Skip to document

Czaa168 - Need to explore my self Analyzing the effectiveness of an intervention is a

Need to explore my self Analyzing the effectiveness of an intervention...
Course

Science, Technology and Society (GEC103)

210 Documents
Students shared 210 documents in this course
Academic year: 2021/2022
Uploaded by:
0followers
0Uploads
0upvotes

Comments

Please sign in or register to post comments.

Preview text

Sin taxes and their effect on consumption,

revenue generation and health improvement: a

systematic literature review in Latin America

Aurelio Miracolo

1

, Marisa Sophiea

2

, Mackenzie Mills

1

and

Panos Kanavos

1,

*

1 Department of Health Policy and Medical Technology Research Group - LSE Health, The London School of

Economics and Political Science, Cowdrey House, Houghton Street, London WC2A 2AE, UK

2 Faculty of Medicine, School of Public Health, Imperial College, London, Medical School Building, St Mary’s

Hospital, Norfolk Place, London W2 1PG, UK

*Corresponding author. Department of Health Policy and Medical Technology Research Group - LSE Health, The London

School of Economics and Political Science, London WC2A 2AE, UK.

E-mail: p.g@lse.ac

Accepted on 13 November 2020

Abstract

Sin or public health taxes are excise taxes imposed on the consumption of potentially harmful

goods for health [sugar-sweetened beverages (SSBs), tobacco, alcohol, among others], aiming to

reduce consumption, raise additional revenue and/or improve population health. This paper

assesses the extent to which sin taxes (a) can reduce consumption of potentially harmful goods,

(b) raise revenue for national health systems and (c) contribute to population health in Latin

America. A systematic literature review was conducted on peer-reviewed and grey literature; end-

points included: impact of raising sin taxes on consumption, ability to raise revenue for health and

the possibility of population health improvements. Risk of bias for each study was assessed. The

synthesis of the literature on sin tax implementation showed improvements in all three endpoints

across the study countries. Following the introduction of sin taxes or by simulating their potential

impact, nearly all studies explicitly reported that consumption of potentially harmful goods (mainly

SSBs and tobacco) declined; revenue was found to have increased in almost all countries, suggest-

ing that there may be additional scope for further tax increase. Simulated improvements in popula-

tion health have also been shown, by demonstrating a relationship between sin tax increases and

reduction in prevalence of diabetes, stroke, heart attacks and associated deaths. However, sin tax

effects on health would be better quantified over the long-term. Data quality and availability chal-

lenges did place some limitations on sin tax impact assessment. Sin taxes can be effective in reduc-

ing consumption of potentially harmful goods, improve population health and generate additional

revenue. Promoting further research on this topic should be a priority.

Keywords: Sin taxes, health financing, fiscal space, Latin America, earmarked taxation, universal health coverage

Introduction

Background

Sin taxes, or public health taxes, are defined by the World Health Organisation as excise taxes targeting goods that can be detrimental to the health of the population (WHO, 2004). These goods include to- bacco products, alcohol, sugar-sweetened beverages (SSBs), which are

drinks with added sugar, such as soft drinks, tea, flavoured coffee, juice and sports drinks. The harmful impact of these goods is well known and is evidenced by research (Cnossen, 2005); for instance, to- bacco consumption is linked to an increased risk of developing cardio- vascular disease (CVD), respiratory disease, cancer and other non- communicable and chronic diseases (U. Department of Health and

VCThe Author(s) 2020. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (creativecommons/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

doi: 10/heapol/czaa

Review

Health Policy and Planning, 36, 2021,

Advance Access Publication Date: 22 April 2021

790– 810
790

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Human Services, 2014), while elevated SSB consumption is generally associated with an increased risk of developing CVD, metabolic dis- ease and obesity (Maliket al., 2013;Arsenaultet al., 2017). Published evidence has demonstrated the effect of sin taxes on con- sumer behaviour, health outcomes and on revenue generation for health systems (Wrightet al., 2017). Although differences in sin tax application and outcomes are present between low- and middle- income countries (LMICs) vs high-income countries (HICs), evidence has shown that the application of these taxes can have a significant ef- fect on consumption patterns and the well-being of the population, while being financially sustainable (Goodchildet al., 2016). The inverse relationship between increases in sin taxes and con- sumption is also well established for the consumption of SSBs (Colcheroet al., 2017). Research related to health and behaviour con- nected to SSBs intake has been conducted in HICs (Claroet al., 2012) reporting that consumption of SSBs instead of zero-calorie beverages can lead to excess weight and obesity. This has raised concerns over SSB consumption in LMIC settings where research is more limited. From an economic standpoint, excise duties are a form of indir- ect taxation, in that they are levied on goods or services rather than on firms or personal incomes. This gives them greater capacity to shape consumer behaviour. Sin taxes can be applied in two different ways: per unit (defined as a fixed amount for each unit of a good or service sold, such as dollars per kilogram) or ad valorem (levied on spending and set as a percentage of the value added by a firm, as is the case of a value-added tax (VAT)). With the former, the tax is represented by a fixed amount per unit, while with the latter, the tax is made up of a fixed percentage per unit. Sin taxes represent one way of raising revenue and, through that, creating fiscal space (FS). The revenue-generating capacity of sin taxes can help countries increase expenditure by creating additional FS (Heller, 2006), which, in turn, allows countries to direct financial resources to public spending without depressing other items of ex- penditure or by destabilizing budget equilibria. An analytical framework of the possible policies that can be adopted for the creation of FS in the health sector has been estab- lished (Heller, 2006;PAHO, 2015); this includes, first, the promo- tion of conducive macroeconomic conditions; second, a reprioritization of health expenditure; third, the improvement of ef- ficiency in existing health expenditure; fourth, increasing the effi- ciency of tax collection; fifth, a recourse to external aid (grants, loans); and sixth, the creation of new tax revenues through a greater tax burden (PAHO, 2015).Latin American taxation on goods such as tobacco, alcohol and sugar, which are potentially harmful for general health, is considerably lower than the average in Organization for Economic Cooperation and Development (OECD) countries (PAHO, 2015) and, as such, represents a valid policy choice for Latin American countries, since they can simultaneously generate revenue as well as influence consumer behaviour and, by implication, population health.

Latin America is considered an area with relatively high levels of consumption of products which can prove harmful to public health (tobacco, alcohol, saturated fat). Twenty per cent of people under 20years of age are overweight or obese in the region (Cominato et al., 2018), while this percentage exceeds 50% among Mexican and Peruvian adults (Kainet al., 2014;Batiset al., 2016;Colchero et al., 2017). Furthermore, an overall high prevalence in tobacco consumption is recorded in the region: only Ecuador, Peru, Bolivia and Paraguay report a consumption of<500 cigarettes per capita per annum, while in all other Latin American countries tobacco con- sumption ranges between 500 and 1500 cigarettes per capita per annum (Muller, 2008). Given the significant consumption of poten- tially harmful goods, the associated negative impact on health in Latin America, and considering the opportunities outlined in the FS framework (PAHO, 2015), the purpose of this paper is to assess the impact of sin tax implementation in the Latin American region. A systematic literature review is conducted for this purpose. While the impact of sin taxes has been investigated at country level in some Latin American countries (Mejiaet al., 2008;Claroet al., 2012; Curtiet al., 2015;Batiset al., 2016) or countries outside the Latin American region (White and Ross, 2015), including HICs (Wright et al., 2017), comparative evidence of this type of taxation at region- al level, and, specifically, in the Latin American context, where there may be economic and cultural similarities amongst the countries in the region, is missing. While the effect of sin taxes in HICs is well established (Wrightet al., 2017), it is unclear if these findings trans- late to Latin America, where there are differences in policy prior- ities, policy processes and fiscal commitments. There is no study that analyses and pulls together any available evidence on the im- pact of sin tax introduction in Latin America, a continent dominated by middle-income countries, where public investment in health is in the majority of cases low as proportion of gross domestic product (GDP) and where increases in spending are required in order to com- ply with universal health coverage pledges. The paper, therefore, contributes to the discussion of whether sin taxes have any effect on tax revenue and consumption of potentially harmful products, im- pact health impact and, broadly speaking, contribute to healthcare financing.

