REGRESSIVE OR PROGRESSIVE? THE EFFECT OF TOBACCO TAXES IN UKRAINE Authors Alan Fuchs and Francisco Meneses ABSTRACT Tobacco taxes are usually considered regressive as the poorest indi- viduals allocate larger shares of their budget towards the purchase of tobacco related products. However, because these taxes also discourage tobacco use, some of the most adverse effects and their economic costs are reduced, including lower life expectancy at birth, higher medical expenses, increased years of disability among smokers, and the effects of secondhand smoke. This paper projects the effects of an increase in the tobacco tax on household welfare in Ukraine. It considers three price-elasticity scenarios among income deciles of the population. Results show that although tobacco taxes are often criticized for being regressive in the short-run, a more comprehensive scenario that includes medical expenses and working years, the bene- fits of tobacco taxes far exceed the increase in tax liability, benefitting in large measure lower income households. Our results also indicate that lower health expenditure seems to be the main driver because of the reduction in tobacco-related diseases that require expensive treatments. Tobacco taxes are also associated with positive distri- butional effects related to the higher long-term price elasticities of tobacco consumption. JEL Codes: H23, H31, I18, O15 Cover photos (clockwise from left): Western Ukraine, photo by Rodion Kutsaev. Kyiv, Ukraine, Source: Twitter NGO Advocacy Center LIFE (https://twitter.com/LIFECenterNGO/status/887557821017063424) photo by Oleh Andros. Andriivs'kyi descent, Kyiv, Ukraine, photo by Illia Cherednychenko. Lviv, Ukraine, photo by Anton Dee. REGRESSIVE OR PROGRESSIVE? THE EFFECT OF TOBACCO TAXES IN UKRAINE Alan Fuchs and Francisco Meneses1 September 27, 2017 1 Fuchs: Poverty and Equity Global Practice, World Bank, I 4-405, 1818 H Street NW, Washington, DC 20433 (email: afuchs@worldbank. org). Meneses, Duke University and Universidad Adolfo Ibañez (email: fjmeneses@gmail.com). Support for the preparation of this report was provided by the World Bank’s Global Tobacco Control Program, co-financed by the Bill and Melinda Gates Foundation and the Bloomberg Foundation. We are grateful to Patricio Marquez, Konstantin Krasovsky, Tatiana Andreeva, Olena Doroshenko, Paolo Belli, Alexandru Cojocaru and Mikhail Matytsin for providing inputs, comments and support. The findings, interpretations, and conclusions in this research 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. THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS 1 INTRODUCTION On December 19, 2016, the Ukrainian Parliament approved the 2017 budget, which includes a specific excise tax on tobacco products that represents a 40 percent increase over the corresponding tax in 2016, while maintaining a 12 percent ad valorem tax. Because low-income families usually allocate a larger proportion of their budgets to purchase tobacco products and alcoholic beverages, the tax increase would seem to be a regressive policy at first glance. However, a closer look reveals that the expected over- all reduction in tobacco consumption2 associated with the tax increase would -in the long run- reduce the adverse effects of tobacco consumption, including higher medical expenditures and added years of disability among smokers, the negative effect on life expectancy at birth, reductions in the quality of life, and numerous negative externalities among first- and secondhand smokers, thus benefiting former smokers and their families. Meanwhile, a boost in government revenue (paid by those who continue to smoke) ear- marked toward providing social transfers (such as health care or pensions) could further lever the benefits of a tobacco taxation for the poorest households. The increase in the tobacco tax and the subsequent reduction of tobacco consumption could therefore result in potential measurable benefits for different income groups. This paper describes and quantifies the effects of tobacco tax increases on aggregate household welfare through three channels. Channel (1) implies that higher tobacco prices due to higher taxes induce behavioral response in the means of a reduction on tobacco consumption.3 The reduction in consumption is then associated with (2) a reduction in medical expenses, and (3) a rise in income because of the gain in years of employment. To assess the impact of these effects, this paper estimates the price elastic- ity of tobacco, simulates upper- and lower-bound scenarios, and calculates the welfare gains among various population income groups. There is ample and robust evidence linking tobacco consumption with health-related problems.[1, 2] Diseases associated with tobacco use range from lung cancer to stroke and even to congenital malformation in children.[3] In 2010, 7 million early deaths were attributed to tobacco consumption globally.[4] Today, more than 80 percent of the world’s smokers live in low- and middle-income countries, harming health, incomes, earning potential, and labor productivity and undermining human capital accumulation, 2 Most importantly, the number of people who quit smoking or do not start at all. 3 This reduction in tobacco consumption is manifested through the set of people that discontinue smoking, and in the long run, younger individuals that do not start smoking at all. 3 Regressive or Progressive? The Effect • The Tobacco of of Effect Taxes Tobacco in Ukraine Taxes in Ukraine which is critical to sustainable economic growth and social development.[5] In Ukraine, 85,000 deaths are attributed to tobacco consumption yearly.[6] Evidence linking tobacco and health problems has triggered important policy shifts among international organi- zations and policy makers alike, leading to more rigorous restrictions and taxation on the sale and use of tobacco. Accordingly, the World Health Organization (WHO) has made the reduction of tobacco consumption one of its primary goals, thus promoting tobacco monitoring, smoke-free policies, smoking-cessation support programs, relevant health advice, advisory deter- rents, and taxation policies.[7] Among strategies, the tobacco tax seems to be one of the most efficient measures for reducing tobacco consumption and has the added benefit of raising government revenue.[8] The inelastic demand of some tobacco consumers is use- ful in increasing tax revenues, and the higher price elasticity of younger smokers makes the tax an efficient consumption deterrent in the long run.[8, 9] A recurrent concern is the potential regressivity of tobacco taxes because low-income households allocate larger shares of their budget to purchase tobacco products relative to richer households. Nonetheless, in this paper we show that, if indirect (health) effects are included in the calculations, the concern about tobacco tax policies is no longer valid. Instead, the future benefits of nonsmoking outweigh the losses attributed to tobacco taxes among the population in general and among low-income groups specifically.[10, 11] Beyond short-term reductions in household tobacco expenditures, the possible benefits of tobacco tax policies include lower medical expenditures and more healthy life years, both of which could translate into accountable economic benefits that more than offset the losses generated by tax increases when consumers discontinue -or never start- smok- ing. To test these hypotheses, a social welfare framework is used to calculate the effects on various income groups and different price elasticities for tobacco consumption are estimated. To establish a contextual background, section 2 briefly reviews the litera- ture on the health effects of tobacco, tobacco policies, and price elasticities. Section 3 describes the methodology, parameters, and data used to forecast the impact of the tobacco tax. Section 4 presents the estimation results, and section 5 concludes with a discussion on policy implications. 4 // Introduction 5 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS 2 THE LITERATURE a. Tobacco and health During the last century, about 100 million deaths were related to tobacco use.[13] If current trends were to remain constant, about 1 billion people could die from tobacco- related diseases during this century.[14] In Ukraine, 7.2 million adults consume tobacco every day, and, in 2010, over 85,000 deaths were attributed to tobacco consumption.4[15] According to the U.S. Department of Health and Human Services, tobacco consumption is responsible or contributes to many types of cancers, including lung, oral, laryngeal, pancreatic, kidney, cervical, and acute myeloid leukemia.[3] Smoking is related to respira- tory problems such as chronic respiratory symptoms, tuberculosis, influenza, pneumonia, other infections, chronic bronchitis, emphysema, and asthma. It is also associated with cardiovascular diseases, such as aneurysms, stroke, and coronary heart disease, as well as adverse reproductive and developmental effects, such as low birthweight, congen- ital malformation in babies, and complications in pregnancy, along with male sexual dysfunction.[3, 14] The exposure to secondhand smoke has a causal relationship with many respiratory diseases in children and adults. There are more than 4,000 chemicals in tobacco smoke (of which at least 250 are harmful and more than 50 can cause cancer). Although the nature of the causal relationship between secondhand smoke and cancer or its impact on reproduction is not clear, research has strongly connected them.[3,16] Moreover, according to the WHO, secondhand smoke is responsible for over 600,000 premature deaths worldwide. b. Tobacco control policies Globally, antitobacco policies include smoking prohibition in specific locations and com- pletely smoke-free environments, advertising to deter tobacco use, smoking cessation programs, prohibitions on tobacco sales close to schools, and taxation. These various policies have shown diverse effects in tobacco use, tobacco availability and secondhand smoke exposure among the population. WHO argues that entirely smoke-free environments, rather than separate smoking rooms or good ventilation systems, are the only way to prevent the harmful consequences of secondhand tobacco smoke.[5] Smoke-free laws are popular because there is evidence 4 Children refers to individuals under 18 years of age. 7 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine that they improve health outcomes without affecting business. In 2016, these laws ben- efited 19 percent of the world's population.[7] The benefits depended on the breadth of the legislation. For example, prohibiting smoking in all indoor workplaces reduced the exposure to secondhand smoke by 80 percent–90 percent and decreased the incidence of acute respiratory illness (IARC 2009). In 2016, about 33 percent of the world's population had access to smoking cessation sup- port programs, 21 percent more than in 2012.[7] These programs represent the fifth most widespread policy in the world.[5] They significantly raise quitting rates among smokers who want to quit and are more cost-effective compared with other health care programs. [17] Although they are effective, they only treat those addicted individuals who want to be treated. Their presence is also associated with country income, and they are mainly found in high-income countries. Another way to discourage tobacco consumption is through health warning labels on tobacco packages. This is the third most common policy against cigarettes in the world. In 2016, almost 45 percent of the world’s population was being exposed to such labeling. Warning labels are widely supported by the public and may not represent a cost to gov- ernments.[5] However, they have to be regularly updated and changed to remain impact- ful. The use of warnings may influence people against tobacco consumption; however, their use only accounts for a marginal decrease.[18–20] Mass media campaigns that reach large populations represent the most popular and common way to combat tobacco use. In 2016, such campaigns addressed 56 percent of the world's population. People in low-income countries are less likely to be exposed to these campaigns. However, there is limited information about the cost-effectiveness of this approach.[5] Durkin, Brennan, and Wakefield (2012) conclude that mass media aware- ness programs could promote quitting; however, their impact depends on the duration of the campaigns, especially among low-income smokers. It also depends on the mes- sage; information about the adverse health risks of smoking represents the most efficient means to reach users.