase in tobacco taxes depends on the price elasticity of this product across different sectors of LY the population. The price elasticity determines the magnitude of the income shock and the benefits gained from a decline in tobacco consumption. To assess the net welfare gains from this policy, one must look beyond the direct impact on household income and consider other benefits of lower tobacco consumption, includ- ing 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 CCO tobacco taxation to determine if, in the end, the policy is regressive. The results show that, considered by itself, a price increase for tobacco through higher taxes generates negative income variations across all groups in the population because overall prices increase. For white and clove cigarettes, these negative effects are partic- ularly acute in the scenario of lower-bound elasticity, and they are more moderate as THE elasticity expands in absolute terms. Based on the assumptions of a more comprehensive approach, including benefits through lower medical expenses and an increase in working years, the overall monetary effect is estimated. The different types of cigarettes and elasticity assumptions generate four particular results. For white cigarettes, the overall effect is always positive, but not necessarily pro- BELOW gressive. For clove cigarettes, the lower-bound elasticity generates an income loss among all decile groups of the population. The middle-bound elasticity generates income losses across all groups except the first two deciles. The upper-bound elasticity generates NDED 36 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia income gains across all groups, but particularly lower-income groups. This means that the effect of taxes on the different income groups is exacerbated if the variation in price elas- ticities among income deciles is greater. The three price elasticity scenarios mimic the short- versus the long-term effects of tobacco taxes. There is evidence that adult smokers will only present small changes in their behavior if the price increases; the lower-bound elasticity tends to represent this behavior more closely. In contrast, younger groups of the population show more elastic demand, similar to the upper-bound elasticity. After a few decades, one should expect the impact of the tax policy to resemble the upper-bound elasticity scenario. 37 // Discussion REFERENCES Adioetomo, Sri Moertiningsih, Triasih Djutaharta, and Hendratno. 2005. “Cigarette Con- sumption, Taxation, and Household Income: Indonesia Case Study.” Economics of Tobacco Control Paper 26 (February), Health, Nutrition, and Population Discussion Paper Series, World Bank, Washington, DC. Ahsan, Abdillah, Nur Hadi Wiyono, Ayke Soraya Kiting, Triasih Djutaharta, and Flora Anin- ditya. 2013. “Impact of Increasing Tobacco Tax on Government Revenue and Tobacco Consumption.” SEADI Discussion Paper, Support for Economic Analysis Development in Indonesia Project, United States Agency for International Development, Washington, DC. Arnez M. 2009. “Tobacco and Kretek: Indonesian Drugs in Historical Change.” Austrian Jour- nal of South-East Asian Studies 2: 49–69. Atkinson, A. B., and J. L. Skegg. 1973. “Anti-smoking Publicity and the Demand for Tobacco in the UK.” Manchester School 41 (3): 265–82. Barber, Sarah, Sri Moertiningsih Adioetomo, Abdillah Ahsan, and Diahhadi Setyonaluri. 2008. “Tobacco Economics in Indonesia.” International Union against Tuberculosis and Lung Disease, Paris. Bardsley, P., and N. Olekalns. 1999. “Cigarette and Tobacco Consumption: Have Anti-smok- ing Policies Made a Difference?” Economic Record 75 (230): 225–40. Barkat, A., A. U. Chowdhury, N. Nargis, M. Rahman, Pk. A. Kumar, S. Bashir, and F. J. Cha- loupka. 2012. “The Economics of Tobacco and Tobacco Taxation in Bangladesh.” Interna- tional Union against Tuberculosis and Lung Disease, Paris. Behan, D. F., M. P. Eriksen, and Y. Lin. 2005. “Economic Effects of Environmental Tobacco Smoke.” March 31, Society of Actuaries, Schaumburg, IL. Bird, K. 1999. “Industrial Concentration and Competition in Indonesian Manufacturing.” PhD thesis, Australian National University, Canberra. Blecher, E. H., and C. P. van Walbeek. 2008. “An Analysis of Cigarette Affordability.” Interna- tional Union against Tuberculosis and Lung Disease, Paris. BPS (Badan Pusat Statistik, Statistics Indonesia), BKKBN (National Population and Family Planning Board), MOH (Kementerian Kesehatan), and ICF International. 2013. Indonesia Demographic and Health Survey 2012. Jakarta: BPS, BKKBN, MOH, and ICF International. 38 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia CDC (Centers for Disease Control and Prevention). 1998. “Response to Increases in Ciga- rette Prices by Race/Ethnicity, Income, and Age Groups: United States, 1976–1993.” Mor- bidity and Mortality Weekly Report 47 (29): 605. Chaloupka, F. J., and M. Grossman. 1996. “Price, Tobacco Control Policies, and Youth Smok- ing.” NBER Working Paper 5740 (September), National Bureau of Economic Research, Cam- bridge, MA. Chaloupka F. J., T. Hu, K. E. Warner, R. Jacobs, and A. Yurekli. 2000. “The Taxation of Tobacco Products.” In Tobacco Control in Developing Countries, edited by P. Jha and F. J. Chaloupka, 237–72. Oxford, UK: Oxford University Press. 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, Wash- ington, DC. de Bayer, Joy, and Ayda A. Yurekli. 2000. “Curbing the Tobacco Epidemic in Indonesia.” East Asia and the Pacific Region Watching Brief 6 (May), World Bank, 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 Feder- ation, and Ukraine.” Discussion Paper 92765 (October), Health, Nutrition, and Population Discussion Paper Series, World Bank, Washington, DC. https://openknowledge.worldbank. org/handle/10986/21114. 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. ———. 2014. The Health Consequences of Smoking, 50 Years of Progress: A Report of the Sur- geon General. January. Atlanta: Office on Smoking and Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, DHHS. http://www.surgeongeneral.gov/library/reports/50-years-of-progress/full-report.pdf. Djutaharta, Triasih, Henry Viriya Surya, N. Haidy, A. Pasay, Hendratno, and Sri Moertiningsih Adioetomo. 2005. “Aggregate Analysis of the Impact of Cigarette Tax Rate Increases on Tobacco Consumption and Government Revenue: The Case of Indonesia.” Economics of Tobacco Control Paper 25, Health, Nutrition, and Population Discussion Paper Series, World Bank, Washington, DC. 39 // References Doll, R., and A. B. Hill. 1956. “Lung Cancer and Other Causes of Death in Relation to Smok- ing.” British Medical Journal 22 (5001): 1071–81. Fuchs, A., and F. J. Meneses. 2017a. “Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine.” December 27, World Bank, Washington, DC. ———. 2017b. “Are Tobacco Taxes Really Regressive? Evidence from Chile.” World Bank, Washington, DC. ———. 2018. “Tobacco Price Elasticity and Tax Progressivity in Moldova.” World Bank, Washington, DC. Gallus, S., A. Schiaffino, C. La Vecchia, J. Townsend, and E. Fernandez. 2006. “Price and Ciga- rette Consumption in Europe.” Tobacco Control 15 (2): 114–19. Goodchild, M., N. Nargis, and E. Tursan d’Espaignet. 2018. “Global Economic Cost of Smok- ing-Attributable Diseases.” Tobacco Control 27 (1): 58–64. Guindon, G. E., P. Tobin, M. Lim, and D. Boisclair. 2007. “The Cost Attributable to Tobacco Use: A Critical Review of the Literature.” Tobacco Free Initiative, World Health Organization, Geneva. Jha, P., and R. Peto. 2014. “Global Effects of Smoking, of Quitting, and of Taxing Tobacco.” New England Journal of Medicine 370 (1): 60–68. Kostova, Deliana, Frank J Chaloupka, Ayda Yurekli, Hana Ross, Rajeev Cherukupalli, Linda Andes, and Samira Asma.. 2014. “A Cross-Country Study of Cigarette Prices and Affordabil- ity: Evidence from the Global Adult Tobacco Survey.” Tobacco Control 23: e3. Kpodar, K. 2006. “Distributional Effects of Oil Price Changes on Household Expenditures: Evidence from Mali.” IMF Working Paper WP/06/91, International Monetary Fund, Wash- ington, DC. Kristina, S., D. Endarti, Y. Prabandari, A. Ahsan, and M. Thavorncharoensap. 2015. “Burden of Cancers Related to Smoking among the Indonesian Population: Premature Mortality Costs and Years of Potential Life Lost.” Asian Pacific Journal of Cancer Prevention 16: 6903–08. Kristina, S., D. Endarti, C. Wiedyaningsih, A. Fahamsya, and N. Faizah. 2018. “Health Care Cost of Noncommunicable Diseases Related to Smoking in Indonesia, 2015.” Asian Pacific Journal of Cancer Prevention 30 (1): 29–35. Levy, D., L. M. de Almeida, and A. Szklo. 2012. “The Brazil SimSmoke Policy Simulation Model: The Effect of Strong Tobacco Control Policies on Smoking Prevalence and Smok- ing-Attributable Deaths in a Middle Income Nation.” PLoS Medicine 9 (11): e1001336. 40 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Lewit, E. M., and D. Coate. 1982. “The Potential for Using Excise Taxes to Reduce Smoking.” Journal of Health Economics 1 (2): 121–45. Lightwood, James, David Collins, Helen Lapsley, and Thomas E. Novotny. 2000. “Estimating the Costs of Tobacco Use.” In Tobacco Control in Developing Countries, edited by Prabhat Jha and Frank J. Chaloupka, 63–103. Oxford, UK: Oxford University Press. Malson, Jennifer L., Eun M. Lee, Ram Murty, Eric T. Moolchan, and Wallace B. Pickworth. 2003. “Clove Cigarette Smoking: Biochemical, Physiological, and Subjective Effects.” Phar- macology, Biochemestry, and Behavior 74: 739–45. Mason, J., W. Wheeler, and M. J. Brown. 2015. “The Economic Burden of Exposure to Sec- ondhand Smoke for Child and Adult Never Smokers Residing in US Public Housing.” Public Health Reports 130 (3): 230–44. McGhee, S. M., L. M. Ho, H. M. Lapsley, J. Chau, W. L. Cheung, S. Y. Ho, M. Pow, T. H. Lam, and A. J. Hedley. 2006. “Cost of Tobacco-Related Diseases, Including Passive Smoking, in Hong Kong.” Tobacco Control 15 (2): 125–30. Nasrudin, Rus'an, Ledi Trialdi, Djoni Hartono, and Abdillah Ahsan. 