KN#4. Unpacking the empirics behind health tax revenue1 The purpose of this note is to provide policy makers with an overview of the revenue potential for health taxes, including from increases in tax rates and reforms of tax structures. It is one in a series of knowledge notes responding to specific questions around health taxes and key issues raised during health tax reforms. SUMMARY Introducing, reforming and increasing health taxes on tobacco, alcohol, and sugar-sweetened beverages (SSBs) can generate meaningful increases in tax revenues while improving health outcomes over time. Using a novel health tax revenue database that draws on publicly available sources as well as country data, our analysis shows that tobacco and alcohol excise taxes generate an average of 0.6 and 0.3 percent of GDP in tax revenue, respectively. SSB taxes generate significantly less revenue than tobacco and alcohol. Health tax revenues vary widely between countries, affected by tax design including tax structures and tax rates, tax administration and baseline levels of consumption. However, there is no significant variation between high-income and low- and middle-income countries, highlighting the usefulness and importance of health taxes to all countries. Reforms of health taxes, including reforming tax structures and raising tax rates can contribute to significant increases in revenue very rapidly. Tax increases can be sustained over long periods, generating additional tax revenues over long periods of time. Even for countries with relatively high tax rates, tax increases still generate increases in tax revenues when these rates are increased regularly. Health taxes are efficient since they are relatively easy to implement, can generate revenue quickly, and limited distortion to general economic activities. Health taxes are also fiscally progressive once the long-run behavioral effects of reduced consumption, reduced medical expenses, and increased labor productivity are accounted for. The expenditures that are funded by increases in general revenue because of health taxes can further increase the progressive distributional impact of health taxes. WHY FOCUS ON HEALTH TAXES? Health taxes are excise taxes applied to products that account for the underestimation of harm and difficulty of generate negative externalities and internalities, quitting in later life. In the case of SSB taxation, negative most commonly tobacco, alcohol, and SSBs.2 Negative internalities help make a strong case. A growing literature externalities are the costs that accrue to society at large, supports the taxation of SSBs by showing the significant including non-users of these products. High taxes on negative internalities their consumption generates (see tobacco and alcohol may be rationalized to reduce the Alcott et al., 2019a; Alcott et al., 2019b; Gertler et al., costs that are externalized (e.g., healthcare costs, second- 2021). hand smoke, road traffic accidents, and domestic Health taxes are an efficient revenue generating tool violence). Negative internalities are the uninternalized for governments due to their relative ease of costs borne by the user resulting from time-inconsistent introduction, ease of administration, and rapid preferences due to imperfect information, present-bias, revenue generation potential. Health taxes are also less and/or addiction. Higher taxes on tobacco and alcohol are distortionary to broader economic activity than other also justified by considering the internalities: they indirect taxes and, once reductions in healthcare costs discourage people from initiating smoking and drinking, and increased labor productivity are considered, particularly at young ages, and they increase prices to increases in health taxes are found to be progressive. 1 This note was prepared by Evan Blecher, Ceren Ozer and Danielle Bloom (World Bank). The authors wish to thank peer reviewers Mark Goodchild and Itziar Belausteguigoitia (WHO); Patrick Petit (IMF); Chris Lane and Hana Ross (World Bank), as well as Emilia Skrok, Tuan Minh Le, and Violeta Vulovic (World Bank) for comments and input, and Linde Kremer for support (also World Bank). This series is produced via the Global Tax Program Health Tax Project under Task Team Lead Ceren Ozer. 2 Unless specifically indicated, references to tobacco, alcohol, and SSB taxes and revenues throughout the note refer to excise taxes and excise tax revenues on these products. Nevertheless, as it is the case with any tax reform, policy Finance as well as data shared with GTP during country makers who are considering health taxes are often engagements. The database includes all health tax presented with counterarguments to their effectiveness, revenues by product and country, as well as historical including in terms of the magnitude, buoyancy and time series data. At present, the database includes data sustainability of their impact on revenue. on 88 countries. The countries and data sources are listed in Appendix A, noting which products the data are Health taxes are most effective when they are well available for in each country.5 Unless otherwise indicated, designed and implemented; poorly designed and all references to tax revenues in this Knowledge Note are implemented health taxes may not result in the sourced from this database. See Box 1 for a discussion on increases in revenues and improvements in health the various metrics used to interpret revenue data. that policy makers expect. This note aims to review revenue related questions that may come up during the WHAT DO WE KNOW ABOUT THE policy making process and provide evidence to support MAGNITUDE OF HEALTH TAX REVENUES? the development of well-designed and implement health taxes. In doing so, the note unpacks the factors behind While magnitudes vary significantly between the magnitude and buoyancy of health tax revenues and products and countries, one thing is clear- health analyzes why these revenues are higher in some taxes generate meaningful revenues. In 2019, on countries than in others. This is done by assessing the average, tobacco and alcohol taxes generated revenues of impact of tax structures and tax rates on revenue, in the 0.6 and 0.3 percent of GDP, respectively (see Figures 1 context of a policy transmission mechanism.3 In this note and 2, and Table A1 in Appendix A). 2019 is used as the we also touch on a broader range of key considerations most recent year for comparison since tax revenue data including tax administration, the distributional impact, in 2020 and 2021 were subject to significant volatility due longer term sustainability of revenue, and excise tax to the COVID-19 pandemic, while data for 2022 are not revenues in the context of broader tax reforms. Country widely available yet (See Box 2 for an examination of the examples and empirical data are used to respond to the impact of the pandemic on revenues in several countries). most frequently asked questions of policy makers and In this sample there is no significant difference between practitioners, particularly of ministries of finance and tax revenues in high-income (HIC) and low- and middle- authorities. The purpose of this policy note is not to income countries (LMIC). Tobacco revenues averaged 0.6 consider the best practices in health tax policy design. and 0.5 percent of GDP in HICs and LMICs (40 and 41 There is already a broad consensus on best practices countries), respectively, and alcohol revenues averaged published by international organizations including World 0.3 percent in both HICs and LMICs (40 and 38 countries). Bank (2018; 2020), World Health Organization (2017; Revenues from tobacco exceeded alcohol in 48 of 76 2021a; 2023), and International Monetary Fund (Petit and countries (63 percent) for which tax revenue was available Nagy, 2016; Petit et al., 2021). for both commodities in 2019. A note on data and methods: As yet, there is no consistent In most cases, SSB taxes generate significantly less resource for health tax revenue data with different revenue than tobacco and alcohol taxes. Fewer sources covering narrow ranges of countries or products, countries implement SSB taxes,6 and as a result, revenue or not including historical data.4 This Knowledge Note data is only available for 23 countries. Many of the draws on a new resource that is under development. countries for which data are available apply SSB taxes During the production of this note, the World Bank Global alongside broader non-alcoholic beverage taxes, Tax Program (GTP) has started collecting and averaging 0.07% of GDP, with a maximum of 0.19% of standardizing revenue data from secondary sources and GDP. The lower magnitude of revenue is due to relatively supplementing them with data from national sources, lower tax rates, more elastic price elasticity of demand, including publicly available data from Ministries of 3 The tax structure refers to the type of tax (specific or ad valorem), the tax structure. The scope refers to the specific products which are included base, and other characteristics and attributes of the tax. Specific taxes or excluded from the tax. are levied on the volume of the product (i.e., the number of cigarettes or 4 These include OECD Global Revenue Statistics Database, European litres of beer), with the tax base being the unit of volume (i.e., the litres Commission Taxes in Europe Database, and the WHO Report on the of beverage or the volume of sugar or alcohol). Ad valorem taxes are Global Tobacco Epidemic. levied on the value of the product with the tax base being the point in 5 The database is available to World Bank staff but is not published on a the supply chain where the value is established. This can be early in the public forum due to the confidentiality of some national data. supply chain (e.g., the CIF value or ex-factory price) or later in the supply Furthermore, many of the data are easily accessed from sources listed chain (e.g., retail prices). Other characteristics and attributes include in Appendix A. whether the tax is uniform (i.e., applied equally to all products) or tiered (when different rates are applied based on prices or other product 6 WHO (2021a; 2018) report that 168 and 155 countries apply tobacco and characteristics (e.g., length, production volumes, or alcohol or sugar beer excise taxes, respectively, compared to 94 national level SSB excise content), or whether a tax thresholds is applied (i.e., a tier below which taxes, of which 20 apply only to SSBs, the remainder apply to non- no tax is paid). Finally, the scope of the tax is also part of this tax alcoholic beverages that also apply to SSBs (World Bank, 2023b). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 2 often a narrow scope of the tax,7 and tax structures that generate more supply side responses- although these will Box 1 // Measuring and interpreting tax be considered in more detail later. revenue metrics Health tax revenues are an unusually large Measuring and interpreting health tax revenue data is contributor in several countries with distinct challenging, and no single metric can tell the whole story. characteristics. Some small countries record relatively Excise tax revenues are often expressed as a percentage of high tobacco and alcohol revenues, averaging 1.0 and 0.7 GDP, however other denominators like share of tax revenue percent of GDP, respectively, well above global averages.8 may be very useful when considering narrower aggregates like This is not unexpected since small countries tend to health taxes since it controls for the relative size of tax generate relatively high indirect tax revenue.9 Health tax collections. Share of total revenue (including non-tax revenue) revenue shares are even higher in many small island may also be useful in countries which have large non-tax states. For example, tobacco revenue accounts for 3.4 revenues like resource royalties or social security contributions percent of GDP in Nauru10 and alcohol revenue accounts for similar reasons. Data from the World Bank’s World for 1.8 percent of GDP in the Seychelles. In Nauru, it likely Development Indicators is used as a standardized source of stems from challenges in measuring GDP due to limited GDP, tax revenue and total revenue data. capacity and the narrow range of economic activity (see IMF, 2022). When combined with volatility in its primary Indonesia and Luxembourg highlight how these metrics can be industries, tracking changes in economic activity has used together to interpret the magnitude of health tax become more difficult. In the Seychelles, it likely results revenues (Table B1). Tobacco tax revenue as a percentage of from very high levels of domestic consumption (Perdix et GDP is the same in both countries, however tobacco tax al., 1999) supplemented with consumption from a very revenue contributes a much larger share of total tax revenue large tourism sector.11 Countries with low overall tax and total revenue in Indonesia than Luxembourg (2.9 and 3.5 revenues are also likely to have a relatively high share of times, respectively). Tobacco tax revenue as a share of total total tax revenues from tobacco or alcohol excise taxes revenue in Indonesia is amongst the highest in the world and (see Box 1). principally due to the very low total tax and total revenue to GDP ratio. This highlights the need to consider multiple metrics WHAT DETERMINES THE MAGNITUDE OF when analyzing or comparing health tax revenues between HEALTH TAX REVENUES? countries, or even over time. Well-designed and administered tax policies Table B1: Tobacco, total tax, and total revenues in contribute to higher tax revenues. The examples Luxembourg and Indonesia, 2019 showed that higher tax rates and higher levels of use are Luxembourg Indonesia a strong contributor to high revenues, yet this does not Tobacco tax revenue explicitly consider tax structures and tax administration. Percentage of GDP 1.0 1.0 For example, Cambodia and the Philippines have starkly Percentage of total tax revenue 3.8 10.9 different tax structures, with Cambodia applying an ad Percentage of total revenue 2.4 8.5 valorem tax early in the supply chain (an identical tax Total tax revenue structure to Vietnam), while the Philippines implements a Percentage of GDP 26.4 9.8 best practice uniform specific tax. Furthermore, weak tax Total revenue administration can lead to low or declining revenues. In Percentage of GDP 42.0 12.3 South Africa, deterioration of tax administration resulted in relatively low tax revenues compared to the Philippines despite similar smoking prevalence, and similar tax rates and tax structures. These examples are discussed in detail in the subsequent section. 7 Of the 20 countries that have taxes on SSBs only, only 7 applied taxes to 10 Nauru also has a peculiarity in that the excise tax is legally an import all six categories of SSBs that the World Bank (2023b) database covers. duty. However, it is levied as a specific tax of $A 380 per 1000 sticks On the other end of the spectrum, 2 countries only applied to two (WHO, 2021a), a tax structure almost never observed for import duties, categories, highlighting the wide variation in scope of tax. and almost always as an excise. Furthermore, the rate is substantially 8 We use the World Bank’s definition of Small States, i.e., countries with higher than almost all other import duties. Since cigarettes are not populations less than 1.5 million (World Bank, 2023a). manufactured in Nauru and would not likely ever be, the excise tax is effectively implemented through the import duty system for simplicity. 9 Taxes on goods and services accounted for 9.3 percent of GDP in 24 While it may legally be an import duty, it is, in every other respect an small countries in 2019 compared to 8.6 percent in 104 other countries excise tax and we treat it as such. (World Bank, World Development Indicators). When considering only island small states, this rises to 10.2 percent (15 countries), compared to 11 Annual tourist arrivals in 2019 were 428,000 compared to a population 8.5 percent in 112 other countries. of 97,625 in the same year (World Bank, World Development Indicators). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 3 Box 2 // Health tax revenues during the COVID-19 pandemic The COVID-19 pandemic generated one of the largest economic shocks since World War II and had a remarkable impact on tax revenues in many countries. For example, in 2020, total tax revenues declined in real terms in 97 of 105 countries for which data were available, averaging a decline of 10.1 percent. By comparison, in 2019 tax revenues declined in real terms in only 31 of the 105 countries, averaging a decrease of 3.5 percent. Furthermore, comparing tax revenues as a percentage of GDP in 2020 is also skewed due to large declines in GDP in many countries. Excise tax revenues were particularly affected, with declines in excise tax revenues resulting from supply chains disruptions, changes in consumption patterns, and even temporary sales bans in some places. Declines in excise tax revenues were not heterogenous, with different countries and products experiencing varying trends. Several countries implemented temporary sales bans on tobacco and/or alcohol, which had large impact on tax revenues.1 South Africa was one of the most prominent with the tobacco sales ban lasting approximately 5 months (Filby et al., 2022), and three separate alcohol sales bans lasting approximately 5 months in total (Barron et al., 2023). Tobacco and alcohol excise tax revenue declined 50.2 and 30.6 percent, respectively, in real terms in 2020/21 compared to 2019/20. By comparison, total tax revenue and VAT declined by 11.0 and 7.8 percent, respectively, highlighting the dramatic impact the sales bans had on tax revenues. Alcohol revenues recovered quickly, exceeding pre-pandemic levels, however tobacco revenues did not and are still significantly lower than pre-pandemic levels (see Box 3). The ban inadvertently benefited manufacturers who were disproportionately engaged in the illicit market prior to the ban and likely further entrenched the already large illicit cigarette market (Filby et al., 2022; van der Zee et al., 2023). In the United Kingdom, beer sales and tax revenues fell sharply in early 2020, remaining at low levels through the first five months of the year but recovering somewhat before declining rather dramatically until early 2021 (HMRC, 2022c) (Figure B2).1 Beer sales and tax revenue declined 13.2 and 13.6 percent, respectively, in 2020, and were still below the 2019 baseline in 2021. The variation in sales and excise tax revenues closely correlate with lockdown restrictions in the UK which limited the consumption of alcohol due to closures in restaurants, bars, and other establishments that sell beer for on-premises consumption.1 On the other hand, cigarettes saw minimal disruption in clearances and tax revenues during the same period, with sales and tax revenue up by 0.1 and 7.3 percent, respectively, in 2020 (HMRC, 2022a). Figure B2: Monthly excise tax revenue from beer in the United Kingdom, 2016-2022 600 500 Tax revenue (millions, GBP) 400 300 200 100 0 Jan-21 Jan-16 May-16 Jan-17 May-17 Jan-18 May-18 Jan-19 May-19 Jan-20 May-20 May-21 Jan-22 May-22 Sep-22 Sep-16 Sep-17 Sep-18 Sep-19 Sep-20 Sep-21 Excise revenue Excise revenue (12 month moving average) Source: HMRC (2022c) HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 4 Specific taxes and mixed tax systems generally result Figure 1 // Tobacco excise taxes as a in higher tax revenues than ad valorem taxes alone. Data from the health tax revenue database shows that percentage of GDP by country, 2019 tobacco tax revenue averaged 0.6 and 0.7 percent of GDP in the 28 and 43 countries with specific and mixed Low- and middle-income countries High-income countries systems, respectively, compared to an average of only Bulgaria 0.2% in the 10 countries with ad valorem taxes only.12 Kosovo North Macedonia Other elements of the tax structure also affect tax Georgia Türkiye revenue yield, including the tax base and tiers. There was Turkiye Armenia insufficient variation of these factors within the tax types Ukraine Egypt, Arab Rep. to generalize results, however they are considered in Mauritius Indonesia Solomon Islands more detail through country examples later. Philippines Bangladesh Argentina Specific taxes will result in more stable and Thailand Kazakhstan predictable tax revenues than ad valorem taxes. Burkina Faso Romania Under ad valorem taxes, the effective tax is the product of South Africa Cambodia the tax rate and tax base. Firms can affect the tax base Malaysia Senegal Niger and thereby the effective tax even when tax rates remain Mali Guyana unchanged. For example, firms can increase or decrease Colombia Rwanda retail prices (when the tax base is located later in the Ecuador El Salvador supply chain) or affect their cost structure (when the tax Congo, Dem. Rep. Dominican Republic Honduras base is located earlier in the supply chain). Furthermore, Côte d’Ivoire Mongolia consumer behavior can also affect the effective tax, Brazil Costa Rica reducing it if they trade down to cheaper products, or Guatemala Panama increasing it when trading up to more expensive Uganda Nicaragua products. Specific taxes are applied to the volume of the Nauru Greece Luxembourg product thereby negating the impact of firm and Czechia Poland consumer behavior on the effective tax. Slovenia Estonia Slovak Republic Specific taxes are also preferred to ad valorem taxes Hungary Cyprus because they are more effective in targeting negative Latvia Malta externalities and internalities and are more effective Seychelles Portugal Australia in improving health. The negative externalities and Lithuania Italy internalities of tobacco, alcohol, and SSBs do not coincide Spain France with the value of the product, but rather the volume of Belgium Chile consumption. For example, a cheaper cigarette does not Austria Finland generate smaller externalities or internalities than a more Germany United Kingdom Uruguay expensive one, and thus a tax on volume is a better proxy Japan Denmark of externalities and internalities than a tax on value. Canada Ireland Specific taxes also increase prices of cheaper products Netherlands Switzerland more in relative terms, therefore reducing price variation Sweden Singapore in the market and reducing incentives to trade down to Norway Korea, Rep. Croatia cheaper brands in response to a tax increase (WHO, Trinidad and Tobago New Zealand 2021b). However, specific taxes need to be regularly United States adjusted to account for inflation to ensure that their 0.0% 1.0% 2.0% 3.0% 4.0% value, and thus tax revenue is not eroded over time.13 Source: World Bank GTP health tax revenue database Uniform taxes also result in higher, more stable, and predictable tax revenues than tiered tax systems. consumption. An exception is when the tiers are based on Firms have similar a similar incentive to adjust the the alcohol or sugar content and generate intentional characteristics of products to shift them into lower tax incentives for firms to adjust the characteristics of tiers, reducing tax revenue, but also undermining the products to shift products into lower tax tiers. While this impact of the tax on increasing prices and reduction will still reduce revenue, it can be a trade-off for a larger 12 Data on tax structures was sourced from WHO (2021a). No comparable 13 The GTP has recently published a Knowledge Note on the topic with a tax structure data was located for alcohol and SSBs. more detailed analysis of the effects of inflation on health taxes, including good practice policy solutions (Lane et al., 2023). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 5 health impact. Tax structures that generate these Figure 2 // Alcohol excise taxes as a intentional trade-offs are discussed in more detail later. percentage of GDP by country, 2019 THE IMPACT OF HEALTH TAX REFORMS Low- and middle-income countries High-income countries ON INCREASING TAX REVENUE Bhutan Increases in taxes result in increases in prices that in turn Mauritius Mongolia reduce sales and consumption. The relationship between Cambodia Thailand prices and consumption is defined by the price elasticity Dominican Republic Rwanda of demand.14 Since the price elasticity of demand for Solomon Islands South Africa tobacco and alcohol is relatively inelastic and since taxes Guyana Colombia are only a share of the price, tax revenues increase El Salvador Philippines Türkiye despite declining consumption. This policy transmission Tunisia Uganda mechanism – and thus the impact of tax increases on tax Burkina Faso North Macedonia revenue – is influenced by a variety of factors, including Ecuador Bulgaria tax policy and tax administration, as well as several Armenia Honduras parameters including tax pass-through rate.15 Congo, Dem. Rep. Georgia Costa Rica Health tax reforms including increasing tax rates and Kosovo Kazakhstan reforming tax structures can raise substantial tax Panama Guatemala revenue, quickly and efficiently. Beginning in 2012, the Côte d’Ivoire Mali Philippines embarked on an ambitious series of alcohol Equatorial Guinea Brazil Senegal and tobacco tax reforms (Kaiser et al., 2016) that Egypt, Arab Rep. Romania consolidated several price-based tiers on cigarettes and Argentina Niger beer in an upward manner. The reforms culminated in a Seychelles Nauru uniform specific tax being achieved in 2017 (Figure 3; the Latvia Estonia same figure for beer is shown in Figure A1 in Appendix Lithuania Finland A).16 Between 2012 and 2020, alcohol and tobacco excise Poland United Kingdom revenues increased by 140 and 270 percent in real terms, Norway Ireland Slovak Republic respectively (Figure 4). Increases in tax revenue were very Sweden Greece rapid, with tax revenues increasing by 24 and 112 percent Czechia Australia on alcohol and tobacco, respectively, in the first year Slovenia Korea, Rep. (2012 to 2013), amounting to 0.4 percent of GDP and Japan Trinidad and Tobago accounting for the entire increase in total tax to GDP ratio New Zealand Hungary in that year (15.1 to 15.5 percent). Between 2012 and Belgium Cyprus 2020, total tax revenue to GDP increased from 15.1 to France Denmark 17.8 percent with the incremental revenue from tobacco Malta Chile Netherlands and alcohol taxes contributing approximately one third of Singapore Portugal the increase. Luxembourg Uruguay Spain In addition to rapidly increasing tax revenues, Austria Germany sustained tax increases can generate additional tax United States Canada revenues over long periods of time. Since the early Italy Switzerland 1990s, South Africa has experienced large increases in Croatia tobacco and alcohol tax revenues following consistent 0.0% 0.5% 1.0% 1.5% 2.0% annual increases in tobacco and alcohol taxes. Notably, Source: World Bank GTP health tax revenue database South Africa has a well-designed tax structure, employing specific taxes that are uniformly applied within each product category. Between 1990/91 and 2021/22, cigarettes taxes increased by 641 percent in real terms 14 The impact of prices on consumption (defined by the price elasticity of 15 The tax pass-through rate is the magnitude by which a tax change demand) is a critical part of the policy transmission mechanism since influences prices. It measures the change in price in response to a this ultimately determines the impact on both health and revenue. The change in the tax rate. price elasticity of demand is the percentage change in consumption 16 Between 2012 and 2020, cigarette taxes increased by 1,245 and 29 resulting from a percentage change in price. Inelastic demand means percent on the lowest and highest tiers, respectively. Beer taxes that the decline in consumption is less than proportional to the increase increased by 173 and 38 percent on the lowest and highest tiers, in price. This is discussed in more detail later in this section. respectively. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 6 while tax-paid sales declined by 72 percent (Figure 5).17 Even though sales declined, tax revenues increased by Figure 5 // Tobacco excise taxes and tax 103 percent. Similar patterns are observed for alcohol taxes, although the magnitudes of tax and revenue revenue in South Africa, 1960-2022 increases varies between different products. Between 25 25 Revenue (billions, constant 2022 ZAR) Excise tax per pack (constant 2022 ZAR) 1990/91 and 2021/22, beer taxes increased by 95 percent 20 20 in real terms coinciding with tax revenues increasing by 182 percent (Figure 6). Taxes on spirits increase by 146 15 15 percent in real terms, coinciding with tax revenue 10 10 increasing 322 percent (Figure A2 in Appendix A).18 Despite challenges with tax administration and the over 5 5 shifting of taxes (see Box 3 for a more detailed analysis of 0 0 health taxes in South Africa), tobacco and alcohol tax 1960 1970 1980 1990 2000 2010 2020 increases in South Africa have resulted in substantial Excise per pack Revenue increases in revenues. Note: Years represent fiscal years (e.g., 1968 is the 1968/69 fiscal years). Source: National Treasury (2023); Statistics South Africa (CPI) Figure 3 // Cigarette excise taxes in the Philippines, 2012-2020 Figure 6 // Beer excise taxes and tax revenue in South Africa, 1968-2022 50 PHP per pack (constant 2020 prices) 45 25 350 40 Revenue (billions, constrant 2022 ZAR) Excise/LAA(20L) (constant 2022 ZAR) 35 300 20 30 250 25 15 200 20 15 150 10 10 5 100 5 0 50 2012 2013 2014 2015 2016 2017 2018 2019 2020 0 0 Low Medium High Premium 1968 1977 1986 1995 2004 2013 2022 Excise/LAA Excise/20L Revenue Note: Low (<₱5 per pack), Medium (₱5-6.5 per pack), High (₱6.5-10 per pack), and Premium (>₱10 per pack). Note: Years represent fiscal years (e.g., 1968 is the 1968/69 fiscal years). Source: Kaiser et al. (2016); Republic of the Philippines (2012; 2017); World The tax base was changed from a volumetric tax to a tax per litre of Bank World Development Indicators (CPI) absolute alcohol in 1998/99; volumetric taxes have been scaled to 20 litres of beer approximating equivalence between 1997/98 and 1998/99. Figure 4 // Alcohol and tobacco excise tax Source: National Treasury (2023); Statistics South Africa (CPI) revenue in the Philippines, 1994-2020 Figure 7 // Tobacco excise tax base and 20% 250 effective tax in Vietnam, 2006-2016 Revenue (billions, constant 2020 PHP) 16% 200 Percentage of GDP 6,000 80% 12% 150 70% VND/pack (constant prices) Ad valorem excise tax rate 5,000 60% 8% 100 4,000 50% 3,000 40% 4% 50 30% 2,000 0% 0 20% 1994 2000 2006 2012 2018 1,000 10% Health tax (LHS) Other tax (LHS) Health tax (RHS) 0 0% 2006 2008 2010 2012 2014 2016 Source: OECD (2022); World Bank World Development Indictors (GDP; CPI) Tax base Excise tax Excise tax rate (percentage of base) Source: Blecher and Le (2018) 17 While not included in the figures, sales volumes are imputed from the complicated tax history while being a relatively smaller contributor to tax revenue data by dividing tax revenue by the specific tax rate. alcohol tax revenue (ZAR 6.3 billion compared to ZAR 20.8 and 12.9 18 Similar increases in wine taxes and revenues have occurred, however billion for beer and spirits, respectively), it is not included in the analysis. still wine, the largest contributor to wine tax revenue, was not taxed in the 1980s, with tax being reintroduced only in 1991/92. Since wine had a HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 7 Tax increases under an ad valorem tax may not depends on the tax structure, including the type of generate the same degree of tax revenue, especially tax and the tax base. In addition to the tax structure, tax when applied early in the supply chain. Vietnam is a pass-through is also a determined by the market typical example, applying an ad valorem tax on cigarettes structure and the price elasticity of demand (Keen, 1998). based on the ex-factory price. The tax rate increased from Under specific taxes, the greater the degree of over 55 to 65 percent in 2008, and 65 to 70 percent in 2016. shifting, the smaller the increase in revenue. Conversely, Figure 7 shows the average tax base (i.e., average ex- the greater the degree of under shifting, the larger the factory price) and effective tax of cigarettes between 2006 increase in revenue. Under ad valorem taxes, the greater and 2016 in real terms. The trend shows how the effective the degree of over shifting, the larger the increase in tax did not increase in response to the increase in the tax revenue, and the later in the supply chain the tax base is, rate. Rather, the tax base decreased and the effective tax the larger the increase in revenue because of the also decreased slightly. As a result, average prices fell in multiplier effect. Furthermore, specific taxes are more real terms (by 31 percent between 2007 and 2016) likely to be over shifted than ad valorem. See Appendix B (Euromonitor, 2023). The reason for the decline in prices for a more detailed discussion of tax pass-through. was the under shifting of tax increases,19 supported by Increases in tobacco and alcohol taxes result in the decline in the tax base and effective tax as well as the relatively large increases in tax revenues because rest of the supply chain not increasing margins (these they are nearly always inelastic. The example of margins actually decreased).20 Excise tax revenues tobacco tax increases in South Africa highlighted this increased from VND 13.5 to 13.8 billion between 2013 and since the decline in consumption was less than 2019, however this amounted to a decline of 11 percent in proportional to the increase in prices and goes to show real terms (Ross, 2021). The lack of increases in revenue how tax revenue increases even when consumption was further compounded by cigarettes sales increasing by declines. While the price elasticity varies between and 2 percent between 2006 and 2016 (Euromonitor, 2021). even within products, between countries and over time, Ad valorem taxes can result in increasing tax several systematic reviews have found that alcohol is revenues even when tax rates remain unchanged, inelastic (see Table 2 for selected estimates). Earlier although this is dependent on the tax base. This is systematic reviews showed that alcohol is more inelastic called the “multiplier effect” and occurs when the tax base in HICs compared to LMICs (Sornpaisarn et al., 2013; increases, resulting in an increase in the effective tax Nelson, 2013) although a more recent study finds no without an increase in the tax rate (Keen, 1998). The significant difference between LMICs and HICs (Guindon implications for tax policy are dependent on the location et al., 2022). Beer is generally found to be more inelastic of the tax base. For example, when the tax base is located than other alcoholic beverages. Surprisingly, there are later in the supply chain, the multiplier effect can provide fewer systematic reviews of tobacco, however price a hedge against inflation since increases in retail prices will increase Table 2 // Price elasticity estimates from systematic reviews the tax base and thus the effective tax. However, if the tax base is Product Category Countries Estimate Reference located early in the supply chain Tobacco Cigarettes HICs -0.2 to -0.6, clustering around -0.4 NCI (2016) increases in retail prices may have LMICs -0.2 to -0.8, clustering around -0.5 limited impact on the tax base and Alcohol Beer Global -0.30 Guindon et al. (2022) effective tax. Tax revenues may Wine -0.60 benefit from industry led price Spirits -0.65 increases although the multiplier Beer Mostly HICs -0.30 effect itself may itself act as a Wine -0.45 disincentive for industry led price Spirits -0.55 increases. All alcohol LMICs -0.64 Beer -0.50 The degree to which taxes are Other -0.79 passed through can also affect tax SSBs SSBs Global -1.59 Andreyeva et al. (2022) revenue, however the impact 19 Tax increases are said to be fully passed through when the change in 20 These are several potential causes for this, including cost cutting, but it price is the same as the change in the tax. Tax increases can also be over may also highlight the quality gradient. There is a widely held consensus shifted when the price increases by more than the increase in tax for both tobacco (WHO, 2021b) and alcohol (WHO, 2017) that ad valorem (increasing margins). Tax increases are under shifted when the price taxes encourage a more diverse quality or perceived quality of product increases by less than the increase in tax (decreasing margins as the compared to specific taxes. supply chain absorbs some of the tax increase). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 8 elasticity estimates vary between LMICs and HICs, with Box 3 // Health taxes in South Africa most estimates in HICs between -0.2 and -0.6, clustering South Africa implements specific taxes on tobacco and around -0.4, and between from -0.2 to -0.8, clustering alcohol, with a uniform rate on cigarettes, and specific taxes around -0.5 in LMICs (NCI, 2016). Price elasticity is based on alcohol content for beer and distilled spirits. Wine explained in more detail in Appendix B. taxes are applied as volumetric rates across three separate The demand for SSBs is more elastic than tobacco and categories (still, sparkling, and fortified wine). Since the early alcohol meaning that SSB taxes have a relatively 1990s, tobacco and alcohol tax rates have been increased smaller tax revenue potential. A systematic review significantly and regularly. While tax increases have slowed in across a range of countries finds SSBs to be elastic, with recent years, they have continued to increase in nominal an elasticity estimate of -1.59 (Andreyeva et al., 2022). terms to maintain their real value. Tax increases have led to Systematic reviews have not yet identified differences large increases in prices, reductions in sales volumes, and between HICs and LMICs, however a cross-sectional study increases in tax revenues on alcohol. found that SSBs were less elastic in the lowest-income Since the mid-2010s, tax revenues from tobacco taxes have countries (Muhammad et al., 2019). The most significant declined because of weak tax administration, not weak tax reason that SSBs are found to be elastic, or more elastic policy. Tax revenue in 2021/22 was 45 percent lower than its than tobacco and alcohol, is the greater availability of peak in 2014/15. Independent estimates show that illicit substitutes, including water. While there are significant cigarettes increased from less than 10 percent of the market challenges in access to safe and affordable drinking water in 2009 to nearly one third by 2017 (Vellios et al., 2020). A in many countries, leading to the inclusion of access to targeted intervention against illicit cigarettes that began in safe and affordable water in the Sustainable Development 2013 was subsequently shut down alongside other deliberate Goals (United Nations, 2018), potable water provides a attempts to undermine the governance of tax administration readily available substitute when consumers are faced (Van Walbeek, 2020). As was highlighted in Box 2, tobacco tax with higher prices.21 administration was vulnerable and the temporary sales bans fed an already large illicit market, with the lack of any security MOVING FROM THE SHORT- TO LONG- features on packs increasing the vulnerability to tax evasion RUN (Vellios et al., 2022). Tobacco tax administration in South Africa suffers from many challenges that continue to In the long run, successful health tax policy will lower undermine tax revenue collection. consumption, raising the question of whether tax revenues may begin to decline at some point in the Tobacco tax increases have also resulted in dramatic declines future. This is supported by economic theory that in tobacco use, with adult smoking prevalence declining from teaches us that as time goes by the price elasticity is going 32 percent in the early 1990s (Van Walbeek, 2002) to 18 to become less inelastic as consumers can find more or percent in 2012 (Reddy et al., 2015), leading to significant better alternatives, or as firms respond. This is confirmed declines in mortality and morbidity. Tobacco’s contribution to by a meta-analysis of long-run price elasticities for all-cause mortality declined from 13 to 8 percent of all tobacco (Gallet and List, 2003).22 The implications is that deaths, and 8 to 5 percent of all disability adjusted life years the demand response to price increases with time, which lost (DALYs) between 1990 and 2019 (GBD, 2019). Using a lowers the tax revenue, but also increases the impact on synthetic control analysis Chelwa et al. (2016) show that 78 health. From a practical perspective, this raises several percent of the decline in per capita cigarette sales between questions. First, how elastic (or inelastic) is demand for 1990 and 2004 was attributable to tax increases. tobacco, alcohol, and SSBs in the long-run. Second, how During much of the 1990s and early 2000s, tax increases might authorities mitigate the risk of declining revenue were over shifted to retail prices (Linegar and van Walbeek, (keeping in mind that the lower consumption of products 2018) leading to faster consumption decline and lower than with externalities and internalities will bring economic expected tax revenue increases. Alcohol tax increases were benefits in terms of lower health care costs and higher also over shifted, making the increases in tax revenues even labor productivity). And third, what is the time horizon more impressive (Russel and van Walbeek, 2016). that we need to consider. These questions are best answered by considering trends in good practice countries who have well-designed tax structures and high tax rates resulting from a long history of raising taxes. 21 The price elasticity is only one reason that SSBs have a smaller revenue 22 The price elasticities analyzed in the previous section are short-run potential than tobacco and alcohol. Other reasons include lower tax elasticities. rates and tax structures. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 9 Even countries with relatively high taxes can continue Figure 8 // Tobacco excise taxes and excise to experience increases in revenues from raising tax revenues in Australia, 2001-2022 taxes. Australia has some of the highest cigarette taxes 20 30 and prices in the world and has continued to raise taxes Revenue (billions, constant 2022 AUD) Excise per pack (constant 2022 AUD) 18 in recent years.23 Figure 8 shows tobacco tax revenues in 16 25 real terms since 2001 and how increases in tax rates have 14 20 12 resulted in increased tax revenues. Deliberate increases 10 15 occurred in 2010, and annually from 2013 to 2020, in 8 10 6 addition to twice-annual automatic adjustments to 4 5 account for inflation and income growth.24 This tax policy 2 0 0 coincided with large declines in tobacco use. Daily adult 2001 2006 2011 2016 2021 smoking prevalence declined from 22.3 to 10.7 percent Excise tax per pack Excise and customs revenue between 2001 and 2021, respectively (Greenhalgh et al., 2021). Ignoring the disruptions in tax revenues in 2021 Notes: Excise tax per pack is measured at year end, 20 cigarettes per pack and 2022 due to the COVID-19 pandemic (see Box 2), Source: Scollo and Bayley (2022); ABS (2022) Australia shows that even the highest taxed countries can still generate substantial increases in tax revenues when Figure 9 // Cigarette and hand-rolled tax rates are increased. tobacco excise tax revenue and volumes In contrast, countries with relatively high taxes that in the UK, 2000-2022 do not increase taxes are at risk of experiencing 14 70 declines in tax revenues in the long-run. This is Revenue (billions, constant 2022 GBP) Sales (billions of sticks or equivalent) 12 60 because countries are likely to experience long term 10 50 trends in declining consumption precisely because of 8 40 prior tax increases. An example of this situation is the 6 30 4 20 United Kingdom (UK) where both cigarette taxes and 2 10 prices are relatively high.25 While the UK has also 0 0 continued to raise taxes, it has not done so as 2000 2004 2008 2012 2016 2020 aggressively as Australia. Between 2000 and 2022, tax per Cigarettes (volume) RYO (equivalent volume) pack increased by a total of 59 percent (in real terms), Cigarettes (revenue) Cigarettes + RYO (revenue) coinciding with a decline in tax revenue from cigarettes of Source: HMRC (2022a); ONS (2022a)) 38 percent as cigarettes sales volumes declined by 61 percent (Figure 9). However, only part of the decline in Figure 10 // Cigarette sales and smoking revenue was due to the secular decline in cigarette prevalence in the UK, 2000-2022 consumption, with substitution to other tobacco products playing a significant role, notably a significant increase in 80 30% Sales (billions of sticks or equivalent) sales of hand-rolled tobacco (HRT) used to make “roll- 70 25% Smoking prevalence (adult) 60 your-own” cigarettes (RYO). HRT sales increased by 246 50 20% percent between 2000 and 2022, however off a 40 15% significantly smaller base. A large portion of this 30 10% 20 substitution was likely due to the relatively higher taxes 10 5% on cigarettes compared to HRT (average excise per pack 0 0% 2000 2004 2008 2012 2016 2020 of cigarettes was GBP 6.95 compared to an average of GBP 4.57 on RYO in 2022).26 The significant difference in Cigarettes (volume) RYO (equivalent volume) taxes is despite taxes on HRT having increased more Smoking prevalence (16+) Smoking prevalence (18+) rapidly than cigarettes (95 percent in real terms since Note: The definition of age in smoking prevalence data was changed from 2000). Once revenue from HRT is included, the decline 16 years and older to 18 years and older from 2021. ONS (2022b) only backdated changed to the historic time series to 2011. Both series are was only 21 percent indicating that a portion of the shown for the overlapping years. Source: HMRC (2022a); ONS (2021; 2022b) 23 In 2020, using the WHO benchmark of the most sold brand of cigarette, 24 Automatic adjustment for inflation until 2014, and for nominal wages Australia had the highest price and second highest excise tax in the thereafter. world (WHO, 2021a). This example is somewhat different to the earlier 25 In 2020, using the WHO benchmark of the most sold brand of cigarette, example of South Africa which showed the buoyancy of revenue over the UK had the 6th highest price and 4th highest excise tax in the world long periods, however tax rates in South Africa are still substantially (WHO, 2021a). lower than Australia. The example of Australia is to highlight a country 26 A RYO cigarette tends to use approximately 0.75g of fine-cut tobacco with very high tax rates in the long-run. (Gallus et al., 2015). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 10 decline in cigarette revenue was due to a substitution to real terms) between 2000 and 2021 (HMRC, 2022c). Some HRT.27 countries are less concerned about long-run revenue losses. New Zealand will phase out cigarette sales, Over the long run, declines in tax revenue may be less banning the sale of cigarettes to anyone born after 2008, than expected or only occur when accompanied by aiming to eliminate smoking.30 There is concern that as dramatic declines in consumption. For example, the countries achieve success with health tax policies and set decline in tax revenue in the UK coincided with large tax rates at high levels, that the remaining smokers or declines in cigarette sales and smoking prevalence (Figure drinkers will become a vulnerable group in society, 10).28 The decline in tax revenue is less consequential whether that may be persons with lower incomes31 or when considering the impact on health resulting from mental health or substance abuse issues32, or those declining tobacco use since reducing the negative without adequate support to successfully change externalities and internalities generates a net benefit for behaviors. Countries should pay attention to vulnerable society. Declining revenue should also be netted against populations when considering health tax reforms and the savings to the budget from a reduction in expenditure consider appropriate complementary non-tax policy on direct medical costs (and potentially elsewhere, for interventions to support their cessation and behavior example, in the judicial system). For example, in South change. Africa, estimates were that alcohol costs the national and provincial governments ZAR 17.2 billion in 2009/10, which WHAT ARE KEY CONDITIONS FOR was more than the total tax collections from alcohol in INCREASED TAX REVENUES? that year (National Treasury, 2014). While, in the United States, Lightwood and Glantz (2016) found that a 10 Policy makers and practitioners are faced with many percent relative decline in smoking was followed by an challenges when considering health tax reforms. expected decline in healthcare expenditure of US$ 80 While they will likely face opposition, particularly from the billion in the following year.29 Thus, even if tax revenues manufacturing industries, there are other substantive were to decline in the long run, they cannot be considered challenges that are important to reflect on. These include in isolation of the dramatic impact that health taxes can the impact of tax administration on revenues, particularly have on the expenditure side of the budget. how poor or weak tax administration can undermine Long-run declines in tax revenue are only likely to revenue generation. Arguments are often made that occur in the distant future. As the examples of South health taxes are regressive, yet studies show that tax Africa and Australia showed declines in excise tax increases are progressive once the long-run impact of revenues are unlikely to occur for the foreseeable future. reduced consumption are accounted for. In addition, However, if countries are concerned about the long run revenue generated by health taxes can be directed to sustainably of revenue there is ample opportunity to seek ensure even more progressive distributional outcomes. alternative revenue sources, even from other health And finally, while increasing taxes is argued to increase taxes. In the case of the UK, declining tobacco tax revenue revenue, this is not always the case. As has already been has been partially offset by the implementation of an SSB alluded to, health taxes are sometimes not designed with excise that generated GBP 334 million in the 2021/22 revenue generation in mind, but rather to affect fiscal year (HMRC, 2022b), and by large increases in incentives for producers or consumers to avoid the tax. alcohol excise taxes, that increased revenue by 28% (in 27 The rise in popularity of electronic nicotine delivery systems (ENDS) income populations. Van Walbeek (2002) shows that the decline in presents a similar challenge, however the impact on revenue is yet to be smoking prevalence in South Africa was significantly steeper among determined. Heated tobacco products (HTP) are taxed in the UK, poorer compared to richer groups between 1993 and 2002. However, he although tax paid sales and revenue data do not show a significant notes that evidence from the UK shows that smoking prevalence impact yet. Electronic cigarettes are not taxed and while prevalence is declined more rapidly in higher- rather than lower socio-economic significant (7.7 percent of adults) a quarter of cigarette smokers also use groups since the 1960s. While it may be argued that the South African electronic cigarettes (ONS, 2022b), potentially limiting the revenue case has a greater attribution to tax increases, it nevertheless highlights impact. the need to focus attention lower-income or socio-economic groups 28 Importantly, smoking prevalence estimates include illicit tobacco, and when considering health taxes. See also the discussion on progressivity the decline in smoking prevalence highlights how declines in cigarette in the next section. sales were not a result of smokers substituting to illicit cigarettes in 32 Unhealthy alcohol use is often found to coincide with mental health response to higher taxes. Blecher (2019) highlights how official data on disorders (see Palzes et al. (2020) for a typical example), however illicit tobacco in the UK shows declining illicit cigarettes sales declined causality can run in either direction or be bidirectional. While little is sharpy since the early 2000s as taxes increased. known about price elasticities of these groups, systematic reviews show 29 The study estimated the value to be US$ 63 billion in 2012 prices. that price sensitivity declines as volume or concentration of drinking Converted to 2022 prices using CPI-U from the Bureau of Labor increases (see Guindon et al., 2022). However, evidence also shows that Statistics. tax/price increases reduce initiation (for example, see Sornpaisarn et al., 2015) and thus would reduce the potential for future unhealthy alcohol 30 https://www.bbc.com/news/world-asia-63954862 use. Nonetheless, the higher prevalence of unhealthy alcohol use 31 Lower income populations are relatively more price sensitive than among vulnerable populations highlights the need for focussed higher income populations and thus the relative impact of tax increases attention when considering health tax policy. should favor greater behavior change among them compared to higher- HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 11 These challenges are considered in more detail in this of the same systems and methods, particularly those in section. other areas of indirect tax like customs and VAT. Furthermore, improvements and reforms in one area of Tax administration tax administration likely affects others. While health taxes may receive particular attention, concerns about tax Health taxes are relatively easier to administer than compliance should not be seen in isolation or as a barrier many other taxes. Excise taxes are usually collected to health tax policy reforms and should be coordinated from a small number of taxpayers early in the supply with broader efforts to improve the effectiveness of tax chain, in most cases at the point of importation or the administration. factory gate, making tax administration relatively simple and cost effective. Application of technological solutions Health taxes have unique tax administration like track-and-trace systems are increasing the efficiency challenges and barriers that should be recognized and reducing costs of tax administration, increasing tax including a unique and unusual incentive for firms to compliance, while simultaneously increasing revenue exploit illicit trade and weak tax structures to avoid collection. tax increases. There is a long history of firms overstating the levels and scale of illicit trade in tobacco to generate a Tax evasion can undermine health tax revenues; causal link between tax increases and tax compliance, countries with weak tax administration are while also actively engaging in illicit trade to undermine particularly vulnerable to tax avoidance and evasion tax policy or for other strategic purposes like market entry on high-risk products and tax increases may not (WHO 2021b). Independent research points to no or increase revenue as much as expected. There are many limited impact of tax increases on illicit cigarette trade forms of tax evasion, varying by product. Tobacco is at (Ross and Blecher, 2019). Authorities should take an relatively higher risk of tax evasion due to its relatively independent view, ensuring that independent estimates higher value to volume and weight ratio, and (often) has of tax compliance inform policy and evaluation while higher taxes (see WCO, 2021).33 While it is argued that tax recognizing a conflict of interest between firms and tax increases may increase the incentives for tax evasion it is policy. notable that countries with higher tax rates often tend to have lower rates of tax evasion. While this may not seem Health taxes are progressive when the long-run intuitive, countries with more significant governance and impact of behavior changes are accounted for corruption challenges tend to have lower tax rates, and countries with higher rates of tax have a greater incentive World Bank research shows that health taxes can be to invest in improving tax administration and compliance fiscally progressive in the long-run when the (Joossens et al., 2010). behavioral effects of reduced consumption are accounted for. A common argument against health taxes Specific taxes and uniform taxes are easier to collect, is that they are regressive since the burden of the tax particularly in low resource tax administration increase falls disproportionately on the poor. This is a settings, resulting in higher and more stable narrow argument that relies on the average rather than revenues. Ad valorem taxes generate increased marginal tax incidence. Since poorer income groups are opportunities for tax evasion through undervaluation and relatively more price sensitive (i.e., more price elastic) under invoicing (WHO, 2021b; Petit et al., 2021). than richer income groups the anticipated change Conceptually, it is easier for authorities to estimate a behavior (i.e., increased cessation or reduced intensity) is volume of product than ascertain a value, especially when larger in poorer than richer income groups. Furthermore, levied early in the supply chain. Weak or under resourced savings in avoided health costs and increases in labor authorities may be more vulnerable, lacking the resources productivity and extended working life disproportionately to effectively administer ad valorem systems. Additionally, favor the poor thereby making health taxes progressive in tax administrations in both HICs and LMICs can employ the long run. This has been demonstrated for tobacco, modern technological tools like track-and-trace systems alcohol, and SSB taxes in 14 countries using the World that can track volumes accurately and effectively. Bank’s Extended Cost Benefit Analysis (ECBA) (see Fuchs Challenges in tax administration and reforms to and Pierola (2022) for a summary of these results). It improve tax administration are not unique to health should be noted that almost all these studies were taxes. Health tax administration will suffer if general tax conducted in LMICs, many of which have very low tax administration is weak since they share and rely on many rates. As previously discussed, there are concerns that 33 Tobacco tax evasion is often referred to as illicit trade, which can involve the smuggling, counterfeit, illicit/cheap white cigarettes, unbranded tobacco, bootlegging, and under declaration (Ross and Blecher, 2019). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 12 countries with very high tax rates and declining use may the sales of lower alcohol or sugar beverages and encounter different equity challenges among vulnerable decrease sales of higher alcohol or sugar beverages.34 populations in the long-run. Since the lower alcohol or sugar content products will attract a lower effective tax than products with higher Tax revenues generated by health taxes contribute to alcohol or sugar content, revenue will be expected to the progressive impact of government expenditures decline over time as this shift occurs. in most countries. The analysis of health tax incidence ignores the overwhelmingly progressive impact of Alternatively, firms may respond by reformulating expenditures supported by higher health tax revenue. products to reduce the alcohol or sugar or shifting While tax revenue is largely fungible, World Bank research advertising from higher to lower alcohol or sugar in South Africa has shown that government expenditure products. Evidence from South Africa suggests that a was overwhelmingly progressive, with just over half of reform of the tax base on beer from the volume to government expenditure devoted to social spending. alcohol content, combined with increases in the tax rate, Furthermore, it shows that the tax system is progressive resulted in dramatic shifts in advertising from higher to although excise taxes on alcohol and tobacco were found lower alcohol beers that coincided with a significant – on average – to be regressive (Inchauste et al., 2015). decline in the average strength of beer consumed and This study however, did not consider the impact of total alcohol consumption from beer (Blecher, 2015).35 reduced externalities and internalities. Once reductions in More recently, South Africa implemented an SSB tax with medical costs and increases in working years were similar incentives in the tax structure. The excise tax is accounted for in an ECBA study on tobacco taxes in South based on the grams of sugar per 100ml that exceed a Africa, the distributional impact of tobacco tax increases threshold of 4g/100ml. This generates an incentive for over time was progressive (Fuchs et al., 2018). When the firms to lower sugar content to reduce their tax liability. impact of the progressivity of the increased fiscal Since implementation, revenue from the SSB tax declined expenditure was accounted for (as shown by Inchauste et from ZAR 3.3 to 2.3 billion (2018/19 and 2021/22 fiscal al., 2015), the net impact of tobacco tax increases is even years) even though the tax rate has remained unchanged more progressive. (National Treasury, 2022). Furthermore, this also indicates a significant decline in the volume of sugar consumed When revenue isn’t the goal: reformulation and other from SSBs, thus achieving the policy goal. Similar evidence health objectives on reformulation has been observed in the UK where a tiered system exists with a threshold below which no tax Some health tax reforms may be designed in a is paid (5g/100ml), and a tier above which a higher tax is revenue neutral manner. For example, a non-alcoholic paid (8g/100ml). Initial estimates were that the tax would beverage tax that includes taxes on both SSBs and non- raise GBP 520 million, with revenues declining over time SSBs might be reformed to reduce or eliminate the tax on as both producers and consumers were expected to shift non-SSBs while increasing the tax on SSBs. This reform behavior (HMRC, 2016). The tax was announced nearly would have significant health benefits and have little or two years in advance of implementation to provide no impact on revenue if tax structures are well designed producers with sufficient time to reformulate products to and tax rates smartly set. In 2014, Finland reformed their reduce sugar content to reduce or eliminate their tax non-alcoholic beverage tax, increasing the tax on SSBs liability. The market response was significantly greater relative to sugar-free beverages (Thow et al., 2022). than initially anticipated, with dramatic reformulation More sophisticated tax designs may generate occurring even before the tax was implemented, resulting incentives for consumers to substitute towards lower in a reduction in the official revenue estimate to GBP 240 alcohol or sugar beverages. For example, specific taxes million (HM Treasury, 2018). Fifty days prior to on alcohol and SSBs may be based on the alcohol or implementation, the proportion of beverages with sugar sugar content rather than the volume of the beverage. content above the threshold had declined by 42 percent, This would result in lower effective taxes on lower alcohol whereas 50 days after implementation the proportion of or sugar content products. Firms may respond by passing beverages with sugar content above the threshold had through tax differentials to consumers which will increase declined by 67 percent (Scarborough et al., 2020). The tax 34 Lower alcohol or sugar beverages will become relatively cheaper than products with higher alcohol or sugar content. 35 Even though this system has coincided with a reduction in the average alcohol strength of beer and a reduction in alcohol consumption (Blecher, 2015), the tax rate has increased by 67 percent in real terms between 1998/99 and 2021/22, while tax revenues have increased by 113 percent during the same time (National Treasury, 2022). HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 13 raised GBP 318 million in the first year (HMRC, 2023), highlighting the significant impact that tax structures with strong incentives for firms to reformulate products to reduce sugar content can have on revenue, with a clear tradeoff between revenue and health impact. POLICY CONSIDERATIONS Several policy relevant lessons can be drawn from the cross-country evidence and experience with respect to tax revenues when implementing health taxes: • Health tax revenues are supported by well-designed policy. Specific taxes and uniform systems are more likely to generate larger and more stable tax revenues, while the use of ad valorem taxes and tiers are likely to result in smaller revenue generation or even a lack of meaningful revenue increase resulting from a tax increase. • The short-run impact on revenue of well-designed health tax reforms will be overwhelmingly positive in most, if not all countries. The impact of tax increases on revenue is very rapid, and increases can be sustained over long periods of time. • Questions regarding the sustainability of revenues should be acknowledged although it should not be considered a barrier to raising health taxes. In the long-run, the impact on revenue may weaken or decline. However, even countries with relatively high excise taxes continue to experience increases in tax revenues when increasing taxes. • Long-run declines in tax revenues should not be viewed as a policy failure but rather as a policy success, resulting in lower mortality, morbidity, and ultimately economic costs. • Health taxes generate unique tax administration challenges and revenue may be undermined by tax evasion. Improvements in tax administration may be required to support health tax reforms to ensure an optimal revenue impact. Concerns regarding tax compliance should not be a barrier to excise tax reforms. • Revenue generation is not the only goal of health taxes, and consideration should also be given to their impact on health. Furthermore, some health taxes or health tax reforms may be designed in a revenue neutral manner or to generate incentives for consumers or firms to reduce their tax liability and thus with revenue expectations declining over time. • While health taxes are an important source of revenue, it is important to be reminded that the revenues they generate should not be considered in isolation from the large economic savings and reductions in mortality and morbidity due to reduction in consumption of tobacco, alcohol, and SSBs. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 14 APPENDIX A: ADDITIONAL DATA Table A1 // Tobacco and alcohol tax revenues Tobacco tax revenue (percentage of) GDP Tax revenue Total revenue All HICs LMICs All HICs LMICs All HICs LMICs Count 81 40 41 69 37 32 69 36 33 Mean 0.6 0.6 0.5 3.0 2.8 3.2 1.9 1.6 2.3 Median 0.4 0.5 0.2 2.4 2.8 1.6 1.4 1.6 1.0 Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Maximum 3.4 3.4 2.3 11.2 9.6 11.2 8.5 3.3 8.5 CV 1.04 0.94 1.15 0.88 0.64 1.05 0.94 0.48 1.07 Alcohol tax revenue (percentage of) GDP Tax revenue Total revenue All HICs LMICs All HICs LMICs All HICs LMICs Count 78 40 38 66 37 29 66 36 30 Mean 0.3 0.3 0.3 1.8 1.3 2.3 1.2 0.8 1.7 Median 0.2 0.2 0.3 1.1 0.8 2.0 0.7 0.6 1.1 Minimum 0.0 0.0 0.0 0.1 0.1 0.1 0.1 0.1 0.1 Maximum 1.8 1.8 1.2 7.6 3.9 7.6 5.8 2.6 5.8 CV 0.98 1.07 0.91 0.90 0.80 0.84 1.00 0.81 0.87 Note: The CV (coefficient of variation) is ratio of the standard deviation to the mean and is a standardized measure of dispersion that shows the extent of variability in relation to the mean of the population. Source: World Bank GTP health tax revenue database Figure A1 // Beer excise taxes in the Figure A2 // Spirits excise tax and excise tax Philippines, 2012-2020 revenue in South Africa 1968-2023 50 16 400 Revenue (billions, constrant 2022 ZAR) PHP per pack (constant 2020 prices) Excise/LAA (constant 2022 ZAR) 45 14 350 40 12 300 35 30 10 250 25 8 200 20 6 150 15 4 100 10 5 2 50 0 0 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 1968 1977 1986 1995 2004 2013 2022 Low Medium High Premium Excise/LAA Revenue Note: Low (<₱14.5 per liter), Medium (₱14.5-22 per liter), and High (>₱22 per Note: Years represent fiscal years (e.g., 1968 is the 1968/69 fiscal years). A liter). uniform rate has been applied since 1998/99. Prior to that, multiple categories Source: Kaiser et al. (2016); Republic of the Philippines (2012; 2017); World with different, albeit similar tax rates, were applied. For simplicity, only one Bank World Development Indicators (CPI) category from each period is shown. Whiskey is used from 1968/69 to 1973/74 (other categories were wine brandy, grape brandy, and dop brandy), and “other spirits” from 1974/75 to 1997/98 (other categories were wine, sugarcane, and grain spirits. Source: National Treasury (2023); Statistics South Africa (CPI) HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 15 APPENDIX B: TECHNICAL ANNEX Tax pass-through explained alcohol tend for be more inelastic, they are more likely to be over shifted, particularly when specific taxes are As previously indicated, when specific taxes are employed applied since over shifting does not affect the effective over shifting results in a smaller increase in tax revenue tax. This is supported by empirical evidence, alcohol tends than a fully passed through tax increase. This is because to be fully passed-through or over shifted (Shang et al., over shifting will increase retail prices by more and result 2020; Nelson and Moran, 2020) whereas SSBs which are in sales decreasing by more than if the same tax increase generally elastic or more elastic are more likely to be was fully passed through. Since the effective tax is not under shifted (Andreyeva et al., 2022). However, since affected by the over shifting, the relatively lower sales over shifting of ad valorem taxes results in a larger volumes will result in relatively lower revenue. The effective tax, the incentive for firms to over shift tax opposite is also true and under shifting results in a larger increases is smaller, making ad valorem taxes less likely to increase in revenue since the increase in retail prices will be over shifting, and more so when applied later in the be smaller, resulting in a larger decrease in sales supply chain. compared to when taxes were fully passed through. The example of ad valorem cigarette taxes in Vietnam Furthermore, the degree of over and under shifting will showed how tax