[21] In Ukraine, several tobacco control policies have been adopted in the last years. In 2005, the Law on Measures to Prevent and Reduce the Use of Tobacco Products and their Harmful Impact on the Health of the Population was adopted. On 15 March 2006, Parlia- ment ratified the WHO Framework Convention on Tobacco Control (FCTC), implemented fully on September 4, 2006 .[22] According to the WHO Reports on the global tobacco epidemic, between the Second (survey data collected in 2006 or earlier) and the Third Reports (survey data collected in 2009 or earlier), Ukraine has demonstrated one of the fastest declines in smoking prevalence in the world: age and sex standardized current tobacco smoking prevalence declined from 45% to 32%. Per the national reports, daily smoking prevalence in Ukraine 8 // The Literature decreased from 37.2% in 2005 to 25.5% in 2010. Ukraine has followed best international tobacco control practices, but the success has been achieved without governmental funding for tobacco control activities. Ukraine has almost not used those strategies which require even moderate national resources like quit lines or other cessation services. The decline in smoking prevalence hence potentially resulted from the tobacco control legislation first adopted in 2005 and amended later, which included extension of smoke free policies; step-by-step tobacco advertising bans; large health warnings and other measures; however, most of these policies were implemented between 2005 and 2007.[23] Those legislative measures which came into force in late 2012 including (1) tobacco adver- tising ban — on September 16, 2012, (2) introduction of large (50% of the pack surface area) graphic health warning on tobacco packaging — on October 4, 2012, (3) smoke-free policies in restaurants and other public and workplaces — on December 16, 2012, have resulted in significant decrease of cigarette sales in Ukraine in 2013.[24] c. Tobacco taxes Prohibiting certain practices (e.g., tobacco marketing or indoor smoking) has a limited effect because after such practices are prohibited, they cannot be prohibited further. Increases in tobacco excise taxes are not subject to such constraints; excise taxes can continue to be increased, even if the tax rate is already very high.[25] Tobacco taxation is considered one of the most efficient measures to reduce tobacco consumption; as a secondary benefit, they also increase government revenue.5[8] Because both effects are desirable from a pol- icy standpoint, the use of taxes is considered to be economically justified. Additionally, the higher price elasticity of young people makes taxes a good way to fight tobacco use because taxes will significantly reduce consumption in the long run. Institutions such as the Inter-American Development Bank (IADB 2010), WHO (2008), the International Agency for Cancer Research (IARC 2011), and the World Bank (1999) and authors such as Levy et al. (2014) have associated price increases with significant declines in tobacco consumption.[8,26–29] These authors, as well as the WHO, estimate that higher taxes are responsible for almost half the decline in smoking.[30] However, the effects of these pol- icies mainly depend on the type of taxes. For example, ad valorem taxes are based on prices; so, tobacco companies can potentially avoid higher taxes by cutting on providers and setting lower prices. For this reason, consumption levels and tax revenue depend on the industry pricing strategy. Alternatively, specific excise taxes establish a fixed tax amount, although the tax amount must be adjusted periodically for inflation to accomplish their mission and are associated with the risk of encouraging contraband sales.[5] The taxation system in Ukraine, as in other countries, uses both types of tobacco taxes, that is, specific and ad valorem excise. 5 Tobacco tax increases have also been associated with a rise in contraband and illegal tobacco sales, reducing the expected increase in government revenue (Jha and Chaloupka 2000) [56] 9 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine d. The price elasticity of tobacco consumption The extent of price elasticity is crucial in calibrating the effect of taxation systems because it determines the sensitivity of demand to a change in tobacco prices. In general, tax increases generate larger shifts of tobacco purchase and consumption among low- and middle-income populations than in high-income groups (WHO 2015a). [5] There is an extensive literature estimating the relationship between tobacco prices and consumption. Guindon (2013) provides a broad review of 26 international studies; a few should be mentioned directly.[31] Chaloupka and Grossman (1981) and Lewit and Coate (1981) estimate the elasticity among the under-18 population in the United States at, respectively, −1.44 and −1.31.[33, 34] Among adults ages 18 years or older, Chaloupka (1991) and Lewit and Coate (1981) estimate the elasticity in the United States at between −0.27 and −0.42, respectively.[34, 35] For all 52 countries in the European region, Gallus et al. (2006) estimate a price elasticity of −0.46 using national yearly aggregated data.[35] For Hungary, the price elasticity is estimated at between −0.44 and −0.37 and, for the United Kingdom, at −0.5.[37, 38] For Poland and Turkey, tobacco price elasticities have been estimated at, respectively, −0.4 and −0.19 in the short run (−0.7 for long-run elastic- ity in Poland).[39, 40] For India, cigarette price elasticities have been estimated for differ- ent income groups, finding −0.83 and −0.26 for the lowest and highest income groups, respectively.[40] For South Africa, Van Walbeek (2002) estimates prices elasticities for different income quintiles, controlling for income changes, and finds elasticities of −1.39 and −0.81 for the poorest and richest income quintile, respectively.[41] For Taiwan, China, price elasticities of tobacco have been estimated at −0.29.[42] Research conducted in such countries as USA, UK, Canada, Bangladesh, China, and Indo- nesia has indicated that smoking prevalence among men and women in lower socioeco- nomic groups is more responsive to the changes in cigarette prices; however, in countries such as Egypt, Bulgaria, and Turkey the evidence is mixed [IARC 2011]. In Ukraine, we see little difference in response to sharp price increase among different SES groups in short- term perspective, while in medium-term perspective (5 years) tobacco tax hikes have higher impact on smoking prevalence rates among younger and poorer.[23] There are two important factors involved in determining tobacco price elasticities: income and age. People in low-income groups have more elastic demands relative to medium- and higher-income groups.[14] At the same time, younger groups in popula- tions are more responsive to price and thus tax increases because they tend to be less nicotine dependent, more affected by peer effects, and possess less disposable income. [14] Studies in the United States have consistently shown that younger groups have higher elasticities relative to older groups.[32,33,43] 10 // The Literature e. Further costs of tobacco: life, work, and medical expenses The major costs of tobacco consumption beyond the direct price are associated with public and private health care costs. Tobacco-related health care costs can be either direct or indirect. Direct costs include the monetary value of the consumption of goods and services motivated and, in many cases, compelled because of tobacco use. These are divided into health care costs (hospitalization, medication, medical supplies, equipment, and so on) and non–health care costs (job replacements for sick smokers, insurance, cleaning up the cigarette ash and stubs, packaging, and smoke residue of smokers, and so on). Goodchild et al. (2016) estimate that the global economic cost of tobacco-related diseases is equivalent to 1.8 percent of the world’s gross domestic product (GDP).[57] For the United States, the direct health care costs associated with tobacco-related diseases are estimated at 1.1 percent of GDP or 8.7 percent of annual health care spending.6[44] Meanwhile, the indirect costs include the loss of productivity because of lost working days related to smoking illnesses and the value of the lives prematurely lost. Both effects are incorporated in the disability-adjusted life years indicator.[45] Focusing on health care costs, Lightwood et al. (2000) estimate the cost of tobacco use. They suggest that the gross health care cost in high-income countries fluctuates between 0.1 percent and 1.0 percent of GDP.[46] In terms of price elasticities, limited data inhibit accurate estimates in low- and middle-countries, but the authors argue that the price elasticity could be as high as those in high-income countries. Meanwhile, Verguet et al. (2015) analyze the health effects of a price increase in China.[12] Their research concludes that a 50 percent rise in prices would result in 231 million years of life gained over 50 years, with a significant impact in the lowest income quintile. Pichón-Riviere et al. (2014) estimate that tobacco use in Chile will reduce life expectancy by nearly 4.0 years among women and 4.3 years among men.[47] There would also be about 379,000 life days lost, which is more than a thousand years.7 The research represented in this paper takes advantage of the current literature on health care costs. It draws information and ideas from Marquez et al. (2017) and others who estimate the long-term health care costs of tobacco in Ukraine and adds new estimates and measures of the incidence of disease and of income distribution.[54] From a meth- odological standpoint, this paper follows the methods described by Fuchs and Meneses (2017).[48] 6 Estimated cost US$169.3 billion divided by GDP (2010); US$14.96 trillion equals 1.13 percent. 7 Days lost refer to years of life lost (YLL) because of premature mortality. Another indicator is years lost because of disability (YLD) among people living with poor health and its consequences. Usually disability-adjusted life years = YLL + YLD. 11 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine f. Costs and benefits that are not included Some well-researched costs and benefits are not covered in this paper. Secondhand smoke has been shown to be an important societal cost of smoking, affecting the health of adults and children.[49] In one example, from the state of Indiana, the health-related costs of secondhand smoke have been estimated at more than US$1.3 billion yearly, US$201 per-capita. However, this paper does not include the cost of secondhand smoke in the analysis because of the lack of detailed information required about smoking within households or the workplace. As discussed earlier, one potential channel in which tobacco taxes can improve income inequality is the possible use of tax revenues on progressive policies such as direct income transfers or services. These policies might involve expansions in health care, social welfare, and education expenditures. Although earmarking tax revenues for specific projects -such as health care, social welfare or education- is a common practice in some countries, this paper does not include them in the assessment as they depend on a myr- iad of factors that include political decisions. Therefore, this paper covers only benefits that directly arise from tobacco tax policy. 12 // The Literature 13 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS 3 MODEL The impact of the tobacco tax in Ukraine is estimated using a social welfare framework similar to the framework applied elsewhere in the literature.