2013. “Tobacco Eco- nomic of Indonesia: Poor Households’ Spending Pattern, Tax Regressivity, and Economic Wide Impact of Tax Simplication.” Working Paper in Economics and Business III, 2, Depart- ment of Economics, Faculty of Economics, University of Indonesia, Jakarta. Ng, M., M. K. Freeman, T. D. Fleming, M. Robinson, L. Dwyer-Lindgren, B. Thomson, A. Wollum, et al. 2014. “Smoking Prevalence and Cigarette Consumption in 187 Countries, 1980–2012.” JAMA: Journal of the American Medical Association 311 (2): 183–92. NIHRD (National Institute of Health Research and Development). 2009. Study on Medical Expenditures and Burden of Major Tobacco Attributed Diseases in Indonesia. Jakarta, Indonesia: National Institute of Health Research and Development. Öberg, M., M. S. Jaakkola, A. Woodward, A. Peruga, and A. Prüss-Ustün. 2011. “Worldwide Burden of Disease from Exposure to Second-Hand Smoke: A Retrospective Analysis of Data from 192 countries.” Lancet 377 (9760): 139–46. 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, Sakthivel, Swati Srivastava, and Anup Karan. 2015. “Price Elasticity of Tobacco Products among Economic Classes in India, 2011–2012.” BMJ Open 5 (12): e008180. Stavrinos, V. G. 1987. “The Effects of an Anti-smoking Campaign on Cigarette Consump- tion: Empirical Evidence from Greece.” Applied Economics 19: 323–29. 41 // References Sumner, M. T. 1971. “The Demand for Tobacco in the UK.” Manchester School 39 (1): 23–36. Szilágyi, Tibor. 2007. “Higher Cigarette Taxes: Healthier People, Wealthier State, the Hungar- ian Experience.” Central European Journal of Public Health 15 (3): 122–26. http://apps.szu.cz/ svi/cejph/archiv/2007-3-07-full.pdf. Townsend, J. 1987. “Cigarette Tax, Economic Welfare, and Social Class Patterns of Smoking.” Applied Economics 19: 355–65. Townsend, J., P. Roderick, and J. Cooper. 1994. “Cigarette Smoking by Socioeconomic Group, Sex, and Age: Effects of Price, Income, and Health Publicity.” British Medical Journal 309 (6959): 923–27. Verguet, S., C. L. Gauvreau, S. Mishra, M. MacLennan, S. M. Murphy, E. D. Brouwer, R. A. Nugent, K. Zhao, P. Jha, and D. T. Jamison. 2015. “The Consequences of Tobacco Tax on Household Health and Finances in Rich and Poor Smokers in China: An Extended Cost-Ef- fectiveness Analysis.” Lancet Global Health 3 (4): e206–e216. WHO (World Health Organization). 2007. “Impact of Tobacco-Related Illnesses in Bangla- desh.” WHO, Geneva. ———. 2011. “Economics of Tobacco Toolkit: Assessment of the Economic Costs of Smok- ing.” WHO, Geneva. http://apps. who.int/iris/bitstream/10665/44596/1/9789241501576_eng. pdf. ———. 2015a. Global Report on Trends in Prevalence of Tobacco Smoking, 2015. Geneva: WHO. http://apps.who.int/iris/bitstream/10665/156262/1/9789241564922_eng.pdf. ———. 2015b. WHO Report on the Global Tobacco Epidemic 2015: Raising Taxes on Tobacco. Geneva: WHO. ———. 2017a. “Tobacco: Key Facts.” Tobacco Fact Sheet, WHO, Geneva. ———. 2017b. WHO Report on the Global Tobacco Epidemic, 2017: Monitoring Tobacco Use and Prevention Policies. Geneva: WHO. ———. 2017c. “WHO Report on the Global Tobacco Epidemic, 2017, Country Profile: Indonesia.” WHO, Geneva. http://www.who.int/tobacco/surveillance/policy/country_profile/ idn.pdf. Witoelar, Firman, Pungpond Rukumnuaykit, and John Strauss. 2005. “Smoking Behavior among Youth in a Developing Country: Case of Indonesia.” Economic Growth Center, Yale University, New Haven, CT. Witt, S. F., and C. L. Pass. 1981. “The Effects of Health Warnings and Advertising on the Demand for Cigarettes.” Scottish Journal of Political Economy 28 (1): 86–91. 42 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia World Bank. 1999. “Curbing the Epidemic: Governments and the Economics of Tobacco Control.” Report 19638 (May), Development in Practice Series, World Bank, Washington, DC. ———. 2017a. “The Economics of Kretek Rolling in Indonesia.” Health, Population, and Nutrition Global Practice, World Bank, Washington, DC. ———. 2017b. “The Economics of Clove Farming in Indonesia.” Health, Population, and Nutrition Global Practice, World Bank, Washington, DC. ———. 2017c. “The Economics of Tobacco Farming in Indonesia.” Health, Population, and Nutrition Global Practice, World Bank, Washington, DC. Wynder, E. L., and E. A. Graham. 1950. “Tobacco Smoking as a Possible Etiologic Factor in Bronchiogenic Car Cinoma: A Study of Six Hundred and Eighty-Four Proved Cases.” Journal of the American Medical Association 143: 329–36. Zheng, Rong, Patricio V. Marquez, Abdillah Ahsan, Yang Wang, and Xiao Hu. 2018. “Ciga- rette Affordability in Indonesia: 2002–2017.” Global Tobacco Control Program (May), World Bank, Washington, DC. 43 // References 44 OVER MORE THA A DECADE, THE INDONESIAN GOVERNMENT H BEEN GRADUALL RAISING EXCISE TAXES ON TOBAC PRODUCTS. YET, THE TAXES REMAIN BELOW T AN ANNEXES A. Model and Methodology This section describes the partial equilibrium approach used to simulate the impact on HAS Annexes consumption because of an increase in the price of cigarettes. This approach is used to Annexes Annexes Annexes A. evaluate Model theMethodology and Annexes first-order Annexes effects of a change in prices. It relies mainly on household expen- Annexes Annexes A. A. ModelModel and and Methodology Methodology diture patterns. TheA. focusModel isandon Methodology the impacts of a rise in the price of cigarettes because this A. A. Model Model and and Methodology Methodology This A. section Model and describes Methodology the A. partial Model and Methodology equilibrium approach used to simulate the impact on consumption has This been This section focus the describes section describesof the the potential thepartialpartial reform of equilibrium equilibrium tobacco approach approach taxes. used used to to simulate simulate the impact the impact onon consumption consumption because of an increase This insection the price describesof the partial cigarettes. This equilibrium approach approach is used toused evaluate to simulate the the first-order impact on consum This This section sectionThis describes describes section the the describes partial partial the equilibrium equilibrium partial approach approach equilibrium used used approach to to simulate simulate used to the the simulate impact impact theon on impactconsumpt consumpti approach used to simulate the impactThis because effects because section on consumption of a of of an describes change an increase increase in the partial because prices. in of It in the anthe price price equilibrium relies increase mainlyof cigarettes. ofin on cigarettes. approach the price household This This used approach approach to simulate cigarettes. ofexpenditure is used Thisis the approach patterns.used to impact evaluate to evaluate The on is focus the consumption used isthe first-order tofirst-order on evaluate the the on co first-o To assess effects effects of the of a adistributional because changebecause change in of of because in prices. anan prices. impact increase increase It of relies It an reliesof in inan increase mainly the the mainly increase on price price inon the of householdof in household the cigarettes. cigarettes. price price of of This This cigarettes. expenditure expenditure cigarettes, approach approach This patterns. patterns. the is is approach The simula- used used The focus to to is focus evaluate evaluate used is on is to onthe the the evaluate the first-ord first-orthe because of a an increase in the price of cigarettes. This approach used to is focus of evaluate the first-order LY ettes. This approach is used to evaluate the first-order impacts of rise in the effectsprice of of a change cigarettes prices. It in because relies this has mainly been the on household the expenditure potential patterns. reform of The focus is o tion allows impacts impacts of for of a aeffects effects differences rise risein the in of ofthe aa change effects price change in price of ofthe ofa in in cigarettes prices. prices. responses change cigarettes in ItIt because relies relies across prices. because mainly It thismainly consumption relies thishas has on onmainly been household household been the on the deciles focus focus expenditure expenditure householdof to the of reflect the potential patterns. patterns. expenditure potentialthe reform The The patterns. reform of focus focus of The is on is on t focu on household expenditure patterns. The effects focus of tobacco taxes. isa change on the in prices. impacts It of relies a rise mainly in the on price household of cigarettes expenditure because patterns. this has The been focus the focus is on of the the potential refor impacts impacts of of impacts aa rise rise inin of have the the price price a risebecause of in the priceof cigarettes cigarettes because because of cigarettes this this because has has been been the the focus focus of of the the potential potential reform reform fact impacts cause this has been the focus of the potentialtobacco tobacco that of apoor reform taxes. risetaxes. ofin households the tobaccoprice oflikely taxes. cigarettes different this price has elasticities been the focus relativeof this the tohas been reform households potential the focus of of the potential tobaccotobacco tobacco taxes. taxes. taxes. tobacco To assess with taxes. morethe distributional resources. The impactdifferent of anelasticities, increase in the combined price of with cigarettes, the initial the simulation consumption allows for ToTo differencesassessassess in the the distributional the distributional To responses assess acrosstheimpact impact of of distributional consumption an an increaseincrease impact deciles in of the in reflect to the an price price increase of of the cigarettes, in fact cigarettes, the that pricethe poor the of simulation simulation cigarettes, households allowsallows the likely forfor simulation allow patterns differences differences across in To To the in deciles, theassess assess responses To responses explain the the assess whether distributional distributional acrossacrossthe distributional consumption consumption a impact impact price of of reform deciles an impact an deciles increase increase to ofwill to an reflect be in in more the increase reflect the the thefact price price