increases were not over shifted under an affect the magnitude of the relatively lower and higher ad valorem tax. As tax rates increased, the tax base revenue. declined resulting in the effective tax declining slightly, The opposite occurs when ad valorem taxes are coinciding with declining prices. The increase in tax rate employed, with over shifting resulting in a larger increase had a negligible impact on tax revenues. Contrast this to in tax revenue than a fully pass through tax increase. This South Africa where specific tax increases on cigarettes is because the over shifting results in a larger increase in and beer were over shifted, resulting in larger prices the effective tax rate for the same tax increase. Even increases and sales declines than if tax increases were though retail prices will increase by more when tax fully pass through. While revenue was lower than if the increases are over shifted compared to tax increases tax increases were fully passed through, revenue still being fully passed through, the relatively higher effective increased significantly since the over shifting had no tax will result in higher tax revenue. The opposite is also impact on the effective tax and the marginal impact of the true for under shifting of tax increases under ad valorem over shifting on prices (and consequently sales) was small taxes, with smaller revenue increases compared to a fully compared to the impact of the tax increase on prices (and passed through tax increase. Once again, the degree of consequently sales). over and under shifting will affect the magnitude of the The less competitive or more concentrated a market the relatively higher and lower revenue. higher the expected tax pass-through since firms have The location of the tax base in the supply chain has a greater pricing power. Conversely, over shifting is less significant impact in an ad valorem system. The later in likely (or under shifting is more likely) in more competitive the supply chain the tax base, the larger the effect on the or less concentrated markets where firms are more likely effective tax rate. If the tax base in later in the supply to compete on price. Products that are more inelastic are chain, then the relationship between the tax base and tax more likely to see higher degree of tax pass-through (over pass through is much closer. For example, if the tax is shifting) because the suppliers can compensate for falling based on the retail price and the tax rate increase, a quantity demanded with higher margins, whereas decline in the net-of tax price would imply an under shift, products that are more elastic are more likely to witness an increase in the net-of-tax price would imply an over lower degree of passed-through (under shifting). shift. If the tax base is located early in the supply chain, a shrinkage of the tax base to maintain the effective tax Price elasticity of demand would only result in an under shift if various margins later in the supply chain (e.g., at the wholesale or retail level) Policy makers and practitioners are concerned that remain unchanged. In general, ad valorem taxes generate reductions in consumption will reduce tax revenue. The weaker incentives for the suppliers to over shift the tax empirical evidence shows that tax revenues will increase increase since doing so will increase the effective tax. when tax increases drive down consumption due to the inelastic nature of demand. Demand is said to be inelastic The more inelastic the product, the more likely taxes are when the reduction in consumption is less than to be over shifted since there is a stronger incentive for proportional to the increase in price. For example, if the tax to be over shifted to compensate for declining consumption declines by 6 percent in response to a 10 volumes by increasing profit margins. Since tobacco and percent increase in price, the price elasticity of demand is HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 16 -0.6 and the demand is inelastic.36 If consumption were to decline by 12 percent after a 10 percent increase in price, then demand is elastic since the decline in consumption is more than proportional to the increase in price. The less elastic the demand for a product, the larger the increase in excise tax revenue for a tax increase, all else being held constant. Price elasticity will influence the tax revenue after a tax increase through its impact on sales volumes after a price change. Ultimately, the magnitude of revenue increase is primarily determined by the magnitude of the tax increase. Table A2 shows a conceptual example of this in practice. The less elastic the product, the less responsive the quantity demanded to higher tax (and price), and the higher the revenue. The example also highlights that increases in taxes generate positive incremental tax revenue even when a product has a unitary elasticity or is elastic since the tax is just a share in the price, and a relatively small share. Table A2 // Conceptual example of the impact of price elasticity of tax revenue Very inelastic Somewhat Unitary Elastic inelastic Price elasticity of demand -0.4 -0.6 -1.0 -1.2 Quantity 1,000 1,000 1,000 1,000 Excise tax (per unit) 0.20 0.20 0.20 0.20 Price (per unit) 1.00 1.00 1.00 1.00 Excise tax revenue 200 200 200 200 Excise tax increase (per unit) 0.10 0.10 0.10 0.10 New excise tax (per unit) 0.30 0.30 0.30 0.30 New price (per unit) 1.10 1.10 1.10 1.10 Percentage change in price 10% 10% 10% 10% Percentage change in quantity -4% -6% -10% -12% New quantity 960 940 900 880 New tax revenue 288 282 270 264 Increase in tax revenue 88 82 70 64 Note: Tax increase are assumed to be fully passed through to prices. The point elasticity formula is used in the example. 36 If the price elasticity is -0.8, then the 10 percent increase in price results price, however it is said to be less inelastic since the same price in an 8 percent decline in consumption. This is still inelastic since the increases results in a larger decline in consumption. decline in consumption is still less than proportional to the increase in HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 17 APPENDIX C: DATA SOURCES Country Product Source Alcohol Tobacco SSB Argentina ✓ ✓ ✓ OECD Armenia ✓ ✓ National data Australia ✓ ✓ *Australian Tax Office37 Austria ✓ ✓ EC Bangladesh ✓ National data Belgium ✓ ✓ ✓ EC; OECD Bhutan ✓ ✓ OECD Brazil ✓ ✓ OECD Bulgaria ✓ ✓ EC Burkina Faso ✓ ✓ OECD Cambodia ✓ ✓ National data Canada ✓ ✓ OECD Chile ✓ ✓ ✓ *Servicio de Impuestos Internos38; OECD Colombia ✓ ✓ OECD Congo, Dem. Rep. ✓ ✓ ✓ OECD Costa Rica ✓ ✓ ✓ OECD Cote d'Ivoire ✓ ✓ OECD Croatia ✓ ✓ EC Cyprus ✓ ✓ EC Czech Republic ✓ ✓ EC Denmark ✓ ✓ ✓ EC; OECD Dominican Republic ✓ ✓ OECD Ecuador ✓ ✓ ✓ OECD Egypt, Arab Rep. ✓ ✓ OECD El Salvador ✓ ✓ ✓ OECD Equatorial Guinea ✓ OECD Estonia ✓ ✓ EC Finland ✓ ✓ EC France ✓ ✓ EC Georgia ✓ ✓ OECD Germany ✓ ✓ EC Greece ✓ ✓ EC Guatemala ✓ ✓ ✓ OECD Guyana ✓ ✓ OECD Honduras ✓ ✓ ✓ OECD Hungary ✓ ✓ EU Indonesia ✓ National data Ireland ✓ ✓ EU Italy ✓ ✓ EU Japan ✓ ✓ OECD Kazakhstan ✓ ✓ OECD Korea, Rep. ✓ ✓ OECD Kosovo ✓ ✓ National data Latvia ✓ ✓ EC Lithuania ✓ ✓ EC Luxembourg ✓ ✓ EC Malaysia ✓ ✓ OECD 37 https://www.ato.gov.au/About-ATO/Research-and-statistics/In- 38 https://www.sii.cl/sobre_el_sii/serie_de_ingresos_tributarios.html detail/Taxation-statistics/ HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 18 Mali ✓ ✓ ✓ OECD Malta ✓ ✓ EU Mauritius ✓ ✓ ✓ OECD Mongolia ✓ ✓ OECD Morocco ✓ OECD Nauru ✓ ✓ OECD Netherlands ✓ ✓ EC New Zealand ✓ ✓ OECD Nicaragua ✓ ✓ OECD Niger ✓ ✓ OECD North Macedonia ✓ ✓ National data Norway ✓ ✓ ✓ OECD Panama ✓ ✓ ✓ OECD Philippines ✓ OECD Poland ✓ ✓ EC Portugal ✓ ✓ EC Romania ✓ ✓ EC Rwanda ✓ ✓ *Rwanda Revenue Authority Annual Reports39; OECD Senegal ✓ ✓ ✓ OECD Seychelles ✓ ✓ ✓ OECD Singapore ✓ ✓ OECD Slovak Republic ✓ ✓ EC Slovenia ✓ ✓ EC Solomon Islands ✓ ✓ OECD South Africa ✓ ✓ ✓ *National Treasury Budget Review40 Spain ✓ ✓ EC Sweden ✓ ✓ EC Switzerland ✓ ✓ OECD Thailand ✓ ✓ *Thai Revenue Department Annual Report41 Tokelau ✓ ✓ OECD Trinidad and Tobago ✓ ✓ ✓ OECD Tunisia ✓ OECD Turkiye ✓ ✓ National data* Uganda ✓ ✓ ✓ OECD Ukraine ✓ National data United Kingdom ✓ ✓ ✓ *HMRC Tobacco Bulletin42, Alcohol Bulletin43, and Soft Drinks Industry Levy Bulletin44 United States ✓ ✓ Office of Management and Budget (federal)45; US Census Bureau (state and local)46; Uruguay ✓ ✓ ✓ OECD OCED Venezuela, RB ✓ ✓ OECD 39 https://www.rra.gov.rw/en/public-information/annual-reports 40 https://www.treasury.gov.za/documents/national%20budget/2023/review/FullBR.pdf 41 http://webinter.rd.go.th/publish/24602.0.html 42 https://www.gov.uk/government/statistics/tobacco-bulletin 43 https://www.gov.uk/government/statistics/alcohol-bulletin 44 https://www.gov.uk/government/statistics/soft-drinks-industry-levy-statistics 45 https://www.whitehouse.gov/omb/historical-tables/ 46 US Census Bureau (2000), updated annually. Annual Survey of State and Local Government Finances, 1977-2020. Compiled by the Urban-Brookings Tax Policy Center. Washington, DC: Urban-Brookings Tax Policy Center. Tobacco available at: https://www.taxpolicycenter.org/statistics/state-and-local-tobacco-tax- revenue Alcohol available at: https://www.taxpolicycenter.org/statistics/state-and-local-alcohol-tax-revenue HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 19 REFERENCES African Business (2020) Coronavirus threaten Southern Africa’s tobacco industry. 28 May 2020. Available online: https://african.business/2020/05/economy/covid-19-threatens-southern-africas-tobacco-industry Allcott H, Lockwood B, Taubinsky D (2019a) Regressive Sin Taxes, with an Application to the Optimal Soda Tax. Quarterly Journal of Economics. 134(3): 1557-1626. Allcott H, Lockwood B, Taubinsky D (2019b) Should We Tax Soda? An Overview of Theory and Evidence. Journal of Economic Perspectives. 33(3): 202-227. Andreyeva T, Marple K, Marinello S, Moore TE, Powell LM (2022) Outcomes following taxation of sugar-sweetened beverages: A systematic review and meta-analysis. JAMA Network Open, 5(6):e2215276. Australian Bureau of Statistics (ABS) (2022) Consumer Price Index, September Quarter 2022. Canberra: Australian Bureau of Statistics. Action on Smoking and Health (ASH) (2022) Smoking costs society £17bn – £5bn more than previously estimated. Press release, 14 January 2022. London: Action on Smoking and Health. Barron K, Parry C, Bradshaw D, Dorrington R, Groenewald P, Laubscher R, Matzopoulos R (2023) Alcohol, violence and injury- induced mortality: Evidence from a modern-day prohibition. Review of Economics and Statistics. Forthcoming Blecher E (2015) Taxes on tobacco, alcohol and sugar sweetened beverages: Linkages and lessons learned. Social Science and Medicine. 136-137: 175-9. Blecher E, Le TT (2018) Vietnam Country Factsheet: Tobacco Tax Structures. Chicago: Tobacconomics, Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago. Blecher E (2019) Case Studies in Illicit Tobacco Trade: United Kingdom. Chicago: Tobacconomics, Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago. Brown J, Kirk-Wade E, Baker C, Barber S (2021) Coronavirus: A history of English lockdown laws. London: House of Commons Library. Chelwa G, van Walbeek C, Blecher E (2016) Evaluating South Africa’s tobacco control policy using a synthetic control method. Tobacco Control. Euromonitor (2021) Cigarettes in Vietnam. Euromonitor International. Euromonitor (2023) Cigarettes in Vietnam. Euromonitor International. Filby S, van der Zee S, van Walbeek C (2022) The temporary ban on tobacco sales in South Africa: lessons for endgame strategies. Tobacco Control. 