Methods

Approach and endpoints

A systematic literature review (SLR) has been conducted to investi- gate the impact of sin taxes in the Latin American region. The geo- graphical scope of the study included the South American continent, the Spanish-speaking countries of continental central America and excluded the Caribbean region. Three endpoints were considered: first, a consumption endpoint, examining whether the application of excise taxes has had any effect on the demand for goods (i. SSBs, unhealthy food, tobacco, alcohol); second, a revenue endpoint,

KEY MESSAGES
  • This is the first systematic literature review assessing the effect of sin taxes on consumption, fiscal space generation and their impact on population health in Latin America.
  • Reduction in harmful goods consumption (81% of studies), positive effects on revenue generation (71%) and on health outcomes (82%) are key outcomes.
  • There is still room for further tax increases where sin taxes have been adopted.
  • Further research is needed to improve data collection for a more comprehensive analysis of the impact of sin taxes

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Data synthesis

Findings are grouped under the three endpoints, (1) effect on con- sumption, (2) effect on revenue generation and (3) health impact. In each of the three endpoints, there was a further division, where possible, relating to the type of good (e. SSBs, tobacco, alcohol). The study’s PROSPERO identification number is CRD42018096210.

Results

Study characteristics

The PRISMA flowchart (Figure1) shows the number of studies included in our review and how they are arrived at. In the initial stage of the systematic review, 1321 studies were found across all databases. Following the screening process and by applying the ex- clusion criteria, 34 studies were included in the review. Of the 34 included studies, 27 addressed the consumption end- point, 6 the health endpoint and 10 the revenue generation end- point; 9 studies addressed multiple endpoints. There were no randomized control trials (RCTs) amongst the included studies.

With regards to the intervention, 13 studies focused on SSBs and high energy density foods. This included the excise tax on SSBs (1 peso/L) and the 8% sales tax on foods implemented in Mexico in January 2014 and the SSBs excise tax in Brazil. Twenty-three studies were related to the taxation of tobacco products. Countries involved in the analysis included Mexico, Argentina, Brazil, Uruguay, Ecuador, Peru, Colombia and Panama. Two studies analysed the intervention on a continental and multi-country level (Garce ́set al., 2014;Goodchildet al., 2017 ). Studies on tobacco focused mainly on the change in de- mand for tobacco, the impact on price caused by the tax imple- mentation and the main features related to the demographic and epidemiological context in which these policies are operating. Alcohol was assessed in just one study, together with the analysis of tobacco demand in Ecuador (Chave ́z, 2016). The SLR included mostly observational studies and to a lesser extent narrative reviews. Most of the included literature focused on studies analysing consumption, and the main goods of interest were, first, tobacco, its demand, and the role of illicit trade and, second, SSBs and their impact on all three endpoints. Studies displayed sig- nificant variety in the population included, data sources, and

Figure 1PRISMA flowchart.

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

evaluation methods for the specific tax of interest, as well as the evaluation of the specific tax of interest. Country differences in tax- ation systems, sin tax structure and levels of stakeholder involve- ment have added complexity to our analysis. The dominance of observational studies and the absence of other study designs (e. RCTs) is the result of the type of argument addressed and the re- quirement of wide population cohorts, which represent the national trend and must not be criticized as a source of low-quality evidence (Pindyck et al., 2018). Table2 outlines the characteristics of included studies (endpoint, publication outlet, national setting, population, data sources indicator of interest).

Effect on consumption

SSBs and unhealthy foods The effect of sin taxes on consumption of SSBs was addressed by 12 studies. Nine of these were related to the implementation of SSBs taxes in Mexico, two focused on taxation of high-sugar content bev- erages in Chile and one investigated the potential relationship be- tween SSB prices and levels of consumption in Brazil. The literature focused on Mexico due to the high levels of SSB consumption. Before tax implementation, Mexico had the highest worldwide soft drinks consumption (163 litres per capita) in 2011 (Colcheroet al., 2016b). In January 2014, Mexican government introduced a tax of 1 Mexican peso per litre on all sugary non- alcoholic beverages, i. sodas, flavoured waters, sweetened dairies, teas and energy drinks with added sugars, but excluded drinks con- sisting of 100% juice and beverages with artificial sweeteners (Claro et al., 2012). This caused an 11% price increase in carbonated SSBs and circa a 10% price increase in non-carbonated SSBs, compared with prices in 2013 (Colcheroet al., 2016a). At the same time, Mexico introduced an 8% ad valorem tax on non-essential highly energy-dense foods (with at least 275 calories per 100g) (Colchero et al., 2016b). Six studies analysed the changes in consumption caused by the implementation of the SSB tax (1 peso/l) in Mexico. The common aim of these studies was to understand how consumer behaviour would change following the tax introduction. This was achieved by investigating different data sources, notably, Nielsen’s Mexico Consumer Panel services (henceforth Nielsen Panel), that collects data on households’ monthly purchases and covers 63% of the Mexican population, and the Mexican National Health and Nutrition Survey based on questionnaire responses and manufactur- ing sector data, particularly the ‘Economic Behaviour of the Industries in the Country’ (EMIM) database. All six studies high- lighted that the introduction of the specific SSB tax increased the price of SSB products approximately by 10% in 2014 compared with 2013. Results from one study (Colcheroet al., 2017) showed a decrease in SSB purchases of 5% in 2014 and 9% in 2015 (aver- age reduction of7% in 2014–15) compared with the 2012– period. Another study (Colcheroet al., 2016b) based on the same source found a change in SSB purchases of6% in 2014 compared with 2012–13. The reduction was higher in low socioeconomic sta- tus (SES) groups, relative to medium and high SES groups (9% vs5% vs5%, respectively). Another study (Nget al., 2019), based on the Nielsen Panel, divided the study population in four groups encompassing all possible consumers of taxed and untaxed beverages: (1) those who had higher (H) purchases of taxed (T) bev- erages and lower (L) purchases of untaxed (U) beverages (HTLU— and whose consumption choices were considered unhealthier), (2) those who had higher (H) purchases of taxed (T) and higher (H) pur- chases of untaxed (U) beverages (HTHU—whose consumption