inregressive, fact the that of of that cigarettes, cigarettes, price poor poor more of households households the the cigarettes, simulation simulation likelythe likely allows f allows simulation To assess ease in the price of cigarettes, the simulation the allows distributional for differences impact in the of an increase responses have different price elasticities relative to households with more resources. The different elasticities, in across the price consumption of cigarettes, deciles the to simulation reflect the allows fact that for poor households neutral, havehave or different more different differences differences priceprogressive. price in differences elasticities inthe elasticities theresponses responses inpatterns the across across responses consumption consumption across consumption deciles deciles to to reflect reflect deciles to the the factthat fact reflect that the poor poor fact households households that poor househ lik lik like on deciles to reflect the fact that poordifferences households combined in with the likelythe responses have initial across relative different consumption consumption price relative to to households households elasticities deciles across relative towith deciles, with reflect to more households explainmore the resources. fact resources. whether that with The poor amore The price different different households resources. reform elasticities, elasticities, likely willThe be different elastic combined combined with have have withthe the different different initialhave initial price price different consumption consumption elasticities elasticities price patterns patterns relative relative elasticitiesacross across to to households households relative deciles, deciles, to with with households explain explain more more whether whether with aresources. resources. more a price price The The resources. reform reform different different will The elasticit elasticiti bedifferent have useholds with more resources. The different more different elasticities, regressive, price more elasticities combined neutral, relative with or more theto households initial progressive. consumption with more patterns resources. across The deciles, different explain elasticities, whether be awill price reform w The more loss more of regressive, regressive, real consumption combined combined more more with combined neutral,with neutral, the the or arisinginitial initial with or more more thefrom consumption consumption initial progressive. the progressive. price consumption increase patterns patterns across across patterns in a product deciles, deciles, across explain explain i is obtained deciles, whether whether explain aa price whether reform price reform a pricewill will refo reform with combined ns across deciles, explain whether a price will be the initialmoreconsumption regressive, more patterns neutral, across or more deciles, explain whether a price reform will be progressive. as follows: regressive, more regressive, more more more neutral, more regressive, neutral, more or more or moreor neutral, progressive. progressive. more progressive. essive. more The loss regressive, more neutral,arising of real consumption or more from progressive. the price increase in a product i is obtained as follows: The The loss lossof of real real consumption consumption The loss of real consumption arisingarising from from thethe price price arising increase increase from in a the a product inproduct price increase i isi obtained is in obtained a product as follows: as follows: i is obtained as follows The loss The loss The of of !" real real loss + consumption consumption of∆" real ' ∗ ∆)* arising consumption, arising from from arising the the price price from (A1) theincrease increase price in product in aaproduct increase in a ii isis obtained product obtained as follows: as i is obtained follows: as f The e price increase in a product i is obtained asloss of real consumption arising follows: #$ !"!" from#$ the )*,,price ∆)*∆)increase , * , in a product (A1) i is (A1) obtained as follows: #$ + + ∆" #$ ∆" #$ '#$ ∗' ∗ ∆) )*,,!" )*,,#$ + ∆"#$ ' ∆)∗ *) ∆) * , ∆) (A1) )* where "#$ is the share of product i in total ∆)*!" household !" #$ + #$ +∆" ∆" expenditure !" #$ #$ ''∗ #$ + ) ∗ ∆" ,,*,, * for , ∗ a household '(A1) * in a (A1) (A1) decile j; (A1) ∆.# is , (A1) !"#$ + ∆" #$ ' ∗ ) , ) *,, #$ *,, )*,, where where where " " is is the is the theshare share whereshare of of product of product product is the i in share i in total inoftotal totalhousehold product household household i in expenditure totalexpenditure expenditure household for for a a household for household expenditure a household in for ina a a deciledecile in price household aj; ∆. j; ∆. # is# is in a decile j; ∆ CCO ,, the price increase; #$ #$ and ∆"#$ " is #$ the change in the *,, consumption of the good that depends on the where where " " #$ is where is the the share share is of the of product product share of i i in in product total total i household household in total expenditure expenditure household for for expenditure aa household householdfor a in in household aa decile decile in j;∆. j; ∆. usehold expenditure for a household in wherethe decilethe "#$ elasticity a decileprice ofprice j;isthe ∆. increase; the# increase; share is the product. the and of priceand price 10 ∆" #$ product ∆" increase; is " #$ #$i #$ increase; is the in thechange total and and change in household ∆" #$ the in is the the the consumption consumption expenditure change change in in theof for the of the a thegood household consumption consumption good that that independs of adepends ofdecile the thegood onj; good the on ∆. that # price the is price depends onathe dec# elasticity elasticity of of thethethe the price price product. product. the increase; increase; 10 10 price and and increase; ∆" ∆" and isis the the change change ∆"#$ is 10 in in the change the the consumption consumption of of the the good good that that depends depends on on the the pr pr pri the consumption of the good that depends the price that on increase; the depends price on and elasticity the ∆" #$ is price of the the elasticity product. change of in #$ 10 the #$ consumption theproduct. of thein goodthe consumption that depends on of the the good price that depends o Change elasticity elasticity of the of 10 elasticity the product. product. 10 10 of the product.10 elasticity in of tobacco the product. expenditure Change Change in in tobacco tobacco Change expenditure expenditure in tobacco expenditure CHANGE To estimate IN TOBACCO Change Change the variation in in Change in tobacco tobacco cigarette EXPENDITURE expenditure expenditure tobacco in consumption expenditureafter the price increase, the model considers the Change in tobacco expenditure To To estimate estimate the thevariation To variation in cigarette in cigarette consumption consumption (after)after the theprice price increase, increase, the themodel model ofconsiders considers thethe consider change in prices To estimate the (∆. # ),estimate variation the in tobacco the variation cigarette price elasticity consumptionin cigarette /$afterconsumption for decile the price j, after and increase, the the price share the increase, model cigarette the model change change in in prices To estimate To prices estimate ( ( ), To ), the the thethe estimate variation variation tobaccotobaccothe in in price price variation cigarette cigarette elasticity elasticity in consumption consumption ( cigarette ( ) ) for for decile consumption after after decile j, the the j, and price price and after the the increase, increase, theshare share price of the the of increase, model model cigarette cigarette considers considers t To mption after the price increase, the model estimate considers expenditure the inthe variation period ∆. change 0 (∆. #in"#$ in # ). prices cigarette The change (∆.# ), the consumption tobacco in expenditure after the / price $ / for elasticity price $ increase, each (/ household $ ) for the model indecileconsiders each j,decile and the is the the share model of ciga co considers the change change change inin change prices prices prices in ( (( ∆.∆. ).prices ), ),), the the the tobacco tobacco tobacco (elasticity # ),in the price price price tobacco elasticity elasticity elasticity price (( / / )) ) for for decile decile decile j j ,,, andand and the the share share j, decile of of cigare cigare elasticity (/$ ) for decile j, and the share expenditure change presentedexpenditure of in asprices cigarettea sharein(∆. period in period the # ), total expenditure of 0 0 (tobacco "#$ (" ). in expenditure #$ The period The ## change price 0 and ∆. change(" in #$ ). The averaged (expenditure expenditure / $ ) change by for decile decile tofor for in j,elasticity each each and expenditure quantify $$household the the (/ householdfor) for share $overall inof each decile ineach each decile cigarette household impact, andisin as the share iseach dec o presented the presented share of as a asexpenditure expenditure share cigarettea share of total of expenditure expenditure in expenditure in total period period expenditure in 00 (( " period period " and ). ).and 00 The The averaged ( ( " change change averaged ). ). The The by inin decile bychange change expenditure expenditure decile to in in quantify to quantify expenditure expenditure for for the each each theoverall for household household overall each each impact, impact, in in householdas each each as decile indecile eac in expenditure for each household follows: expenditure in each decile is presented in period 0 ("#$ ). asThe a share change of total #$#$ expenditure in expenditure #$ andeach for averaged household by decile in each to quantify decile is the overall impa follows: follows: presented presented as aa share as presented share as of ofshare a total of total expenditure expenditure total and expenditure and averaged averaged by decile by decile toto quantify quantify the overall the overall impact, theimpact, presented d averaged by decile to quantify the overall household as ain impact, share as each follows: of total decile expenditure is presented and as a averaged share of by total decile toand expenditure quantify averaged and the by overall averaged decile impact, by to quantify as overall follows:follows: follows: follows: ABCD decile to quantify ∆ Expenditure the overall :,;=((1 impact, + ∆P)!1 as + ε; ∗ ∆P'-1) follows: EFGHI JKLJMN:GOPJ ABCD ABCD (A2) ∆ Expenditure ∆ Expenditure =( :,; ∆ =( ( 1 ( + 1 + ∆P :,; Expenditure:,;=( )! ∆P 1 )! + 1 ε+ ∗ ; (;ε ∆P'∗ ∆P' 1 + ∆P)! -1) -1)  1 + ε EFGHI JKLJMN:GOPJ ∗ ∆P' EFGHI JKLJMN:GOPJ CD -1)  AA ABCD (A2) (A2) (A2) ; BCD ABCD ∆ ∆ Expenditure Expenditure ∆ Expenditure :,; :,; =( =( ((1 1 ++ ∆P ∆P=( )! )! ( 111 +++ ε ∆PAε; ; ∗ ∗ )! BCD ∆P' ∆P' 1 + -1) -1) ε  ∗ ∆P' CD -1)  BCD CD EFGHI JKLJMN:GOPJ ABCD CD (A2) (A2 (A2) ε ∗ ∆P'-1) ; (A2) Medical expenses ∆ Expenditure :,; =( ( 1 + ∆P )! 