[12,50] The potential changes in household welfare induced by an increase in tobacco taxes is estimated by consid- ering three factors: (1) the rise in tobacco expenditures because of the tax increase, (2) the reduction in medical expenses because of lower tobacco consumption, and (3) the change in incomes because of reduced mortality, leading to additional work years.8 The aggregated effect of the tax policy is estimated as follows: Income effect = change in tobacco expenditure (A) + lower medical expenses (B) + rise in income (C) (1) The basis of the analysis rests on the Ukrainian Household Living Conditions Survey 2012. The survey is designed to measure general patterns of expenditure, including expendi- ture in tobacco products. Data limitations do not permit a simulation of the exact price increases by brand, but this may be accomplished using the aggregate prices paid by households. A partial equilibrium model allows the distributional effects of the tobacco tax to be assessed, resulting in an estimation of the first-order effects of these policies. A partial equilibrium approach is then used, and the change in prices is evaluated, mainly by rely- ing on household expenditure patterns. This decision implies that only the first-order response is assessed and that additional behavioral changes among economic agents are first-order not first-order first-order first-order response covered, response such response response is assessed as is and the is assessedassessed thatexpansion is assessed and and and that additional inthatthe additional that behavioral consumption additional additional behavioral behavioral behavioral changes of other changes goods. changes among changes among economic These among among assump- economic economic agentseconomic are agents agents agents not areare not covered, are notnot co covered cove such as the e is assessed and that additional such suchsuch as as tions behavioral the expansion asexpansion the imply the in changes expansion expansion that the theamongin the in model consumption in the the uses economic consumption consumption consumption oftheother agents share of goods. are other of of tobacco notof other othergoods. These goods. goods. covered, These These These consumption assumptions assumptions assumptions assumptions in imply thatimply household the imply imply that budgets modelthatthat the the model the uses the uses model model u th uses sion in the consumption share of tobacco of other share share share per goods.of consumption tobacco of priceof These tobacco tobacco increases. consumption consumption consumption in assumptionshousehold The loss in in in imply realinincome household household household budgets that the model budgets per budgets budgets price arising per from uses per price per increases. price the price increases. increases. price increases. The loss in increases The The realThe loss products loss in loss income real in inincome i =real arising 1, real income income from arisin arising from arising f onsumption in household price increases budgetsprice price price …, per isincreases increases in increases nproducts price in in obtained i= increases.in1,products products products by …, in The =is i 1, loss = =real i …, in1, 1, …, obtained n …, n obtained is is by income n obtained is obtained by by arising by from products i = 1, …, n is obtained by * * * ∆&' ∆&'∆&'∆&' * " " + ∆" " " "∗ "+ " ∆ + , ∆ " +" ∆ ∗" & " ,∗ & ∗ , , (2) (2)(2)(2) &',) ',)&',) ',) * ∆& " + ∆" ∗ ' , " (2) where "&where iswhere ',) where where the share " is " the is "of isshare the the the share product share share of of iproduct of in of product productproduct total i total i household ini intotal in in totaltotal household household household household expenditure, expenditure,expenditure, expenditure, expenditure, and ∆" and and is and the and ∆ ∆ " is ∆ is the is " the percent is " the the percent percent price percent percent price 9increase price increase. price inc increa hare of product i in total Therefore, Therefore, household Therefore, Therefore, if 10 price percent expenditure, if 10 increase. if of 10if percent 910 percent the Therefore, and percent total of of the budget if of the the total 10 is total total budget destined percent budget budgetofisthe is destineddestined destined is total for 9 cigarettes, forfor cigarettes, budget for is cigarettes, cigarettes, example, destined forfor for andfor example, example, example, the cigarettes, and price the and forand of price thethe price cigarette of of price cigarettes of cig cigare ∆ " is the percent price increase. ercent of the total budget risesis by rises 10 destined risesrises by percent, for example, by10 by10percent, the cigarettes, 10 andpercent, real percent, the loss fortheprice the real in example,the ofloss real income real loss loss in and cigarettes in income in amounts income income the price rises amounts byamounts to 10percent. 1 of amounts to to cigarettes percent, to percent. 1 ∆ 1 is 1 percent. the percent. the " real ∆ loss ∆ changein " is ∆ is" the " the income ischange in the change change consumption amounts in in in consumption the of of consumption consumption of th t, the real loss in income taxed taxedtaxed good, which amounts to taxed to 1good, good, good, which depends∆ 1 percent. which which depends on" the depends is the is depends the price on change change the on on the the price elasticity in price price elasticityelasticity elasticity of the specific consumption inconsumption of the of ofofof the specific the good. thethe specific specific good. good.good. taxed good, which depends on the depends on the price elasticity of the specific good. price elasticity of the specific good. Tobacco Tobacco Tobacco expenditures Tobacco expenditures expenditures : expenditures The : The variation : The : variation The invariation tobaccovariationin in in tobacco tobacco tobacco consumption consumption consumption consumption after the after tax after the after the increase taxthe tax increase is tax increase increase estimated is is is estimated estimated onbased estimated based based a basedono : The variation in tobacco consideration consideration consideration consideration consumption of the after change of of the the of the in the change prices tax change change increase (in∆ prices in inisprices ), prices the (∆ ), ( estimated the (∆ tobacco ), the ∆ the tobacco ), price tobacco based tobacco price elasticity on a price priceelasticity elasticity elasticity , and the , and ,the , and tobacco andthe the tobacco tobacco expendituretobacco ofexpenditure expenditure decile of of expenditure of deci de i in he change in prices (∆ ), theperiodi in tobacco0i ( iOther period in in period period price 0( elasticity studies 0 8 (0 ( have also .,evaluated ) and ) the0 . productivity ). ). tobacco 0 0 loss,expenditure of decile disability costs, externalities, and so on. Because of the availability of relevant data, 0 this paper focuses solely on medical expenses and income changes associated with shifts in mortality. 0 ). 9 For a detailed discussion of the methodology, see Coady et al. (2006) and Kpodar (2006).[51,52] 10 10 10 10 ∆ ∆ 1 ∆ ∆ " = ( + = "∆ " ( ="1 1 (=+ +1( ∆ + 1 + 1 ∗∆∆ ∆ +− 1 +1)1 + ∗ ∆ ∆ ∗− ∆1) − 1) ∗− ∗∗ 1) ∗ ∗ "C "C "C (3) "C (3) (3) (3) 10 " = ( 1 + ∆ 1 + ∗ ∆ − 1) ∗ "C 15 (3) The The change inchange The The change change tobacco in tobacco tobacco inexpenditure in tobacco expenditure expenditure expenditure is dividedis by is divided divided divided isthe byexpenditure total the by by total thethe total expenditure total expenditure expenditure for for each forfor each decile eacheach decile group decile decilegroup group i, thereby group i, thereby iobtaining , thereby i, thereby ob obtainin obtain a total a comparable acco expenditure is divided by the a expenditure comparable comparable a comparable per per household household per per household household measure for each measure of measure the decile measure of iof change group the of the in the change , thereby change change tobacco in tobacco tobacco inexpenditure in tobacco obtaining expenditure expenditure expenditure relative torelative relative the relative total the to to the the to expenditure total totaltotalofexpendi expenditure expendituro household measure of theeach decile changeeacheach group, in each decile tobaccodecile as group, decile group, group, follows: follows: as as expenditure as follows: follows: relative to the total expenditure of (EF∆G (EF∆G EFH∗∆G EFH∗∆G IE)∗JK&L*M"NOPL IE)∗JK&L*M"NOPL such as price the expansion increases in products in the= consumption of in other bygoods. These assumptions imply that nthe model uses by the per price increases. The loss in ireal price1, …, income n is increases obtained arising products from price i= 1, …, n increases in obtained i is products by = 1, …, is obtained share of tobacco consumption in household budgets ∆& per price increases. The loss in real income arising from first-order response is assessed and *that additional behavioral " + ∆" ∗ by' , changes * among (2) economic ∆&' *agents are not covered, ∆&' price increases in products i = 1, …, n "is obtained &',) " " + ∆" ∗ & , " " + ∆" ∗ & (2) , (2) such as the expansion in the consumption of other goods. These assumptions imply ',) that the model uses ',) the (2) ∆&' whereof share the share isRegressive "tobacco first-order consumption of product response where in * iEffect household in istotal +assessed ∆ household budgets ∗and ,that per additional expenditure, price increases. i the in behavioral andThe (2) of"loss ∆ changes is the in real among percent income economic price arising increase. agents from 9 not covered, arepercent or Progressive? The " " " is of the Tobacco "share Taxes &',) of where inproduct Ukraine " is total share household product expenditure, i in total and household ∆ " is the expenditure, and price ∆increas " is th price increases Therefore, if 10 in such products percent as of the i = the expansion 1, total Therefore, …, n is budget if in obtained10 the is consumption destined percent by of Therefore, for the of cigarettes, total if otherbudget 10 forgoods. percent isexample, destined of These the and assumptions for total the price cigarettes, budget isofimplyfor cigarettes destined that example, the for model and the cigarettes, uses price for theof cigaret example, ld expenditure, and ∆" is the percent price increase. 9 where rises by " ispercent, 10 the share share theof of real tobacco product iconsumption in total household in household to expenditure, budgets and per ∆ price is increases. the to percent The price loss in real ∆increase. income 9 arising from d for cigarettes, for example, andloss the in rises price *by income 10of percent, amounts cigarettes the ∆&' real rises 1by percent. loss 10in income percent, ∆ " is the the amounts " change real loss 1in in consumption percent. income amounts of to " is the the change 1 percent. in consumption ∆" is the chan of Therefore, if 10 price percent increases of the total in products budget " + ∆ is i " = ∗1, …, destined , n for is obtained cigarettes, by for (2) example, and the price of cigarettes taxed good, which to 1 percent. ∆" is the change in depends on taxed the " consumptionprice good, which elasticity of depends of & the taxed ',) the specific on good, the price good. which elasticity depends of on the the specific price elasticity good. of the specific good. rises by 10 percent, the real loss in income amounts to 1 percent. ∆ ∆& " is the change in consumption of the he specific where good. the share of product i in total household " is which * " expenditure, " + ∆" ∗ & and ' , ∆" is the percent(2) price increase.9 taxed good, Tobacco expenditures Tobacco :depends The variation expenditures: on the Tobacco The price invariation elasticity tobacco expenditures in of Tobacco consumption tobacco the specific consumption after good. the aftertax increase the tax increase is estimated is esti- based on a Therefore, if 10 percent of the total budget is destined for cigarettes, for example, and the pricethe : The variation expenditures in tobacco : The consumption variation in after tobacco tax increase of consumption is estimated after the tax based increase on ',) cigarettes mption consideration after the mated of the tax increase based where change on is is ain prices consideration consideration estimated the share ( ∆ ), based the ofof of amounts producttobacco the the on a itochange change price consideration in1total in in elasticity prices prices household ( of∆ the ), , , and the thechange the tobacco tobacco tobacco in price prices expenditure elasticity ( ∆ ), the , tobacco of and decile the price tobacco expenditure elasticity , and the of dec tob rises Tobacco by expenditures 10 percent, : The the real " variation loss inin income tobacco consumption percent. after the ∆"tax is expenditure, theincrease change is in and estimated ∆" is the of consumption based on percent the a price increase. 9 o price i inelasticity period 0( , and the Therefore, tobacco )i.ifin period expenditure 10 percent 0 ( of decile decile the total i in period budget ) . 