choices were also considered unhealthier), (3) those who had lower (L) purchases of taxed (T) and lower (L) purchases of untaxed (U) beverages (LTLU—whose consumption choices were considered healthier) and (4) those who had lower (L) purchases of taxed (T) beverages and higher (H) purchases of untaxed (U) beverages (LTHU—whose consumption choices were also considered health- ier). The study compared the pre-tax behaviour of these groups with their consumption levels after the SSB tax implementation. Among others, results showed that, following the SSB tax implementation, the HTLU-unhealthier and HTHU groups (both considered to be ‘unhealthy’ in their consumption choices), reduced their consump- tion of taxed beverages both in absolute and relative terms and, at the same time, increased their consumption of untaxed beverages. It has been shown that the greatest effect of this consumption shift from taxed to untaxed beverages was observed in the lowest socioe- conomic group. A further study (Colcheroet al., 2016a) using an al- ternative data source, notably, manufacturing industry data (EMIM) analysed the changes in SSB and plain water sales in 2014 and 2015 (using the pre-tax period, 2007–13, as a counterfactual). Results suggested a decrease in SSB per capita sales of 7% and an increase of 5% in plain water per capita sales in the 2014– period compared with the counterfactual, reporting an association of the tax implementation with the changes in per capita sales. Overall, results of the studies assessing SSB tax implementation in Mexico reported a decrease in the consumption of taxed SSBs, and that the tax mildly shifted purchases towards untaxed beverages or other products. Some studies (Colcheroet al., 2016b, 2017 ;Wright et al., 2017) pointed out that effects of tax implementation may be more substantial in the long-term rather than the short-term. This would be because human habit formation is gradual and changing behaviour in light of increased taxation may take time to shape (Colcheroet al., 2017;Wrightet al., 2017). Additionally, following tax implementation consumers may switch to cheaper untaxed bev- erages and this pattern could be better seen over the longer term (Colcheroet al., 2016b). The results (measures, intervention and counterfactual) included in the above studies were adjusted for dif- ferent indicators, mainly seasonality of beverage consumption and socioeconomic factors. Without such adjustments, the results would have been biased by temporary factors. Ortega-Avilaet al.(2018)examined how the implementation of the tax was perceived by a cohort of adolescents. A qualitative study explored the awareness and perception on the introduction of the SSB tax within a cohort of Mexican adolescents, reported most of them were unaware of this policy and that they perceived the 1 peso/ l increase as not high enough to shift their preferences and SSB con- sumption patterns. For those interviewed, alternatives to costly SSB products would mainly be homemade drinks. The study underlined that the impact of the tax could be misperceived by some segments of the population and that this would represent a limitation in changing citizens’ attitudes towards these products. Another study (A ́lvarez-Sa ́nchezet al., 2016) focused on the awareness of Mexicans on the SSB tax introduction. Based on questionnaire sur- vey data of>6,000 adults, the study found that people’s awareness and decrease in consumption were directly proportional, i. people who were aware of the tax introduction were more inclined to de- crease their SSB intake. Three studies (Batiset al., 2016;Taillieet al., 2017;Herna ́ndez et al., 2019) focused on the 8% ad valorem tax on non-essential and energy foods in Mexico. One study (Batiset al., 2016) analysed the difference in the volume purchase of taxed and untaxed packaged food between observed data in 2014 and their respective

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Table 2

(continued)

Author(s)

Endpoint

Year Journal

Setting

Number of participants

Data source

Participant age

Tax/indicatorof interest

Goodchild

et al.

(2017)

Consumption/

revenue

2017

Pan American

Journal ofPublic Health

Latin America

andCaribbean

Data on number of partici-

pants not specified. Dataon taxes and prices for a20-cigarette pack werederived from the WHOreport on the global to-bacco epidemic. Share oflicit (i. tax paid) cigar-ette retail sales in eachcountry was collectedfrom GlobalData andEuromonitor

WHO Bulletin 2016,

WHO report on theglobal tobacco epidem-ic, Global Data andEuromonitorInternational

Data on participants’ age

not specified.

Tobacco excise

taxes

Guerrero-Lopezet al.

(2013)

Consumption

2013

Salud Publica de

Mexico

Mexico

66,684 individuals fromENSA 2000; -

70,297 individuals fromENSANUT 2006; and -

67,786 individuals fromENSANUT 2012

Encuesta Nacional de

Salud (ENSA) 2000;Encuesta Nacionales deSalud y Nutricio

`n

(ENSANUT) 2006 and 2012

Population of



10 years

of age

Tobacco

Herna ́ndezet al.

(2019)

Consumption

2019

Preventive

Medicine

Mexico

154,777 individuals, col-

lected in five waves ofENIGH survey (2008,2010, 2012, 2014 and2016)

National Household

Income andExpenditure Surveys(ENIGH)

General population

8% tax on energy-

dense, nutrient-poor foods

Iglesias (2016)

Consumption/

revenue

2016

Revista

Sudamericanade SaludPublica

Brazil

Not applicable

Not applicable

Not applicable

2011 tobacco tax

Iglesias

et al.

(2017)

Consumption/

revenue

2016

Tobacco Control

Brazil

37,317 individuals

GATS (Global Adult

Tobacco Survey) 2008and 2013 survey

Adults of



18 years of

age

2012 cigarette ex-

cise tax

Jan

et al.

(2014)

Health

2014

PLoS One

Panama

2,191 individuals

National Institute of

Statistics and Census ofPanama

Population of



30 years

of age

TTI

Kostova

et al.

(2014)

Revenue

2014

Tobacco Control

Brazil,

Mexico,Uruguay

45,838 individuals

GATS (Global Adult

Tobacco Survey)

Population of



15 years

of age

Tobacco

Maldonadoet al.

(2016)

Consumption

2016

Pan American

Journal ofPublic Health

Colombia

Estimates on country

population

National Directorate of

Taxes and Customs ofColombia

Population of



15 years

of age

Tobacco

Martinez

et al.

(2015)

Consumption

2013

British Medical

Journal

Argentina

Number of participants not

specified. Cigarette

Ministry of Economics

and Production of

Adults of



14 years of

age

Tobacco tax

(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Table 2

(continued)

Author(s)

Endpoint

Year Journal

Setting

Number of participants

Data source

Participant age

Tax/indicatorof interest

consumption dataderived from total salesof cigarettes to the publicreported by the Ministryof Economics andProduction. Populationdata were collected fromthe INDEC

Argentina; NationalInstitute of Statisticsand Census (INDEC)

Nakamuraet al.

(2018)

Consumption

2018

PLoS Medicine

Chile

2,836 households

Kantar Worldpanel

General population

SSBs

Ng

et al.

(2019)

Consumption

2018

Public Health

Nutrition

Mexico

6,089 households

Nielsen’s Mexico

Consumer PanelServices

General population

SSBs

Ortega-Avilaet al.

(2018)

Consumption

2018

Public Health

Nutrition

Mexico

29 individuals

Semi-structured

interviews

Adolescents (from 15–

19 years old)

SSBs

Rodriguez-Iglesias

et al.

(2016)

Consumption

2016

Argentine Journal

of Cardiology

Argentina

Number of participants not

specified. Intensity andprevalence of tobaccouse were derived fromthe National Risk FactorSurvey

National Risk Factor

Survey (2005, 2009and 2013)

Participants’ age not

specified.

Tobacco tax

Saenz-de-Mieraet al.

(2010)

Consumption

2010

Tobacco Control

Mexico

1,079 individuals

International Tobacco

Control PolicyEvaluation Survey(ITC-Mexico)

Adults of



18 years of

age

Tobacco

Szklo

et al.