1 + ε ; ∗ ∆P' -1) :,;  EFGHI JKLJMN:GOPJ ; EFGHI JKLJMN:GOPJ EFGHI JKLJMN:GOPJ (A2) CD EFGHI JKLJMN:GOPJ CD CD EFGHI JKLJMN:GOPJCD Medical Medical expenses expensesMedical expenses CD expenses MedicalMedical Medical expenses expenses The change Medical in medical expenses from expenses tobacco-related diseases is estimated using equation (A3), where the The The cost ofchange change treatment medical in medical in The expenses expenses change of tobacco-related from in medicalfrom tobacco-related tobacco-related expenses diseases for from income diseases diseases tobacco-related decileis Q is isestimated estimated diseases obtained using isequation using from equation estimated (A3), administrative (A3), using where where (A3), w equation the cost the cost of of Thechange change The treatment treatment of of The in in medical medical tobacco-related tobacco-related change in expenses expenses diseases diseases medical from from expensesfor tobacco-related tobacco-related income for income from decile decile diseases Qdiseases tobacco-related isQ obtained obtained is isestimated is estimated diseasesfrom from is using using administrative administrative estimated equation equation (A3),whe (A3), wh data.change The -related diseases is estimated using equation The (A3), in medical cost where the cost of treatment expenses of tobacco-related from of tobacco-related tobacco-related medical expenses is diseases distributed diseases is estimated across for income using equation income decile decile QQ is obtained (A3), according where to fromequation using administr ( data. data. The The cost the the of cost cost cost tobacco-related of of of the treatment treatment tobacco-related cost of of of medical tobacco-related tobacco-related medical treatment expenses expenses of is distributed is diseases diseases tobacco-related distributed for for income income across across diseases income for decile decile income income QQ decile is is obtained decileobtained decile Q according Q Q is from accordingfrom obtainedto administrat administrat administrati to from adm ases for income decile Q is obtained from cost of treatment the administrative share of households data. The consume of tobacco-related that cost of tobacco-related diseases in tobacco for medical income decile expenses decile Q. Equation is distributed obtained Q is (A3) shows fromthe across income administrative income gainsdecile Q accordi the the share share of of households data. data. The The households cost cost that of of that tobacco-related tobacco-related consume consume tobacco tobacco medical medical in in decile expenses expenses decile Q. Equation is distributed is . Equation distributed (A3) (A3) across across income income decile decile according QQ according associated data. penses is distributed across income decile Q The with cost according ofthe to decrease data. the share tobacco-relatedin ofThe medical cost households medical of expensestobacco-related expensesthat because consume is of reduced distributedmedicalQacross tobacco expenses tobacco in decile income is shows Q.shows distributed consumption Equation decile Qthe the inincome income across the (A3) according long showsgains income to gains thedecile income Q ac associated term. associated with the the with the share share the of decrease decrease the associated of households households in share in with the medical medical of that that decrease consume consume expenses expenses households in because that tobacco tobacco because consumeof reduced of in reduced in decile tobaccodecile tobaccoinQQ. tobacco. Equation Equationconsumption consumption decile Q . (A3) (A3) Equation shows shows in in the the long (A3) the the long showsincome income the ga ga gai inc cco in decile Q. Equation (A3) shows the 10 share the income of gains households that consume tobacco in medical decile .expenses Equation Q(2006). because (A3) shows of reduced tobacco the income consumption in the gains For a term. term. detailed associated associated discussion of the with with methodology, term. associated the the decrease decrease see Coady with et in in al. the decrease medical medical (2006); Kpodar in ofexpenses expenses medical because because expenses of of reduced becausereduced tobacco tobacco of reduced consumption consumption tobacco in in the the lo lo long consumption THE associated ses because of reduced tobacco consumption with in the longthe decrease in medical expenses because reduced tobacco consumption in the term. term. term. term. 10 For a detailed discussion of the methodology, see Coady et al. (2006); Kpodar (2006). (2006); Kpodar (2006). 46 10 For 10 For a detailed a detailed discussion discussion 10 For aof of the the methodology, methodology, detailed discussion ofsee see the Coady Coady et al. et (2006); methodology, see Kpodar al. (2006); Kpodar Coady (2006). et al.(2006). 10 detailed Foraadetailed 10 For 10 For a detailed discussion of the For discussion discussion ofthe of methodology, themethodology, seeCoady see Coady etal. et (2006); al.(2006); Kpodar(2006). Kpodar (2006). a detailed discussion methodology, see Coady et al. (2006); Kpodar (2006). et al. (2006); Kpodar (2006). of the methodology, see Coady 10 oady et al. (2006); Kpodar (2006). 25 25 25 25 25 uct i in total household expenditure for a household in a decile j; ∆.# is e product. A. 10 Model and Methodology the change in the consumption of the good that depends on the price Annexes To estimate the variation in cigarette consumption after the price increase, the model considers the bacco expenditureThis section describes the partial equilibrium approach used to simulate the impact on consumption A. Model and Methodology change in prices (∆.# ), the tobacco price elasticity (/$ ) for decile j, and the share of cigarette ure because of an increase in the price expenditure of cigarettes. in period This 0 approach ("#$ ). The used to in is change evaluate expenditure the first-order for each household in each decile is he variation in cigarette effects of consumption a Distributional change in prices. after It the relies price mainly increase, on household the model expenditureconsiders the patterns. The focus is on the This section The describes Effects the partial presented of Tobacco equilibrium as Taxation: a share approachThe of Evidence total used ofexpenditure White to simulate and Clove and averaged Cigarettes the impact on consumption in by decile Indonesia to quantify the overall impact, as ces (∆. garette the # ),impacts consumption tobacco of a rise after pricein the the elasticity price price of increase, (/ $ ) for cigarettes follows: the model decilebecause jconsiders , and thisthe has theshare been the of cigarette focus of the potential reform of n because periodprice0 ("#$ ). of The an changeincrease in the price of for in expenditure cigarettes. each household This approach in each is useddecile toisevaluate the first-order obacco tobacco elasticity taxes. (/$ ) for decile j, and the share of cigarette effects of a change in prices. It relies mainlyto on household the expenditure patterns. as The focus is on the . share of The change total in expenditure expenditure and for averaged each household by decile in quantify each decile ∆ Expenditure overall impact, is ABCD impacts of a rise in the price of cigarettes because this has been :,;=((the 1 + focus ∆P)!1of + the ∆P'-1)EFGHI JKLJMN:GOPJ ε; ∗ potential reform of CD (A2) xpenditure To and assess averaged by decile toimpact the distributional quantify of the an increaseoverall impact, in the price as of cigarettes, the simulation allows for tobacco taxes. differences in the responses across Medical consumptionexpenses deciles to reflect the fact that poor households likely =(MEDICAL EXPENSES ABCD ∆ Expenditure have (1 + ∆P :,;different )!1 + price ε; ∗ ∆P'-1) elasticities EFGHI JKLJMN:GOPJ relative to households with more (A2) resources. The different elasticities, To assess the distributional ABCDimpact of an increase in the price of cigarettes, the simulation allows for CD combined 1 + ε; ∗ ∆P'-1)EFGHI JKLJMN:GOPJ ( 1 + ∆Pdifferences ( )! The withchange the in initial medical consumption The expenses patterns change in medical from tobacco-related across (A2) expenses from deciles, explain diseases whether tobacco-related is estimated a price reform diseases using iswill equa- be estimated using equation (A3), where nses in the responses across CD consumption deciles to reflect the fact that poor households likely more regressive, tion (A3), more where neutral, the or more the cost cost of progressive. of treatment treatment of of tobacco-related tobacco-related diseases diseases for for income income decile decile Q is obtained from administrative have different price elasticities relative to households with more resources. The different elasticities, obtained is with from data. administrative The cost of data. across tobacco-related The cost of tobacco-related medical expenses medical is distributed expenses across is income decile Q according to combined medical expenses the initial consumption patterns deciles, explain whether a price reform will be The lossfrom of real tobacco-related consumption diseases the share arising is from estimated ofthe households price using increase equation that in consume (A3), a product where tobacco i is obtained in decile Q. Equation (A3) shows the income gains as follows: atment ofmore regressive, tobacco-related distributed more diseases neutral, across for or more income income decile progressive. decile Q is obtained according from toin the administrative shareexpenses of households that consume s from tobacco-related diseases is estimated associated using with equation the ∆)* (A3), where decrease medical because of reduced tobacco consumption in the long of tobacco-related co-related diseasestobacco medical for income expenses in deciledecile Q is is !" distributed term. . Equation obtained #$ + ∆" ' #$(A3)across from∗ , shows income administrative decile Q according(A1) the income gains associated with the decrease to )*,, ouseholds ed medicalThe thatloss expenses consume of real tobacco consumption is distributed Annexes in decile across arising income Equation Q. from the price decile Q(A3) increase according showsto the in a income productgains i is obtained as follows: h the decrease where in " in #$ is the medical medical share expenses expenses product ofbecause because of i in total reduced of reduced household tobacco tobacco expenditure consumption consumption for in thea household long in thein long a decile term. j; ∆.