0 is ( . destined for cigarettes, ) . for example, and the price of cigarettes taxed good, which consideration of the change on depends 0 the price in prices (∆elasticity ), the tobacco of theprice specific 0 elasticity good. , and the0 tobacco expenditure of decile rises by 10 percent, the real loss in income amounts to 1 percent. ∆" is the change in consumption of the i in period 0 (0 ). Tobacco expenditures : The good, taxed variation ∆ which in " = depends tobacco 1 + ∆ on the consumption 1+ price ∗ ∆ ( ∆ elasticity after −∆ 1) the ∗ of the taxspecific increase "C good. 10 is estimated (3) based on a ∆ 10 (3) "C " = ( 1 + ∆ 1 + ∗ ∆ " =− ( 1)1+ ∗ ∆ 1+ ∗ "C − 1) ∗ consideration of the change ∗ ∆ − 1) ∗ "C 10 in prices (∆ ), the tobacco price elasticity , and the tobacco expenditure of decile (3) The change i in period 0 in tobacco Tobacco expenditure ∆ ( expenditures )The . changeis " := The divided ( in 1+ variation tobacco∆the by 1 totalexpenditure + Thein ∗ tobacco expenditure ∆ − 1) change isconsumption in divided for each by ∗ tobacco decile the expenditure after total 10 the tax increase is estimated based on a "C group expenditure is dividedi, thereby (3) byfor obtaining theeach total decile expendituregroup i, thereby for eachobtain decile 0 a comparable The per change consideration household in tobacco measure a of comparable expenditure the ofchange the per change is in divided prices household ina ( tobacco comparable by ∆ ),the measure the totaltobacco expenditure of per expenditure the household price change relative for elasticity in to measure each tobacco the ,decile and total of expenditure the the expenditure changetobacco relative in expenditure of tobacco to the of total expenditure decileexpenditure relative tal expenditure for each decile group i, thereby obtaining eachchange The decile in tobacco group, group ias,in follows: thereby expenditure period to obtaining 0 ( each is expenditure decile dividedby a comparable group, ).the total per household expenditure for measure aseach ofdecile the change group in i, thereby obtaining tobacco expenditure relative the total 0 asof follows: each decile group, follows: 10 a comparable ∆ per household " = measure of ( the1 +change ∆ 1 in + ∗ ∆ − tobacco 1) ∗ expenditure "C to the total expenditure relative (3) of tobacco expenditure relative to the total (EF∆G expenditure EFH∗∆G IE)∗JK&L*M"NOPL of each decile 'Q group, (EF∆G as EFH∗∆G follows: IE)∗JK&L*M"NOPL each decile group, as follows: ∆ . = ∆ . = ' ∆ . (4) = (EF∆G 'Q EFH∗∆G IE)∗JK&L*M"NOPL'Q (4) RSNTU JK&L*M"NOPL RSNTU JK&L*M"NOPL' 10 RSNTU JK&L*M"NOPL' ∆G EFH∗∆G The change in tobacco IE)∗JK&L*M"NOPL 'Q expenditure (4) is divided by ∆ the total " = expenditure ( 1 + ∆ 1 for + each ∗ ∆ decile − 1) group ∗ i , thereby obtaining "C (3) (EF∆G EFH∗∆G IE)∗JK&L*M"NOPL'Q a comparable RSNTU JK&L*M"NOPL' This gives us the per household ∆ . change measure in the This proportiongives ofus the change of the = in in tobacco change tobacco expenditure,the expenditure proportion that is the relative change toin the tobacco total (4) expenditure consumption of This gives RSNTU JK&L*M"NOPL us the ' of tobacco change in the expenditure, proportion of that tobacco is the expenditure, change that isconsumpti in tobacco the chang each in decile relation group, to the The as household follows:change in budget. tobacco expenditure is divided by the total expenditure for each decile group i, thereby obtaining expenditure, that is the change inin relation tobacco to the household consumption in relation budget. to the household budget. This gives us the change a comparable in the proportion per household of tobacco measure (EF∆G of expenditure, the change EFH∗∆G IE)∗JK&L*M"NOPL that in istobacco the'Qchange expenditure in tobacco relative consumption to the total expenditure of gives This ∆ . each us the decile change group, in the as follows: proportion = of tobacco expenditure, that is the change (4) in in relation Medical to the expenses : The household change budget. in Medical medical expensesexpenses : The from change Medical tobacco-related RSNTU JK&L*M"NOPL inexpensesmedical : The ' diseases is estimated in equation (5), expenses change from in medical tobacco-related expensesdiseases from tobacco-related is estimated indiseases equation is ( tobacco consumption in relation to the household budget. obtaining the cost of the treatment obtaining of tobacco-related the cost of the (5), obtaining diseases treatment the=for of income cost tobacco-related (EF∆G decile EFH∗∆G i fromdiseases IE)∗JK&L*M"NOPL Pichón-Riviere for income et decile al. i from for income decile et Pichón-Riviere i m tobacco-related diseases isin estimated in equation that of the change treatment of tobacco-related diseases 'Q This gives us the change the ∆ . proportion of tobacco expenditure, is the in tobacco consumption (4) d diseasesMedical (2014) for expenses and adjusts income Medical : The it decile change according expenses: i from in (2014) The to medical the Pichón-Riviere change expenditure and adjustsexpenses in medical etit al. from survey. according (2014) expenses tobacco-related andto from theadjusts expenditure tobacco-related it diseases RSNTU JK&L*M"NOPL according survey.is toestimated diseases the ' is in expenditure esti- equation survey. (5), in relation to the household budget. of tobacco-related diseases for income decile i from Pichón-Riviere et al. vey. obtaining the mated cost of the treatment This in equation gives us (5), the change obtaining ( EFH∗∆G inthe the cost proportionof the treatment IE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY of tobacco of ( EFH∗∆G expenditure, tobacco-related IE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY ' that diseases is ( the change in tobacco for IE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY consumption ∆ . . to the = =(5) EFH∗∆G ' (2014) and adjusts it according expenditure ∆ . survey. . RSNTU JK&L*M"NOPL = ∆ . . (5) ' RSNTU JK&L*M"NOPL' Medical expenses income YN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY : The in relation change decile ' i from to in the medical household Pichón-Riviere expenses budget. et al. from (2014) tobacco-related and adjusts ' it diseases according is to estimated the expendi- in equation RSNTU JK&L*M"NOPL (5), ' (5) ( EFH∗∆G IE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY' SNTU JK&L*M"NOPL' obtaining theture Equation 5 shows the incomecost of ∆ . survey. the treatment . gains associated Equation of =5 tobacco-related shows withthe the income diseases reduction gains forin medical associatedincome with decile expenses the i from because reduction Pichón-Riviere (5)ofinlower medical tobacco et al. expenses because of lower tobac Equation RSNTU JK&L*M"NOPL 5 shows ' the income gains associated with the reduction in medical expenses (2014) consumption and adjusts in the Medical it according long expenses term. to Although :the The change expenditure the calculation in medical survey. is notexpenses realistic from in the tobacco-related short term because diseases it assumes is estimated that in equation (5), e reduction in medical expenses because consumption of lowerin the long tobacco term. Although consumption in the the long calculation term. Although is not realistic the calculation in the short is not term because realistic in it theassumes short te t Equation the effects 5 of shows obtaining the tobacco-related income thegains diseasecost of associated willthe treatment immediately with the of reduction diminish tobacco-related inwith medical the diseases expenses reduction for because in income tobacco of decile lower consumption, i tobacco from Pichón-Riviere et al. n is not realistic in the term because it the short . ∆ . effects (of assumes that the effects of tobacco-related disease will EFH∗∆G =according tobacco-relatedIE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY disease will immediately ' diminish with the reduction (5) immediately diminish with the reduction in tobacco consumptio consumption while, in practice, in the (2014) this long outcome and adjusts term. Although would it the calculation require athisfew to years. the expenditure is RSNTU JK&L*M"NOPL 11 realistic not survey. ' in the short term because it assumes that 11 y diminish with the reduction in while, tobacco practice, in consumption, outcome while, in would practice, require this outcome a few years. would 11 require a few years. ears.11 the effects of tobacco-related disease will immediately diminish with the reduction in tobacco consumption, ( EFH∗∆G IE)∗XSYN RPLTN.RSZT[[S \LUTNLM ]"YLTYLY ' Equation while, 5 shows this the income gains ∆ . associated . with the = reduction in medical expenses because of lower tobacco (5) Increasein inpractice, theEquation length of working 5 outcome shows life the: would The income impact require Increase in the length of working gains on a few incomes associated years. Increase from 11 with life:in the The the the length rise reduction impact in RSNTU JK&L*M"NOPL ofthetheon working number in incomes medical life of : The years ' expenses from impact of the rise employment in the number is fromof years rise of employmen consumption in the long term. Although the calculation is not realistic in short term because iton assumesincomes that the in the number then estimated. because In the of lower baseline, tobacco then the estimated. income consumption In lost the in is then because the baseline, long estimated. of term. the disabilityAlthough income In the or lost death the baseline, calculation because associated the of is income not disabilitywith tobacco lostbecauseor because death associated with tobac es from the the rise in effects of the number Equation tobacco-related of5years shows disease ofthe employment willincome immediately gains associated diminish with with thethe reduction reduction inin medical tobacco expenses consumption, of of lowerdisability tobacco or death Increase consumption in the length is of estimated working (equation life : The consumption impact 6). 12 The ison years incomes estimated lost arefrom (equation distributedthe rise 6). inacross The the number years each lost decile of years proportionately are distributed of employment to the is consumption is estimated (equation 6). shortacross each are decile proportionately that eachto ecause of in realistic disability or indeaththe shortassociated term because with itaassumes tobacco that the effects 12 of tobacco-related 12 The years disease lost distributed across de while, practice, consumption this outcome in the would long require term. Although few years. 11the calculation is not realistic in the term because it assumes then estimated. t are distributed will across In each the baseline, decile proportionatelythe income lost to the because of disability or death associated with tobacco immediately the effects diminish of tobacco-related with the reduction disease in willtobacco immediately consumption, diminish while, within the practice, reduction in tobacco consumption, consumption is estimated (equation 9 For a detailed discussion of the methodology, 9 For a detailed see 6). 12 discussion Coady The et al. years(2006)of the lost and 9 methodology, For are Kpodar distributed a detailed (2006).[51,52] see discussionCoady across ofal. et the each (2006) methodology, and decileKpodar seeproportionately (2006).[51,52] Coady et al. (2006) and to the Kpodar (2006).[51,52] 10 Another expressionthis outcome might while, Increase in the length of working life:Another be in would ∆ Expenditure practice, 10 require The impact = this ∆C∆P a outcome few expression + on ∆CP years. mightCincomes + would ∆PC 11 . be ∆ Expenditure C Another 10 require fromexpression the=rise a few ∆C∆P years. in+the might ∆CP 11 number C + ∆PCC . be ∆ Expenditure of years = ∆C∆P of+employment∆CPC + ∆PCC . is Kpodar (2006).[51,52] 11 Other studies have forecast the pass-through 11 Other between the decline inthe tobacco consumption and the effect on medical between expenditures. These estimates then 9 a detailed Foralso may estimated.discussion differentiate In ofthe the effect the baseline, methodology, associated with the see studies people income Coady who have etstop forecast al. (2006) lost consuming and because pass-through 11 Other Kpodar tobacco studies of versus disability between have (2006).[51,52] forecast people the who or decline the do death pass-through not in tobacco start associated because consumption of thewith the tax and decline policies. tobacco the in effect tobacco Because on consumption medical expenditures. and the effectTheseon estim me Increase in Increase the length in may the of length also working differentiate of working life: the The effect life : impact The associated may also impact on withincomes differentiate on people incomes from who the effectstop the from consuming associated rise the inwith rise the tobaccopeople in numberversus the who number people stop who consuming of do not years start because tobacco of of the tax versus people employment who policies. is do not Becstart bacco consumption of consumption and is 10 Another expression might data restrictions, estimated these the effect be assumptions ∆ Expenditure on medical (equation cannot of be expenditures. data = 6). in used∆C∆P12 restrictions, Thethis These+ ∆CP years paper. theseestimates+ ∆PC C assumptions lost C are of . datacannot distributed restrictions, be used thesein across this assumptionspaper. eachcannot decile used in this paper. to the be proportionately ng tobacco12 Other 11 Income studies versusispeopleof assumed have years who to do be of forecast then equal not employment the start to the pass-through because estimated. average 12 Income of the isis between In then consumption tax policies. the assumed the estimated. of decline baseline, to each Because be in equal 12In household tobacco the to Income the the baseline, per consumption income average decile. is assumed lost consumption the and to be the becauseincome equal effect of each to the on of lostmedical household disability average because expenditures. per decile. consumption of death or ofTheseeach estimates associated household with tobacco per decile. may also differentiate the effect associated with people who stop consuming tobacco versus people who do not start because of the tax policies. Because sehold of 9 For a detailed data per decile. these consumption disability discussion restrictions, ofor thedeath assumptions methodology, associated cannot isbe estimated see used Coadyinwiththis et al.