(2018)

Consumption

2018

American Journal

of PublicHealth

Brazil

Gathers information from

different surveys

GATS (Global Adult

Tobacco Survey)Brazil; VIGITEL;Secretariat of FederalRevenues of Brasil

General population

Illicit tobacco

consumption

Taillie (2017)

Consumption

2017

Preventive

Medicine

Mexico

6,089 households collected

in three waves

Nielsen Company’s

Mexico ConsumerPanel

General population

SSBs

Studies using simulation techniques

Bardach

et al.

(2016)

Health/revenue

2016

Revista Peruana

de MedicinaExperimental ySalud Pu` blica

Peru

30,000,000 observations

(individuals as part of aMonte Carlo simulation)

Databases (Medline,

Embase, Central,SocINDEX, Econlit,LILACCS, NBER,CRD)

Adults of



35 years of

age

Tobacco

James

et al.

(2019)

Health/revenue

2019

Tobacco Control

Colombia

Number of participants not

specified. Median cigar-ette price obtained from

Administrative

Department ofNational Statistics of

Participants age’ not

specified. Median cig-arette price obtained

Tobacco tax

(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

counterfactual (2012–13). The study showed that, in 2014, the con- sumption on purchased food was 467g/per capita/year, compared with the 492g consumed food predicted by the counterfactual, with the mean volume of taxed food purchases decreasing by 5%. At the same time, non-significant variation was found between observed and counterfactual volumes of untaxed food purchases. A difference in consumption between SES groups was detected as well. For the low SES, there was a decrease of 10%, while for medium SES the decrease stood at 5%. Interestingly, no change in con- sumption was found in the high SES. However, the study pointed out that it was difficult to infer a causality between the tax imple- mentation and the consumption changes due to database limitations in terms population representativeness (data mainly concentrated in urban areas), and the 2 years’ counterfactual could be considered limited in evidencing changes in consumption patterns. Results from the second study (Herna ́ndezet al., 2019) were in the same direc- tion, recording a decrease of 5% on taxed food purchases in 2014–16 compared with 2008–12. At the same time, untaxed food consumption increased by 2% during the same period. The last study focused on the 8% ad valorem tax in Mexico (Taillieet al., 2017) and was based on the Nielsen’s panel. It ana- lysed how different types of households (low/high income) and con- sumers (with healthy/unhealthy behaviours or diet) reacted to this tax, by implementing a pre–post study design (2012, prior to tax im- plementation, to 2015, post-tax implementation). The study reported that the total volume of taxed products purchased declined by 4% in 2014 and by 14% in 2015, while the untaxed purchase changes were higher in 2014 (þ2%) but declined in 2015 (4%). The household subgroup analysis reported that, in the post-tax implementation period (2014–15) compared with the pre- tax period (2012–13), the low-income household group consump- tion decreased by 1%, the high-income household consumption (i. those purchasing a lot of both taxed and untaxed products), decreased by 1%; consumers, whose consumption patterns were considered to be ‘unhealthy’ (i. consuming more taxed products and less untaxed products) decreased their total consumption by 4%, while consumers whose consumption patterns were consid- ered to be ‘healthy’ (i. consuming more untaxed products and less taxed products) registered no differences in the post-tax period. Overall, the study reported a higher decrease in the second year after the implementation, compared with the first. The authors argue that this could be caused by many factors, probably by a gradual shift in consumer habits or by awareness campaigns on the harmful health impact of these products. The major gap between healthy and un- healthy households in consumption patterns might be explained by the fact that healthy consumers are already less inclined to buy harmful foods compared with those used to buy them. The study confirmed a trend of reduction in the consumption of energy-dense ultra-processed foods after tax implementation in Mexico. Two studies (Caroet al., 2018;Nakamuraet al., 2018) analysed the impact of the “Impuesto Adicional a las Bebidas Analcoho`licas” (IABA) related to SSBs in Chile, which was implemented in October 2014. Specifically, in 2014, there was an increase in the tax rate from 13% to 18% on beverages with high levels of sugar (H-SSBs), defined as beverages with>6 g of sugar per 100ml. Conversely, there was a tax decrease for beverages containing<6 g of sugar per 100ml. Both studies showed a decrease of H-SSBs consumption in the post-increase period, compared with the pre-increase period. Caroet al.(2018)reported a monthly per capita decrease in H-SSBs purchases of 3% by volume, and 4% by calories, while the volume of L-SSBs increased of 10%, based on a post-increase period from November 2014 to December 2015 and a pre-increase period, as

counterfactual, from 2013 to October 2014 al. (2018)also reported an H-SSBs monthly purchased decrease of 21%, by comparing the post-increase period (November 2014 to December 2015) to a pre-increase period that started in 2011. However, both studies agreed that the small increase in the SSB tax did not impact the population significantly, and that based on the small cohort observed and the short post-tax period it was not pos- sible to assess the causal effect of the tax. In addition to the research focusing on Mexico and Chile, an- other study (Claroet al., 2012) attempted to evaluate price and in- come elasticity related to SSBs in Brazil. Although, strictly speaking, not a taxation study, the study simulated the effects on consumption of a 1% increase in price and 1% increase in income and analysed SSB taxation practices in Brazil; the study reported that a 1% in- crease in the price would cause a 0% reduction in SSBs product consumption. Additionally, changes in family income would influ- ence SSBs consumption: for a 1% increase in family income there would be a corresponding 0% increase in SSBs consumption. Overall, poor households in Brazil would be more than twice as like- ly, relative to wealthy households, to change their consumption pat- terns if price and income changed. The study, however, underlined how these estimates were based only on home food and beverage consumption, approximately accounting for 76% of total household expenditure, leaving almost a quarter of purchasing patterns un- accounted for by the analysis.

Tobacco Fifteen studies evaluated various aspects of tobacco use, i. the ef- fect of tax implementation on consumer behaviour, the role of illicit tobacco product consumption, how price and income elasticity were shaped in each country and how elasticity could potentially change or was found to change following tax implementation. Mexico was included in four studies; the country dealt with a tobacco-related re- form process which commenced after the ratification of the Framework Convention on Tobacco Control (FCTC) in 2004 and lasted for nearly a decade. Mexico is considered to be a country with a heavy burden of tobacco-related ill-health, reporting a smok- ing rate of 14% among Mexican adults (WHO, 2015). Three of the identified studies (Saenz-de-Mieraet al., 2010;Guerrero-Lopez et al., 2013;Reynales-Shigematsuet al., 2015) focused on the effect of the new tax structure (updated to 2011) on tobacco consumption levels, through country-level surveys and self-reported price of ciga- rettes by consumers. The research mainly underlined how smoking rates declined by 30% during 2002–15, how adolescent and adult groups reduced tobacco consumption in response to the specific ex- cise tax introduction, and how the reform process uniformly affected all sociodemographic groups. A narrative review on Argentina (Goodchild et al., 2016) reported that tobacco affordability rose by 100% between 1997 and 2007, whilst the country experienced sharp economic growth. The study offered significant insights on how the introduction of an ex- cise tax on tobacco would significantly reduce smoking prevalence (it was assumed that a 10% price increase would reduce the preva- lence by 3%). Another study (Ferranteet al., 2007) used a tobacco policy simulation model to evaluate how policies introduced in Argentina, relating to advertising, promotion and sponsorship bans, would have an effect on consumption. The study reported that these policies, regardless of the low level of taxes on cigarettes compared with HICs, produced a relative reduction in tobacco consumption in 2004 compared with 2001.