# is consume tobacco in decile Q. Equation !" (A3) + ∆"shows ' ∗ ∆)* income gains the , (A1) the price medical expenses because increase; of and A. reduced ∆" Model #$ is the and 10 #$ For tobacco consumption change a detailed Methodology #$ in ) the discussion *,, consumptionof in the long the of methodology, the good see VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJWB that Coady depends et al. (2006); on the Kpodar price (2006). elasticity of the∆ Medical expenditure product. 10 :,;=(!1 + ε; ∗ ∆P'-1) EFGHI JKLJMN:GOPJCD (A3) where "#$ is the share of product i in total household expenditure for a household in a decile j; ∆.# is This section describes the partial equilibrium approach used to simulate the impact on consumption 25 discussion logy the of theprice Change increase; methodology, in tobacco seeand Coady ∆"#$ expenditure is the et al. (2006); changeKpodar in(2006). the consumption of the good that depends on the price AA reduction reduction because in tobacco in tobacco ofconsumption anconsumption increase in inthe thein price longlong the of run cigarettes. would run would beThis strongly be approach strongly related is to related used to aevaluate a decline to in the the first-order elasticity of thodology, see Coady the product. et al. (2006); Kpodar of 10 (2006). incidence of tobacco-related effects a change diseases. in prices. The It model reliesassumes mainly on that the health household effects of tobacco-related expenditure patterns. The focus is on the e partial equilibrium To estimate decline diseases approach the in will the variation incidence used immediately intocigarette simulate of tobacco-related diminish the impact consumption with theon diseases. consumption after reduction the price The in model increase, tobacco 25 assumes the modelthat consumption. considersthe Although health the this Change in tobacco impacts expenditure of a rise in the price of cigarettes because this has been the focus of the potential reform of n the price change in of assumption cigarettes. prices effects ofis This (∆. approach implausible tobacco-related # ), the tobacco is in the used diseasesprice to short term elasticity evaluate willbecause the immediately (/ ) for first-order $changes decile diminish j in the effects, and with the the share of tobacco-related of reduction in diseases cigarette ∆ tobacco taxes. =( -1) 25VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJWB (A3) ices. It relies expenditure mainly take onin time toperiod Medical expenditure household materialize, ("#$it 0 expenditure The patterns. ). provides change :,; an +in !1upper-bound ε ; expenditure The ∗ ∆P' focus is on estimate forthe each of the household effects of in each tax increases. decile is To estimate tobacco the variationconsumption. in cigarette Although consumption this assumption after the price is implausibleincrease, the EFGHI JKLJMN:GOPJ in the CD short model term because considers the presented as a share of total price of cigarettes because this has been the focus of the potential reform of expenditure and averaged by decile to quantify the overall impact, as change follows: inchanges Increase prices ( ∆.To in the in years ), # of the assess effects working tobacco the distributional of life price tobacco-related diseases elasticity impact ( / of$ ) anfor decile increase take time j in, and the to the price materialize, share of of cigarettes, it cigarette provides the simulation an allows for expenditure A reduction in period in tobacco ∆ differences 0 (" ).consumption The change in in the long expenditure run would for each be strongly VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW household relatedin each to a B decline in the decile is Medical expenditure #$ in the responses :,;=( !1 +across ε; ∗ ∆P' -1) consumption deciles to reflect the fact that (A3) VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW poor households B likely upper-bound incidence estimate of tobacco-related the ∆ ofdiseases. effects The of Medical expenditure tax model ∆ increases. assumes Medical expenditure :,;=( that!1 +the ε; health ∗=( EFGHI JKLJMN:GOPJ ∆P'!1 -1) +  effects εCD;∗ ∆P' ofimpact, tobacco-related -1) VFWG EPJHG.EFYHZZF [JIHGJN \:WJHWJW  The (A3) B (A3) presented ZF [JIHGJN \:WJHWJW The as model a share of total have estimates different the expenditure impact price on and elasticities income averaged relative arising by to from decile households the ABCDto quantify increase with :,; in the more working overall resources. years EFGHI JKLJMN:GOPJ (equation as different A5). CD EFGHI JKLJMN:GOPJelasticities, nal impact LJMN:GOPJCD of follows: an diseases B (A3) increase will in immediately the price of diminish cigarettes, ∆ Expenditure:,;=( 1 + ∆P 1 + ε; ∗ ∆P'-1)EFGHI JKLJMN:GOPJ ( )! thewith the simulation reductionallows in for tobacco consumption. (A2) Although this CD To Aestimate assumption reduction INCREASE isthe increase combined in tobacco implausible IN YEARS in within working the consumption the OF initial short years, WORKING consumption term in the the because years longLIFE of patterns run changes life wouldlost in (YLL across the ) from CD deciles, beeffects strongly tobacco-related of explain related tobacco-related whether to a decline diseases a price diseases in the reform will be ses across consumption deciles to reflect A the fact reduction that in poortobacco households consumption likely in the long run would be strongly related to a decline are take distributed time to of incidence moreacross tobacco-related materialize, deciles regressive, it provides ( Q ) more proportionally diseases.A reduction neutral, an upper-bound or to The model estimate in more tobacco the number progressive. assumes of consumption of that the households the health effects in the ofeffectsthat long tax increases.consume run would of tobacco-related tobacco be strongly related toin a the decline in the icities relative rongly Medical related to households a expenses The (equation decline model A4). in with the estimates more Subsequently, resources. incidence the impact the of The income different tobacco-related incidence on income lost of is elasticities, tobacco-related arising estimated diseases. ABCD The model assumes that the health effects of tobacco-related from as the the diseases.increase average The in income model working per assumes years household that the in health effects of tobacco-related diseases will ∆ Expenditure:,;=( immediately ( diminish 1 + ∆P 1 + ε; ∗ ∆P'-1) )! with the reduction in tobacco consumption. (A2) Although this health effects of consumption tobacco-related patterns decile across Q. Overall, deciles, the model diseases explain whether anticipates willdiseasesimmediately aterm that price will income reform diminish will EFGHI JKLJMN:GOPJ immediately will be with increase diminish as CD the reduction the with number the of yearsin tobacco reduction lost in consumption. because tobacco Although Although consumption. this this assumption Increase (equation in years is A5). The implausible of Tolossworking estimate of real in life the short the consumption increase in because arising working from changes years, the price in the the years increase effects of in of life a tobacco-related lost product (in YLL ) ithe is from obtained diseases as follows: co consumption. eutral, The change or more progressive. of Although prematurein medical this deaths expensesfrom assumption from tobacco is tobacco-related implausible assumption consumption in diseases the implausible isdecline. short is estimated term in of the because using short changes equation term because (A3), whereeffects of tobacco-related changes in the effects of tobacco-related diseases diseases take time to materialize, it provides an upper-bound estimate the effects of tax increases. effects ofMedical tobacco-related the cost expenses tobacco-related of treatment diseases diseases of tobacco-related take time to are distributeddiseases take materialize, time for toacross it provides income materialize, deciles decile ∆)it *an Q upper-bound (provides )isproportionally obtained an estimate from upper-bound to the ofestimate administrative the effects number of theof tax increases. effects of tax increases. The model estimates the impact on income !" #$ + ∆" arising #$ ' ∗the from ,increase in working (A1) years (equation A5). effects ofdata.tax increases. The cost Increase of intobacco-related years of working medicallifetobacco expenses is distributed ) *,,across income decile Q according (A4) is to ption arising The from To ofthe households estimateprice increase ]^_`Qab cde_f the that increase in consume product aIncrease # = in (chh working i# is obtained ∗years, (equation iℎe_d ^k il^`d_f the as years follows:of A4).life Subsequently, # )/n^.opeqQ^a lost (YLL ) from the income tobacco-related (A3),lost esti- diseases the change in medical expenses from tobacco-related diseases is estimated using # share of households that consume in years Increase tobacco of in working in years decile life of working Equation Q. number life(A3) shows equation the income where gains are distributed where across " deciles #$ is the(Q share ) of product proportionally ito inthe total household of expenditure households thatfor a household consume tobacco in a decile j; ∆.