(equation tobacco paper. (2006) andconsumption 6). Kpodar 12 The years (2006).[51,52] islostestimated are distributed (equation across 6). The 12 each decile proportionately to the 10 Income is 12 Another assumed to expression mightbe equal to the average consumption be ∆ Expenditure = ∆C∆P + ∆CP ofCeach + ∆PC household C. per decile. 11 Other studies years have forecast lost are the distributed pass-through between across the each decline decile in tobacco proportionately consumption and to thethe effect number on medical of house- expenditures. These estimates 8 9 For a detailed discussion of the methodology, see Coady et al. (2006) and Kpodar (2006).[51,52] may also differentiate holds the thateffect consumeassociated with tobacco, people who and stop the consuming income tobacco lost is versus estimated people who as do the not start because average income of the tax policies. Because of data restrictions, these 10 Another expression assumptions cannot be mightusedbe this paper. 8 = ∆C∆P + ∆CPC + ∆PCC . in ∆ Expenditure per household. 12 Income is assumed to be Other 11 equal to studies The the effect have average forecast of the consumption tax the pass-through of increase each household betweenis then per the estimated decline in tobacco decile. in relation consumption to the and income the effect on medical expenditures. 8 These estimates may also differentiate the effect associated with people who stop consuming tobacco versus people who do not start because of the tax policies. Because of data restrictions, these assumptions cannot be used in this paper. 12 Income is assumed to be equal to the average consumption of each household per decile. 8 10 Another expression might be ∆ Expenditure=∆C∆P+∆CP_0+∆PC_0. 11 Other studies have forecast the pass-through between the decline in tobacco consumption and the effect on medical expenditures. 8 These estimates may also differentiate the effect associated with people who stop consuming tobacco versus people who do not start because of the tax policies. Because of data restrictions, these assumptions cannot be used in this paper. 12 Income is assumed to be equal to the average consumption of each household per decile. 16 // Model number of households that of gains because consume increased years of and tobacco, the income employment. It islost is estimated expected as the will average that incomes fall asincome per household. The effect the numberof the tax increase of years is then lost increase estimated because in relation of higher to the numbers ofincome gains premature because deaths fromof increased years of employment. It is expected tobacco consumption. that incomes will fall as the number of years lost increase because of higher numbers of premature deaths from tobacco consumption. EFH∗∆G IE ∗pLTPY USYN GLP ]L["UL∗q*[SrLsSYY' Δ Proportional = (6) RSNTU JK&L*M"NOPL' Lastly, total welfare gains are estimated for each income group by adding the results of the reduction of medical Lastly, total welfare gains are estimated for each income group by adding the results treatments, the gains in working years, and the increase in tobacco expenditures (see equation 1). of the reduction of medical treatments, the gains in working years, and the increase in a. Elasticity tobacco expenditures (see equation 1). parameters After the model is defined for the calculation of the impact on income of the tobacco taxes, the estimates in the literature on elasticities, disease prevalence, the cost of medical treatments, and mortality patterns are a. Elasticity parameters examined. Several studies have estimated the tobacco price elasticity in Ukraine and other European countries. After the Within this research, model the defined workisof for the Denisova andcalculation Kuznetsova of the impact (2014) and on income of Krasovsky et the tobacco al. (2002) stands out for taxes, the estimates in the literature on elasticities, disease prevalence, the cost of their use of prices and quantities of tobacco in Ukraine. The parameters estimated by Krasovsky et al. (2002) medi- are an average cal treatments, price of mortality and elasticity −0.24, with patterns are examined. variations by income groupstudies Several have and age. estimated Denisova the and Kuznetsova (2014) generate lower and tobacco priceupper bounds elasticity for this in Ukraine elasticity and to simulate other European the impact countries. of tobacco Within price the this research, rises. work of Denisova and Kuznetsova (2014) and Krasovsky et al. (2002) stands out for their Table 1 shows theof use elasticities prices and estimated by quantities ofDenisova and tobacco in Kuznetsova Ukraine. (2014). Their The parameters elasticities estimated are divided into by Krasovsky 10 income deciles. The average elasticity is similar to Verguet et al. (2015) and to estimates on other countries. et al. (2002) are an average price elasticity of −0.24, with variations by income group and Table 1. age. Denisova Tobacco and Price Kuznetsova (2014) Elasticities, generate by Income lower Decile: and upper Denisova bounds and for this elasticity Kuznetsova (2014) (%) to simulate the impact of tobacco price rises. Price elasticity Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Average Lower bound −36 −34 −32 −27 −22 −15 −14 −10 −7 −3 −20.0 Medium bound Table 1 shows −44 the elasticities −42 −40 estimated −35 −by −23 and 30 Denisova −Kuznetsova 22 −18 (2014). −15Their elas- −11 −28.0 −56are divided Upper bound ticities −54 into −52 10 income−47 deciles. −35 −42 The average −34 elasticity 30 −similar is −27 to −23 Verguet et al. −40.0 Sources: Denisova and Kuznetsova 2014; (2015) and to estimates on other countries. b. Elasticity calculations To 1. Table Tobacco enhance Price the Elasticities, analysis, tobacco Income byprice Decile: are elasticities Denisova estimatedand using Kuznetsova (2014) (%) the household budget survey, 2010–13. The declared prices paid by households and the quantities of cigarettes purchased are used; this allows the Price elasticity of calculation a measure Decile 1 of effective Decile 2 Decile 3 price Decile 4 per cigarette, Decile 5 considering Decile 6 Decile 7 brand Decile pricing 8 variability. Decile 9 Decile 10 Figure 1 shows Average the variation Lower bound in −36 the price −34 of cigarettes −32 estimated −27 −22 using the −15 household −14 survey −10 and the −7 changes −3 in −20.0 using prices national statistical data on Ukraine. As expected, the variation in the prices paid by households reflect price Medium bound −44 −42 −40 −35 −30 −23 −22 −18 −15 −11 −28.0 variations between filtered and nonfiltered cigarettes, showing that households adjust consumption budgets as Upper bound −56 −54 −52 −47 −42 −35 −34 −30 −27 −23 −40.0 prices change. Source: Denisova and Kuznetsova 2014; 17 9 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine b. Elasticity calculations To enhance the analysis, tobacco price elasticities are estimated using the household budget survey, 2010–13. The declared prices paid by households and the quantities of cigarettes purchased are used; this allows the calculation of a measure of effective price per cigarette, considering brand pricing variability. Figure 1 shows the variation in the price of cigarettes estimated using the household survey and the changes in prices using national statistical data on Ukraine. As expected, the variation in the prices paid by house- holds reflect price variations between filtered and nonfiltered cigarettes, showing that households adjust consumption budgets as prices change. Figure 1. Tobacco Price Variation, 2008–13 CIGARETTE PRICES IN UKRAINE Household Budget Survey and Ukraine State Statistics; 2008-100 500 400 300 200 100 2008 2009 2010 2011 2012 2013 Year Self-reported Cigarette Price Domestic Filter Cigarette Non-filter Cigarette Once a measure of the price of cigarettes in Ukraine is obtained, the tobacco price elas- ticity across population groups is estimated (refer to Appendix I for more details). Table 2 shows the tobacco price elasticity across income deciles and other subgroups in the population: total population, household heads ages 25–40, and household location (rural versus urban). 18 // Model Table 2 - Tobacco Price Elasticities, by Income Decile, Age, and Location (%) Price elasticity Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Average Total population −59 −51 −52 −46 −44 −43 −42 −41 −36 −33 −45 Under age 40 −77 −69 −73 −70 −61 −64 −56 −64 −51 −51 −64 Rural −47 −30 −32 −30 −27 −27 −33 −26 −22 −21 −29 Urban −62 −65 −63 −55 −53 −52 −48 −50 −45 −41 −53 Source: Estimates based on data of the household budget survey, 2010–13. Estimates of the average tobacco price elasticity of −0.45 are higher than those calculated by Krasovsky et al. (2002) for Ukraine (-.25). To incorporate different assumptions of price elasticities, lower-bound and upper-bound elasticities are simulated. These estimates show differences between −0.2 and +0.2 relative to the previously estimated elastici- ties. The lower-bound elasticities are similar to Krasovsky et al. (2002) and tend to reflect income groups that typically do not change patterns of consumption, such as rural resi- dents or older population groups. These groups tend to exhibit less change in consump- tion when prices change. Figure 2. Tobacco Price Elasticities, Lower, Medium, and Upper Bounds (%) TOBACCO PRICE ELASTICITY Estimations Using Household Budget Survey, years 2010-2013 0 Price Elasticity of Demand -.2 -.4 -.6 -.8 0 2 4 6 8 10 Income Decile Lower Bound Elasticity Medium Bound Elasticity Upper Bound Elasticity 19 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine The upper-bound elasticity tends to reflect a longer-term scenario, echoing the effect the tobacco tax would have on younger smokers (table 3). After a few decades, these people will represent the majority of the population as older smokers die or quit smoking. The total average effect of the price increase over the long term would therefore be approxi- mated more closely by the upper-bound price elasticity. Table 3. Tobacco Price Elasticities, by Income Decile: Fuchs and Meneses (2017) (%) Price elasticity Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Average Lower bound −39 −31 −32 −26 −24 −23 −22 −21 −16 −13 −25 Medium bound −59 −51 −52 −46 −44 −43 −42 −41 −36 −33 −45 Upper bound −79 −71 −72 −66 −64 −63 −62 −61 −56 −53 −65 Source: Estimates based on data of the household budget survey, 2010–13 Once changes in tobacco expenditure due to price increases are calculated, we proceed to calculate the incidence of tobacco consumption on medical treatments, lost years of work, and related costs. Denisova and Kuznetsova (2014) estimate the incidence of tobacco on medical spending. As a secondary source of information, the Ukrainian gov- ernment and WHO have estimated the disease prevalence in each age-group and the monetary cost of treatment. This information allows a calculation of the average cost of medical treatment for each disease. The cost of lost years of employment can be esti- mated using the household consumption survey (see below). c. Mortality age patterns Early mortality related to tobacco consumption is analyzed to obtain the elasticities. Med- ical events associated with tobacco show a strong relationship with age and the number of years of smoking. Table 4 illustrates age patterns in tobacco-related deaths. Lost years of life are distributed among households as the using the proportion of households that have smokers, and the income (or wages) of this population segment is used to estimate the working years lost. 20 // Model Table 4. Tobacco-Related Deaths Age-group Ischemic Stroke Other COPD Other respira- Lung Other Other disease cardiovascular tory diseases cancer cancers causes diseases 0−1 - 23 50 - 49 - 143 4,299 1−4 - 1 13 - 112 - 62 739 5−14 - 8 20 - 171 4 25 851 15−24 128 66 182 14 438 9 138 4,930 25−34 718 288 964 43 1,251 54 704 13,528 35−54 13,472 5,585 7,056 2,244 13,989 691 2,757 42,064 55−74 118,364 37,654 12,794 9,586 43,274 5,553 2,132 33,808 75+ 181,987 57,724 27,992 2,289 15,306 6,441 767 24,653 TOTAL 314,669 101,349 49,071 14,176 74,590 12,752 6,728 124,872 Source: Calculations using data of Denisova and Kuznetsova 2014. Note: COPD = chronic obstructive pulmonary disease. d. Cost of treatment To estimate the medical costs of tobacco consumption, tobacco-related medical treat- ments and deaths are examined. The Ukrainian Department of Health Statistics and Information provides information on mortality related to tobacco consumption for 2013. As a second source of information, the total number of tobacco-related events are also obtained from Denisova and Kuznetsova (2014) (see table 4).13 Once the estimates of the number of deaths and the incidence of other events have been calculated, we investigate the medical treatments costs for each of these diseases. Table 5 shows estimates of the average cost of medical treatment for tobacco-related diseases in Ukraine. These expenses refer to the cost of treatment incurred by the state and excludes the costs to users. Although Ukraine has an extensive public health care system, not all medical costs are covered by the state. According to the National Health Accounts of Ukraine in 2015, out-of-pocket expenses accounted for 40.5 percent of all medical costs in Ukraine. This cost is covered directly by user expenditures. These costs usually consist of the inputs, medicines, and pharmaceutical components of the treatment. The estimate of the cost of treatment of these diseases therefore includes out-of-pocket expenditures calculated at 40.5 percent of the total expenditure. 