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

The literature also provides evidence on the extent of ‘illicit con- sumption’ of tobacco products and the effect of overall illicit smok- ing prevalence. Illicit consumption refers to consumption of tobacco products not legally purchased (e. counterfeit cigarettes). Three studies (Iglesias, 2016;Iglesiaset al., 2017;Szkloet al., 2018), all from the Brazilian context, estimated how illicit cigarette consump- tion changed after the excise tax implementation in 2012, using na- tional surveys (GATS-Brazil, Vigitel). The studies researched how the excise tax implementation impacted the overall proportion of il- licit cigarette use among smoking population or on illicit smoking prevalence, looking at the general population or focusing on adults aged 18years or older (seeTable2). All studies showed a reduction in smoking prevalence, but at the same time, an increase in illicit consumption from 16% in 2008 to 32% in 2013 was observed and continued to grow until 2016, when the estimated proportion of illicit consumption reached 42.8% al.(2015)analysed whether a price increase in tobacco products would encourage smokers to consume cheaper tobacco products in Uruguay, by switching their consumption to illicit tobacco products. The study reported that a 10% price increase would increase by 4% the probability of consuming roll-your-own cigarettes over more expen- sive manufactured legal cigarettes, suggesting that it is relevant to narrow different tobacco product prices in order to successfully re- duce overall consumption. The last point of the tobacco consumption analysis is related to the price and income elasticity of demand, whether the demand for tobacco products is elastic or inelastic and whether tobacco prod- ucts are normal and necessary goods. Data from five countries (Argentina, Colombia, Ecuador, Mexico, Peru) were identified and based on the evidence provided, both price and income were found to shape household or individual behaviour. Specifically, across all five countries, demand for tobacco products was found to be inelas- tic (with price elasticity of demand<1, indicating low responsive- ness to price changes; e. a 10% increase in the final price of tobacco products would result in a decrease in consumption by <10%). This could occur for various reasons, mainly related to con- sumer information on the new price, the level of addiction or lack of awareness of the risk related to tobacco products. In terms of the re- sponsiveness of the demand for tobacco products to a change in in- come, captured by the income elasticity of demand, the evidence from all five countries showed that with an increase in income, to- bacco consumption increased less than proportionally. The reported results confirmed that tobacco products are normal goods (income elasticity of demand being>0, with consumers raising consumption levels as their purchasing power increases) (Pindycket al., 2018); they were also found to be ‘necessities’ (income elasticity of demand being>0 but<1) (Table3).

Alcohol Only one study (Chave ́z, 2016) analysed alcohol consumption, and the effect of price elasticities of demand for tobacco and alcohol. The study reported a higher effect based on the price elasticity of de- mand for tobacco (0) compared with alcohol (0). The study also assessed the elasticity compared with Chilean total ex- penditure based on the quantity and quality of the goods, finding that the elasticity of alcohol consumption relating to total expend- iture was 0 (compared with 0 for tobacco consumption), mean- ing that the variation in the quantity of consumed alcohol was relatively inelastic compared with the tobacco when total expend- iture increased. If total expenditure declined, high-quality cigarettes and alcohol consumption would also decline, the latter being more sensitive to expenditure changes.

Effect on revenue generation

Tobacco Nearly all studies on revenue generation (9 out of 10) focused on revenues from tobacco taxation. Two studies approached this topic by considering multiple Latin American countries. One of them (Goodchildet al., 2017) examined the effect of tax increases on weighted average prices, revenue generation and volume. On aver- age, a 50% tobacco tax increase across the Latin American region would raise weighted average tobacco product prices by 28%, gen- erate US$7 million revenue (þ32%), and reduce the volume con- sumed by 7%; this trend would be traced in nearly all Latin American countries. The other study that considered the entire re- gion (Garce ́set al., 2014) did not analyse a potential implementa- tion but, rather, compared how Central American countries adapted to the FCTC directives. The analysis showed an overall gap that needed to be filled, due to the political and economic complexity of the area, and a lack in prioritization of research on legislation related to tobacco. Six studies analysed the revenue effect of tobacco taxation at country level. Two of these (Iglesias, 2016;Iglesiaset al., 2017) studied how the implementation of two alternative taxation systems (either an ad valorem, or a mix of specific and ad valorem) allowing manufacturers of tobacco products to choose from in the Brazilian context impacted fiscal revenue and, as a consequence, changed lev- els of illicit consumption. In the Brazilian tobacco tax reform to- bacco producers could choose between two regimes: a general regime, similar to the taxation system prevailing since 1999, where the ad valorem rate would have been 45% of the consumer price; and a special regime with a mix of specific and ad valorem rates. The latter has a lower ad valorem rate that could not be higher than 15%. Results were uniform in both studies: although revenue

Table 3Price elasticityand income elasticityof demand for cigarettes in select Latin American countries

Study name Publication year Study country Price elasticity of demanda Income elasticity of demandb

Chave ́z 2016 Ecuador 0 N/A Gonzalez-Rozadaet al. 2016 Peru 0 0. Iglesiaset al. 2017 Argentina 0 0. Jime ́nez-Ruizet al. 2008 Mexico 0 0. Maldonadoet al. 2016 Colombia 0 0. Martinezet al. 2013 Argentina 0 0.

aDemand is considered to be inelastic if for an increase in price by 1%, demand declines by<1%. In this case, the price elasticity of demand is negative and be-

tweenb 1 and 0. If the income elasticity of demand is>0, this indicates a normal good; if the income elasticity of demand is<1 a good is a ‘necessity’, i. where income rises by 1% but demand for that good rises by<1%; if the income elasticity of demand is>1, then the good in question is a luxury. Source:The authors from the literature.

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

randomized studies. Twenty-eight out of 34 studies reported at least a medium/unclear or high risk of bias in at least one of the seven dimensions we have considered (confounding; selection of partici- pants; intervention classification; deviation from intended interven- tion; missing data; outcome measurement; and selection of reported results). Most of the medium/high risk of bias were related to the outcome measurement (13 studies reported high risk, while 5 reported medium/unclear risk), followed by missing data (10 studies reported high risk, 2 reported medium/unclear) and deviation from intended intervention (9 studies reported high risk while 2 reported medium/unclear). Conversely, only 2 studies reported risk of con- founding bias (1 high risk and 1 medium/unclear), and 3 reported intervention classification bias (0 high risk and 3 medium/unclear). Results showed a relevant presence of moderate or high risk of bias

specifically in the missing data and the outcome of measurement domains. Missing data bias was primarily due to the lack of infor- mation on geographical coverage, production chain (manufacturer or retailer data), economic and social indicators. Bias in outcome measurements, due to self-reported data and underestimation of intervention and/or comparators, were often linked to a vague com- position of data. A more detailed description of risk of bias is avail- able inTable4(and more detailed information is provided in Supplementary Appendix Table SA1).