# is !"#$ the + ∆"cost' of associated mated ∗ treatment ∆) with * , as the the of average decrease tobacco-related in income medical(A1) per household diseases expenses for because income in of decile decile reduced Overall, Q .tobacco is obtained the from consumption model anticipates administrative in the long The (equation)*,,model the price increase; estimates the(!1 impact + and on ∆" income is  the change arising from in B the the consumption increase in B working of the (A5) good years that depends (equation in on years A5). the price ofA4). Subsequently, the ε; income lost is estimated as the average income per household #$ uFPv:Mw xJHPW ∗EFGHI yKLJMN:GOPJ data. term. The cost ∆ Income tobacco-related : =The model medical ∗ ∆P' estimates expenses -1) #$ the is impact distributed on income across income arising from decile Q the accordingincrease toin increase working (equation A5). that To estimate decile Q income . Overall, will the elasticity the increase increase of the model in as the working product. anticipates The number 10 model years, that of the income years estimates yearsEFGHI JKLJMN:GOPJ will lost of thelife increase because impact lostas (YLL B on theof income premature ) from tobacco-related number arising of deaths years from lost from the diseases in working years (equation A5). because eproduct in working the years i in share total The (equation of household totalhouseholds income A5). expenditure that gains To consume in for estimate each ( a tobacco household income the To estimate increase inin group are decile a decile the in Qworking . j estimatedEquation ; increase ∆. is # in years, by (A3) adding working the years shows the years, of the results the life income lost the of consume years ( YLL gains ofincrease ) life lost from in tobacco-related (YLL) from tobacco-related diseases diseases are premature of tobacco distributed deaths consumption across from deciles tobacco decline. proportionally Q) consumption to decline. the number of households that tobacco L ) from associated tobacco-related tobacco with thediseases in the expenditures, decrease inthe are medical distributed reduction expenses across medical in distributedbecause deciles of treatments, ( reduced Q ) proportionally and tobacco to consumption the number in the of households long that consume tobacco "#$ is the change (equation consumption Change A4). in Subsequently, the of tobacco good the that are expenditure income depends lost on the price isacross estimated deciles asthe ( Q) gain the in working proportionally average income toyears the (equation per number of household 1 inhouseholds that consume tobacco ouseholds 0 term. that in consume tobacco 10 For a detailed discussion of the methodology, the main text). the model (equation A4). see (equation Subsequently, Coady al. (2006); etA4). the Kpodar Subsequently, income (2006). the lost is estimated income lost is as the average estimated as theincomeaverage per household income in per household in decile . Overall, Q]^_`Qab cde_f # = (chh anticipates that # ∗ iℎe_d ^k il^`d_f income will increase # )/n^.opeqQ^a as the number of years (A4) lost because verage income per household To in estimate decile the Q . Overall, variation decile the inQ.cigarette model Overall,decline. anticipates the model anticipates consumption that income after the that # will price increase income increase, will asthe the modelas increase number of the number considers years lost because the of years lost because of premature deaths from tobacco consumption the number of B. penditure years lost because Tobacco Price Elasticity of prematureby Decile of deaths from premature tobacco deaths from consumption tobacco consumption decline. decline. 10 For a detailed discussion change of the in prices methodology, (∆. see # ), Coady the et tobacco al. (2006);price uFPv:Mw xJHPW Kpodar elasticity (2006). (/$ ) Bfor decile(A5) B ∗EFGHI yKLJMN:GOPJ j, and the 25 share of cigarette ∆ Income : = (!1 + ε; ∗ ∆P'-1) expenditure in ]^_`Qab cde_f # =period (chh # 0∗ ( " ). iℎe_d ^k il^`d_f #$ The change EFGHI JKLJMN:GOPJ # in expenditure )/n^.opeqQ^a B # for each household(A4) in each decile(A4) is n in cigarette consumption Let #{ be after the defined aspricethe increase, ]^_`Qab cde_f the group model # = considers (chh # ∗ iℎe_d ^k il^`d_f the # )/n^.opeqQ^a asaverage quantity smoked per day andby individual in income decile d; n # the Thez total income presented gains in each a share income of total ]^_`Qab cde_f are expenditure estimated by = #averaged (chh# by adding ∗the Q iℎe_d ^k il^`d_f results decile to of the quantify # )/n^.opeqQ^a increase the in impact, overall # as (A4) Q^a the # tobacco price average (A4) elasticity price tobacco expenditures, per (/ ) for cigarette decile (unit j, and value the reduction in medicaluFPv:Mw xJHPW the of share tobacco of use); treatments, cigarette | the consumption and the gain in working decile of individual years (equation 1 25 Q ; and follows:$ ∆ Income # B ∗EFGHI yKLJMN:GOPJ for: = (!1 -1) (A5) of B The total income gains in each + εincome; ∗ ∆P'  each are group estimated by uFPv:Mw xJHPWB ∗EFGHI yKLJMN:GOPJ ("#$ ). The change }#{ in the the in main individual expenditure text). characteristics. each Then, household the ∆ Income smoking in : ∆= Income (!1 intensity decile + ε; ∗ EFGHI JKLJMN:GOPJis∆P' equation -1)  B adding is written the as results follows: uFPv:Mw xJHPW the B B ∗EFGHI yKLJMN:GOPJ (A5) N:GOPJB : = ( !1 + ε ; ∗ ∆P' -1)  EFGHI JKLJMN:GOPJB B (A5) total expenditure (A5)The and total averaged income increase in tobacco expenditures, by gains decile into each quantify income the the group overall reduction are estimated impact, in medicalas by adding treatments, the results and the of the increase EFGHI JKLJMN:GOPJ gain in in B The total income The total gains =(income in ∆P each income group are estimated A are by adding the results of the increase in increase in +gains +in each income group estimated by adding the results of the BCD B. tobacco Tobacco expenditures, Price Elasticity the ln ∆ Expenditure zreduction #{ = by Decile~ + in ~Ä medical pan :,; (1 ∗+ | )! treatments,~1 Å }#{ ;in ε+ ∗ and ∆P' Ç#{ the -1)gain in (A6) working years (equation (A2) 1 ding the results ofworking the increase years in (equation tobacco 1 in expenditures, the main tobacco text). the #reduction expenditures, the medical reduction treatments, EFGHI JKLJMN:GOPJ in medical and CD the gain in working years (equation 1 treatments, and the gain in working years (equation 1 in the main text). gain in working years (equation 1 in the main in text). the main text). Thez Let empirical be defined analysis Medical ofA as expenses the equation average quantity (7) assumes smoked a log-log per day relationship by individual among Q in income intensity, smoking decile d; n price, the 1 +#{ BCD ure:,;=((1 + ∆P)!B. ε; ∗ ∆P'-1) Tobacco EFGHI JKLJMN:GOPJ Price (A2) and average income. price paz per is Elasticity cigaretteobserved (unit by if Decile and value only of if tobacco the individual use); the from a consumption given decile decile É is of a current individual smoker. ; and B. Tobacco #{ The change Price B.Elasticity Tobacco in medical CD B. Price Tobacco expenses Decile byElasticity from Priceby | # Decile by Decile Elasticity tobacco-related diseases is estimated Q using equation (A3), where }#{ the individual characteristics. Then, the smoking intensity equation is written as follows: Let zmodels Several #{ be be the were defined defined cost tested as ofas thethebefore treatmentaverage average deciding of quantity quantity the onsmoked tobacco-related final smoked per elasticities diseases day day per byfor that individual by income weredecile individual used Q in income in Q the in is model. income obtaineddecile from Table d; n the administrative Let z#{ beLet defined z as defined be the average as thequantity average smokedquantity per day by per smoked individual day by Q in income individual Q decile in income the d; n the d; n decile considers B1average price the data.perThe maincigarette equation cost ln z (unit of = without value tobacco-related ~ + of ~#{ controls, tobacco pan ∗medical | +adding use); ~ | expenses } # demographic the + consumption Ç is distributed controls (A6) decile across and of then individual income adding decile Q ; and Q according to vidual Q in income penses from tobacco-related decile decile demographic d; n the the diseases controls average is and average estimated year #{ price price fixed per per average using effects. Äcigarette cigarette equation price per#(unit (unit (A3), value cigaretteÅvalue where #{ of of (unit tobacco tobacco #{ value of use); use); | tobaccoas # the the consumption use);consump- decile of individual |# the consumption decile of individual Q; and Q ; and }#{ the individual the share characteristics. of households Then, that the smoking consume intensity tobacco inequation decile Q. is written Equation (A3) follows: showsisthe income gains mption decile ofdiseasesindividual Q;income and }#{ the individual the characteristics. individual Then, characteristics. the smoking Then, the intensity smoking equation intensity written equation asis follows: written tobacco-related ion is written tion decile The empiricalassociated as follows: for of individual analysis of decile Q equation with ;is theand obtained} decrease #{ the (7) assumes from in individual administrative a log-log medical characteristics. expenses relationship because Then, among of reduced the smoking smoking tobacco intensity, inten- consumptionprice, in the long as follows: o-related medical and expenses income. is distributed term. is observed across ln z income = if and only ~ + decile ~ pan Q according ifÄthe individual∗ | + ~ to } from + Ç a given #{ ∗decile (A6) sity equation paz is #{ written as follows:#{  ln z # =Å ~ + ln #{ ~ z#{Ä= pan ~ + |#~+ ~Åis É }#{ + Ç#{ smoker. a current (A6) (A6) that consume tobacco in decile Q. Equation (A3) shows the income Ä pan ∗ |# + ~Å }#{ + Ç#{ #{ gains (A6) ase in medical expenses The empirical because of reduced analysis of beforetobacco equation consumption (7) assumes a in the log-log long relationship among smoking intensity, price, Several models were tested The empirical deciding analysis on the final of analysis equation elasticities (7) assumes that were a log-log used in the relationship model. Table among smoking intensity, price, B1and income. 