POVERTY & EQUITY NOTES JULY 2020 · NUMBER 25 Increasing Taxes on Tobacco in Low and Middle-Income Countries: Hurting or Saving the Poor? Alan Fuchs Tarlovsky and María Fernanda González Icaza Policy makers hesitate to increase tobacco taxes over concerns about taxes being regressive and potentially increasing poverty and inequality. This note summarizes a set of studies of the effects of raising tobacco taxes in 11 low and middle-income countries using an extended cost-benefit analysis (ECBA) and harmonized national household budget survey data and introduces the TOBACTAX Tool. The studies find that demand price elasticities for tobacco products are larger among lower-income households and that the poor receive the largest long-term gains from tobacco taxation. Tobacco taxes have progressive long-term effects due to lower medical expenses and added years of productive life, which contribute to poverty reduction in most countries studied. TOBACTAX Tool can help replicate such analyses elsewhere. Tobacco is a major public health priority (WHO Evidence on the effects of taxing tobacco in 1999). Smoking is the second leading cause of death and developing countries remains scant. Nearly 80 percent disability worldwide, killing over 8 million people each of the world’s smokers live in low and middle-income year (Ng et al. 2014; WHO 2017), and triggering an array countries. Yet, to the authors’ knowledge, this is the first of policy initiatives. Tobacco taxation is considered the cross-country analysis on the distributional welfare most efficient intervention (World Bank 1999): by effects of taxing tobacco based on harmonized boosting prices, taxes induce smokers to quit and household consumption and other international data. discourage potential consumers. Evidence suggests that The sample of 11 countries in this study represents over higher taxes are responsible for almost half the decline 860 million people, or 15 percent of smokers worldwide. in smoking worldwide (WHO 2014). The retail prices of tobacco products represent only Methodology a fraction of overall costs from smoking to societies. In addition to health consequences, smoking cuts The ECBA 1 Equation 1 quantifies medium to long-terms earnings potential and labor productivity, and lowers benefits as tobacco taxes rise, based on 3 effects. human capital accumulation and economic growth (WHO 2015a). Goodchild, Nargis, and Tursan d’Espaignet Equation 1: ECBA Conceptual Framework (2018) estimate that tobacco-related diseases accounted for 5.7 percent of global health expenditure in 2012 and Net Change in Change in Change in years that total economic costs of smoking, including health Income expenditures and productivity losses, were equivalent to Effect tobacco = expenditure + medical expenses + oflife working lost 1.8 percent of the world’s gross domestic product (GDP). (A) (B) (C) Yet, policy makers hesitate to raise tobacco taxes The immediate effect of raising taxes on cigarettes because of concerns that poorer households may face increases household tobacco expenditures (A). 2 Taxes higher costs. discourage smoking and health outcomes and thus improve, at least 2 indirect benefits to households: changes in direct medical expenditures (B) the Global Burden of Disease (GBD) Project. For the (hospitalization, pharmaceuticals, medical equipment: shares of total health expenses financed through out-of- Lightwood et al. 2001), and improvement in years of pocket health expenditures, we used the World 3 working life lost (C). Development Indicators (WDI) database. Table 1: Share of cigarettes in household expenditures, Data by decile (%) Decile This study includes 11 countries: Bangladesh, Bosnia 1 2 3 4 5 6 7 8 9 10 and Herzegovina, Chile, Georgia, Indonesia, Bangladesh 5.4 5.1 4.9 4.8 4.6 4.5 4.3 4.2 4.1 3.6 Moldova, Mexico, South Africa, the Russian Bosnia and 8.1 8.0 7.9 7.9 7.8 7.8 8.3 7.4 7.2 7.0 Federation, Ukraine, and Vietnam. These countries Herzegovina vary widely in income, sociodemographic characteristics, Chile 4.7 4.1 3.9 3.2 2.8 3.2 2.8 2.4 1.9 1.5 and tobacco consumption patterns. They range from Georgia 14.0 12.5 11.5 10.8 10.3 10.3 9.9 8.7 8.1 6.8 lower-middle-income countries (Bangladesh has the Indonesia 11.6 12.7 12.8 13.0 13.1 13.2 12.7 11.9 10.9 8.8 lowest GDP per capita) to upper-middle (Mexico, Russia) and high-income countries (Chile). The main source of Mexico 10.1 5.3 5.6 5.0 4.6 4.6 4.0 