Discussion

This SLR identified and assessed the impact of sin taxes on goods that are considered to be harmful from a public health perspective in

Table 4Sin taxes in the Latin American context: summary of risk of bias assessment

Bias due to confoundingBias in selection of participants into the studyBias in classification of interventionsBias due to deviations from intended interventionsBias due to missing dataBias in measurement of outcomesBias in selection of the reported result

1 Alvarez-Sanchez et al.

(2018)

2 Bardach et al. (2016)

3 Batis et al. (2016)

4 Caro et al. (2018)

5 Chàvez (2016)

6 Claro, Popkin et al. (2012)

7 Colchero, Guerrero et al.

(2016)

8 Colchero, Popkin et al.

(2016)

9 Colchero, Rivera et al. (2017)

10 Curti et al. (2005)

11 Ferrante et al. (2007)

12 Garces et al. (2013) N/A N/A N/A

13 Gonzalez-Rozada et al.

(2016)

14 Goodchild et al. (2017)

15 Guerrero-Lopez et al. (2013)

16 Hernandez et al. (2019)

17 Iglesias (2016) N/A N/A N/A

(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Latin American countries from 2000 to 2018, by analysing the evi- dence based on three endpoints: effect on consumption, effect on revenue and health impact and is the first that is doing so in the Latin American region. Twenty-three out of 27 studies examining consumption effects confirmed that the application of a sin tax was inversely related to consumption levels. In the case of SSB tax in Mexico and its effect on consumption, this has been analysed by seven studies, six of them confirming the inverse relationship be- tween tax introduction and consumption levels. Evidence from 10 studies analysing the revenue endpoint is aligned in supporting ex- cise tax implementation or increase in the region to support add- itional revenue generation in a sustainable manner, providing, among others, case studies focused on Argentina, Brazil, Colombia, Mexico and Peru. Finally, five out of six studies focusing on the like- ly impact on health showed through a series of simulation models that potential sin tax implementation or increase would avert thou- sands of deaths, particularly from CVD and cancer, as well as lead to hundreds of thousands of additional LYG in a relatively short timeframe a summary of sin tax effect(s) or impact(s) and the extent of the effect(s) or impact(s) reported by

each study. None of the studies reported a negative effect or impact on any of the three endpoints. Results and conclusions on the association between sin tax im- plementation or increase and decrease in the consumption of harm- ful goods for public health, improved population health conditions or new sources of revenue in Latin America are aligned and compat- ible with findings from the literature in other geographical areas. An earlier systematic review (Wrightet al., 2017) with different criteria analysing 102 studies, focused on how consumption levels and rev- enue generation could be affected by public health taxes. This re- view did not focus on a specific geographical area, but the vast majority of the studies included came from HICs. Nevertheless, it confirmed the effectiveness of sin taxes as a tool for reducing harm- ful goods consumption, while revenue collection would be depend- ent on a variety of factors, e. the effectiveness of taxation in changing behaviour. Another recent systematic review (Redondo et al., 2018) has analysed results from 17 studies examining how the impact of taxes could shape SSB consumption. Likewise, the inverse relationship between SSB consumption and taxation levels was con- firmed. Our study reinforces all these findings particularly with

18 Iglesias, Skzlo et al. (2016)

19 James et al. (2019)

20 Jan et al. (2014)

21 Jiménez- Ruiz et al. (2008)

22 Kostova et al. (2014)

23 Maldonado et al. (2016)

24 Martinez et al. (2013) N/A N/A N/A

25 Nakamura et al. (2018)

26 Ng et al. (2018)

27 Ortega et al. (2017)

28 Reynales-Shigematsu et al.

(2015)

29 Rodriguez-Iglesias, Rios et

al. (2016)

N/A N/A N/A

30 Rodriguez-Iglesias, Schoj et

al. (2017)

31 Saenz-de-Miera, et al. (2010)

32 Sanchez- Romero et al.

(2016)

33 Szklo et al. (2018)

34 Taillie et al. (2017)

Source: The authors from the literature.

The table above summarizes the risk of bias level of each study. Green, yellow and red dots, respectively, indicate low, moderate/unclear and high risk of bias according to each domain. Source:The authors from the literature.

Table 4(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Table 5

(continued)

First author(s)

Otherendpoint(s)

Tax/indicatorof interest

Setting

Tax effect/impact (

1

/-)

a

Extent of the impact

Iglesias

et al.

(2016) Revenue

2012 cigarette ex-

cise tax

Brazil

Positive

Cigarette tax increase reduced smoking prevalence. Daily manufactured cigarette smoking prevalence rates

moved from 13% in 2008 to 10% in 2013. However, illicit manufactured consumption moved from16% to 31% in the same period.

Jime ́nez-Ruiz

et al.

(2008)

Revenue

VAT (Tobacco) Mexico

Positive

Total price elasticity is



0, meaning that a 10% in the cigarette price would lead to a



5% decrease in

average cigarette consumption. Higher prices would reduce household smoking largely in terms of smok-ing prevalence.

Maldonado

et al.

(2016)

No

Tobacco

Colombia

Positive

Tobacco demand is sensitive to price and income. Demand price elasticity is



0, while income elasticity

is 0. Study supports a higher taxation on tobacco to reduce consumption in the country.

Martinez

et al.

(2015)

No

Tobacco tax

Argentina

Positive

Long-term income elasticity was 0, while own-price elasticity was equal to



0, meaning that a 10%

increase in the growth of real income led to an increase of tobacco consumption of 4%, while a 10%price increase produced a fall of 3% in cigarette consumption. Short-term income elasticity was 0,while short-term own-price elasticity of cigarette demand was



  1. Study suggests how findings on elas-

ticity provide positive evidence on tobacco tax increases.

Nakamura

et al.

(2018)

No

SSBs

Chile

Mildly positive

After tax implementation, there was a significant decrease in the monthly purchased volume of higher-taxed

sugary soft drinks by 21%. Reduction in soft drink purchasing was higher in higher socioeconomicgroups and in higher pre-tax purchasers of SSBs. However, evaluation did not involve a randomized de-sign, therefore results cannot demonstrate a causal inference.

Ng

et al.

(2019)

No

SSBs

Mexico

Positive

After SSBs tax implementation, there has been a reduction in consumption of SSBs, particularly among high

SSBs purchasers, compared with the pre-tax period.

Ortega-Avila

et al.

(2018)

No

SSBs

Mexico

Neither positive

nor negative

Limited awareness among adolescents of the SSB tax implementation. Impact of the tax could be misper-

ceived by some segments of the population and that this would represent a limitation in changing citizens’attitudes towards these products.

Reynales-Shigematsu

et al.

(2015)

Health

Tobacco tax

Mexico

Positive

Tobacco consumption gradually decreased in the 2002–13 period as a result of policies implemented since

2002 (FCTC ratification in 2002, tobacco tax increases from 2007 to 2011).

Rodriguez-Iglesiaset al.

(2016)

No

Tobacco

Argentina

Positive

Cigarette tax increase that raises prices by 10% would lead to a decrease in smoking consumption of 3%.

Authors suggest pursuing this policy option as price of cigarettes in the country is one of the lowest in theworld.

Rodriguez-Iglesiaset al.

(2017)

Revenue

Tobacco

Argentina

Positive

Long-term price elasticity is



0, meaning that a 10% increase in real prices reduces cigarette consump-

tion by 2% per quarter. Long-term income elasticity is 0, meaning that a 10% increase in real in-come raises consumption by 4% per quarter. There is wide margin for excise tax implementation inthe country in order to effectively reduce consumption

Saenz-de-Mieraet al.

(2010)

No

SPST and VAT

Mexico

Mildly positive

2007 VAT tax increase, although indirectly, likely led to a general decrease in tobacco consumption.

Szklo

et al.

(2018)

No

Illicit tobacco

consumption

Brazil

Neither positive

nor negative

Illicit cigarette use increased in Brazil, both overall and across two socioeconomic groups of smokers who

did not stop smoking, after a new cigarette excise tax implementation. Illicit consumption needs to becarefully considered as a potential consequence of excise tax increase or implementation.