31: 694-700. Fuchs A, Del Carmen G, Mukong A (2018) Long-run Impacts of Increasing Tobacco Taxes: Evidence from South Africa. Policy Research Working Paper 8369. Washington DC: World Bank. Fuchs A, Pierola (2022) The Distributional Impacts of Health Taxes. Washington, DC: World Bank. Gallet C, List J (2003) Cigarette demand: a meta-analysis of elasticities. Health Economics. 12: 821-835. Gallus S, Lugo A, Ghislandi S, La Vecchia C, Gilmore A (2014) Roll-your-own cigarettes in Europe: use, weight and implications for fiscal policies. European Journal of Cancer Prevention. 23(3): 186-92. Gertler P, Gracner T, Miranda R, Seira E (2021) Internalities and the effectiveness of taxing sugar-sweetened beverages. Working Paper. Global Burden of Disease Study (GBD) (2019) Results. Seattle: Institute for Health Metrics and Evaluation (IHME). Available from http://ghdx.healthdata.org/gbd-results-tool. Greenhalgh E, Bayly M, Scollo M (2021) Prevalence of smoking-adults. In Greenhalgh E, Scollo M, Winstanley M (eds.) Tobacco in Australia: Facts and issues. Melbourne: Cancer Council Victoria. Guindon G, Zhao K, Fatima T, Garasia S, Quinn N, Baskerville N, Paraje G (2022) Prices, tobacco and alcohol use: a systematic umbrella review. Addiction. 117: 3004-3023. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 20 HM Revenue and Customs (HMRC) (2016) Soft Drinks Industry Levy Consultation document. London: HM Revenue and Customs. HM Revenue and Customs (HMRC) (2022a) Tobacco Bulletin, October 2022. London: HM Revenue and Customs. HM Revenue and Customs (HMRC) (2022b) Soft Drinks Industry Levy statistics commentary, September 2022. London: HM Revenue and Customs. HM Revenue and Customs (HMRC) (2022c) Alcohol Bulletin, October 2022. London: HM Revenue and Customs. HM Revenue and Customs (HMRC) (2023) Soft Drink Industry Levy Bulletin, May 2022. London: HM Revenue and Customs. HM Treasury (2018) Soft Drinks Industry Levy comes into effect. 5 April 2018. News report: https://www.gov.uk/government/news/soft-drinks-industry-levy-comes-into-effect Inchauste G, Lustig N, Maboshe M, Purfield C, Woolard I (2015) The Distributional Impact of Fiscal Policy in South Africa. Policy Research Working Paper 7194. Washington DC: World Bank. International Monetary Fund (IMF) (2022) Republic of Nauru Staff Report for the 2021 Article IV Consultation and Informational Annex. Washington DC: International Monetary Fund. Joossens L, Merriman D, Ross H, Raw M (2010) The impact of eliminating the global illicit cigarette trade on health and revenue. Addiction. 105(9): 1640-9. Kaiser K, Bredenkamp C, Iglesias RM (2016) Sin tax reform in the Philippines: transforming public finance, health, and governance for more inclusive development. Washington DC: World Bank Group. Keen M (1998) The Balance between Specific and Ad Valorem Taxation. Fiscal Studies. 19(1): 1-37. Lane C, Blecher E, Ozer C, Bloom D, Prinz D (2023) Health Taxes and Inflation. Global Tax Program Health Taxes Knowledge Note Series. Washington: World Bank. Lightwood J, Glantz S (2016) Smoking Behavior and Healthcare Expenditure in the United States, 1992-2009: Panel Data Estimates. PLoS Medeicine. 13(5): e1002020. Linegar D, van Walbeek C (2018) The effect of excise tax increases on cigarette prices in South Africa. Tobacco Control. 27(1): 65-71. Muhammad A, Meade B, Marquardt DR, Mozaffarian D (2019) Global patterns in price elasticities of sugar-sweetened beverage intake and potential effectiveness of tax policy: a cross-sectional study of 164 countries by sex, age and global-income decile. BMJ Open, 9:e026390. National Treasury (2014) A review of the taxation of alcoholic beverages in South Africa. Pretoria: National Treasury, Republic of South Africa. National Treasury (2022) Budget Review. Pretoria: National Treasury, Republic of South Africa. Nelson J (2013) Meta-analysis of alcohol price and income elasticities - with corrections for publication bias. Health Economics Review, 3(1):17. Nelson J, Moran J (2020) Effects of Alcohol Taxation on Prices: A Systematic Review and Meta-Analysis of Pass-Through Rates. The B.E. Journal of Economic Analysis & Policy. 20(1): 21. Organization for Economic Co-operation and Development (OECD) (2022) Global Revenue Statistics Database. OECD Stat. Paris: Organization for Economic Co-operation and Development. Office for National Statistics (ONS) (2021) Adult Smoking Habits in the UK: 2020. London: Office for National Statistics. Office for National Statistics (ONS) (2022a) Consumer price inflation, UK: October 2022. London: Office for National Statistics. Office for National Statistics (ONS) (2022b) Adult Smoking Habits in the UK: 2021. London: Office for National Statistics. Palzes V, Parthasarathy S, Chi F, Kline-Simon A, Lu Y, Weisner C, Ross T, Elson J, Sterling S (2020) Associations Between Psychiatric Disorders and Alcohol Consumption Levels in an Adult Primary Care Population. Alcohol, Clinical and Experimental Research. 44(12): 2536-2544. Perdix J, Bovet P, Larue D, Yersin B, Burnand B, Paccaud F (1999) Patterns of alcohol consumption in the Seychelles Islands (Indian Ocean). Alcohol and Alcoholism. 34(5): 773-785. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 21 Petit P, Mansour M, Wingender P (2021) How to apply excise taxes to fight obesity. How to notes. World Bank: International Monetary Fund. Petit P, Nagy J (2016) How to design and enforce tobacco excises? How to notes. Washington: International Monetary Fund. Reddy P, Zuma K, Shisana O, Kim J, Sewpaul R (2015) Prevalence of tobacco use among adults in South Africa: Results from the first South African National Health and Nutrition Examination Survey. South African Medical Journal. 21;105(8): 648-55. Republic of the Philippines (2012) Tax Reform for Acceleration and Inclusion (TRAIN) Act (2012). Manila: Republic of the Philippines. Republic of the Philippines (2017) Act Restructuring the Excise Tax on Alcohol and Tobacco Products (2017). Manila: Republic of the Philippines. Reuters (2020) Thailand extends alcohol ban, health ministry says some measures could be eased. 20 April 2020. Available online: https://www.reuters.com/article/us-health-coronavirus-thailand-cases/thailand-extends-alcohol-ban-health- ministry-says-some-measures-could-be-eased-idUSKBN2220FY Ross H (2021) Lost Funds: A Study on the Tobacco Tax Revenue Gap in selected ASEAN countries. Bangkok, Thailand: Southeast Asia Tobacco Control Alliance. Ross H, Blecher E (2019) Ilicit Trade in Tobacco Products Need Not Hinder Tobacco Tax Policy Reforms and Increases . Tobacconomics White Paper. Chicago: Tobacconomics, Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago. Russell C, van Walbeek C (2016) How does a Change in the Excise Tax on Beer Impact Beer Retail Prices in South Africa? South African Journal of Economics. 84: 555-573. Scarborough P, Adhikari V, Harrington R, Elhussein A, Briggs A, Rayner M, Adams H, Cummins S, Penney T, White M (2020) Impact of the announcement and implementation of the UK Soft Drinks Industry Levy on sugar content, price, product size and number of available soft drinks in the UK, 2015-19: A controlled interrupted time series analysis. PLoS Medicine. (17)2: e1003025. Scollo M, Bayly M (2022) Tobacco taxes in Australia. In Greenhalgh E, Scollo M, Winstanley M (eds.) Tobacco in Australia: Facts and issues. Melbourne: Cancer Council Victoria. Sebeelo T (2023) Contested Terrains? The Politics of Alcohol Bans, Drinking Contexts, and COVID-19 in Botswana. Contemporary Drug Problems. Online first. Shang C, Ngo A, Chaloupka FJ (2020) The pass-through of alcohol excise taxes to prices in OECD countries. European Journal of Health Economics, 21(6):855-867. Sharma N, Chopra M, Bauld L, Nazar GP, Joshi N, Chugh A, Mohan S, Mohan D, Ali M, Mohan V, Tandon N, Narayan V, Reddy K, Prabhakaran D, Arora M (2023) Impact of a tobacco sales ban on the frequency of tobacco consumption in India during the COVID-19 pandemic. Tobacco Induced Diseases. (28)21: 51. Singh S, Sharma P, Balhara Y (2021) The impact of nationwide alcohol ban during the COVID-19 lockdown on alcohol use- related internet searches and behaviour in India: An infodemiology study. Drug and Alcohol Review. 40: 196-200. Sornpaisarn B, Shield K, Cohen J, Schwartz R, Rehm J (2013) Elasticity of alcohol consumption, alcohol-related harms, and drinking initiation in low- and middle-income countries: A systematic review and meta-analysis. International Journal of Alcohol and Drug Research, 2(1), 45-58. Sornpaisarn B, Shield K, Cohen J, Schwartz R, Rehm J (2015) Can pricing deter adolescents and young adults from starting to drink: an analysis of the effect of alcohol taxation on drinking initiation among Thai adolescents and young adults. Journal of Epidemiology and Global Health. 5: S45-57. Steyn K, Bradshaw D, Norman R, Laubscher R, Saloojee Y (2002) Tobacco use in South Africans during 1998: the first demographic and health survey. Journal of Cardiovascular Risk. 9(3): 161-70. Thow AM, Rippin H, Mulcahy G, Duffey K, Wickramasinghe (2022) Sugar-sweetened beverage taxes in Europe: learning for the future. European Journal of Public Health. 32(2): 273-280. HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 22 United Nations (2018) Sustainable Development Goal 6: Synthesis Report 2018 on Water and Sanitation. New York: United Nations. U.S. National Cancer Institute and World Health Organization (NCI and WHO) (2016) The Economics of Tobacco and Tobacco Control. National Cancer Institute Tobacco Control Monograph 21. Bethesda: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, and Geneva: World Health Organization. Van der Zee K, Filby S, van Walbeek C (2023) When Cigarette Sales Suddenly Become Illegal: Evidence From an Online Survey of South African Smokers During COVID-19 Lockdown. Nicotine and Tobacco Research. 25(5): 325-330. Vellios N, van Walbeek C, Ross H (2020) Illicit cigarette trade in South Africa: 2002–2017. Tobacco Control. 29: s234-s242. Vellios N, van Walbeek C, Ross H (2022) Measuring the illicit cigarette market in the absence of pack security features: a case study of South Africa. Tobacco Control. 31:580-585. Van Walbeek C (2002) Recent trends in smoking prevalence in South Africa – Some evidence from AMPS data. South African Medical Journal. 92(6): 468-72. Van Walbeek C (2020) Case Studies in Illicit Tobacco Trade: South Africa. Tobacconomics Fact Sheet. Chicago: Tobacconomics, Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago. World Bank (2018) Economics of Tobacco Taxation Toolkit. Washington: World Bank. World Bank (2020) Tax on Sugar-Sweetened Beverages: International Evidence and Experiences. Washington: World Bank. World Bank (2023a) World Bank Group Support to Small States. Washington: World Bank. World Bank (2023b) Global SSB Tax Database. Washington: World Bank. World Customs Organization (WCO) (2021) Illicit Trade Report 2021. Brussels: World Customs Organization. World Health Organization (2017). Resource tool on alcohol taxation and pricing policies. Eds. Sornpaisarn B, Shield K, Österberg E, Rehm J. Geneva: World Health Organization. World Health Organization (WHO) (2018) Global status report on alcohol and health 2018. Geneva: World Health Organization. World Health Organization (WHO) (2021a) WHO report on the global tobacco epidemic 2021: addressing new and emerging products. Geneva: World Health Organization. World Health Organization (WHO) (2021b) Technical Manual on Tobacco Tax Policy and Administration. Geneva: World Health Organization. World Health Organization (WHO) (2023) WHO manual on sugar-sweetened beverage taxation policies to promote health diets. Geneva: World Health Organization. This Knowledge Note Series is funded by the World Bank’s Global Tax Program (GTP). More information: https://www.worldbank.org/en/programs/the-global-tax-program globaltaxprogram@worldbank.org HEALTH TAXES • GLOBAL TAX PROGRAM • WORLD BANK • Unpacking the empirics behind health tax revenue Page 23