4.3 3.6 2.8 data are national household budget surveys in each Moldova 5.4 5.3 4.9 5.1 4.7 4.5 5.9 4.8 5.3 4.9 country. For several countries, the study also used Russian 5.6 2.7 2.4 2.0 2.1 1.8 1.5 1.5 1.4 1.0 4 Federation datasets and work of other researchers and studies. If South Africa 4.5 4.1 4.2 3.8 4.6 4.5 5.1 4.4 3.3 1.8 available, we used surveys with nationally representative data for 2016. Otherwise, we used the most recent Ukraine 7.4 5.5 5.0 5.2 5.6 4.9 4.1 3.8 4.2 3.8 publicly available datasets. Vietnam 1.8 1.7 1.6 1.5 1.4 1.3 1.3 1.1 1.0 0.7 Sources: National budget surveys, circa 2016. In most countries, poorer or middle-income Note: Shares are conditional on household reporting spending on households spend a larger share of their budgets on cigarettes, and relative to the harmonized household consumption tobacco. In half the sample (Bosnia and Herzegovina, aggregate. Deciles based on per capita consumption aggregate. Georgia, Mexico, Moldova, Russia, and South Africa), people in the wealthiest decile are most likely to smoke. However, lower-income households allocate larger Tobacco Demand Price Elasticity shares of consumption to tobacco. Except in Bosnia and Herzegovina, the richest 10 percent consistently allocates the smallest share. Indonesia and Vietnam Evidence suggests that price elasticities of demand exhibit the highest smoking prevalence rates in the for tobacco may be higher in lower-income settings sample; on average, 64 percent of households in (World Bank 1999). The distributional and welfare Indonesia and 58 percent of households in Vietnam effects of tobacco taxation ultimately depend on the consume tobacco. responsiveness of consumers to changes in tobacco prices (WHO 2011a). An equivalent price change of 10 To maintain data consistency across countries, the percent in low and middle-income countries would likely analysis adapted direct medical expenditures from the result in a 6 percent average fall in demand (IARC 2011). calculations of Goodchild, Nargis, and Tursan Less dependence on nicotine, larger peer effects, and d’Espaignet (2018), who used data on the cost of illness limited disposable income may also make younger to estimate the direct medical costs of smoking- people less likely to buy tobacco when prices increase attributable diseases in 2012. We collected data on (Jha and Peto 2014). mortality, years of life lost, and morbidity for 2016 from July 2020 · Number 25 2 Figure 1: Illustrates the estimated price elasticities of • Reduce medical expenditures across all deciles demand for cigarettes by decile and countries • Lowers workers’ mortality modestly in all deciles 0.0 and countries -0.2 • Adds working years • Reduce poverty -0.4 Doubling the price of tobacco could lower the total -0.6 number of extreme poor by 1.1 to 2.0 million, equivalent -0.8 to lifting 1.7 percent to 2.0 percent of the poor out of extreme poverty. -1.0 -1.2 Table 2 approximates the changes in the number of poor 1 2 3 4 5 6 7 8 9 10 in each country if the price of cigarettes were increased Decile by 100 percent. Bangladesh Bosnia and Herzegovina Chile Georgia Indonesia Moldova Table 2: Simulated changes in number of poor Mexico Russian Federation Ukraine Vietnam Country Including all Including out-of-pocket South Africa tobacco-related tobacco-related Sources: Estimates based on national socioeconomic surveys. medical expenses medical expenses Elasticities in Chile and Ukraine have been simulated based on Fuchs Bangladesh −559,680 −439,974 and Del Carmen 2018; Fuchs and Meneses 2017a. See the Bosnia and Herz. 0 0 supplemental material for details. Chile 0 0 Note: In most cases, a multiple time cross-section model with time Georgia −4,697 −2,183 fixed effects has been used. Demographic controls include the age, Indonesia −987,656 −483,317 education, and gender of the household head, the share of individuals Moldova 0 0 by age-group in each household, and urban status. Deciles are based Mexico −135,942 −103,030 on the harmonized per capita household consumption aggregate. Russian Fed. −101,803 −39,262 South Africa −149,426 12,098 Consistent with previous empirical results, the Ukraine 0 0 Vietnam −97,006 −19,494 analysis finds elasticities to be greater in lower- Total −2,036,209 −1,075,163 income deciles. Demand responses are inelastic in all Source: Simulations based on national budget surveys, circa 2016. countries (absolute values below 1). 