Taillie (2017)

No

SSBs

Mexico

Positive

Total volume of taxed products purchased declined by 4% in 2014 and by 14% in 2015, while the un-

taxed purchase changes were higher in 2014 (

þ

2%) but declined in 2015 (



4%). The study con-

firmed a trend of reduction in the consumption of energy-dense ultra-processed foods following taximplementation.

(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

Table 5

(continued)

First author(s)

Otherendpoint(s)

Tax/indicatorof interest

Setting

Tax effect/impact (

1

/-)

a

Extent of the impact

Revenue

Bardach

et al.

(2016)

Health

Tobacco

Peru

Positive

Model results report that, with a 50% price increase of tobacco products, Peru would collect 3 billion

of Peruvian Sol (approximately US$ 1 billion) in the 10 years following price increase.

Garce ́s

et al.

(2014)

No

Tobacco

Central America Positive

Excise taxes implementation or increase on tobacco products is encouraged in Latin America, as tobacco

taxes in the region are overall far lower than the levels recommended by FCTC.

Goodchild

et al.

(2017)

No

Tobacco excise

taxes

Latin America and

Caribbean

Positive

A 50% tobacco tax increase across the Latin American region would raise weighted average tobacco prod-

uct prices by 28%, generate US$ 7 million revenue (

þ32%), and reduce the volume consumed by 7%.

This trend would be traced in nearly all Latin American countries.

Iglesias (2016)

Consumption Tobacco tax

Brazil

Positive

2011 reform increased revenues by 20% compared with pre-tax year, despite legal sale contraction. The re-

form succeeded in revert the previous policies of low taxation on tobacco products, generating revenuesfor the country.

Iglesias

et al.

(2016) Consumption 2012 cigarette ex-

cise tax

Brazil

Positive

Raising excise taxes increased government revenues. Cigarette excise tax collection doubled between 2006

and 2013, from US$1 billion to US$2 billion. However, illicit tobacco trade increased.

Kostova

et al.

(2014)

No

Tobacco

Brazil, Mexico,

Uruguay

Neither positive

nor negative

The study suggests that a uniform high excise tax is more likely to reduce range of cigarette prices compared

with a tiered tax structure (i. where cheaper cigarettes are taxed at a lower rates than more expensivecigarettes) in each of the 15 considered countries (all LMICs). Levels of excise taxes are one the maincomponents of tobacco prices.

James

et al.

(2019)

Health

Tobacco tax

Colombia

Positive

Tax increase was legislated in December 2016, tripled the specific excise taxes and increased VAT by 3%,

leading to a 70% relative price increase in tobacco products. The study simulated that, after a full intro-duction of the new increases, in 20 years the net annual gains in tax revenue would be of COP$1 bil-lion (approximately US$364 m) compared with the pre-tax net annual gains (2016).

Jime ́nez-Ruiz

et al.

(2008)

Consumption VAT

Mexico

Positive

Study simulated that, holding other factors constant, a 10% price increase of tobacco products would yield

a rise in revenue by 15%, providing strong arguments for increasing cigarette taxes

Rodriguez-Iglesiaset al.

(2017)

Consumption Tobacco

Argentina

Positive

Study simulated that, even in a conservative scenario, despite the changes in real income and the final prices

of cigarettes even a 100% price increase in a low-revenue scenario would be beneficial for revenues andsustainable for the market.

Sa ́nchez-Romeroet al.

(2016)

Revenue

SSBs

Mexico

Positive

Simulation was based on two different scenarios, i. a 10% and a 20% reduction in SSB consumption, and

by taking into account any potential replacement for calorie compensation. Simulation results reportedthat, with a 10% reduction 983 million international dollars would have been saved in a time span of9 years, while a 20% reduction would have led to a saving of 1 billion international dollars.

Health

Bardach

et al.

(2016)

Revenue

Tobacco

Peru

Positive

Micro-simulation model estimated that a 50% price increase of tobacco products would lead in the next 10

years to almost 14 000 avoided deaths, 6210 avoided cardiovascular events and 5361 avoided new cancercases.

Ferrante

et al.

(2007)

Consumption Tobacco excise

tax

Argentina

Positive

Developed a simulation model to assess how tax increases in the retail price of tobacco would impact avoid-

able deaths. Two tax increase scenarios on the retail price were adopted: one at 75% (compared with tax-ation at 68% in 2007, leading to an overall 28% price increase) and one at 85% (with a final priceincrease of 113%). With a 75% increase, 1899 deaths per year would be avoided over a 20-year period(2004–24), and 2911 deaths would be prevented in the 2024–34 period. With an 85% increase, 7581deaths per year would be avoided until 2034.

(continued)

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

price through excise taxation is the most cost-effective single meas- ure in order to reduce the demand for tobacco and contribute smok- ing cessation improvement (WHO, 2003). These international guidelines are interfacing with a complex regional scenario, which is characterized by a particularly challenging epidemiological reality, significant levels of production of alcohol, tobacco and sugar and a timid political consensus over guidelines such as the ones by WHO in some countries. The consumption level of harmful goods, the related burden of disease and the difficulties in the tax structure reform in Mexico were a clear example of how consumer habits, state of health and state regulation can have significant impact on outcomes in the health of the population and in the long-term sustainability of the health system. At the same time, even with a partial reform com- pared with what FCTC recommended, evidence from Mexico showed how a wider approach that included taxation and an organ- ized set of other measures, could lead to a sensible improvement in all the endpoints considered. The role of data collection and research related to sin taxes and their impact represented another relevant point which emerged from our study. Funded studies included in our systematic literature re- view received grants only from government organizations [e. National Institutes of Health (USA), the Brazilian Ministry of Health], international organizations (e. the World Bank), non- Latin American non-governmental organizations (e. Bloomberg Philanthropies) or academic institutions (e. University of South Carolina). Of course, in many countries, general research may be conducted by manufacturing industries (Chave ́z, 2016;Iglesias, 2016 ) and, as such, could represent a source of bias as it represents corporate interests. Academic research leveraging country-level data appears to be limited as the only relevant sources are national sur- veys, often not including rural areas or relying on self-reporting methods. This creates a high risk of bias for researchers and policy- makers and has been already highlighted with regards to beverage industry statistics, which can be misleading and ‘fail to account for population or economic growth’ (Colcheroet al., 2017). The same was observed in the case of tobacco and how industry lobbying activities can strongly influence policy-making. Studies exist discus- sing how the tobacco industry concurred with the non- implementation of tobacco taxes in many parts of the world, despite the robust scientific evidence supporting their implementation (Jha and Chaloupka, 2000). The case of Argentina may represent the most telling example in Latin America of the relationship between the state and the tobacco industry that is weighted in favour of man- ufacturers’ aims. For these reasons, the implementation of sin taxes varies across settings based on the specific targeting of goods, the effective amount of tax, the choice between ‘per unit’ vs ‘ad valorem’ taxes, and the use of potential FS created beside the underlying broader rationale that justifies sin tax implementation. Research has emphasized that the specific country framework with regards to overall health state, socioeconomic composition, consumer habits, policy-making processes and orientations determines the most ef- fective pathway for a successful sin tax implementation in the case of SSBs (Brownellet al.,2009;Claroet al., 2012). That said, a spe- cific definition of what products are targeted is necessary to avoid side-effects in consumption, such as provoking the use of other similar harmful goods (i. goods of dubious quality that might not be captured by the tax reform, such as low-quality foods or beverages). The definition of the appropriate amount of tax is another controversial decision. Research on cardiovascular risk in young