13 Although determining the exact cause of each particular disease is not possible, the medical community has agreed on the prob- abilities of disease occurrence by age, gender, and tobacco consumption. The resulting tables of probabilities, along with tobacco incidence, are used to estimate the deaths related to tobacco. 21 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine Table 5. Cost of Medical Treatments Per Case Disease Tobacco-related Tobacco-related mor- Governmental Out-of-pocket morbidity, men, % bidity women, % cost, US$ cost, US$ Ischemic heart disease 24 2 584 389 Stroke 25 2 684 456 Other cardiovascular diseases 25 2 438 292 Respiratory diseases (COPD) 56 17 244 163 Lung cancer 91 27 633 422 Other cancers 44 3 633 422 Source: Calculations using data of Denisova and Kuznetsova 2014. For tuberculosis WHO 2004; Vasalal et al. 2008.[53] Note: US$1.00 = Hrv 8 for years 2013. COPD = chronic obstructive pulmonary disease Table 6 shows the annual medical costs of tobacco-related mortality. The results are taken from calculations based on the data in tables 4 and 5. After calculating the costs related to diseases, the cost of the treatment of illnesses that are related to tobacco consumption are calculated (table 7). Cardiovascular diseases, cardiovascular surgeries, other respiratory diseases (pneumonia and bronchitis), other cancers, and tuberculosis are included as major causes of the costs associated with medical care. Data on these costs have been collected based on various academic, national, and international sources. 22 // Model Table 6. Total Medical Cost of Tobacco-Related Cases Cost item Ischemic Stroke Other car- COPD Other Lung Other Total disease diovascular respiratory cancer cancers diseases Government cost (US$) 584 684 438 244 244 633 633 Out-of-pocket cost (US$) 389 456 292 163 163 422 422 Tobacco attributed Men, % 24 25 25 56 29 91 44 Women, % 2 2 2 17 3 27 3 Cases, men 32,720 10,069 5,476 6,643 10,766 7,902 2,207 75,782 Cases women 3,567 1,222 543 393 1,124 1,099 51 7,999 Total tobac- co-attributed cases 36,286 11,290 6,019 7,036 11,890 9,000 2,259 83,781 Total govern- ment cost (US$) 21,191,269.28 7,722,360.00 2,636,370.18 1,716,808.40 2,901,145.36 5,697,101.28 1,429,826.73 43,294,881 Total out-of- pocket cost (US$) 4,127,512.85 5,148,240.00 1,757,580.12 1,144,538.93 1,934,096.91 3,798,067.52 953,217.82 28,863,254 Note: COPD = chronic obstructive pulmonary disease. Table 7. Medical Costs of New Tobacco-Related Cases Indicator Cardio- Cardiovascular Other Other Tuberco- Total vascular surgeries respiratorya cancersb losis diseases Government cost (US$) 90 834 91 633 398 Out-of-pocket cost (US$) 60 556 61 422 265 Total cases 2,245,864 10,085 312,424 147,074 30,819 Tobacco attributed, % 29 29 25 26 9 Tobacco attributed cases 651,301 2,925 78,106 38,239 2,774 Total government cost (US$) 58,617,050 2,439,041 7,107,646 24,205,439 1,103,382 93,472,558 Total out-of- pocket cost (US$) 39,078,034 1,626,027 4,738,431 16,136,959 735,588 62,315,039 Sources: Cardiovascular surgeries, number of procedures: Sokolov et al. Register of percutaneous coronary interventions: a comparative analysis, reperfusion therapy in Ukraine, Survey PKV 2015. Journal Heart and vessels, 2015, issue 3: pages 7–29. Calculation of the average price of one stent was based on public information on the Ministry of Health website on the number of stents in 2017 and the total budget for this purpose (http://moz.gov.ua/ua/portal/pre_20170809_a.html). The market cost of the supplies kit for coronary angiography was taken from the price-list of private clinics, “Clinics of New Technologies,” based at the National Institute of Cardiovascular Surgery, M. M. Amosov, Academy of Medical Sciences of Ukraine, http://www.cnt-amosov.com.ua/index.ukr.php. Costs are taken from “Where the money goes and how to get more with scarce resources in the Ukraine’s health care: Report on findings of PETS/QSDS survey,” joint report of the World Bank, United Nations Development Program, and Kyiv School of Economics, 2017. Tuberculosis data taken from Vassall et al. 2008. a. Pneumonia and bronchitis. b. All other cancers, but lung cancer. 23 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine The total estimated out-of-pocket cost of tobacco is over US$90 million, similar to the esti- mates of the government costs calculated by Denisova and Kuznetsova (2014). e. Summary of descriptive statistics Table 8 summarizes the most important indicators, including total monthly expenditure from the household consumption survey and the incidence of household tobacco pur- chases. The cost of the medical treatment of tobacco-related diseases is estimated as a proportion of monthly income for each income decile. The share of income lost because of years of employment lost resulting from tobacco-related mortality is then estimated. Table 8. Baseline Descriptive Results, Household Survey 2013 Indicator Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Household expenditure (US$) 311 398 431 458 518 523 576 620 685 1282 Proportion tobaccoa 2.5 2.5 2.4 2.2 2.2 2.2 2.1 1.9 1.8 1.5 Households that smoke (%) 40 44 42 41 41 39 42 43 38 41 Woman-headed households (%) 55 52 51 56 54 57 58 57 54 57 Age, household head 51 50 51 53 52 53 54 54 53 51 Percentage HH with a Child 3-6 years of age 28 26 20 16 15 10 11 8 7 4 a. Proportional to total budget per household in each decile 24 // Model 25 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS 4 RESULTS The three scenarios in the tobacco price elasticity, lower bound, medium bound, and upper bound, are presented in table 3. These three scenarios allow an understanding of the ways results could change under different assumptions. a. Tobacco price increase As a first step, the income changes for each income decile arising from the increase in tobacco prices are estimated based on low-, medium-, and upper-bound elasticity. Using equation (4) and tables 3 and 8, one can calculate the effects of the tobacco price increase. For example, given the lower-bound elasticity (−0.39) in table 3, the proportion of tobacco expenditure among the first decile (2.5 percent) in table 8, and a price increase of 25 per- cent, there was increased expenditure of 0.32 percent. This represents a loss in welfare among consumers because they have to devote a higher proportion of their incomes to purchase the same amount of tobacco and reduce their consumption of other goods. The results for all income deciles and elasticity scenarios are shown in table 9. Table 9. Direct Effects of the Price Increase Because of Taxes (%) Price shock scenario Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Complete pass-through 0.62 0.62 0.60 0.56 0.56 0.55 0.53 0.48 0.45 0.37 Low-bound elasticity 0.32 0.38 0.36 0.38 0.39 0.39 0.39 0.35 0.36 0.31 Medium elasticity 0.16 0.22 0.21 0.24 0.25 0.25 0.25 0.23 0.25 0.22 Upper-bound elasticity 0.01 0.07 0.06 0.10 0.11 0.12 0.12 0.11 0.14 0.13 Sources: Proportion of household budget. Based on data of the 2013 household budget survey. Note: The table shows the share of total household budget for each decile. Complete pass-through refers to elasticity equal to zero; consumers pay all the increased prices. Across the three elasticities, the direct effect of the tobacco tax is a welfare loss, but in none of the cases does the shock seem to be regressive. In the low-, medium-, and upper-bound elasticity scenarios, the effect of the price increase is progressive, affecting higher-income groups in a higher proportion (figure 3). To show the effect of the elas- ticities on prices, table 9 includes the estimates of a complete pass-through scenario, whereby the increase in prices is completely passed to consumers without a reduction in consumption. Only in this case is the price shock regressive, affecting the lower-income deciles to a greater degree. 27 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine Figure 3. Direct Expenditure Effect: Direct Effect of Tobacco Taxes INCREASE IN EXPENDITURE: DIRECT EFFECT OF TAXES (Increased of expenditure due to tobacco tax) 0 Income Gains (%) -.2 -.4 -.6 0 2 4 6 8 10 Income Decile CI 95% Medium Bound Elasticity Direct pass through Lower Bound Elasticity Upper Bound Elasticity b. Medical expenses The yearly medical costs associated with tobacco consumption are estimated, assuming a direct medical impact on health. Although this assumption is unrealistic in the short run, the long-run reduction of tobacco consumption would tend to behave according to this pattern, whereby a reduction in tobacco consumption would be strongly related to a reduction in tobacco-related diseases and thus a reduction in medical costs (table 10). Health care expenditures are estimated using equation (5) and tables 3 and 8. Table 10 – Reduction in Medical Costs (%) Price shock scenario Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Low-bound elasticity 0.43 0.30 0.24 0.16 0.13 0.10 0.09 0.08 0.04 0.02 Medium elasticity 0.64 0.50 0.40 0.28 0.24 0.19 0.18 0.16 0.10 0.06 Upper-bound elasticity 0.86 0.70 0.55 0.41 0.35 0.28 0.26 0.24 0.16 0.10 Sources: Proportion of household budget. Based on data of the 2013 household budget survey. Note: The table shows the share of total household budget for each decile. 28 // Results The reduction in tobacco consumption would have a positive effect on income through reduced medical treatments. The saving in expenditure iwould vary between 0.43 and 0.02 percentage points of the household income in the case of the lower-bound elasticity assumption, between 0.64 and 0.06 percentage points in the case of the medium-bound elasticity, and between 0.86 and 0.1 percentage points in the case of the upper-bound elasticity (figure 4). These results show the importance of the elasticity assumptions; they also stress the relevance of the possible elasticity variations across income groups. Figure 4. Reduction in Expenditure: Because of the Reduction in Medical Expenditures REDUCTION IN EXPENDITURE: MEDICAL COSTS OF TOBACCO TAXES (Reduction of Medical Expenditures) .8 Income Gains (%) .6 .4 .2 0 0 2 4 6 8 10 Income Decile CI 95% Lower Bound Elasticity Medium Bound Elasticity Upper Bound Elasticity Source: Author estimation using a price shock of 25% c. Income gains because of increased years of employment The cost of working life lost because of tobacco consumption is estimated based on the assumption that there is a direct impact of lower tobacco use on health and thus work-generated income. The impact on the income of each income decile is calculated using the age pattern of mortality and estimating the years of life lost. The welfare effect is then estimated using the lower-, medium-, and upper-bound elasticity by decile variation. The 218,658 deaths attributed to tobacco consumption are distributed using the occur- rence of mortality profile.14 For each death, the number of potential years of work are calcu- lated, and the lost working years are divided across the deciles according to tobacco con- 14 Numbers base don 2013 data. 29 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine sumption. Using equation (6) and tables 3 and 8, one may calculate the impact of the tax increase on the increase in the years of employment. For example, in the first