5 However, Note: Negative numbers represent a reduction in the estimated responsiveness is greater in lower-income households. number of the poor. Positive numbers represent a rise. All simulations have been calculated based on the harmonized per capita consumption Elasticity estimates range from −1.05 among the poorest aggregate and the international extreme poverty line of US$1.90 per 10 percent of the population in Bangladesh to −0.05 person per day (PPP 2011 dollars). among the wealthiest decile in Bosnia and Herzegovina. Conclusions Study Results There is reason for optimism about the long-term We assume a tax increase that raises the prices of economic and health benefits of increasing tobacco tobacco products by 100 percent in each country. While taxes. While households will likely experience short-term different patterns emerge across countries, our main welfare losses, in addition to health benefits, taxes on findings are that higher tobacco taxes: tobacco generate medium and long-term economic • Decrease household disposable income, without benefits, including higher labor productivity and fewer behavioral adjustments July 2020 · Number 25 3 medical bills, which offset short-term household welfare decile, and it assesses the impacts of increasing taxes losses. on tobacco on household income and distributional outcomes. The analysis found net benefits for 80 to 90 percent of the population analyzed. Despite differences in the The TOBACTAX tool can be accessed here. magnitude and distribution of the net welfare effects across countries, the distributional impact of raising taxes on cigarettes is generally progressive: across Next Steps countries, lower-income households capture the largest relative benefits. Extending the comparative ECBA study to other countries requires more data sources. Including Introducing TOBACTAX Tool for Global Analysis countries with the highest numbers of smokers—such as China and India— will add to the understanding of the global implications of raising taxes on tobacco for long- Comprehensive policy strategies are needed term welfare and distributional outcomes. related to tobacco usage and taxation. Inducing changes in tobacco consumption that translate into ABOUT THE AUTHORS net social gains requires country-specific policies that Alan Fuchs Tarlovsky is a Senior Economist in the meet that country’s consumer responses, especially Poverty and Equity Global Practice and a Global among youth and at-risk groups. Lead of the Fiscal and Social Policy Global Solutions Group at the World Bank. The Poverty and Equity Global Practice has María Fernanda González Icaza is a Consultant in developed a user-friendly tool—TOBACTAX—to the Poverty and Equity Global Practice. aid policy practitioners and academics in estimating the welfare and distributional impacts of implementing tobacco tax reforms. TOBACTAX contributes to global knowledge by applying the ECBA methodology to any country with available household survey data and minimal data requirements. Using the STATA software, TOBACTAX estimates the price-elasticities of tobacco by income 1 This study relies on the ECBA methodology adapted from previous applications by Fuchs and Meneses (2017a, 2017b, 2018), Pichón-Riviere et al. (2014), and Verguet et al. (2015). 2 Secondhand smoke exposure is excluded from ECBA. The income gains that households may derive from potential increases in public expenditure linked to greater revenue collection. Potential earmarking of tax revenues are also excluded. 3 There are several other benefits from cessation of smoking on productivity and human capital endowments and returns not captured in the ECBA. Labor and productivity costs of tobacco include reduced output per worker because of illness on the job, tobacco-related work absenteeism, and forgone labor force participation to care for family members sick from smoking. 4 Fuchs and Del Carmen 2018; Fuchs, Del Carmen, and Genoni 2018; Fuchs, Del Carmen, and Mukong 2018; Fuchs, Matytsin, and Obukhova 2018; Fuchs and Meneses 2017a, 2017b, 2018; Fuchs, Orlic, and Cancho 2019. 5 With the exception of the first decile in Bangladesh. This note series is intended to summarize good practices and key policy findings on Poverty-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank, its board or its member countries. Available for download at the World Bank Publications, Documents & Reports site. July 2020 · Number 25 4