adults (Duffeyet al., 2010) concluded that only high rates of tax- ation would produce a significant change in consumption; this is consistent with recommendations made by many of the studies in this study. The decision of adopting a ‘per unit’ vs ‘ad valorem’ tax is usual- ly at the forefront of the debate. A per-unit sin tax is easier and more flexible to implement from a government regulation perspec- tive than an ad valorem tax; generally, LMICs are encouraged to im- plement per-unit taxes because of limitations in law enforcement or administrative capacity. The disadvantages of this type of tax relate to the frequency and timing of revisions in order to ensure the tax remains effective. In fact, manufacturers can try to adjust the burden of the tax on some segments of the production process and, through that, reduce the price increase of the product to the consumer. At the same time, consumers can shift their consumption to lower- priced goods or to other similar products, keeping in mind, however, that for some products (e. tobacco and alcohol) substitution may be difficult. An ad valorem tax has its advantages because it easily adjusts to inflation changes, it is more visible and directly payable to the authorities. Notwithstanding the discussion on the relative merits of per unit or ad valorem taxes, the predicted revenue from the tax is uncertain and requires careful monitoring. An analysis of European Union countries has demonstrated that per-unit taxes have a better yield than ad valorem taxes on retail cigarette prices (Delipalla and O’Donnell, 1998;Goodchildet al., 2017). However, the overall preference of one tax over the other depends on specific country fea- tures and the specific objectives of policy-makers. Other studies (Wrightet al., 2017;Whiteheadet al., 2018) suggest that specific taxes would be more effective in reducing the consumption of cer- tain goods since an ad valorem tax could potentially shift consump- tion to cheaper and lower-quality goods, or induce manufacturers to reduce prices in order to maintain consumption levels. A final, broader, point of discussion relates to the implications of sin tax introduction on choice and on its regressive nature. Two studies (Brownellet al., 2009;Claroet al., 2012) claim that the benefits of sin taxes, especially on health outcomes, outweigh their dis-benefit on choice. With regards to their regressive nature, products subject to sin taxes, such as cigarettes, tend to cause greater harm to lower socioeconomic groups, and although the latter are impacted finan- cially more heavily than higher socioeconomic groups, the incentive to behavioural change is greater.

Study limitations

The results of this study reflect Latin American countries’ economic, political and epidemiological features and reality; therefore, the results may not be generalizable to other geographical regions. Furthermore, the analysis compared countries with significant dif- ferences in regulation, epidemiological frameworks and economic conditions, while many studies analysed policy changes in just a few countries. Additionally, most studies analysed sin tax impact within a relatively short space of time and lack long-term evidence. Despite all the above, the broader results of this study are consistent with most of the recent academic literature and underline many potential benefits of sin tax implementation in middle-income countries.

Conclusion

This study has confirmed the role of sin taxes in the Latin American context as a valid policy option for reducing consumption of harm- ful goods, generating additional revenue and potentially improving

Downloaded from academic.oup/heapol/article/36/5/790/6246142 by guest on 23 April 2022

health outcomes. The majority of studies reported that implementa- tion of sin taxes in Latin America resulted in reductions in harmful goods consumption, increases in revenue generation and a positive— albeit simulated—effect on health outcomes. The results on the risk of bias assessment and the analysis of the included studies suggested that future work on this topic would require more accurate data collection processes that go beyond weak study designs that may be susceptible to high risk of bias. This would require an increase in efforts to pro- mote research and address stakeholder interests. Apart from improv- ing data collection, a broader general effort is necessary in producing research on this topic; Latin American countries are gradually inves- t

Was this document helpful?

Czaa168 - Need to explore my self Analyzing the effectiveness of an intervention is a

Course: Science, Technology and Society (GEC103)

210 Documents
Students shared 210 documents in this course
Was this document helpful?
Sin taxes and their effect on consumption,
revenue generation and health improvement: a
systematic literature review in Latin America
Aurelio Miracolo
1
, Marisa Sophiea
2
, Mackenzie Mills
1
and
Panos Kanavos
1,
*
1
Department of Health Policy and Medical Technology Research Group - LSE Health, The London School of
Economics and Political Science, Cowdrey House, Houghton Street, London WC2A 2AE, UK
2
Faculty of Medicine, School of Public Health, Imperial College, London, Medical School Building, St Mary’s
Hospital, Norfolk Place, London W2 1PG, UK
*Corresponding author. Department of Health Policy and Medical Technology Research Group - LSE Health, The London
School of Economics and Political Science, London WC2A 2AE, UK.
E-mail: p.g.kanavos@lse.ac.uk
Accepted on 13 November 2020
Abstract
Sin or public health taxes are excise taxes imposed on the consumption of potentially harmful
goods for health [sugar-sweetened beverages (SSBs), tobacco, alcohol, among others], aiming to
reduce consumption, raise additional revenue and/or improve population health. This paper
assesses the extent to which sin taxes (a) can reduce consumption of potentially harmful goods,
(b) raise revenue for national health systems and (c) contribute to population health in Latin
America. A systematic literature review was conducted on peer-reviewed and grey literature; end-
points included: impact of raising sin taxes on consumption, ability to raise revenue for health and
the possibility of population health improvements. Risk of bias for each study was assessed. The
synthesis of the literature on sin tax implementation showed improvements in all three endpoints
across the study countries. Following the introduction of sin taxes or by simulating their potential
impact, nearly all studies explicitly reported that consumption of potentially harmful goods (mainly
SSBs and tobacco) declined; revenue was found to have increased in almost all countries, suggest-
ing that there may be additional scope for further tax increase. Simulated improvements in popula-
tion health have also been shown, by demonstrating a relationship between sin tax increases and
reduction in prevalence of diabetes, stroke, heart attacks and associated deaths. However, sin tax
effects on health would be better quantified over the long-term. Data quality and availability chal-
lenges did place some limitations on sin tax impact assessment. Sin taxes can be effective in reduc-
ing consumption of potentially harmful goods, improve population health and generate additional
revenue. Promoting further research on this topic should be a priority.
Keywords: Sin taxes, health financing, fiscal space, Latin America, earmarked taxation, universal health coverage
Introduction
Background
Sin taxes, or public health taxes, are defined by the World Health
Organisation as excise taxes targeting goods that can be detrimental to
the health of the population (WHO, 2004). These goods include to-
bacco products, alcohol, sugar-sweetened beverages (SSBs), which are
drinks with added sugar, such as soft drinks, tea, flavoured coffee,
juice and sports drinks. The harmful impact of these goods is well
known and is evidenced by research (Cnossen, 2005); for instance, to-
bacco consumption is linked to an increased risk of developing cardio-
vascular disease (CVD), respiratory disease, cancer and other non-
communicable and chronic diseases (U.S. Department of Health and
V
CThe Author(s) 2020. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
doi: 10.1093/heapol/czaa168
Review
Health Policy and Planning, 36, 2021,
AdvanceAccessPublicationDate:22April2021790–810
790
Downloaded from https://academic.oup.com/heapol/article/36/5/790/6246142 by guest on 23 April 2022