decile, assum- ing the upper-bound elasticity, the income increase would be 0.01 percent of income. Table 11 shows the results for all deciles using the three elasticity scenarios. Table 11. Years of Working Life Lost and Income Increase (%) Price shock Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 scenario Low-bound elasticity 0.0005 0.0004 0.0004 0.0003 0.0003 0.0003 0.0003 0.0003 0.0002 0.0002 Medium elasticity 0.0008 0.0007 0.0007 0.0006 0.0006 0.0005 0.0005 0.0005 0.0004 0.0004 Upper-bound elasticity 0.0010 0.0010 0.0010 0.0009 0.0008 0.0007 0.0008 0.0008 0.0007 0.0007 Sources: Proportion of household budget. Based on data of the 2013 household budget survey. Note: The table shows the share of total household budget for each decile. The results show that the reduction in tobacco consumption and the expected increase in years of potential work have positive impacts on welfare. In the first scenario, the gains are evenly distributed across income deciles. However, elasticities vary across deciles, generat- ing an important impact on lower-income groups (figure 5). Figure 5. Income Gains, Production during Years Lost INCOME GAINS: PRODUCTION DURING YEARS LOST (Production during years lost by income decile) .001 Income Gains (%) .0008 .0006 .0004 .0002 0 2 4 6 8 10 Income Decile CI 95% Lower Bound Elasticity Medium Bound Elasticity Upper Bound Elasticity Source: Author estimation using a price shock of 25% 30 // Results d. Net effects: total distributional impacts Once the effects of tobacco tax policy on prices, medical expenditures, and increased years of employment are calculated separately, one may examine the bigger picture. Based on a lower-bound elasticity, the results show a mixed effect of tobacco tax policy. The effect is progressive in that it has a smaller impact on the lower-income groups of the population relative to the higher-income groups, but the overall effect is negative (table 12; figure 6). Apparently, a population that is not as sensitive to tobacco price changes will not reduce consumption sufficiently to allow health and work benefits to offset cost increases. This is exactly what happens in Ukraine. Table 12. Total Net Effect (%) Price shock scenario Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Low-bound elasticity 0.11 −0.08 −0.12 −0.22 −0.26 −0.29 −0.29 −0.27 −0.32 −0.29 Medium elasticity 0.48 0.28 0.19 0.05 −0.01 −0.06 −0.08 −0.07 −0.15 −0.16 Upper-bound elasticity 0.86 0.63 0.49 0.31 0.24 0.16 0.14 0.12 0.02 −0.02 Sources: Proportion of household budget. Based on data of the 2013 household budget survey. Note: The table shows the share of total household budget for each decile. Figure 6. Total Direct and Indirect Expenditure Effect of Tobacco Taxes TOTAL INCOME EFFECT: DIRECT AND INDIRECT EFFECT OF TAXES (tobacco price increase, medical expenditure and working years gained) 1 Income Gains (%) .5 0 -.5 0 2 4 6 8 10 Income Decile CI 95% Lower Bound Elasticity Medium Bound Elasticity Upper Bound Elasticity Source: Author estimation using a price shock of 25% 31 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine In the case of a medium-bound elasticity, the tax has a progressive impact because of a positive effect on lower-income groups and a negative effect on higher-income groups. In the case of the upper-bound elasticity, the tax would have positive and progressive distributional effects, benefiting lower-income groups in larger proportion and having a negative effect on the highest income decile. 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. In conclusion, under a low tobacco price elasticity scenario, the overall effect is negative for all income groups. In the medium-bound scenario, the results are mixed, being posi- tive among lower-income groups and negative outcomes among higher-income groups. In the case of an upper-bound elasticity scenario, there would be income gains among the first eight deciles of the population, and a progressive pattern. Moreover, the impact would be particularly important among lower-income deciles. Furthermore, the assump- tions in this model do not include other possible policies, such as smoking cessation pro- grams, antismoking advertising, youth outreach, or policies financed through the new tax revenue. Therefore, these results are in line with the literature, showing the important role that taxation plays in lowering tobacco usage. 32 // Results 33 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS 5 DISCUSSION There has been extensive research on the negative effects of tobacco consumption on health and well-being, as well as the benefits of various public policy mechanisms aimed at reducing tobacco use. One of the most efficient ways to deter tobacco use is the imple- mentation of taxes, both ad valorem and specific excise taxes. However, questions remain regarding the net positive effect of these policies and whether tobacco tax increases end up hurting poorer people more, as they happen to be more likely to smoke, to have poor health and less access to insurance and adequate treatment. The question of regressive taxation is particularly important because the welfare effects derived from increased taxes heavily depend on the price elasticity of this item across different sectors of the population. Price elasticity will determine the magnitude of the income shock, as well as the benefits gained because of the reduction in tobacco consumption. Much of the net welfare gain occurs through the reduction in medical costs and the increase in potential working years associated with good health, an effect of lower levels of tobacco consumption. If a tax merely raise prices without reducing purchased quantity, it would fail to be an effective policy. Thus, it is critical to understand the effects of these sort of policies by determining the aggregate welfare gains or losses generated. A respon- sible and comprehensive policy analysis should focus on poorer groups because consump- tion taxes can be regressive and because the poor are also more likely to smoke. One of the main motivations of this paper is to weigh the main costs and benefits of tobacco taxation to determine if, in the end, the policy is regressive or not. Results show that -when considered by itself- a price increase on tobacco through higher taxes would lead to slight tobacco expenditure increases across all population groups simply because of the higher price effect. This effect is more accentuated under the low- er-bound elasticity scenario and more moderate as elasticity increases in absolute terms. Conversely, a more comprehensive approach -including benefits through lower medical expenses and an increase in potential working years- the short-term tax burden is more than compensated. The tax increase shows a progressive pattern in all cases, though the absolute benefits vary. The reduction in medical expenses is the main driver of the increase in net incomes because of the reduction in tobacco-related problems, which require expensive treatments. In all three scenarios based on elasticity, the benefits of the reduced medical costs are greater, particularly among lower-income groups. This is because of the lower income. Assuming that medical expenses are constant across all population groups, the income increases are less among lower-income groups. 35 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine The various elasticity assumptions produce three distinct set of results. The lower-bound elasticity creates losses among most income groups, but follows a somewhat progressive pattern. The medium-bound and upper-bound elasticities lead to income gains among lower-income groups, but losses among higher-income groups. Thus, the effect of tobacco taxes on the various income groups is exacerbated if the variation in price elasticities across income deciles is greater. In all three scenarios, the taxes have a progressive effect on income distribution. The three price elasticity scenarios mimic the short- versus the long-term effects of a tobacco tax. There is evidence that adult smokers will only present small changes in their behavior if faced with price increases; the lower-bound elasticity is likely to measure this situation. In contrast, younger people usually show more elastic demand, that is, demand that is similar to the upper-bound elasticity. After a few decades, one may expect the impact of the tax policy to resemble the upper-bound elasticity scenario, as young people replace older groups in the population. The results provide evidence that support possible preservation or increase in tobacco taxes. The analysis also shows the importance of tobacco price elasticity in assessing the potential effect of this type of public policy. Specifically, this paper suggests that taxation, especially in the short run when price elasticity is lower, should be accompanied by other policies to deter smoking, such as smoking cessation programs or mass media campaigns on the negative effects of tobacco use. Because the effect of and reaction to price changes differ across income groups, specific policies should be targeted at different groups, but focus on low-income households. For example, smoking cessation programs, which tend to be expensive and less accessible to those with fewer resources, could be made more accessible to lower-income groups, along with targeted advertising specifically adapted to different sociocultural contexts. Overall, an integrated policy approach that involves coordi- nation between taxation and behavior change may be the most effective, especially in the short term, while price elasticities are still low. Further research should focus on which combination of public policies is most (cost) effective across income and age-groups, given that price elasticities differ across the population. Future research on the new Ukrainian tobacco tax should also allow for analysis on how price elasticities change in the real world. 36 // Discussion REFERENCES 1 Doll R, Hill AB. Lung cancer and other causes of death in relation to smoking. British medical journal 1956;2:1071. 2 Wynder EL, Graham EA. 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 1950;143:329–336. 3 General S. The health consequences of smoking—50 years of progress: a report of the surgeon general. In: US Department of Health and Human Services. Citeseer 2014. 4 World Health Organization. Tobacco and its environmental impact: an overview. 2017. 5 World Health Organization. WHO report on the global tobacco epidemic 2015: raising taxes on tobacco. World Health Organization 2015. 6 Eriksen M, Mackay J, Ross H, et al. The tobacco atlas. American Cancer Society 2013. 7 World Health Organization. WHO report on the global tobacco epidemic 2017. 8 The World Bank. Curbing the epidemic: governments and the economics of tobacco control. Development in Practice Series 1999;8:196. 9 Chaloupka FJ, Cummings KM, Morley CP, et al. Tax, price and cigarette smoking: evidence from the tobacco documents and implications for tobacco company marketing strategies. Tobacco Control 2002;11:i62–i72. 10 Debrott Sánchez D, others. Economía del Control del Tabaco en los países del MERCOSUR y Estados Asociados: Chile. Organización Panamericana de la Salud 2006. 11 Denisova I, Kuznetsova P. The effects of tobacco taxes on health: An analysis of the effects by income quintile and gender in Kazakhstan, the Russian Federation, and Ukraine. Published Online First: 2014 12 Verguet S, Gauvreau CL, Mishra S, et al. The consequences of tobacco tax on household health and finances in rich and poor smokers in China: an extended cost-effectiveness analysis. The Lancet Global Health 2015;3:e206–e216. 13 Peto R, Lopez AD. The future worldwide health effects of current smoking patterns. Tobacco and public health: Science and policy 2004;:281–286. 14 Jha P, Peto R. Global effects of smoking, of quitting, and of taxing tobacco. New England Journal of Medicine 2014;370:60–68. 15 Eriksen M, Mackay J, Ross H, et al. The tobacco atlas. American Cancer Society 2013. 16 Health UD of, Services H, others. The health consequences of smoking: a report of the Surgeon General. 2004. 37 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine 17 Cromwell J, Bartosch WJ, Fiore MC, et al. Cost-effectiveness of the clinical practice recommend- ations in the AHCPR guideline for smoking cessation. Jama 1997;278:1759–1766. 18 Borland R. Tobacco health warnings and smoking-related cognitions and behaviours. Addiction 1997;92:1427–1435. 19 Fathelrahman AI, Omar M, Awang R, et al. Smokers’ responses toward cigarette pack warn ing labels in predicting quit intention, stage of change, and self-efficacy. Nicotine & Tobacco Research 2009;11:248–253. 20 Levy D, de Almeida LM, Szklo A. The Brazil SimSmoke policy simulation model: the effect of strong tobacco control policies on smoking prevalence and smoking-attributable deaths in a middle income nation. PLoS medicine 2012;9:e1001336. 