10 For a detailed discussion The of the empirical methodology, see Coady of equation et al.a (7) (2006); assumes Kpodar decile (2006). Éais log-log smoker. among smoking intensity, price, relationship considers the paz #{ is observed main equation if and income. and without only if the controls, individual adding from demographic given controls and a current then adding p among smoking intensity, price, and paz #{ is observed income. paz#{ isifobserved and only if if andthe onlyindividual if the from individual a given from decilea given É is a current decile É issmoker. a current smoker. demographic controls and year fixed effects. iven decile É is a Several current models smoker.were tested before deciding on the final elasticities that were used in the model. Table Several models Several were modelstestedwere before tested deciding before on the final deciding on elasticities the final that were used that in theused model. 25 Table the methodology, see B1 Coady 47 model.considers // et al. (2006); Annexes the Kpodarmain equation (2006). without controls, adding demographic controls and elasticities then adding were in the model. Table hat were used in the Table B1 considers the B1 considers main equation without controls, adding demographic controls and then adding demographic controls and year fixed effects. the main equation without controls, adding demographic controls and then adding 26 raphic controls and then adding demographic controls and demographic year fixed controls and effects.year fixed effects. 25 ∆ Income: = (!1 + ε; ∗ ∆P'-1) B (A5) the results of the The total income gains in each income group are estimated by adding EFGHI JKLJMN:GOPJ uFPv:Mw xJHPWB ∗EFGHI yKLJMN:GOPJB B increase in + ε; ∗ ∆P'-1) EFGHI JKLJMN:GOPJB (A5) The tobacco expenditures, thetotal income reduction in gains in each medical income treatments, group and gainestimated the are byyears in working adding the results (equation 1 of the increa in the main text). tobacco expenditures, the reduction in medical treatments, and the gain in working years (equat income group are estimated by adding the results of the increase in in the main text). ction in medical treatments, and the gain in working years (equation 1 B. Tobacco Price Elasticity by Decile B. Tobacco Price Elasticity by Decile Decile Let z#{ be defined as the average quantity smoked per day by individual Q in income decile d; n the Let z#{ average price per cigarette value of as be defined (unit the average tobacco use); | quantity smoked per day by individual Q in income decile d; n # the consumption decile of individual Q ; and ge quantity smoked per day by individual Q inindividual }#{ the income decile average price characteristics. per cigarette (unit value d; n the Then, the smoking intensity equation of tobacco use); consumption |# the as is written follows: decile of individual Q value of tobacco use); |# the consumption decile of individual} the individual #{ Q ; and characteristics. Then, the smoking intensity equation is written as follows: s. Then, the smoking intensity equation is written as follows: ln z#{ = ~ + ~Ä pan ∗ |# + ~Å }#{ + Ç#{ (A6) ln z#{ = ~ + ~Ä pan ∗ |# + ~Å }#{ + Ç#{ (A6) The empirical analysis of equation (7) assumes a log-log relationship among smoking intensity, price, = ~ + ~Ä pan ∗ |# + ~Å }#{ + Ç#{ The(A6) empirical analysis The of equation empirical (7) assumes analysis of equationa log-log (7) assumesrelationship a log-log among smoking relationship among smoking intensity, p and income. paz#{ is observed if and only if the individual from a given decile É is a current smoker. intensity, price, and and income. on (7) assumes a log-log relationship among smoking intensity, price, income. paz #{ is is observed observed if if and and only only ififthe theindividual individual from from a given a decile given É is a current sm Several if and only if the individual from a givendecile models were tested currentsmoker. decile É is a current before smoker. deciding on the final elasticities that were used in the model. Table B1 considers the main Several models equation were tested without before controls, deciding adding on the final demographic elasticities controls and that were then used in the model. T adding demographic B1 considers the main equation without controls, adding demographic controls and then ad ore deciding on the final elasticities that were used Several incontrols the model. models wereand year fixed Table tested beforeeffects. deciding on final elasticities that were used in the the effects. demographic controls and year fixed n without controls, adding demographic controls and then adding model. Table B1 considers the main equation without controls, adding demographic con- fixed effects. trols and then adding demographic controls and year fixed effects. Table B1. Regression Results, by Model 26 PRICE SHOCK NO CONTROLS CONTROLS CONTROLS AND FIXED SCENARIO 26 EFFECTS WHITE CLOVE ALL WHITE CLOVE ALL WHITE CLOVE ALL Log price −0.69 −0.59 −0.60 −0.69 −0.60 −0.60 −0.77 −0.64 −0.65 (0.022) (0.0058) (0.0056) (0.0227) (0.006) (0.0059) (0.0233) (0.0061) (0.006) Decile 2*Log price 0.07 0.05 0.05 0.07 0.05 0.05 0.07 0.05 0.06 (0.0072) (0.0014) (0.0014) (0.0075) (0.0015) (0.0014) (0.0075) (0.0015) (0.0014) Decile 3*Log price 0.09 0.08 0.08 0.10 0.08 0.09 0.10 0.09 0.09 (0.0066) (0.0014) (0.0014) (0.0069) (0.0014) (0.0014) (0.0069) (0.0014) (0.0014) Decile 4*Log price 0.12 0.10 0.10 0.12 0.10 0.10 0.12 0.11 0.11 (0.0062) (0.0014) (0.0013) (0.0065) (0.0014) (0.0014) (0.0065) (0.0014) (0.0014) Decile 5*Log price 0.14 0.11 0.12 0.14 0.12 0.12 0.14 0.12 0.12 (0.0062) (0.0014) (0.0014) (0.0065) (0.0014) (0.0014) (0.0064) (0.0014) (0.0014) Decile 6* Log price 0.15 0.13 0.13 0.16 0.13 0.14 0.16 0.14 0.14 (0.0061) (0.0014) (0.0014) (0.0065) (0.0014) (0.0014) (0.0065) (0.0014) (0.0014) Decile 7* Log price 0.16 0.13 0.14 0.17 0.14 0.14 0.17 0.15 0.15 (0.0061) (0.0014) (0.0014) (0.0064) (0.0015) (0.0014) (0.0064) (0.0015) (0.0014) Decile 8* Log price 0.17 0.14 0.14 0.18 0.15 0.15 0.18 0.16 0.16 (0.0061) (0.0015) (0.0014) (0.0065) (0.0015) (0.0015) (0.0065) (0.0015) (0.0015) Decile 9*Log price 0.18 0.14 0.15 0.19 0.16 0.16 0.20 0.17 0.17 (0.006) (0.0015) (0.0015) (0.0065) (0.0016) (0.0016) (0.0065) (0.0016) (0.0016) Decile 10*Log price 0.19 0.15 0.16 0.21 0.18 0.18 0.22 0.18 0.19 (0.0061) (0.0016) (0.0016) (0.0067) (0.0018) (0.0017) (0.0067) (0.0018) (0.0017) Education HH −0.01 −0.01 −0.01 −0.01 −0.01 −0.02 (0.0017) (0.0005) (0.0005) (0.0018) (0.0005) (0.0005) 48 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Table B1. Regression Results, by Model (cont.) PRICE SHOCK NO CONTROLS CONTROLS CONTROLS AND FIXED SCENARIO EFFECTS WHITE CLOVE ALL WHITE CLOVE ALL WHITE CLOVE ALL Gender HH −0.05 −0.17 −0.18 −0.06 −0.18 −0.19 (0.0256) (0.0085) (0.0082) (0.0259) (0.0085) (0.0082) Age HH 0.01 0.01 0.01 0.01 0.01 0.01 0.0008 0.0002 0.0002 0.0008 0.0002 0.0002 Share ages <5 0.82 0.87 0.87 0.87 0.90 0.90 (0.0624) (0.0171) (0.0166) (0.0623) (0.0171) (0.0166) Share ages 6-17 0.59 0.62 0.62 0.63 0.64 0.65 (0.0468) (0.0118) (0.0115) (0.0466) (0.0118) (0.0115) Share ages 18-49 0.47 0.59 0.62 0.47 0.60 0.63 (0.0426) (0.0103) (0.0101) (0.0426) (0.0103) (0.01) Urban −0.11 −0.09 −0.09 −0.10 −0.09 −0.08 (0.0135) (0.0039) (0.0037) (0.0135) (0.0038) (0.0037) Dummy Year 2016 0.17 0.13 0.14 (0.014) (0.0039) (0.0038) Constant 9.26 8.84 8.88 8.61 8.06 8.05 9.00 8.28 8.28 (0.13) (0.034) (0.0332) (0.142) (0.0381) (0.0371) (0.1446) (0.0385) (0.0375) Observations 31,903 333,848 359,185 30,629 317,559 341,961 30,629 317,559 341,961 R-Squared 0.1 0.1 0.1 0.11 0.13 0.13 0.12 0.13 0.14 Source: Estimates based on Indonesia National Socioeconomic Surveys 2015–16. Note: Demographic controls age, highest level of education, and gender of the household head; share of individuals by age- group in each household; and urban status. The model also controls for year fixed effects. Only 2015–16 were used to estimate elasticities because previous surveys asked about the number of cigarette packets consumed, whereas the most recent surveys ask about quantities of individual cigarettes. HH = household head. 49 // Annexes INDICATOR DECILE 1 DECILE 2 DECILE 3 DECILE 4 DECILE 5 DECILE 6 DECILE 7 DECILE 8 DECILE 9 DECILE 10 Average Household Monthly 161,615 216,099 261,429 309,522 364,929 431,781 515,455 633,208 831,527 1,603,566 ALL 532,922 ing on tobacco products. per capita Expenditure (Rupia) Proportion of exp. on all 5.4 6.4 6.8 6.8 7.1 6.7 6.5 6.0 5.2 2.8 5.3 tobacco products a Proportion of exp. white 0.1 0.2 0.3 0.3 0.4 0.3 0.4 0.4 0.4 0.4 0.3 cigarettes Table C.1. Descriptive Statistics Proportion of exp. clove 4.3 5.4 6.0 6.2 6.5 6.1 6.0 5.5 4.8 2.5 4.8 cigarettes C. Descriptive Statistics, 2010 Households that smoke any 69.1 71.2 70.2 69.3 70.0 66.8 67.7 64.2 60.8 50.2 66.0 tobacco product (%) Households that smoke white Note: Deciles were created using per capita household expenditure. 