21 Durkin S, Brennan E, Wakefield M. Mass media campaigns to promote smoking cessation among adults: an integrative review. Tobacco control 2012;21:127–138. 22 World Health Organization, Ministry of Health in Ukraine. Tobacco Control in Ukraine. 2009. 23 Krasovsky K. Sharp changes in tobacco products affordability and the dynamics of smoking prevalence in various social and income groups in Ukraine in 2008–2012. Tobacco induced diseases 2013;11:21. 24 Konstantin Krasovsky, Tatiana Andreeva, Alla Grygorenko, et al. Tobacco Control in Ukraine. Second National Report. 25 U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, et al. The Economics of Tobacco Control. 2016. 26 Ramos A. Economía del control del tabaco en los países del Mercosur y Estados Asociados: Uruguay. (Published Online First: 2006). 27 World Health Organization, others. WHO report on the global tobacco epidemic, 2008: the MPOWER package. 2008. 28 International Agency for Research on Cancer, World Health Organization. Evaluating the effective ness of smoke-free policies. IARC Press, International Agency for Research on Cancer, 2009. 2009. 29 Levy D, Rodríguez-Buño RL, Hu T-W, et al. The potential effects of tobacco control in China: projections from the China SimSmoke simulation model. Bmj 2014;348:g1134. 30 World Health Organization. Guidelines for the Implementation of Article 6 of the WHO FCTC. 2013. 31 Guindon GE. The impact of tobacco prices on smoking onset: a methodological review. Tobacco control 2013. 32 Chaloupka FJ, Grossman M. Price, tobacco control policies and youth smoking. National Bureau of Economic Research 1996. 38 // Discussion 33 Lewit EM, Coate D. The potential for using excise taxes to reduce smoking. Journal of health economics 1982;1:121–145. 34 Chaloupka F. Rational addictive behavior and cigarette smoking. Journal of political Economy 1991;99:722–742. 35 Gallus S, Schiaffino A, La Vecchia C, et al. Price and cigarette consumption in Europe. Tobacco control 2006;15:114–119. 36 Townsend J, Roderick P, Cooper J. Cigarette smoking by socioeconomic group, sex, and age: effects of price, income, and health publicity. Bmj 1994;309:923–927. 37 Szilágyi T. Higher cigarette taxes–healthier people, wealthier state: the Hungarian experience. Central European journal of public health 2007;15. 38 Yürekli A, Önder Z, Elibol M, et al. The economics of tobacco and tobacco taxation in Turkey. Paris: International Union Against Tuberculosis and Lung Disease 2010;5. 39 Ciecierski CC, Cherukupalli R, Weresa MA. The economics of tobacco and tobacco taxation in Poland. Paris: International Union Against Tuberculosis and Lung Disease 2011;:8–11. 40 Selvaraj S, Srivastava S, Karan A. Price elasticity of tobacco products among economic classes in India, 2011–2012. BMJ open 2015. 41 Van Walbeek CP. The distributional impact of tobacco excise increases. South African Journal of Economics 2002;70:258–267. 42 Lee J-M. Effect of a large increase in cigarette tax on cigarette consumption: an empirical analy sis of cross-sectional survey data. Public health 2008;122:1061–1067. 43 Centers for Disease Control and Prevention (CDC. (1998). Response to increases in cigarette prices by race/ethnicity, income, and age groups--United States, 1976-1993. MMWR. Morbidity and mortality weekly report, 47(29), 605. 44 Xu X, Bishop EE, Kennedy SM, et al. Annual healthcare spending attributable to cigarette smoking: an update. American journal of preventive medicine 2015;48:326–333. 45 World Health Organization. WHO technical manual on tobacco tax administration. World Health Organization 2010. 46 Lightwood J, Collins D, Lapsley H, et al. 4 Estimating the costs of tobacco use. Published Online First: 2000. 47 Pichón Riviere A, Bardach A, Caporale J, et al. Carga de Enfermedad atribuible al Tabaquismo en Chile. Documento Técnico IECS 2014. 48 Fuchs Tarlovsky A, Meneses FJ, others. Are tobacco taxes really regressive? evidence from Chile. The World Bank 2017. 49 Mason J, Wheeler W, Brown MJ. The economic burden of exposure to secondhand smoke for child and adult never smokers residing in US public housing. Public Health Reports 2015;130:230–244. 39 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine 50 Pichón Riviere A, Bardach A, Caporale J, et al. Carga de Enfermedad atribuible al Tabaquismo en Chile. Documento Técnico IECS 2014. 51 Coady DP, El Said M, Gillingham R, et al. The magnitude and distribution of fuel subsidies: evidence from Bolivia, Ghana, Jordan, Mali, and Sri Lanka. Published Online First: 2006. 52 Kpodar K, Djiofack C. The Distributional effects of oil price changes on household income: Evidence from Mali. Journal of African Economies 2009;19:205–236. 53 Vassall A, Chechulin Y, Raykhert I, et al. Reforming tuberculosis control in Ukraine: results of pilot projects and implications for the national scale-up of DOTS. Health policy and planning 2008;24:55–62. 54 Meng Y, Brennan A, Purshouse R, et al. Estimation of own and cross price elasticities of alcohol demand in the UK—a pseudo-panel approach using the Living Costs and Food Survey 2001– 2009. Journal of health economics 2014;34:96–103. 55 Joyce Dargay, Petros Vythoulkas. Estimation of a Dynamic Car Ownership Model: A Pseudo-Panel Approach. Journal of Transport Economics and Policy 1999;33:287–301. 56 Angus Deaton. Panel data from time series of cross-sections. Journal of Econometrics 1985;30:109–26. 57 Seth H. Giertz. Panel Data Techniques and the Elasticity of Taxable Income. 2008. 58 Marno Verbeek, Francis Vella. Estimating dynamic models from repeated cross-sections - Google Scholar. Journal of Econometrics 2005;127:83–102. 40 // Discussion 41 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOBA TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKIN TOBACCO CONS APPENDIX I Elasticity Estimation: To estimate the price elasticity of tobacco, per income decile. We obtain four years of household consumption surveys for Ukraine. We use the Household Appendix I Budget survey from Ukraine for the years 2012, 2013, 2014 and 2015. There methodology used is a repeated cross section analysis, and as the survey is nationally representative at Elasticity Estimation: decile level, To estimate it could the price be used elasticity of tobacco, as a replacement of a panelper income data decile. set, under theWe obtain four years of assumption household consumption surveys for a Ukraine. We use the Household Budget survey from Ukraine for the that income deciles are comparable during each year and represent the same group of the years 2012, 2013, 2014 and 2015. There methodology used is a repeated cross section analysis, and as the population. These data sets have been translated and standardized by the World Bank. survey is nationally representative at decile level, it could be used as a replacement of a panel data set, under the assumption For all data that we calculate setsdeciles income and check during are comparable each consumption, the tobacco income year and represent thelevel sameandgroup of the population. These data income calculate sets have beenand deciles effective translated andprice paid per quantity. standardized The four by the World surveys Bank. Forare allmerged data sets we calculate and check the tobacco consumption, income level and calculate income deciles and effective into a single data set to estimate a single regression. Using official statistics, we obtain the price paid per quantity. fourof The rate inflation surveys are merged the country into for these a single years data set and deflate to estimate prices, to make singleall a them regression. comparable.Using official statistics, we obtain the inflation rate of the country for these years and deflate prices, to make them all comparable. In the surveys for Ukraine we have quantity consumed and total price paid per package. Therefore, we estimate the effective price paid per package, in each purchase, in each for Ukraine In the surveyshousehold. Wewe have quantity eliminate outlierconsumed andstandard that are three total price paid perfrom deviations package. Therefore, the mean, underwe estimate the effective price paid the assumption thatper package, these in each purchases purchase, tend to reflectindata each household. problems. WeWe eliminate estimate outlier that the price are three standard deviations elasticity of demandfromof the underthe mean, using tobacco assumption thefollowing that these purchases tend to reflect data equation: problems. We estimate the price elasticity of demand of tobacco using the following equation: EC Ln( ) . = E " ∗ " *Ln(price) (7) We estimate the lower un upper bound and estimate the reductions in consumption. To estimate the lower We estimate the lower un upper bound and estimate the reductions in consumption. To bound we usually look for assumptions on would could be the price elasticity of rural or population that is estimate highly addicted the lower to tobacco. For thebound upperwebound, usuallywelook for assumptions look on would for the elasticity could groups be the price of younger of the elasticity of rural or population that is highly addicted to tobacco. For the upper bound, population, that may reflect the longer-term scenario. In the case of Ukraine, the difference among these groups then towebelook around 0.16 for the -0.19 in elasticity absolute of groups Therefore youngervalues. we use a of the population, 0.2may that band for our reflect thelower lon- bound an upper bound scenarios. ger-term scenario. In the case of Ukraine, the difference among these groups then to be around 0.16 -0.19 in absolute values. Therefore we use a 0.2 band for our lower bound an upper Table A1 bound scenarios. - Tobacco Price Elasticities, by Income Decile, Age, and Location (%) Price elasticity Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Average Total population −59 −51 −52 −46 −44 −43 −42 −41 −36 −33 −45 Under age 40 −77 −69 −73 −70 −61 −64 −56 −64 −51 −51 −64 Rural −47 −30 −32 −30 −27 −27 −33 −26 −22 −21 −29 Urban −62 −65 −63 −55 −53 −52 −48 −50 −45 −41 −53 Source: Estimates based on data of the household budget survey, 2010–13. Repeated cross section estimation of elasticities versus panel data has been discussed in the literature, and using subgroups of the population, and following them, has been recalled as pseudo-panel data approach. These methodology has been used to estimate price elasticities for products like alcohol or cars.[54,55]. For a comparison of panel data and cross section techniques please review Deaton 1985, Giertz 2008 and Verbeek and Bella 2005. [56–58] 43 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine Table A1 - Tobacco Price Elasticities, by Income Decile, Age, and Location (%) Price Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile Average elasticity 10 Total population −59 −51 −52 −46 −44 −43 −42 −41 −36 −33 −45 Under age 40 −77 −69 −73 −70 −61 −64 −56 −64 −51 −51 −64 Rural −47 −30 −32 −30 −27 −27 −33 −26 −22 −21 −29 Urban −62 −65 −63 −55 −53 −52 −48 −50 −45 −41 −53 Source: Estimates based on data of the household budget survey, 2010–13. Repeated cross section estimation of elasticities versus panel data has been discussed in the literature, and using subgroups of the population, and following them, has been recalled as pseudo-panel data approach. These methodology has been used to estimate price elasticities for products like alcohol or cars.[54,55]. For a comparison of panel data and cross section techniques please review Deaton 1985, Giertz 2008 and Verbeek and Bella 2005. [56–58] 44 // Appendix 1 APPENDIX II We estimate the model for four specific groups of the population: rural, urban, population under 40 years old, and over 40 years old. For the elasticities estimated for these popula- tions (in Table A1), the mean results are as follows. We can see that that for most income groups of the rural population, the effect of the tobacco tax is negative, but the effect is still progressive. On the other hand, for urban as well as for households with heads of young age, the effects tend to be positive as well as progressive. Table A2 - Total Net Effect (%), by Income Decile, Age, and Location (%) Price elasticity Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Rural 25.9 -9.3 -11.8 -16.2 -22.0 -24.2 -17.3 -22.2 -26.5 -23.4 Urban 53.9 52.1 35.2 16.7 10.3 4.0 -1.0 1.5 -7.1 -10.3 Under age 40 81.9 59.1 50.3 36.4 20.3 17.5 7.7 15.4 -2.1 -3.8 Source: Proportion of household budget. Based on data of the 2013 household budget survey. Note: The table shows the share of total consumption for each decile. 45 THIS PAPER DESC AND QUANTIFIES EFFECTS OF TOB TAX INCREASES O AGGREGATE HOU WELFARE THROU THREE CHANNEL IS AMPLE AND RO EVIDENCE LINKI TOBACCO CONS