20.2 2.2 2.4 2.8 3.7 3.2 3.7 4.2 4.4 6.4 3.5 cigarettes (%) Source: Estimates based on Indonesia National Socioeconomic Survey 2010. Households that smoke clove 43.7 53.6 57.2 59.5 60.8 59.9 61.2 58.7 55.9 43.8 55.4 cigarettes (%) Woman-headed households (%) 13.5 13.4 15.0 14.3 15.4 15.8 16.1 15.6 15.4 17.8 15.2 Age, household head 49.6 49.5 49.6 49.1 49.2 48.9 48.6 48.5 48.0 47.2 48.8 Share of kids ages 0-5 in 12.6 10.6 9.5 9.6 8.6 8.0 7.8 7.8 6.9 4.9 8.6 household (%) Share of individuals ages 6-17 25.5 23.7 22.2 21.8 21.3 21.0 19.6 18.6 17.1 14.3 20.5 in household (%) Share of adults ages 18-49 in 39.4 41.3 43.6 43.8 44.5 45.8 47.8 48.9 51.9 54.4 46.1 household (%) Share of individuals ages 50+ 22.5 24.4 24.7 24.8 25.6 25.1 24.8 24.7 24.1 26.4 24.7 a. Proportional to total budget per household in each decile; households that smoke = 1 if a household member reports spend- in household (%) 50 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia D. Robustness Checks The reduction in medical expenses and the rise in income from longer lives and greater productivity because of reduced morbidity are unlikely to have an immediate impact on household income. Instead, the effects are likely to materialize later in the life cycle of a smoker induced to quit as a result of higher taxes. To incorporate this in the analysis, dis- counted values of medical costs savings and years of life lost are included in the analysis. A discount rate of 3 percent is used for medical costs, assuming the costs of smoking materialize 5 or 10 years after the tax increase. Similarly, discount rates of 3, 5, and 10 per- cent are applied to years of life lost. The overall results hold with the discounted values, though the impact on income is less pronounced the higher the discount rate applied (figures D1 and D2; tables D1–D4). Table D1. Net Effect on Household Expenditures, White Cigarettes, 3 Percent Discount (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.16 0.14 0.12 0.12 0.10 0.09 0.08 0.08 0.08 0.10 Medium elasticity 0.24 0.24 0.23 0.24 0.22 0.21 0.20 0.22 0.24 0.24 Upper-bound elasticity 0.32 0.34 0.33 0.36 0.34 0.34 0.33 0.37 0.40 0.38 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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; a 3 percent discount rate has been applied to the years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in five years. Table D2. Net Effect on Household Expenditures, Clove Cigarettes, 3 Percent Discount (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity −0.50 −0.83 −1.01 −1.11 −1.18 −1.19 −1.11 −0.99 −0.79 −0.40 Medium elasticity 0.09 −0.18 −0.34 −0.44 −0.52 −0.57 −0.54 −0.50 −0.40 −0.19 Upper-bound elasticity 0.68 0.47 0.32 0.22 0.13 0.06 0.02 0.00 −0.01 0.02 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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; a 3 percent discount rate has been applied to years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in five years. 51 // Annexes Figure D1. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes, (Tobacco Price Increase, Medical Expenditure, and Working Years Gained) (3 percent discount rate) A. WHITE CIGARETTES .4 INCOME GAINS % .3 .2 .1 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound B. CLOVE CIGARETTES Lower Bound Upper Bound .5 INCOME GAINS % 0 -.5 -1 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. Years of life lost have been estimated using all deaths from tobacco-related diseases; a 3 percent discount rate has been applied to years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in five years. 52 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia Figure D2. Total Income Effect: Direct and Indirect Effects of Tobacco Taxes Tobacco Price Increase, Medical Expenditure, and Working Years Gained) (10 percent discount rate) A. WHITE CIGARETTES .3 INCOME GAINS % .2 .1 0 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound B. CLOVE CIGARETTES Lower Bound Upper Bound .5 INCOME GAINS % 0 -.5 -1 -1.5 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. Years of life lost have been estimated using all deaths from tobacco-related diseases; a 10 percent discount rate has been applied to years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in 10 years. 53 // Annexes Table D3. Net Effect on Household Expenditures, White Cigarettes, 10 Percent Discount (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity 0.13 0.11 0.09 0.08 0.06 0.05 0.04 0.04 0.03 0.04 Medium elasticity 0.20 0.19 0.18 0.18 0.16 0.15 0.14 0.15 0.16 0.15 Upper-bound elasticity 0.27 0.28 0.27 0.28 0.26 0.26 0.25 0.27 0.29 0.26 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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; a 10 percent discount rate has been applied to years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in 10 years. Table D4. Net Effect on Household Expenditures, Clove Cigarettes, 10 Percent Discount (%) PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound elasticity −0.56 −0.88 −1.05 −1.15 −1.22 −1.23 −1.14 −1.02 −0.82 −0.42 Medium elasticity 0.01 −0.25 −0.41 −0.51 −0.58 −0.62 −0.59 −0.54 −0.44 −0.22 Upper-bound elasticity 0.57 0.37 0.23 0.14 0.05 −0.01 −0.04 −0.06 −0.06 −0.02 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. 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; a 10 percent discount rate has been applied to years of life lost. Similarly, for medical cost savings, a 3 percent discount is used, assuming benefits materialize in 10 years. 54 The Distributional Effects of Tobacco Taxation: The Evidence of White and Clove Cigarettes in Indonesia E. Simulation: Tobacco Tax Revenue Redistributed to All Households If tobacco tax revenues were redistributed to households through direct transfers, this could offset health expenditures and potentially benefit lower-income households the E. Simulation: Tobacco Tax Revenue Redistributed to All Households most because the transfers would represent a larger share of the total expenditure of If these tobacco households. tax revenues Toweretest this hypothesis, redistributed a 10 percent to households through tax increase direct is assumed transfers, to be health this could offset expenditures and potentially benefit lower-income households the most because the equivalent to 10 percent of total tobacco tax excise revenue; this amount is redistrib- transfers would represent a larger share of the total expenditure of these households. To test this hypothesis, a 10 percent tax increase is uted to assumed to all households be equivalent to 10equally. 11 percent of This effect total is added tobacco asrevenue; tax excise a fourth component this in the anal- amount is redistributed to all households equally.11 This effect is added as a fourth component in the analysis, as follows: ysis, as follows: Income effect = change in tobacco expenditure (A) + lower medical expenses (B) + rise in income (C) + direct tax revenue transfer (D) (E1) !"#$%& ()* +$,. (#)./0$# = (()* +$,$.3$/(5&)6753/$ℎ569/)/Total expenditureH (E2) As expected, this additional revenue benefits poorer households the most (figure E1, panel b). Thus, an increase in tobacco taxes, accompanied by some redistribution, even if it is not earmarked for the poor, could result in a progressive policy As expected, outcome. this additional revenue benefits poorer households the most (figure E1, panel b). Thus, an increase tobacco in – Figure E1 accompanied taxes, Effect: Total Income by some All Tobacco redistribution, even if Products a. Tobacco it is not price for earmarked increase, medical the poor, expenditure, could b. Tobacco result in a progressive price increase, policy outcome. medical expenditure, working years gained working years gained and direct tax revenue transfer Table E1. Net Effect on Household Expenditures (Including Direct Tax Revenue Transfer), All Tobacco Products PRICE SHOCK DECILE SCENARIO 1 2 3 4 5 6 7 8 9 10 Lower-bound Source: elasticity Estimates based 1.42 on Indonesia National 0.47 Socioeconomic 0.05 2016.−0.27 Survey −0.49 −0.62 −0.66 −0.63 −0.55 −0.26 Note: Estimates assume a price shock of 25 percent. The redistribution of 10 percent of tobacco tax excise revenue to households is Medium assumed elasticity in panel b. 2.08 1.19 0.79 0.49 0.27 0.10 0.01 −0.03 −0.05 0.04 Upper-bound elasticity 2.73 1.92 1.54 1.25 1.02 0.83 0.68 0.57 0.45 0.34 Table E1 – Net Effect on Household Expenditures (Including Direct Tax Revenue Transfer), All Tobacco Products Source: Deciles Price scenario based on Indonesia National Socioeconomic Survey 2016. shock Estimates 1 shock of Note: Estimates assume a price 2 25 percent 4 3 and a redistribution 6 of tobacco 5of 10 percent 7 tax excise8 revenue9 10 Lower-bound elasticity to all households. 1.42 0.47 0.05 −0.27 −0.49 −0.62 −0.66 −0.63 −0.55 −0.26 Medium elasticity 2.08 1.19 0.79 0.49 0.27 0.10 0.01 −0.03 −0.05 0.04 Upper-bound elasticity 2.73 1.92 1.54 1.25 1.02 0.83 0.68 0.57 0.45 0.34 Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent and a redistribution of 10 percent of tobacco tax excise revenue to all households. 11 In 2017, tobacco excise tax revenues were approximately Rp 58.2 trillion. See “Government to Revise Indonesia's Tobacco Excise Tax Policy,” Indonesia Investments, Jakarta, August 22, 2017, https://www.indonesia- investments.com/finance/financial-columns/government-to-revise-indonesia-s-tobacco-excise-tax-policy/item8118?. In 2017, tobacco excise tax revenues were approximately Rp 58.2 trillion. See “Government to Revise Indonesia's Tobacco Excise Tax 11 31 Policy,” Indonesia Investments, Jakarta, August 22, 2017, https://www.indonesia-investments.com/finance/financial-columns/govern- ment-to-revise-indonesia-s-tobacco-excise-tax-policy/item8118?. 55 // Annexes Figure E1. Total Income Effect: All Tobacco Products A. TOBACCO PRICE INCREASE, MEDICAL EXPENDITURE, WORKING YEARS GAINED .5 INCOME GAINS % 0 -.5 -1 -1.5 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound B. TOBACCO PRICE INCREASE, MEDICAL EXPENDITURE, WORKING YEARS Lower Bound Upper Bound GAINED AND DIRECT TAX REVENUE TRANSFER 2 INCOME GAINS % 1.5 1 .5 0 -.5 1 2 3 4 5 6 7 8 9 10 DECILE CI 95% Medium Bound Lower Bound Upper Bound Source: Estimates based on Indonesia National Socioeconomic Survey 2016. Note: Estimates assume a price shock of 25 percent. The redistribution of 10 percent of tobacco tax excise revenue to households is assumed in panel b. 56